Tropical Truth(s)
Tropical Truth(s) The Epistemology of Metaphor and other Tropes
Edited by Armin Burkhardt and Brigitte Nerlich
De Gruyter
ISBN 978-3-11-023020-8 e-ISBN 978-3-11-023021-5 Bibliographische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliographie; detaillierte bibliographische Daten sind im Internet über http://dnb.d-nb.de abrufbar. © 2010 Walter de Gruyter GmbH & Co. KG, Berlin / New York Druck und Einband: Hubert & Co, Göttingen Gedruckt auf säurefestem Papier. Printed in Germany.
www.degruyter.com
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1. Metaphor, Simile and Truth Nicolaas T. Oosthuizen Mouton (Copenhagen, DK) Metaphor, empiricism and truth: A fresh look at seventeenth-century theories of figurative language . . . . . . 23 Pedro José Chamizo Domínguez (Málaga, E) / Brigitte Nerlich (Nottingham, GB) Metaphor and truth in Rationalism and Romanticism . . . . . . . . . . . . . . . . . . . . . . 51 Emmanuelle Danblon (Brussels, B) Persuasion: between trope and truth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 John A. Barnden / Alan M. Wallington (Birmingham, GB) Metaphor and its unparalleled meaning and truth . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Mark Lee (Birmingham, GB) Truth, metaphor and counterfactual meaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Carla Bazzanella / Lucia Morra (Torino, I) ‘Metaphorical’ truth conditions, context, and discourse . . . . . . . . . . . . . . . . . . . . 137 Masa-aki Yamanashi (Kyoto, J) Metaphorical modes of perception and scanning. A comparative study of Japanese and English . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Cornelia Zelinsky-Wibbelt (Hanover, D) Natural Language Processing: Minds, brains, and programmes . . . . . . . . . . . 177 Neal Norrick (Saarbrücken, D) Pear-shaped and pint-sized. Comparative compounds, similes and truth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Paul Georg Meyer (Aix-la-Chapelle, D) ‘Money is ruthlessly finding its own level’: Metaphor and metonymy in verb semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Contents
VI
2. Metonymy, Synecdoche and Truth Armin Burkhardt (Magdeburg, D) Between poetry and economy. Metonymy as a semantic principle . . . . . . . . 245 Antonio Barcelona (Córdoba, E) Metonymy in conceptualization, communication, language, and truth . . . . 271 Brigitte Nerlich (Nottingham, GB) Synecdoche: A trope, a whole trope, and nothing but a trope? . . . . . . . . . . . . 297
3. Other Tropes and Truth Wolfgang Braungart (Bielefeld, D) Eironia urbana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 Herbert L. Colston (Kenosha, Wisconsin/USA) Irony, analogy and truth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 Armin Burkhardt (Magdeburg, D) Euphemism and truth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355 Kenneth Holmqvist (Lund, S / Jarosław Płuciennik (Lodz, PL) Princes Antonomasia and the Truth: Two types of metonymic relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 Rita Brdar-Szabó (Budapest, H) / Mario Brdar (Osijek, HR) “Mummy, I love you like a thousand ladybirds”: Reflections on the emergence of hyperbolic effects and the truth of hyperboles . . . . . . . 383
Introduction The main inspiration to this book came, initially, from the well-known passage in Nietzsche’s “On truth and lie in an extra-moral sense”: What then is truth? A moveable host of metaphors, metonymies, and anthropomorphisms: in short, a sum of human relations which have been poetically and rhetorically intensified, transferred, and embellished, and which, after long usage, seem to a people to be fixed, canonical, and binding. Truths are illusions which we have forgotten are illusions; they are metaphors that have become worn out and have been dried of sensuous force, coins which have lost their embossing and are now considered as metal and no longer coins. (Nietzsche 1873/2006: 117)
If truth is based on metaphors (and other tropes) which have become standard expressions by canonization, and if metaphors (and other tropes) can be neither true (nor false), then our sentences too can be neither true (nor false); from which follows the nihilistic view that there is no truth at all. Such a radical view is a direct challenge to epistemology. For if it was correct, all scientific and philosophical striving would automatically lose its foundation. Considering the importance of this question, it is amazing that it has never been the topic of thorough philosophical investigation. It has, however, been touched upon time after time: The history of exploring connections between tropes and truth began in Antiquity and reached a peak in the 18th and 19th centuries when various scholars attempted to craft an overarching ‘philosophy of metaphor’ and also engaged in reflection on metaphor and truth (see Nerlich & Clarke 2001). These thinkers contributed to a philosophy of the metaphoric according to which metaphor was neither just a ‘figure of speech’, nor just a poetic fiction or decoration, but was regarded as underlying the structure and evolution of human thought and language. The whole movement of analysing metaphor in the context of ordinary life, language and thought, which began during the 17th and 18th centuries and particular with the work of Giambattista Vico, seems to have reached a first pan-European peak in the 1830s when linguists, philosophers, literary theorists and some rhetoricians alike proclaimed that figures of speech are, as Ortony (1975) was to say some 150 years later, not only nice, but necessary. For the English rhetorician and philosopher Benjamin Humphrey Smart, for example, they are “essential parts of the original structure of language; and however they may sometimes serve the purpose of falsehood, they are on most occasions, indispensable to the effective communication of truth. It is only by [these] expedients that mind can unfold itself to mind; – language is made up of them; there is no such thing as an express and direct image of thought.” (Smart 1831: 210) For him, as for many after him, tropes and figures of speech “are the original texture of language, and that from which whatever is now plain at first arose. All words are originally tropes; that is
2
Armin Burkhardt / Brigitte Nerlich
expressions turned [...] from their first purpose, and extended to others” (ibid.: 214). A similar view had been expressed by Jean Paul in Germany in 1804. He wrote in his Vorschule der Ästhetik: Der bildliche Witz kann entweder den Körper beseelen oder den Geist verkörpern. Ursprünglich, wo der Mensch noch mit der Welt auf Einem Stamme geimpfet blühte, war dieser Doppel-Tropus noch keiner; jener verglich nicht Unähnlichkeiten, sondern verkündete Gleichheit: die Metaphern waren, wie bei Kindern, nur abgedrungene Synonymen des Leibes und Geistes. Wie im Schreiben Bilderschrift früher war als Buchstabenschrift, so war im Sprechen die Metapher, insofern sie Verhältnisse und nicht Gegenstände bezeichnet, das frühere Wort, welches sich erst allmählich zum eigentlichen Ausdruck entfärben musste. (Quoted in Biese 1893: 12) The figurative witticism can either give a spiritual dimension to the body or a bodily one to the spirit. Originally, when humans were still at one with the world, this double trope did not yet exist; at that time humans did not compare dissimilar things but proclaimed similarities. As in writing hieroglyphs preceded alphabetical writing, so metaphor came first in speaking, insofar as it designated relations and not objects; and only gradually did metaphor loose its colour and became a ‘proper’ expression. (Translation BN)
This seems to have been the spark that ignited a whole bonfire of metaphor analyses in Germany and abroad. One can find allusions to the claim that the literal is but the pallid remnant of the figurative in Gustav Gerber, Friedrich Nietzsche, Alfred Biese, Gertrude Buck, and many more. Another of Jean Paul’s aphorisms became even more widely known and has to be quoted yet again in this context: “Daher ist jede Sprache in Rücksicht geistiger Beziehungen ein Wörterbuch erblasseter Metaphern. (With regard to mental relations every language is a dictionary of faded metaphors) (translation BN)” (Jean Paul 1962–1977[1804]: 184) Johannes Bauer echoes Jean Paul’s view when he writes, like many others: Die Ursprache ist noch unentfärbte Bildlichkeit. In dieser Zeit gibt es noch keine Prosa, weil jedes Wort schon durch seine Wurzel und seine Zusammensetzung einen poetischen Eindruck erregt, weil jede Anschauung schon ein Gedanke, jede Bezeichnung ein Versuch zu dichten ist. (Bauer 1878, I: 9) The original language is still one of un-discoloured images. At that time there is no prose, as each word through its root and its composition stimulates a poetic impression, as each perception is already thought and each designation an attempt at writing poetry. (Translation BN)
This was written in 1878 at the beginning of two decades of furious metaphorical research activity, another peak in a meta-metaphorical wave of activity which continued well into the beginning of the 20th century. Two of the most important thinkers of that era were Gustav Gerber and Friedrich Nietzsche, whom we already evoked at the beginning of this introduction (cf. also Embden 2005). Whereas Nietzsche is well-kown for this views on truth and metaphor, Gerber is less known, but should be.
Introduction
3
According to Gerber, it is not the case that words are just labels for concepts. And it is also wrong to say that the words which have become conventional through use have a ‘literal’ meaning and that new meanings are ‘figurative’. The conventional meanings only appear to be literal through habituation. They once were as ‘figurative’ (bildlich) as the new ‘figurative’ meanings. As in the case of the first use of a root word, new word meanings are only understood by reference to the context in which they are used, but also, in this case, by reference to our accumulated knowledge of the meaning relations they entertain with other words in the language (Gerber 1884: 104; Gerber 1871–74, II, 1: 21). Here the linguistic or symbolic frame or field supplements the situational one. ‘Meaning’ is always adjusted or adapted by reference to both. For Gerber, “the relativity of concepts and the pliability of meanings are tied to the living use of language” (cf. Cloeren 1988: 155).1 As we shall see, such a theory of the evolution and change of words and concepts was directly endorsed by Nietzsche in his philosophy of language. He claimed that what we regard as truth is nothing but a metaphor to which we have become accustomed: There is no ‘real’ expression and no real knowing apart from metaphor. [...] The most accustomed metaphors, the usual ones, now pass for truths and as standards for measuring the rarer ones. The only intrinsic difference here is the difference between custom and novelty, frequency and rarity. Knowing is nothing but working with the favorite metaphors, and imitating which is no longer felt to be an imitation. (Nietzsche 1872/1999: 50)
In this view, eternal truths and essences go out of the window, together with pre-established concepts. As Gerber wrote: “Nichts ist falscher, als anzunehmen, dass wir durch die Sprache die Dinge in der Welt bezeichnen.” (“Nothing is more wrong than to assume that we use language to designate things in the world”; translation BN) (Gerber 1871–74, I: 248) What is left is the insight that truth is conventionalised metaphor and that meaning is use, as Wittgenstein would later say. That Nietzsche read and appreciated Gerber and integrated many of his thoughts into his reflections on language, truth and knowledge has been variously noted (cf. Ungeheuer 1983, Meijers 1988, Meijers & Stingelin 1988, Gerhardt 1992, and Schumacher 1997). Schumacher (1997: 18) points out that Nietzsche refers to Gerber when he calls the outcome of the confrontation between man and world: “eine andeutende Übertragung, eine nachstammelnde Übersetzung in eine ganz fremde Sprache” (“a transliteration that manages to allude, a stuttering translation into a completely foreign language”; translation BN). For Nietzsche, as for Gerber (and Humboldt before them) language ————— 1
Cloeren compares this approach to Wittgenstein’s “theory of family resemblances and the overlapping of meanings, as well as his view of language as a form of life” (ibid.; cf. Gerber 1884: 161–162).
4
Armin Burkhardt / Brigitte Nerlich
mediates between individual and society, man and world, nature and freedom, mind and body. This is why Nietzsche liked Gerber’s schema of the evolution of language and elaborates it: What is a word? It is the copy in sound of a nerve stimulus. But the further inference from the nerve stimulus to a cause outside of us is already the result of a false an unjustifiable application of the principle of sufficient reason. If truth alone had been the deciding factor in the genesis of language, and if the standpoint of certainty had been decisive for designations, then how could we still dare to say “the stone is hard,” as if “hard” were something otherwise familiar to us, and not merely a totally subjective stimulation. We separate things according to gender, designating the tree as masculine and the plant as feminine. What arbitrary assignments! How far this oversteps the cannons of certainty! We speak of a “snake”: this designation touches only upon its ability to twist itself and could therefore also fit a worm. What arbitrary differentiation! What one-sided preferences, first for this, then for that property of a thing! The various languages placed side by side show that with words it is never a question of truth, never a question of adequate expression; otherwise, there would not be so many languages. The “thing in itself” (which is precisely what the pure truth, apart from any of its consequence, would be) is likewise something quite incomprehensible to the creator of language and something got in the least worth striving for. This creator only designates the relations of things to men, and for expressing these relations he lays hold of the boldest metaphors. To begin with, a nerve stimulus is transferred into an image: first metaphor. The image, in turn, is imitated in a sound: second metaphor. And each time there is a complete overleaping of one sphere, right into the middle of an entirely new and different one. (Nietzsche 1873/2006: 116)
In his book on language and cognition, entitled Kant and the Platypus (Eco 1999), Eco takes up this passage and elaborates it in turn. He shows how Nietzsche developed a new theory of concepts, according to which concepts are but metaphors gone stale. [...] we believe that we know something about the things themselves when we speak of trees, colors, snow, and flowers; and yet we possess nothing but metaphors for things – metaphors which correspond in no way to the original entities [...] Every word instantly becomes a concept precisely insofar as it is not supposed to serve as a reminder of the unique and entirely individual original experience to which it owes its origin; but rather, a word becomes a concept insofar as it simultaneously has to fit countless more or less similar cases – which means, purely and simply, cases which are never equal and thus altogether unequal. Every concept arises from the equation of unequal things. Just as it is certain that one leaf is never totally the same as another, so it is certain that the concept “leaf” is formed by arbitrarily discarding these individual differences and by forgetting the distinguishing aspects. This awakens the idea that, in addition to the leaves, there exists in truth the “leaf”: the original model according to which all the leaves were perhaps woven, sketched, measured, colored, cured, and painted – but by incompetent hands, so that no specimen has turned out to be a correct, trustworthy, and faithful likeness of the original model. (Nietzsche 1873/2006: 117; cf. Eco 1999: 44)
Introduction
5
How can one escape from this conceptual straight-jacket woven from dead metaphors? According to Eco, this is only possible through “a permanent poetic revolution” (ibid.: 46), because, as Nietzsche noted, art “continually confuses the conceptual categories and cells by bringing forward new transferences, metaphors and metonymies. It continually manifests an ardent desire to refashion the world which presents itself to waking man, so that it will be as colorful, irregular, lacking in results and coherence, charming, and eternally new as the world of dreams.” (Nietzsche 1873/2006: 121) Another big name which cannot be left unmentioned in this context is that of Fritz Mauthner. In his Beiträge zu einer Kritik der Sprache which, besides Hugo von Hofmannsthal’s well-known letter of Lord Chandos, is one of the milestones of modern language skepticism, he not only insists that language is language use (“Sprache ist Sprachgebrauch” (1923 [1901], I: 24), but also holds that, because of its abstractness, it is a main means of misunderstanding (“ein Hauptmittel des Nicht-Verstehens”). According to him the use of language with its “zerfahrene Worte” (“scatterbrained words”; translation BN) can never lead to complete mutual comprehension: “Durch die Sprache haben es sich die Menschen für immer unmöglich gemacht, einander kennen zu lernen.” (“Humans have forever given up the possibility of understanding each other when they invented language”; translation BN) (Ibid.: 56) Language is not fit for communication and understanding. In later sections of his book he shows that one of the main reasons for this is that language has been infected by metaphor from its origin: Jedes einzelne Wort ist geschwängert von seiner eigenen Geschichte, jedes einzelne Wort trägt in sich eine endlose Entwicklung von Metapher zu Metapher. Wer das Wort gebraucht, der könnte vor lauter Fülle der Gesichte gar nicht zum Sprechen kommen, wenn ihm nur ein geringer Teil dieser metaphorischen Sprachentwicklung gegenwärtig wäre; ist sie ihm aber wieder nicht gegenwärtig, so gebraucht er jedes einzelne Wort doch nur nach seinem konventionellen Tageswerte, als Spielmarke, und gibt mit diesen Spielmarken nur einen imaginären Wert, gibt niemals Anschauung. (Ibid.: 115) Each single word is pregnant with its own history, every single word carries with it an endless chain of development from metaphor to metaphor. Whoever uses the word could, given this fullness of visions, never even start to speak if they were aware of this metaphorical development; however, not being aware of this metaphorical history, they use a word only in its current conventional value; they use it as a conventional coin to convey some imaginary value that never gets even close to the real intuition that they want to convey. (Translation BN)
For Mauthner, as for Nietzsche and Jean Paul, whom he explicitly quotes in this context (ibid., II: 456), metaphor is the principle of language, and the process of metaphor creation and semantic fading has always been the principle of its lexical growth: Die zwei oder die hundert “Bedeutungen” eines Wortes oder Begriffes sind ebenso viele Metaphern oder Bilder, und da wir heute durchaus von keinem Worte eine Urbedeutung kennen, da die erste Etymologie unendliche Jahre hinter unserer
6
Armin Burkhardt / Brigitte Nerlich Kenntnis von ihr zurückliegt, so hat kein Wort jemals andere als metaphorische Bedeutungen. (Ibid., II: 451) The two or the hundred ‘meanings’ of a word or a concept are just as many metaphors or images, and as we never really know the original meaning of a word nowadays because its etymology is forever shrouded in the mists of time and therefore not knowable, so every word can never have anything but metaphorical meanings. (Translation BN)
As a consequence, human reason, thinking and language are “metaphorical through and through” (“durch und durch metaphorisch”, ibid., II: 463). Language, therefore, “is” metaphor (“Sprache ist Metapher”, ibid., II: 453), and as metaphors do not reach up to the extralingual reality, language cannot do so either: it must “als Erkenntnismittel [...] stets unfruchtbar bleiben [...], immer nur bereit, das Wirkliche gesellig zu beschwatzen” (“as epistemic tool [...] be forever useless [...], remaining for ever just a way of gossiping about real life”; translation BN) (ibid., II: 454). And as Mauthner takes metaphor in a very broad “Aristotelean” sense, according to which the term also covers metonymy, synecdoche, hyperbole and even irony (cf. ibid., II: 459–60), the same seems to apply to all other linguistic phenomena that are called tropes according to traditional rhetoric. 19th-century and early 20th-century reflections on metaphor and truth, on conceptualisation and perception, and on the literal-figurative distinction have some counterparts in the modern philosophy of metaphor. We shall concentrate here on three points of overlap: (1) the rejection of the comparison theory of metaphor; (2) a new conception of truth and knowledge; and (3) the rejection of the literal-figurative dichotomy. As Lakoff and Johnson wrote in Metaphors We Live By, their view that “metaphors can create similarities [,] runs counter to the classical and still most widely held theory of metaphor, namely the comparison theory” (Lakoff & Johnson 1980: 153). In their view, metaphor does not only passively exploit pre-existing similarities via comparison, it can actively create new realities. This too goes against the traditional view of metaphor, especially the view that metaphor is merely a linguistic or poetic device. In this new view metaphor is seen instead as “a means of structuring our conceptual system and the kinds of everyday activities we perform” (ibid.: 145). Finally, similarities are no longer regarded as objectively given, but as subjectively constructed, as based on the experience of people (ibid.: 154). This view goes against the grain of any objectivist theory cognition. Lakoff and Johnson conceded, that this ‘new’ metaphorical theory of truth and knowledge is not so new, but they did not appreciate quite how old it actually is. It should be obvious from this description that there is nothing radically new in our account of truth. It includes some of the central insights of the phenomenological tradition, such as the rejection of epistemological foundationalism, the stress on the centrality of the body in the structuring of our experience, and the importance of that structure in understanding. Our view also accords with some of the key elements of Wittgenstein’s later philosophy: the family-resemblance account of cate-
Introduction
7
gorization [...], and the emphasis on meaning as relative to context and to one’s own conceptual system. (Ibid.: 182)
Let us now briefly look at a book on figurative language written by some of the most eminent scholars in this field, Figurative Language and Thought (Turner, Cacciari, Gibbs & Katz 1998). In his chapter, devoted to the “literal versus figurative dichotomy” (see Turner 2005), Turner points out that in his previous work he offered demonstrations that the commonsense dichotomy between “literal” and “figurative” is a psychological illusion. There is no doubt that some products of thought and language feel literal while others feel figurative. We have reactions, and they are motivated, but these motivations do not come from fundamental differences of cognitive operations. “Literal” and “figurative” are labels that serve as efficient short-hand announcements of our integrated reactions to the products of thought and language; they do not refer to fundamentally different cognitive operations. He argues that we judge a mental or verbal connection to be literal or figurative depending on “the degree to which the conceptual connection or the linguistic expression is generatively entrenched. The greatest degree of generative entrenchment for a conceptual connection occurs when it becomes established as a central part of basic category structure: for example, a woman is a human being.” Here we are dealing with ‘metaphors’ which, according to Nietzsche (and Jean Paul, Gerber, and Mauthner), have been reduced to “schemata and concepts. Thence a pyramidal order of castes and ranks, laws and delimitations, constructed entirely by language, an immense ‘Roman columbarium,’ the graveyard of intuition.” (Nietzsche 1873/2006: 118; quoted by Eco 1999: 45, see above) Like Nietzsche and Gerber, Turner assumes that conceptual entrenchment is graded, with category structures that are so entrenched that they seem ‘literal’ at one end and others that appear to be ‘figurative’ at the other, and certain well-worn, but not quite dead metaphors in between. As Nietzsche said: “The only intrinsic difference here is the difference between custom and novelty, frequency and rarity. Knowing is nothing but working with the favourite metaphors, and imitating which is no longer felt to be an imitation” (1872/1999: 51). th At the end of the 19 century, just as at the beginning and the end of the th 20 century, it was therefore argued that we can only understand how our conceptual and semantic knowledge is structured and how it changes with use if we accept that “metaphor is not, as we have been taught, an isolated phenomenon, a ‘freak’ in literature, more or less inexplicable, an arbitrary ‘device’ of the writer, but a genuine expression of the normal process of thought at a certain stage in its development, consonant with the ordinary laws of psychology and interwoven with all our common experiences.” (Buck 1971 [1898]: 69, italics added) This is only a brief overview of some high points in the history of the insight into the “metaphorical basis of all thinking” (“metaphorische Grundlage
8
Armin Burkhardt / Brigitte Nerlich
alles Denkens”, Mauthner 1923, II: 462) which led to the search for tropical truth. Some other high points will be discussed in the first, historical, chapters of this book. The outline shows that, with the exception of Nietzsche and Mauthner, when the question of tropical truth was raised at all, it has always been asked with regard to metaphor alone. Even if one does not over-extend the concept of metaphor, like Mauthner, it has to be feared, however, that the same problem of their truth function raises its head with regard to all the other tropes as well, i.e. for metonymy, synecdoche, irony, euphemism, hyperbole and others (which cannot be dealt with in this book). The main idea of this volume, therefore, is to define and examine the different tropes from the perspective of different approaches and disciplines and to propose, from different points of view, answers to questions pertaining to the relation between tropes and truth, thereby broadening the scope of traditional research investigating tropes and truth separately and focusing, for the most part, on metaphor. The book therefore wants to problematise not only the relation between metaphor/metonymies and truth, as envisaged by Nietzsche, but between tropes and truth in general and thus, perhaps, earmark the problem of ‘tropical truth’ as a problem for linguistics and epistemology. If tropes, esp. metaphors, which one might, from one point of view describe as abuses of the semantic and pragmatic rules of language, are at the basis of our cognition, what does this mean for the possibility of our sentences to be true or false and, therefore, for human knowledge in general? Is our truth intrinsically metaphorical (as proposed by Burkhardt [1987: 51]), based on cultural and anthropological foundations, or even tropical? If metaphor and other tropes cannot be true or false but only felicitous or infelicitous this would also provokes the question whether they should be treated as constatives or rather as performatives. However, all tropes seem to contribute solely to the propositions they are part of and, therefore, must be taken as forms of “propositional indirectness” (cf. Burkhardt 1986: 390–5) which can only affect the illocution of the respective utterance very indirectly. By no means does this exclude Jacob L. Mey’s view according to which over and above the question “[w]hat does a particular metaphor express, and how, there is […] another question that needs to be asked: How felicitous is a particular metaphor in a particular context (e.g., solving a problem, obtaining consensus, elucidating a difficult subject matter, and so on)?” (Mey 2001: 62) For “felicitous” here is but another word for ‘adequate’ or ‘effective’ and not intended in the “illocutionary” sense it had within the original framework of speech act theory. A related “pragmatic” account which must be mentioned here is that of James Bono who discusses the role of metaphor in science. He writes: Metaphors are neither true nor false, but rather are simply more or less productive or even counter-productive. To ask whether a metaphoric description is “true” or “false” risks losing sight of the work done by a particular metaphor and misconstruing the very nature of metaphor itself. The very question presupposes that we can – with enough hard work and perseverance – substitute a neutral, “literal” de-
Introduction
9
scription for a metaphoric one. While metaphoric usage may provoke attempts to specify the exact ways in which, for example, “nature” is a “book,” “DNA” is a “code,” “disease” is “warfare,” or the “brain” is a “computer,” such attempts neither exhaust the range of possible implications contained in a metaphoric description, nor provide a sufficient basis for judging a metaphor true or false. To claim that one particular specification of a metaphor is true or false is not to judge the metaphor itself true or false. Thus, while it might be possible to conclude that a top-down, rules-based, algorithmic model of the brain – as in early AI modeling of human intelligence – is seriously flawed, inadequate, or even false, such a conclusion cannot be extended to the metaphor, “the brain is a computer,” itself. Rather, metaphors operate precisely to provoke repeated attempts to imagine, and enact, how a given phenomenon (“nature,” “DNA,” “disease,” or the “brain,” to name only our examples) is like something else. In a sense, then, we may say that the success of a metaphor rests simply in its ability to generate more or less productive leads, strategies, and models for engaging, representing, acting upon, and thus “understanding” features of the world in which we happen to become interested. At a given moment and in a given scientific context, a specific metaphor may, in fact, prove productive, unproductive, or even counter-productive. All metaphors, however, are subject to (narrative) redescription within different or changing contexts over time: consequently, metaphors themselves evolve, produce shifting meanings, and perform different kinds of work to access and map certain aspects of nature. (Bono 2004)
In such a conception, as in Lakoff and Johnson’s, metaphor (and maybe other tropes as well) is taken to be just a cognitive tool the human mind has developed to enlarge the scope of its understanding. However if it is “productive” in suggesting that and showing how “a given phenomenon [...] is like something else” it must not only be adequate to the relevant purpose but also correspond to the facts in a way. Otherwise one would have to say that metaphor (and other tropes) is one of the tools of human perception which would necessarily imply that truth, in the sense of an independent yardstick, is but a chimera. In the face of such a messy situation it should not be surprising that the question of what truth itself is must be left open here. The discussion of this problem would fill at least another volume. For the purposes of this book it may suffice to define truth as the central epistemic feature of a proposition, describing a state of affairs, which is proved or generally taken to be the case, or, to put it in Wittgenstein’s words, truth is “the existence of states of affairs” (1961: 2.). Admittedly, this definition does not answer the crucial question as to how truth may be proved or verified and how we come to believe a piece of information to be true, be it by provable correspondence with the facts, by common consent or by coherence with the prevailing system of knowledge. However, it helps to understand that truth, as a feature of propositions, is what we reasonably consider to be the case. And instead of seeking for an objective truth which is dependent on (scientific) verification one might adopt William James’ “pragmatic” account of truth according to which true ideas “point to or lead towards” a corresponding reality and “must yield satisfaction as their result” (1975: 104). And satisfactory, in James’s sense, are, e.g., beliefs “in
10
Armin Burkhardt / Brigitte Nerlich
other men’s minds, in independent physical realities, in past events, in eternal logical relations” which posit the assumed reality, or the feeling of hope, the cessation of doubt, and consistency between a certain idea and the “entire rest of our mental equipment” (ibid.: 105). Truth, in this view, is also a matter of common conceptualization which is part of the “inherited background on which I distinguish between true and false” (“der überkommene Hintergrund, auf welchem ich zwischen wahr und falsch unterscheide”; Wittgenstein 1970: § 94). “Tropical” sentences might turn out to be at best true in the sense that they satisfy our cognitive and communicative needs. More detailed answers, however, depend on the diversity of tropes and the approaches used to study them and are, therefore, left to the ingenuity of our authors. The book is divided into three sections. The first one is dedicated to the queen of tropes, i.e. metaphor, the second to its sisters, metonymy and synecdoche, and the third one deals with other noble members of the trope family. The chapter on metaphor is opened by a critical outline of the history of its theoretical conceptions. Nicolaas T. Oosthuizen Mouton’s (Roskilde, DK) lively chapter brings deeper historical reflection to the study of metaphor. Nerlich and Clarke (2001) had highlighted similarities between the ‘modern’ programme of metaphor research initiated by George Lakoff and Mark Johnson and 19th-century philosophical reflections on metaphor. Mouton goes back further and studies in great detail the debate about metaphor that took place in the 17th century. He objects to the claim made in recent years by prominent scholars of metaphor (see Lakoff & Johnson 1999) that classical empiricists rejected metaphor per se and that they did so because no niche could be found for metaphor in a philosophy dominated by a correspondence theory of truth. He also refutes Johnson’s saying (1981: 11) that ‘scientific-minded philosophers’ like Locke were simply giving metaphors “one beating after another”. This was not the case, at least not if one is thereby to understand that they were against all and every use of metaphor. Through a close reading of the original texts, Mouton comes to the conclusion that 17th-century philosophers did not reject metaphor per se, but that they were rather objecting to what they perceived as widespread abuses of metaphor, especially by the alchemists. The chapter should be required reading for anybody interested in the history of metaphor analysis and the philosophy of metaphor. Pedro José Chamizo Domínguez (Málaga, E) and Brigitte Nerlich (Nottingham, GB) also contribute to deepening the historical reflection on metaphor and truth by going back even further and debunk another philosophical preconception about metaphor. It is widely thought that in contrast with both empiricist and rationalist philosophers, romantic literary critics and philosophers, inspired in part by Giambattista Vico and Jean Jacques Rousseau, rediscovered metaphor (together with the power of imagination and creativity), but “tended to portray poetry, and metaphor along with it, as essentially different from everyday language: both poetry and metaphor came to be seen as expressing an ‘emotive’, rather than a ‘cognitive’ meaning” (Leezenberg
Introduction
11
2001: 1). Here again one has to be careful, as the ‘emotive’/‘cognitive’ distinction was already well established in empiricist and rationalist philosophical discourse and as the view that metaphor was pervasive in everyday life and thought was discussed from César Chesneau Du Marsais, at the end of the 18th century, up to Friedrich Nietzsche and beyond. In their chapter, Chamizo Domínguez and Nerlich discuss rationalist philosophers’ views on metaphor and truth and explore closely the distinctions made by some philosophers between the emotive and the cognitive, the figurative and the literal, the female and the male, the subjective and the objective, all linked to two fundamental dichotomies: clarity vs. obscurity and truth vs. obfuscation. In a second part they provide an overview of the discussion of metaphor from Vico and Rousseau up to Nietzsche who tried to reduce truth to metaphor. In her chapter Emmanuelle Danblon (Brussels, B) links the past, namely, Greek insights into the philosophy and rhetoric of persuasion, where issues of truth and persuasion are intricately mixed, to modern epistemology, where both are dissociated. She sets herself the task of reconnecting epistemology with rhetoric, metaphor with truth, the literal with the figurative, via mimesis and fiction. She especially studies the importance of the ‘as if’. As Kurt Vaihinger said in his Philosophy of as if: “All cognition is the perception of one thing through another” (Vaihinger 1924 [1870]: 29). Truth, then, emerges from an interaction between the two. The next chapter, written by John A. Barnden and Alan M. Wallington (Birmingham, GB) provides evidence to support two main arguments, that are independently supported (and may independently succeed or fail) but that are also linked. First, the authors argue that not all metaphorical sentences, clauses or other potentially proposition-bearing linguistic units should be regarded as individually having a meaning in terms of the target, as opposed to conspiring with other constituents to lead to a target meaning (a metaphorical meaning). Thus, the concentration on the meaning of individual metaphorical sentences (or sometimes clauses) in the technical literature on metaphor semantics, even when the consideration takes into account contextual effects, is too limiting. Rather, metaphorical meaning is in general a matter of chunks of discourse, of no definite size or shape. But then, they claim, it is only consistent to expect that meaning in general, not just metaphorical meaning, should be regarded as attached to discourse chunks, not individual sentences, except that we have assumed that we can assign a lexical meaning to each sentence. None of this militates against the semantic/pragmatic importance of any sentence, clause, etc. that does not get its own (non-lexical) meaning: it can still be crucial in building the overall meaning of a discourse chunk. Overall they argue for a loose and distant connection between linguistic entities on the one hand and meanings and truth values on the other. Indeed in the case of truth their arguments lead to the suggestion that talk of truth of lexical entities such as sentences is only an approximation in general, adopted for reasons of practical convenience, although often the approximation can be a good one.
12
Armin Burkhardt / Brigitte Nerlich
Mark Lee (Birmingham, GB), like Mouton and Chamizo Domínguez and Nerlich, sets out to overthrow an old philosophical dogma, in this case the classical account of counterfactual reasoning based on the truth-conditions of the counterfactual expression. He argues that such accounts are misguided and ignore the figurative aspect of counterfactuals. He shows that rather than being truth-conditional, counterfactual meaning is essentially analogical. An example of a counterfactual is: “If everybody had left the conference early then I’d be talking to an empty room”. The discussion of such counterfactuals or the counterfactual reasoning that structures their understanding sheds light on metaphorical reasoning, which has also been explored in the previous chapter by Barnden and Wallington. Understanding how metaphorical counterfactuals, such as “If John had put that idea to one side, then he would have been happy”, ‘work’ is therefore of particular interest in this context. Lee explains that there is an important distinction between metaphorical and counterfactual reasoning. Metaphorical reasoning often involves treating a statement which is literally false as true and then drawing conclusions about its implications. Counterfactual reasoning involves understanding the connotation of a trivially true statement through one which has as it components two false statements and then drawing a conclusion from its verification. In both cases, the concept of truth is contextually dependent on the particular conceptual space within which reasoning proceeds. Therefore, truth is not an objective or context free notion but instead one sensitive to the particular demands of the space it is in. Proceeding from some ideas proposed in Burkhardt (1990) and following some basic assumptions of Austin’s theory of speech acts, particularly his doctrine of “infelicities”, Carla Bazzanella and Lucia Morra (Torino, I) develop a pragmatic account of metaphor and the question of its possible truth. In their view metaphor is a purposefully deviant use of language which (at least in the case of “creative” metaphors) leads to a context-dependent interpretive process creating provisional and, therefore, undefined meanings. The authors analyse the interaction between “global” and “local” contexts in the process of metaphor interpretation and propose different conditions for its success. Rejecting e.g. Searle’s “substitutionalist” account of metaphor, which implies the idea that in order to establish the truth conditions of a metaphorical utterance another sentence explaining the speaker’s intended meaning should be constructed, they come to the conclusion that the concept of truth conditions does not sensibly apply to metaphor and must be replaced by that of appropriateness conditions. In Masa-aki Yamanashi‘s (Kyoto, J) contribution the question of tropical truth is treated from a completely different perspective, namely from the point of view of contrastive cognitive semantics. The author compares English and Japanese with regard to metaphors of visual perception. His comparative analysis of Japanese sentences and their English equivalents leads to the interesting conclusion that, while both languages, by conceptualizing the visual field in terms of a container or of tactile perception, make use of the same
Introduction
13
conceptual metaphors, still differ a great deal in detail. Whereas it is common in English to conceptualize the visual field as a container directly, the term me (‘eye(s)’) in Japanese metonymically represents a visual field which in turn is conceptualized as a container. While in English metaphors of the type SEEING IS TOUCHING the term eye can be the subject of the respective metaphorical utterance, in Japanese it can only occur as the grammatical object. And while metaphors in which visual perception is construed in terms of visual sense impressions directly reaching the eyes are quite common in Japanese, many metaphors of this kind would be unacceptable in English. “Duality” phenomena, such as the conceptualization of perception as either the movement of the perceiver’s eye to the target or vice versa, are frequent in Japanese, but not in English which prefers a conceptualization of vision in which the target of perception moves to the perceiver. Using Lakoff’s Invariance Principle the author also gives good reasons why visual perception may be conceptualized in terms of tactile perception in both directions and why auditory and olfactory perception cannot. On the basis of this analysis of the intercultural differences in metaphorical conceptualization (and its grammatical consequences) Yamanashi comes to the conclusion that there is no objective truth, but that truth itself, as a cognitive process is deeply rooted in our body’s interactions with the world, and therefore intrinsically metaphorical. In her chapter Cornelia Zelinsky-Wibbelt (Hanover, D) focuses on the computer metaphor which is now almost commonplace in cognitive psychology and examines its legitimacy as a model for explaining the human mind. As such a metaphor implies the existence of an analogy between brain/mind and the computer with its hard- and software, the crucial question is whether it is both possible and necessary for the computer to know the ultimate nature of the meanings which humans associate with linguistic utterances. The author points out that speakers grasp sentence meanings mainly by top-down inferences which are based on the overall discourse context and that, therefore, hybrid natural language processing systems composed of different subsystems are needed to provide adequate simulations of linguistic utterances. By focusing on the relevant bits of encyclopedic information, human intelligence is capable of speeding up understanding in an economical way. Moreover, the computational model of the mind does not cope with human linguistic creativity and flexibility with regard to grammar, word formation, and semantic and pragmatic interpretation, e.g. of metonymic expressions and conversational implicatures. In contrast to the human mind, the representations of artificial systems consist exclusively of system-internal relations which are purely syntactic, because their ability of recursive reasoning lacks the human ability of meta-representation, which is self-referential. The author points out from a principled philosophical as well as from an empirical linguistic perspective that both the dualistic and the monistic view of the brain/mind relation are reductionist. By showing that the application of computer metaphors to the brain/mind relation is often linguistically unwarranted. Zelinsky-
14
Armin Burkhardt / Brigitte Nerlich
Wibbelt also demonstrates that metaphorical utterances cannot be true in any strict sense. Neal R. Norrick (Saarbrücken, D) examines an interesting borderline-case between simile, metaphor and hyperbole, namely comparative compounds with suffixal elements like shaped, sized, colored, scented as their second component. On the basis of material from well-known computerized corpora he points out that compounds like pint-sized or pear-shaped, depending on the context, may have a literal and a tropical reading which is based on stereotypes (and prototypes). The latter is hyperbolic if, as in man-sized, “appropriateness for” is expressed, or metaphorical if, as in sparrow-sized, the position of the “object” on one scale is compared to that of the (prototypical) “vehicle” on another. Though never construed as literal truth claims, everyday utterances containing comparative compounds in their literal sense may be true or false. With their “appropriate for” readings hyperbolic uses may also be taken to have a truth-value which, despite of the apparent lack of precision, will be commonly accepted in communicative practice. Comparative compounds with scalar meanings, however, are in a way impressionistic, have a potential for ambiguity, and with regard to their truth-value, therefore, must be regarded as context dependent. An examination of botanic text books shows that scientific literature cannot do without using a variety of comparative compounds, but even the preference of terms like chordate instead of heart-shaped cannot hide the fact “that our highly vaunted concept of scientific truth ultimately comes down to subjective human judgments of similarity founded on basiclevel concepts and culturally salient prototypes”. The main topic of Paul Georg Meyer’s (Aachen, D) contribution is the emergence of polysemous meaning structures in transitive verbs by metaphorical and metonymic processes. After having shown that event, agent, and object are the ingredients of any transitive situation, an analysis of the different contextual meanings of open makes evident that the readings of the object and processes expressed by the verb are often affected by abstract or even metonymic agent meanings. A minute comparative examination of the different uses of find and stumble leads to the insight that while the achievement of finding is always cognitive, its object may be physical or cognitive. Stumbling, however, can be physical or cognitive and, instead of the physical one, may also have an abstract object. The metaphorical meanings of both verbs seem to have arisen from the same kind of meaning shift from physical to cognitive, but, by focusing either on the process of finding or on the object found, still show disjunct sets of truth conditions. The last section explains the metaphorical sentence Money finds its own level as the result of a complex interaction of metaphors and metonymies concerning the event, agent and object positions. Armin Burkhardt (Magdeburg, D) takes us from metaphor to metonymy which, with synecdoche, is the topic of the second section. In his chapter “Between poetry and economy. Metonymy as a semantic principle” he argues that ancient rhetoric did not separate the tropes clearly enough from one an-
Introduction
15
other. This is particularly the case for metonymy (e.g. glass for ‘content of the glass’) and synecdoche (e.g. Fritz or the German for ‘the Germans’). On the basis of a frame-theoretical model of explanation, it is shown that the former must be defined as the use of a word designating an object, person or state of affairs to refer to their parts or aspects and vice versa, while the latter must be limited to the set-subset relation. Analyses of examples prove that literary metonymy is a creative shift within the same frame. In everyday language, however, the economic purpose of focusing and abbreviating comes more to the fore. Using the framework of historical semantics and historical contrastive analysis, the distinction between a designation-oriented (“onomasiological”) subtype and a word-form-oriented (“semasiological”) subtype of metonymy is made and fruitfully applied. The author shows that metonymy is not only a rhetorical trope, but a substantial cognitive principle of language and other sign systems. And as long as its referential function is successful, i.e. as long as the interlocutors know what the metonymic expressions are an elliptical shorthand for, there is no reason to think that sentences containing them could not be true or false. Antonio Barcelona’s (Córdoba, E) chapter provides an overview of the role of metonymy in thought, communication, language and truth, from a cognitive linguistic perspective and explains the use of the term metonymy in this tradition in which it is linked to ‘idealised cognitive models’ on the one hand and ‘pragmatics’ on the other. The bulk of the chapter is devoted to reviewing the numerous areas, linguistic and nonlinguistic, in which metonymy can be claimed to play a major role. The relation between metonymy and truth is complex, but Barcelona points out that metonymy, as a mechanism for conceptual activation, is in principle unaffected by truth value judgments. But since metonymy is one of the major factors in conceptualization, it may determine in part the cognitive context (e.g. by means of a metonymic model) on the basis of which a given sentence will be judged as true or false; on the other hand, metonymy is a major factor guiding pragmatic inferences, which may be confirmed or cancelled on the basis of later information. Brigitte Nerlich’s (Nottingham, GB) chapter guides the reader through a brief history of the concept of synecdoche. The chapter has two aims: To summarise the past and present fate of synecdoche, and to make French linguists aware of the cognitive, and cognitive linguists outside France aware of the French rhetorical tradition in which this figure of speech is discussed. The chapter also tries to elucidate the relation between synecdoche and truth, which, as in the case of metonymy, is rather difficult to establish. The chapter charts the rises and falls of synecdoche in the ‘old rhetoric’, in historical semantics, in Roman Jakobson’s work, in the ‘new rhetoric’ as developed in France, and finally in the framework of cognitive semantics, where she especially refers to the work carried out by Ken-ichi Seto who has made the most sustained effort so far at studying synecdoche from a cognitive linguistic perspective.
16
Armin Burkhardt / Brigitte Nerlich
Wolfgang Braungart’s (Bielefeld, D) article “Urban irony” opens the last chapter of the book which is dedicated to other tropes. Irony is a trope which is common in everyday life as well as in the literature reflecting it. Taking up a relevant quote from Friedrich Schlegel, the author points out that irony presupposes sociability and that, therefore, the concepts of irony and urbanity are intrinsically connected. This idea can be traced back to the philosophy and literary theory of Romanticism and further to the writings Cicero and Quintilian. As a consequence of increasing urbanization, however, and in combination with the Enlightenment conception of reason as basically social, conversational and processual, irony has become the genuine poetic feature of modernity. Braungart also shows that irony is the most “performative” of all tropes, as its interpretation is extremely context-dependent (and, thus, dominated by the rhetorical category of aptum). It is described as a highly civilised and “witty” way of telling an allegedly controversial truth without running the risk of direct confrontation. Always on the verge of lie, irony is “simulated insincerity” and, therefore, can neither be expressed “properly”, nor can it be considered, in any strict sense, as either true or false. It may only be appropriate or not and will always be true and false at one and the same time. The relation between irony and truth has always intrigued scholars. Herbert Colston’s (Wisconsin-Parkside, USA) chapter discusses this relation in great depth. His chapter provides an in-depth study of the intricacies of the relationship between tropes and truth in general and irony and truth in particular, exploring not only the relation between tropes and truth though but also between tropes, truth an the emotive and social aspects of communication. The chapter then focuses in particular on the relationship between truth and a particular type of “tropical” language – ironic analogies, such as “Calling Chilies just another steakhouse is like saying the Great Wall of China is just a fence”. Their use and understanding are linked to the negotiation of social status between interlocutors. Overall, Colston claims that if there is any viability to the philosophical notions of a search for meaning, a cooperative principle, relevance and others, then the simple existence of tropical language must support a connection between tropes and truth. Unless one would argue that all tropes are outright attempts to create falsity, or to unnecessarily place layers of obstacles around a core sense of communicated truth, then some such relationship(s) must hold. Armin Burkhardt (Magdeburg, D), in his second contribution to this volume, investigates the working of euphemism. After a short outline of the history of the concept and its definitions, a “veiling” and a “concealing” subfunction of euphemisms are distinguished. While the former subfunction is caused by religious or social taboos and may have a moral background, the latter must be taken as means of deliberate hiding. In a next step the author shows that euphemisms may either concern the whole sentence or just a single word or expression. Important subtypes of euphemisms of the “syntactic” kind are referential vagueness or the use of the passive voice without mentioning an agent. “Lexical euphemisms”, by contrast, may be either “abstrac-
Introduction
17
ting” or (particularly when based on metaphors and metonymies) “positivising”. Referring to Hermann Paul, euphemisms are then also presented as forms of semantic change which, under the influence of social taboos, may even end in a chain of substitutions of one euphemism by another which was termed the “euphemism treadmill” by Stephen Pinker. After having distinguished between meiosis, which is an indirect speech act, and euphemism, which is not, Burkhardt shows that with regard to their truth or falsity abstracting and positivising euphemisms must be judged in different ways: while the former are only less concrete or informative than their non-euphemistic equivalents, the latter counterfactually add a positive feature to a deplorable propositional content and, therefore, must always partly be false. Kenneth Holmqvist (Lund, SE) and Jaros!aw P!uciennik (Lodz, PL) explore a rather neglected trope, antonomasia, and its relation to truth. They link their ‘story’ of antonomasia in a creative way to Cervantes’ story of Princess Antonomasia in his well-known novel Don Quixote. Antonomasia is classically defined as is the substitution of any epithet or phrase with a proper name and vice versa. Holmqvist and P!uciennik refine this definition in various ways. They distinguish between two types of antonomasia. The first kind is similar to periphrasis, where a proper name is substituted by an epithet. This kind of antonomasia was often used in ancient times when talking about mythical and epical characters. They call this metonymical antonomasia. They call the second type metaphorical antonomasia because here a different mechanism is at work. A proper name is used in order to describe a very different kind of reference. This type of antonomasia highlights analogical features shared by the paragon (a paragon, in Lakoff’s terms, is a specific example that comes close to embodying the qualities of the ideal) and the instance of the type. The authors link their study of antonomasia to a new way of looking at culture. They argue that antonomastic names function as a resource for the en-culturation of newly encountered phenomena. If we refer to Leonard Cohen as the Lord Byron of rock music, we treat a popular singer as a famous romantic poet elevating him and popular songs to a higher level of culture. Antonomasia is, they say, essentially humanistic and anti-computational. It highlights the anti-computational nature of language and shows that language might instead be understood as a kind of game of make-believe or as a simulation tool. A computer will not find all the references to the word “king”, if we call him “His Majesty”. Both kinds of antonomasia discussed in this chapter require a special kind of historical and contextual knowledge to be understood or indeed ‘deciphered’. To find the ‘true’ meaning of an antonomasiologically encrypted reference we have to go beyond the linguistically given and engage in the culture in which it is used. Rita Brdar-Szabó (Budapest, H) and Mario Brdar (Osijek, HR), in their chapter “‘Mummy, I love you like a thousand ladybirds’: Reflections on the emergence of hyperbolic effects and the truth of hyperboles”, define, typologise and analyse examples of hyperbole from a cognitive linguistic point of view, or more specifically from the point of view of mental space theory, as
18
Armin Burkhardt / Brigitte Nerlich
developed by Fauconnier and Turner, and Michael Israel’s work on semantic scalarity or scalar reasoning. Their chapter tries provide an answer to two related questions concerning hyperbole: Can hyperbole be regarded as a trope or a figure? And: Can hyperbolic expressions be considered as just another case of flouting one or more of Gricean maxims (notably the maxims of quality and quantity) in rhetoric? The answers to these questions are necessarily very complex, but can be summarised as follows: pure hyperbole is hardly possible as such, as hyperbolic effects are a basically contextually induced and pragmatically embedded phenomenon. Hyperbolic effects can be identified as such when taking into account the expectations of the recipient and the communicative aims of the producer of the utterance, which also means that the problem of the truth of hyperboles should be approached from the perspective of the basic interpretation possibilities of a text as a whole. The work on this volume has taken a lot of time and was interrupted very (indeed: too) often by other (mostly official) duties. It would take another book to tell the story of the genesis of this one (which was, in a way, initiated by the discussions of a seminar (Oberseminar) held by Neal Norrick at Braunschweig University in 1993). Therefore, we do not only have to thank our authors for their contributions but also for their patience. But there is a wellknown German saying Gut Ding will Weile haben (‘it takes time to do a thing well’). And in this sense we are quite confident that the result will justify the time it has taken to achieve it. Málaga and Nottingham, March 2009
References Bauer, Johannes: Das Bild in der Sprache [The Image in Language]. 2 vols. Ansbach 1878; 1889. Biese, Alfred: “Das Metaphorische in der dichterischen Phantasie” [The Metaphorical in the Poetic Imagination], in: Zeitschrift für vergleichende Litteraturgeschichte NF 2,1889, 318–39. – Die Philosophie des Metaphorischen [The Philosophy of the Metaphoric]. Hamburg and Leipzig: Leopold Voss 1893.
Introduction
19
Bono, James: “Between Semiotics and Geometry: Metaphor, Science, and the Interdisciplinary ‘Trading Zone’.” MLA Annual Conference, Philadelphia, December 28, 2004, Session Performing Science: Metaphor, Material Practices, Invention, and Exchange(s). Manuscript. Buck, Gertrude: The Metaphor: A Study in the Psychology of Rhetoric. Folcroft, PA: Folcroft Library Editions 1971[1898]. Burkhardt, Armin: Soziale Akte, Sprechakte und Textillokutionen. A. Reinachs Rechtsphilosophie und die moderne Linguistik. Tübingen: Max Niemeyer 1986. – “Wie die ‘wahre Welt’ endlich zur Metapher wurde. Zur Konstitution, Leistung und Typologie der Metapher”, in: Conceptus 21, 1987, 39–67. – “Searle on metaphor”, in: Burkhardt, Armin (ed.): Speech Acts, Meaning and Intentions. Critical Approaches to the Philosophy of J.R. Searle. Berlin–New York: W. de Gruyter 1990, 303–35. Cloeren, H. J.: Language and Thought: German Approaches to Analytic Philosophy in the Eighteenth and Nineteenth Centuries. Berlin–New York: Walter de Gruyter 1988. Eco, Umberto: Kant and the Platypus: Essays on Language and Cognition. London: Secker & Warburg 1999. Embden, Christian C.: Nietzsche on Language, Consciousness and the Body. Chicago etc.: University of Illinois Press 2005. Gerber, Gustav: Die Sprache als Kunst [Language as Art]. 2 vols. Bromberg: Mittler’sche Buchhandlung 1871–74. – Die Sprache als Kunst [Language as Art]. First vol. (Second enlarged edition). Berlin: Gaertner 1885[1871]. – Die Sprache und das Erkennen [Language and Cognition]. Berlin: Gaertner 1884. James, William: The Meaning of Truth. Cambridge, Mass.–London: Harvard University Press 1975. Jean Paul [Johann Paul Friedrich Richter]: Vorschule der Ästhetik. [Introduction to Aesthetics]. In Werke [Works], edited by Norbert Miller. Abt. 1, Vol. 5. Darmstadt: Wissenschaftliche Buchgesellschaft 1962–1977[1804], 7–330 Lakoff, George, Mark Johnson: Metaphors We Live By. Chicago: University of Chicago Press 1980. Mauthner, Fritz: Beiträge zu einer Kritik der Sprache [Contributions to a Critique of Language]. 3 vols. Leipzig: Felix Meiner 1923[1901/02], 3rd ed. Mey, Jacob L.: Pragmatica. An Introduction. Oxford: Blackwell 2001, 2nd edition. Nerlich, Brigitte, David D. Clarke: “Mind, meaning, and metaphor: The philosophy and psychology of metaphor in nineteenth-century Germany”, in: History of the Human Sciences 14:2 (2001), 39–61. Nietzsche, Friedrich: “On Truth and Lies in a Nonmoral Sense (1873)”, in: Pearson, Keith Ansell, Duncan Large (eds.): The Nietzsche Reader. Oxford: Blackwell, 2006, 114–23. – Nietzsche, F. (1872/1999). “The philosopher: reflections on the struggle between art and knowledge”, in: Breazeale, D. (ed. and translator): Philosophy and truth: selections from Nietzsche’s notebooks of the early 1870’s. Amherst, NY: Humanity Books, 3–58.
20
Armin Burkhardt / Brigitte Nerlich
Ortony, Anthony: “Why Metaphors are Necessary and not Just Nice”, in: Educational Theory 25/1, 1975, 45–53. Smart, Benjamin Humphrey: An Outline of Sematology: Or an essay towards establishing a new theory of grammar, logic, and rhetoric. London: Richardson 1831. (Published anonymously.) Turner, Mark: “The Literal Versus Figurative Dichotomy”, in: Coulson, Seana, Barbara Lewandowska-Tomaszczyk (eds.): The Literal and Nonliteral in Language and Thought. Frankfurt/Main–Berlin–Bern–New York–Paris–Wien: Peter Lang 2005, 25–52. – , Cristine Cacciari, Raymond W. Gibbs, Albert N. Katz: Figurative Language and Thought. Oxford: Oxford University Press 1998. Wittgenstein, Ludwig: Tractatus Logico-Philosophicus. Translated by D. F. Pears and B. F. McGuinness. London: Routledge & Kegan Paul 1961. – Über Gewißheit. Ed. by G.E.M. Anscombe and G.H. von Wright. Frankfurt/Main: Suhrkamp 1970.
1. Metaphor, Simile and Truth
Nicolaas T. Oosthuizen Mouton (Copenhagen, DK)
Metaphor, empiricism and truth: A fresh look at seventeenth-century theories of figurative language
0. Introduction 1. Modern portrayals of classical empiricist conceptions of metaphor 2. Classical empiricists on the necessity of metaphor 3. Empiricist critiques of the alchemists’ abuses of metaphor 4. So what is the legacy of 17th-century debates about metaphor, then? 5. Conclusion “In the post-medieval development of empiricist and rationalist systems, it is mistrust, rather than appreciation, that dominates philosophical accounts of metaphor. During the rise of empiricist epistemologies in the sixteenth and seventeenth centuries, metaphors suffered one beating after another at the hands of ‘scientific-minded’ philosophers. It is especially important to understand the typical empiricist stand on metaphor, for it is essentially the same view as that held today by the inheritors of the empiricist legacy.” Mark Johnson (1981: 11)
0. Introduction In recent decades, the classic empiricist stand on the role of metaphor in “discourses that pretend to inform and instruct” – to borrow John Locke’s felicitous phrase – has attracted a considerable amount of attention. Almost without exception, the empiricist legacy is portrayed as ill-founded and indefensible.1 Thus Ted Cohen informs us that “there has been a very strong line in Western ————— 1
This is not the case among intellectual historians who happen to be interested in empiricist attitudes toward metaphor (e.g. Clark 1998, Vogt 1993, Vickers 1984). My generalization rather pertains to scholars working within the mainstream of contemporary research on metaphor. (Exceptions to this rule include e.g. Jäkel 1997, Musolff 2005.)
24
Nicolaas T. Oosthuizen Mouton
philosophy, especially in that strain running from British empiricism through Vienna positivism, which has denied to metaphors and their study any philosophical seriousness of the first order” (Cohen 1978: 3). Having cited a few de-contextualized remarks, Cohen concludes that “although these remarks of Hobbes and Locke may seem remote, their import has prevailed until quite recently” (ibid.: 5). Mark Johnson concurs, and adds a touch of hyperbole: “During the rise of empiricist epistemologies in the sixteenth and seventeenth centuries, metaphors suffered one beating after another at the hands of ‘scientific-minded’ philosophers. It is especially important to understand the typical empiricist stand on metaphor, for it is essentially the same view as that held today by the inheritors of the empiricist legacy” (Johnson 1981: 11). In similar vein, Lakoff and Johnson (1980: 190–191) inform us that the empiricists treated metaphor with “contempt”, and they did so because of an irrational “fear” of metaphor.2 And Mary Hesse (1987: 76) insists that modern empiricism subscribes to a “literalist dogma”, according to which “all informative language is univocal, that is, literal”, and adds that one of the main sources of this mainstream empiricist dogma can be traced “back to a particular form of the correspondence theory of truth that accompanied the rise of modern science in the seventeenth-century”. We are then advised to dismiss both the dogma and the assumptions that purportedly gave rise to it. Undoubtedly, the legacy of empiricism contains lots of ideas that we may as well donate to needy members of poverty-stricken traditions, or commit to the flames. It would be prudent, however, to go about the task of sorting through our inheritance without haste, lest we end up ditching items that are still valuable. In the case at hand, I agree that it is important to understand the classic empiricist stand on metaphor, and I agree that its legacy is still alive. But I do not think that the standard portrayal of the original empiricist stance is accurate: philosophers like Locke had considerably more sophisticated views of metaphor than their contemporary critics give them credit for. And since the initial position is misconstrued, its legacy is also misunderstood. Moreover, I do not think we are particularly well served by the kind of heavyhanded rhetoric that Mark Johnson resorts to when he claims that metaphor “suffered one beating after another at the hands of ‘scientific-minded’ philosophers”. A metaphor cannot “suffer”, and the metaphorical suggestion that they can suffer at best serves to disguise condemnation as description. As the late Clifford Geertz advised, “there are better things to do with even a defective inheritance than trash it” (Geertz 2000: 18). My main aim is not to defend this defective inheritance, however, but simply to understand it, and to show how classical conceptions of how metaphors ought to be used shaped the ways in which they were – and still are – —————
2
For good measure, Lakoff and Johnson add that the empiricists also feared emotion, imagination, and subjectivism (ibid).
Metaphor, empiricism and truth
25
actually used. In particular, I hope to show that a better understanding of the
empiricist legacy may help us to make better sense of a remarkable feature of the way in which modern scientists use metaphors: they tend to spell out, in considerable detail, the differences between the source- and target domains of the metaphors they happen to be using. If I say that this is remarkable, it is because it is surely not common to do so in most other contexts. It would be rather unusual to encounter a politician, for example, who felt obliged to draw attention to all the ways in which markets are unlike invisible hands, or policy initiatives unlike bridges to the future, or nations unlike organisms, or whatever her favorite metaphors happen to be. Yet in scientific contexts, it is quite common to do just that. For example, modern social scientists make liberal use of the notion that collective entities – states and societies, cities and corporations – are biological organisms.3 But typically, they barely finish introducing the idea before they rush to point out various ways in which collective entities are not like organisms at all. Thus Herbert Spencer extended and elaborated the metaphor to an astounding degree, but he also made considerable efforts to spell out “the leading differences between societies and individual organisms” (Spencer 1996: 273, emphasis added). Or, to give a more recent example, consider the following remarks by Michael Hannan and John Freeman, the two founding fathers of modern ‘Organizational Ecology’: Organizational ecology and evolution are more complicated than comparable processes in bio-ecology for several reasons. We have already mentioned the fact that forms of organization are not coded in inert genetic material. Individual organizations can and sometimes do change their forms. In addition, information about building structure does not pass from parent to offspring. There is often no clearcut parallel to a parent. Moreover, there is no reason why individual organizations cannot live forever. This means that an organization can contribute to future generations directly, by persisting. (Hannan & Freeman 1989: 143, emphasis added)
It turns out that this is a recurrent feature rather than an isolated incidence. Already in 1977, when Hannan and Freeman first presented their perspective on the “population ecology” of organizations, they dedicated a whole section ————— 3
Already in antiquity, of course, it was fairly common to think of collective entities as living organisms, and to talk of their parts – patricians and plebeians, soldiers and statesmen – as interdependent organs. Thus Plato argued that the well-ordered city resembles a human individual, insofar as a wound to any of its parts makes the whole city suffer (Republic, 462c–d), while Aristotle postulated that in states and organisms alike, disproportionate growth of any part of the body causes the organism to perish or to “change into the form of some other animal” (Politics, 1302b– 1303a). But neither Plato nor Aristotle made any effort to spell out the ways in which collective entities are unlike organisms. This is by no means the only differences between ancient and modern versions of the social organism metaphor, but a more detailed discussion is beyond the scope of this paper.
26
Nicolaas T. Oosthuizen Mouton
to the theme of “Discontinuities in Ecological Analysis”, in which they spelled out a string of differences between organizations and organisms (Hannan & Freeman 1977: 936–8). Subsequent contributors to the tradition expanded that list considerably. This remarkable practice raises interesting questions: Whence the ideal that scientists ought to spell out the differences between the source- and the target domains of the metaphors they use? And did this ideal influence the ways in which they actually use metaphors? I will argue that the ideal is part of our empiricist legacy, and that it did have a practical impact. I will build my case as follows. In the first part, I briefly review the scant evidence which modern critics of classical empiricist conceptions of metaphor cite in support of their case. In the second part, I present counter-evidence indicating that many empiricists held metaphor, rightly used, to be ubiquitous, unavoidable, and useful. But this does not explain away the fact that they did, after all, also make some rather critical remarks about metaphor. In the third part, I argue that the empiricists did not object to the use of metaphor per se, but rather to what they perceived as widespread abuses of metaphor, especially by the alchemists. In the final part, I examine the impact of such debates about how metaphors ought to be used on the ways in which they actually were used, and try to spell out the implications for any attempt to formulate a theory about the role of metaphor in science and other “discourses that pretend to inform or instruct”. 1. Modern portrayals of classical empiricist conceptions of metaphor It is not difficult to find evidence that could support a case like Mark Johnson’s. For if calling metaphors “perfect cheats” or “powerful instruments of error and deceit” count as instances of metaphors “suffering a beating”, then metaphors did indeed suffer the odd beating at the hands of seventeenth-century philosophers. Johnson (1980: 11–3) presents five pieces of evidence. Three stem from the writings of Thomas Hobbes.4 (For present purposes, let us ignore the fact that Hobbes was hardly an empiricist.) Johnson opens his case by citing Hobbes’ assertion that one of the ways “in which men abuse speech occurs when they […] use words metaphorically; that is, in other sense than that they are ordained for; and thereby deceive others”. This is followed by references to Hobbes’ likening of metaphors and other “senseless and ambiguous words” to “ignes fatui”. Reasoning upon them, warned Hobbes, leaves us “wandering amongst innumerable absurdities”. Johnson’s third ex—————
4
The instances of Hobbes’ complaints about metaphor cited by Johnson can be found in Leviathan, part 1, chapters 4 and 5. See also De Cive, 12, 12. An excellent discussion of Hobbes’ account of metaphor can be found in Musolff (2005).
Metaphor, empiricism and truth
27
hibit consists of yet another remark of Hobbes’, in which he condemned “the
use of metaphors, tropes, and other rhetorical figures, instead of words proper. For though it be lawful to say, for example in common speech, the way goeth, or leadeth hither or thither; the proverb says this or that, whereas ways cannot go, nor proverbs speak; yet in reckoning, and seeking of truth, such speeches are not to be admitted”. Johnson then cites Bishop Berkeley’s sermon that “a philosopher should abstain from metaphor”. Finally, there is John Locke’s eloquent critique of eloquence: But yet if we would speak of things as they are, we must allow that all the art of rhetoric, besides order and clearness; all the artificial and figurative application of words eloquence hath invented, are for nothing else but to insinuate wrong ideas, move the passions, and thereby mislead the judgment; and so indeed are perfect cheats: and therefore, however laudable or allowable oratory may render them in harangues and popular addresses, they are certainly, in all discourses that pretend to inform or instruct, wholly to be avoided; and where truth and knowledge are concerned, cannot but be thought a great fault, either of the language or person that makes use of them. What and how various they are, will be superfluous here to take notice; the books of rhetoric which abound in the world, will instruct those who want to be informed: only I cannot but observe how little the preservation and improvement of truth and knowledge is the care and concern of mankind; since the arts of fallacy are endowed and preferred. It is evident how much men love to deceive and be deceived, since rhetoric, that powerful instrument of error and deceit, has its established professors, is publicly taught, and has always been had in great reputation: and I doubt not but it will be thought great boldness, if not brutality, in me to have said thus much against it. Eloquence, like the fair sex, has too prevailing beauties in it to suffer itself ever to be spoken against. And it is in vain to find fault with those arts of deceiving, wherein men find pleasure to be deceived. (Essay, 3.10.34)
Clearly, this is first and foremost a critique of the art of rhetoric, rather than of metaphor. One may add that it is a rather rhetorical critique of rhetoric, and that Locke, who once taught rhetoric at Oxford (Milton 1994: 32), surely must have had his tongue firmly in his cheek when he wrote that passage. In short, I agree with what Paul de Man had to say about this particular passage: “There is little epistemological risk in a flowery, witty passage about wit like this one, except that it may be taken too seriously by dull-witted subsequent readers” (De Man 1978: 16). Seen thus, the proper response to the passage is to follow the example set by Gottfried Wilhelm Leibniz, in his New Essays, namely to answer irony with irony by congratulating Locke with the eloquence of his critique of eloquence: “you seem to be fighting eloquence with its own weapons, having at your command an eloquence which is superior to the deceptive kind you are attacking” (Leibniz 1996: 3.10.34). For the sake of the argument, however, let us suppose that Johnson’s interpretation is the correct one, and that we have just witnessed metaphor “suffer a beating” at the hands of Locke. Even if we accept Johnson’s interpretation of that piece of evidence, I do not think his case is particularly strong: he overlooked a con-
28
Nicolaas T. Oosthuizen Mouton
siderable amount of counter-evidence that does not fit very well with the idea that seventeenth-century empiricists simply rejected metaphor per se. But let us first briefly look at Mary Hesse’s argument that it was a dogma of empiricism that all scientific language is purely literal, and that this idea “goes back to a particular form of the correspondence theory of truth that accompanied the rise of modern science in the seventeenth-century” (Hesse 1987: 76). The gist of that theory, according to Hesse, was that “scientific language should be like a mirror of nature, which has a unique name for each unique essence, and its grammatically correct sentences will correspond to the laws of nature connecting these essences” (ibid). Hesse does not spell out why this view should leave no room for metaphor, but it is easy enough to fill in the missing pieces of the argument. For if scientific language is supposed to have a unique name for each unique essence, and if metaphor involves “giving the thing a name that belongs to something else”, then metaphor is incompatible with scientific language for two reasons. Firstly, to use a metaphor would amount, from this perspective, to giving two names to a unique essence (its proper name and the imported name). Secondly, by using a metaphor, a single name ends up being attached to two essences (insofar as the name now refer both to the source- and the target domain of the metaphor, to use modern parlance). Either way, it undermines the commitment to a language in which there is one unique name for each unique essence. According to Hesse, an obsession with scientific language, thus understood, was a feature that tied together Bacon, Galileo, Leibniz, the Royal Society – indeed, “seventeenth-century natural philosophers”, tout court. Let us assume, for the moment, that this was indeed the case, and take a look at what Umberto Eco (1995: 238) called “the most complete project for a universal and artificial philosophical language that the seventeenth-century was ever to produce”, namely John Wilkins’ Essay towards a Real Character, and a Philosophical Language (1641). Now, it is common to claim that “in his Essay, Wilkins soberly marshaled the case against metaphor” (Ormsby-Lennon 1988: 332). Yet if Wilkins thought that the perfect language should be free of tropes, why then did he try to design mechanisms that would enable him to incorporate metaphor, metonymy, and synecdoche in the universal language? (He tried to do so by means of particles that, when attached to a term, would indicate that it ought to be taken in a figurative sense.) Admittedly, his attempt to account for tropes was not very successful. Indeed, as Eco (1995: 247–8) points out, many of the problems with Wilkins’ overall project resulted from, or were made more visible by, his attempt to incorporate tropes. The important point here is not that his project failed, however, but rather that its aim was patently not to get rid of tropes altogether. Wilkins aimed to enable users to immediately identify tropes as tropes; to avoid mistaking metaphors for literal statements of identity. But that is an altogether different matter than trying to banish them. Hence, even if we assume that all seventeenth-century philosophers were indeed committed to the idea of a universal language, it would not
Metaphor, empiricism and truth
29
amount to a demonstration that they thereby automatically rejected tropes per
se. The second problem with Hesse’s account is that it is simply not true that all seventeenth-century philosophers were committed to the ideal of a universal language. Aarsleff (1982: 263) observes that few members of the Royal Society actually had much faith in universal language schemes such as Wilkins’. In this regard, they were heavily influenced by Locke, who insisted that no man “can pretend to attempt the perfect Reforming the Languages of the world, no not so much as that of his own Country, without rendring himself ridiculous” (Essay, 3.11.2). Indeed, we may take Locke’s views on the words-ideas-things triangle as a rather good counter-example to Hesse’s claim that seventeenth-century philosophers held that scientific language should have a unique name for each unique essence. After all, Locke mentions the supposition that words stand for ‘real essences’ as a common abuse of language, and insists that it is “preposterous and absurd” to think that our “names stand for ideas we have not, or (which is all one) essences that we know not, it being in effect to make our words the signs of nothing” (Essay, 3.10.21). When it comes to the “real essences of things”, Locke argued, we are like the “gazing Country-man” who saw the great clock in Strasbourg and had to remain satisfied with observing its outward appearances, rather than like the expert who knows “all the Springs and Wheels and other contrivances within” (Essay, 3.6.3).5 On Locke’s view, words serve as “knots” with which we “tie” various simple ideas, representing the outward qualities of things, into arbitrary “bundles” (Essay, 3.5.10) – arbitrary in the sense that there are usually numerous other combinations that could have been tied together. To more than these convenient products of “the workmanship of the understanding”, Locke insisted, we cannot aspire when it comes to knowledge of the essence of things. In short, no language can ever be constructed that would contain only one unique name for each unique essence, because we do not have epistemic access to the essence of things. Hence, even a perfectly literal scientific language would be as incapable of signifying the “real essences” of things as metaphorical lan————— 5
I do not intend to get embroiled in the long-standing controversies about the details of Locke’s account of “real essence”. Suffice it to note here that he seems to shift between making ontological claims to the effect that there are not “a certain number of forms or molds, wherein all natural things that exist are cast” (Essay, 3.3.17); epistemological claims holding that we do not and cannot know the “real essence” of natural things; factual claims that we do not in fact classify things according to their presumed ‘real essences’ (see e.g. Essay, 3.6.20); and a ‘constructivist’ thesis to the effect that in the case of what Locke calls ‘mixed modes’, real and nominal essence are in fact identical, for the simple reason that they only have a nominal essence – i.e. they are only names holding together complex abstract ideas that exist only in the mind (see e.g. Essay, 3.5.1).
30
Nicolaas T. Oosthuizen Mouton
guage. Whatever the reasons underlying Locke’s critique of metaphor may have been, it could not have been the reasons Hesse stated. Let us turn to Hesse’s favorite witness when it comes to the Royal Society’s views on metaphor, namely “the Society’s official historian Thomas Sprat, [who] wrote that its aim was to confine its writing and talking to a ‘close, naked, and natural way of speaking’” (Hesse, ibid). Here, it is worth noting that the passage does not explicitly mention tropes at all: it is Hesse’s assumption that a “naked” and “natural” way of speaking is one free of metaphor. Actually, Sprat argued that while it is tempting to condemn rhetoric as such, one should keep in mind “that it is a Weapon, which may be as easily procur’d by bad men, as good”, and that it would therefore be rather shortsighted of ‘good men’ to give their opponents a monopoly on that weapon, thereby leaving themselves completely defenseless. Indeed, Sprat insisted that rhetoric was at first no doubt an admirable instrument in the hands of wise men; when [tropes and figures of speech] were only employ’d to describe Goodness, Honesty, Obedience; in larger, fairer, and more moving Images: to represent Truth, cloth’d with Bodies; and to bring Knowledg back again to our very senses, from whence it was at first deriv’d to our understandings. (Sprat 1959: 111–2)
Two features of this passage are noteworthy. Firstly, Sprat appears to be objecting not to the use of tropes as such, but to what he perceives as abuses of tropes. Hesse herself provides us with a fairly good clue of who the perceived abusers were when she quotes Seth Ward objecting to the “windy impositions of magic and astrology, of signatures and physiognomy”. She cites this passage as more evidence that both Sprat and Ward were enemies of metaphor per se who held that “truth value inheres only in the literal, the univocal, the exact, the cool, the impersonally objective” – a claim patently contradicted by Sprat’s statement that tropes can “represent Truth”. Arguably, Ward aimed at a smaller target, namely the abuse of metaphor by alchemists, astrologers, and other adepts of occultist practices. I will pursue that claim in a later section. The second interesting feature of Sprat’s remarks on tropes is his claim that they can be employed to “represent Truth”, and “bring knowledge back again to our very senses, from whence it was at first derived to our understandings”. Similar remarks were quite common among “scientific-minded” seventeenthcentury thinkers, who increasingly realized that metaphor is a necessary feature of scientific inquiry. 2. Classical empiricists on the necessity of metaphor Variations on the idea that tropes could “bring knowledge back again to our very senses, from whence it was at first derived”, surfaced frequently in the seventeenth-century. “Whenever we wish to signify clearly something that is understood or can be understood by the mind”, noted Johann Clauberg (1622–
Metaphor, empiricism and truth
31
65), “we use such modes of speaking by which in general the perceptions of
corporeal organs are properly designated (i.e., those which we attribute to them), or their sensible objects, e.g., from seeing and visible things we transfer them to the intellect and intellectual things”.6 Similarly, in 1674 the Jesuit philologist Pierre Besnier noted that “if we compare them to their first origin, most of our words are nothing but metaphors”.7 In support of his claim, he provided a list of examples, such as the tendency to talk about the mind as the kind of thing that can ‘weigh’ options, that ‘untangle’ or ‘unravel’ evidence, that ‘discover’ truth. Besnier concluded that “in order to explain the actions of the most spiritual of worlds, we make use of images that are actually corporeal in their first origin, though most of them have lost their proper signification to assume another that is purely figurative”. It is to such texts, Lia Formigari (1988: 114) argues, that we can trace the fact that “eighteenth-century philosophers unfailingly resorted to [metaphor] in order to account for both the origin and the operations of language”. But if seventeenth-century philologists had an impact on eighteenth-century philosophers, it is arguably because Locke brought their ideas into the mainstream of seventeenth-century philosophy.8 The following passage comes about as close as one can get to saying, without using the word ‘metaphor’, that no discourse about “things that fall not under our senses” – atoms, spirits, minds – would be possible without metaphor: It may also lead us a little towards the original of all our notions and knowledge, if we remark how great a dependence our words have on common sensible ideas; and how those which are made use of to stand for actions and notions quite removed from sense, have their rise from thence, and from obvious sensible ideas are transferred to more abstruse significations, and made to stand for ideas that come not under the cognizance of our senses; v.g. to imagine, apprehend, comprehend, adhere, conceive, instil, disgust, disturbance, tranquillity, &c., are all words taken from the operations of sensible things, and applied to certain modes of thinking. Spirit, in its primary signification, is breath; angel, a messenger: and I doubt not but, if we could trace them to their sources, we should find, in all languages, the names which stand for things that fall not under our senses to have had their first rise from sensible ideas. (Essay, 3.1.5)
————— 6
7 8
As quoted in Aarsleff (1982:67). To support his point, Clauberg gave examples from ordinary language, noting that we often talk about “grasping” a point, having a “sharp-sighted” mind, and so forth. In short, he gave examples quite similar to the list of UNDERSTANDING-IS-SEEING metaphors in Metaphors We Live By (Lakoff & Johnson 1980: 48). Besnier (1674: 38–9), as quoted in Aarsleff (1982: 82, footnote 74). See also Formigari (1988: 115–6). Aarsleff (1982) indicates Clauberg as one of Locke’s sources. Formigari (1988: 116) also argues that both Locke and Vico read Clauberg, and possibly also Besnier.
32
Nicolaas T. Oosthuizen Mouton
As Leibniz noted in the course of a lengthy commentary on the passage: “This analogy between sensible and insensible things, which has served as the foundation for figures of speech, is worth exploring” (Leibniz 1981, sections 276– 8). He was not alone in deciding that the topic was worth exploring.9 The influence of Locke’s remarks is clearly seen in Bishop Berkeley’s remark that “speech [is] metaphorical more than we imagine, insensible things of their modes, circumstances etc being expressed for the most part by words borrowed from things sensible”.10 Similar observations concerning the prevalence of metaphor in philosophical discourse appear in Berkeley’s Three Dialogues between Hylas and Philonous.11 Hylas asks how Philonous can say that “objects are in your mind, as books in your study: or that things are imprinted on it, as the figure of a seal upon wax. In what sense […] are we to understand those expressions?”. To this, Philonous replies that “when I speak of objects as existing in the mind, or imprinted on the senses, I would not be understood in the gross literal sense; as when bodies are said to exist in a place, or a seal to make an impression upon wax”. Hylas is quick to reply that this makes Philonous guilty of an abuse of language, yet Philonous denies that this is the case: “nothing is more usual than for philosophers to speak of the immediate objects of the understanding as things existing in the mind”. He continues to note that “most part of the mental operations [are] signified by words borrowed from sensible things, as is plain in the terms comprehend, reflect, discourse, &c., which, being applied to the mind, must not be taken in their gross, original sense”. In short, Berkeley was neither blind to his own metaphors, nor to those of his fellow empiricists. Nor yet did he think that it was an abuse to rely on metaphor. In one of Samuel Johnson’s letters to Berkeley, dated 5 February 1730, Johnson remarks: “It is true, those expressions (modifications, impressions, etc.) are metaphorical, and it seems to me to be no less so, to say that ideas ————— 9
10 11
Apart from the evidence I discuss below, Nerlich and Clarke (1996) and Land (1974: 19) point out that the passage became a catalyst to Condillac, to Charles de Brosses and to John Horne Tooke, who interpreted it as opening up the prospect of understanding the mind through an etymological analysis of language – indeed, that etymology would repeat epistemology. Eventually, Locke’s nascent insight into the metaphorical nature of language was also noticed and used by Kant: Nerlich and Clarke (1996, section 1.4) point out that Kant knew this quote well and used it in The Critique of Judgement (§ 49). See also Jäkel (1997). See Berkeley’s Notebook B, 176. The notebooks were only posthumously published in 1871. ‘Philonous’ is generally taken to represent Berkeley’s own idealist views, while Hylas serves as the materialist opposition. Hence, if Philonous says that he wishes not to be understood ‘in the gross literal sense’ on these matters, one can reasonably assume that this is Berkeley’s own view on the matter, and hence, that he was quite aware of being rather far into metaphorical territory.
Metaphor, empiricism and truth
33
exist in the mind, and I am under some doubt whether this last way of speak
ing don’t carry us further from the thing, than to say ideas are the mind variously modified; but as you observe, it is scarce possible to speak of the mind without a metaphor.” That is, Johnson attributes to Berkeley the view that metaphors are necessary rather than nice (at least when speaking of ‘mind’), and Johnson himself admits, grudgingly, that Berkeley is probably right in saying so. The basic chain of reasoning that led to this grudging acceptance of metaphor is spelled out in the fourth book of Locke’s Essay Concerning Human Understanding, in a section entitled “In things which sense cannot discover, analogy is the great rule of probability” (Essay, 4.16.12). Locke opens the section by noting that, up to that point, he had been mostly concerned with “matters of fact, and such things as are capable of observation and testimony”. Such matters, he argues, can be settled within the broad daylight of reason, and leads to “the certainty of true knowledge”. But there is another class of propositions, he admits, about which men “entertain opinions with variety of assent”, and which take us into the twilight of probability. These involve beliefs about things that do not fall “under the reach of our senses […] [and which] are not capable of testimony”. He enumerates a list of different kinds of things that “come not under the cognizance of our senses”: firstly, immaterial beings like spirits and angels; secondly, “the manner of operation in most parts of the works of nature, wherein, though we see the sensible effects, yet their causes are unknown, and we perceive not the ways and manner how they are produced”; and thirdly, material beings which are either too small (e.g. ‘corpuscles’) or too remote (e.g. aliens on other planets) for us to notice them. Such entities, being beyond the reach of our senses, obviously create epistemological problems for a philosopher insisting that it is sense-experience which “supplies our understandings with all the materials of thinking” (Essay, 2.1.2.). Locke admired the impressive scientific feats of his day, and wanted to admit as valid parts of human knowledge such theories as his good friend Robert Boyle’s ‘corpuscularian hypothesis’. This left him with somewhat of a predicament, given that Boyle’s hypothesis postulated unobservable entities in order to explain the behavior of things that we can observe. To understand the role of metaphor and analogy in Locke’s solution to the problem, it is instructive to note that Boyle himself noted that “I make frequent use of similitudes, or comparisons; and therefore, I think myself here obliged to acknowledge once for all, that I did it purposely […] they are not always bare pictures and resemblances, but a kind of argument […] [or] analogous instances which do declare they nature or way of operating of the things they relate to, and by that means do in a sort prove, that, as it is possible, so it is not improbable, that the thing may be such as it is represented” (Boyle, as quoted in Mulligan 1994: 254). Locke’s defense of the claim that we can have some knowledge about things which cannot be experienced directly via the senses takes the same
34
Nicolaas T. Oosthuizen Mouton
direction as Boyle’s. In forming opinions about such matters, Locke concludes, analogy is “the only help we have, and it is from that alone we draw all our grounds of probability […] a wary reasoning from analogy leads us often into the discovery of truths and useful productions, which would otherwise lie concealed” (4.16.12). In cases where we cannot explain what we see in terms of other observable entities, we are forced to rely on analogical conjectures that yield probable explanations. This, we typically do by trying to reckon how what we observe can “more or less hold to [other] truths that are established in our minds, and as they hold proportion to other parts of our knowledge and observation”. Even though Locke credits metaphor with being capable of leading us “into the discovery of truths”, his general drift rather seems to be that “a wary reasoning from analogy” leaves us in the sphere of the probable. It does not provide “clear and certain knowledge”, but such knowledge appears to be rather limited anyway, and we have no choice but to rely on analogies if we are to cope with this world, given the limitations of our senses. If man had nothing to direct him “but what has the certainty of true knowledge”, Locke tells us, he would “be often utterly in the dark, and in most of the actions of his life, perfectly at a stand, had he nothing to guide him in the absence of clear and certain knowledge” (Essay, 4.14.1).12 Rather than to complain about this state of affairs, Locke argues that God, “in the greatest part of our concernments, has afforded us only the twilight of probability” – and that is all we need to get along in practice. This leaves us with an account similar to Hans Blumenberg’s (1994) view of ‘man’ as a ‘Mängelwesen’, who needs metaphors exactly because he is a creature of deficiency. Given the evidence just reviewed, it is surely reasonable to conclude that many seventeenth-century philosophers perceived metaphors and analogies to be pervasive, necessary, and useful. Yet how can such a conclusion fit with the evidence that led Mark Johnson to conclude that metaphor received “one beating after another” at the hands of these same philosophers? How shall we account for these contradictory attitudes? 3. Empiricist critiques of the alchemists’ abuses of metaphor We are not necessarily dealing with anything as troublesome as a genuine contradiction. Arguably, what modern commentators tend to interpret as wholesale rejections of metaphor per se are better understood as limited critiques of what was perceived to be abuses of metaphor. More specifically, —————
12
The point is similar to Lakoff and Johnson’s argument that “though questions of truth do arise for new metaphors, the more important questions are those of appropriate action” (Lakoff & Johnson 1980: 158).
Metaphor, empiricism and truth
35
many seventeenth-century critiques of metaphor were arguably directed at
what was perceived to be widespread abuses of metaphor by the alchemists and, more generally, the inhabitants of those curious regions of early modern thought that have variously been described as “the occult tradition” (Vickers 1984), “Hermeticism” (Yates 1964, Merkel & Debus 1988), “Western esotericism” (Faivre & Hanegraaff 1995), or “the Renaissance episteme” (Foucault 1994). Already quite early in the seventeenth-century, we find Daniel Sennert commenting thus on the alchemists’ use of the Microcosm-Macrocosm analogy: “the Analogie of the great and little World is extended too large by the Chymists, because they make not an Analogie, but an identity, or the same thing”.13 A few decades later, we find Robert Boyle raising a related complaint, namely that the alchemists take unreasonable liberty […] of playing with names at pleasure […] if I were oblig’d in this dispute, to have such regard to the phraseology of each particular Chymist, as not to write any thing which this or that Author may not pretend, nor to contradict this or that sense, which he may give as occasion serves to his Ambiguous Expressions, I should scarce know how to dispute, nor which way to turn myself. For I find that even Eminent Writers (such as Raymund Lully, Paracelsus, and others) do so abuse the termes they employ, that as they will now and then give divers things one name, so they will oftentimes give one thing many names. (Boyle 1661: 200).
Samuel Parker similarly complained about the linguistic abuses of the Rosicrucians.14 In those circles, Parker complained, people “pretend to be Nature’s Secretaries, and to understand all her Intrigues […] and yet put up with nothing but Rampant Metaphors, and Pompous Allegories, and other splendid but empty Schemes of Speech […]” (Parker 1666: 73). For present purposes, two features of these passages are relevant. Firstly, they are all directed against a specific category of perceived abusers of metaphor: the alchemists make unreasonable use of metaphor. Secondly, they all try to establish that the alchemists make unreasonable use of metaphor. This points to – or at least leaves open – the possibility that some ways of using metaphor are reasonable after all. For Sennert does not criticise the alchemists for using metaphors and analogies, but for “extending them too large”; Parker does not complain that the followers of the Rosy Cross use metaphors, but that they trade in “nothing but rampant metaphors”; and Boyle does not say ————— 13 14
Sennert (1619/1662: 26), as quoted in Vickers (1984: 141). Vickers gives a lengthy discussion (ibid.: 136–43) of the relevant parts of Sennert’s critique of Paracelsus. Yates (1972/2002) and Ormsby-Lennon (1988) discuss Rosicrucianism in some detail.
36
Nicolaas T. Oosthuizen Mouton
that one should not take any liberties in playing with names, but rather that the alchemists take “unreasonable liberties”. Indeed, Boyle explicitly defended the responsible use of similes and metaphors. (He often referred to them as ‘comparisons’, a practice that is still not uncommon among subscribers to the comparison theory of metaphor.) As Mulligan points out, Boyle saw “the metaphoric mode […] as central to his own epistemology” (Mulligan 1994: 256). According to Boyle, metaphors do not merely provide a means to make one’s prose more entertaining, but are capable of doing genuine cognitive labor: I presume it will be taken notice of, that, in the following treatise, as well as in divers of my other writings […] I make frequent use of similitudes, or comparisons; and therefore, I think myself here obliged to acknowledge once and for all, that I did it purposely. And my reasons for this practice were, not only because fit comparisons are wont to delight most readers and make the notions they convey better kept in the memory […] but I was induced to employ them chiefly for two other reasons […] comparisons fitly chosen, and well applied […] usefully serve to illustrate the notions […] by placing them in a true light. And […] apposite comparisons not only give light, but strength to the passages they belong to, since they are not always bare pictures and resemblances, but a kind of argument […] analogous instances do declare the nature of way of operating of the things they relate to, and by that means do in a sort prove that, as it is possible, so it is not improbable, that the thing may be such as it is represented.15
As for his characterization of the alchemists’ uses and abuses of metaphor, Boyle suggested that anyone who reckons that he “mistakes the Controversie” should read the alchemists’ writings, and then judge. To see what Boyle had in mind, consider William Gratacolle’s Names of the Philosopher’s Stone, published nine years before Boyle’s Sceptical Chymist.16 Gratacolle lists more than a hundred – I lost count and interest at that number – different names of The Stone. The following selection of examples should serve to give an impression of the whole: a living spirit; a fugitive servant; a marvellous father; a menstruating woman; the beginning of the world; Gold; Sun; a Dragon; a Dragon which eateth his tayle; the dregs of the belly; earth found on the dunghill putrefied, or in horse dung; a dead body; most strong vinegar; a Serpent; a Toad; the green Lion; the quintessence; Camelion; most vild black; blacker than black; Virgins milke; a Basilisk; Jupiter; earth; fire; aire; all things; Chaos. Gratacolle paraded these names, one after the other, without any accompanying explanation. We are merely informed that these names “all of them signifie but one thing, our stone, our brasse [followed by yet another list of —————
15 16
As quoted in Mulligan (1994: 254). The list was included in Five treatises of the Philosophers’ Stone (1652).
Metaphor, empiricism and truth
37
names]”. Or consider Paracelsus’ musings on the topic of the Philosopher’s
Stone: “Summarily, then, the matter of the Philosophers’ Stone is none other than a fiery and perfect Mercury extracted by Nature and Art; that is, the artificially prepared and true hermaphrodite Adam, and the microcosm: That wisest of the philosophers, Mercurius, making the same statement, called the Stone an orphan” (The Aurora of the Philosophers, chapter 16.) In short, occultist Thomas Vaughan was not exaggerating when he casually observed that “that the same thing should have a thousand names is no news to such as have studied the philosopher’s stone” (as quoted in Ormsby-Lennon 1988: 332). At this point, it is maybe useful to step away from the details for a moment, and try to look at the bigger picture. What we are witnessing is basically a clash of two entirely different cognitive styles. If one ignores Foucault’s questionable claim that “in any given culture and at any given moment, there is always only one episteme that defines the conditions of possibility of all knowledge, whether expressed in a theory or silently invested in a practice” (Foucault 1994: 168), then one might say that we are witnessing a clash between what Foucault labeled “the Renaissance Episteme” and “the Classical Episteme”. As far as I can tell, the patterns of thought that he labeled thus coexisted for a rather lengthy period, and while numerous individuals apparently slotted into both with equal ease, the relationship was essentially antagonistic. For present purposes, the main difference between the two styles is their attitude towards similarity. More precisely, one cognitive style is characterized by a blind reliance on resemblance, while the other insists that resemblance is “the occasion of error” unless it is analyzed into relations of identity and difference (ibid.: 51, 67). This broad interpretive framework is quite helpful in rendering intelligible the reasons why the likes of Locke put so much emphasis on the so-called “faculty of distinction”. Consider section 31 of Locke’s Of the Conduct of the Understanding, entitled ‘Distinction’, in which Locke describes and distinguishes between two “vicious excesses”. The first involves a tendency to “fill the head with abundance of artificial and scholastic distinctions”, which eventually cause all order to “crumble into dust” under its weight. For the sake of brevity, let us refer to this tendency as ‘hair-splitting’ – or better still, simply ‘splitting’. Locke notes that splitting is especially prevalent amongst “the schoolmen”. The second “vicious excess” is the exact opposite of the first, and involves “an aptness to jumble things together wherein can be found any likeness” (ibid). Locke reckons that this is “a fault in the understanding […] which will not fail to mislead it”. Now, whereas ‘the schoolmen’ were explicitly mentioned as arch-splitters, no prototypical ‘jumblers’ are explicitly identified. Yet Locke’s depiction of “jumbling” is almost identical with modern historians’ descriptions of the alchemists’ cognitive style as “a generally unifying type of thought, [with] a tendency to ‘lump’ rather than to ‘split’, an urge toward unity, integration” (Jevons 1964: 155).
38
Nicolaas T. Oosthuizen Mouton
What exactly does “jumbling” have to do with metaphor? Well, Locke’s critique of ‘jumbling’ was first and foremost a critique of similarity; and one’s opinions about similarity can hardly fail to have implications for one’s opinions about metaphor. Locke himself connected the two issues explicitly. Having posed his warning against the dangers of jumbling in general, he immediately proceeded to identify similes, metaphors, and allegories as “near kin” within an extended family of jumbling errors. The main danger of metaphors and similes, Locke warns, is that if one let the mind “upon the suggestion of any new notion run immediately after similes to make it clear to itself”, the result is all too often that the relevant mind fails to notice that metaphors “always fail in some part” (Conduct, section 32). The same point is made in Locke’s discussion of the “faculty of discernment” in the Essay. About this faculty, Locke tells us that it is the one “whereby [the understanding] perceives two ideas to be the same, or different” (Essay, 2.11.1). He categorically states that there can be no knowledge without discernment: “insofar as this faculty is in itself dull, or not rightly made use of, for the distinguishing one thing from another, so far our notions are confused, and our reason and judgment disturbed or misled” (Essay, 2.11.2). It prevents us from putting together ideas “with quickness and variety, wherein can be found any resemblance”, and requires a form of judgment which always first “[separates] carefully, one from another, ideas wherein can be found the least difference, thereby to avoid being misled by similitude, and by affinity to take one thing for another. This is a way of proceeding quite contrary to metaphor and allusion […]” (ibid). Now, at first this last remark may seem to be a clear-cut a rejection of metaphor (per se). Indeed, it may appear to amount to a call for banishing similarities altogether. But if you read these passages carefully, without haste, you’ll notice that what troubles Locke is first and foremost the tendency to “run immediately” after similarities, to put ideas together “with quickness”. A mind that is prone to jumbling is not deemed suspect because it assembles ideas, as the passage may seem to imply, nor yet because it assembles them on the basis of similarities. It is suspect because it assembles with haste, chases after similarities without taking note of dissimilarities.17 Similarly, the man of judgment is not praised because he refrains from assembling ideas – be that via metaphor or not – but because he first patiently separates one from another ideas wherein can be found “the least difference”, and only then reassembles them. Thereby, he avoids “taking one thing for another” on account of overlooking the differences that make similar things similar, rather than identical. Locke is basically echoing Daniel Sennert’s remark that “the Analogie of the ————— 17
Two sections of the Conduct (16 and 25) are dedicated to a discussion of ‘haste’. In both instances, haste is held responsible for an impressive range of cognitive errors.
Metaphor, empiricism and truth
39
great and little World is extended too large by the Chymists, because they
make not an Analogie, but an identity, or the same thing” (as quoted in Vickers 1984: 141). In short, seventeenth-century debates about the role of metaphor in “discourses that pretend to inform and instruct” were part of a broader clash between two entirely different cognitive styles. One of the crucial respects in which they differed pertained to their respective attitudes towards similarity. To the one, resemblances were a legitimate source of knowledge; to the other, it was a source of error unless analyzed into clear relations of identity and difference. A rather important point follows from this. Gentner and Jeziorski (1993: 448), who made a detailed comparison of how alchemists like Paracelsus and scientists like Boyle used metaphors, put it very well: “[Although] an appreciation of similarity (including metaphorical similarity) is almost surely universal in human cognition, what to do with this sense of similarity is not. Opinions on how to tame the raw perception of likeness have varied” (Gentner & Jeziorski 1993: 448). They make a second important point that fits very neatly into the case that I have been trying to make thus far: the alchemists’ and the scientists’ different opinions on how metaphors ought to be used, in principle, had a considerable impact on how the members of the two groups actually used them, in practice. More specifically, Gentner and Jeziorski compared the two groups’ metaphors in terms of criteria such as structural consistency and systematicity. Among other things, they found that the scientific tradition’s analogies display one-to-one correspondences between the sourceand target domains that are conspicuously absent in the occult tradition. This follows quite naturally from the former’s insistence that similarities and differences should be spelled out explicitly, and the latter’s tendency to run after resemblances. 18 We may infer two further lessons from the previous two. Firstly, if opinions on how metaphors ought to be used had an impact on how they were actually used, and if such opinions have changed over time, then it follows that we should expect to find differences in how particular metaphors actually evolved during different eras. After all, if it makes a difference whether you ————— 18
One could add that the emphasis on the importance of paying attention to ‘the least difference’ between ideas served to make scientists much more aware of metaphors. Indeed, one may make the stronger claim that advocates of the ‘discern first, liken later’ policy reaffirmed a necessary condition for understanding metaphors as metaphors. For if a metaphor is, basically, a cognitive tool that enables us to speak and think of one kind of thing as if it was a different kind of thing, then it follows that we will only understand ourselves to be speaking metaphorically as long as we are aware of the fact that we are in fact dealing with different kinds of things. Failing that, we will not understand one kind of thing ‘as if’ it was another kind, but simply ‘take one thing for another’, as Sennert and Locke warned.
40
Nicolaas T. Oosthuizen Mouton
pay attention to the differences between the source- and target domains, and if the thinkers of one era largely ignore them while the inhabitants of the next feel obliged to pay attention to “the least difference”, then a metaphor may well survive through both periods, but it will be treated differently and therefore develop in different directions. Secondly, while this does not rule out the idea that the ability to use metaphors is part of a universal human “conceptualizing capacity” (Lakoff 1987), it does seem very unlikely that this is the whole story. The fact that we all have a natural ability to use metaphors would not explain why early modern scientists felt obliged to pay attention to “the least difference” between the source and target domains of the metaphors they used, whereas differences were pretty much ignored by the alchemists. It would be more useful to say that we are dealing with an evolving “style of reasoning” (Hacking 2002: 159–99; Crombie 1994; Fleck 1979), which gradually emerged within particular communities as its members negotiated among themselves what ought to count as legitimate ways to argue, and what not. 4. So what is the legacy of seventeenth-century debates about metaphor, then? As I noted at the beginning of this chapter, commentators like Ted Cohen and Mark Johnson agree, firstly, that the “scientific-minded” philosophers of the seventeenth-century gave metaphors “one beating after another” and “denied to metaphors and their study any philosophical seriousness of the first order” (Cohen 1978: 3; Johnson 1981: 11). By now, the problems with this (mis) characterization of seventeenth-century views are hopefully obvious. Secondly, they agree that “although these remarks of Hobbes and Locke may seem remote, their import has prevailed until quite recently” (Cohen 1978: 5). Or, as Johnson puts it, “it is especially important to understand the typical empiricist stand on metaphor, for it is essentially the same view as that held today by the inheritors of the empiricist legacy” (Johnson 1981: 11). Both Johnson and Cohen imply that the inheritors of the empiricists’ legacy consisted, first and foremost, of positivist philosophers. Thus Johnson (1981: 16) informs us that “twentieth-century Anglo-American thinking about metaphor has been emasculated, narrowed, and inhibited by logical positivist views of language and is therefore either hostile or patronizing toward figurative expression”. Johnson then proceeds to promise that, in order “to show the breadth of this baleful influence, I want to describe the original positivist attack on metaphor” (ibid). Curiously, the only evidence that Johnson cites in support of the claim that such an attack ever took place consists of a single quote from Ogden and Richards’ The Meaning of Meaning (1923), which is hardly a typical positivist treatise. Similarly, Ted Cohen (1978: 5) confidently pronounces that “the works of many twentieth-century positivist philosophers and others either state or imply that metaphors are frivolous and inessential, if not dangerous and logically perverse, by denying to them (1) any capacity to
Metaphor, empiricism and truth
41
contain or transmit knowledge; (2) any direct connection with facts; or (3) any
genuine meaning”. Yet Cohen does not supply a single specific example of a positivist philosopher who actually proposed such views. While absence of evidence is not evidence of absence, talk about a “positivist attack on metaphor” surely ought to be backed up by a few clear-cut examples of such “attacks” by card-carrying positivists. I have found little evidence that would support such a thesis. At any rate, given that Cohen and Johnson arguably misunderstood the classical “empiricist stand on metaphor”, it might be a good idea to take a second look at its legacy, as well. One may begin by noting that, if the likes of Johnson and Cohen were right in holding that the classical empiricists proposed a total ban on metaphor in science, then we may reasonably conclude that “the empiricist legacy” actually had very little practical impact. After all, scientists continued to make liberal use of metaphor. By way of contrast, if my portrayal of the typical seventeenth-century position on how metaphors ought to be used is correct, and if those views had any influence on scientific practice, then we should expect to find, inter alia, a greater concern to avoid “running after resemblances” without pausing to spell out the differences between the source- and target domains of a metaphor. In what follows, I will briefly make a two-pronged case in support of my view of the empiricist legacy. Firstly, I will briefly provide some evidence showing that the way in which modern social scientists use metaphors do in fact conform to this expectation – not always, to be sure, but often. Secondly, I will argue that they use metaphors in this way because the classical empiricists’ ideas about how metaphors ought to be used in science became institutionalized in most scientific disciplines, and thus continue to influence the way in which metaphors are actually used. Taken together, those two arguments in turn add up to an answer to one of the questions that I posed in the introduction to this article. As you may recall, I noted that one of the vexing aspects of the development of the social organism metaphor is that while its advocates extended the metaphor remarkably far, they also went to considerable lengths to spell out, as Spencer (1996: 273) put it, “the leading differences between societies and individual organisms”. I also noticed that this is a rather strange practice: in most contexts, people would surely not feel obliged to spell out all the respects in which the sourceand target domains of the metaphors they happen to be using are dissimilar. This led me to the simple question: why is this focus on the differences between the source- and target domains so prevalent in modern social scientists’ uses of the metaphor? By now, it should hopefully be clear why I think that the legacy of the classical empiricists may provide an answer to that question. But let me first briefly document that Spencer did indeed go to considerable lengths to spell out the differences between the source- and target domains of the metaphor, and that this was by no means an isolated incidence. As far as Spencer is concerned, he not merely dedicated roughly five pages of his essay on The Social Organism (1860/1996) to a discussion of the
42
Nicolaas T. Oosthuizen Mouton
differences between societies and organisms, but returned to the issue in later works like The Study of Sociology (1873), where he added a number of new differences to the list. In the 1860 text, he pointed out that societies, unlike organisms, have no specific external forms; do not form a continuous mass; and do not have mobile parts (Spencer 1996: 273–7). “Perhaps the most important distinction”, he added, “is that while in the body of an animal only a special tissue is endowed with feeling, in a society all the members are endowed with feeling” (ibid). Similarly, in the 1873 text, he pointed out, inter alia, that there is no clear-cut counterpart in the social domain to biological species, and that this has significant implications for students of the social organism: “Biology deals with numerous individuals of a species, and with many species of a genus, and by comparing them we can see what traits are specifically constant and what generically constant; and the like holds more or less with the other concrete sciences. But comparisons between societies, among which we may almost say that each individual is a species by itself, yield much less definite results: the necessary characters are not thus readily distinguishable from the accidental characters” (Spencer 1873: 101). Now, it may of course be the case that Spencer’s focus on difference was an idiosyncratic feature of a personal style of reasoning. But if we turn to subsequent manifestations of the metaphor in twentieth-century sociology, for example, we find the same thing. Consider the work of the so-called “organizational ecologists” mentioned in my introduction. For present purposes, suffice it to say that its proponents basically tried to shift sociologists’ attention away from the question of how individual organizations adapt to their environments, to the question of how environmental selection affects entire populations of organizations. The best way to answer that question, they argued, is to “follow the lead of the bioecologists” (Hannan & Freeman 1977: 934). Already in their first article, however, Hannan and Freeman (1977: 936–8) dedicated an entire section to the theme of “Discontinuities in Ecological Analysis”. They began the section by noting that “utilization of models from ecology in the study of organizations poses a number of analytic challenges involving differences between human and nonhuman organizations with regard to their essential ingredients” (1977: 936). This was followed by fairly detailed discussions of a number of specific differences, such as the fact that “biological analyses are greatly simplified by the fact that most useful information concerning adaptation to the environment […] is transmitted genetically […] and with near invariance”. But there is no obvious counterpart in the organizational world to the genetic transmission of information, and information is hardly ever transmitted “with near invariance” from one organization to another. Subsequent contributors added to the list of disparities by pointing out, inter alia, that organizations, unlike organisms, are potentially immortal (Carroll 1984: 74–5), and that, “unlike biological organisms, organizations do not have offspring, though some have suggested that ‘spin-off’ organizations may be analogous” (McKelvey & Aldrich 1983: 114).
Metaphor, empiricism and truth
43
Now, by pointing out these differences, the advocates of organizational ecology created considerable difficulties for themselves. For example, since they were basically trying to apply biological (Darwinian) models to organizational problems, and since net mortality rates play a crucial role in the models they borrowed, the fact that organizations are potentially “immortal” obviously constituted a considerable challenge. Why then did they create such difficulties for themselves? Why not simply use metaphors the way politicians do, and ignore all the ways in which nations are unlike organisms, markets unlike invisible hands, policy initiatives unlike bridges to the future? Arguably, the best explanation is to be found in the fact that, throughout the formative years of organizational ecology, methodological discussions in the social sciences frequently contained remarks that remind the reader rather a lot of Locke’s warning that, when resorting to metaphor, one should take care to pay attention to “the least difference” and thereby “avoid being misled by similitude, and by affinity to take one thing for another” (Essay, 4.14.4). Thus Abraham Kaplan (1964: 266), in his reflections on “The Conduct of Inquiry: Methodology for Behavioral Science”, sternly warned that “no two things in the world are wholly alike, so that every analogy, however close, can be pursued too far”. Here is a more elaborate version of the same argument, taken from an article in the Administrative Science Quarterly – which also happens to be one of the journals in which organizational ecologists have published a lot of their work:
To use metaphors precisely we must specify which important properties the relevant items do not share. It is not enough to say what the similarities are in terms of shared properties. To speak only of similarities leaves open the possibility of implied identity, which should be ruled out if no identity is intended […] The point at which a metaphor stops being of positive heuristic value and starts to become misleading is difficult to detect. (Pinder & Bourgeois 1982: 642–3)
These prescriptions mirror the ones made by Locke in at least five respects.19 Firstly, like the classical empiricists, Pinder and Bourgeois did not object to ————— 19
There is also at least one important difference. Locke’s formulations frequently imply that one has to enumerate all the differences – including “the least difference” – between the source- and target domains. By way of contrast, Pinder and Bourgeois hold that “to use metaphors precisely we must specify which important properties the relevant items do not share” (Pinder & Bourgeois 1982: 642–3). This is an important qualification, for there are obviously incredibly many differences between the domains that feature in any modestly interesting metaphor. The fact that some differences are ignored while others become the subject of recurrent debates adds considerably to the complexity of the issue that we need to explain. Unfortunately, Pinder and Bourgeois neither explain why exactly certain differences are important while others are irrelevant, nor how to tell them apart in practice.
44
Nicolaas T. Oosthuizen Mouton
the use of metaphor per se, but to “the unconstrained use of tropes (such as similes, analogies, and metaphors) in the development and presentation of formal theory”: in their view, this would impede “the progress of administrative science” (Pinder & Bourgeois 1982: 641, emphasis added). That this should not be construed as an attempt to completely banish metaphor, the authors emphasized on the very first page of the article: “We wish to acknowledge openly that it is virtually impossible to eschew metaphors in regular discourse. Of necessity, therefore, the present paper will be replete with them.” (Ibid.) Secondly, their claim that “to use metaphors precisely we must specify which important properties the relevant items do not share” may well have been taken directly from Locke, even though they do not mention him. For as you may recall, Locke also insisted that a sound faculty of judgment always takes care to first “separate carefully, one from another, ideas wherein can be found the least difference” (Essay, 4.14.4). Thirdly, Pinder and Bourgeois’ reasons for insisting that one should specify the differences between the source- and target domains of a metaphor are roughly the same as Locke’s. Locke, as you may recall, insisted that metaphors and similes “always fail in some part”, and suggested that by paying attention to difference, one could “avoid being misled by similitude, and by affinity to take one thing for another” (Essay, 4.14.4). Similarly, Pinder and Bourgeois warned that people often overlook the “critical difference between metaphor and identity” (1982: 642). To say that x is identical to y, they pointed out, means that any property that x has, y has as well. But while metaphors are patently not identity statements, Pinder and Bourgeois continued to observe, people frequently argue as if they are, and simply assume that, because the source domain contains a certain property, so does the target domain. To counter this tendency, Pinder and Bourgeois’ (ibid.: 643) insist that “even when the object used as a metaphor shares a number of defining characteristics with organizations, each inference about organizations must be tested on its own merits and not simply assumed to hold because it holds in the metaphor”. Fourthly, like their seventeenth-century predecessors, Pinder and Bourgeois’ critique of perceived abuses of metaphor was subsequently equated with a rejection of metaphor per se (see e.g. Morgan 1983) – a position from which they then distanced themselves emphatically: “At various places in his paper, Morgan states or implies that we would advocate the total abolition and avoidance of tropes in administrative science. We do not advocate such a position, for, as we noted more than ten times in our paper, figurative terminology is ubiquitous in ordinary language” (Bourgeois & Pinder 1983: 611). Even so, that is the position that they have become associated with. Finally, the very fact that their position was attacked from a different position points to an important point, namely that styles of reasoning – including the style of metaphorical reasoning that became institutionalized in most sciences after the seventeenth-century – are neither unchanging nor unchallenge-
Metaphor, empiricism and truth
45
able. On the contrary, one could make a good case that in fields such as “or
ganization studies”, the kind of ideals which Pinder and Bourgeois championed has become a minority position over the course of the last two decades, first and foremost due to the impact of Gareth Morgan’s enormously popular Images of Organization (1986). It is beyond the scope of this chapter to analyze the ideas propagated in that book, or to examine its impact. Suffice it here to say that Locke would have cringed if he had to witness the way in which the members of the relevant field run after resemblances, without pausing to pay attention to any differences. The more important point here is that the rules of metaphorical reasoning can always be re-negotiated by the members of a community, which makes it rather difficult to generalize about the epistemology of metaphor. 5. Conclusion Since this chapter turned out to be rather long, and covered a lot of rather strange territory, let me conclude with a very brief recapitulation of the main points. The “scientific-minded” philosophers of the seventeenth-century did not try to banish metaphor from science altogether, as modern commentators like Mark Johnson and Ted Cohen suggest. Rather, they criticized specific abuses of metaphor, and tried to formulate guidelines for how to use metaphor in a defensible fashion. Their main prescription pertained to the importance of paying attention to the differences between the source- and target domains. Those ideals about how metaphors ought to be used influenced how metaphors were actually used. Insofar as their views on metaphor survived, that is the kind of legacy we should look for. If we look for it, we do indeed still find very similar views on how metaphor ought to be used in scientific contexts, and they do have an impact on how metaphors are used in practice. Modern proponents of the notion that collective entities are organisms, for example, have by now offered remarkably long lists of respects in which collective entities are not like organisms. This does not simply reflect a natural “conceptualizing capacity”, but rather a historically situated style of reasoning that evolved via debate. A crucial feature of styles of reasoning, as Ian Hacking (2002: 159–77) noted, is that they supply the very criteria which a statement has to satisfy in order to count as a candidate for truth-or-falsehood. Insofar as that is the case, the kind of historical reconstruction that I offered in this chapter offers the most obvious way in which to address questions pertaining to the epistemology of metaphor. It does not provide a universal answer to the question of whether metaphors can be true, but simply makes explicit the origins and evolution of the criteria to which the members of a particular epistemic community happen to appeal when they have to decide whether a given metaphorical utterance is true or false or meaningless. The folks on the other side of the mountain may well have different ideas about how to tell
46
Nicolaas T. Oosthuizen Mouton
whether a metaphor is meaningless or meaningful, and if the latter, whether it is true or false. References Aarsleff, Hans: From Locke to Saussure: Essays on the Study of Language and Intellectual History. Minneapolis: University of Minnesota Press 1982. Aristotle: The Complete Works of Aristotle: The Revised Oxford Translation. Ed. by J. Barnes. Princeton: Princeton University Press 1984. Berkeley, George: Three Dialogues between Hylas and Philonous. Kessinger Publishing 1713/2005. http://books.google.com/books?id=uCGHNJwV8MsC. – Posthumously published notebooks A & B. 1707–08. [First printed 1871] Retrieved online at the International Berkeley Society’s website: http://www. georgeberkeley.org.uk/ Blumenberg, Hans: “An anthropological approach to the contemporary significance of rhetoric”, in: Baynes, Kenneth, James Bohman, Thomas McCarthy (eds.): After Philosophy: End of Transformation? Cambridge, MA: The MIT Press 1994, 429– 58. Bourgeois, V. W., C.C. Pinder: “Contrasting philosophical perspectives in administrative science: A Reply to Morgan”, in: Administrative Science Quarterly, 28(4) (1983), 608. Boyle, Robert: The Sceptical Chymist. 1661. Retrieved online at: http://www.library.upenn.edu/etext/collections/science/boyle/chymist/200.html. Carroll, G. R.: “Organizational ecology”, in: Annual Review of Sociology, 10 (1984), 71–93. Clark, Stephen H.: “‘The whole internal world his own’: Locke and metaphor reconsidered”, in: Journal of the History of Ideas, vol. 59 (1998), 241–65. Cohen, Ted: “Metaphor and the cultivation of intimacy”, in: Critical Inquiry, Vol. 5, No. 1, Special Issue on Metaphor, 1978, 3–12. Crombie, A. C.: Styles of Scientific Thinking in the European Tradition. 3 Volumes. London, Duckworth 1994. De Man, Paul: “The epistemology of metaphor”, in: Critical Inquiry, 5 (1978), 13–30. [Reprinted in S. Sacks (ed.) (1979) On Metaphor. Chicago: University of Chicago Press.] Descartes, René: The Philosophical Writings of Descartes, Volume 1. Edited by John Cottingham, Robert Stoothoff, and Dugald Murdoch. Cambridge: Cambridge University Press 1985. Eco, Umberto: Semiotics and the Philosophy of Language. Bloomington, IN: Indiana University Press 1984. – The Search for the Perfect Language. London: Fontana Press 1995. Faivre, Antoine: “Renaissance hermeticism and the concept of Western esotericisim”, in: Van den Broeck, Roelof, Wouter J. Hanegraaff (eds): Gnosis and Hermeticism from Antiquity to Modern Times. Albany: SUNY Press 1998. Formigari, Lia: Language and experience in 17th-century British Philosophy. Amsterdam: John Benjamins 1988. Foucault, Michel: The Order of Things: An Archaeology of the Human Sciences. New York: Vintage Books 1970/1994.
Metaphor, empiricism and truth
47
Geertz, Clifford: The Interpretation of Cultures. New York: Basic Books 1973/2000.
Gentner, Dedre, M. Jeziorski: “The shift from metaphor to analogy in Western science”, in: Ortony, Andrew (ed.): Metaphor and Thought. Cambridge, UK: Cambridge University Press 1993, 447–80. Glucksberg, Sam, Boaz Keysar: “How metaphors work”, in: Ortony, Andrew (ed.): Metaphor and Thought. Cambridge, UK: Cambridge University Press 1993, 401– 24. Gratacolle, William: Five treatises of the Philosophers’ Stone. London 1652. Extract available online at: http://www.alchemywebsite.com/gratacol.html. Gutting, Gary: Michel Foucault’s Archaeology of Scientific Reason. Cambridge, UK: Cambridge University Press 1989. Haack, Susan: “‘Dry truth and real knowledge’: Epistemologies of Metaphor and Metaphors of Epistemology”, in: Hintikka, Jaakko (ed.): Aspects of Metaphor. Amsterdam: Kluwer Academic Publishers 1994, 1–22. Hacking, Ian: “Styles of scientific reasoning”, in: Rajchman, John, Cornel West (eds.): Post-Analytic Philosophy. New York: Columbia University Press 1985, 145–65. – “Locke, Leibniz, language and Hans Aarsleff”, in: Synthese, 75/2 (1988), 135–53. – Historical Ontology. Cambridge, Mass.: Harvard University Press 2002. Hannan, M. T., J. Freeman: “The population ecology of organizations”, in: The American Journal of Sociology 82(5) (1977), 929–64. – “Where do organizational forms come from?”, in: Sociological Forum 1(1) (1988), 50–72. Hesse, Mary: “Ayer and the philosophy of science”, in: Gower, Barry (ed.): Logical Positivism in Perspective: Essays on Language, Truth, and Logic. London: Croom Helm 1987, 69–88. Hobbes, Thomas: Leviathan: A Critical Edition / by G.A.J. Rogers and Karl Schuhmann. Bristol: Thoemmes Continuum 1660/2003. Jäkel, Olaf: “Kant, Blumenberg, Weinrich: Some forgotten contributions to the cognitive theory of metaphor”, in: Gibbs, Raymond W., Gerard J. Steen (eds.): Metaphor in Cognitive Linguistics. Amsterdam–Philadelphia: John Benjamins 1997, 9– 27. James, William: The Principles of Psychology. 1890. Retrieved online at http://psychclassics.yorku.ca/James/Principles/ Jevons, R.F.: “Paracelsus’s two-way astrology”, in: British Journal for the History of Science, 2 (1964), 139–55. Johnson, Mark: “Introduction: Metaphor in the philosophical tradition”, in: Johnson, Mark (ed.): Philosophical Perspectives on Metaphor. Minneapolis: University of Minnesota Press 1981, 3–47. Johnson, Samuel: Letter to George Berkeley, 5 February 1730. Retrieved online at the International Berkeley Society’s website: http://www.georgeberkeley.org.uk/ Kaplan, Abraham: The Conduct of Inquiry: Methodology for Behavioral Science. New Brunswick, N.J.: Transaction Publishers 1964/1998. Lakoff, George: Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: The University of Chicago Press 1987. – “The contemporary theory of metaphor”, in: Ortony, Andrew (ed.): Metaphor and Thought. Cambridge, UK: Cambridge University Press 1993, 202–51. – , Mark Johnson: Metaphors We Live By. Chicago: The University of Chicago Press 1980.
48
Nicolaas T. Oosthuizen Mouton
Leibniz, Gottfried Wilhelm: New Essays on Human Understanding. Translated and edited by P. Remnant and J. Bennet. Cambridge, UK: Cambridge University Press 1981. Levin, Samuel R.: Metaphoric Worlds: Conceptions of a Romantic Nature. New Haven: Yale University Press 1988. Locke, John: An Essay Concerning Human Understanding. Ed. by Peter Nidditch. th Oxford: Clarendon Press 1690/1989. (17 edition, 1775, available at http://books.google.com/books?id=6QYOAAAAYAAJ, 30th edition, 1849, available at http://books.google.com/books?id=YxwGAAAAQAAJ. – Of the Conduct of the Understanding. London 1706 (posthumous publication). [New York: Teachers College Press 1966.] Retrieved from http://www.ilt.columbia.edu/publications/CESdigital/locke/conduct/toc.html Losonsky, Michael: “Locke on meaning and signification”, in: Rogers, G.A.J. (ed.): Locke’s Philosophy: Content and Context. Oxford: Clarendon Press 1994, 123–41. Maclean, Ian: “Foucault’s Renaissance Episteme reassessed: An Aristotelian counterblast”, in: Journal of the History of Ideas, Vol. 59, No. 1, Jan. (1998), 149–66. McKelvey, B., H. Aldrich: “Populations, natural selection, and applied organizational science”, in: Administrative Science Quarterly 28(1) (1983), 101–28. Merkel, Ingrid,, Allen G. Debus (eds.): Hermeticism and the Renaissance: Intellectual History and the Occult in Early Modern Europe. London: Associated University Presses 1988. Milton, J.R.: “Locke at Oxford”, in: Rogers, G.A.J. (ed.): Locke’s Philosophy: Content and Context. Oxford: Clarendon Press 1994, 29–47. Morgan, Gareth: “More on Metaphor: Why we Cannot Control Tropes in Administrative Science”, in: Administrative Science Quarterly, 28(4) (1983), 601. – Images of Organization. Beverly Hills: Sage 1986 Mulligan, Lotte: “Robert Boyle, ‘Right reason’ and the meaning of metaphor”, in: Journal of the History of Ideas, 55 (1994), 235–57. Musolff, Andreas: “Ignes fatui or apt similitudes? The apparent denunciation of metaphor by Thomas Hobbes”, in: Hobbes Studies 18 (2005), 96–113. Nerlich, Brigitte: Semantic Theories in Europe, 1830–1930. Amsterdam: John Benjamins 1996. – , David D. Clarke: “Mind, meaning, and metaphor: The philosophy and psychology of metaphor in nineteenth-century Germany”, in: History of the Human Sciences, 14:2 (2001), 39–61. Ormsby-Lennon, Hugh: “Rosicrucian linguistics: Twilight of a Renaissance tradition”, in: Merkel, Ingrid, Allen G. Debus (eds.): Hermeticism and the Renaissance: Intellectual History and the Occult in Early Modern Europe. Washington–London– Toronto: Associated University Presses 1988, 311–44. Paracelsus: Hermetic and Alchemical Writings of Paracelsus. Translated by Arthur Edward Waite. Kila Montana: Kessinger Publishing 2002. http://books.google.com/books?id=3dgsJJkVQKkC. Parker, Samuel: A Free and Impartial Censure of the Platonick Philosophie. Oxford: Printed by W. Hall, for Richard Davis 1666, 2nd ed. Pinder, C. C., V.W. Bourgeois: “Controlling tropes in administrative science”, in: Administrative Science Quarterly 27(4) (1982), 641. Ricoeur, Paul: The Rule of Metaphor. Multi-disciplinary Studies of the creation of meaning in language. Translated by R. Czerny with K. McLaughlin and J. Costello. London: Routledge 1978. Spencer, Herbert: The Study of Sociology. London: Henry S. King 1873.
Metaphor, empiricism and truth –
49
Essays: Scientific, Political and Speculative. (12 Volumes). London: Routledge 1891/1996. [Reprint of 1891 edition.] Sprat, Thomas: The History of the Royal Society of London for the Improving of Natural Science. London: J Martyn 1667. [New ed., London: Routledge & Keegan Paul 1959.] Original edition now avialble via http://books.google.com/books?id=YTYJAAAAQAAJ. Stutterheim, C.F.P.: Het Begrip Metaphoor. Een Taalkundig en Wijsgerig Onderzoek. Amsterdam: Acad. Proefschrift 1941. Vickers, Brian: “Analogy versus identity: The rejection of occult symbolism, 1580– 1680”, in: Vickers, Brian (ed.): Occult and Scientific Mentalities in the Renaissance. Cambridge: Cambridge University Press 1984, 95–163 – “On the function of analogy in occult science”, in: Merkel, Ingrid, Allen G. Debus (eds.): Hermeticism and the Renaissance: Intellectual History and the Occult in Early Modern Europe. Cranbury, N.J.: Associated University Presses 1988, 269– 92. Vogt, Philip: “Seascape with fog: Metaphor in Locke’s Essay”, in: Journal of the History of Ideas, 54 (1993), 1–18. Walker, William: Locke, Literary Criticism, and Philosophy. Cambridge, UK: Cambridge University Press 1994. Way, Eileen: Knowledge Representation and Metaphor. Amsterdam: Kluwer Academic Publishers 1991. Yates, Frances: Giordano Bruno and the Hermetic Tradition. Chicago: University of Chicago Press 1964. –
The Rosicrucian Enlightenment. London: Routledge 1972/2002.
Pedro José Chamizo Domínguez (Málaga, E) / Brigitte Nerlich (Nottingham, GB)
Metaphor and truth in Rationalism and Romanticism
0. 1. 1.1 1.2 2. 3. 4.
Introduction Rationalism Man and language Ideal versus actual communication Between rationalism and romanticism Romanticism and beyond Conclusion
0. Introduction Since it is still a widely held belief that “[b]efore the twentieth century, metaphor has rarely commanded the attention of philosophers” (Leezenberg, 2001: 1), the aim of this article is to explore how metaphor and other tropes did command the attention of rationalist and romantic philosophers. As Nicolaas T. Oosthuizen Mouton has shown in his chapter (see this volume, 23–49) with reference to certain misunderstandings surrounding John Locke’s position on metaphor, this is certainly far from true. Not only so-called empiricist philosophers, such as Locke, debated the advantages and disadvantages of metaphor, so-called rationalist philosophers did so as well, as we will show in this chapter (keeping in mind that the rationalism/ empiricism divide is far more shallow than generally assumed).1 It is also widely thought that, in contrast with both empiricist and rationalist philosophers, romantic literary critics and philosophers, inspired in part by Giambattista Vico (1725) and Jean Jacques Rousseau (1781) rediscovered metaphor (together with the power of imagination and creativity), but “tended to portray poetry, and metaphor along with it, as essentially different from everyday language: both poetry and metaphor came to be seen as expressing an ‘emotive’, rather than a ‘cognitive’ mean————— 1
Although other analogous accounts on the topic (Johnson 1985: 3–47; and Bustos Guadaño 2000: 33–51) had paid attention to empiricist philosophers, such as Hobbes or Locke’s reflection on metaphor, the opinions of rationalist philosophers remain almost unexplored. In fact, neither Johnson nor Bustos Guadaño allude to them in their respective accounts. Nevertheless, as we will try to show here, the topic has been widely considered inside the rationalist tradition where an initial contempt of metaphor gradually lead to a positive appreciation by, for example, a rationalist philosopher such as Leibniz.
52
Pedro José Chamizo Domínguez / Brigitte Nerlich
ing” (Leezenberg 2001: 1). Here again one has to be careful, as the ‘emotive’/‘cognitive’ distinction was already well established in empiricist and rationalist philosophical discourse and as the view that metaphor was pervasive in everyday life and thought was discussed from César Chesneau Du Marsais, at the end of the 18th century, up to Friedrich Nietzsche and beyond. In the following we shall first discuss rationalist philosophers’ views on metaphor and truth. This also involves exploring more closely the distinctions made by some philosophers between the emotive and the cognitive, the figurative and the literal, the female and the male, the subjective and the objective, all linked to two fundamental dichotomies: clarity vs. obscurity and truth vs. obfuscation. As one might expect, the rationalist/empiricist account of metaphor is rooted in classical accounts, but, more surprisingly, it also anticipates other accounts on the topic in the 20th century. As neither rationalists nor empiricists provide any particular definition of metaphor, one may assume that they both share the Aristotelian definition of metaphor as “the application of a strange term either transferred from the genus and applied to the species or from the species and applied to the genus, or from one species to another or else by analogy” (Aristotle 1932: 1457b). However, as we are tying to show in this chapter, they also anticipate, obviously using their own particular jargon, some ideas about metaphor which would become prominent in the 20th century, namely, 1) that any metaphoric utterance entails a “flouting” or “categorial falsity” (Grice 1989: 34); 2) that there is no metaphorical meaning and, consequently, no metaphorical truth (Davidson 1984: 245–64); and 3) that the language we use in order to refer to the body and the mind is metaphorically intermingled (Lakoff/Johnson 1980). In a second part we provide an overview of the discussion of metaphor from Vico and Rousseau up to Nietzsche (for a more detailed account of the status of metaphor in the 19th century, see Nerlich/Clarke [2001]). 1. Rationalism Although the two main philosophical systems in the history of modern philosophy, rationalism and empiricism, are usually presented as being opposed to each other, there are many theoretical and thematic overlaps between the two. This is particularly true for the rejection of rhetoric and the use of tropes in philosophical discourse. It should be stressed that this rejection is rooted in a host of other convergent or divergent philosophical theses and historical contingencies. Rationalism and empiricism shared, for example, a specific conception of language which made them, on the whole, regard figures of speech as communicative and cognitive vices, but they were opposed in so far as rationalists believed in innate ideas and empiricists believed ideas were acquired through the senses. Using José Ortega y Gasset’s distinction between
Metaphor and truth in Rationalism and Romanticism
53
‘ideas’ and ‘beliefs’ (Ortega 1934: 377–409),2 one can say that their general conception of language and tropes belongs to the sphere of beliefs or unconscious knowledge, while the subject of innatism belongs to the sphere of ideas or conscious knowledge. We shall focus here on the sphere of beliefs, namely the rejection of tropes and figures of speech. Let us analyse the thinking that underlies that rejection, explicitly and consciously shared by rationalists and empiricists. For both rationalists and empiricists the object of philosophy is the problem of knowledge, or, to borrow Ernst Cassirer’s term Das Erkenntnisproblem (1906–1907).3 And this in three ways. Firstly, the aim of philosophy was to justify science and give it a firm grounding. This can only be achieved if one can attain a type of knowledge which is based on certainty and is free of any reasonable doubt. Secondly, the aim of philosophy was to establish on this basis a science which would allow man to subjugate nature (Francis Bacon’s natura parendo vincitur) and to become nature’s master (René Descartes’ nous rendre comme maîtres et possesseurs de la nature). And thirdly, philosophy was supposed to construct a language which would be similar to the one used by the sciences (especially mathematics, geometry, and so on), that is, a language which would be clear and straightforward (free of tropes and hyperbatons, a figure of speech in which words that naturally belong together are separated from each other for emphasis or effect), and in which words would be used according to their proper (literal) meanings and according to an ideal syntactic order. 1.1 Man and language It is well known that for rationalists man is a composite of mind (res cogitans, in Cartesian jargon) and body/extension (res extensa, in Cartesian jargon), with the mind (cogito) being its most important component, from the cognitive and ontological point of view. However, despite trying to disregard the body (extension) when thinking about man, rationalist philosophers found that —————
2
3
According to Ortega y Gasset ‘ideas’ (ideas, in Spanish) are those opinions, theses, or thoughts which are consciously known by people of a given epoch and, consequently, they are explicitly discussed, accepted, or rejected. By contrast, ‘beliefs’ (creencias, in Spanish) are those opinions, theses, or thoughts which are unconsciously known by people of a given epoch and, consequently, they are accepted as “truths”. Ideas can become beliefs, and vice versa, with the passage of time. For instance, the existence of God was probably a belief in the Middle Ages and, consequently, nobody discussed it or called it into question. By contrast, the existence of God became an idea in the Enlightenment and, consequently, it was called into question and continues to provoke debate until today. For the purpose of this paper we disregard Maréchal’s (1949) opposing point of view. According to Maréchal the core problem in modern philosophy is a metaphysical problem. But perhaps Cassirer and Maréchal were both right, because metaphysical problems and epistemological problems are always linked.
54
Pedro José Chamizo Domínguez / Brigitte Nerlich
their main problem was not how to justify the existence of the mind, but how to justify the existence of the body. It also became clear to rationalist philosophers that many of our philosophical troubles stem from the fact that our body and mind are inextricably linked. Blaise Pascal was particularly aware of this problem and came to anticipate some of Mark Johnson and Eve Sweetser’s theses on the use of bodily metaphors when speaking about the mind and of mental metaphors when speaking about the body. However, unlike cognitive linguists today, Pascal regarded this mixing of cognitive and bodily spheres in language as a danger to philosophy. But despite expressing such fears, the mind-body metaphors and philosophical paradoxes that Pascal used in his own writing, such as “Man is a but a reed, weakest in nature, but a reed which thinks”, “the geometrical spirit”, and “The heart has its reasons of which reason knows nothing”, and so on, became popular buzz-words amongst the cultural elite of the 18th century and beyond: Et ainsi, si nous [sommes] simples matériels, nous ne pouvons rien du tout connaître, et si nous sommes composés d’esprit et de matière, nous ne pouvons connaître parfaitement les choses simples, spirituelles ou corporelles. De là vient que presque tous les philosophes confondent les idées des choses, et parlent des choses corporelles spirituellement et des spirituelles corporellement. Car ils disent hardiment que les corps tendent en bas, qu’ils aspirent à leur centre, qu’ils fuient leur destruction, qu’ils craignent le vide, qu’[ils ont] des inclinations, des sympathies, des antipathies, qui sont toutes choses qui n’appartiennent qu’aux esprits. Et, en parlant des esprits, ils les considèrent comme en un lieu, et leur attribuent le mouvement d’une place à une autre, qui sont choses qui n’appartiennent qu’aux corps. Au lieu de recevoir les idées de ces choses pures, nous les teignons de nos qualités, et empreignons [de] notre être composé toutes les choses simples que nous contemplons. (Pascal 1976: 84) [“So, if we are simply material, we can know nothing at all; and if we are composed of mind and matter, we cannot know perfectly things which are simple, whether spiritual or corporeal. Hence it comes that almost all philosophers have confused ideas of things, and speak of material things in spiritual terms, and of spiritual things in material terms. For they say boldly that bodies have a tendency to fall, that they seek after their centre, that they fly from destruction, that they fear the void, that they have inclinations, sympathies, antipathies, all of which attributes pertain only to mind. And in speaking of minds, they consider them as in a place, and attribute to them movement from one place to another; and these are qualities which belong only to bodies. Instead of receiving the ideas of these things in their purity, we colour them with our own qualities, and stamp with our composite being all the simple things which we contemplate.” Pascal 2005: 72]
Like Locke before him, Pascal therefore came to the conclusion that the use of figures of speech, particularly when bodily metaphors are used to speak about the mind and mental metaphors are used to speak about the body, is dangerous and a source of confusion and should therefore be avoided in philosophy. However, Pascal’s own use of metaphors was intricately linked up with his philosophical thinking, again just as in the case of Locke.
Metaphor and truth in Rationalism and Romanticism
55
The real question, asked by neither Locke nor Pascal was: Is it really possible to avoid these linguistic mind-body ‘confusions’, if, as Descartes pointed out, deep down, language itself, just like man, is also a compound or mixture of something immaterial (meaning, idea, thought) and something material (sign, sound, word)?: Ainsi, lorsqu’on apprend une langue, on joint les lettres ou la prononciation de certains mots, qui sont des choses matérielles, avec leurs significations, qui sont des st pensées. (Lettre à Chanut, February, 1 1647, in: Descartes 1973: IV, 604) [“Thus, when one learns a language, one combines the letters or pronunciation of certain words, which are material things, with their meanings, which are thoughts.” Our translation].4
The fact that humans and language are this kind of mixture or compound of something spiritual and something material has two important consequences. Firstly, the connection between the signifier and the signified can only be the result of human convention, can only be a human institution: Des mots, qui ne signifient rien que par l’institution des hommes, suffisent pour nous faire concevoir des choses, avec lesquelles ils n’ont aucune ressemblance (Le Monde, in: Descartes, 1973: XI, 4) [“Words, which mean something through human convention, are enough to make us conceive things which don’t resemble them in any way whatsoever.” Our translation].5
Secondly, and most importantly, it allows philosophers to use language as an example in order to explain the problem that is at the core of rationalism: the one about the union of body and mind. This is made eminently clear in the ————— 4
5
The same idea can be found in other places and authors: “The general use of Speech, is to transfer our Mental Discourse, into Verbal; or the Train of our Thoughts into a Train of Words” (Hobbes 1950: 23). The same idea and almost the same words can be found in other authors and places. For instance in the Grammar of Port-Royal: “Parler, est expliquer ses pensées par des signes, que les hommes ont inventez à ce dessein” (Arnauld/Lancelot 1966: 5) [“Speaking ist explaining one’s thoughts by means of signs, that men have invented on purpose.” Our translation]; or “Une des principales choses, que je trouve digne de considération touchant ces signes, est qu’ils n’ont aucune conformité avec les pensées, que l’on y joint par institution. En effet, soit que nous exprimions nos pensées par des gestes, par des discours, ou par des caractères, qui son les trois sortes de signes les plus ordinaires, pas lesquels nous fassions connaître nos pensées, nous voyons bien (si nous y faisons un peu de réflexion) qu’il n’y a rien de moins ressemblant à nos pensées, que tout ce que nous sert à les expliquer.” (Cordemoy 1968: 209) [One of the chief things I finde worthy of consideration touching these signs, is, That they have not any resemblance to the Thoughts, which men joyn to them by institution. And indeed, whether we express our thoughts by gestures, by discourse, or by characters, (which are the three sorts of the most used signs, by which we manifest our thoughts) we cannot by see (if we consider it with some attention) that there is nothing less resembling our Thoughts, than is all that, which serves us to express them. (Cordemoy 1668: 21; original capitals and italics)].
56
Pedro José Chamizo Domínguez / Brigitte Nerlich
following passage from Géraud de Cordemoy’s Discours physique de la parole first published in 1668:6 Mais, ce que je trouve de plus admirable en cela, c’est que cette extrême différence qu’il y a entre ces signes & nos pensées, en nous marquant celle qui est entre notre corps & notre âme, nous donne en même temps à connaître tout le secret de leur union. Au moins il me semble que cette étroite union, que la seule institution des hommes est capable de mettre entre certains mouvements extérieurs, & nos pensées, est, à qui veut y prendre garde, le plus beau moyen de concevoir en quoi consiste véritablement l’union du corps & de l’âme. Car enfin, si l’on conçoit que les hommes puissent par institution joindre certains mouvements à certains pensées, on ne doit pas avoir de peine à concevoir que l’Auteur de la nature, en formant un homme, unisse si bien quelques pensées de son âme à quelques mouvements de son corps que ces mouvements ne puissent être excités dans le corps, qu’aussitôt des pensées ne soient excitées en l’âme; & que réciproquement, dés que l’âme veut que le corps soit mû d’une certaine façon, il le soit en même temps. (Cordemoy, 1968: 210) [But what is most admirable herein is, that this vast difference between those Signs and our Thoughts, doth by marking to us that, which is between our Body and Soul, teach as at the same time the whole Secret of their Union. At least methinks, that that strict union, which is the sole institution of men, is able to settle betwixt certain external Motions, and our Thoughts, is to him that will consider it, the best means to conceive, wherein in truth consists the Union of the Body and the Soul. For certainly, if we do conceive, that men can by institution joyn certain Motions to certain Thoughts it cannot, be hard to conceive, that the Author of Nature, in forming a man, so well unites some Thoughts of his soul to some motions of his Body, that those motions cannot be raised in the Body, but the thoughts must also be forthwith excited in the Soul, and that reciprocally as soon as the Souls will have the Body move after a certain manner, it be so at the same time. (Cordemoy 1668: 22–3; original capitals and italics)]
By using an argument a fortiori, Cordemoy suggests that language can be a reasonable example for and at the same time an explanation of the problem of the union of body and mind. Nevertheless, although language can explain some philosophical problems, language itself becomes (again) a problem. If humans need language in order to communicate their thoughts, it is because humans are both thought and extension. In fact, pure spirits shouldn’t need language as they can communicate their thoughts to humans and/or amongst themselves without using language: Au reste, ce qui se dit de la communication de deux purs Esprits, se doit dire de celle qui peut être entre un esprit uni à un corps, & un esprit qui n’y est pas uni. Car enfin ce qui fait que deux hommes ne se peuvent communiquer leurs pensées sans mouvements, c’est qu’ils ont des corps, & que l’un ne peut être averti par l’autre, qu’à l’occasion des mouvements du corps, auquel son âme est unie. Mais, si l’on suppose que l’un des esprits n’a point de corps, il pourra se rendre présent
————— 6
Cordemoy’s work was translated into English in the year 1668 itself with the convenient title of A Philosophicall Discourse Concerning Speech, Conformable to the Cartesian Principles, in order to point out the philosophical tradition of Cordemoy’s account on speech.
Metaphor and truth in Rationalism and Romanticism
57
par ses pensées mêmes à celui qui aura un corps, comme à celui qui n’en aura pas; & réciproquement l’esprit qui sera uni à un corps, pourra sans le secours de la voix, exprimer ses pensées à tout esprit, qui n’aura pas de corps. (Cordemoy 1968: 249–50) [But next, what hath been said of the Communication of the two meer Spirits, ought to be said of the commerce that may be betwixt a Spirit united to a Body, and one that is not. For certainty what incapacitates two men to communicate their thoughts to one another without motions, is, that they have Bodies, and that the one cannot be advertis’d by the other but by the motions occasione’d by the Body, to which the Soul is united: But supposing that one of the Spirits have no Body, it is capable to render it self preset by its very thoughts to that which hath a Body, as it doth to that which is destitute of a Body; and reciprocally that Spirit, which is united to a Body, will be able, without the intervention of the Voice, to express its thoughts to every Spirit that is Body-less (Cordemoy 1668: 111; original capitals and italics)].
The communication problems that humans have are rooted in the fact that they need the body to communicate with other minds. Ideally, communication should only be amongst minds via the transfer of pure thoughts or what Roy Harris called ‘telementation’ (Harris 1981). However, this is not possible because we are both body and mind and the bodily part of us influences our thoughts. This influence of the body on the mind manifests itself especially through the passions which, as we shall see, really muddy the waters of pure thought transmission from mind to mind. 1.2 Ideal versus actual communication As a result of their conception of language, rationalists dreamt of an ideal type of communication in which the body would have no part. But this was only an ‘ideal’. Even rationalists had to acknowledge that ‘real’ communication functioned very differently. So, while still clinging to their ideal of perfect mental communication, they also began to denounce the defects and imperfections of real communication. Ideally, communication should be one in which the natural order of ideas was directly expressed. Rationalists believed that there was a ‘natural’ order in thought and that this order could be expressed in language. They also held that there was a natural order in the world, and that this order could be known by human beings and expressed by means of laws. The natural order of ideas was believed to be both semantic and syntactic in nature. So, if human beings want to achieve the order of ideas, their language should reproduce the logical, semantic and syntactic order of their thoughts. The criterion for being sure whether they are not wrong in the sphere of ideas is that they conceive them clearly and distinctly. When ideas are conceived clearly and distinctly, they must be expressed in a language which is logically correct as well as free of hyperbatons in the sphere of syntax, and ambiguities and tropes in the sphere of semantics. Words, in turn, should have ‘clear meanings’ according to the ‘simple ideas’ in our minds.
58
Pedro José Chamizo Domínguez / Brigitte Nerlich
We have seen that rationalists had two conceptions of communication, an ideal one and a real one. Communication was either the direct expression of our thoughts in their natural and logical order or else an imperfect shadow of this ideal. In parallel with this two-fold conception of communication they also developed a two-fold conception of language, again an ideal one and a real one. To provide communication with the ideal means for the direct and orderly expression of thought, they dreamt of a universal and rational language, based on logical principles, which was, in turn, opposed to the real natural languages spoken around the globe. Et si quelqu’un avait bien expliqué quelles sont les idées simples qui sont en l’imagination des hommes, desquelles se compose tout ce qu’ils pensent, et que cela fût reçu par tout le monde, j’oserais espérer ensuite une langue universelle fort aisée à apprendre, à prononcer et à écrire, et ce qui est le principal, qui aiderait au jugement, lui présentant si distinctement toutes choses qu’il lui serait presque impossible de se tromper ; au lieu que tout au rebours, les mots que nous avons n’ont quasi que des significations confuses, auxquelles l’esprit des hommes s’étant accoutumé de longue main, cela est cause qu’il n’entend presque rien parfaitement. (Lettre à Mersenne, November, 20th 1629, in: Descartes 1973: I, 81) [“And if someone had properly explained which are the simple ideas, which are in men’s imagination, and from which all they think is composed; and if everybody understood that; then I would dare to hope for a universal language which would be very easy to learn, pronounce, and write; and, most importantly, which would help judgement in so far as it would represent all things so distinctly that it would be almost impossible to make an error; instead of which, conversely almost all the words which we possess have confused meanings, to which the human mind has grown accustomed over a long period of time; which means that it is almost impossible to understand anything perfectly.” Our translation].
The result of the acquisition of that universal language could be that les paysans pourraient mieux juger de la vérité des choses, que ne font maintenant les philosophes. (Ibid.: 81–2) [“peasants would be better judges of the truth of things than philosophers are nowadays.” Our translation].
Ideal communication via an ideal universal language would give direct access to truth – even to peasants; or, even more easily to a peasant whose pure thoughts have not yet been contaminated by the ravages of centuries of civilised language. Nevertheless, Descartes himself was somewhat sceptic as to whether such a language would be possible. For that reason he advised Marin Mersenne, a theologian, philosopher, mathematician and music theorist and author of La Vérité des sciences contre les sceptiques et les pyrrhoniens (Truth of the Sciences against the Sceptics and Pyrrhonians, 1624) with whom Descartes corresponded, in the end not to pursue this idea and he concluded his famous letter to Mersenne by saying that a universal language could only exist ‘in the land of fiction’:
Metaphor and truth in Rationalism and Romanticism
59
Mais n’espérez pas de la voir jamais en usage; cela présuppose de grands changements en l’ordre des choses, et il faudrait que tout le monde ne fût qu’un paradis terrestre, ce qui n’est bon à proposer que dans le pays des romans. (Ibid.: 82) [“But don’t expect it ever to be used; this presupposes wide changes in the order of things, and it would be necessary that the entire world were the heaven on earth, which one cannot conveniently propose except in the realm of novels.” Our translation].
Having come to see that this ideal universal language could only exist in a fictional paradise, rationalists of the Port Royal school of thought began to argue that, instead, natural language should be cleansed of all tropes and hyperbatons, especially when philosophical thoughts are expressed, that is when we speak about ‘purely speculative matters’ (Arnauld/Nicole 1981: I, 96). That means that, since the supposed ideal language looks almost impossible, natural language should be purified in such a way that it at least approximates the ideal language. So, figures (both semantic and syntactic ones)7 should be avoided, and this for several reasons. Firstly, they should be avoided because the thoughts, emotions, and passions of the speakers get mixed up with each other when figures of speech are used, and, as a result, the hearers will receive all three instead of ‘the utterly naked truth’, i.e. the cognitive (propositional) contents of the speaker’s utterances: C’est encore par là qu’on peut reconnaître la différence du style simple & et du style figuré, & pourquoi les mêmes pensées nous paraissent beaucoup plus vives quand elles sont exprimées par une figure, que si elles étaient renfermées dans des expressions toutes simples. Car cela vient de ce que les expressions figurées signifient outre la chose principale, le mouvement & et la passion de celui qui parle, & impriment ainsi l’une & l’autre idée dans l’esprit, au lieu que l’expression simple ne marque que la vérité toute nue. (Ibid.: I, XIV, 96) [“And it is also in this way that one can recognise the difference between the plain style and the figurative
————— 7
The fact that humans can, from early childhood, identify the meanings that underlie the different tropes and hyperbatons is used by Cordemoy as an argument in favour of innatism. The hypothesis of innatism is proven, then, by the fact that two to three year old children are able to understand the meanings of the different figures and grammatical constructions, even when adults do not wish them to understand the meanings of these constructions: “Mais, quelque peine qu’on se donne pour leur apprendre certaines choses, on s’aperçoit souvent qu’ils savent les noms de mille autres choses, qu’on n’a point eu dessein de leur montrer. Et ce qu’il y a de plus surprenant en cela, c’est de voir, lors qu’ils ont deux ou trois ans, que par la seule force de leur attention, ils soient capables de démêler dans toutes les constructions qu’on fait en parlant d’une même chose, le nom qu’on donne à cette chose.” (Cordemoy 1968: 213–14) [But in taking pains to teach them certain things, we often perceive, that they know the names of a thousand other things, which we desseigned not to shew them: And what is most surprising therein, is, to see, when they are 2. or 3. years of age, that by the sole force of their attention they are capable to find out in all the constructions which are made in speaking of one and the same thing, the name, we given to that thing. (Cordemoy 1668: 30–1)].
60
Pedro José Chamizo Domínguez / Brigitte Nerlich one, and why the same thoughts appear to use to be more lively when they are expressed by means of a figure [of speech] than if they were enclose in completely plain expressions. For this derives from the fact that figurative expressions mean, in addition to the main thing, the speaker’s motion and passion, and thus stamp both ideas on the spirit, while the plain expression does not show but the very plain truth.” Our translation].
Secondly, they should be avoided because the use of figures of speech moves us and stirs up our emotions. Now this is something that should be entirely avoided in the expression of philosophical thoughts. To be moved on this occasion would be similar to having mental convulsions: Mais comme le style figuré signifie ordinairement avec les choses les mouvements que nous ressentons en les concevant & en parlant, on peut juger par là de l’usage que l’on en doit faire; & quels sont les sujets auxquels il est propre. Il est visible qu’il est ridicule de s’en servir dans les matières purement spéculatives, que l’on regarde d’un œil tranquille, & qui ne produisent aucun mouvement dans l’esprit. Car puisque les figures expriment les mouvements de notre âme, celles que l’on mêle en des sujets où l’âme ne s’émeut point, sont des mouvements contre nature, & des espèces de convulsions. (Ibid.) [“But since the figurative style usually signifies together with the things represented the motions we experience when we conceive them and speak about them, one can see what use one should make thereof and for which topics this style is the appropriate one. It is clear that it would be ridiculous to use it for purely speculative matters, which are gazed upon with a clear eye, and which do not produce any motion in the spirit. For, as figures express the motions of our soul, it would lead to motions against nature and to, what one could call, convulsions, if one mixed them in with the representation of topics that don’t move the soul.” Our translation].
And thirdly, they should be avoided because the use of figures of speech might make us accept something as true which is entirely false, just because it comes in a beautiful package, so to speak. Entre les causes qui nous engagent dans l’erreur par un faux éclat qui nous empêche de la reconnaître, on peut mettre avec raison une certaine éloquence pompeuse & magnifique, que Cicéron appelle abundantem sonantibus verbis uberibusque sententiis. Car il est étrange, combien un faux raisonnement se coule doucement dans la suite d’une période qui remplit bien l’oreille, ou d’une figure qui nous surprend, & qui nous amuse à la regarder. (Ibid.: III, XX, b 2, 277) [“Among the causes that induce us to make a mistake because a false brightness prevents us from seeing it, one can include with good reason a certain pompous and magnificent eloquence, which Cicero calls abundantem sonantibus verbis uberibusque sententiis.8 For it is strange how false reasoning follows almost imperceptibly a nicely turned phrase which pleases the ear, or a figure that surprises or amuses us.” Our translation].
————— 8
Tusculanarum Disputationum, I, XXVI: “Full of resonant words and copious sentences”.
Metaphor and truth in Rationalism and Romanticism
61
It was generally thought that passions disrupt the natural order of our ideas, and that, likewise, figures of speech disrupt the natural order of speech. Body was opposed to mind, emotion was opposed to cognition, and the figurative was opposed to the literal. In philosophical discourse the only things that mattered were mind, cognition and the literal expression thereof. The other three were troublemakers. They stood between the philosophers and their search for ‘truth’ and the scientific subjugation of Nature. Philosophers were therefore urged to use language ‘properly’, which meant using words in the literal sense, or ‘au sens propre’, uncontaminated by figurative uses or abuses (see Nerlich/Chamizo Domínguez 2003). And with regard to syntax words were supposed to be used according to the natural order which is the one of ideas. Unfortunately, none of the natural languages came anywhere close to this ideal, although it was believed that the French language came very close, as is explicitly asserted in the last paragraph of Port-Royal’s Grammar. J’ajouterai seulement qu’il n’y a guère de Langue qui use moins de ces figures que la notre: parce qu’elle aime particulièrement la netteté, & à exprimer les choses autant qu’il se peut, dans l’ordre le plus naturel & le plus désembarrassé, quoi qu’en même temps elle ne cède à aucune en beauté ni en élégance. (Arnauld & Lancelot, 1966: XXIV, 160) [“I will only add that there is hardly any language that uses these figures less than ours: as it particularly loves clarity and to explain things as far as one can, according to the most natural and unencumbered order; however, this does not deprive it of either beauty or elegance.” Our translation].9
In short, the majority of rationalists, just like the majority of empiricists, regarded figures of speech as cognitive and linguistic vices. On a considéré, par exemple, en ce qui regarde la Rhétorique, que le secours qu’on en pouvait tirer pour trouver des pensées, des expressions, & des embellissements, n’était pas si considérable. L’esprit fournit assez de pensées, l’usage donne les expressions; & pour les figures & les ornements, on n’en a toujours que trop. Ainsi tout consiste presque à s’éloigner de certaines mauvaises manières d’écrire & de parler, & surtout d’un style artificiel & rhétoricien composé de pensées fausses &
————— 9
The same idea about the French language resurfaced from time to time. Antoine de Rivarol could therefore publish a treatise in 1784 entitled De l’universalité de la langue françoise. He declared that “Ce qui n’est pas clair n’est pas français” [Whatever is not clear is not French. Our translation]. Rivarol tried to prove his thesis about the universality and clarity of the French language by saying that it was the only language spoken in Western Europe that adhered to the direct order, that is to say that the subject-verb-object order is favoured in most types of French sentences. A similar opinion has been found (and ironically quoted) by Wittgenstein in the 20th century: “This case is similar to the one in which someone imagines that one could not think a sentence with the remarkable word order of German or Latin just as it stands. One first has to think it, and then one arranges the words in that queer order. (A French politician once wrote that it was a peculiarity of the French language that in it words occur in the order in which one thinks them.” (Wittgenstein 1958: §336). He might have been referring to Rivarol!
62
Pedro José Chamizo Domínguez / Brigitte Nerlich hyperboliques & de figures forcées, qui est le plus grand de tous les vices. (Arnauld/Nicole, 1981: Second Discours, 29) [“For Example, as to what concerns Rhetoric, we consider’d that there is little advantage to be drawn from that Art, for the finding out Thoughts, Expressions and Embellishments. Outwit furnishes us with Thoughts; Use affords us Expression, and for Figures and Ornaments they are many times superfluous, so that all the Benefit from thence consists in avoiding certain evil Habits of Writing and Speaking, especially and Artificial and Rhetorical stile compos’d of false Imaginations, Hyperboles, and forc’d Figures, the most unpardonable of all Vices in an Orator.” Arnauld/Nicole, 1693: 24–5; original capitals and italics]
From the end of the Renaissance onwards10 rhetoric had lost its appeal and was gradually sidelined by ‘serious’ philosophers, because, after all, “Ceux qui ont le raisonnement le plus fort, et qui digèrent le mieux leurs pensées, afin de les rendre claires et intelligibles, peuvent toujours le mieux persuader ce qu’ils proposent, encore qu’ils ne parlassent que bas-breton, et qu’ils n’eussent jamais appris de rhétorique.” (Discours de la Méthode, in: Descartes 1973: VI, 7) [“Those who have the strongest faculty of reason, and who most skilfully direct their thoughts with a view to render them clear and intelligible, are always best able to persuade others of the truth of what they propose, though they should speak only in the language of Lower Brittany, and be wholly ignorant of the rules of rhetoric.” Our translation] 2. Between rationalism and romanticism This view that plain truth even when spoken in dialect is more persuasive than rhetoric and that, conversely, rhetoric only obfuscates and obscures truth was not universally accepted however. Gottfried Wilhelm Leibniz discussed the pros and cons or the vice and virtues of using rhetoric to persuade listeners of the truth of certain propositions and assumed a position between Descartes and one of Descartes’ major opponents on this issue, namely Vico, to whom we shall come next. Leibniz writes in his Nouveaux essais sur l’entendement humain (which was a reply to Locke’s Essay on Human Understanding): PHILALÈTHE. J’avais pensé conclure, mais je me souviens du septième et dernier abus, qui est celui des termes figurés ou des allusions. Cependant on aura de la peine à le croire abus, parce ce qu’on appelle esprit et imagination est mieux re-
————— 10
See, for example, the following text from Michel de Montaigne’s Essais: “Voire mais, que fera-il [the student] si on le presse de la subtilité sophistique de quelque syllogisme: le jambon fait boire, le boire desaltere, parquoy le jambon desaltere? Qu’il s’en mocque. Il est plus subtil de s’en mocquer que d’y respondre.” (1962: I, XXVI, 170) [But what will become of our young gentleman, if he be attacked with the sophistic subtlety of some syllogism? ‘A Westfalia ham makes a man drink; drink quenches thirst: ergo a Westfalia ham quenches thirst.’ Why, let him laugh at it; it will be more discretion to do so, than to go about to answer it (Montaigne 2006)].
Metaphor and truth in Rationalism and Romanticism
63
çu que la vérité toute sèche. Cela va bien dans les discours, où on ne cherche qu’à plaire; mais dans le fond, excepté l’ordre et la netteté, tout l’art de la rhétorique, toutes ces applications artificielles et figurées des mots ne servent qu’à insinuer des fausses idées, émouvoir les passions et séduire le jugement, de sorte que ce ne sont que de pures supercheries, cependant c’est à cet art fallacieux qu’on donné le premier rang et les récompenses. C’est que les hommes ne se soucient guère de la vérité, et aiment beaucoup à tromper et être trompés. Cela est si vrai que je ne doute pas que ce que je viens de dire contre cet art ne soit regardé comme l’effet d’une extrême audace. Car l’éloquence, semblable au beau sexe, a des charmes trop puissants pour qu’on puisse être admis à s’y opposer.11 THÉOPHILE. Bien loin de blâmer contre votre zèle pour la vérité, je le trouve juste. Et il serait à souhaiter qu’il put toucher. Je n’en désespère pas entièrement, parce qu’il semble, Monsieur, que vous combattez l’éloquence par ses propres armes, et que vous en avez même une d’une autre espèce, supérieure à cette trompeuse, comme il y avait une Venus Uranie, mère du divin Amour, devant laquelle cette autre Vénus bâtarde, mère d’un Amour aveugle, n’osait paraître avec son enfant aux yeux bandés. Mais cela même prouve que votre thèse a besoin de quelque modération, et que certains ornements de l’éloquence sont comme les vases des Égyptiens, dont on se pouvait servir au culte du vrai Dieu. Il en est comme de la peinture et de la musique, dont on abuse et dont l’une représente souvent des imaginations grotesques et même nuisibles, et l’autre amollit le cœur, et toutes deux amusent vainement; mais elles peuvent être employées utilement, l’une pour rendre la vérité claire, l’autre pour la rendre touchante, et ce dernier effet doit être aussi celui de la poésie, qui tient de la rhétorique et de la musique. (Leibniz, 1966: III, X, 34, 305–6; original Italics) [“PHILALETHES: I had meant to stop, but I remember a seventh and last abuse, which is that of figurative terms or allusions. Yet this will be hardly regarded as an abuse, because what is called wit and imagination is more easily accepted than the dry truth. This is acceptable in discourses where the aim is merely to please; but really all the art of rhetoric, besides order and clarity, all these artificial and figurative uses of words, serve only to insinuate wrong ideas, move the passions and mislead the judgement; and so indeed are nothing but perfect cheats. Yet it is this defective art which is highly regarded and given the highest esteem. The fact is that men hardly care about truth and they greatly love to deceive and be deceived. So true is this that what I have said against this art will no doubt be regarded as a great boldness on my part; for eloquence, like the fair sex, has such charms which make it almost impossible to oppose it. THEOPHILUS: Far from disapproving of your zeal for the truth, I find it very proper. And it is to be hoped that it may have some effect. I do not entirely despair of that; because, sir, you seem to be fighting eloquence by means of its own weapons, having at your command an eloquence which is superior to the deceptive kind you are attacking. In the same way there was a Uranian Venus, the mother of divine Love, in whose presence that other bastard Venus, the mother of a blind love, did not dare to appear with her blindfolded child. But that very fact proves that your thesis needs to be moderated and that cer-
————— 11
This echoes the following quote from Locke: “Tis evident how much Men love to deceive, and be deceived, since Rhetorick, that powerful instrument of Error and Deceit, has its established Professors, is publickly taught, and has always been had in great Reputation: And, I doubt not, but it will be thought great boldness, if not brutality in me, to have said thus much against it. Eloquence, like the fair Sex, has too prevailing Beauties in it, to suffer it self ever to be spoken against. And ’tis vain to find fault with those Arts of Deceiving, wherein Men find pleasure to be Deceived.” (Locke 1975: III, chapter X).
64
Pedro José Chamizo Domínguez / Brigitte Nerlich tain ornaments of eloquence are like the Egyptians’ vases, which could be used in the worship of the true God. Painting and music are similarly misused: the former is often used to depict fantasies which are grotesque and even harmful, the latter mollifies the heart, the amusement they provide are futile; but they can be usefully employed, one to make the truth vivid and the other to make it affecting; which should also be the effect of poetry, which involves both rhetoric and music.” Our translation].12
Having immersed himself not only in abstract philosophical speculations but also in concrete etymological studies, Leibniz refused to believe that words should or could only have clear and transparent (literal) meanings. He pointed out: The clarity of a word arises from two factors – either from the word in itself or from its context in speech. The clarity of a word in itself, again has two sources – origin and usage. The origin of a word, finally, can be resolved into two factors – the use of the root and the analogy of the derivation made from the root. Usage is the meaning of a word known in common by all who use the same language. Analogy is a meaning reached by shifting, or by derivation, which is likewise known to all who use the same language. For example, the usage or meaning of the word fate is the necessity of events. [...] Mostly, too, usage has arisen from origin by a certain figure of speech. This appears in the given example, since fatum is originally the same as dictum but means in usage what will happen necessarily. Let us see, therefore, whose dictum will happen necessarily; it is manifest that God’s commands alone fit this description. Thus by origin fate is dictum, then by antonomasia or par excellence, the dictum of God, then by synecdoche the dictum of God concerning the future, or the decree of God, and finally by the metonymy of cause, what will happen necessarily, which is the present usage of the word. (Leibniz, Preface to an Edition of Nizolius 1956[1670]: 188; quoted in: Schmitz 1985: 247)
The evolution of language shows that we cannot escape the evils of metaphor, that it is in fact only by ‘virtue’ of figures of speech that language develops and can keep up with the progress of thought. In this sense language is always ‘imperfect’. Were it to become a ‘perfect’ mirror of thought and stayed that way, both language and thought would asphyxiate or, indeed, collapse. 3. Romanticism and beyond But as we said earlier on, one of the clearest opponents of Descartes’ views on metaphor and truth was Giambattista Vico (see Bréhier, 1981, 325). Vico doesn’t deny that ‘clear ideas’ have their uses, but these are rather limited, confined mainly to mathematics and physics – anywhere else, he argues, clarity and distinction are the vice of human reason, rather than its virtue (ibid.) These are sciences which, according to Vico the mind has constructed ————— 12
This text has not been included in (Leibniz 1996).
Metaphor and truth in Rationalism and Romanticism
65
itself, according to a clear and orderly plan, so to speak. Everywhere else clarity and distinction are however, the vice of reason rather than its virtue (ibid.) A clear idea is a finite idea, but: “de ma souffrance, par exemple, je ne puis saisir la forme et la limite; la perception que j’en ai est infinie, et cette infinité témoigne de la grandeur de la nature humaine” [“of my pain, for example, I cannot understand the form or the limit; the perception I have of it is infinite and that infinity demonstrates the grandeur of human nature.” Our translation] (quoted by Bréhier 1981: 325). Vico’s work, his New Science, written in 1725, was largely forgotten during the Enlightenment when philosophers and cultural scientists alike tried to model their methods on the glorious natural sciences and their search for ‘objective truth’. However, his ideas were later taken up by Herder, Goethe, Gerber, Biese and many other German romantic and post-romantic philosophers looking for a method of cultural understanding that would go beyond empiricist methodologies and beyond ‘pure reason’. As Gordon Scott has pointed out, Vico was one of the first to make a distinction between the methods of the historical and the methods of the natural sciences, a distinction resurrected in the 19th century as the distinction between Verstehen and Erklären as distinctive methods used by the Geisteswissenschaften and the Naturwissenschaften, and in the 20th century under the titles of ‘qualitative’ vs ‘quantitative’ methods in the social sciences (see Nerlich, 2004). In Vico’s view, to search for ‘the truth’ in physics is something quite different from searching for the ‘truth’ in history. Only God knows history in its totality, because he made the world, but the specific events of history can be understood by the human intellect [...] because they are made by human actions. The historian shares the quality of humanity with those men, great and small, whose actions create the phenomena of history. This enables him to enter inside the historical process, thus achieving a subjective understanding of it that is more profound than the objective knowledge attainable by the natural scientist, who is compelled to remain outside the phenomena he studies. [...] by this reasoning Vico made the bold claim that one can discover laws of social development that are more certain even than the laws of physics. In calling his book New Science (1725) he meant to contend that the science of history furnishes the most precise and mot irrefutable form of human knowledge. Historians, and other social scientists, have no need to consider themselves inferior to natural scientists, or to try to mimic their methodology, for a categorically different, and superior, methodology is available to them. (Scott 1991: 405)
But Vico’s New Science not only laid out a new methodology for the social sciences, it also proposed “a theory of primitive concept formation and metaphor” (Leezenberg 2001: 56) that has some similarities with Lakoff and Johnson’s (1980) cognitive semantics “which claims that extensive systems of metaphorical concepts derived from bodily experience underlie much of our everyday language use” (ibid.: 65). Vico developed a cyclical three-step theory of the history of nations who are supposed to go through a first, divine, period in which a language of ges-
66
Pedro José Chamizo Domínguez / Brigitte Nerlich
tures is developed; then follows an heroic period in which language is primarily symbolic and people speak in ‘poetic characters’; finally, follows a third period of civil society in which the language has become conventionalised and prosaic. During the heroic age people think in ‘imaginative universals’ (generi fantasici), which are mainly based on projecting bodily and emotional experiences onto unknown inanimate entities. Abstract thought and abstract categorisation have not yet developed. The images that are used in language and thought might appear ‘figurative’ to modern readers, but they were the ‘proper’ language of that time. Only once language is conventionalised in the third period, can one speak of the use of ‘metaphors’ as opposed to the proper or literal usage of certain words; only then can one speak of figurative language. One can however find corollaries between this early ‘poetic logic’ and more modern ‘figures of speech’, of which Vico distinguishes four main ones: metaphor, metonymy, synecdoche and irony: The most luminous figure, and hence the most basic and common, is metaphor. Metaphor is especially prized when, by the metaphysics just described, it confers sense and emotion on insensate objects. The first poets attributed to physical bodies the being of animate substances, endowed with limited powers of sense and emotion like their own. In this way, they created myths about them; and every such metaphor is a miniature myth. This gives us the criterion for dating the origin of metaphors in various languages. For example, all metaphors based on analogies between physical objects and the products of abstract thought must date from and age in which philosophies were just beginning to take shape. We find proof of this in the fact that in every language the terms used in the fine arts and advanced sciences are of rustic origin. Noteworthy too is the fact that in all languages most expressions for inanimate objects employ metaphors derived from the human body and its parts, or from human senses and emotions. Thus, we say head for top or beginning; front or brow, and shoulders or back, for before and behind; [...]. (Vico 1999: 159–60)
After many more examples of this kind, Vico points out: All this follows from Axiom 1: ‘In his ignorance, man makes himself the measure of the universe.’ And in the examples cited, man has reduced the entire world to his own body. Now rational metaphysics teaches us that man becomes all things through understanding, homo intelligendo fit omnia. But with perhaps greater truth, this imaginative metaphysics shows that man becomes all things by not understanding, homo non intelligendo fit omnia. For when man understands, he extends his mind to comprehend things; but when he does not understand, he makes them out of himself and, by transforming himself, becomes them. (Ibid.: 160)
For Vico figures of speech are neither deviant nor merely decorative. They are the natural form of thought and speech. All figures of speech [...] were previously thought to be the ingenious inventions of writers. But my discussion of them proves that they were in fact necessary modes of expression in all the early poetic nations, and originally had natural and proper meanings. These expressions became figurative only later, as the human
Metaphor and truth in Rationalism and Romanticism
67
mind developed and invented words which signified abstract forms, that is, generic categories, comprising various species, or relating parts to a whole. Knowing this, we may begin to demolish two common errors of the grammarians: that prose is the proper form of speech, and poetic speech improper; and that men spoke first in prose and later in verse. (Ibid.: 162)
Jean-Jacques Rousseau echoed Vico when he wrote in his essay on the origin of language: Comme les premiers motifs qui firent parler l’homme furent des passions, ses premières expressions furent les Tropes. Le langage figuré fut le premier à naitre, le sens propre fut trouvé le dernier. On n’appelle les choses de leur vrai nom que quand on les vit sous leur véritable forme. D’abord on ne parla qu’en poësie; on ne s’avisa de raisonner que longtemps après. (...) L’image illusoire offerte par la passion se montrant la première, le langage qui lui répondoit fut aussi le premier inventé; il devint ensuite métaphorique quand l’esprit éclairé reconnoisant sa première erreur n’en employa les expressions que dans les mêmes passions que l’avoient produite. (Rousseau 1995: III, 381–2) [“As man’s first motives for speaking were of the passions, his first expressions were tropes. Figurative language was the first to be born. Proper meaning was discovered last. One calls things by their true name only when one sees them in their true form. At first only poetry was spoken; there was no hint of reasoning until much later. (…) The illusory image presented by passion is the first to appear, and the language, that corresponded to it was also the first invented. It subsequently became metaphorical when the enlightened spirit, recognizing its first error, used the expressions only with those passions that had produced them.” (Rousseau 1986: 13–4)].
And Rousseau was certainly not the only one who, after Vico, turned the received wisdom about metaphor on its head. Vico’s view of metaphor has permeated thinking about figurative language from Goethe to Eco. But what is more, many philosophers of language, such as Gustav Gerber, Friedrich Max Müller, and Alfred Biese picked up Vico’s distinction between the underlying poetic or imagistic (bildliche) nature of all language, and the later use of figures of speech in ordinary language and literature, what Müller called the distinction between radical and poetical metaphor (Müller 1877 [1864]: 388). It was only when the stage of radical metaphor gave place to the stage of poetical metaphor that it made any sense to apply the distinction between the ‘literal’ (or conventional) and ‘figurative’ (unconventional). At the end of the 19th century the literary critic and philosopher Biese drew all the strands of thoughts about metaphor together which had emerged after Vico. Amongst many others he discusses the works of Vico himself (Biese 1893, 8) (who had rejected the literal/figurative distinction and regarded metaphor as a necessary form of intuition),13 Johann Heinrich Lambert ————— 13
About Vico he writes: “[...] er begreift, dass das Metaphorische kein äusserlicher Schmuck, keine Fiktion ist, sondern eine notwendige Form der Anschauung, dass in der Vermenschlichung alles Gegenständlichen die Urpoesie der Völker, der Mythos und die Sprachschöpfung ihren Quellpunkt haben.” (Ibid.: 10) [“He
68
Pedro José Chamizo Domínguez / Brigitte Nerlich
(ibid.: 113) (who had claimed that metaphor is an important tool in constructing the architecture of a language)14, Johann Wolfgang von Goethe (ibid.: 16 and passim) (who had said that language anthropomorphises reason and reality, and that reasoning and categorisation are inherently embodied), Jean Paul (ibid.: 16) (who had written that “jede Sprache [ist] in Rücksicht geistiger Bezeichnungen ein Wörterbuch erblasseter Metaphern” [every language is [...] but a dictionary of pallid metaphors] (Jean Paul 1962– 1977[1804]: 184]), Arthur Schopenhauer (ibid.: 104), (who had argued that all primitive thinking is carried out in pictures) and Gerber (ibid.: 13 and 23), with whom Biese agreed that it is fundamentally wrong to say that literal meaning is different from non-literal meaning on the basis that the latter is ‘figurative’. Summarising what he had learned from all these philosophers of metaphor, Biese declared: “Die Metapher ist [...] kein poetischer Tropus, sondern eine ursprüngliche Anschauungsform des Denkens.” [“Metaphor is (…) is not a poetic trope but an original form of cognitive intuition.” Our translation.] Metaphor is “das eigentliche innerste Schema des Menschengeistes” [“the original inner schema of the human mind”. Our translation] (Biese 1893: VI). During the 19th century it became more and more clear that, as Ortony (1979) was to say a century later, metaphor is not only nice, but necessary and this not only for the expression of emotions but also for the expression of thought and truth. For the post-Lockian rhetorician and philosopher Benjamin Humphrey Smart, for example, metaphors are essential parts of the original structure of language; and however they may sometimes serve the purpose of falsehood, they are on most occasions, indispensable to the effective communication of truth. It is only by [these] expedients that mind can unfold itself to mind; – language is made up of them; there is no such thing as an express and direct image of thought. (Smart 1831: 210; our italics)
For him, as for many after him, tropes and figures of speech became “the original texture of language, and that from which whatever is now plain at first arose. All words are originally tropes; that is expressions turned [...] from their first purpose, and extended to others” (ibid.: 214). Some, like Friedrich Nietzsche, radicalised this view and tried to reduce truth to metaphor. One does not have to go quite so far, but one certainly should think about Nietzsche’s views on the evolution of language and concepts, as Umberto Eco has done in his book on language and cognition, en————––
14
realises that the metaphorical is not just external decoration or fiction, but that, instead it is a necessary form of intuition, that the primordial poems of all people, that myths and the creation of language all have as their origin the anthropomorphisation of concrete objects”. Our translation]. One should stress in this context that Lambert discussed metaphor in the context of a book entitled: Neues Organon oder Gedanken über die Erforschung und Bezeichnung des Wahren und dessen Unterscheidung vom Irrthum und Schein (1764).
Metaphor and truth in Rationalism and Romanticism
69
titled Kant and the Platypus (Eco 1999). He shows how Nietzsche developed a new theory of concepts, according to which concepts are but metaphors gone stale. Quoting Nietzsche, Eco writes: Since nature has thrown away the key, the intellect plays on fictions that it calls truth, or systems of concepts, based on the legislation of language. [...]: we think we talk about (and know) trees, colors, snow, and flowers, but they are metaphors that do not correspond to the original essences. Every word becomes concept as its pallid universality takes the color out of the differences between fundamentally unequal things: thus we think that in correspondence with the multiplicity of individual leaves there exists a primordial “leaf” on “the model of which all leaves have supposedly been woven, drawn, circumscribed, colored, wrinkled, and painted -- but by a clumsy hand -- in such a way that no exemplar would seem to be correct and reliable as a faithful copy of the original shape” [Nietzsche 1895 [1873]: 360). (Eco 1999: 44; our italics).
The debate about knowledge, truth and metaphor, about rationalism, empiricism and romanticism continues today in a heated exchange between intellectual heavy-weights Stephen Pinker and George Lakoff (Pinker 2006; Lakoff 2006). 4. Conclusion We have come a long way from the ‘clear’ and ‘distinct’ ideas of a Descartes or even Locke which were seen as such important vehicles for the communication of ‘truth’. [...] In fact, the truth is a poetically elaborated “mobile army of metaphors, metonymies, and anthropomorphisms” that subsequently gel into knowledge, “illusions whose illusory nature has been forgotten,” coins whose image has been worn away and are taken into consideration only as metal; so we become accustomed to lying according to convention, in a style that is binding for everyone, placing our actions under the control of abstractions, and having reduced the metaphors to schemata and concepts. Thence a pyramidal order of castes and ranks, laws and delimitations, constructed entirely by language, an immense “Roman columbarium,” the graveyard of intuition. (Ibid.: 45; our italics)
It is obvious that such views would sooner or later be disputed and provoke a backlash. This backlash came in the form neo-positivists, truth-conditional semanticists, and transformational grammarians. Metaphor was again seen as a vice of reason and a deviation in language. Words were again supposed to be linked to well-defined (literal) meanings which picked up well-defined states of affairs in the world and the link between the two was the bearer of truth. However, this backlash against metaphor provoked its own backlash in turn in the form of neo-romanticism, namely ‘cognitive semantics’. It is a shame that the proponents of this new view of metaphor as a cognitive and
70
Pedro José Chamizo Domínguez / Brigitte Nerlich
linguistic virtue had in the meantime lost all knowledge of their 19th-century ancestors. Like them they believe that the radical distinction between the literal and metaphoric in grammar and semantics has to be replaced by the view that language (and thought) are metaphorical through and through. In their seminal book Philosophy in the Flesh, Lakoff and Johnson wanted to show that the laws of thought are based on metaphor not logic and that truth itself is a metaphorical construction. Pace rationalism, they claim that: Eliminating metaphor would eliminate philosophy. Without a very large range of conceptual metaphors, philosophy could not get off the ground. The metaphoric character of philosophy is not unique to philosophical thought. It is true of all abstract human thought, especially science. Conceptual metaphor is what makes most abstract thought possible. Not only can it not be avoided, but it is not something to be lamented. On the contrary, it is the very means by which we are able to make sense of our experience. Conceptual metaphor is one of the greatest of our intellectual gifts. (Lakoff/Johnson 1999: 129).
References Aristotle: Poetics, in: Aristotle in 23 Volumes. Vol. 23. Cambridge [MA]: Harvard University Press 1932. Arnauld, Antoine, Claude Lancelot: The Art of Speaking. 1696, article available at: http://0-eebo.chadwyck.com.jabega.uma.es/search/full_rec?ACTION=ByID& SOURCE=pgimages.cfg&ID=V44467 (accessed 11 July, 2008). – Grammaire générale et raisonnée. Ed. by H. E. Brekle. Stuttgart–Bad Cannstatt: F. Frommann 1966 (third, facsimile edition, 1676). Arnauld, Antoine, Pierre Nicole: Logic, or the Art of Thinking. 1693. Article available at: http://0-eebo.chadwyck.com.jabega.uma.es/search/full_rec?ACTION=ByID& SOURCE=pgimages.cfg&ID=V46037 (accessed 11 July, 2008). – La logique ou l’art de penser. Ed. by Pierre Claire/François Girbal. Paris: J.Vrin. 1981. Biese, Alfred: Die Philosophie des Metaphorischen. In Grundlinien dargestellt. Hamburg–Leipzig: Leopold Voss 1893. Bréhier, Émile: Histoire de la philosophie IXe–XXe siècle. Paris: Presses Universitaires de France 1981. Bustos Guadaño, Eduardo de: La metáfora. Ensayos transdisciplinares. Madrid: Fondo de Cultura Económica 2000. Cassirer, Ernst: Das Erkenntnisproblem in der Philosophie und Wissenschaft der neueren Zeit. 2 Vols. Berlin: Bruno Cassirer 1906–1907. Cordemoy, Géraud de: A Philosophicall Discourse Concerning Speech, Conformable to the Cartesian Principles, 1668, article available at http://eebo.chadwyck.com/ search/full_rec?SOURCE=pgimages.cfg&ACTION=ByID&ID=V61559&FILE= &SEARCHSCREEN=param (accessed 10 July, 2008.) – Discours physique de la parole, in: Œuvres Philosophiques. Édition critique par Pierre Clair et François Girbal. Paris: Presses Universitaires de France 1968. (First edition 1668). Davidson, Donald: “What metaphors mean”, in: Inquiries into Truth and Interpretation. Oxford: Clarendon Press 1984, 245–64 [1978].
Metaphor and truth in Rationalism and Romanticism
71
Descartes, René: Œuvres. Ed. by Ch. Adam and P. Tannery. Paris: C.N.R.S.-J. Vrin 1973. Du Marsais, César Chesneau de: Des tropes ou des différens sens dans lesquels on peut prendre un même mot dans une même langue. Paris: chez David 1757. (First edition 1730). Eco, Umberto: Kant and the Platypus: Essays on Language and Cognition. London: Secker & Warburg 1999. Gerber, Gustav: Die Sprache als Kunst. 2 vols. Bromberg: Mittler’sche Buchhandlung. 1871–74. – Die Sprache und das Erkennen. Berlin: Gaertner 1884. – Die Sprache als Kunst. First vol. (Second enlarged edition). Berlin: Gaertner. 1885. (First edition 1871). Grice, Herbert Paul: “Logic and conversation”, in: Studies in the Way of Words. Cambridge, Mass.: Harvard University Press 1989, 22–40 [1975]. Harris, Roy: The Language Myth. New York: St. Martin’s Press 1981. Hobbes, Thomas: Leviathan. Introduction by A. D. Lindsay. New York: E. P. Dutton & Co. 1950. (First edition 1651). Jean Paul [Johann Paul Friedrich Richter]: “Vorschule der Ästhetik”, in: Werke. Edited by Norbert Miller. Abt. 1, Bd. 5. Darmstadt: Wissenschaftliche Buchgesellschaft 1962–1977, 7–330. (First edition 1804). Johnson, Mark: “Introduction”, in: Mark Johnson (ed.), Philosophical Perspectives on Metaphor. Minneapolis: The University of Minnesota Press 1985, 3–47. Lakoff, George: “A response to Stephen Pinker”, 2006, article available at: http://www.powells.com/biblio?show=HARDCOVER:USED:9780374158286:13. 95&page=authorsnote#page (accessed 18 July, 2008.) – , Mark Johnson: Metaphors We Live by. Chicago: Chicago University Press 1980. – Philosophy in the Flesh. New York: Basic Books 1999. Lambert, Johann Heinrich: Neues Organon, oder Gedanken über die Erforschung und Bezeichnung des Wahren und dessen Unterscheidung vom Irrthum und Schein. 2 vols., Leipzig: Wendler 1764.(Repr. Hildesheim: Olms, 1965–69). Leezenberg, Michiel: Contexts of Metaphor. Amsterdam: Elsevier 2001. Leibniz, Gottfried Wilhelm: Nouveaux essais sur l’entendement humain. Ed. By Jacques Brunschvig. Paris: Garnier-Flammarion 1966. (First edition 1765). – New Essays on Human Understanding. Translated and edited by Peter Remnant and Jonathan Bennett. Cambridge: Cambridge University Press 1996. Locke, John: Essay on Human Understanding, ed. by P. H. Nidditch. Oxford: Oxford th University Press 1975. (First edition 1689; 5 edition 1706). Maréchal, Jean: Le point de départ de la métaphysique. Paris: Desclée de Brouwer 1949, 2nd. ed. Montaigne, Michel de: Essais, in: Œuvres Complètes. Textes établis par Albert Thibuadet et Maurice Rat. Introduction et notes par Maurice Rat. Paris: Gallimard [1580] 1962. – The Essays. Translated by Charles Cotton, 2006, article available at http://www.gutenberg.org/dirs/3/6/0/3600/3600.txt (accessed 14 July, 2008). Müller, Friedrich Max: The Science of Language. Ninth ed, in: 2 vols. London: Longmans, Green, and Co. 1877. (First edition 1864). Nerlich, Brigitte: “Coming full (hermeneutic) circle: The debate about methods in psychology”, in: Mixing Methods in Psychology, edited by Zazie Todd, Brigitte Nerlich, Suzanne McKoweon, and David D. Clarke. London: Routledge 2004, 17– 36.
72 –
Pedro José Chamizo Domínguez / Brigitte Nerlich
, David Clarke: “Mind, meaning, and metaphor: The philosophy and psychology of metaphor in nineteenth-century Germany”, in: History of the Human Sciences 14:2 (2001), 39–61. – , Pedro J. Chamizo Domínguez: “The use of literally: Vice or virtue”, in: Annual Review of Cognitive Linguistics 1 (2003), 193–206. Nietzsche, Friedrich: Ueber Wahrheit und Lüge im aussermoralischen Sinne. Grossoktav-Ausgabe. Leipzig: C. G. Naumann 1895. (First edition 1873). Ortega y Gasset, José: Ideas y creencias, 1934, in: Obras Completas. Vol. V. Madrid: Alianza-Revista de Occidente 1983, 377–409. Ortony, Antony: “Beyond literal similarity”, in: Psychological Review 86 (1979), 161–80. Pascal, Blaise: Pensées, in: Œuvres Complètes. Ed. by Jacques Chevalier. Paris: Gallimard [1660] 1976. – Pensées. Translated by W. F. Trotter, 2005, article available at http:// www.ccel.org/ccel/pascal/pensees.i.html (Accessed 11 July, 2008.) Pinker, Stephen: “Block that metaphor”, in: The New Republic (2006), article available at: http://pinker.wjh.harvard.edu/articles/media/2006_09_30_thenewrepublic. html (accessed 18 July, 2008.) Rivarol, Antoine de: De l’universalité de la langue françoise. Berlin–Paris: Bailly 1784. Rousseau, Jean Jacques: Essai sur l’origine des langues où il est parlé de la mélodie et de l’imitation musicale (texte établi et annoté par Jean Starobinski), in: Œuvres Complètes. Vol. V. Ed. by Bernard Gagnebin, Marcel Raymond et al. Paris: Gallimard 1995. (First published 1781). Rousseau, Jean Jacques and Johann Gottfried Herder: On the Origin of Language. Translated by John H. Moran and Alexander Gode. Chicago: Chicago University Press 1986. Schmitz, H. Walter: “Die durchgängige Tropisierung der Sprache. Über einen Aspekt von ‘Zeichen im Wandel’”, in: Historiographia Semioticae: Studien zur Rekonstruktion der Theorie und Geschichte der Semiotik, ed. by Klaus D. Dutz andPeter Schmitter. Münster: MAkS Publikationen 1985, 241–70. Smart, Benjamin H[umphrey]: An Outline of Sematology: Or an Essay towards Establishing a New Theory of Grammar, Logic, and Rhetoric. London: Richardson 1831. (Published anonymously) Vico, Giambattista: The New Science of Giambattista Vico (Rev. ed.; T.G. Bergin and M.H. Fisch, eds. and trans.). Ithaca, NY: Cornell University Press 1948. (First published 1725). Wittgenstein, Ludwig: Philosophical Investigations. Translated by G.E.M. Anscombe. Oxford: Blackwell 1958.
Emmanuelle Danblon (Brussel, B)
Persuasion: between trope and truth
0. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
Introduction Truth and persuasion A short history of truth Truth and metaphor The birth of fictionality Ontogenetic development From semantic skills to rhetoric functions Working on the notions Literal and figurative meaning A stratified rationality Shared fiction and persuasion The ritual of rhetoric
0. Introduction Around 350 B.C., in Athens, a young courtesan named Phryne was charged with impiety and had to appear before the court. The orator Hyperides conducted the case for the defence. As he noticed that his discourse has no persuasive power on the judges’ mind, he decided1 to lift her tunique, revealing thus her nudity to the assembly; according to the legend, this gesture persuaded the judges of Phryne’s innocence. In this view, Phryne’s beauty – which was supposed to equal that of Aphrodites herself – is the evident sign of her innocence. This is, obviously, a parabole about the power of persuasion, and especially about the links between truth and persuasion. In this paper, I will try to better understand the paraoxical links between truth and persuasion, i.e. the difficulty, for the modern mind to admit that an evidence feeling does not necessarily implies anything about truth. Indeed, in Phryne’s myth, we all have a firm intuition as to the psychological effect such as we may imagine what Phryne’s beauty alleged by triggered on the assembly. But, at the same time, we would be at a loss to explain this feeling, and, to express literally the relation which could hold between beauty and innocence, i.e. between an aesthetic judgement and an ethical judgement: ??“She is beautiful, so she is innocent”, ??“Smart people are generally good”, ??“Ugli—————
1
According to another version of this legend, Phryne accomplished this act by herself; this detail is not crucial for our claim.
74
Emmanuelle Danblon
ness is a sign of guilt”, etc. Everyone has the strong feeling – at least in our modern societies – that such principles do not provide any suitable ground for moral standards, and especially, for juridical rules. Yet, this imagery keeps an extraordinary psychological power, which is plainly operative in fiction, where beauty and ugliness enable us to discriminate between the pure and young heroin and the bad witch. Of course, if those feelings were devoid of any factual content, this would mean that we simply cannot ground our rules and norms, including truth; which would lead us to some form of ethical and epistemological relativism. In such a view, we should not mind to know whether Phryne was really innocent: what does matter is the fact that she was acquitted, thanks to her beauty. As the Sophist Gorgias said: “Truth is nothing, but beauty is persuasive, thanks to seduction”. 1. Truth and persuasion As it is well known, the Sophists’ interest in understanding the causes of persuasion was harshly condemned by Plato. According to him, indeed, every rhetorical discourse, in order to reach truth, has to refrain from appealing to emotion. This is in harmony with the ultimate moral of Phryne’s legend; after the acquittal, a decree forbade any further appeal to pity – i.e. any further argumentum ad misericordiam – in trials that defendant would be allowed to be exposed to the judges’ view. For his part, Aristotle held to an ethical conception of rhetoric where persuasion could be seen as the very sign of truth, inasmuch as men are naturally attracted by truth. In contrast to Plato, he had a rather optimistic confidence in our everyday intuitions. However, both Plato and Aristotle implicitely conceive of as an epistemological criterion. While truth is an exact correspondance to reality, the verisimilitude produced by persuasion appears as a kind of pseudo-truth (Plato) or quasi-truth (Aristotle): For, in fact, the true and that which resembles it come under the purview of the same faculty, and at the same time men have a sufficient natural capacity for the truth and indeed in most cases attain to it; wherefore one who divines well in regard to the truth will also be able to divine well in regard to probabilities. (Rhet, I, 1, 1355a)
Our modern conception of epistemology strongly dissociates truth from persuasion. The former is a semantic criterion that we use as a regulatory principle when we attempt to say something about the world; the latter is a psychological effect which is still not clearly understood. In this article, I will try to shed some light on persuasion, by studying it from a rhetorical perspective, thanks to some contemporary linguistic and cognitive theories. This will lead us to consider the concept of truth in a diachronic perspective, which could
Persuasion: between trope and truth
75
help us to capture its complexity, especially in rhetoric2. As we will see, this reflections are strongly related to the of fiction and tropes. 2. A short history of truth Let us first turn to the history of truth. This will help us to better understand our “anthropological” relation with this notion. Indeed, it seems that, through the centuries, the concept of truth, which is probably dependent on a frame of thinking (see Detienne 1967), evolved parallel to rationality and cognitive skills. The first notion conveied by the Greek word aletheia during the “prehistory” of philosophy was intimately linked with poetry, with the sung and rhythmic speech of oral tradition. As they were inseparable from memory, oral cognitive skills required the use of suitable linguistic techniques: poets did not compose with words, but with structural formulas that had to be lexically filled afterwards; Homeric poetry provides a splendid example of this. Since inspired poets held records of the whole community, it followed that only some happy few retained the memory of ancient times. The twofold function of the poet – praising immortal heroes or brave warriors and blaming enemies – corresponded to the highest political power in these societies. Therefore, in Archaic Greece, poetical an political speech was one and only thing. Praise was sometimes qualified as etumos which means “genuine”, “real” or “true”; but the words for “accuracy” or “precision” in the knowledge of a poem or a ritual were alethes and aletheia. Truth and authenticity thus played a complemetary role: since poetical truth escaped both debate and proof, aletheia did not amount to a correspondance between a proposition and a state of affairs, but to an agreement between one judgement and other judgements. Accordingly, in the oldest conceptions of truth, all the attributes encompassed in this concept bear on discourse performance (See Detienne 1967: 51). This probably stems from the fact that an oral society does not reason on language as an object and, therefore, does not develop any semantic metalanguage3. In other words, the initial concept concept of truth pertained to the language of reality, and turned into a semantic notion in a second time only. The main feature of speech in oral tradition is efficiency: thanks to the magic of words, poetical praise or blame and oraculary discourse give rise to reality; speech does not reflect events but appears as one of the conditions that have to be fulfilled for events to occur. As in our modern performative speech acts (see Searle and Vanderveken 1985, Searle 1995), the creation of reality and the enunciation of truth go hand in hand. This absolute efficiency of speech precludes any theoretical distinction between an assertive and a —————
2 3
About the proposition to see in the rhetoric model from a genealogic viewpoint, see Danblon (2002). About the features of the oral mode of thouhgt by comparison with literacy, see Olson (1996), Goody (1984), Donald (1991).
76
Emmanuelle Danblon
declarative speech act. Quite logically, such a truth-in-act assigns a crucial role to persuasion – in Greek Peitho. Truth – aletheia – needs Peitho in order to reach the audience’s assent. 3. Truth and metaphor Those features are also underlined by Rodrigo (1998) emphasizes that, while modern tradition rests on an articulation between language and truth, we still have not totally broken with the archaïc concept of truth. Now, I would like to suggest that this archaic concept presupposes a mode of thought that is either “iconic” or “indexical” mode of thought – to use a Peircean terminology. Let me explain this point more preciesely. According to Rodrigo (1998), both mimesis and metaphor are, in Aristotle’s view, “paradoxical institutions of truth” in that they present facts in their absence – which relies on an iconic mode of representation. However, for Aristotle the poet is a legend maker, which incites Ricœur (1975) to translate the Greek muthos by the French expression “mise en intrigue” rather than “récit”; this choice of Ricœur’s reveals the indexical/actional dimension of mimesis. In any case, metaphor is usually conceived of as a sharing more features with iconic and indexical signs than with symbolic signs. In other words, Understanding metaphor and trope apparently amounts to understanding the language of persuasion without taking any account of its symbolic dimension in the Peircean serie. By contrast, I will assume that the use of metaphor as a metaphor requires a reflective consciousness of language, and consequently the full semiotic capacity of Peircean symbolic sign. In order to support this claim, let us first go back to the turning point at which this new attitude towards language may have emerged. According to some authors (Donald 1991, Vernant 1985), the transition from a mythical to a theoretical or modern thought proved crucial for the evolution of rationality. Vernant claims that this transition occurred, around the 7th and 6th centuries B.C., within a frame where social and cognitive evolution convergently produced a secularization of speech. This event correlates with the emergence of rhetoric and philosophy on the one hand, and with the emergence of law and history on the other hand. At the same time, language becomes an autonomous object of investigation. From now on, the idea that a “performance” may have a magical creative power will be interpreted in metaphorical terms. Poetry begins to exist as poetry and looses its religious and sacred dimension. As a consequence, the concept of truth progressively separates from the criterion of “evidence”, and thus from inspiration. This happens when writing as a public technical tool makes it possible for a new appraisal of language to emerge.
Persuasion: between trope and truth
77
4. The birth of fictionality The evidence available supports the hypothesis that historiography and the notion of a fiction appeared simultaneously in the history of thought. According to Schaeffer (1999, 2000) the aesthetic attitude relies on extremely complex abilities that emerged very late in the evolution of mankind. Schaeffer claims that the ultimate origin of fiction may be found in lures, which cause organisms to perceive one thing “as if” it were another thing. Natural lures – e.g., homochromy in some butterflies – are evolutionarily older than Intentional ones (see also Baudonnière 1997); yet, they provoke something like a “false belief”4 in the receptor – e.g., in the case of butterflies, in a potential predator. Thus we may use this primitive model to describe higher order situations. A delusion – e.g. a white lie – consists in acting intentionally in order to create the illusion that one thing is another thing. Now, fiction also proceeds “as if” one thing was another thing, but – and this point is crucial – without any deceptive intention; on the contrary, fiction is essentially grounded on the mutual knowledge of the fictional intention. Now, according to Schaeffer again, what made it possible for fiction to emerge as a cultural acquisition is a capacity to dissociate a mimetic operator from a vital function. Being in fictional immersion means being in a “divided mental state” (Schaeffer 1999: 192): one sees an object while knowing that it is a “mimeme”. This implies a stratification between a pre-attentional state and an attentional treatment of the representations at hand: fictional immersion presupposes the efficiency of pre-attentional lures, but it excludes any state of illusion – and therefore, any intent of delusion – at a conscious level of beliefs. With a real lure or a real delusion, beliefs too are deluded as happens in the case of ordinary illusion. Moreover, as underlined by Olson (1996, see also Gombrich 1971), everyone may have the experiment of a divided mental state: if you stick a needle in someone’s eye on a photograph, you necessarily experience a negative reaction, while knowing that the real person is not injured at all. This model may help us to characterize more precisely well-known anthropological data which show that members of traditional societies may entertain “contradictory beliefs”, e.g., when they admit the ubiquity of ghosts, while assuming that ubiquity is physically impossible.5 As we will see later, the distinction I have just introduced are necessary to capture the essence of rhetoric and persuasion, without indulging in Plato’s condemnation: persuasion may be efficient without being manipulatory.
————— 4
5
Of course, I do not want to assign “beliefs” to organisms like butterflies. I am just trying to underline the presence, from this stage on, of something which is more complex than transparent perception. This example comes from Boyer (1990).
78
Emmanuelle Danblon
5. Ontological development This short historical survey gains some supports from ontogenetic studies in the emergence of language and cognition. If we assume that ontogenesis sums up – at least in part – the phylogenetic evolution, (Donald 1991, Tomasello 1999, Lyons 1995), it is generally admitted that the ability to interpret figures of speech like metaphors only emerges around the age of 7 or 8, i.e. when children begin to use written language (see, e.g. Olson 1996, Bruner 1996, Colston/Gibbs 2002). Thus, in the ontogenesis too, literacy seems to involve a new mode of thought, thanks to which people acquire the capacity to have a distant look at language. In this respect, Olson (1996) reports an experiment carried out by Luria which showed that litterate children are able to perform inferences based on premisses devoid of any intuitive empirical content, while illiterate children cannot. More precisely, when confronted with a syllogism, illiterate people show a propensity to ground their responses on the factual content of the utterance, while literate people assess the premisses validity of the reasoning independently of the content. This led Luria to assume that literacy brings logical thought. From his own part, Olson supposes that literacy brings about the awareness of logical thought, even though, without literate consciousness, logical skills are used in an actional mode. Besides producing these logical and argumentative abilities, literacy gives rise to a gap between language and the world that allows one to exploit the “semantic plurality of language” (Dominicy 1989) in order to develop new rhetorical tools, viz. to distinguish between a literal and a figurative meaning, between saying and meaning, or the letter and the spirit of an utterance. Pre-literate children have difficulties in understanding a figure of discourse like irony, surely because of their inability to discriminate between “what is said” and “what is meant” (Olson 1996, Colston/Gibbs 2002). 6. From semantic skills to rhetoric functions The awareness of the conventional nature of language, and the resultant gap between language and the world, provide the foundations of both a modern (semantic) conception of truth and a flexible vision where the “semantic plurality of language” instead of being considered as a weakness, opens for human cognition the way to a new set of skills. It is important to capture more precisely the implications of such an acquisition, since “the myth of linguistic adequacy” remains vivid in human modes of thought. Indeed, if linguistic adequacy is taken as an epistemological criterion, it follows that in an ideal use of language, utterances should be evident, i.e. undisputable, and that argumentation reduces to poor substitute for demonstrative proof. For instance, the rationality of scientific or juridical statements, should be modelled on that of logical and mathematical utterances, where notions are assumed to be univocal. Apart from the fact that the
Persuasion: between trope and truth
79
semantics of natural language does not work with univocal notions, such an ideal of clarity and precision leads to a reductionist view of language use and argumentation. This point was emphasized by Perelman and OlbrechtsTyteca (1969), who claimed that the fuzziness of notions ensures the very possibility of the argumentative process. Indeed, the participants generally agree on a fuzzy notion in a first time. But very often, very often a need of clarification arises within the process itself because the implementation of the notion at hand leads to some incompatibilities. Notice that clarification does not matter per se, since a clarification process has no absolute endpoint. What does matter, in fact, is that the range of the notion be restricted in order to allow decision to take place. Hence, the clarification process is not exhaustible. Obviously, such a strategy relies on a mature conception of language and thought, where people accept conventionality without trying to go back to a (mythic) adequacy of speech. This point has been emphasized by Perelman, who showed that what made an extended agreement possible on the foundamental charters of our modern times was the very fact that their notions could be differently interpreted in various value systems. 7. Working on the notions Let us now analyze more precisely the way in which the range of fuzzy notions will be restricted during the argumentative process. This will help us to underline the intimate link between this critical process and the rhetorical techniques of persuasion. In Perelman’s view, clarifying a notion consists in dissociating two aspects, within this very notion: the first one which is labelled “apparent” will be ruled out; the second one, which labelled as “real”, will be retained so as to lead us to a conclusion. For example, which in the notion of equality, one may dissociate the “equality by right” from the “equality de facto”, and thus create a notional hierarchy that will determine the ideological and political nature of the conclusion. As I argued elsewhere (Danblon 2002), it turns out when analyzing argumentative discourses, that the relevant dissociation is that between the “letter” and the “spirit” of the notion (see also Goyet 1996). Indeed, since notions, far from being a simple reflection of reality, rely on our conventional representations of it, they cannot be qualified as “apparent” or “real”. Perhaps Perelman was the victim, at this level, of the “myth of linguistic adequacy”. However, while the dissociation technique rests on our awareness and manipulation of the distance between language and thought, it remains rhetorically efficient to present the product of dissociation in terms of “apparence” and “reality”, because a persuasion effect will be better obtained when the orator uses a specific vocabulary, characterized by iconicity and evidence. As we will see later, rhetoric effects frequently stem from the orator’s capacity of going back – even fictionally – to an ancient iconic relation between language and thought.
80
Emmanuelle Danblon
8. Literal and figurative meaning Conventionality thus assigns to argumentation a new set of rational functions which bring into play the same semantic skills as those that are used in metaphor and, more widely, in figurative meaning. Ever since Plato, metaphor has been condemned because of its essential inadequacy to the facts, while some authors – e.g. Hegel – have acknowledged its crucial importantce for the expression of mental or sensory ideas (see Cooper 1986). Anyway, both positions stick to a conception of metaphor where the expression is, in Peircian terms, iconical or indexical, without any symbolic dimension. In fact, since the very beginning of the reflection on language, the tense relationships between metaphor and truth have always involved iconicity, or the lack of it, as a central issue. Our short history of truth helps us now to better understand this situation. Indeed, the epistemological deadlock that the notion of metaphorical truth has reached may be overcome since if the iconicity of metaphor is put at its right place in the genealogy of representations, i.e., not at some orginial place but at the point where the semantic “plurality of language” has become plainly functional. In this respect, Cooper (1986) points out that, as soon as we distinguish between literal meaning and metaphoric (or figurative) meaning, we can separate truth from representation and thus admit that metaphors have a heuristic power, even if they have to be criticized afterwards. In our modern conception of truth, metaphor may be thought of as a symbolic expression that use the icon “as if” it were the one and only possible way to express an idea. Cooper emphasizes that metaphor are neither true nor false: even though we may appreciate the capacity of a given metaphor to provide an adequate representation of facts, this does not mean that we assess it as true. Davidson claim that metaphor is “seeing one thing as another thing” instead of “seeing that P”. Now, the “seeing as” phrase captures the iconic dimension of metaphor, which always refers to another object without refereing to a proposition, as is the case when one “sees that P”. Of course, this iconicity is presented “as if” it were independent of any symbolic sign, and this does not amount to producing a lure but to creating a poetical and conventional fiction, based on our capacity to dissociate between the letter and the spirit of a word, between saying and meaning. 9. A stratified rationality At this stage, we may try to understand more precisely the links between truth and persuasion. The historical perspective that I am adopting here will allow me to put forth a general hypothesis about the stratification of our modern rationality. At a phylogenetic level, Merlin Donald (1991) distinguishes between three stages – more precisely, three transitions – characterized by specific linguistic uses, each of which led to the emergence of specific skills in human rationality.
Persuasion: between trope and truth
81
The following chart shows how Donald’s model can be integrated to the framework developed in this paper (see also Danblon 2002): Transition
Cognitive skills
Epistemological Semiotic mode mode Mimesis Immediate per- Sensory eviden- Icon ception, rhythm, ce mime Oral language Actional model, Myth, oral and Index narration actional concept of truth Written language Inferential rea- Science, seman- Symbol soning, critique tic concept of truth In order to properly interpret this chart, two additional assumptions should be made. The first claims that we possess a synchronic memory of our cognitive evolution in that this evolution progressed by accumulating the different transitions distinguished rather than by eliminating the older ones. Second, such an cumulative mode of evolution, instead of relegating old capacities to some kind of primitive or old-fashioned form of mind, triggered the emergence of new, more complex, capacities thanks to the reflexive consciousness acquired of the difference between language and the world. Actually, we may assume that rhetoric sprang from this cognitive and cultural evolution. 10. Shared fiction and persuasion Let us remember that, according to Schaeffer, fictional immersion presupposes the efficiency of pre-attentional lures while excluding any state of illusion at the conscious level of beliefs. This is exactly what happens when we are moved by a fiction. We do not believe that the events represented happen in the actual world, but our emotion is actual, exactly “as if” we were really experiencing those events. Pleasure plays a central role in our reception of fiction, but since we can escapefrom the domain of primary functions, it becomes independent of any type of truth-conditional relation to reality. Schaeffer underlines, moreover, that the development of the imaginative capacity is inversely proportional to an actional effect. In other words, thanks to this ability, the immediate reaction gives way to a fictional immersion posture that involves a reduction of the psychological tension because it neutralizes the reactive loop between the individual and the surrounding reality. Fiction thus allows us to inhibit reactions that would normally be triggered. Significantly enough, it appears that children who have difficulties in integra-
82
Emmanuelle Danblon
ting the fictional process display a strong propensity to respond to stimulations by acting, whereas children who do not resort to an “as if” play. This means that by entertaining a divided mental state, we fictionalize our representations, realizing them in an “as if” mode. Analogically, rhetoric persuasion can be viewed as an actual and strong emotive effect based on this fictional ability to act and react “as if” we went back to our ancient rationality, where the main epistemological mode was truth-in-action, where saying was acting. Sonce we ritualize this evident and narrative truth thanks to our imaginative fictional capacity, it follows that, pace Plato, persuasion, far from being reprehensible in se relies on a shared ability to play with representations.6 11. The ritual of rhetoric Finally, let us turn to the very mechanism that will operate in rhetoric thanks to this stratification of rationality. Recall that the rhetorical institution began to exist when our consciousness of the conventionality of language appeared both to argumentative reasoning and to fiction. Thus, we may assume that rhetoric combines these two abilities within a specific ritual that I will try to describe now. As we know, Aristotelian rhetoric divides into three genres: deliberative, forensic and epideictic. It is commonly admitted that the epideictic has a particular status within this a rhetorical institution: it seems to be the only genre to avoid argumentative reasoning in its discursive realizations. Epideictic speeches, mainly rely on “amplification”, i.e., on a peculiar way of “describing” events in a poetical mode. This feature gives rises to an apparent paradox: while pertaining to rhetoric, amplification does not correspond to any argumentative technique which would amount to a reasoning, with an argument, a justification and a conclusion. Yet, we are now able to account for this puzzling situation. Since criticism and fiction appeared simultaneously as human cognitive skills, one may argue that they were both implied by a new semantic function which comes to grips with the gap between language and thought. Now, criticism is the privileged argumentative method in both the deliberative and the forensic genres, where the main strategy at work consists in dissociating notions in order to rule out inconsistencies and to reach a decision. But, at the ————— 6
Note that there is no question of ethical innocence here; of course rhetoric can be be manipulative, when it uses this fictional capacity to induce persuasion in order to reach covert aims. But this phenomenon has nothing to do with persuasion in se. The same holds for the difference between white lie and irony: both are grounded on our capacity to dissociate from language and reality, but, while the former results from a covert strategy, the latter is produced in accordance with an overt fictional strategy.
Persuasion: between trope and truth
83
same time, this dissociation technique crucially reveals the conventional nature of notions and, hence, their potential weakness as warrants for juridical and political decision. Epideictic discourse typically aims at restoring concord between citizens: indeed, amplification creates a shared fiction, “as if” we went back to our ancient rationality where everything were evident, where people had no reflexive consciousness of the gap between words and facts, between trope and truth. More precisely, it turns out that the dissociation technique is the exact converse of amplification whereas dissociation limits the range of a notion through a distinction between the letter and the spirit a of law or rule, amplification makes use of poetical tools – rhythm, rhymes, repetitions and other tropic(al) techniques – in order to suppress any conflicting components within the notion at hand, “as if” this notion still possessed the evidence at hand in our ancient conception of truth. Thus, the rhetorical institution includes a ritual mechanism that makes room for a deep critical attitude by compensating it with the help of a formal shared fiction that allow us to emotionally ground the most important values of our community. While criticism crucially works on a syntagmatic line, amplification operates on a paradigmatic series, thanks to parallelism and repetition. This shows that amplification favours an iconic way of expression, “as if” each instance referred directly to another instance, without ever referring to a drastically different state of affairs, as it happens with symbolic signs. Perhaps this is may be why such a persuasive effect –which may evoke “sublime” – is such as one has the feeling that “that is it”. Conversely, a dissociation movement will lead the audience to judge that “that is true”. The former is a judgement in terms of iconical adequation to facts, the later is a judgement in terms of symbolic (and hence, conventional) correspondance with facts. But the iconic judgement is fictional, i.e., it has no influence on the beliefs that has been created in the critique process, which never prevents from provoking a deep, collective emotion: this is persuasion. References Aristotle: Complete works, ed. by J. Barnes. Princeton (N.Y.): Princeton University Press (Bollingen Series, 71) 1985. Baudonnière, Pierre-Marie: Le mimétisme et l’imitation. Paris: Flammarion 1997. Boyer, Pascal: Tradition as Truth and Communication. A Cognitive Description of Traditional Discourse. Cambridge: Cambridge University Press 1990 Bruner, Jerome: “Frames for thinking. Ways of making meaning”, in: Olson, David R. (ed.): Modes of Thought. Cambridge: Cambridge University Press 1996, 93–105. Colston, Herbert, L., Raymond W. Gibbs: “Are Irony and Metaphor Understood Differently ?”, in: Metaphor and Symbol, 17, 1 (2002), 57–80. Cooper, David, E.: Metaphor. Oxford–New York: Blackwell 1986. Danblon, Emmanuelle: Rhétorique et rationalité. Essai sur l’émergence de la critique et de la persuasion. Éditions de l’Université de Bruxelles, Collection “Philosophie et Société”. 2002.
84
Emmanuelle Danblon
Detienne, Marcel: Les maîtres de vérité. Paris: François Maspero 1967. Dominicy, Marc: “De la pluralité sémantique du langage. Rhétorique et poétique”, in: Poétique, n° 80 (1989), 499–514. – “Sur l’épistémologie de la poétique”, in: Histoire, Epistémologie, Langage 13, I (1991), 151–74. Donald, Merlin: Origins of the Modern Mind. Three Stages in the Evolution of Culture and Cognition. Cambridge, Mass.–London: Harvard University Press 1991. Goody, Jack: The logic of Writing and the Organization of Society. Cambridge–New York: Cambridge University Press 1986. Gombrich, Ernst H.: L’art et l’illusion. Psychologie de la représentation picturale, traduction par Guy Durand. Paris: Gallimard 1971. Goyet, Francis: Le sublime du lieu commun. L’invention de la rhétorique dans l’Antiquité et à la Renaissance. Paris: Champion 1996. Havelock, Eric, A.: Preface to Plato. Oxford: Blackwell 1963. Lyons, William: Approaches to Intentionality. Oxford: Clarendon Press 1995. Olson, David, R. (ed.): Modes of Thought. Cambridge: Cambridge University Press 1996a. – The World on Paper. The Conceptual and Cognitive Implications of Writing and Reading. Cambridge: Cambridge University Press 1996b. Perelman, Chaïm/Olbrechts-Tyteca, Lucie: The New Rhetoric. A Treatise on Argumentation. Notre Dame, (Ind.)–London: University of Notre Dame Press 1969. Ricœur, Paul: La métaphore vive. Paris: Seuil 1975. Riffaterre, Michael: Fictional Truth. London: The John Hopkins University Press 1990. Rodrigo, Pierre: “Mimesis et métaphore chez Aristote: deux institutions paradoxales de la vérité”. En torno a Aristoteles. Homaje al Profesor Pierre Aubenque. Université de Santiago de Compostela 1998, 269–83. Schaeffer, Jean-Marie: Pourquoi la fiction ? Paris: Seuil 1999. Searle, John R.: The Construction of Social Reality. New York: The Free Press 1995. – , Daniel Vanderveken: Foundations of Illocutionary Logic. Cambridge: Cambridge University Press 1985. Tomasello, Michael: The Cultural Origins of Human Cognition. Cambridge, Mass.– London: Harvard University Press 1999. Vernant, Jean-Pierre: Mythe et pensée chez les Grecs. Paris: Maspero 1985.
John A. Barnden / Alan M. Wallington (Birmingham, GB)
Metaphor and its unparalleled meaning and truth
0. 1. 1.1 1.2 1.3 1.4 1.5 2. 2.1 2.2 2.3 3. 3.1
Introduction Some assumptions, terminology and preliminary observations Restricted aim Concerning discourse participants and meanings Concerning consciousness Concerning the two sides in metaphor Concerning source/target parallelism Non-parallelism (Non-)parallelism in some existing accounts Conversational cat-flaps Other examples Non-parallelism, source-scenario holism and metaphorical meaning Sub-sentential holism: avoiding separate metaphorical meanings for constituents 3.2 Holism beyond sentence boundaries 4. Restrictions on inference in interpretation 4.1 Target-side, source-side and literal-case inferencing 4.1.1 The case of mapping-based accounts 4.1.2 The case of non-mapping-based accounts 4.2 Bringing a generator or believer into the picture 5. Combining the consequences; and truth 6. Summary 0. Introduction This article arises indirectly out of the development of a particular approach, called ATT-Meta, to the understanding of some types of metaphorical utterance. Information on the theoretical approach and the implemented computer program that partially realizes it can be found in Barnden (2001a,b, 2006a,b, 2008), Barnden, Glasbey et al. (2004), Barnden and Lee (2001) and Lee and Barnden (2001). However, the specifics of the approach are not the focus of the present article, which concentrates on some general issues that have informed, or arisen from, the development of the approach. The article connects those issues to the questions of metaphorical meaning and truth. A large part of the exploration of metaphor in fields such as Cognitive Linguistics and natural language Pragmatics takes metaphor to rest on com-
86
John A. Barnden / Alan M. Wallington
plex mappings between a target subject matter and a source subject matter (see, e.g., Lakoff 1993, 2008). This is one main way in which some aspects of the source, generally including some structural aspects, can have parallels in the target. It is typical for much, and perhaps the whole, of the meaning of a metaphorical expression to be explained by means of these parallels. We will talk about “parallelism” rather than “isomorphism”, or even “analogy”, because “isomorphism” implies a strict one-to-one correspondence between items in the source scenario and items in the target scenario, and analogy theorists often hold that analogy rests on isomorphism. We wish not to prejudge the question of whether looser, messier forms of parallelism are sometimes needed. Also, we will talk about parallels rather than mappings, because as explained below, our discussion will embrace the question of parallelism that may be discernible, albeit implicitly, in accounts of metaphor that are not presented as being based on mappings. Three main cases of such accounts are Relevance Theory (RT: Sperber & Wilson 2008, Wilson & Carston 2006; also Vega Moreno 2007; going back to prior RT proposals about metaphor in Sperber & Wilson 1986, 1995), the categorization-based or “class-inclusion” approach (Glucksberg 2008, Glucksberg & Keysar 1990), and Ritchie’s CLST account: (Context-Limited Simulation Theory: Ritchie 2006). One main issue we will address is that of substantial non-parallelism that can exist between source and target subject-matters in metaphor: more precisely, the issue that there may be source aspects that are exploited by the utterance and have a deep effect on metaphorical interpretation but that do not themselves have a natural parallel in the target subject-matter (although we will point out that artificial parallels can generally be created). This is part of a broader point that the content of metaphorical discourse should often be seen as being derived in a rather holistic way from several metaphorical bits of the discourse, which conspire to describe some source-domain scenario, rather than being derived by putting together metaphorical meanings of each metaphorical bit (even if each bit could in principle be assigned its own separate metaphorical meaning). As a result, we suggest that it is misguided to think of the propositions making up the content (or, if you like, “meaning”) about the target scenario that is being described is a matter of metaphorical meaning of specific grammatical units such as sentences, clauses or other constituents that a traditional semantic theory would assign propositional meaning to. Rather, grammatical units (that are to be taken metaphorically) have meanings in source-domain terms; content in target-domain terms is derived from the source-domain scenario depicted by those units and fleshed out through inference; and the target-domain content is only (in general) fuzzily relatable to particular grammatical units. This view is on a spectrum at whose extreme point we could place Davidson’s (1984) view that metaphors only have literal (i.e., source-domain) meanings, with other effects on the understander not being a matter of propositional content. However, on our view it is proper to take metaphorical discourse as having non-literal meaning, couched as a collection of propositions, among other things possibly; it is just
Metaphor and its unparalleled meaning and truth
87
that the propositions are not to be thought of (in general) as meaning of specific sentences, clauses or other grammatical units rather than of a possibly only fuzzily delineated piece of discourse. A second main concern of the article is the question of how much, and what types of, inference are involved in the derivation of metaphorical meaning. While our own ATT-Meta approach accords with, amongst other accounts, Relevance Theory (RT) as to the centrality of inference in metaphor interpretation, we disagree with RT on its claim that metaphor interpretation is a matter of (a relatively high degree of) concept “broadening” and (often) “narrowing”. We argue that almost all inference can in fact be theoretically redescribed as concept broadening and/or narrowing, and what is left over should be included in the metaphor interpretation process anyway. Thus, the RT claim about broadening and narrowing really just says that inference, of unrestricted type, is involved in metaphor understanding, in accord with what ATT-Meta claims. In the course of the argument, and now in no way in contrast to RT, we stress that there is no a priori limit to the amount of inference that might be done in the course of deriving metaphorical meaning. This is not to say that the amount of reasoning is always or even usually large, but just that the matter cannot be prescriptively circumscribed. The plan of the article is as follows. Section 1 states some background assumptions and terminology. Section 2 summarizes some prominent ways in which non-parallelism has appeared in metaphor theory, and goes on with further illustration and analysis of the extent and type of non-parallelism that can exist. Section 3 explores some consequences of non-parallelism for the question of meaning (we leave truth till later in the article). In this section we argue the point mentioned above about metaphorical content arising holistically from discourse and not a matter of specific grammatical units. Section 4 turns to the second main concern, namely the amount and types of inference allowed during metaphor interpretation. Section 5 puts the considerations of the previous two sections together in a mutually reinforcing way. Section 6 concludes. 1. Some assumptions, terminology and preliminary observations 1.1 Restricted aim It is fair to say that all accounts of meaning and truth of metaphor are provisional and programmatic. For instance, the special issue of the Mind and Language journal on metaphor in 2006 (vol. 21 no. 3) contains and alludes to intense debate on: how much of metaphor interpretation is “semantic” as opposed to “pragmatic”; on “what is said” by a metaphorical utterance; on the relationship of metaphorical meanings of utterances to their literal or conventional meanings; and on the relationship of metaphorical meaning to pragmatic phenomena such as implicatures. Such discussions, and even extensive
88
John A. Barnden / Alan M. Wallington
related works such as Stern (2000), only touch upon a few salient examples of metaphor (see also some comments in Steen 2008), and have not amounted to full, stress-tested accounts covering the large range of types of metaphor and issues concerning metaphor. Even accounts of individual examples only address their meaning tentatively and partially. Not surprisingly, therefore, we do not purport to present in this article a specific, fully worked out account or meaning or truth as opposed to discussing certain issues and considering the implications of various phenomena and assumptions. 1.2 Concerning discourse participants and meanings We will use “generator” to mean the writer or speaker of an utterance and “understander” to mean an agent understanding it, be that agent a hearer or a reader (and be it a person or an AI system). We usually avoid the common but possibly misleading practice of referring to “speakers” and “hearers”, as metaphorical phenomena need not be the same in speech and text, and the question of decomposition into grammatical units is much more fraught for speech than for text. We also avoid the word “addressee” as an understander of an utterance is not necessarily an addressee. Under the heading of “meaning” of an utterance we will generally mean what many researchers would call “content”; but in any case we will take meaning to consist of specific propositions that are or might be conveyed in some way by an utterance. The main distinction we are drawing here is with more indirect semantic notions such as Stern (2000, 2008)’s notion of the “metaphorical character” of a linguistic entity (word, phrase, etc.), which is a function from contexts to contents (contents consisting of specific propositions, and that may or may not be true of any given world circumstance). So, in this article we cast those contents, but not the metaphorical character itself, as a form of meaning. One question for us will be what restriction if any should be placed on the derivation processes that lead to meanings, where there may be different types of meaning and thus different restrictions. We will use the term lexical meaning (short for lexical-compositional meaning) of an utterance for a proposition (or collection of propositions) that could arise from a given decomposition of the utterance into words or phrases, and compositionally stitching together one lexical sense of each of those words or phrases. We assume that an understander possesses an internal lexicon relating words (and many multi-word phrases, e.g. idioms) to senses. What counts as a lexical meaning is relative to a particular lexicon, be that an internal (mental) lexicon or an external one such as a dictionary. Clearly, when discussing a particular understander, what is of most interest is lexical meanings according to that particular understander’s lexicon. We avoid the closely related term “conventional meaning” (see, e.g., Camp 2006’s useful explanation of different meaning terms) as there may be idiosyncratic differences between different internal or external lexicons as to what senses are
Metaphor and its unparalleled meaning and truth
89
actually included, even when the various senses in question are generally agreed to exist. We also avoid the term “literal meaning” as an alternative to lexical meaning, because an internal lexicon may include senses that might be dubbed metaphorical by some metaphor theories, just as dictionaries do, so that an appropriate metaphorical meaning of an utterance might just be a lexical meaning. However, to focus our discussion of metaphorical meaning, we will put aside that lexical-meaning case as being relatively unproblematic, and assume that the understander is (for some reason, such as the demands from or influences from context) finding, or seeking to find, some meaning different from any available lexical meaning, one that we as theoreticians would call metaphorical (e.g., because inter-domain mappings are used, or certain types of adjustment of concepts or categories are used as in Relevance Theory or the class-inclusion theory, or ...). We will thus, for simplicity, restrict the term metaphorical meaning or metaphorical interpretation to refer to such a meaning, and will also use metaphorical interpretation to refer to the process for finding it. Finally, we will use the term literal meaning, relative to a given theory of what figurative (metaphorical, metonymic, ironic, etc.) language is, and relative to a particular lexicon, to mean a meaning derived from a non-figurative lexical meaning where the derivation does not amount to figurative interpretation according to that theory. As a special but possibly common case, a literal meaning can be a lexical meaning, but we are allowing for literal derivation to involve certain processes of inference, etc., as for example in Relevance Theory. For simplicity of discussion, and to parallel most discussions of metaphor, we will talk of the literal meaning of an utterance, or the lexical meaning, although in reality we should be talking about a range of literal or lexical meanings that are in some way relevant to a discussion of a metaphorical meaning of an utterance, even once contextual considerations have been taken into account. As is also usual in discussions of metaphoric we will have to make, for the sake of argument, hopefully-reasonable assumptions about what lexical and literal meanings particular example utterances have, as well as about what metaphorical meanings they might have. Some authors have objected to the notion that metaphorically interpreted utterances necessarily have lexical meanings, on the basis that the only possible candidates would be semantically anomalous in some way. So, for example, in the case of “The ATM swallowed my credit card” (see, e.g., Stern 2006), it would be claimed that no meaning arises from composition of the words’ lexical meanings (if we assume that, for instance, the only lexical sense of “swallow” is to do with an animal orally ingesting something). Our own feeling is that there is a lexical meaning, albeit an absurd one. We will adopt this as a working assumption in the article. The jury is still out on the issue, and if the verdict ultimately went against us we would seek to adjust some of our arguments suitably.
90
John A. Barnden / Alan M. Wallington
In any case, notice carefully that the above notion of metaphorical meaning does not assume that a metaphorical meaning is derived from some lexical meaning of the whole utterance, although the notion allows for this possibility. As pointed out by many researchers, it is at least in principle possible that a metaphorical meaning is (at least sometimes) worked out from the lexical meanings of parts of the utterance taken separately, without composing those lexical senses, or only partially composing them. 1.3 Concerning consciousness We will often be discussing inferences steps performed by an understander, knowledge used, processing strategies pursued, etc. However, we make no assumption that these are present to the understander’s consciousness, although in some cases they may be. 1.4 Concerning the two sides in metaphor Although our own ATT-Meta account of metaphorical interpretation rests on mappings, it does not rest on an assumption that a mapping goes between different domains or other sectors of the conceptual landscape possessed by an understander, under any technical notion of domain (etc.) that has so far been put forth in the literature. For example, in the case of metaphorical utterances that rest upon talking of an idea as if it were a physical object (“They kicked the idea around”, “She pushed the idea to the back of her mind”, etc.) we use in our system a mapping between physical operation upon an idea so viewed and mental usage of the idea. However, while this can loosely be described as mapping between physical subject-matter and mental subjectmatter (or between the physical domain and the mental domain, under a loose, intuitive usage of the word “domain”), we amongst other researchers are sceptical that any precise technical sense of “domain” will suffice as a way of clarifying or characterizing what mappings in metaphor do (see the considerations in, for instance, Barcelona 2002, Haser 2005, Kittay 1989, and Peirsman & Geeraerts 2006). In any case, the present article seeks to encompass the implications of theories other than our own, including theories such as Relevance Theory, class-inclusion theory and Ritchie’s CLST that do not involve mappings, and which therefore experience less pressure in the first place to be concerned with domains or other such conceptual sectors. We will say throughout that, in a case of metaphorical interpretation, the lexical and literal meanings of an utterance are about or within a source subject matter (or just the source, for short) but that the metaphorical meaning is about/within a target subject matter (target for short). Source-side inference is inference within the terms of the source (as long as the inference does not count as moving to metaphorical interpretation by the theory at hand); target-
Metaphor and its unparalleled meaning and truth
91
side inference is inference within the terms of the target. A lexical or literal meaning describes some scenario (situation) within the source, or at least part of some scenario; source-side inference can elaborate that scenario. We call such a scenario a source scenario. We call the target situation actually being described by the utterance – that is, by the metaphorical meaning of the utterance – the target scenario. Given that there is no precise delineation of what a subject matter is, the notions we are defining in this paragraph are necessarily only loose ones, but are nevertheless useful as a foundation for discussion. Finally, remembering that a lexical meaning of an utterance can itself already be metaphorical, in involving metaphorical lexical senses of component words or phrases, notice carefully that a source-side scenario can in that way itself be metaphorical. (And there can be chained non-lexical metaphor, so that sources intermediate in chains are themselves metaphorical with respect to subject matter later in the same chain.) Thus, the important distinction is not between metaphorical meaning and literal meaning, but between target scenarios and source scenarios and thus between metaphorical meaning (in the above sense) and lexical meaning and further source-side content derived from lexical meaning. 1.5 Concerning source/target parallelism We introduced above our use of the term “parallelism” to cover explicit or implicit source/target analogy, where the analogy is generally very partial, and can be fuzzy. Now, it is often observed that even when the understander already knows a parallel between parts of the source and target subject matters, an utterance may contain terms that refer to aspects of the source for which the understander knows no parallel. Hence, one question raised is that of whether the understanding of the utterance should seek a parallel for those aspects. In, for instance, “The idea was skulking in the caverns of John’s mind” the understander may be familiar with a partial parallel between minds and physical regions or terrains, and a partial parallel between ideas and physical (perhaps animate) objects, as these parallels are very commonly used in discourse, but may not know – in fact, probably does not know – a parallel for the specific concepts of “cavern” and “skulking”. The question is then whether or not the interpretation of the utterance needs to establish a target parallel for these so-far-non-paralleled source aspects. The notion of non-paralleled source aspects (which includes what we have referred to elsewhere as “map-transcending” aspects) is strongly related to, but not identical to, the concept of “unused parts” of the source-domain and to the concept of “metaphorical entailments”. We avoid these alternative terms as both are potentially misleading and get at a more limited concept than sheer lack of parallelism. It would be possible to misunderstand “unused parts” to refer to parts that had never been used in previous discourse. By contrast, lack of parallelism does not imply this novelty. It would, for instance, be possible
92
John A. Barnden / Alan M. Wallington
for the notion of caverns of the mind often to have been mentioned in discourse, but for there nevertheless to be no parallel for it known to a particular understander or any understander. Indeed it is likely that there is nothing specific about the mind that corresponds to, or could correspond to, physical caverns, which we claim are used in our example merely to emphasize qualities such as unavailability and perhaps menace of the idea. We avoid the “entailments” terminology largely because the term is used to talk about entailments of some specific conceptual metaphor (or set of conceptual metaphors) that is under discussion, whereas lack of parallelism is not relative to particular conceptual metaphors; and in the extreme case of a metaphorical utterance that does not exploit any already-known parallels at all there would be no sense in talking about entailments of conceptual metaphors. Equally, the “unused parts” terminology assumes that there is a specific source domain at hand. 2. Non-parallelism 2.1 (Non-)parallelism in some existing accounts While much of the emphasis in mapping-based accounts in Cognitive Linguistics and Pragmatics is on source/target parallels that the understander is already familiar with, another prominent line of work, better represented in Psychology and AI, is on forms of metaphor interpretation that involve discovering some source/target parallelism from scratch (i.e., creating mappings on the fly). Thus, psychological or AI models of analogy construction such as SME (Falkenhainer, Forbus & Gentner 1989) and ACME (Holyoak, Novick & Melz 1994) have been applied to metaphor (Gentner & Bowdle 2008, Gentner, Falkenhainer & Skorstad 1988, Holyoak & Thagard 1989). The AI approaches to metaphor of Wilks (1978) and Fass (1997) have also been based on mapping creation, as has one part of Hobbs’s (1990) approach. Now, mapping-based accounts are not in fact necessarily committed to finding a parallel for (all) non-paralleled aspects of utterances. Rather, they can instead seek as far as possible to connect non-paralleled aspects to aspects that do already have a parallel. The non-paralleled aspects can thereby still have a deep effect on the metaphorical interpretation, and even be at the crux of the interpretation. For instance, suppose a particular idea is being metaphorically viewed as a physical object, and the understander knows a parallel between, on the one hand, physical hiddenness of physical objects and, on the other hand, mental unusability of ideas (that are being viewed as physical objects). Then, in the case of the utterance “The idea was in a cavern in John’s mind”, the understander may be able to infer (source-side) from the fact that the idea (as physical object) is in a cavern in John’s mind that it is physically hidden, and can therefore use the known parallel to conclude that the idea is not mentally usable. This type of inferential linkage to known parallelism has
Metaphor and its unparalleled meaning and truth
93
long been central in our own approach to metaphor (Barnden 1998, 2001a,b, 2006a,b, 2008, Barnden, Glasbey et al. 2004, Lee & Barnden 2001). It has also appeared strongly in several other AI accounts, such as (another part of) Hobbs’s account (Hobbs 1990, 1992), and in Narayanan’s account (Narayanan 1999), which is part of the Neural Metaphor Theory developed by Lakoff and co-workers (see Lakoff 2008 for a summary). It has long been an aspect of Conceptual Metaphor Theory (see, e.g., Lakoff 1993) but not developed at the level of detail of the discussions of inference in accounts such as Narayanan’s, Hobbs’s and our own. The idea is also extensively used in Langlotz’s (2006) approach to metaphorical idioms. Langlotz generally casts the steps of inferential linkage as instances of “metonymy within metaphor” (a type of metaphtonomy in the sense of Goossens 1990) but we can take the metonymical acts as a special case of inference. However, there is a contrast between different flavours of accounts that use inferential linkage to known parallels. On the one hand, our own account emphasizes that if adequate meaning can be found by inferentially linking the utterance to source items with known parallels then parallels should not be sought for the non-paralleled aspects of the utterance (unless there is a special need to do so, or the understander has some special level of curiosity about such possible parallels). Thus, there is a tendency towards minimizing the amount of new parallelism sought. On the other hand, the Neural Theory of Metaphor involves a maximization of bindings principle (Lakoff 2008) that would seem to lead to the zealous creation of new parallels as opposed to a tendency to avoid them. As noted in the Introduction, some major approaches to metaphor such as RT, class-inclusion theory and Ritchie’s CLST do not involve mappings, and are therefore not obviously concerned with parallelism. These approaches have significant differences from each other, but for our purposes they all rest on the same core idea. (And the similarity between RT and class-inclusion theory is explicitly noted by Wilson & Carston 2006). This idea is that metaphorically used terms introduce lexical senses which, if not themselves already applicable to relevant aspects of the target, nevertheless inferentially imply or in some other way associatively activate features, categories or other conceptual items that do apply to the target. (In fact, ATT-Meta includes a version of this idea in its “View-Neutral Mapping Adjuncts” aspect.) To take an example from Wilson & Carston (2006), consider (1) Caroline is a princess. Suppose that the understander accesses a royal-personage lexical sense of “princess”. We will label this concept as PRINCESS. Suppose further that encyclopaedic knowledge attached to PRINCESS contains or implies, perhaps through much inference, the fact that, at least typically or stereotypically, princesses are [over-]indulged, spoilt, unwilling to do hard physical work, etc. We can cast this as the activation of a concept PRINCESS* of over-indulged, spoilt, etc. people. This concept need not have existed before in the under-
94
John A. Barnden / Alan M. Wallington
stander’s mind, in which case it is a newly-formed, ad hoc concept. Caroline is then asserted to be covered by this concept. The upshot for our purposes is that the notion of PRINCESS does not have an already-known parallel in the target subject matter (a subject matter to do with the activities and attitudes of people in general) and yet evidently is responsible, via PRINCESS*, for the metaphorical interpretation. An important point that will connect to an argument in Section 4.1.2 is that it is by no means clear here why a concept PRINCESS* has to be considered at all, rather than just to say that features such as OVER-INDULGED are asserted of Caroline. We will be concerned later that this may just be a terminological matter. But the question for now is whether the sort of process just outlined constitutes giving PRINCESS a target-subject matter parallel, consisting perhaps of the ad hoc concept PRINCESS* itself, is more vexed, and is somewhat obscured by the fact that both target and source are to do with people. A more telling example from this point of view is (2) Sally is a block of ice, analysed in Wilson & Carston (2006) in terms of, amongst other things, a concept of HARD physical objects. We suppose that part of the metaphorical interpretation is that Sally has a psychological sort of hardness (amounting to, e.g., a lack of sympathy for others). Wilson and Carston give two possible routes for the interpretation process to reach psychological hardness via physical hardness. Both routes involve, in somewhat different ways, a concept HARD* which subsumes both the subconcept of hard physical objects and the subconcept of psychologically-hard people. Thus, HARD* itself cannot be regarded as a psychological parallel for HARD. However, one of the suggested interpretive routes also involves first the activation of the concept HARD** of psychologically-hard people. This concept can of course be viewed as a target subject-matter parallel of HARD. In the processing route in question, HARD** is already known to the understander and is associated indirectly with HARD in the understander’s mind, because Wilson and Carston suppose that the word “hard” links to both HARD and HARD**. (However, they do not explain why this word comes into the processing at all: it’s neither in the utterance nor in the encyclopaedic knowledge package they specify, so perhaps the idea is that concepts can activate internal representations of words during the course of understanding utterances). Thus, according to the processing route in question, the understander is actually exploiting a known parallel, albeit not one specified by what would normally be regarded as a “mapping”. (Note also that there is a question as to why HARD* is constructed, if HARD** is what is relevant – see Tendahl & Gibbs 2008.) In the other suggested processing route, it is HARD* that is already known by the understander to be a superordinate concept of HARD. If Sally is simply regarded as instantiating HARD*, then the process does not involve finding or creating a parallel for HARD. However, Wilson and Carston imply that it is
Metaphor and its unparalleled meaning and truth
95
possible that from HARD* the subconcept HARD** is dynamically produced as the one to put Sally under, rather than putting her directly under HARD*. In this scheme, the understander is creating an on-the-fly parallel (HARD**) for HARD. To summarize, non-mapping-based accounts of metaphor can deal (to some extent at least) with cases where the utterance involves a not-alreadyparalleled source aspect. First of all, as we have just been discussing, an inferred concept such as HARD may be important for the interpretation but (depending on exactly how the theoretical account now goes) may not already have a parallel in the target and furthermore may not be given such a parallel in the course of interpretation. But secondly, and more dramatically, the notion of physical ice accessed by the word “ice” in the example above is not itself given any target parallel (by either of the two interpretive routes above), nor is the physical notion of a “block” of something. Yet clearly the physicalice notion is crucial to the metaphorical interpretation, and the physical-block notion is also important because it emphasizes, or at least clarifies, the physical hardness: after all, ice can be crushed or slushy. In this discussion based on RT, and also applying with suitable changes to other accounts such as the class-inclusion account and CLST, we have used copula metaphors (A-is-B metaphors) because that is typically what the authors in question do (though several non-copula examples are treated in, for instance, Sperber & Wilson 2008). But copula metaphors bias the consideration of non-parallelism, because after all in saying that Caroline is a princess or Sally is a block of ice we can always suppose that there is a source scenario containing an individual if hypothetical royal-family princess or physical block of ice, and say that Caroline is the parallel of that princess and that Sally is the parallel of the ice-block. Similarly when individual entities are not involved, as in “Everyone is a moon” (an example discussed below) we can postulate that each individual person corresponds to some moon (not necessarily different ones) or that the person category corresponds to the moon category. Non-parallelism comes out more clearly in non-copula metaphors such as the example used above of “The idea was skulking in the caverns of John’s mind.” Although again we can postulate an animate entity (doing the skulking) in the source scenario, and thus parallel to the idea in the target, there is no pressure from the form of the sentence itself to suppose that there is anything in the target scenario that is the parallel of caverns or of the notion of skulking. The following subsections explore more extensively the non-parallelism that can arise in metaphor, without going in any detail into the complex question of how different existing approaches might treat it. The examples serve as a prelude to making some points about the meaning of metaphors in Section 3.
96
John A. Barnden / Alan M. Wallington
2.2 Conversational cat flaps Barnden, Glasbey et al. (2004) extensively analysed the following rather vivid and literary example: (3) I tried not to run down Phil too much – I felt bad enough as it was, what with screwing his girlfriend and all. But it became unavoidable, because when Jackie expressed doubts about him, I had to nurture those doubts as if they were tiny, sickly kittens, until eventually they became sturdy, healthy grievances, with their own cat-flaps which allowed them to wander in and out of our conversation at will. (John Hornby: High Fidelity. Harmonsworth: Penguin 1995: 12) The claim is that it would be a mistake – a waste of time, even if success were possible – to try to find parallels for the following source-scenario aspects (and possibly others): the cat-ness of the doubts and grievances their sickliness the cat-flaps the cats’ complete control over the wandering (cf. the “at will”). The most obvious case here is the cat-flaps. The point about these is that they support the point (which is also explicitly stated) that the cats can wander in and out at will, this wandering corresponding to the grievances appearing intermittently in the conversation. It would seem pointless to try to work out some specific aspect of the conversation or some other part of the target scenario that corresponds to the cat-flaps. One could perhaps say that the cat-flaps are parallel to opportunities that the conversation provides for the grievances to appear intermittently in the conversation. But, it is already established that the grievances intermittently appear; so of course there must be opportunities for them to do so! If X happens then there must have been an opportunity for X to happen. Bringing in the opportunities as such as part of the meaning of the discourse, let alone establishing them as a parallel for the cat-flaps is just pointless extra work. Equally, the cats’ own complete control over their wandering has itself no parallel with any feature of the grievances corresponding to the cats. In particular, it cannot be parallel to control exerted by the grievances, because grievances are not the sort of thing that can exert control. Rather, the unrestrictedness and autonomy suggest, within the source scenario, that it is not the narrator (or anyone else) who is bringing the cats in and out: that is, the cats’ movements are not controlled by the narrator (etc.). It is this non-controlledness-by-the-people that has a parallel: it is parallel to the (alleged) noncontrolled-ness-by-the-people with which the grievances appear in and disappear from the conversation.
Metaphor and its unparalleled meaning and truth
97
The main point of the kittens’ sickliness (and tininess), arguably, is to buttress the metaphorical notion of “nurturing” the kittens. This nurturing itself does plausibly have some parallel in the target scenario. Now, if the kittens are sickly and tiny then nurturing is especially natural and morally good. Perhaps therefore the point of the sickliness and tininess is to portray the speaker’s action in converting doubts to grievances as natural and morally good. More securely, it can be inferred in the source scenario that the kittens were helpless and were likely to die if not nurtured. From this, we might suppose that in the target scenario the doubts were likely to disappear unless the narrator acted. So, the tininess and sickliness certainly lead to source-scenario conclusions that map to the target in some way, generating important information there, but that does not mean that they themselves, or the steps connecting them to those conclusions, map to the target. Finally for the above list, the cat-ness is totally irrelevant except in that it underlies familiar domestic scenarios of nurturing kittens, cats using cat-flaps, and so on. So it contributes to the development of a coherent source scenario, but there is no reason at all to suppose that some specific properties of cats (aside from the being-nurtured, the wandering, etc.) are to be mapped to the target. 2.3 Other examples An example we have treated in various papers (e.g., Barnden 2001a, Barnden, Glasbey et al. 2004, Barnden & Lee 2001) and whose understanding we have implemented using the ATT-Meta system is the following: (4) In the far reaches of her mind, Anne knew Kyle [her husband] was having an affair, but ‘to acknowledge the betrayal would mean I’d have to take a stand. I’d never be able to go back to what I was familiar with,’ she says. Not until eight months had passed and she finally checked the phone bill did Anne confront the reality of her husband’s deception. (L. Gross: “Facing up to the Dreadful Dangers of Denial”, in: Cosmopolitan 216 (3), USA ed., March 1994) Here we claim that the point of the “far reaches” is to generate the source-side conclusion that Anne’s conscious self (metaphorically conceived as a person located in the middle of Anne’s mind-space) has great difficulty in physically operating upon the idea that Kyle was having an affair (this idea being metaphorically cast as a physical object physically located in the far reaches of the mind-space). There is no need to, and it would probably be very difficult to, find some component or aspect of Anne’s mind that parallels the far reaches themselves. We could postulate some such component, and say that the affairthought was in some sense “in” it, but what would be the point? The message of the utterance is the fact that Anne is in a mental state of having difficulty in
98
John A. Barnden / Alan M. Wallington
consciously involving the affair-thought in her thinking. And this difficulty comes from the difficulty of, in metaphorical terms, physically operating upon the idea. The “far reaches” are just a source-side tool for conveying this physical difficulty. Barnden (2006b) discusses the degree of non-parallelism in fifteen further examples taken from real discourse. One of those examples is: (5) This all means that general managers have cricks in their necks from talking down to the Community Health Councils and District Health Authorities, and up to Regions and the Department. (Goatly 1997: 162; from The Daily Telegraph) This arguably involves a familiar metaphorical view of ORGANIZATIONAL CONTROL AS VERTICAL POSITION. The managers control the Councils and Authorities, and are controlled by the Regions and the Department. In the source scenario, the managers get cricks in their necks because of their contortions. The managers therefore experience physical suffering, and hence emotional suffering. The causation and the emotional suffering can map to the target by ATT-Meta mapping mechanisms discussed in Barnden (2001b, 2006b) and Barnden & Lee (2001). Alternatively, in, say, an RT account the inferred emotional suffering can be applied directly to the managers (and perhaps the causation can be dealt with somehow). Whatever the particular account, there is no justification for postulating that the understander’s prior existing knowledge already encompasses parallels for necks, cricks, or physical suffering, and there is no need for the understander to construct such parallels on the fly in order to come to an intuitive interpretation. Hence, those aspects of the source, and the way they are connected, are not paralleled in the target. They serve merely to imply the existence of emotional suffering caused by having to deal with the various entities mentioned in the passage.1 A further example discussed in Barnden (2006b) is (6) it was that mechanical sort of smile that suggested gears and pulleys (Goatly [1997: 181]; from a corpus). There is no need to have or find a parallel for the gears and pulleys, which serve merely to express an extremity on a dimension of machine-likeness (automaticity, artificiality, etc.). An understander might well build a mental image of the person’ head containing gears and pulleys, but that does not of itself imply that those gears and pulleys have parallels in, say, the actual physical structure of the head. However, if understanders infer a jerky type of movement from the gears and pulleys, they might ascribe a similar jerky style ————— 1
A variant interpretation of the passage is that the managers get real cricks in their necks, from the stress cuased by their dealings with the managers, etc. But we are not talking about this interpretation.
Metaphor and its unparalleled meaning and truth
99
to the person’s smile apparatus, and this style would then serve as a parallel for the hypothetical machine’s style of movement. 3. Non-parallelism, source-scenario holism and metaphorical meaning In this section we first draw out some suggestions that non-parallelism leads to for nature of the connection between linguistic entities such as sentences and their meanings (in the metaphor case and more generally). We believe that the profundity of these consequences has yet to be fully appreciated. First, we need to make a point that is hopefully obvious but that appears rarely to be stressed despite its importance. For understandable reasons, most technical discussions of metaphor meaning use quite short sentences as examples, such as (1) “Caroline is a princess”, or the commonly-used example abstracted from Shakespeare’s Romeo and Juliet of “Juliet is the sun” (see, e.g., Stern 2000), or non-copula examples such as “My car drinks gasoline” (see Fass 1997, Wilks 1978). In discussing (4) in previous work, we have generally abstracted out the sentence “In the far reaches of her mind, Anne knew that Kyle [her husband] was having an affair” rather than dealing with the actual full example. Thus, one can be seduced into thinking that the main issue for metaphor meaning studies is the meaning of metaphorical sentences, and that target-side propositions that are part of the content of metaphorical discourses are components of the meanings of whole metaphorical sentences. But, of course, consideration of sentences needs to be generalized to, at least, sub-sentential units such as clauses. For one thing, a metaphorical clause might be coupled with a non-metaphorical clause, as in “Juliet is the sun and I love her”. Also, clauses are not the only sub-sentential unit type that might have propositional content (i.e., cashed out meaning-wise as a set of propositions); so, for example, in “Caroline had a hissy fit in her princessy way” the prepositional phrase “in her princessy way” could be taken to have as part of its content the proposition that Caroline is princessy. Similarly, in the cat-flap example (3), we find the segment “the sturdy, healthy grievances, with their own cat-flaps”; here the prepositional phrase “with their own cat-flaps” could be treated as providing the source-side proposition that the cats have their own cat-flaps, leading to the question of whether there is some parallel target-side proposition (such as that the grievances have opportunities for appearing intermittently). Some researchers have indeed attended to metaphorical clauses as opposed to sentences, e.g. in discussion of metaphor within speech-reporting and attitudinal contexts (e.g., “Romeo said that Juliet is the sun”, “Romeo believes that Juliet is the sun”, cf. Stern 2000; also see van Dijk 1980; see also Camp 2005 for other sorts of case). Moreover, part of the point of Stern’s “Mthat” operator for delivering the metaphorical character of an expression is that it can be applied to any constituent of a sentence or to a whole sentence. How-
100
John A. Barnden / Alan M. Wallington
ever, our emphasis in this article is on a large consequent issue that is raised, namely the question of what precisely are the grammatical units, in any particular case of metaphorical utterance, that should be given a metaphorical meaning? The point is that, once we have realized that we ought at least some of the time be giving metaphorical interpretations to metaphorically-meant sub-sentence components, and then stitching together those interpretations, the question is whether that is always what we should do. We will proceed, in the next subsection, to argue that the answer is No: that it is often more sensible to stitch together that source-side meanings of separate metaphoricallymeant grammatical components, and extract target-domain content somehow from a (relatively) holistically created source-domain scenario. Creating separate metaphorical meanings for the separate constituents would be a distraction: a waste of work, where moreover the work (even when doable) would often be artificial or difficult. We will not be answering the question of how you tell whether it is appropriate to derive separate metaphorical meanings for the constituents or not. (We would suggest that contextual effects are important, in that the need to link the metaphorical sentence to surrounding discourse will govern what sorts of information are derived from the sentence and thus influence the treatment of constituents.) Our task is to show that this is an issue worthy of discussion and that it is often sensible not to derive separate metaphorical meanings. Such non-derivation constitutes a form of non-parallelism, not necessarily in the sense that a parallel (a separate metaphorical meaning) could not be derived, but at least in the sense that such derivation is avoided. 3.1 Sub-sentential holism: avoiding separate metaphorical meanings for constituents Consider the following example: (7) Everyone is a moon, and has a dark side which he never shows to anybody. (Attributed to Mark Twain by Brians 2003: 74) The question is: should we derive a metaphorical meaning for the clause “Everyone is a moon” and a metaphorical meaning for the clause “[Everyone] has a dark side which he never shows to anybody” and then combine these meanings? We would strongly suggest not. Surely the second clause is an indication of what it is about being a “moon” that we should attend to. (There is the separate point that the passage is conveying that there is something bad about the non-shown side. This aspect does not seem connected to the moon, but instead relies on a commonly-used metaphorization of bad stuff as darkness.) Certainly, the second clause can be given a metaphorical meaning. In the course of doing this we can use the property of Earth’s physical moon (and similar moons elsewhere, perhaps) that it has a dark half that cannot be
Metaphor and its unparalleled meaning and truth
101
seen (by the naked eye).2 This consideration reinforces the never-showing aspect of the second clause. But this reinforcement is on the source side. Having taken effect, it of course serves to reinforce the target-side message, so that the first clause is indirectly having an effect. But the point is that there is no need to give the first clause its own metaphorical meaning. Indeed, it would be quite hard to do this: either the operation would involve using the second clause for guidance as to what is meant, in which case there hardly seems any point considering the first clause at all by itself, or the operation would involve taking the clause in isolation of the second, in which case (unless surrounding discourse context could help) we have the usual problem of the indeterminacy of metaphor (cf. e.g. Stern, 2000) – note that the moonness could imply something like subservience to something else that is being portrayed as the Earth. Actually, the first clause has a deeper effect than just reinforcing the never-showing. The moon also has a bright side, at least some of which we can normally see, and which indeed (unless obscured by cloud) is extremely salient in the night sky. Thus, a more elaborated interpretation of the sentence could include the notion that everyone also has a side that is (in part) usually very much apparent. This new message cannot come from just the second clause, because although the mention of a dark side weakly suggests a nondark side, there is no warrant for taking that side to be bright and salient. But, the fact that the message cannot come just from the second clause alone is a not a reason for saying that the first clause should be given its own metaphorical meaning, but is rather a reason to say that a source-side scenario should be constructed from the clauses, and then target-side meaning should be extracted form the scenario as appropriate. The second clause, “[Everyone] has a dark side which he never shows to anybody” is of course itself composed of two clauses, “[Everyone] has a dark side” and (effectively) “he never shows it to anybody”. There is certainly a case here for saying that these two clauses contribute separate target-side messages, especially as the second sub-clause is going out of the person-asmoon conception unless it is personifying the moon (giving us the complex situation of a chained metaphor). The darkness and the non-showing connect to each other source-side, but that is not in itself a reason for saying that they do not have separate target-side significance. Thus, we have a mixed situation where in one case it is sensible to treat constituents somewhat separately as regards metaphorical meaning, whereas in another case in the same sentence it is not sensible to do that. Consider now the next example: —————
2
One may wonder whether Mark Twain is mixing up the idea of a dark side with the idea of the hidden side of the moon, which of course is usually partially lit by the sun and indeed wholly lit once a month, although we can never see it. The actual dark side is changing form moment to moment, so does not actually fit what seems to be Mark Twain’s intended meaning very well. See also Brians (2003: 74).
102
John A. Barnden / Alan M. Wallington
(8) Suddenly, as if the movement of his hand had released it, the load of her accumulated impressions of him tilted up, and down poured in a ponderous avalanche all she felt about him. (From Virginia Woolf, To the Lighthouse) We interpret this as using a metaphor of a truck that has a tilting platform for carrying material such as gravel. We concentrate on the clauses “the load of her accumulated impressions of him tilted up” and “down poured in a ponderous avalanche all she felt about him”. We suggest that these two clauses together paint a source scenario in which “accumulated impressions” are like gravel sliding off the tilted platform of the truck. It is difficult see what separate target-side meaning the first clause could have. Rather, it clarifies the nature of the “avalanche”: the truck-load presumably slides completely off – it’s not like an ordinary mountain avalanche where just some snow slides to a lower point. The tilting also perhaps conveys a degree of suddenness or quickness in the process, aiding and abetting the adverb “suddenly” in the passage. And the tilting goes along with the hand-movement in the first part of the passage, just a truck operator might pull a lever. Thus the clause “the load ... tilted up” contributes important coherence, clarification and emphasis to the source-side scenario, rather than having a metaphorical meaning all of its own. (We emphasize though that such a meaning could be concocted with sufficient theoretical conjuring.) Finally for this subsection, we will briefly some aspects of the following lengthy and complex example (which is not necessarily to be regarded as a felicitous use of metaphor!): (9) The lid of my consciousness has become thinner, thanks to my trials. I now begin to see beyond my former daily perceptions, of a closed material world. Gleams and flashes come to me: the close sense of great movement in the world just beyond our own, the sense of His Presence and Intention hovering, focused, above us---almost, I may say, like a cloud. Yes, like a cloud, which gathers slowly and thickens, blotting out the light of the sun with its intention, with the moist deposit for earth that it contains. (Jane Rogers: Mr Wroe’s Virgins, London: Faber & Faber 1991: 270) Much as with the moon example, the part of the passage beyond the first clause, “the lid of my consciousness has become thinner, thanks to my trials”, serves to explain what that clause is getting at. The first clause does rely on the familiar metaphorical view of the MIND AS CONTAINER, adding to it the aspect of a lid, but it is difficult see what its import is. Notice that the lid is not opened – it just becomes “thinner”. It is hard to see what this amounts to in target terms without reading on. Although once one has read the later sentences one can go back and stipulate that some part of or the whole of their message is the metaphorical meaning of the first clause. But why would we
Metaphor and its unparalleled meaning and truth
103
want to suppose this happens, other than satisfying a perceived theoretical need to give each constituent its own target-side meaning? The thinning of the container lid is presumably intended to buttress some sort of openness on the source side (though it is a bit of a puzzle, given that a thin lid can still be opaque, etc.) 3.2 Holism beyond sentence boundaries In the previous subsection we largely confined attention to holism across constituents within a single sentence. But note that in the last example, (9), there is a sentence break after what we called the first clause. Thus, the phenomenon we are talking about is not particularly about constituents within the same sentence, even if that this is perhaps how it normally arises in practice. Also, the example shows that a whole metaphorical sentence can be one that we suggest should not be given its own metaphorical meaning. Indeed, even with other examples we have looked at, we can chop them up into separate sentences, as in this variant of (7): (10) Everyone is a moon. Everyone has a dark side which he never shows to anybody. Presumably this stylistic change should make no essential difference to theoretical questions of metaphorical meaning. Yet judging by the metaphor literature one would thing that there was an imperative to give the sentence “Everyone is a moon” a metaphorical meaning. We can pull the same sort of trick with (8), getting (11) The load of her accumulated impressions of him tilted up. Down poured in a ponderous avalanche all she felt about him. It was as if the movement of his hand had released it. The change should not increase the pressure to supply a separate meaning for “The load of her accumulated impressions of him tilted up”. Similarly, consider the following stylistic variant of part of (3): (12) I had to nurture those doubts. They were tiny, sickly kittens. Eventually they became sturdy, healthy grievances. They had their own cat-flaps. These allowed the cats to wander in and out of our conversation. They did this at will. Presumably, no matter what might think of the literary merits of this new version, any metaphor theory that could provide a metaphorical meaning of (3) should be able to provide one for (12). Consider one of its sentences, “They [the grievances/cats] had their own cat-flaps”. Our points in Section 2.2
104
John A. Barnden / Alan M. Wallington
about the non-parallelledness of the cat-flaps are not essentially affected by switching to the more fragmented (12). The problem is that given non-parallelism in (3) and hence (12), it is difficult to see what metaphorical meaning the sentence “They had their own cat-flaps” could have, unless the understander goes to the lengths of assuming that the cat-flats are parallel to particular types of event arising in the conversation that provide special opportunities for the grievances to be conveyed or cease to be conveyed. To the extent that this is an optional elaboration, it’s equally optional for the sentence “They had their own cat-flaps” in (12) to have a metaphorical meaning by itself. Rather, (12) merely emphasizes the at-willness of the metaphorical cats’ movements within the source scenario, and thereby indirectly supports the inference of non-controlledness of the grievances by the conversational participants. This at-willness by the cats and non-controlledness by the conversation participants is conveyed more explicitly by the last two sentences of (12). Thus, we conclude that “They had their own cat-flaps” does not necessarily by itself convey anything specific about the target scenario. It just conspires with other sentences to lead within source terms to a conclusion about non-controlledness. Thus non-controlledness is then transferred to the target. So, in sum, the crucial target-scenario features – the non-controlledness – arises holistically from the source-side scenario supported by several sentences rather than being pieced together from individual target-side meanings derived from the individual sentences. Turning to (4), we can consider the following stylistic variant of part of it: (13) In a part of her mind, Anne knew Kyle was having an affair. This part was way out in the far reaches. Here, the overall effect of the two sentences in (13) is roughly at least the same as the corresponding part of (4). The second sentence serves just to clarify the nature of the source scenario that is partly described by the first sentence, rather than having a target meaning all of its own (although we could no doubt concoct an artificial one). The effect of (13) is from the two sentences taken together as painting a source scenario from which a target scenario can be derived. It is not that both sentences have their own separate target meanings that are then combined to find an overall metaphorical meaning for (13). Certainly, we can argue that the first sentence in (13) does have some sort of target meaning, weaker than the effect (13) as a whole. For instance, the first sentence could be taken to mean that Anne’s knowledge is not held with a standard degree of conscious awareness. But the second sentence now steps in to clarify the nature of the “part” mentioned by the first sentence and thereby to provide a stronger and more definite meaning for (13) as a whole. What we abstract form these examples and those of the previous subsection is that it is misguided to think of the primary issue in the area of the meaning of metaphor to be the meaning of sentences. Sentences are the wrong
Metaphor and its unparalleled meaning and truth
105
unit, and indeed there is no clear grammatical unit at all, in general. Rather, meaning arises out of discourse in way that is much less tied to grammatical boundaries, and dependent on the specifics of any given stretch of discourse. In particular it can arise from the cooperative effect of more than one clause or sentence. It is just that, often, some metaphorical meaning of some sort can indeed be extracted from a single metaphorical sentence or sub-sentential unit. But this special case has no special significance, other than that of having distracted metaphor researchers from the holistically discourse-based, as opposed to sentence-based, nature of metaphor, even when the researchers are in other respects very sensitive to contextual effects on metaphor meaning. And when, as for the first sentence of (13), a single sentence is quite readily assigned a metaphorical meaning, this meaning can then be refined in the light of other metaphorical sentences in a way that relies on refining the source scenario, rather than operating just within the target. The source refinement then leads to a parallel target refinement. Part of what we are saying here is that an aspect of the target scenario may arise from a source-scenario aspect that results from inference processes supported by several sentences. That source-scenario aspect may not be readily attributable to just one sentence. Thus, in (13) the conclusion that Anne has a low degree of conscious awareness of the idea of her husband having an affair arises from a conclusion, in source terms, that the idea as physical object is highly inaccessible physically. This inaccessibility is inferred both from the information that Anne’s state of knowing is physically located in some part of her mind (so, we assume, the idea itself is located there too) and that that part is in the far reaches her mind: drawing from both sentences in (13). A form of discourse that is an extreme case of the phenomena we are drawing attention to is allegory, if we can take it to be a type of extended metaphor (contra Crisp 2005). Many sentences or multi-sentence passages in a lengthy allegory may serve to enrich the source scenario to make it more vivid, believable, story-like, coherent etc. rather than having any discernible, separate significance for the target. However, our claim is that this phenomenon is much more widespread in metaphor. We should also mention that we perceive blending accounts of metaphor as being implicitly friendly to the arguments of the present section (and indeed other aspects of this article). Finally, there are consequences of this discussion for the meaning of nonmetaphorical language. It would on the face of it be anomalous if it turned out that sentences and clauses were the right unit for propositional meaning in non-metaphorical cases but not the right unit in metaphorical cases (especially if there is no rigid division between metaphorical and non-metaphorical language, as argued for instance in RT: Wilson & Carston 2006, Sperber & Wilson 2008). We would thus tentatively suggest that meaning and truth are, across the board, much more holistically dependent on discourse chunks of no particular shape than on grammatical units as has usually been assumed.
106
John A. Barnden / Alan M. Wallington
4. Restrictions on inference in interpretation In this section we turn to a different set of considerations, but ones that have an important interaction with the non-parallelism issue, as we will show in Section 5. Several accounts of metaphor interpretation allow for possibly elaborate chains of inference or association of various sorts (as in RT, Ritchie’s CLST, ATT-Meta and other AI accounts such as those of Hobbs and Narayanan). We will use “inference” here as a term encompassing the various ways in which interpretation can move from one mental item to another. The question therefore arises of whether one can discover a priori limits to the amounts or broad types of inference that should be allowed to contribute to the meaning of a metaphor. Stern (2000) seems to restrict the inferences to what he calls “massociations”, which include associations by similarity and exemplification. As regards RT, Wilson and Carston (2006) and Sperber and Wilson (2008) explicate metaphor in terms of conceptual broadening and narrowing processes (that also go on in literal language, hyperbole, etc.). By contrast, ATT-Meta, the Blending account (Fauconnier & Turner 1998, Turner & Fauconnier 2000) and the accounts by Hobbs and Narayanan are much more liberal about the type and amount of inferences allowed in constructing metaphorical meanings. Inference-heavy accounts such as these, together with RT and Ritchie’s CLST lead in a particular direction, we will now suggest. We will first argue, in Section 4.1, that target-side inferencing in metaphor interpretation should be no more or less restricted, a priori, than source side inferencing; and since on our own approach we place no a priori restriction on the latter, we suggest we should place none on the former. We will also argue that in RT’s apparent restriction to broadening and narrowing is not in fact a restriction at all: virtually all inference can be cast theoretically as some combination of broadening and narrowing, and what is left out from this is also needed in metaphor interpretation. We then go on in Section 4.2 to argue, briefly, that our liberal stance on amounts and types of inference does not raise problems for accounting for what generators of metaphorical utterances intend to convey or for what is believed by an agent who holds a metaphorically-couched belief.. Although, as noted, some metaphor accounts seek to restrict the types of inference involved, the accounts mentioned to not seek to place any a priori restrictions on the amount of inference. Thus, the amount-of-inference aspect of the present section is not put forward as a corrective to other accounts. However, the point is important to make as it has a significant role in Sections 4.2 and 5. Now do we count the relevance-seeking principles of RT, which can serve to limit the overall amount of effort the understander puts in (essentially, the understander puts in just the effort to achieve an adequate amount of relevance for his/her current purposes) as an a priori restriction on amount of
Metaphor and its unparalleled meaning and truth
107
inference, because there is no limit placed by the theory in general on the amount of inference that might be required to achieve adequate relevance. 4.1 Target-side, source-side and literal-case inferencing We will split the discussion between metaphor approaches that use mappings and ones that do not. 4.1.1 The case of mapping-based accounts The starting point of this subsection is that approaches such as our own (ATTMeta) and Hobbs’s do not restrict the amount and type of inference that can take place in order to link lexical meanings of utterances to conceptual elements that the available mappings can handle. For instance, in our approach we would take the physical location of an idea in the far reaches of Anne’s mind to imply, ultimately, that Anne’s conscious self only has a very low degree of ability to physically operate upon the idea. This conclusion is then acted upon by a particular mapping to create the conclusion that Anne only has a very low degree of ability to use the idea in her conscious thoughts. The process of linking the physical location of the idea to the very low degree of physical operability is quite an elaborate one, and in particular uses some common-sense knowledge about physical space and objects. Such a linking process can use any type of relevant knowledge or inference step. Similarly, in the case of the cat-like grievances in example (3), there is a need to infer the non-controlledness (on the part of the conversation participants) as regards the movements of the cats. We assume that this non-controlledness maps over to the target, to apply to the grievances. Within the source, the non-controlledness is inferred from the at-willness on the part of the cats themselves: presumably the at-willness implies that the cats experience no limitation on their movements and enforcement of movements, and this then implies that the conversation participants are in particular not limiting or enforcing the movements. Furthermore, the extra emphasis on the atwillness within the source that comes form the mention of the cat-flaps rests on reasoning based upon knowledge about the nature of cat-flaps. From just these two examples it is becoming implausible that any a priori limit can exist on the amounts or types of source-side inferencing done to link lexical meanings to mappable content. Of course, time and effort constraints that circumstances place upon the understander serve to provide ad hoc limitations. The question now is what amounts and types of target-side inferencing can happen whenever the interpretation process steps over into the target subject matter via the action of a mapping. We would tentatively suggest that it would be misguided to restrict the amounts or types. More precisely and generally, our discussion will argue that the target side and source side should
108
John A. Barnden / Alan M. Wallington
be equally restricted or unrestricted. We argue this in two halves that combine to give the result: (A) that restrictedness on the target side should be equal to restrictedness of inferencing in literal interpretation (we will call this “literalcase restrictedness”); and (B) that, in turn, literal-case restrictedness should be equal to source-side restrictedness. As regards (A), we note that in literal interpretation, the lexical meaning of the utterance already places us in what we can call the target subject matter, and then some inferencing may occur to modify or enrich the situation described by the lexical meaning. It is difficult to see why different restrictions should apply to such inferences from those that apply to target-side inferencing in metaphor interpretation: the only difference between the literal process and the metaphorical process is the directness with which they land on the target side. Of course, time and effort constraints may affect literal and metaphorical interpretation differently, because the latter may require source-side as well as target-side inference, but this is not an issue of a priori restrictions on any one aspect of overall inferencing. And of course a particular case of literal interpretation and a particular case of metaphorical interpretation will be seeking to connect in different ways to surrounding discourse, so there may well be differences in the actual inferences that come to be performed, as a result of such contextual differences. It is helpful here to consider the (very common) case of sentences that can be literally true in some contexts and metaphorically true in others. An (invented) example is (14) The hippos are wallowing in their swamp which could be uttered in a context where the hippos metaphorically stand for some people and the swamp metaphorically stands for, say, some project they are involved in. Clearly, what target-side inferences are actually performed to understand (14) in such a context will be very different from the (“targetside”) inferences performed in a context where (14) is being taken literally. In the former case the inferences will be about people, project activities, etc., whereas in the latter they will be about hippos, the swamp, etc. But this matter does not impinge on the question of a priori restrictions on amounts or types of inference, where by type we are not referring to subject matter but rather the general form of the inference (m-association, concept broadening, narrowing, deduction, induction or whatever). Turning to (B), on the other hand, what is at issue for (14) is whether different restrictions should apply to inferencing about hippos and swamps when the utterance is taken literally from when it is taken metaphorically. It would, for instance, be odd to say that the metaphorical process could do unlimited inferencing of that sort but the literal process could not, or that, conversely, the literal process was more unrestricted than the source-side of the metaphorical process.
Metaphor and its unparalleled meaning and truth
109
Thus, if both (A) and (B) are true, then source and target sides of inferencing in metaphorical interpretation should be equally restricted (and therefore the target side should be unrestricted if the source side is). Nevertheless, before moving on we will raise a possible objection to (B). The objection is based on the point that the useful inferencing actually done on the source side is constrained by what mappings are available. For an inference to be useful in helping to develop a target-side meaning, it must match some mapping or must lead to some mapping that does so. Consider again the hippo example (14). Now, it is said that hippopotamuses cause more deaths in Africa than any other animal because they upset boats. A source-side inference made by someone familiar with this fact might be that the hippos could cause somebody’s death. However, unless there is a mapping that already applies to this proposition, or to some source-side consequence of the proposition, or some such mapping is now invented, this inference will not contribute to the targetside consequences. Thus, the very nature of metaphor supplies its own qualitative constraint on the nature of inferencing on the source side, and therefore there is no reason to be concerned about a disparity with the inferencing done in the case of literal interpretation. In particular, it is reasonable to say that in metaphorical interpretation we allow, as part of meaning construction, reasoning steps within the source side that would not be allowed as part of meaning construction in the literal case. The other side of the coin is that what steps are usefully done is constrained by constraints that simply do not exist for literal interpretation. To put it another way, there are extra opportunities for inferencing opened up by the need to link to mappings, but the inferencing does need to lead to things that can indeed be mapped. However, we suggest that this objection does not work, for the following reason. It does support the idea that the particular inference actions usefully made are constrained on the source side by the availability of particular mappings and their relevance in the light of getting a coherent understanding of the overall discourse (and indeed the particular mappings may serve to influence or even guide what inferences are actually made), but it does not in principle show that the amounts or types of inference are constrained a priori. Thus, while the constraints from the mappings make source-side inference special in one qualitative sense, it is not a sense that bears upon the matter of the sort of restrictedness that our argument for (B) gets at. Also, the idea that available mappings in a sense grant a licence to draw out source-side inferences from the lexical meaning, over and above inferences allowed in getting a literal meaning from the lexical meaning, is at odds with the notion that the literal meaning is itself presumably formed for reasons of establishing a connection with context. Thus the only difference between the pressures applied by context on the inferencing is that in the metaphorical case the pressure is transmitted via mappings into the source side whereas in the literal case it is transmitted directly within that same subject matter.
110
John A. Barnden / Alan M. Wallington
4.1.2 The case of non-mapping-based accounts For simplicity and definiteness, we will concentrate here on RT again, as a specific but reasonably representative case of non-mapping-based but inferentially heavy accounts. In RT, the meaning extracted form a metaphorical utterance is claimed to consist of inferential actions that broaden or narrow the concepts under which entities are classified, such as broadening and narrowing a literal, royal-family concept of PRINCESS to another concept PRINCESS* of people who are over-indulged, etc. Given RT’s emphasis on the continuity of metaphorical interpretation with other forms of interpretation including literal, it is even more the case that we should align any inferential restrictiveness applied to literal interpretation of, say, (14) with the restrictiveness applied to metaphorical interpretation of it. We will now argue that actually to say that metaphorical interpretation involves (a particularly marked extent of) broadening and narrowing actually says nothing more than that metaphorical interpretation involves (a particularly marked extent of) inferencing in general. This is because the notions of broadening and narrowing actually cover most types of inference, and to the extent that there’s anything left, the residue is something that RT should seek to include within metaphorical interpretation anyway. Suppose some form of inference (deductive, abductive, inductive, by-analogy or whatever) leads from some premises P1,...Pn to a conclusion C. We can always regard the inferencing as a matter of doing some broadening and possibly some narrowing. To take a simple case first, suppose the premises and conclusion are about some single particular entity E. For instance, there could be two premises P1 and P2, where P1 is that E is a student and P2 that E is young. The conclusion C could be the plausible conclusion that E is (financially) poor. Now consider the concept Prem of young students, i.e. the concept covering things F that satisfy both premises in place of E. Also consider the concept Concl of entitles that are poor, i.e. the concept covering entities F that satisfy C with F in place of E. Then the inference can be regarded as an action of moving from the concept Prem to the concept Concl. If the inference is absolutely watertight (i.e., any young student is poor), then that move is one of broadening, as Concl is strictly more inclusive than Prem. Otherwise, it is one of both broadening and narrowing. Going from the concept of young students to the concept of poor people is no different in quality from going from the concept PRINCESS of royal princesses to the concept PRINCESS* of overindulged (etc.) people. This equivalence of quality may initially be obscured by the practice of using, for convenience, labels such as PRINCESS*. Such labels may surreptitiously and erroneously suggest (to the reader of a paper using them) a connection to some other concept such as PRINCESS that is stronger, more immediate or more fundamental than the connection between the concept of poor people to the concept of young students. It is important to remember that
Metaphor and its unparalleled meaning and truth
111
the move from PRINCESS to PRINCESS* is a matter of encyclopaedic knowledge and can involve inference. The general case of our point is as follows. Let’s say the premises P1,...Pn and conclusion C are about some particular entities E1,...,Ek: in other words of some list of entities <E1,...,Ek>. Consider any conceivable ordered sequence of the form
that satisfies P1,...,Pn in place of <E1,...,Ek> (that is, when E1,...Ek in the premises are replaced by entities F1,...,Fk, the modified premises are true). We can then define the concept Prem of those entity sequences that do satisfy the premises. We can similarly define the concept Concl derived from the inference’s conclusion C. Then, Concl is either more inclusive than Prem, in which case we have broadening, or covers only some of the entities covered by Prem (including, we assume, the actual entity sequence in consideration,. <E1,,,,Ek> ) in which case we have a case of narrowing and broadening. To take an example, suppose we have the following premises: Abigail loves Boris; Boris loves Carla; the named entities here are people; and they are all different from each other. Imagine an inference to the conclusion that Abigail is jealous of Carla. Consider the concept Prem of three-person sequences where A,B,C are all different, A loves B and B loves C, and consider the concept Concl of three-person sequences where A an C are different and A is jealous of C. (It doesn’t matter about B). Then, to say that the conclusion in question follows from the premises in question is equivalent to saying that Concl is more inclusive than Prem or some restriction of Prem, i.e. that Concl and Prem are related by broadening and possibly narrowing. And this “conceptification” of inferences is not just some arbitrary trick we are doing – or at least no more arbitrary than the use of concepts in RT’s account of metaphor already is. This point is obscured by the is-a form of examples such as “Caroline is a princess”. In another examples such as “Sally is icy” – to vary example (2) – it is only a theoretical construction to bring in concepts of things made of ice, or hard things, or whatever, rather than, say, to deal throughout in terms of properties such as iciness or hardness, or propositions that could be expressed in logical form by logical forms like madeof(Sally, ice) and is-hard(Sally). What our conceptification of inference does not reduce to broadening/narrowing is of course the case where Concl actually covers the same entities as Prem. This would arise in inferences like going from Jim being a bachelor to or from Jim being an unmarried male adult person (putting aside, for example, arguments about whether monks are bachelors). However, it would seem advisable to propose that such equivalences are usable in the course of metaphor interpretation. Thus, to say that metaphor interpretation involves broadening and narrowing says little if anything over and above saying that it uses inference. To turn to another way in which inferences can be classified, RT excludes implicatures from the inferencing done during metaphor interpretation, where implicatures are defined in RT (c.f. Sperber & Wilson 2008) as generator-
112
John A. Barnden / Alan M. Wallington
intended implications of the explicit content of an utterance (where explicit content includes the metaphorical meaning of an utterance in terms of broadened/narrowed concepts). Our arguments do not bear against the exclusion of implicatures. An important feature of implicatures is that they use, as a premise, the fact of utterance: the fact that the generator has chosen to make the particular utterance under consideration and not some other. (Wearing 2006 points out that this leads to an objection to taking implicatures to be part of metaphorical meaning, because metaphors can be in attitudinal contexts, something we will discuss in Section 4.2.) It seems that the broadening and narrowing done during metaphorical interpretation is always applied to the concepts such as PRINCESS linked directly to lexical items within the utterance, not concepts involving aspects of the fact of utterance (although the understander’s choice of which particular concept(s) attached to an ambiguous lexical item to attend to can be affected by any sort of contextual factor). Thus, metaphor interpretation in RT excludes the fact of utterance from taking part in the premises on which the broadening and narrowing work. However, the exclusion of the use of particular sorts of premise does not conflict with our arguments in this section. That is, the type of inferential liberalism we advocate does not of itself dictate that implicatures and the fact of utterance must be allowed to contribute to a notion of metaphorical meaning. 4.2 Bringing a generator or believer into the picture Section 4.1 only addressed what happens in interpretation in “bare” mode, so to speak – i.e. when the understander is taking an utterance at face value – and not, for example, considering what the generator intended to convey by it. Such communicative intentions are of course a central concern in pragmatics. The main point we wish to make is that the freedom of inference advocated in Section 4.1 as part of interpretation does not imply that all the inferences we discussed the understander making would also be attributed by her to the generator. Let us call the understander Undine and the generator George, and let us take George’s utterance to be (14). In considering what George intended in uttering it, Undine can consider what inferences George might himself draw on the basis of (14) in uttering it, and presume that that is what George means. Ideally, she would suitably take into account what she knows or conjectures about George’s views about physical hippos, etc., the particular people being referred to as hippos, and so forth. Since Undine only has limited knowledge of George’s mind, and only limited capacity for wondering what it contains, it is likely, on the one hand, that the amount of inferencing that she ascribes to George is considerably more limited than what she would do in understanding (14) for her own purposes, and, on the other hand, the inferencing she ascribes to George may conceivably go off in different directions because of her knowledge of beliefs that George has that she herself does not hold. Thus, it may be that Undine infers from (14) that the “hippo” people are
Metaphor and its unparalleled meaning and truth
113
engaging in FaceBook chat, but does not thereby hold that George infers it, and therefore does not hold that George intended to convey it. Additionally, even when she assumes that George inferred something X, she may take into account how uncertain the inferencing to X is, how difficult the inferencing is, how extensive it is, etc. before concluding that George intended to convey X. In summary, to talk about a particular conclusion X being part of an understander’s actual interpretation of a metaphorical utterance when she takes it as face value for her own purposes has no firm implications for X being part of the content associated with the utterance for other purposes. Notice, however, that Undine might not in fact consider George’s thoughts in any detailed way. In many ordinary circumstances, especially where she (consciously or unconsciously) considers her own interpretation of (14) to be obvious and immediate to everyone, she may just assume that it is what George intended to convey, even when she has no particular reason to think that George shares with her all the beliefs needed to draw the inferences involved. Undine’s considerations about what George may have intended help in analysing at least some cases of when Undine says something like one of the following, as a response to George’s utterance of (14): (15) I agree (16) I disagree (17) George says that the hippos are wallowing in their swamp. In such cases it is possible that Undine means to agree with, disagree with or report what she supposes George to have intended to convey by (14). There may also be additional restrictions on inferencing that are implied by the verbs “agree”, “disagree” and “say”. Of course, another thing Undine might be up to in uttering (15) or (16) is to be alluding to (some portion of) her own inferences, ignoring any consideration of something different that George himself intended. (She may not have occurred to her that he might have intended something different, or she may be exploiting his utterance to her own ends, and so on) Or, she may be meaning “says” in a very narrow way, so that she is merely reporting that George said something lexically close to or identical to “the hippos are wallowing in their swamp”. While our account allows for an understander not ascribing all her own inferences to the generator, or not including them in (dis)agreements and reports of the utterance, the general thrust of Section 4.1 still holds. The source-side inference and the target-side inference (and literal-case inference if the utterance in question were taken literally) are not differentially affected in any a priori way by the consideration of George’s own inferences or by any additional inferential restrictions involved in her uses of “agree” or “says”. In considering George’s intentions, the overall amount of inference she does on his behalf may be less than she would do in interpreting (14) for herself. But this restrictiveness does not of itself impinge in a special way on any particular
114
John A. Barnden / Alan M. Wallington
aspect of inference. In any case, as pointed out in Section 4.1 there will always be overall time and effort restrictions anyway. A further consideration in the semantics and pragmatics of metaphor has been embedding of metaphor within attitudinal contexts such as belief contexts, as in the following: (18) George believes that the hippos wallowed in their swamp. This type of embedding of metaphor has seen some detailed discussion in recent years (Camp 2005, Stern 2000, 2006,) and brief discussion going back longer (van Dijk 1980), and is an important concern in our own research (Barnden 1989, 1999 and forthcoming). Overall, however, attitudinal embedding has been a rather neglected issue in the metaphor field even though crucial in the philosophy of meaning generally (see, e.g., Linsky 1983). One (“opaque” or de-dicto) interpretation of (18) involves attributing to George a use in his thought processes of the metaphorical view of the people and their project as hippos and a swamp. Another (“transparent” or de-re) interpretation is to take the generator (Undine) of (18) to be using the metaphor to describe the situation of the “hippo” people engaged in their project; George is reported to believe this target situation to have occurred, but it is left open how George himself thinks of the situation. It is the opaque reading that concerns us here. (There are also classical de-re/de-dicto issues surrounding the use of the description “The hippos”, putting aside its metaphoricity, but we do not go into this matter here. In addition, an opaque/transparent divide holds also for the said-that case in (17), but we implicitly adopted the opaque interpretation above.) We suggest that (18) can be treated in a way analogous to the treatment of George’s intention in uttering (14). That is, in understanding (18) we consider George to be having a thought that is couched metaphorically, viewing the people concerned as hippos and viewing their behaviour as wallowing. (Note, however, that we are not necessarily bringing in any involvement of George with the sentence (14), and do not even assume that George is an English speaker.) If we bother to and have time, we may take into account any special knowledge that we have about George’s beliefs. Once again, if we do try to take into account the possible differences of George from ourselves, what we infer on George’s behalf is likely to be more limited than what we infer when just understanding (14). And, again, any additional restrictiveness of inferencing does not differentiate between target-side and source-side inferencing. There is a possible problem, however. If inferencing that we surmise to be conducted by George (for instance when we are considering his intentions in uttering (14), or when we are understanding (18), is more restricted in quantity than the inferencing we would do in bare understanding of (14), it could in principle happen in a particular case that the inferencing is not enough to do a full job, or any job, of linking up a lexical meaning of “the hippos are wallowing in their swamp” to mappings or to reach a sufficiently broadened/
Metaphor and its unparalleled meaning and truth
115
narrowed concept, etc. In such a case, only part of the task of metaphor understanding can be done, and only a partial metaphorical interpretation, or no metaphorical interpretation, can be constructed. So, does this show that actually we have to allow all the required source-side inference to go through, to ensure that mappings are used or broadening done, etc., with that result that there is after all an a priori difference between the source-side and the targetside as regards amount of inferencing? The answer is no, because it is always on the cards in discourse that we may fail to understand matters such as what a generator intends to convey by something, or the full nature of someone’s reported belief, and a reason for such failure can be lack of time or unwillingness to exert the required effort. In interpreting (18) we may need to be content with the information that George has some thought that that can be metaphorically couched as (14). This is not much different in principle to being only partially informed of George’s belief state by a sentence such as “George believes that the hippos are doing their least favourite thing” where the that-complement is to be taken non-metaphorically and we have no idea what George might think is the hippos’ least favourite thing. We give another reason for being comfortable to a negative answer to the above question in the next section. 5. Combining the consequences, and truth We have presented the holism considerations of section 3 and the inferenceliberality considerations of section 4 separately, as neither argument as presented depends on the other. However, we now claim that the two sets of considerations support each other. This confluence then sets the stage for discussing the notion of truth, with special but not exclusive attention to the metaphor case. First, let us consider the support of section 4 for section 3. It is of course only in very special cases that we can talk of inferences arising from just an individual sentence. Usually, background knowledge is needed, or information arising from other parts of the discourse containing the sentence. For instance, the reason that an understander might take (14) “The hippos are wallowing in their swamp” to imply that the people are engaging in FaceBook chat could depend not only on prior knowledge about the people concerned, the nature of the project, etc., but could also use extra information derived from the discourse itself. But then it is already at best an approximation and at worst a gross misattribution to say that the FaceBook conclusion is inferred from (14). And, the more broadly one allows the net of inference to spread from a sentence during understanding, the more we can expect that any given conclusion will actually depend on multiple parts of the discourse. Thus, section 4 supports the suggestion of section 3 that the sentence, or closely related units such as the clause, are not the right unit for meaning attributions, and indeed that there is no rigid unit in grammatical terms.
116
John A. Barnden / Alan M. Wallington
Secondly, the support of section 3 for section 4. The more that one considers multi-clause collections in the spirit of section 3, the more that one needs to bring in processes of inference to link the clauses together in understanding, and the more that the notion that the inferences can be limited (in extent, or to particular types) loses plausibility. There is another, more special way in which Section 3 supports Section 4. Recall the question posed near the end of Section 4. The question worries about sentences failing to receive metaphorical interpretations because of additional restrictions on inferencing imposed by considering a generator or believer George. But, the point Section 3 is that we do not assume that all metaphorical sentences should be given interpretations anyway. If George utters (12) (the cat-flap example chopped up into sentences), then we should not assume that George intends to say anything about the target scenario just by his utterance of the sentence “They had their own cat flaps”. Equally, someone might tell us that “George believes they had their own cat flaps” as part of an explanation of George’s beliefs about a situation such as that described by (12). We should not assume that there is a portion of the target scenario that is being metaphorically described by the that-complement here. Thus, the fact that inferential restrictions could lead to a lack of metaphorical interpretation is no special worry and does not raise the need for an exception to the inferential liberality argued for in section 4.1. Now we turn to the question of truth. First, it would be perfectly possible to define a notion of truth-for-understander(-on-this-occasion) based on meaning-for-understander(-on-this-occasion), where the latter is just the sum total of all the conclusions that the understander happens to infer from the piece of discourse being considered. We just say that whatever piece of discourse is deemed to have some meaning M (in that sense), that piece is true if and only all the component propositions in M are true. However, it is not clear that that notion would be of much use in practice. The bigger M can be, the more unlikely it is that it would all actually be true. And indeed, it is not clear why the understander him/herself should be concerned about whether the whole of M is true as opposed to being concerned about whether particular parts of M are true. After all, in normal circumstances an understander is not concerned about the exact properties of pieces of discourse but about what consequences they may have for her. From these considerations we are inclined to take the position that it is in general only a convenient approximation to think, sometimes, of linguistic entities (pieces of discourse) as having truth values, even for particular understanders on particular occasions. Rather, it is the inferred propositions forming the meanings of discourse pieces that have truth values, properly speaking. Talk of discourse pieces themselves having truth values is, when useful at all, just a rough shorthand for talking about the propositions in their meanings having truth values; and degrees of truth of a piece would in practice normally be more interesting than black-or-white true/false values, because of the possible multiplicity of propositions in the meaning of the piece. But, the more
Metaphor and its unparalleled meaning and truth
117
limited a piece of discourse happens to be, and the smaller the set of inferred propositions, the more useful and justifiable for practical purposes it becomes to say that that piece itself has a truth value. If the piece is the sentence “It is 5pm” and the only proposition the understander infers is that it is now approximately 5pm (allowing for loosening during the interpretation process), then we can easily and usefully say as a convenient shorthand that the sentence is true (for the understander on this occasion) on the basis of that proposition being true. Or, we as theoreticians can take the notion that the sentence “It is 5pm” being true just to be convenient simplification of the real truth, namely that some segment of discourse containing that sentence has as its understander meaning some set of propositions containing the proposition that it is now about 5pm, and that proposition amongst other inferred ones is true. On the other hand, if Undine says “That’s true” in response to George’s (14), “The hippos are wallowing in their swamp“, the situation is very much as when she says “I agree”. One option for what is now operative is the set of propositions she surmises is formed by George. Pedantically, we could take Undine to be saying that all those propositions are true. But actually there is no need to assume that when someone says “That’s true” that they agree with absolutely everything they take to be covered by the “That”, any more than when someone says “That car is red” they mean that every single bit of the car is red. And an understander’s attribution of truth may be influenced not just by what proportion of propositions in the meaning are true but also by such matters as her judgments of the relative importance, salience, relevance, newness or interestingness of the propositions. 6. Summary Considerations of metaphor have led us to a view of propositional meaning that is holistic in having only fuzzy connections to grammatical entities such as clauses sentences in discourse, and is liberal in the amount of inferencing involved; and to a view of truth as generally applicable to linguistic entities only in a fuzzy and approximative sense. The considerations involved in our arguments apply to non-metaphorical as well as to metaphorical discourse, but metaphor brings to the fore certain issues that would otherwise be easier to turn one’s gaze from. The article has made two main arguments, that are independently supported (and may independently succeed or fail) but that are also linked. First, we have argued that as an aspect of the non-parallelism that metaphor can exhibit, not all metaphorical sentences, clauses or other potentially proposition-bearing linguistic units should be regarded as individually having a meaning in terms of the target, as opposed to conspiring with other constituents to lead to a target meaning (a metaphorical meaning). Thus, the concentration on the meaning of individual metaphorical sentences (or sometimes clauses) in the technical literature on metaphor semantics, even when the
118
John A. Barnden / Alan M. Wallington
consideration takes into account contextual effects, is too limiting. Rather, metaphorical meaning is in general a matter of chunks of discourse, of no definite size or shape. But then it is only consistent to expect that meaning in general, not just metaphorical meaning, should be regarded as attached to discourse chunks, not individual sentences, except that we have assumed that we can assign a lexical meaning to each sentence. None of this militates against the semantic/pragmatic importance of any sentence, clause, etc. that does not get its own (non-lexical) meaning: it can still be crucial in building the overall meaning of a discourse chunk. We have also argued that target-side inference in metaphor interpretation should be no more restricted than source-side inference, so that if we are very liberal about the latter we must be very liberal about the former. The argument also suggests that the inferencing allowed in deriving literal from lexical meanings should be of the same liberality as the source-side and target-side inferencing in metaphor. We tentatively proposed, in fact, that the understander’s own actual meaning for an assertive utterance on a particular occasion should just be the collection of propositions derived form the utterance by the understander on that occasion. However, this meaning is not necessarily the same as the meaning the understander would surmise that the generator of the utterance intended; the understander may (if being sufficiently careful and being prepared to put in enough effort and time) conjecture what propositions the generator would associate with the utterance. Such conjecturing is affected by what assumptions the understander makes about how the generator’s beliefs differ from his/her own, and is affected by the amount of time and effort the understander puts in, but is not restricted a priori in terms of extent or type of inferencing the understander ascribes to the generator. We showed that the argument from non-parallelism and the arguments concerning liberality of inference reinforce each other, leading overall to a loose and distant connection between linguistic entities on the one hand and meanings and truth values on the other. Indeed in the case of truth our arguments lead to the suggestion that talk of truth of lexical entities such as sentences is only an approximation in general, adopted for reasons of practical convenience, although often the approximation can be a good one. Acknowledgments This research was supported by grants GR/M64208 and EP/C538943/1 from the Engineering and Physical Sciences Research Council of the UK.
Metaphor and its unparalleled meaning and truth
119
References Barcelona, Antonio: “Clarifying and applying the notions of metaphor and metonymy within cognitive linguistics: An update”, in: Dirven, R., R. Pörings: Metaphor and Metonymy in Comparison and Contrast. Berlin–New York: Mouton de Gruyter 2002, 207–77. Barnden, John A.: “Towards a paradigm shift in belief representation methodology” in: Journal of Experimental and Theoretical Artificial Intelligence, 1(2) (1989), 133–61. – “Combining uncertain belief reasoning and uncertain metaphor-based reasoning”, in: Procs. Twentieth Annual Meeting of the Cognitive Science Society. Mahwah, N.J.: Lawrence Erlbaum 1998, 114–19. – “An implemented system for metaphor-based reasoning, with special application to reasoning about agents”, in: C.L. Nehaniv (ed.): Computation for Metaphors, Analogy, and Agents, Lecture Notes in Artificial Intelligence, Vol. 1562. Springer 1999, 143–53. – “Uncertainty and conflict handling in the ATT-Meta context-based system for metaphorical reasoning”, in: Akman, V., P. Bouquet, R. Thomason, R.A. Young (eds.): Procs. Third International Conference on Modeling and Using Context. Lecture Notes in Artificial Intelligence, Vol. 2116. Berlin: Springer 2001a, 15–29. – Application of the ATT-Meta metaphor-understanding approach to selected examples from Goatly. Technical Report CSRP-01-01, School of Computer Science, The University of Birmingham, U.K. 2001b. – “Artificial intelligence, figurative language and cognitive linguistics”, in: Kristiansen, G., M. Achard, R. Dirven, F.J. Ruiz de Mendoza Ibáñez (eds.): Cognitive Linguistics: Current Applications and Future Perspectives. Berlin: Mouton de Gruyter 2006a, 432–59. – “Consequences for language learning of an AI approach to metaphor”, in: Salazar, J., M. Amengual, M. Juan (eds.): Usos Sociales del Lenguaje y Aspectos Psicolingüísticos: Perspectivas Aplicadas. Palma, Mallorca: Universitat de les Illes Baleares 2006b, 15–57. – “Metaphor and artificial intelligence: Why they matter to each other”, in: Gibbs, R.W. Jr. (ed.): The Cambridge Handbook of Metaphor and Thought. Cambridge, U.K.: Cambridge University Press 2008, 311–38. – “Cognitive linguistics must nest inside itself: Considerations arising from an AI project on metaphor”, in: M. Brdar (ed.): Converging and Diverging Trends in Cognitive Linguistics. John Benjamins (forthcoming). – Sheila R. Glasbey, Mark G. Lee, Alan M. Wallington: “Varieties and directions of inter-domain influence in metaphor”, in: Metaphor and Symbol, 19(1) (2004), 1– 30. – , Mark G. Lee: “Understanding usages of conceptual metaphors: An approach and artificial intelligence system”, in: Tech. Rep. CSRP-01-05, School of Computer Sci., University of Birmingham, UK. 2001. Brians, Paul: Common Errors in English Usage. Wilsonville: Franklin, Beedle & Associates, Inc. 2003. Camp, Elisabeth: “Josef Stern, Metaphor in Context”, in: Noûs, 28(4) (2005), 715–31. Camp, Elisabeth: “Contextualism, metaphor, and what is said”, in: Mind and Language, 21(3) (2006), 280–309. Crisp, Peter: “Allegory, blending, and possible situations”, in: Metaphor and Symbol, 20(2) (2005), 115–31.
120
John A. Barnden / Alan M. Wallington
Davidson, Donald: Inquiries into Truth and Interpretation. Oxford, UK: Oxford University Press 1984. Falkenhainer, Brian, K.D. Forbus, Dedre Gentner: “The Structure-Mapping Engine: algorithm and examples”, in: Artificial Intelligence, 41(1) (1989), 1–63. Fass, Dan: Processing metaphor and metonymy. Greenwich, Connecticut: Ablex 1997. Fauconnier, Gilles, Mark Turner: “Conceptual integration networks”, in: Cognitive Science, 22(2) (1998), 133–87. Gentner, Dedre, Brian Bowdle: “Metaphor as structure-mapping”, in: Gibbs, R.W. Jr. (ed.): The Cambridge Handbook of Metaphor and Thought. Cambridge, U.K.: Cambridge University Press 2008, 109–28. – , Brian Falkenhainer, Janice Skorstad: “Viewing metaphor as analogy”, in: Helman, D.H. (ed.): Analogical reasoning. Dordrecht: Kluwer 1988. Glucksberg, Sam: “How metaphors create categories – quickly”, in: Gibbs, R.W. Jr. (ed.): The Cambridge Handbook of Metaphor and Thought. Cambridge, U.K.: Cambridge University Press 2008, 311–38. – , Boaz Keysar: “Understanding metaphorical comparisons: beyond similarity”, in: Psychological Review, 97(1) (1990), 3–18. Goatly, Andrew: The language of metaphors. London–New York: Routledge 1997. Goossens, Louis: “Metaphtonymy: the interaction of metaphor and metonymy in expressions for linguistic action”, in: Cognitive Linguistics, 1 (1990), 323–40. Haser, Verena: Metaphor, metonymy and experientialist philosophy: Challenging cognitive semantics. Berlin–New York: Mouton de Gruyter 2005. Hobbs, Jerry R.: Literature and cognition. CSLI Lecture Notes, No. 21, Stanford University 1990. – “Metaphor and abduction”, in: Ortony, A., J. Slack, O. Stock (eds.): Communication from an Artificial Intelligence Perspective: Theoretical and Applied Issues. Berlin: Springer-Verlag 1992, 35–58. Holyoak, Keith J., L. R. Novick, E.R. Melz: “Component processes in analogical transfer: mapping, pattern completion, and adaptation”, in: Holyoak, K.J., J.A. Barnden (eds.): Advances in Connectionist and Neural Computation Theory, Vol. 2. Norwood, N.J.: Ablex Publishing Corp. 1994, 113–80. – Paul Thagard: “Analogical mapping by constraint satisfaction”, in: Cognitive Science, 13(3) (1995), 295–355. Kittay, Eva F.: Metaphor: its cognitive force and linguistic structure. (Paperback ed.) Oxford, U.K.: Clarendon Press 1989. Lakoff, George: “The contemporary theory of metaphor”, in: Ortony, Andrew (ed.): Metaphor and Thought. Cambridge, UK: Cambridge University Press 1993, 2nd ed. – “The neural theory of metaphor”, in: Gibbs, R.W. Jr. (ed.): The Cambridge Handbook of Metaphor and Thought. Cambridge, U.K.: Cambridge University Press 2008, 17–38. Langlotz, Andreas: Idiom creativity: A cognitive-linguistic model of idiom-representation and idiom-variation in English. Amsterdam–Philadelphia: John Benjamins 2006. Lee, Mark G., John A. Barnden: “Reasoning about mixed metaphors with an implemented AI system”, in: Metaphor and Symbol, 16(1&2 (2001), 29–42. Linsky, Leonard: Oblique contexts. Chicago: University of Chicago Press 1983. Narayanan, Srini: “Moving right along: A computational model of metaphoric reasoning about events”, in: Proceedings of the National Conference on Artificial Intelligence (AAAI ’99). AAAI Press 1999, 121–8.
Metaphor and its unparalleled meaning and truth
121
Peirsman, Yves, Dirk Geeraerts: “Metonymy as a prototypical category”, in: Cognitive Linguistics, 17(3) (2006), 269–316. Ritchie, L. David: Context and connection in metaphor. Basingstoke, U.K–New York: Palgrave Macmillan 2006. Sperber, Dan, Deirdre Wilson: “A deflationary account of metaphor”, in: Gibbs, R.W. Jr. (ed.): The Cambridge Handbook of Metaphor and Thought. Cambridge, U.K.: Cambridge University Press 2008, 84–105. Steen, Gerard: “The paradox of metaphor: Why we need a three-dimensional model of metaphor”, in: Metaphor and Symbol, 23(4) (2008), 213–41. Stern, Josef: Metaphor in Context. Cambridge, Mass.–London: Bradford Books, MIT Press 2000. – “Metaphor, literal, literalism”, in: Mind and Language, 21(3) (2006), 243–79. – “Metaphor, semantics, and context”, in: Gibbs, R.W.Jr. (ed.): The Cambridge Handbook of Metaphor and Thought. Cambridge, U.K.: Cambridge University Press 2008, 206–79. Tendahl, Markus, Raymond W. Gibbs, Jr.: “Complementary perspectives on metaphor: Cognitive linguistics and relevance theory“, in: Journal of Pragmatics 40(11) (2008), 1823–64. Turner, Mark, Gilles Fauconnier: “Metaphor, metonymy, and binding”, in: Barcelona, Antonio (ed): Metaphor and Metonymy at the Crossroads: A cognitive Perspective. Berlin–New York: Mouton de Gruyter 2000, 133–45. van Dijk, Teun A.: “Formal semantics of metaphorical discourse”, in: Ching, M.K.L., M.C. Haley, R.F. Lunsford (eds.): Linguistic Perspectives on Literature. London: Routledge and Kegan Paul 1980, 115–38. Vega Moreno, Rosa E.: Creativity and convention: The pragmatics of everyday figurative speech. Amsterdam/Philadelphia: John Benjamins 2007. Wearing, Catherine: “Metaphor and what is said”, in: Mind and Language, 21(3) (2006), 310–32. Wilks, Yorick: “Making preferences more active”, in: Artificial Intelligence, 11(1978), 197–223. Wilson, Deirdre, Robyn Carston: “Metaphor, relevance and the ‘emergent property’ issue”, in: Mind and Language, 21(3) (2006), 404–33.
Mark Lee (Birmingham, GB)
Truth, metaphor and counterfactual meaning
0. 1. 2. 3. 4. 4.1 4.2 4.3 4.4 4.5 5. 6.
Introduction A classical truth-based account of counterfactual reasoning Metaphorical reasoning Counterfactual reasoning Metaphorical counterfactuals One-sided metaphorical counterfactuals Double-sided metaphorical counterfactuals Denying the metaphor Extending the metaphor Specifying the metaphor Understanding metaphorical counterfactuals Further discussion “Metaphors ask us to imagine the world in a new way, while conditionals may ask to imagine a new world.” Cohen (1998)
0. Introduction There is a theory of language which argues that the syntax of an utterance acts as an input to a context-free, truth-conditional semantic representation (e.g. see Lewis 1972) which may be extended by additional reasoning to capture its pragmatic (or intended) meaning (e.g. Grice 1975). Henceforth, this view will be referred to as the “Classical” view of meaning. Two types of phenomena are troublesome to this theory: metaphors and counterfactuals. Metaphorical utterances present a problem since the theory cannot explain neither the systematic nature nor the pervasiveness of metaphorical language. Often, metaphors are literally false yet present little difficulty in their interpretation. Similarly, truth-conditional views of counterfactuals have failed to match our intuitions. As will be explained, logically, all counterfactuals are literally true yet this does not accord with our intuitions as to what is an acceptable or unacceptable counterfactual expression. This chapter will argue that both difficulties are due to metaphors and counterfactuals sharing the same cognitive processes.
124
Mark Lee
The similarity between metaphors and counterfactuals has also been pointed out by Cohen (1998) who argues that there are two main types of philosophical argument: metaphorical and counterfactual. For example, a philosopher might describe one concept in terms of another in a metaphorical argument or ask “what if” question and then come to a conclusion based on its answer in a counterfactual argument. Cohen argues that many philosophical misunderstandings are due to philosophers mixing metaphorical and counterfactual arguments. Because counterfactual reasoning requires reasoning about a novel world (one where the antecedent is true, unlike the real world) the content of the counterfactual must be semantically unambiguous so that the hearer can imagine the correct world within which to interpret the counterfactual. Therefore, because metaphors are “semantically loose” Cohen argues that an argument which mixes metaphors and counterfactuals is likely to be philosophically nonsensical. However, in mundane language it is apparent that counterfactuals are often metaphorical. Furthermore, as this chapter will argue, there is a rich interaction between the two types of trope. This should be of no surprise if the account detailed below is correct since both phenomena require the same cognitive mechanisms in their interpretation and therefore can be easily mixed. In the next section I will review a classical account of counterfactual reasoning based on the truth-conditions of the counterfactual expression. I will argue that such accounts are misguided and ignore the figurative aspect of counterfactuals: rather than being truth-conditional, counterfactual meaning is essentially analogical.
1. A classical truth-based account of counterfactual reasoning As an aid to discussion, consider the counterfactual: (1) If everybody had left the conference early then I’d be talking to an empty room. Notice that both the antecedent and the consequent are implied to be false (e.g. not everybody had left the conference early and the speaker was not talking to an empty room). Traditionally, the falseness of the antecedent is implied as a presupposition of the counterfactual since it satisfies Strawson’s test (1952) in that if the antecedent were actually true then the counterfactual could not be assigned a truth value and would therefore be meaningless. The falseness of the consequent is a conversational implicature. For example, it could be cancelled by then uttering:
Truth, metaphor and counterfactual meaning
125
(2) ... but I am anyway. The presupposed falseness of the antecedent presents a problem for a truthconditional account. Since any conditional with a false antecedent is trivially true, then according to classical logic, all counterfactuals are tautological. This does not match our intuitions since we clearly distinguish between reasonable and non-reasonable counterfactuals. More specifically, there must be a plausible story imaginable which causally links the truth of the antecedent to the truth of the consequent. This is termed the problem of verification: verifying that if the antecedent were true, then it would be reasonable to assume the consequent would also be true. Many approaches to the verifying counterfactuals have been suggested. The majority follow the spirit of Given: if A then B i. set up a pretence space where A is true. ii. remove any proposition which contradicts A. iii. construct meaning of the counter factual by verifying that A entails B within the pretence space. (Brée 1982)
Step ii is essential since if the antecedent is added to the set of true propositions then a contradiction will arise (since the negation of the antecedent is present). Furthermore, it is not sufficient to remove just the negation of the antecedent: in addition, any proposition which entails the negation of the antecedent must also be removed. This step must be repeated till the set of true propositions is consistent. An obvious answer would be to remove any proposition which would be false if the antecedent was true. However, as Goodman (1947) points out, this process can be expressed as another counterfactual and so a circularity is introduced. Goodman termed this the contenability problem i.e. the problem of specifying which propositions were contenable with the counterfactual. Goodman proposes that a counterfactual is acceptable, if and only if the antecedent conjoined with relevant true statements about the attendant circumstances leads by way of true general principle to the consequent. (Ibid.: 37)
The contenability problem therefore is the problem of finding the correct set of relevant true statements which do not contradict the antecedent but support the entailment of the consequent. Many formal semantic solutions have been suggested to the contenability problem, a representative solution being Lewis’ (1972) possible-world semantics approach. Given a counterfactual all the possible worlds where the antecedent is true are ordered by their similarity to the real world. A counterfactual is reasonable by Lewis’ account just in case all of the possible worlds
126
Mark Lee
where the antecedent is true which are close to the real world, also feature the consequent being true. Lewis’ account of similarity is based on ranking the importance of the difference between two possible worlds based on whether the difference between the worlds involves the widespread violation of laws of nature, a factual difference over a large region of space/time, a minor violation of a law of nature or a difference in some particular fact respectively. There are several obvious problems with such a notion. First, similarity judgements are inherently contextual. Any two (distinct) worlds will share many different properties and fail to share many other different properties. Therefore, rather than a general notion of similarity, what is required is a notion of specifying “similar in what way” i.e. what is a relevant similarity and what is a relevant difference. It could be argued that relevance is specified by the requirements of the causal relationship between the antecedent and the consequent. However, this is also not a satisfactory solution due to its circularity. The set of propositions true in the counterfactual world must be specified before the causal relationship is examined.1 The second problem is related to the first: counterfactuals and the contexts in which they are uttered often do not provide the necessary detail to specify the counterfactual world they refer to. This can be due to the often figurative aspect of counterfactuals; e.g. consider Quine’s example: (3) If Julius Caesar was in charge of the Korean War then he’d have used the atom bomb. (Quine, quoted by Lewis [1973]) Example 3 is a blend of two different domains: the “Roman empire” and the “Korean war”. However, the counterfactual does not explicitly specify which aspects of each domain should be considered when verifying the counterfactual. Instead, the hearer must imagine a possible causal relationship between the antecedent and the consequent in the counterfactual world and then supply the details of what is mapped from the respective domains (i.e. Caesar’s ruthlessness blended with modern day war technology). However, consider the possible counterfactual reply: (4) No, he would have used slings and arrows. In this counterfactual, Caesar’s lack of knowledge of military war technology is imported into the counterfactual world to deny the verification of the first counterfactual. Notice though that the validity of either counterfactual is not affected since both construct novel (and therefore separate) worlds. Also, there is no intuitive way to order such counterfactual worlds by their similarity to the real world, since both are highly figurative and fantastical worlds.
————— 1
It is worth pointing out at this point that the problem of specifying what is relevant between two worlds is analogous to the problem in metaphor interpretation where two concepts are compared.
Truth, metaphor and counterfactual meaning
127
2. Metaphorical reasoning What follows is a brief outline of a model of metaphorical reasoning which has been computationally implemented in a more general account of reasoning. However, the precise details are not required for the present purposes of this chapter (but see Barnden & Lee [1999] for further information; see also Barnden & Wallington, this volume, 87–123). Following Lakoff (1992), metaphor expressions operate at a conceptual rather than linguistic level. That is to say, a metaphorical expression specifies a mapping between two different conceptual domains and that the same metaphor can be manifested in different linguistically realized expressions. For example, expressions like boiling with rage, hot under the collar are examples of the same metaphor, EMOTIONS ARE HEAT. The majority of metaphors found in mundane discourse are conventionalized and therefore, the mappings they involve between conceptual domains are also conventionalized so that for a given metaphor there are a finite number of mappings between the source and the target. This allows for metaphors to be systematic. For example, assuming there is a mapping between emotional intensity and temperature we can analyze expressions such as he cooled down or his blood ran cold as manifestations of the same metaphor as above but as expressing an opposite connotation. Metaphorical reasoning proceeds as follows. Given a metaphorical expression, a space is created within which the propositional content of the expression is treated as literally true. For the remainder of this chapter, such spaces will be referred to as metaphorical spaces. Within this space, reasoning is performed in terms of the source domain of the metaphor until a connotation can be mapped over to the target domain via the conceptual mapping. A simple example using the metaphor introduced above will illustrate this, (5) John’s anger boiled over. A metaphorical space is constructed within which the utterance is treated as literally true. This metaphorical space contains domain-specific knowledge about temperature and that things that are boiling are hot. Therefore, it can be inferred that if anger is boiling then it is hot. Given that there is a conceptual mapping between temperature and emotional intensity, the result of this inference can be mapped over to the target domain to suggest that John’s anger increased. This is represented in Figure 1.
128
Mark Lee
Restricting all metaphorical reasoning to a localized space means that an expression can be reasoned about in terms of its truth within the conceptual metaphor regardless of whether it is literally true. In addition the mapping between the two spaces provides an elegant way to deal with the uncertainty of many metaphors since any transfer of information can be treated as an uncertain inference which can be overruled by more certain information in either space. As we shall see, such methods of dealing with the importation of uncertain information can also be applied to the problem of counterfactual reasoning.
3. Counterfactual reasoning What follows is a brief sketch of an account of counterfactual reasoning which has been implemented within the same general reasoning framework used above for metaphorical reasoning. A more detailed computational exposition of counterfactual reasoning is given in (Lee 2001). Following Fauconnier (1997) a counterfactual of the form If antecedent then consequent constructs a mental space within which the counterfactual is reasoned about. For the rest of this chapter I will refer to such spaces as counterfactual spaces. Like a metaphorical space, a counterfactual space need only be internally logically consistent and therefore there is no requirement for a counterfactual space to be true to the real world. In particular, within the counterfactual space, the antecedent is assumed to be true. Verifying the counterfactual involves reasoning about whether the consequent is entailed by the antecedent. Since initially, the counterfactual space is empty of any other proposition this involves the selective importation of true statements from the real world and any conceptual domain referred to by the counterfactual. Inconsistency is avoided by insisting that the only statement which is certainly true within the
Truth, metaphor and counterfactual meaning
129
counterfactual space is the truth of the antecedent. Therefore, any statement which contradicts the antecedent is denied. Such an approach avoids the contenability problem since it builds a sketch of a counterfactual world within which the counterfactual is true rather than attempting to reason about a possible world where any contradictory statement must be first removed. Such a sketch also only specifies the counterfactual world in enough detail to verify the counterfactual and therefore allows for the under-specification of any counterfactual. Once a counterfactual has been verified, its connotation to the target domain (i.e. the subject of conversation) can be determined by drawing an analogy between the counterfactual space and the target domain. This is done by finding an inference chain analogous to an inference in the counterfactual space which can be made in the target domain. This approach to counterfactual can be illustrated by a simple example. Given the conditional, (6) If John had sold his shares, he would have made a profit. and assuming the following facts are known to be true F1 John has shares to sell. F2 John has not sold his shares. F3 The value of his shares has increased since he bought them. F4 Selling shares results in gaining the current value of the shares. The conditional is counterfactual since F2 contradicts the antecedent. Therefore, counterfactual reasoning proceeds as follows: A counterfactual space is constructed and linked to the conceptual space of John’s finance. Within the counterfactual space, the antecedent is treated as true: C1 John sold his shares For the counterfactual to be verified, the consequent must follow. Therefore, an attempt is made to prove the consequent within the counterfactual space: C2 John made a profit Since the space only contains the antecedent, further information must be imported from the source domain via an importation rule.
130
Mark Lee
Importation: Import any fact from any connected source domain into the counterfactual space as required but at a lower level of certainty. This allows for the facts F1, F3, and F4 to be imported into the counterfactual space at a lower level of certainty and thus verify the counterfactual. However, F2, if imported, is denied by the stronger certainty rating of the antecedent C1. This is shown in Figure 2.
The consequent follows from the antecedent together with the set of imported facts thus verifying the counterfactual. One possible connotation can be drawn from the counterfactual by considering the analogy between the inference chain of John selling his shares and thereby making a profit in the counterfactual space and John not selling his shares and therefore presumably not making a profit in the real world and thus implying that John should have sold his shares. There is an important difference between the approach sketched above and classical accounts of counterfactual reasoning. In the classical view, counterfactual reasoning is a case of reasoning about the truth conditions of a counterfactual over possible worlds. Such worlds, though imaginary, are fully specified (i.e. the set of truthful propositions within a possible world are specified). However, in the approach argued above, counterfactual reasoning involves the mapping for propositions into a space which is not fully specified and therefore there is no constraint on whether there is indeed a fully specified possible world within which the counterfactual could be true. Therefore, truth within the counterfactual space is subjective rather than objective in that further specification of the space could invalidate the verification of the counterfactual.
Truth, metaphor and counterfactual meaning
131
4. Metaphorical counterfactuals The proceeding two sections have sketched accounts of metaphorical and counterfactual reasoning based on the same notion of mapping conceptual information between domains. In addition both accounts have involved the mapping of information from the real world to some fictional world where the content of the statement is treated as literally true and then the inference resulting is mapped back to the real world as a connotation of the expression. It is worthwhile considering what happens when both tropes are mixed. i.e. metaphorical counterfactuals where either or both the antecedent or consequent is a metaphor. As previously noted in the introduction, Cohen (1998) argues that the mixing of metaphors and counterfactuals creates serious problems of interpretation in philosophical argument. This is because metaphors are “semantically loose”, that is to say their interpretation is highly contextually dependent and so if a metaphor occurs within a counterfactual then because of its underdetermined semantic content there are too many possible counterfactual worlds to which it refers to for it to be objectively interpreted. However, the account above suggests that indeterminacy is a central aspect of both tropes and therefore cannot be avoided. Cohen’s work is prescriptive in that he argues that philosophers should attempt to avoid metaphorical counterfactuals. However, counterfactuals have no special status in mundane discourse, and so feature metaphorical language as often as any other type of expression. However, as we shall see, the degree of metaphoricity depends on the construction of the counterfactual. In particular, we can distinguish between two types: one-sided metaphorical and double-sided metaphorical counterfactuals.
4.1 One-sided metaphorical counterfactuals One-sided metaphorical counterfactuals are counterfactuals where either the antecedent or the consequent is metaphorical. For example, consider the following two counterfactuals (7) If John had put that idea to one side, then he would have been happy. (8) If Bill had been late again, Sheryl would have blown her top. In (7), the antecedent is metaphorical whilst in (8), the consequent is. From an initial analysis of counterfactuals from Cobuild’s Bank of English corpus, it appears that one-sided metaphorical counterfactuals almost always involve highly conventionalized conceptual metaphors. The only exceptions to this are where a metaphor is previously introduced in the discourse and then reused by the counterfactual. This accords with Cohen’s analysis since if an
132
Mark Lee
one-sided counterfactual involves a highly conventionalized metaphor then the conceptual mappings between source and target will be relatively fixed and therefore allow for the hearer to construct the correct counterfactual world.
4.2 Double-sided metaphorical counterfactuals Double-sided metaphorical counterfactuals feature a metaphor in both the antecedent and consequent. Typically, the antecedent and the consequent employ the same conceptual metaphor. In such cases there is a strong interaction between the two metaphor usages. Such interactions perform a specific function within the counterfactual. As an example, three functions are described below.
4.3 Denying the metaphor Often the metaphorical expression in the consequent is used to deny the conceptual metaphor introduced by the antecedent. This typically occurs when the antecedent is either highly conventionalized or idiomatic. For example, (9) If time was money then the poor would be rich. The antecedent expresses a highly familiar conceptual metaphor, TIME IS MONEY, in a paraphrase of the idiomatic expression. However, the consequent employs the same conceptual metaphor but with a less familiar mapping (that of the “poor” being owners of “time/money”) to suggest a counterexample to the metaphor since the poor being rich is a contradiction. Denying the overall metaphorical view is a common function of doublesided metaphorical counterfactuals. Typically, the antecedent involves an idiomatic expression while the consequent shows a novel mapping which provides the counterexample. Typically, such counterfactuals are used to connotate that a particular metaphorical view is not appropriate for a given situation.
Truth, metaphor and counterfactual meaning
133
4.4 Extending the metaphor Double-sided counterfactual metaphors also allow less familiar (but not novel) conceptual metaphors to be extended by introducing a metaphor in the antecedent and then providing a novel mapping in the consequent. For example, (10) If war was a market place then the soldiers would be coins. The metaphor WAR AS MARKETPLACE is a less familiar metaphor but is not novel: it is often used in news paper reports dealing with the economics of military conflict. The metaphor is introduced first in the antecedent and then the consequent extends the metaphor by introducing a novel mapping between soldiers in the source domain and currency in the target domain.
4.5 Specifying the metaphor Double-sided metaphorical counterfactuals can also be used to introduce and novel metaphors. For example, (11) If wishes were horses then beggars would ride. The metaphor expressed in the antecedent is clearly novel. Furthermore, without any context, it cannot be sensibly interpreted. However, by introducing a second mapping of the novel metaphor in the consequent, the correct interpretation can be found by attempting to verify the counterfactual using the suggested mappings. This is detailed further in the next section.
5. Understanding metaphorical counterfactuals The approaches to metaphorical and counterfactual reasoning can also be combined to deal with metaphorical counterfactuals such as described above. To illustrate this, re-consider example (11). As before, the counterfactual prompts the creation of a counterfactual space. However, since both the antecedent and the consequent are metaphorical, a metaphorical space is nested within this space. Within the metaphorical space the two metaphorical expressions are treated as being valid mappings between the source domain (the counterfactual space) and the target (the metaphorical space) i.e. Wishes = horses
134
Mark Lee
Beggars = riders Let us assume that it is common knowledge that riders possess horses and this fact is imported into the counterfactual space. possess(riders, horses) According to the mapping expressed in the antecedent, horses can be substituted with wishes so that riders possess wishes. Since the consequent identifies beggars as riders then the counterfactual can be verified by making the inference within the counterfactual space that: possess(beggars, wishes) This inference can then be transferred back to the real world as a connotation of the counterfactual. This reasoning process is shown in Figure 3.
6. Further discussion The sections above have described a model of reasoning about both metaphors and counterfactuals and mixes of the two tropes. As previously discussed, the treatment of metaphors and counterfactuals have each posed problems for the classical view of meaning where literal truth is primary to
Truth, metaphor and counterfactual meaning
135
figurative meaning. It is therefore important to consider how the relationship of “objective” truth is to the approach. Metaphors are often literally false yet can communicate rich and truthful information about the real world. If the model of reasoning described above is correct, then metaphors do so by forcing the hearer to treat them as literally true in an alternative metaphorical space which has conventionalized mappings between the source and target domains. Counterfactuals on the other hand, are always literally true since any conditional with a false antecedent is tautological. However, this clearly does not match our intuitions concerning counterfactuals. Instead, certain counterfactuals are treated as reasonable if they can be verified by showing that if the antecedent was indeed true then the consequent would follow. This is done by constructing a counterfactual space – where the false antecedent is treated as true and then constructing a sketch of a counterfactual world where the consequent follows. After the verification of the counterfactual, its connotation is then derived by drawing an analogy between the counterfactual space and the topic of discourse. Therefore, there is an important distinction between metaphorical and counterfactual reasoning. Metaphorical reasoning often involves treating a statement which is literally false as true and then drawing conclusions about its implications. Counterfactual reasoning involves understanding the connotation of a trivially true statement though one which has as it components two false statements and then drawing a conclusion from its verification. In both cases, the concept of truth is contextually dependent on the particular conceptual space within which reasoning proceeds. Therefore, truth is not an objective or context free notion but instead one sensitive to the particular demands of the space it is in. References Barnden, John A, Mark G. Lee: “An implemented context system that combines belief reasoning, metaphor-based reasoning and uncertainty handling”, in: Proceedings Of the Second International and Interdisciplinary Conference on Modelling and Using Context (Context’99). Trento 1999. Brée, David S: “Counterfactuals and Causality”, in: Journal of Semantics 1.2 (1982). Fauconnier, Gilles: Mappings in Thought and Language. Cambridge: Cambridge University Press 1997. Goodman, Nelson: “The problem of counterfactual conditional sentences”, in: Journal of Philosophy, 44 (1947), 113–28. Lakoff, George: “The Contemporary Theory of Metaphor”, in: Ortony, Andrew (ed.): Metaphor and Thought. Cambridge: Cambridge University Press 1993, 2nd ed., 202–51. Lee, Mark G.: A Computational Account of Conceptual Blending Within Counterfactuals. Cognitive Science Research Papers (CSRP–01–10), School of Computer Science, University of Birmingham 2001. Lewis, David K.: Counterfactuals. Cambridge, Mass.: Harvard University Press 1973. Strawson, Peter Frederick: Introduction to Logical Theory. London: Methuen 1952.
Carla Bazzanella / Lucia Morra (Torino, I)
‘Metaphorical’ truth conditions, context, and discourse*
0. 1. 1.1 1.2 2. 2.1 2.2 2.3 2.4 2.4.1 2.4.2 2.4.3 3. 3.1 3.2 3.2.1 3.2.2 3.2.3 3.3 3.3.1 3.3.2 3.3.3 4.
Introduction Truth and successfulness Metaphor as deviance? Degrees and dimensions of success. Towards context Context and metaphor Global and local context Metaphor and global context Metaphor and local context Possible interaction between global and local context in metaphor Coincidence Partial coincidence No coincidence: When metaphor turns out to be false Metaphor in discourse Interactional and textual levels. How metaphor develops in discourse Conditions of success/failure Consistency Uptake and recognition When metaphors turn out to be “unhappy” Metaphor and dialogic interaction Accepting and appropriating metaphor Refusing metaphor Understanding/misunderstanding metaphor Conclusion
0. Introduction How can one speak about “‘Metaphorical’ truth conditions”? There is an ambivalence in “metaphorical”: one could mean that s/he is dealing with ‘figurative’, ‘translated’ truth conditions1, or2 that the focus is on the truth conditions which are to be assigned – or not assigned – to metaphor. ————–– *
1 2
Though all sections have been planned and revised jointly by both authors, Carla Bazzanella has written the sections 0, 1, 2.1, 2.2, 2.3, 3, 4; Lucia Morra has written the section 2.4. In Burkhardt’s (1990: 326) words: “The contents of the sentences that contain (‘living’ or ‘dead’) metaphors can be valid only metaforice”. Or is to be intended here as inclusive.
138
Carla Bazzanella / Lucia Morra
Once the second meaning is chosen, and metaphorical truth conditions are ‘translated’ into appropriateness conditions (as already proposed by Burkhardt in 19903), the problem is to relate them to context, in a vein similar to the one adopted by Austin with regard to performative utterances: Besides the uttering of the words of the so-called performative, a good many other things have as a general rule to be right and to go right if we are to be said to have happily brought off our action. What these are we may hope to discover by looking at and classifying types of case in which something goes wrong […]: the utterance is then, we may say, not indeed false but in general unhappy. And for this reason we call the doctrine of the things that can be and go wrong on the occasion of such utterances, the doctrine of the Infelicities. (Austin 1962: 14)
The role of context (essential not only in a pragmatic perspective, see e.g. Recanati 1995, Levinson 2000, Givòn 2005) is especially crucial with regard to metaphorical meaning, given both its context-dependence4 and its “coactivation patterns”5. Here, context is held responsible for the “successful” use of metaphor, and will be further specified by distinguishing a “global” (e.g. historical, cultural background) and a “local” context (including its actual use in being deployed, the activation of relevant features and its ‘co-text’, see § 2.1), the latter being partially created by the metaphor itself. Within the recent perspective (cf. e.g. Leezenberg 2001) which claims that the motivation of the metaphorical application of a particular predicate seems to fall outside the reach of semantics6, conditions of both success and infelicity in using metaphor in discourse (taken to mean both spoken and written text) will be proposed here on both interactional and textual levels, and the possible modalities of interaction which follow the use of metaphor will be investigated. In conclusion, the literal/non literal distinction will be rejected, and, in a pragmatic perspective, appropriateness conditions will be claimed to apply not only to metaphor, but to every use of language.
————––
3
4 5
6
“[…] metaphorical expressions are neither true nor false, but at best appropriate or inappropriate to their object.” (Burkhardt 1990: 325) Stern (2000: 14) rightly argues: “[…] that the key to its [=metaphor’s] satisfactory semantic analysis is to embrace its context-dependence.” “In live metaphoric use, both the core node-cluster (literal meaning), and the relevant marginal nodes are co-activate, the first by the expression itself, the second by the discourse context.” (Givón 2005: 84) “A semantic account cannot explain [...] why a specific metaphorically applied expression determines a specific property; this holds in particular for novel and cross-categorial metaphors.” (Leezenberg 2001: 251)
‘Metaphorical’ truth conditions, context, and discours
1.
139
Truth and successfulness
1.1 Metaphor as deviance? Traditionally, from Aristotle until the last century, metaphor has been considered as a deviant use of language7, as it is not based on a direct correspondence between words and worlds. Its “non-denoting”, “senseless/nonsensical”8,“mistake”9, and “parasitic” features have been stressed by relating it to fictional and non-literal language in general. It cannot be denied that metaphor, even though considered as a form of predication, is a rather peculiar one, since “[…it] consciously ascribes properties to objects that (most obviously) do not really have them” (Burkhardt 1990: 320). This kind of explicit violation10 has been explained in different ways (to quote a few: by ellipsis11, by Grice’s (1989) implicatures, by Searle’s (1979) communicative intentions of speakers, cognitive interaction 12), and from different perspectives: substitution views, comparison views, and interaction/cognitive views13. The topic of truth conditions for metaphors as non-literal language has been explicitly tackled, amongst others, by Searle (1979), and discussed by Burkhardt (1990), who rightly criticizes Searle’s view that another sentence explaining the speaker’s intended meaning should be constructed to establish the truth conditions of the metaphorical utterance. In fact, “this sentence would only describe its own truth conditions, not those of the metaphorical utterance” (Burkhardt 1990: 323), and has a circular function, “since it explains speaker’s utterance meaning in terms of truth conditions and vice versa.” (Ibid.) ————––
7
8
9 10 11 12 13
“Cicero, [in] De Oratore […] made the Latin leap explicit as ‘transilire’ […]. Quintilian [in] The Institutio Oratoria […] details metaphorical translatio under the trope, literally the ‘turn of a phrase’, that abides by the same basic principle of violated content selection.” (Gumpel 1988: 57) “[…] metaphorical sentences can be regarded, in the Wittgensteinian sense (cf. 1961: 4.461ff.), as ‘senseless’ (if taken as self-contradictions), or even as ‘nonsensical’ (if taken as sentences that, for lack of corresponding state of affairs, could not have a truth value at all)” (Burkhardt 1990: 324). Cf. Ryle (1949) for the ontological category mistake in relation to metaphor. According to Weinrich (1967: 6), metaphor is the use of a word in a counterdetermining context, that is, a violation of the expected contextual determination. Based on the view that metaphor is a shortened form of comparison, ellipsis has long been considered crucial for metaphor in traditional rhetoric. Cf., from Bühler’s (1934) Überblendungsprojektion (= fading projection) to Turner and Fauconnier’s (1995) blending. For a useful analysis of these three sets of views in the treatment of metaphor, and in particular of Searle’s position in relation to the problem of truth conditions in metaphor, cf. Burkhardt (1990).
140
Carla Bazzanella / Lucia Morra
1.2 Degrees and dimensions of success. Towards context Some time ago, Austin (1961: 130) asked a rhetorical question: “Is it true or false that Belfast is north of London? That the galaxy is the shape of a fried egg? That Beethoven was a drunkard? That Wellington won the battle of Waterloo?” In asking these questions, he was referring to different kinds of statements which relate words and worlds, and even if he was not directly dealing with metaphor (although we can consider as a metaphor the use of “a fried egg” to describe the shape of the galaxy), his answer, in our opinion, can be extended to other uses of language, metaphor included: “There are various degrees and dimensions of success in making statements: the statements fit the fact always more or less loosely, in different ways on different occasions for different intents and purposes. What may score full marks in a general knowledge test may in other circumstances get a gamma.” In order to better evaluate the degrees and dimensions of success of a given metaphor, it is therefore necessary to consider its purposes (as is necessary with regard to any utterance, cf., among others, Wittgenstein 1953: § 88). Using metaphor in everyday language covers a range of functions: expressive, rhetorical-argumentative, explicative, heuristic-predictive, creative (cf., among others, Ortony 1979, Cacciari 1999)14. Both the interactionist (cf., e.g. Bühler 1934, Richards 1936, Black 1962), and the cognitive view (cf., among others, Lakoff & Johnson 1980, Kittay 1987, Indurkhya 1992, Bazzanella & Casadio 1999, Bazzanella et al. 2008, Bazzanella 2009) focus on the “bridging” function of metaphor, which is used to cope and categorize new experiences by exploiting the looseness of facts and categories. Unlike literal meaning, metaphor allows one to cope better with novelty: “[…] even the most adroit of languages may fail to ‘work’ in an abnormal situation or to cope, or cope reasonably simply, with novel discoveries […].” (Austin 1961: 130) The frequency of metaphors not only in everyday language, but also in scientific language – thought to be more ‘approaching the truth’ – can be explained on this basis: “There are reasons to believe that all the typical features of metaphorical language (such as vagueness, ambiguity, polysemy, a shift of meaning, etc.) appear in science [...] in the service of its most vital aim, that is the gaining of new knowledge.” (Radman 1995) Given that the multifarious uses of metaphor ask for different conditions of success, to be valued with reference to the specific function of the metaphor used, the correspondence should be established not only with the external real world, but also with the specific context which is introduced and created by the metaphor itself15. ————––
14 15
The possible reduction to only rhetorical and cognitive functions is not at issue here. In Brünner’s (1987) words, quoted by Burkhardt (1990: 326): “[…] our ‘truths’ are true ‘only relative to the reality defined by the metaphor’.” Eco (2000: 144) speaks about “legittimazione ‘de dicto’”, i.e. ‘de dicto’ legitimization.
‘Metaphorical’ truth conditions, context, and discours
141
Lakoff and Johnson (1980: 3), while stating that: “[…] truth is always relative to a conceptual system that is defined in large part by metaphor”, did not stress what Burkhardt (1990) defines as the constitutive character of metaphor16: i.e. the use of metaphor activates a complex context17 against which to evaluate its relevance, understandability, and, eventually, success. The right way of assessing metaphor should therefore be based on the criterion of success/appropriateness, and not truth. A pragmatic perspective, which stresses the role of context in language, should allow us to consider all uses of language, metaphorical and “parasitic” ones included, by applying the same set of rules, i.e. appropriateness conditions. In Austin’s (1962: 143) words: “Again, in the case of stating truly or falsely, just as much as in the case of advising well or badly, the intents and purposes of the utterance and its context are important; what is judged true in a school book may not be so judged in a work of historical research.” We propose that an answer to our initial problem, how to reject the “deviance view” of metaphor, lies in applying not truth but appropriateness conditions to metaphor and to replace the correspondence view by a view that considers words in context and in discourse. 2. Context and metaphor 2.1 Global and local context In order to analyze the complex role which context plays both in metaphor use and metaphor understanding, it seems useful to distinguish between global and local contexts18. The two levels are actually nested and integrated in any utterance (cf. Bazzanella 1998, 2008), since context is a dynamic process (cf., among others, Mey 1993), but separating them may help with the analysis: – By global context we refer to the a priori or given components of context, i.e. socio-linguistic (such as the physical setting, the social role of participants, cf., among others, Hymes 1974), pragmatic (such as their mutual knowledge/beliefs/intentions, etc.; cf., among others, Levinson 1983), and
————––
16
17
18
“Metaphorical expressions, or rather the speaker’s use thereof, do not assert the truth of the propositions containing them, they rather constitute it.” (Burkhardt 1990: 327) Cfr. Bazzanella (2002) for a discussion of the complex relationship between context and comprehension; Stern (2000) and Leezenbeg (2001) with regard to metaphor in context, and see below. Cf. Bazzanella (1998), Akman and Bazzanella (2003), for this distinction, which aims to integrate two possibly competing notions, and for a more detailed discussion of this topic.
142
Carla Bazzanella / Lucia Morra
cultural parameters which we inherit from and share with our community (cf., Givón’s 1989: 74, “generic focus”19). – By local context we refer to the activated or relevant components of context, i.e. both the ongoing and the shared prior discourse/text20. It is dynamically related both to the current structural21 and linguistic environment in which the metaphor is embedded (the latter being labelled cotext), and constructed in the ongoing interaction/discourse/text as it becomes relevant to it (cf., amongst others, Sperber & Wilson 1986). While the global context is independently shared by the members of a given community, the local context takes shape in the actual process of communicating. Thus, the distinction between the two levels of context, i.e. global and local, is mainly functional: the global context provides the general conditions for communicating, the local one provides the particular conditions for the current act of communication. Furthermore, distinguishing them is relevant in order to establish the “metaphorical truth conditions”: by using a metaphor, a local context is created in which the features of a metaphor which are relevant to the ongoing verbal interaction are selected and interact with the global context (see 2.4). 2.2 Metaphor and global context Both the global and local context affect use, comprehension, and appropriateness of a given metaphor. Let us start by exemplifying this with regard to global context, and with a particular kind of metaphor, that is, metaphorical proper names: (1) Quelques petits Davids vont finir par vaincre un Goliath qu’ils combattent depuis 21 ans. ‘Some small Davids will beat a Goliath whom they have struggled against for 21 years.’ ————––
19
20
21
“Knowledge of (and beliefs concerning) the so-called real world (including society and culture), assumed by the speaker to be held by the hearer, as a member of the same speech community (‘culture’), and manifest first and foremost in the commonly-held lexicon.” (Givón 1989: 74) Our notion of local context includes both Givón’s (1989) “deixis focus” and “discourse focus”. By deixis focus, he refers to shared speech situation, which includes deixis, socio-personal relations, and Speech-Act Teleology, while discourse focus in his words “includes the specific propositions comprising the uttered text, but also whatever other entailed propositions the speaker assumes that the hearer can derive from the text by whatever means.” (1989: 86) For example, in perception, local context is activated in relation to the background, other physical stimuli, etc.
‘Metaphorical’ truth conditions, context, and discours
143
(2) La scène française de notre époque manque de Molières. ‘The French stage of our times lacks some Molières.’ In 1 and 2, taken from Jonasson (1991: 64), David, Goliath, and Molière “are included in our cultural background, which stimulates their ‘extended’ use, but they are extraneous to other cultural backgrounds, e.g. Indonesian or Turkish.” (Bazzanella 2001) Metaphorical proper names refer to a configuration of features, which are known on an encyclopedic basis, and which can be used and understood thanks to shared knowledge22. A “system of associated commonplaces” (cf. Black 1962: 40) is activated according to a given cultural system, i.e. a common global context. With regard to our main issue (i.e. metaphorical truth conditions), in the cases quoted above and in similar ones, the David, Goliath, and Molière who are referred to are not the “actual” personae of the external, historical world, but have become the most adequate shorthands of definite descriptions in order to refer to a person who has given characteristics: “[…] le référent original était bien porteur du Npr [= Nom propre] métaphorique, tout en étant devenu ensuite leur membre par excellence, le prototype d’une catégorie désignée par ce Npr” (Jonasson 1991: 76) [‘while the metaphorical Npr [proper Name] coincided with the original referent, further on it became the prototype of a category designated by this Npr.’]. What has to be assessed is not the referential truth of the predication (which is patently false), but the satisfactoriness23 of the specific lexical/ cultural choice in personifying the configuration of features we want to evoke, by exploiting the shared encyclopedic knowledge and extending it (cf. Bühler 1934).
————––
22
23
As Jonasson (1991: 70) points out: “[…] à presque tous les Npr [= Nombres propres] métaphoriques utilisés est associé un référent, réel ou fictif, bien connu dans la communauté linguistique”[‘a real or fictitious referent, well known by a linguistic community, is associated with almost all the Npr [proper names] metaphorically used’]. Austin stated some time ago (1962: 149): “[…] truth and falsity are […] not names for relations, qualities, or what not, but for a dimension of assessment – how the words stand in respect of satisfactoriness to the facts, events, situations, &c., to which they refer.”
144
Carla Bazzanella / Lucia Morra
2.3 Metaphor and local context A specific local context is created by the metaphor itself, which selects the relevant features and affects the utterance meaning24. Following Glucksberg and Keysar (1993: 411–2), “the properties selected in a metaphor depend on the topic of conversation”, but the topic of conversation is a dynamic concept (cf. Bublitz 1988), which shifts also on the grounds of the new frames which are activated. To quote Leezenberg’s (2001: 287) analysis of (3): (3) Some lawyers are sharks. The perspective serves as a goal for the categorization, and as such, it strongly constrains the features that may be involved in the metaphorical application. Needless to say, the ad hoc categorization may very well clash with [the] existing correlative structure determined by other theories. Thus, the ad hoc categorization of a lawyer as a member of the concept shark under the perspective of behavior towards victims clashes with the ‘scientific’, causal and taxonomic biological categories. But the perspective involved already indicates that this biological theory is not the one employed in the categorization. In other words, perspectives both trigger and constrain the metaphorical interpretation of a term, and allow for an analysis of metaphor that appeals to ‘context selection’ [our emphasis].
We are dealing here with what some have called “double activation”, or, in Leezenberg’s (2001) terms, two thematic dimensions, which both play a role in establishing the meaning of metaphor. 2.4 Possible interaction between global and local context in metaphor The interaction between the global and local context created via metaphor can differ; in brief, three forms of interaction can be posited: – coincidence – partial coincidence – no coincidence. 2.4.1 Coincidence The local context created by a metaphor can coincide with the global context, in the sense that the contextual frame activated by it is not at odds in any way with the assumptions shared and embedded in the global context: the two thematic dimensions postulated by Leezenberg (2001) apply perfectly. In this ————––
24
Following Weinrich (1967: 9), contextual determination tunes and changes the place of a word on a hypothetical semantic scale.
‘Metaphorical’ truth conditions, context, and discours
145
case, the referents share the maximum degree of isomorphism, that is to say, they have all their features in common: the metaphor works as a referential description. Consider the following two statements: (4) Grace is a child. (5) Ann is a child. As Grace is 8 years old, and Ann is 62, but behaves like a child, “child” is not a metaphor in 4, while it is in 5. 2.4.2 Partial coincidence To be a metaphor, an expression must activate a local context which is in some way at odds with the global one which is involved in the interaction. How this requirement is fulfilled is a matter of degree. The specific contextual frame activated by a given metaphor can easily match the assumptions shared and embedded in the global context. This is revealed when Leezenberg’s (2001) two thematic dimensions collide just a little bit; i.e. the referents, while ontologically different, share a high degree of isomorphism, in other words they have several features in common. In some way, the high degree of coincidence facilitates the “metaphorical transfer”, via a straight association (at the same time, the several possible bridges between the two referents can also render the metaphor ambiguous). To quote an example, in cases such as (6), which refers to hair lotion, the syntagmatic context (or cotext) is successfully exploited in order to achieve the advertiser’s goal of impressing the reader, by referring to the similarities between problems/trees/hair. (6) Affronta il problema dalla radice. Face the problem from the root. In some cases, local and global context coincide only partially: here the specific contextual frame activated by the metaphor parallels the global one only with regard to very few, particular features of the referents which the metaphor correlates. The hypothesis of a “partial metaphorical cover” has been proposed by Bazzanella (1999a: 156) to account for the fact that only one or a few features of the source domain occur in the target domain: a metaphor does not activate all the features which are relevant in the source domain (or in the prototype which is referred to by using the metaphor), but only some of them, with an “adventitious choice of material” (Glucksberg et al. 1997). In some cases of partial coincidence, the success of a metaphor is guaranteed only if the intended feature is activated and no other features are triggered by it.
146
Carla Bazzanella / Lucia Morra
Consider the expression “buon padre di famiglia” (bonus paterfamilias), which in Italian legal language denotes the standard of “diligence” required in obligations, such as in the following fragment taken from the Civil Code: (7) Nell’adempire l’obbligazione il debitore deve usare la diligenza del buon padre di famiglia (c.c. 1176). [‘Fulfilling an obligation, every citizen must apply the diligence used by the bonus paterfamilias’]. Being a standard of behavior, the metaphor applies not only to fathers, but also to women, singles, etc.: its meaning is not (or should not be; cf. Calabresi [1985]) connected with the male or the fartherly properties of the image chosen to express it. If this were so, the metaphor would be successful only when applied to certain members of the society. Nor would it be successful, especially in the modern business world, if the properties were meant to be any of the other stereotypical features commonly associated with the character of the good father, i.e. prudence, conservatism, defense of traditional values, and so forth. The metaphor can only be successful if it is understood as referring to the behavior of the Roman bonus paterfamilias, who behaved like a citizen of the Roman society, and for whom the link between the kind of diligence needed by the administration of the society and the one required by the care due to his household was strong: a kind of man who has very little in common with the modern father of a family. This does not mean that nowadays the metaphor does not work any more: far from it, it is used exactly because it provides judges with the “ideal” of an expected behavior determined by the legal knowledge25 or encyclopaedia26 they share. For the very reason that in a legal context it activates just one feature, referring to a particular kind of behavior, the metaphor of the buon padre di famiglia is used with success as a standard for judging the behavior of every citizen, and as a standard for business life.
————––
25
26
Actually, the meaning of the standard, while substantially homogeneous, differs slightly in every judicial interpretation or legal textbook (cf. Hibbits [1994], Bazzanella & Morra [2004]), Morra et al. (2006). The ‘concretion’ of features which has found an adequate representation in ‘buon padre di famiglia’ remains adequate until the features are the ones requested by specific purposes, when the “generic focus” (see 2.1) changes and the lexeme is replaced, as in other codes (cf. Bazzanella & Morra [2004]). Lexical change in metaphorical expressions is affected both by global context (e.g. historical, cultural background) and by local context (e.g. the purpose of its use).
‘Metaphorical’ truth conditions, context, and discours
147
2.4.3 No coincidence: When a metaphor turns out to be false As 2.4.2 has shown, metaphor can be successful if global and local context diverge, even if they diverge to a notable extent. This feature reflects the deep difference between metaphors and referential expressions, that simply cannot be true if they depict a world different from the current one. The case still has to be considered in which metaphor, while functioning as a proper rhetorical trope, is false. In other words, while Burkhardt (1990: 310) says: “One most important feature of metaphorical utterances is often overlooked: they cannot be true or false at all, and therefore do not express any truth conditions”, we think that metaphor can be false when not even one feature is shared by the correlated referents. This occurs when local and global context do not coincide in any way, or when a speaker uses a metaphor which cannot possibly evoke the state of affairs in question. This may be the result of a clash between the individual set of beliefs and the beliefs shared by the other members of the community (on a global level, such as in (8), a fictitious example, where A is talking with B, a color-blind subject, about a woman wearing a green dress): (8) A: What a beautiful dress! B: I don’t know … It makes her look like a lobster… A: A lobster?! It’s green!!! In this case we could say that what the metaphor predicates (the colour red conveyed by the metaphor lobster is not the colour of the dress commented upon) is false; in other words, the correspondence at issue is not between what has been labeled, e.g. the tenor and vehicle of metaphor, but between what is asserted via metaphor and the state of affairs. Furthermore, during translation (a very revealing issue also with regard to metaphor, cf. e.g. Wikberg 2004), significant problems can arise with regard to what a metaphorical predication really corresponds to, as is the case with the Quranic verse 3:106 reported sub (9) in the English literal equivalence: (9) [‘Some faces will be white and some faces will be black’] As Zahri (1999: 141) points out: Taking the communicative and cultural implications of the color terms ‘white and black’ in the Quranic verse into consideration, one should not translate them into English color terms, ‘white and black’ respectively, as many translators did, since the communicative function of these two terms in such a collocation (with faces) does not coincide with their function in the target language. ‘Black face’ in English does not designate the original connotation of disgrace, but expresses anger or fury; and the same applies to ‘white faces’ which does not express glorification or honouring in English as it does in its Quranic context i.e. the contextual meaning
148
Carla Bazzanella / Lucia Morra of ‘black face’ in Arabic is different from the contextual meaning of ‘black face’ in English, and therefore can not be a translation equivalent.
Given its contextual sensitivity, a metaphor (like any utterance) cannot be right in all circumstances, for any purpose, for any audience, etc27. 3. Metaphor in discourse 3.1 Interactional and textual levels. How metaphor develops in discourse Let us now turn to the successful conditions of metaphor in discourse. As we pointed out above, context is a dynamic process, and global and local levels obviously interact, both in verbal and non-verbal performances28. A metaphor produced in discourse establishes both textual and inter-textual connections, by using already introduced metaphors, and creating new ones, in a continuous interplay of both old/new information and different domains (or “thematic dimensions”). In the next section, we will demonstrate that both textual and intertextual connections are critical for a metaphor to be successful, together with the “interactional component”, that is, the uptake of the utterance and the recognition of a speaker’s “intended meaning”. 3.2 Conditions of success/failure We would like to propose three conditions for a metaphor to be successful in discourse; in brief: – Consistency – Uptake – Recognition. Paralleling the above-mentioned distinction between textual and interactional level, the first condition of success proposed here applies to the textual level, the second and third to the interactional level.
————––
27 28
“We aim at the ideal of what would be right to say in all circumstances, for any purpose, to any audience, &c.” (Austin 1962: 145) For an analysis of non-verbal metaphors, see e.g. Indurkhya (1992).
‘Metaphorical’ truth conditions, context, and discours
149
3.2.1 Consistency Metaphor, as any linguistic and cognitive device, develops in discourse (cf. Bazzanella 1999a, 2008), and textual, intratextual, and intertextual connections should be respected in order to achieve the requested adequateness. Leaving aside the global context (i.e. the set of given features which in a sense constrain, but also support, the use of metaphors, by providing shared knowledge and beliefs; see above), consistency of metaphorical uses should be ruled out in the development of discourse on the grounds of the local activated context (cf. Akman & Bazzanella 2003: 324). 3.2.2 Uptake and recognition The terms uptake and recognition have been proposed, respectively, by Austin (1962) and Grice (1989), and we are applying them here to the analysis of metaphor in order to better understand its conditions of success. Uptake. The securing of uptake is an initial, necessary step to enable a metaphor (just like any other linguistic utterance) to be happily/successfully performed. In Austin’s (1962: 116) words the effect to be achieved on the audience “amounts to bringing about the understanding of the meaning and of the force of the locution”. Recognition. Recognition of a speaker’s intended meaning, which might include the recognition of an implicature (as is the case for metaphor in the Gricean account), is another factor that makes communication successful. 3.2.3 When metaphors turn out to be “unhappy” Metaphors are not always successful; as Ortony (1980: 364) states: “[…] metaphors are like jokes; good ones can be very successful, but bad ones can be disastrous.” – If uptake and recognition are not achieved, metaphors turn out to be unhappy, and we are facing a failure (not necessarily a breakdown) in communication, as we shall see below. – If consistency is not respected, the follow-ups of the inconsistency will show mainly on a linguistic level, by creating an infelicity which can be solved metalinguistically, by asking, e.g. “But why did you say that before …?” or can also be disregarded, especially in spoken discourse. In written text, especially in literature, the attention of the reader, and her/his opportunity to reread the text can be more easily affected by inconsistency. Kruger (1993: 25) exemplifies this problem by showing the interdependence of intratextual components in the use and translation of metaphor in Fiela’s child, i.e. an Afrikaans novel. More specifically: “[…] characterisation is determined by the global interaction of metaphorically qualified arguments
150
Carla Bazzanella / Lucia Morra
reflecting the semantic domains of farms, forest and sea”, but in the translation this interaction is not maintained. To quote an example produced by Kruger: the Afrikaans expression reported sub (10): (10) Hulle het jou klaar hokgejaag, Fiela Komoetie… corresponds to the English: ‘They have already chased you into a cage’. The English translator (Matthee [1986], quoted in Kruger [1993]), while choosing ‘cornered’ instead of ‘chased into a cage’, destroys the necessary interrelations between character, discourse and environment29: “Although ‘cornered’ conveys Fiela’s sense of helplessness and invokes connotations of ‘no escape’, ‘trapped’ and ‘being threatened’, the concretisation of her feelings, her dehumanisation and the disregard of her human dignity have been lost in the translation.” (Kruger 1993: 27). Furthermore, the intratextual “net” of metaphorical cross-references, which in the original text were meant to describe and construe the character in a specific way, is not respected: ‘Hulle was haar aan die vaskeer’ (Matthee 1985, 48) was translated as ‘They were driving her into a corner’ (Matthee 1986, 63), instead of something like: ‘they were ensnaring her’. This latter translation would have sustained the source text image of Fiela being an animal trapped in a cage – an image which is of a global significance in the novel. Similarly, ‘hulle pak nog net die opening toe’ (Matthee 1985, 48) was translated as ‘What they are doing now in closing in on you’ (Matthee 1986, 64), instead of say, ‘What they are doing now is fencing you in’ which, again, would have reinforced the global specification of Fiela as one who has been dehumanised. (Kruger 1993: 28)
3.3 Metaphor and dialogic interaction Let us now briefly analyze the kind of verbal interaction which, triggered by metaphor, takes place between interlocutors, and how the various possibilities of this interaction can affect the metaphor’s value in discourse. We have seen that the uptake of the utterance and the recognition of the speaker’s “intended meaning” are necessary preconditions in order for the metaphor to be “happy”, but in the course of the dialogic interaction, this use can trigger subsequent modalities of interaction which are different. Three modalities of interaction are possible in principle: 1 – Acceptance 2 – Refusal 3 – Coming to an understanding. ————––
29
“Most of the metaphorical expressions in Fiela’s speech reflect her rural environment.” (Kruger 1993: 27).
‘Metaphorical’ truth conditions, context, and discours
151
3.3.1 Accepting and appropriating metaphor Metaphor can be accepted by other interlocutors either implicitly (by going on with the discourse in the same direction as the preceding speaker who used the metaphor, which thus results in being taken up and recognized) or explicitly, by means of appropriation, i.e. by repeating and applying it subsequently (cf. Bazzanella 1999a). In a successful, happy verbal interaction, the understanding of a metaphor (i.e. uptake and recognition) is evident when the metaphor is further exploited or enhanced. To quote an example, in (11), taken from a discussion during a science lesson in a primary school (cf. Bazzanella 1999a), first Elisa repeats her own metaphor to be friends with and refines it (you could still distinguish it), then another child (Edoardo) uses it, and finally the teacher, after some turns, summarises the discussion, and in a sense “ratifies” the metaphor30: (11) Insegnante. Se dovessimo provare a fare delle famiglie di comportamento chi mettereste insieme? Chi potremmo mettere con lo zucchero? Elisa. Possiamo mettere insieme le cose che sono amiche con l’acqua, che riescono a entrare nell’acqua, come lo zucchero e il sale. Francesco. Forse il cacao solubile. Elisa. Il cacao non era tanto amico, perché si riusciva ancora a distinguere. Edoardo. L’alcool era amico, perché poi non lo distinguevamo più. [...] Insegnante. Allora possiamo mettere anche i colori in questa famiglia. Ma attenzione: tutti i colori sono amici con l’acqua? [Teacher. If we wanted to try and make some families based on behavior, what could we put together? Who could we put together with sugar? Elisa. We can put together things that are friends with water, that can go into water, like sugar and salt. Francesco. Maybe soluble cocoa. Elisa. Cocoa wasn’t such a good friend with water, because you could still distinguish it. Edoardo. Alcohol was friends with it, because we couldn’t distinguish it any more. [...] Teacher. So we can put colors in this family, too. But mind: are all colors friends with water? ————––
30
For the complex process of negotiation and joint construction of meaning which is going on in the quoted discussion, also by ‘exploiting’ metaphors, cf. Bazzanella (1999a).
152
Carla Bazzanella / Lucia Morra
3.3.2 Refusing metaphor Instead of showing uptake of a given metaphor, the interlocutor may refuse it, by negating the propositional content, i.e. the transfer of features, such as in (12), extracted from a classroom discussion on similarities and differences between vegetables and animals in a primary school (cf. Bazzanella 1999a): (12) Caterina. [...] un po’ è come un animale la pianta perché i rami... sono le braccia Marco. Le piante non parlano. [Caterina. [...] a tree is a bit like an animal because its branches... are the arms.] [Marco. Trees don’t speak.] In this case, the metaphor turns out to be unhappy, and the same thing happens when the metaphor is misunderstood. 3.3.3 Understanding/misunderstanding metaphor Understanding (coming to understanding in Weigand’s [1999] words) is a complex phenomenon which involves many processes, tasks, and devices, and which works (or does not work, or partially works) on several layers, and to different degrees (cf. Vendler 1994). Though we accept what is usually referred to as the “intuitive strength” of metaphor, metaphors are not always “correctly” understood, in the sense that speaker’s intentions are not recognized, as in (11), taken from Bazzanella 1999b: (13) A: Coraggio. Andiamo nella sala bomboniera. B: Bomboniera? È un po’ grande per essere una bomboniera. A: No. Mi riferivo al fatto che sia chiusa e che manchi l’aria. A: Well. Let’s go into the bonbonnière room. B: Bonbonnière? It’s rather big for a bonbonnière. A: No. I was referring to the fact that it’s closed and that there’s no air in it.] Speaker A, by using the metaphor bonbonnière, wants to refer to the enclosed feeling of the very large room where a congress is taking place, while speaker B does not recognize these properties as relevant31 to the chosen metaphor, ————––
31
In our opinion, relevance is not always ‘spontaneously perceived’ by the interlocutor, as D’Hulst (1992: 42) seems to envisage: “Le processus de cognition thèorique se plie […] à des contraintes pragmatiques qui incitent l’émetteur à choisir des metaphores spontanément perçues comme pertinentes par le récepteur.” [‘The
‘Metaphorical’ truth conditions, context, and discours
153
which is, in her view, adequate only for something small. By making her doubt explicit, she triggers A’s intended meaning (I was referring to the fact that it’s closed and that there’s no air in it). In the case of idioms or “dead metaphors” the relevant features have been selected and “agreed upon” by the community (see § 3), but in creative metaphors, where the relationship is not taken for granted (thus creating possible misunderstandings or difficulties in comprehension, as in (11), where interlocutors referred to a different prototype of bonbonnière), the intended meaning has to be recognized on the basis of contextual data32, and sometimes negotiated. The complex process of coming to an understanding in dialogic interaction develops in several steps, misunderstanding included (cf. Weigand 1999, Dascal 1999); the latter demands the need for an interactional cycle of negotiation, and eventually leads to acceptance or refusal/agreement or disagreement (cf. Bazzanella & Damiano 1999). In fact, as Kurth (1997: 6) notes, the linguistic associations which are conveyed by metaphor entail also evaluation/judgement: “A metaphor’s meaning is not restricted to linguistic associations. It also implies a certain way of perceiving the object, often including evaluation or judgement.” 4. Conclusion We are now in a position to resume our discussion on “metaphorical” truth conditions: instead of considering metaphor and non-literal uses of language as deviant phenomena, we are inclined to stress the “vagueness” of “literal language” (cf. Recanati 1995), which itself requires contextualization for speakers’ meaning/intentions to be successfully recognized. We therefore completely agree with Leezenberg’s (2001: 304) conclusion: The notion of literal meaning, then, expresses an ideal of academic discourse rather than the reality of everyday communication. Word meanings are contextdependent, imprecise, and variable; there is no theoretical notion that can usefully serve as a counterpart of our folk concept of literal meaning. […] Literal meaning, in other words, is a myth: as convenient as it may be, it is, in the final analysis, a fiction.
If this is so, it would follow that the distinction between literal and non-literal meaning should be abolished. ––––––––––
32
process of theoric cognition undergoes […] some pragmatics constraints which stimulate the speaker producing metaphors that are spontaneously perceived by the addressee as relevant.’] Following Weinrich (1967: 6), “The smaller the meaning extension and the larger therefore the intension (as in the case of proper names) are, the stronger the expected contextual determination is. Correspondingly, the larger the extension and the smaller the intension (as with abstract nouns) are, the weaker the expected contextual determination is.”
154
Carla Bazzanella / Lucia Morra
As Norrick (1999: 191–2) convincingly argues, the same strategies that we use to solve contradictions, paradoxes, incongruities, etc. in conversation are adopted in understanding metaphors as well: “[…] contextual analysis of figurative utterances alongside outright incongruities and clarifications can provide evidence for real psychological strategies of interpretation. Such interpretive strategies independently required at the discourse level could eliminate the need for any narrowly semantic rules for figures.” If, following Lakoff and Johnson (1980: 3), “[…] our ordinary conceptual system is […] metaphorical in nature”, metaphor could be said to exploit the tight link between experience and language inside a contextual frame which can be given (as in “dead” metaphors) or new (as in creative ones). What is at stake then, as said before, is appropriateness, and not truth conditions, which appear to be insufficient for describing the complex role of metaphor in language. Metaphor (and literal language as well) is to be taken as mainly appropriate in discourse, in its multifarious contextual constraints, as it is “[…] not an awkward way of communicating one’s convictions and beliefs (about the properties of objects) to another. […It] is, on the contrary, a very subtle means of transcending the realm of trivial everyday truth and falsehood.” (Burkhardt 1990: 320, 324) References Akman, Varol, Carla Bazzanella: “The complexity of context”, in: Akman, Varol, Carla Bazzanella (eds.): Special issue on Context, Journal of Pragmatics, 2003. Austin, John Langshaw: Philosophical Papers. London: Clarendon Press 1961. – How to Do Things with Words. Oxford: Oxford University Press 1962. Bazzanella, Carla: “On context and dialogue”, in: Cmejrkovà, Svetla et al. (eds.): Dialogue in the Heart of Europe. Tübingen: Niemeyer 1998, 407–16. – “Metaphor in classroom interaction”, in: Rigotti, Eddo (ed.): Rhetoric and Argumentation. Tübingen: Niemeyer, 1999a. 237–46. – “La metafora tra mente e discorso”, in: Bazzanella, Carla, Claudia Casadio: “Prospettive sulla metafora.” Lingua e stile XXXIV, 2 (1999b), 150–8. – “Metaphor and context: some issues”, in: Kronning, Hans et al. (eds.): Langage et référence. Mélanges in honour of Kerstin Jonasson. Acta Universitatis Upsaliensis: Upsala, 2001, 39–49. – “The significance of context in comprehension. The ‘we case’”, in: Edmonds, Bruce, Varol Akman (eds.): Special issue of FOS (Foundations of science) Context in context, 7. Dordrecht: Kluwer Academic Publishers 2002, 239–54. – “Il radicamento contestuale della metafora”, in: Casadio, Claudia (ed.): Vie della metafora: linguistica, filosofia, psicologia. Sulmona: Editore Prime Vie, 2008, 79– 95. – Special issue of Paradigmi: La forza cognitiva della metafora. Milano: Angeli 2009. – , Claudia Casadio (eds.): “Prospettive sulla metafora”, in: Lingua e stile XXXIV, 2 (1999), 149–226. – “The interactional handling of misunderstanding in everyday conversations”, Journal of Pragmatics, 31-6 (1999), 817–36.
‘Metaphorical’ truth conditions, context, and discours –
155
, Lucia Morra: “Metafora e linguaggio giuridico: alcune riflessioni”, in: Bernini, Giuliano, Ferrari, Giacomo, Pavesi, Maria (eds.): Atti del 3° congresso di studi dell’AITLA. Perugia: Guerra 2004, 191–212. – , Lucia Morra, Rossi Piercarlo: “L’approccio cognitivo alla metafora nel linguaggio giuridico”, in: Caterina, Raffaele (ed.): Fondamenti cognitivi del diritto. Milano: Pearson Paravia Bruno Mondadori 2008: 157–75. Black, Max: Models and Metaphors. Ithaca, N.Y.: Cornell University Press 1962. Bublitz, Wolfram: Supportive Fellow-Speakers and Cooperative Conversations. Amsterdam: Benjamins 1988. Bühler, Karl: Sprachteorie. Jena: Fischer 1934. Burkhardt, Armin: “Searle on metaphor”, in: Burkhardt, Armin (ed.): Speech Acts, Meaning, and Intentions. Critical Approaches to the Philosophy of John R. Searle. Berlin–New York: de Gruyter 1990, 303–35. Cacciari, Cristina: “La metafora: un ponte fra il linguaggio e l’esperienza percettiva”, in: Bazzanella, Carla, Claudia Casadio (ed.): “Prospettive sulla metafora.” Lingua e stile XXXIV, 2 (1999), 159–66. Calabresi, Guido: Ideals, Beliefs, Attitudes, and the Law. Private Law perspectives on a Public Law Problem. Syracuse, N.Y : Syracuse University Press 1985. Dascal, Marcelo: “Introduction: some questions about misunderstanding”, in: Journal of Pragmatics, 31-6 (1999), 753–62. D’Hulst, Lieven: “Sur le rôle depps métaphores en traductologie contemporaine”, in: Target, 4:1 (1992), 33–51. Eco, Umberto: “Dove sta Capuccetto Rosso?”, in: Usberti, G. (ed.): Modi dell’oggettività. Saggi in onore di Andrea Bonomi. Milano: Bompiani 2000, 137– 57. Givòn, Talmy: “Mind, Code and Context. Essays”, in: Pragmatics. Hillsdale, N.J.: Erlbaum 1989. Glucksberg, Sam, Boaz Keysar: “How metaphors work”, in: Ortony, Andrew (ed.): Metaphor and Thought. Cambridge: Cambridge University Press 1993, 2nd ed., 401–24. – , Matthew S. McGlone, Deanna A. Manfredi: “Property attribution in metaphor comprehension”. Journal of memory and language, 36, 1997, 50–67. Grice, H. Paul: Studies in the Ways of Words. Cambridge, Mass.–London: Harvard University Press 1989. Gumpel, Lisolette: “Meaning and metaphor: the world in verbal translation”, in: Müller-Vollmer, Kurt (ed.): Translating Literatures, Translating Cultures. Berlin: Erich Schmidt Verlag 1988, 47–78. Hibbitts, Bernard J.: “Making sense of metaphors: visuality, orality and the reconfiguration of American legal discourse”, in: Cardozo Law Review, 16.2 (1994), 229– 356. Hymes, Dell: Foundations in Sociolinguistics. Philadelphia: University of Pennsylvania Press 1974. Indurkhya, Bipin: Metaphor and Cognition: an Interactionist Approach. Dordrecht: Kluwer 1992. Jonasson, Kerstin: “Les noms propres métaphoriques: construction et interprétation”, in: Langue française, 92 (1991), 64–81. Kittay, Eva Feder: Metaphors. Its Cognitive Force and Linguistic Structure. Oxford: Blackwell 1987. Kruger, Alet: “Translating metaphors in narrative fiction”, in: Perspectives: Studies in Translatology (1993), 23–30. Kurth, Ernst-Norbert: “Altered images: aspects of metaphor translation”, in: International Journal of Translation, vol. 9, n. 1-2 (1997), 5–20.
156
Carla Bazzanella / Lucia Morra
Lakoff, George, Mark Johnson: Metaphors We Live by. Chicago: The University of Chicago Press 1980. Leezenberg, Michiel: Contexts of Metaphor. Amsterdam: Elsevier 2001. Levinson, Stephen C.: Pragmatics. Cambridge: Cambridge University Press 1983. – Presumptive meanings: the theory of generalized conversational implicature. Cambridge, Mass.: MIT Press 2000. Mey, Jacob: Pragmatics. An Introduction. Oxford: Blackwell 1993. Morra, Lucia, Piercarlo Rossi, Carla Bazzanella: “Metaphor in legal language: clarity or obscurity?”, in: Wagner, Anne, Sophie Cacciaguidi (eds.): Clarifying Legal Drafting: Practice and Tools. Berne: Peter Lang 2006, 141–74. Norrick, Neal R.: “Paradox and metaphor: a discourse approach”, in: Bazzanella, Carla, Claudia Casadio (eds.): “Prospettive sulla metafora”, in: Lingua e stile XXXIV, 2 (1999), 191–200. Ortony, Andrew (ed.): Metaphor and Thought. 1979. Cambridge: Cambridge University Press 1993, 2nd ed. – “Some psycholinguistic aspects of metaphor”, in: Honeck, Richard, Robert Hoffman: Cognition and Figurative Language. Hillsdale, N.J.: Erlbaum 1980, 69–83. Radman, Zdravko: From a Metaphorical Point of View: A Multidisciplinary Approach to the Cognitive Content of Metaphor. Berlin–New York: de Gruyter 1995. Recanati, François: “The alleged priority of literal interpretation”, in: Cognitive Science, 19 (1995), 207–32. Richards, Ivor Armstrong: The Philosophy of Rhetoric. Oxford: Oxford University Press 1936. Ryle, Gilbert: The Concept of Mind. London: Hutchinson 1949. Searle, John Rogers: “Literal meaning”, in: Erkenntis, 13 (1979), 207–24. Sperber, Dan, Deirdre Wilson: Relevance. Cambridge, Mass.: Harvard University Press 1986. Stern, Josef : Metaphor in Context. Cambridge, Mass.–London: The MIT Pres 2000. Turner, Mark, Gilles Fauconnier: “Conceptual integration and formal expression”, in: Journal of Metaphor and Symbolic Activity, 10:3 (1995), 183–204. Vendler, Zeno: “Understanding misunderstanding”, in: Jamieson, Dale (ed.): Language, Mind and Art. Dordrecht: Kluwer 1994, 9–21. Weigand, Edda: “Misunderstanding: the standard case”, in: Journal of Pragmatics, 316 (1999), 763–85. – , Marcelo Dascal (eds.): Negotiation and Power in Dialogic Interaction. Amsterdam: Benjamins 2001. Weinrich, Harald: “Semantik der Metapher”, in: Folia linguistica. Acta Societatis Linguisticae Europeae, 1 (1967), 3–17. Wikberg, Kay : “English Metaphors and Their Translation: the Importance of Context”, in: Aijmer, Karin, Anna-Brita Stenström (eds.): Discourse Patterns in Spoken and Written Corpora. Amsterdam: Benjamins 2004, 245–65. Wittgenstein, Ludwig: Philosophische Untersuchungen. Oxford: Blackwell 1953. Zahri, Maysoon: “Metaphor translation”, in: International Journal of translation, vol. XI, 1-2 (1999), 103–44.
Masa-aki Yamanashi (Kyoto, J)
Metaphorical modes of perception and scanning A comparative study of Japanese and English
0. 1. 1.1 1.2 1.3 2. 3. 4. 5. 6. 7.
Introduction Metaphors of perception and cognition Visual field as container Visual perception and touch Perception as reception Duality in visual perception Sense modalities and perceptual metaphors Cognitive constraints on target domain Seeing based on transformation of the visual field Visual metaphors and tropical truth Conclusion
0. Introduction Natural language has some means to provide a more or less effective and natural way to cope with the external world. Metaphor should serve as one of the effective means to achieve that purpose. Much of our cognitive reality is understood in metaphorical terms and our conceptualization of the physical world can be said to be partly metaphorical in nature. Metaphor, in this respect, plays a significant role in determining what is real for us. It is generally the case that metaphorical mappings vary in universality; some of them are highly likely to be universal or at least widespread across languages, while others might turn out to be language specific. Strictly speaking, our conceptual systems change over time. Some researchers hypothesise, however, a range of general conceptual metaphors which would stay the same over time.1 Visual metaphors might constitute a crucial subset of those metaphors, for our perceptual mechanism is one of the most essential means by which to understand the external world. The ways in which we perceive and construe the world are directly or indirectly reflected in the conceptual system of natural language through a variety of visual metaphors. The main objective of this paper is to compare a wide range of visual metaphors involved in Japanese —————
1
For the detailed analyses of conceptual metaphors in English, see, among others, Lakoff and Johnson (1980) and Lakoff (1993a, 1993b). See also Yamanashi (1988) for a comparative study of Japanese and English metaphors.
158
Masa-aki Yamanashi
and English. Metaphors of this sort reflect in a natural way the conceptual systems which underlie both languages. The comparison and examination of such metaphors, in this respect, contribute to the elucidation of the linguistic and cognitive status of visual metaphors, and further the exact status of tropical truth of conceptual metaphors in general. 1. Metaphors of perception and cognition Visual metaphors or metaphors of perception represent some of the most general conceptual metaphors which might characterize the underlying conceptual systems across languages. One of the most typical metaphors which contribute to the creation of conventional visual expressions involves container metaphors. Container metaphors are widespread in natural languages. One such metaphor has to do with visual fields. 1.1 Visual field as container In many languages, visual fields are conceptualized as containers, which means that when we take a look at some object or entity, our field of vision delimits some domain or boundary in which the object or entity in question can be identified. This can be illustrated by the following Japanese examples:2 (1) a. Teki no hikooki ga sikai ni haitte-ki-ta. enemy GN plane NOM view into enter-come-PST ‘The enemy’s plane came into view.’ b. Hune wa suguni sikai kara kie-ta. ship TOP soon sight out-of disappear-PST ‘The ship soon went out of sight.’ c. Kare no herikoputaa wa mada sikai ni haitte-ko-nai. he GN helicopter TOP yet view into enter-come-not ‘His helicopter does not come into view yet.’ ————— 2
See Yamanashi (1995, 1997) for the linguistic status of Japanese visual expressions based on container metaphors of this sort. Metonymically Extended Meanings
me
(i) eyesight (ii) gaze, glance (iii) sight, view (iv) visual field
Table 1
Metaphorical modes of perception and scanning
159
(2) a. Yosoo mo sinakatta kesiki ga me ni haitte-ki-ta. guess even do-not-PST landscape NM eyes into enter-come-PST ‘An unexpected landscape came into view.’ b. Kodomo ga aruite iru no ga me ni haira nakat-ta. child NM walking be CM NM eyes into enter not-PST ‘I failed to notice a child walking.’ c. Kado o magaru to takai tatemono ga me ni hait-ta. corner AC turn when tall building NM eyes into enter-PST ‘Turning the corner, I saw a tall building.’ As is clear from these examples, Japanese has two crucial terms relating to visual perception: sikai ‘visual field’ and me ‘eye(s)’. The examples with the former term (i.e. those in 1 above) basically correspond to the following visual expressions in English (cf. Lakoff & Johnson 1980: 30). (3) a. The ship is coming into view. b. I have him in sight. c. I can’t get all of the ships in sight at once. d. There’s nothing in sight. e. He’s out of sight now. f. That’s in the center of my field of vision. The visual field in these examples is conceptualized as a container through such words as view, sight, field of vision. The Japanese term me in (2) above, however, does not directly represent a visual container. The word me designates a body-part which serves as one of our crucial organs of perception. This term metonymically represents a visual field, which in turn is conceptualized as a container.3 It is interesting to note that in Japanese the term me cognitively serves as a reference-point which indirectly designates a visual field. This can be illustrated in terms of the following reference-point construction (see Langacker [1993] for the basic cognitive status of this construction.) In Figure 1, the circle (labeled C) represents the conceptualizer, R is the reference point, and T the target (that is, the entity that the conceptualizer uses R to establish mental contact with). The dashed arrows represent the mental path the conceptualizer follows in getting to the target. And, the ellipse (labelled D) indicates a domain which corresponds to the set of potential targets to which a particular reference point affords direct access.
————— 3
The Japanese body-part term me has at least the following metonymically extended meanings (cf. Yamanashi 1995: 222, 1997: 839).
160
Masa-aki Yamanashi
<me> R
T
D
C
Figure 1
Consider in this connection the following English examples: (4) a. *The ship is coming into {my/our} eye(s). b. *I have him in {my/our} eye(s). c. *He’s out of {my/our} eye(s) now. d. *That’s in the center of {my/our} eye(s). These examples have a body-part term eye(s), which can logically serve as a reference-point which indirectly designates a visual field as its target. In contrast to Japanese instances in (2), however, the corresponding English body term eye(s) cannot serve as a reference-point to trigger a visual field as its target. This can be confirmed by the fact that those examples in (4) are not acceptable visual expressions in English.4 1.2 Visual perception and touch The reification of a visual field as a container is one of the general and natural means by which our visual or perceptual experience can be linguistically encoded. Another way to express such experiences is a conceptual metaphor which is grounded in our tactile experience. Lakoff pointed out that “there is a general conceptual metaphor that SEEING IS TOUCHING, in which the eyes are limb-like projections that can reach out and touch things.” (Lakoff 1993b: 230) This metaphor involves a set —————
4
For a more detailed discussion of visual expressions of this sort, see Yamanashi (1997). There exist, however, some visual expresssions in which the body term eye(s) is used. Brigitte Nerlich pointed out such metaphorical expressions as In my eyes, this is wrong. She also noted that there is a show “Stars in their eyes” in which ordinary people adopt the role of a celebrity and sing their song, etc. Strictly speaking, however, metaphorical expressions of this sort are qualitatively different from those in (4) in that the body term eye(s) in the former case is idiomatically used in the more abstract sense.
Metaphorical modes of perception and scanning
161
of systematic conceptual mappings based on the following source and target domains: (i) (ii) (iii) (iv) (v)
Source: Limbs can be directed Target: Vision can be directed, Source: A limb can go in only one direction at a time Target: Vision can go in only one direction at a time, Source: Limbs can extend from the body to other objects Target: Vision can move from the body to other objects, Source: Tactile perception occurs when a limb touches an object Target: Visual perception occurs when the eye-gaze touches an object, Source: Limbs can pick out objects Target: Vision can pick out objects. (Lakoff 1993b: 231)
The existence of such conceptual metaphors can be confirmed by the following examples: (5) a. My eyes picked out every detail of the pattern. b. Their eyes met. c. His eyes are glued to the TV. d. I can’t take my eyes off of her. e. She was undressing him with her eyes. f. They made eye contact. (Ibid.: 230) In Yamanashi (1997: 840) the following examples were examined to lend support to the existence of conceptual metaphors based on tactile experience: (6) a. Gabriel’s eyes, ..., wandered to the wall above the piano. (James Joyce, The Dead: 128) b. His eyes moved to the chair over which she had thrown some of her clothes. (Ibid.: 157) (7) a. His eyes went over the mass of swarming white bodies climbing up the screen toward him. (John Steinbeck, The Snake: 418) b. She had moved over in front of the new cage; her black eyes were on the stony head of the snake again. (Ibid.: 418) c. Her eyes were on the snake where it lay still. (Ibid.: 418) (8) Automatically he reached in his pocket for the big black knife, but it was not there. His eyes searched the ground. (John Steinbeck, Flight: 473) (9) During the circus performance I kept my eye more on Hautboy than on the celebrated clown. (Herman Melville, The Fiddler: 29) In (6)a–b, (7)a and (8) the body-part term eyes cooccurs with motion verbs such as wander, move, go over, etc. This kind of subject cannot be literally
162
Masa-aki Yamanashi
taken to move to the designated target in question.5 What moves to the target here is a limb-like visual projection that can reach out and touch the target. Examples of this sort again show that seeing or visual perception is metaphorically taken to be a kind of tactile experience. This can be further supported by the examples in (7b–c) and (9), where the body-part term eye(s) cooccurs with a preposition on, which basically designates a tactile relation.6 (10) a. *Taro no me wa kabe no syasin ni ugoite-it-ta. Taro GN eye TP wall GN picture to move-go-PST ‘Taro’s eyes moved to the picture on the wall.’ b. *Kanozyo no me wa hodoo no rakugaki ni tassi-ta. she GN eye TP sidewalk GN graffiti to reach-PST ‘Her eyes reached the graffiti on the sidewalk.’ c. *Otoko no me wa basu o matte-iru onna ni idoo-si-ta. man GN eye TP bus ACC wait-be woman to shift-do-PST ‘The man’s eyes shifted to the woman waiting for a bus.’ (11) a. Taro wa kabe no syasin ni me o tome-ta. Taro TP wall GN picture on eye ACC lay-PST ‘Taro laid his eyes on the picture on the wall.’ b. Kanozyo wa hodoo no rakugaki ni me o otosi-ta. ————— 5
Consider in this connection the following examples pointed out by Gruber (1967): (i) John sees across the room. (Ibid.: 941) (ii) a. Bill thought he could see into the room, b. It is easy to see through this glass. c. The baby bird saw over the rim of the nest. (Ibid.: 937) Here, visual perception is figuratively reified as a kind of motion. The prepositions used in these examples constitute a subset of those prepositions which can cooccur with motion verbs in general. It is also interesting to note in the following example the coocurrence of the preposition beyond with a perception verb see: “I see before me a window; beyond that some trees ...; beyond that the Atlantic Ocean; beyond that is Europe; beyond that is Asia. ...” Idiomatic Phrases cast an eye over keep an eye on lay eyes on run an eye over set eyes on turn one's eyes on take one's eyes off
Table 2 6
(Boulding 1956: 3) The following visual expressions are also available in English. The use of such prepositions as on, off, etc. in these instances serves as further evidence for the tactile nature of visual metaphors.
Metaphorical modes of perception and scanning
163
she TP sidewalk GN graffiti on eye ACC drop-PST ‘She dropped her eyes on the graffiti on the sidewalk.’ c. Otoko wa basu o matte-iru onna ni me o tuke-ta. man TP bus ACC wait-be woman to eye ACC attach-PST ‘The man directed his attention to the woman waiting for a bus.’ It must be noted in passing that the corresponding Japanese body-part term me cannot be used as a subject to cooccur with verbs of motion (e.g. ugoku ‘move’, iku ‘go’, ugoi-te-iku ‘go-move’, etc.), as is clear from (10). Sentences of this sort are not conventionally acceptable visual expressions in Japanese. Visual expressions in (11), on the other hand, are conventionalized expressions, where the term me serves as an object of such verbs as involving the notion of touch or attachment (e.g. tomeru ‘lay ~ on’, otosu ‘drop ~ on’, tukeru ‘attach ~ to’, etc.).7 1.3 Perception as reception Another way to express our visual experiences is a tactile metaphor in which perception is construed in terms of visual sense impressions which reach the eyes. In this metaphor, perception is metaphorically conceptualized as a kind of reception. The following Japanese examples illustrate metaphors of this type: (12) a. Kono mati de wa takusan-no huroosya ga me ni tuku. this town in TP a lot-GN vagrants NM eye to attach ‘There are lots of vagrants in this town.’ b. Resutoran no kanban ga me ni tomat-ta. restaurant GN sign NM eye on settle-PST ‘I happened to see a sign of the restaurant.’ c. Hito no me ni hure-nai tokoro ni kore o kakusi-nasai ! people GN eye on touch-not place in this ACC hide-do ‘Hide this so that people can not find it !’ In each of these examples, visual perception is construed in terms of visual sense impressions directly reaching the eyes (i.e. me), which can be confirmed by such expressions as tuku ‘attach to’, tomaru ‘settle on’, hureru ‘touch’. Note that the direct English translations of (12) are not acceptable (e.g. *Vagrants attached to our eyes in this town, *A sign of the restaurant settled on my eyes, etc.). This does not mean, however, that there is a lack of tactile metaphors based on perception as reception in English, as can be shown by —————
7
The fact that tactile experience plays a crucial role in visual expressions can be further supported by the following similes in Japanese: e.g. nameru yoo na sisen ‘gaze like licking’, nazoru yoo na sisen ‘gaze like touching’ (cf. Yamanashi 1995: 223).
164
Masa-aki Yamanashi
the following examples which are based on the metaphor of perception as reception (Lakoff 1993b: 232): (13) a. The view blew me away. b. The view knocks me over. c. The mountain peeked at me through the fog. (13)a and (13)b mean that the sense impressions figuratively strike the viewer with visual force. (13)c contains a personification, in which the sight of the mountain is metaphorically personified as coming through the fog to the viewer. The fact that the viewer in these examples is a perceptual receiver can be further supported by the following examples (e.g. *From my office the view blew me away, *From my office the view knocks me over, *From my office the mountain peeked at me through the fog (ibid: 232). The from-phrase in each example apparently implies that the direct object (i.e. me) is to be interpreted as having a source role. The unacceptability of these examples, however, shows that this is not the case. Facts of this sort can be explained in a natural way if the viewer in question is metaphorically conceptualized as a receiver. 2. Duality in visual perception Visual metaphors observed so far represent a special case of so-called “duality” phenomena (cf. Lakoff 1993a). One of the most typical instances for such phenomena is a case of time metaphor. Take, for example, the following pair in Japanese: (i) Tosi no se ga dandan tikazuite kita (‘The end of the year is gradually coming up on us’), (ii) Dandan tosi no se ni tikaduite iku (‘We are gradually coming up on the end of the year’). The pair here should constitute one of the widespread cases of duality where the passage of time is metaphorically conceptualized as the relative motion between time and its observer. In the former instance, i.e. (i), an entity reified as future time is moving toward its observer. By contrast, in the latter, i.e. (ii), the observer is moving toward future time which is reified as a location. Basically the same kind of duality can be observed in the following visual metaphors in Japanese. (Some of them have already been examined in the previous sections.) (14) (i) a. Doositemo soko no ookina keeki ni me ga iku. necessarily there GN big cake to eye NM go ‘I can’t help getting attracted to the big cake.’ b. Doositemo soko no ookina keeki ga me ni tuku. necessarily there GN big cake NM eye to attach ‘I can’t turn my eyes away from the big cake.’ (ii) a. Kanozyo wa kabe no syasin ni me o muke-ta. she TP wall GN picture to eye ACC turn-PST ‘She turned her eyes to the picture on the wall.’
Metaphorical modes of perception and scanning
165
b. Kabe no syasin ga kanozyo no me ni hure-ta. wall GN picture NM she GN eye on touch-PST ‘She had a glimpse of the picture on the wall.’ (iii) a. Kare wa ookina sugi no ki ni me o tenzi-ta. he TP tall cedar GN tree to eye ACC shift-PST ‘He turned his eyes to a tall cedar.’ b. Ookina sugi no ki ga kare no me ni tomat-ta. tall cedar GN tree NM he GN eye on settle-PST ‘He caught a glimpse of a tall cedar.’ The first sentence of each pair in (14) is based on a visual metaphor in which the perceiver’s eye (i.e. the perceiver’s gaze, to be more exact) moves toward the target of perception and finally touches it (cf. (i) in Figure 2).8
(i) <eyes> (ii) <eyes>
(gaze)
(sense stimuli)
Figure 2
By contrast, the corresponding second sentence in (14) is based on a visual metaphor where the target of perception (i.e. the sense stimuli thereof) comes to touch the perceiver (or the perceiver’s eyes) (cf. (ii) in Figure 2).9 In this respect, the visual metaphors in (14) can be analyzed as a special case of duality phenomena. Metaphorical mappings involved in such phenomena represent a typical cognitive process called ‘figure-ground’ reversal in that (i) the mappings in question are based on relative motion between two —————
8
9
Ano otoko kara me o hana su-na (‘Don’t take your eyes off from that man!’) is another crucial example, which confirms the existence of tactile metaphors of this sort. The visual perception shown in the first version (i.e. (i)) in Figure 2 can be taken to be a kind of scanning. This kind of scanning is different from the scanning involved in such linguistic expressions as the following (Langacker 1990: 17–19). (i) Vanessa is sitting across the table from Veronica. (ii) The new highway {goes/runs/climbs} from the valley floor to the senator’s mountain lodge. In the latter case, the perceiver (or the conceptualizer’s perceptual organ) is not linguistically encoded, whereas in the former case the perceptual organ (e.g. me ‘eye’) is linguistically encoded. Put differently, their basic difference lies in whether the conceptualizer is linguistically foregrounded or backgrounded. For a general discussion of the scanning involving a backgrounded conceptualizer, see Langaker (1987, 1990), Talmy (1988) and Yamanashi (1995).
166
Masa-aki Yamanashi
entities, (ii) one of them is fixed, while the other moves, and (iii) which of the two entities gets fixed or moves depends on the conceptualizer’s cognitive capacity of focusing. It should be noted that Figure 2 illustrates the basic dual relationship of Japanese visual expressions. The first version of Figure 2, i.e. (i), has English counterparts, as was already pointed out in (6-9) (cf. Section 1.2). The second version of Figure 2, i.e. (ii), however, does not have direct English counterpart expressions, as is clear from the examples in (4) (cf. Section 1.1) (e.g. a. *The ship is coming into {my/our} eye(s), b. *I have him in {my/our} eye(s), c. *He’s out of {my/our} eye(s) now).
<eyes>
(sense stimuli)
Figure 3
The English counterparts of such expressions have in the object position such terms as sight, view, which represent a visual field as a container (cf. Figure 3). That this is the case can be supported by such examples as The ship is coming into view, I have him in sight, There’s nothing in sight, He’s out of sight now. In English, these examples contrast with visual expressions such as From our house we can see Mt. Fuji, From my room I can see the top of the university tower, She can’t take her eyes off the delicious cake. The former examples involve a metaphor of vision in which the target of perception moves to the perceiver, while in the latter the perceiver’s gaze moves to the target of perception.10 In this respect, they can be said to constitute a case of metaphorical duals.
—————
10
The existence of visual metaphors of the former type can be indirectly supported by the unacceptability of the following sentences (Lakoff 1993a: 232): (i) *From my office, the view blew me away. (ii) *From my office, the view knocks me over. (iii) *From my office, the mountain peeked at me through the fog. The main sentence of each example metaphorically conceptualizes the perceiver as a visual goal, while the from-phrase conceptualizes the perceiver as a visual source. That is why the above sentences are unacceptable.
Metaphorical modes of perception and scanning
167
3. Sense modalities and perceptual metaphors So far we have examined some crucial aspects of visual metaphors observed in Japanese and English. Of special interest are those visual metaphors which are based on our tactile experience. It is interesting in this connection to compare visual metaphors with those involved in other sense modalities. The facts observed so far (especially those in section 1.2 and 1.3) lead us to predict that tactile metaphors, which characterize our visual experiences, will further be involved in other sense modalities such as auditory and olfactory modalities. Some body-part idioms in Japanese lend partial support to this prediction: (15) (i) a. hana ni tuku nose DT attach ‘attach to the nose’ b. hana o iru nose ACC penetrate ‘penetrate the nose’ (ii) a. ko-mimi ni hasamu tiny-ear DT put-in ‘happen to hear’ b. mimi-zawari ear grating ‘grate on the ear’ Here, auditory and olfactory perceptions are metaphorically expressed in terms of tactile experience, as reflected in the basic meanings of the verbs in (15) (e.g. tuku ‘attach’, iru ‘penetrate’, hasamu ‘put-in’, etc.). This can be further supported by the following examples, where auditory and olfactory perceptions are figuratively expressed in terms of verbs of contact or hitting: (16) a. Kare wa sono kyooretu-na nioi ni utinomes-are-ta. he TP the strong smell by struck-be-PST ‘He was struck dumb by the strong smell.’ b. Tyoosyuu wa mina sono hagesii oto ni utinomes-are-ta. audience TP all the heavy sound by struck-be-PST ‘The audience were all struck dumb by the heavy sound.’ It must be noted, however, that auditory and olfactory expressions are more limited than visual expressions in their figurative usage. As was examined in the previous section, visual expressions basically involve duality. This being the case, it is not logically impossible to expect the existence of duality in auditory and olfactory domains of perception. This, however, turns out not to be true, as is clear from the following examples: (17) a. Taro wa otera no yane ni me o muke-ta. Taro TP temple GN roof to eye ACC turn-PST
168
Masa-aki Yamanashi
‘Taro turned his eyes to the roof of a temple.’ b. *Taro wa koosui no bin ni hana o muke-ta. Taro TP perfume GN bottle to nose ACC turn-PST ‘*Taro turned his nose to the bottle of perfume.’ c. *Taro wa sono razio ni mimi o muke-ta. Taro TP the radio to ear ACC turn-PST ‘*Taro turned his ear to the radio.’ The auditory and olfactory expressions in (15) and (16) represent typical instances in which the sense stimuli of the target moves to the perceiver. This leads us to expect the existence of complementary cases in which the perceiver (or perceiver’s sensory organ) moves to the target of perception. The sentence in (17)a illustrates such a complementary case. In contrast, however, the corresponding auditory and olfactory expressions in (17)b and (17)c are not acceptable. Facts of this sort show that duality cannot be found in auditory and olfactory domains of perception. The observations made here apparently seem incompatible with the generality of tactile metaphors which are ubiquitously involved in such cognitive domains as vision, smell and sound. In the case of (17), for example, the source domain of touch can be metaphorically mapped onto the target domain of vision; but the former domain cannot be metaphorically mapped onto the target domain of smell and sound. Why is this the case? The answer lies in the mode of perception involved in these sense domains. Lakoff (1993a) pointed out that the mode of visual perception is “directed” to the target, i.e., visual metaphors map the “directability” of limbs (cf. (i)–(v) in Section 2.2) onto that of vision. By contrast, the mode of auditory and olfactory perceptions is “not directed”, i.e., the sense stimuli of sound and smell are perceivable from all directions at once. These perceptual constraints lead us to give a natural account of the unacceptability of the sentences in (17)b and (17)c above. They both contain a verb mukeru, whose mode of action is “directed”. This, however, runs counter to the fact that auditory and olfactory perceptions are “not directed” in the way visual perception is. Given such constraints on the mode of perception, the acceptability of (18)a can also be predicted as a natural consequence. 4. Cognitive constraints on the target domain The facts observed in the previous section have some significant theoretical implications. They lend support to the validity of the Invariance Principle proposed by Lakoff (1993b: 215–6): “Metaphorical mappings preserve the cognitive topology [...] of the source domain, in a way consistent with the inherent structure of the target domain.” This is a consequence of the fact that the knowledge structure of the target domain automatically imposes inherent limitations on what can be mapped from the source domain. What this principle requires is so called “target domain overrides” (i.e., the inherent limita-
Metaphorical modes of perception and scanning
169
tions of the target domain should cancel any knowledge of the source domain inconsistent with it). This principle leads us to understand in a natural way why auditory and olfactory domains of perception cannot allow such linguistic expressions as (17)b and (17)c above.11 The principle further predicts the acceptability of the following Japanese phrases which are based on our perceptual experiences of vision, sound, and smell. (18) a. ~ ni me o otosu to eye ACC drop ‘drop one’s eyes to ~’ b. ~ ni me o tenziru to eye ACC shift ‘shift one’s eyes to ~’ c. ~ ni me o nagekakeru to eye ACC cast ‘cast one’s eyes to ~’ (19) a. *~ ni mimi o otosu to ear ACC drop ‘drop one’s ear to ~’ b. *~ ni mimi o tenziru to ear ACC shift ‘shift one’s ear to ~’ c. *~ ni mimi o nagekakeru to ear ACC cast ‘cast one’s ear ~’ (20) a. *~ ni hana o otosu to nose ACC drop ‘drop one’s nose to ~’ b. *~ ni hana o tenziru to nose ACC shift ‘shift one’s nose to ~’ c. *~ ni hana o nagekakeru to nose ACC cast ‘cast one’s nose to ~’ The verbs otosu, tenziru, nagekakeru designate an action which involves the mode of “direction”. In (18–20), these verbs cooccur with a direct object based on perception. It should be noted, however, the auditory and olfactory ————— 11
Basically the same holds for the following English examples of auditory and olfactory perception (Lakoff 1993a: 234). (i) *He ran his ears over the crowd. (ii) *He ran his nose over the shoreline. (iii) *His nose was glued to the mountaintop. (iv) *He couldn’t take his ears off the violinist.
170
Masa-aki Yamanashi
domains of perception are inherently ‘not directed’. Therefore, the metaphorical mappings of the directability onto hearing and smell are necessarily overridden. That is why the examples in (19) and (20) are unacceptable. This kind of consequence is exactly what is predicted by the Invariance Principle. It is interesting in this connection to note the surface distribution of such phrases as (i) {a. me o tukeru ‘attach one’s eyes to’, b. *mimi o tukeru ‘attach one’s ear to’, c. *hana o tukeru ‘attach one’s nose to’} and (ii) {a. me o hanasu ‘take one’s eye off’, b. *mimi o hanasu ‘take one’s ear off’, c.*hana o hanasu ‘take one’s nose off’}. Though the verbs tukeru and hanasu as such do not directly reflect the mode of direction, they involve this kind of mode as one of their presuppositional meanings. Thus, basically the same kinds of constraints (i.e. constraints to be imposed on auditory and olfactory domains of perception) predict in a natural way the acceptability of (i)a–c and (ii)a–c above. 5. Seeing based on transformation of the visual field Observations made so far show that natural language provides a variety of effective and flexible means to express the ways in which we perceive and construe the external world. Metaphor serves as one of the most effective means to achieve this purpose. Much of our understanding of the external world is characterized by metaphorical modes of perceptual experience which should constitute the cognitive foundation of natural language. Linguistic phenomena observed in the previous sections illustrate some interesting ways in which our visual experience is expressed in linguistic terms. The following examples also illustrate another case where we visually perceive the external world: (21) a. Yagate buraku no hazure ni de-ta rasiku, miti ga sakyu no ryosen ni kasanari, sikai ga hirakete, hidarite ni umi ga mie-ta. ‘At last he seemed to be coming to the outskirts of the village, and the road lay atop the ridge of sand dunes; the view opened out, and to his left he could see the sea.’ (Abe Kobo The Woman in the Dunes: 168) b. ... sikai ga hirakete ... visual-field NOM open-out ... ‘... the view opened out ...’ (22) a. Onaziyoona huukei ga saigen-mo-naku tuzuku no da. Sorekara, totuzen sikai ga hirakete, tiisana buraku ga araware-ta. ‘The unchanging landscape stretched endlessly on. Then suddenly, the perspective broadened and a hamlet came into sight. (Abe Kobo The Woman in the Dunes: 8) b. ... totuzen sikai ga hirakete ... suddenly, visual-field NOM open-out ‘... suddenly, the perspective broadened ...’
Metaphorical modes of perception and scanning
171
(23) a. Hatizyoo doori o nisi ni mukai, katura-oohasi ni itaru to, kyuuni sikai ga hirogaru. Sore wa nanboku ni nagareru Katura-gawa ga huukei o nobiyaka-ni site-iru kara da. ‘Heading west on Hachijo Street and reaching the Katsura Bridge, a vista suddenly opens as the expansive flood plains of the south-flowing Katsura River come into view.’ (Minoru Senda Kyo no Ikizukai: 58) b. ... kyuuni sikai ga hirogaru ... suddenly visual-field NOM open ‘... a vista suddenly opens ...’ What is interesting about (21–23) is the expression sikai ga hirake-ru (cf. (21)b and (22)b) or sikai ga hirogaru (cf. (23)b). This kind of expression literally means that the visual field opens out (or the perspective broadens), but it figuratively means that one comes to be in a position where a vista opens. The opposite meaning can be indirectly conveyed by such expressions as sikai ga (or kara) kireru in (24)b and (25)b: (24) a. Kisoku-tadasiku kizamare-ta huumon o yokogitte susumu to, huini sikai ga kirete, hukai hora-ana o miorosu gake-giwa ni tatte iru no dat-ta. ‘Suddenly his line of vision was cut off, and he stood on the verge of a cliff looking down into a deep cavity.’ (Abe Kobo The Woman in the Dunes: 18) b. ... huini sikai ga kirete ... suddenly visual-field NOM cut-off ‘... suddenly his line of vision was cut off ...’ (25) a. Mayoi nagara hurikaette miru to, saiwai hinomi wa kodakai suna no ryuuki ni saegi-rare-te, sikai kara kirete iru. ‘Worried, he turned and looked behind him; fortunately, the fire tower was cut off from view by a slight rise in the sand.’ (Abe Kobo The Woman in the Dunes: 169) b. hinomi wa ... sikai kara kirete iru ... fire-tower TOP visual-field from cut-off be ‘the fire tower ... was cut off from view ...’ The expression sikai ga (or kara) kireru literally means that a visual field (or line of vision) is cut off, but it metaphorically means that one comes to be in a position where a vista disappears. The visual expressions examined so far, i.e. those in (21)b–(25)b, basically serve as some kind of metaphorical expressions. However, they are conventionally established expressions of metaphor which have to do with our visual experience. In contrast, we find some live metaphors which figuratively convey our visual experience of the external world. Consider the following examples:
172
Masa-aki Yamanashi
(26) a. Sato no miti ni deru to, titi no kakurete-iru amadera wa tikurin ni kakurete simat-ta. ‘When she reached the village road, the convent where her father had secluded himself was hidden from the road by a grove of bamboo.’ (Yasunari Kawabata The Old Capital: 35) b. amadera wa tikurin ni kakuretesimat-ta convent TOP bamboo into hide-itself-PAST ‘the convent ... was hidden ... by a grove of bamboo’ (27) a. Ame no oku kara, matu rasii mono ga tyoku tyoku kao o dasu. Dasu ka to omoo to, kakureru. Ame ga ugoku no ka ki ga ugoku no ka, yume ga ugoku no ka, nanto-naku husigi na kokoro-moti da. ‘At times within the filmy depths of haze, shadowy shapes of what might have been pine trees showed themselves, only to hide again in an instant. Whether it was the rain or the trees that was moving, or whether the whole thing was merely the unreal wavering of a dream, I did not know.) (Soseki Natsume The Three-Cornered World: 15) b. matu ... ga tyoku-tyoku kao o dasu pine-tree NOM at-times face ACC show ‘at times ... pine trees showed themselves ...’ The sentence in (26)b, i.e. amadera wa tikurin ni kakuretesimat-ta, literally means that the convent has hidden itself in a grove of bamboo, which in effect serves as a kind of personified expression. We know that it is impossible for the convent to hide itself literally in the actual world. What this sentence metaphorically means, therefore, is that the convent in question went out of sight. How about the personified sentence in (27)b, i.e. matu ga tyoku tyoku kao o dasu? The literal meaning of the sentence is that pine trees show their faces. This sentence figuratively means that pine trees comes into view. Sentences of this sort are special live metaphors which indirectly express our experience of visual perception. It must be noted that such sentences are not colloquial expressions; they are usually not used in ordinary conversation. We typically find sentences of this sort in the texts of literature or rhetorical discourse. In any event, the linguistic phenomena observed in this section show the existence of creative communicative power in natural language.
Metaphorical modes of perception and scanning
173
6. Visual metaphors and tropical truth In the previous sections, we have examined a variety of visual metaphors based on the perceptual organs of our body. Now one of the interesting questions we should ask is to what extent visual metaphors are universal. Metaphors in natural languages vary; some of them are widespread from language to language, some language-specific. The examination of Japanese and English visual expressions in this paper shows the existence of duals of visual metaphors in both languages (cf. Section 2). One type of metaphorical dual involves a mode of vision in which the target of perception moves to the perceiver, while in the other the perceiver’s gaze moves to the target of perception. It was also observed that both languages have visual metaphors which are based on tactile experience (cf. Section 3). Japanese and English visual expressions, however, show some differences in the linguistic coding of perceptual organs. In the case of English metaphorical duals, the body term eye(s) can cooccur with motion verbs to indicate the movement of the perceiver’s gaze to the target of perception, whereas the corresponding Japanese body term me cannot cooccur with motion verbs to indicate such movement (cf. Section 1.2). It is also interesting to note some basic differences in the linguistic coding of visual fields between the two languages. In Japanese, the body term eye(s) as well as shikai (= sight, view) can be used to metaphorically express a perceptual field as a container, while in English only the latter terms (i.e., sight, view) can metaphorically express a perceptual field as a container (cf. Section 1.1). Despite such differences, however, visual metaphors in both languages are grounded in common bodily experiences (e.g. tactile experience) (cf. Sections 1.2–1.3). Tactile experience constitutes part of the universal (embodied) base of our understanding of the world through visual metaphors. Due to the existence of such common bodily experience, the tropical truth underlying different languages can be guaranteed in various communicative contexts. This fact supports the general view of embodied cognition, i.e. the view that cognitive processes are deeply rooted in our body’s interactions with the world. The embodied basis of cognition means that there can be no absolutely objective truth; there are only metaphorical or tropical truths – i.e., just the kind of truths which are relative to our embodied and imaginative cognition of the external (or internal) world. Truth acquires its empirical validity through embodiment, especially via perceptual and motor capacities. The detailed examination of perceptual metaphors in this paper lends empirical support to this thesis. It must be noted, however, that truth is not absolutely (or purely) subjective. We all have pretty much the same embodied capacities (e.g. our ability to perceive, our ability to move and manipulate objects, etc.). This fact keeps truth from being purely subjective. It is based on the common nature of human beings and in this sense, therefore, it may be called intersubjective. 7. Conclusion
174
Masa-aki Yamanashi
The ways in which we perceive and construe the external world are directly or indirectly reflected in the conceptual system of natural language through a variety of perceptual metaphors. They constitute some of the most crucial metaphors which characterize the underlying conceptual systems across various languages. In this paper, we examined a wide range of perceptual metaphors involved in Japanese and English, with special reference to those involved in vision and other basic sense modalities such as sound and smell. Despite the fact that these languages are different in their historical origins, perceptual metaphors manifested in both languages are grounded in basically common bodily experiences; especially their common bases turned out to be tactile experiences. This fact suggests that our domain of tactile experience is highly likely to be one of the candidates which empirically constitute the universal base of our cognition and conceptualization of the external world. The examination of perceptual metaphors made in this paper is far from exhaustive. The facts clarified in the paper, however, will contribute to elucidating the linguistic and cognitive status of perceptual metaphors and the exact nature of tropical truth of conceptual metaphors. Visual metaphors examined in this paper also show that the conceptual metaphor of the saying Seeing is believing is a crucial starting point for the investigation of the tropical truth of natural language. Abbreviations used in glosses AC COP CM DT GN
accusative copula complement marker dative genitive
LC MM NM PST TP
locative modality marker nominative past topic marker
References A) Sources for the data Abe, Kobo: The Woman in the Dunes. Translated by E. Dale Saunders. Tokyo: Charles E. Tuttle 1967. Joyce, James: “The Dead”, in: The Best Short Stories of the Modern Age. Ed. by Douglas Angus. Greenwich, Conn.: Fawcett 1962, 119–59. Kawabata, Yasunari: The Old Capital. Translated by J. Martin Holman. Tokyo: Charles E. Tuttle 1988. Melville, Herman: ‘The Fiddler.’ 50 Great American Short Stories. Ed. by Milton Crane. New York: Bantam Book 1965, 28–34. Natsume, Soseki: The Three-Cornered World. Translated by Edwin McClellan. London: Arena Boo 1984. Senda, Minoru: Kyo no Ikizukai. Kyoto: Mitsumura Suiko Shoin 1999. Steinbeck, John: “The Snake”, in: Great American Short Stories. Ed. by Wallace Stegner & Mary Stegner. New York: Dell 1957, 410–22.
Metaphorical modes of perception and scanning –
175
“Flight.” Short Story Masterpieces. Ed. by Robert Penn Warren & Albert Erskine. New York: Dell 1954, 454–74. B) Linguistic literature Boulding, Kenneth E.: The Image. Ann Arbor: The University of Michigan Press 1956. Gruber, Jeffrey S.: “Look and See”, in: Language 43(4) (1967), 937–47. Lakoff, George: “The Invariance Hypothesis”, in: Cognitive Linguistics 1(1) (1990), 39–74. – “The Metaphor System and Its Role in Grammar”, in: Papers from the TwentyNinth Regional Meeting of Chicago Linguistic Society. Chicago: Chicago Linguistic Society 1993a, 217–41. – “The Contemporary Theory of Metaphor”, in: Ortony, Andrew (ed.): Metaphor and Thought , Cambridge: Cambridge University Press 1993b, 2nd ed., 202–51. – , Mark Johnson: Metaphors We Live By. Chicago: The University of Chicago Press 1980. Langacker, Ronald W.: Foundations of Cognitive Grammar. Vol.1. Stanford: Stanford University Press 1987. – “Subjectification”, in: Cognitive Linguistics 1(1) (1990), 5–38. – “Reference-Point Constructions”, in: Cognitive Linguistics, 4(1 (1993), 1–38. Talmy, Leonard: “The Relation of Grammar to Cognition”, in: Rudzka-Ostyn, Brygida (ed.): Topics in Cognitive Linguistics. Amsterdam: John Benjamins 1988, 165– 205. Yamanashi, Masa-aki: “Metonymic Interpretation and Associative Processes in Natural Language”, in: Nagao, M. (ed.): Language and Artificial Intelligence. Amsterdam: North-Holland 1987, 77–86. – Hiyu to Rikai [Metaphor and Understanding]. Tokyo: The University of Tokyo Press 1988. – Ninchi Bunpooron [Theory of Cognitive Grammar]. Tokyo: Hitsuji-Shobo. 1995. – “Visual Perception, Tactile Experience, and Spatial Cognition”, in: Ukaji, M. et al. (eds.): Studies in English Linguistics. Tokyo: Taishukan 1997, 837–45. – Ninchi Gengogaku Genri: Principles of Cognitive Linguistics. Tokyo: Kuroshio Publishers 2000. Zipf, George K.: The Psychobiology of Language. New York: Houghton Mifflin 1935.
Cornelia Zelinsky-Wibbelt (Hannover, D)
Natural Language Processing: Minds, brains, and programmes
1. 2. 3. 4. 4.1 4.2 4.3 4.4 4.5
Introducing the computer metaphor Metaphors, analogy, and truth Extending the computer metaphor to natural language processing The meaning of “meaning” Bottom-up vs. top-down processing Discourse context and relevant meaning Deterministic vs. non-monotonic linguistic processing The expert paradox From implicit to explicit representations Reference by mental models The brain metaphor Recursive reasoning and the first-person perspective 4.9 The embodied mind 4.10 The environmental metaphor 5. Conclusions 1.
Introducing the computer metaphor
Ordinary language use universally abounds with machine metaphors, such as the following: (1) My mind just isn’t operating today! (2) My mind is a bit rusty today. (3) My answering machine is talking to me. These metaphors rely on a folk psychological model which compares the the mysterious functioning of the human mind with that of a machine (cf. Lakoff 1987: 344f.).1 This naive view of the human mind has a long tradition and goes back to antiquity which started to carve up the world into two distinct —————
1
These examples illustrate two directions in which this comparison is achieved. Examples (1) and (2) explain the functioning of the human mind by that of a machine while example (3) personifies the machine through the attribution of human abilities. We will see within the course of this paper that these mutual comparisons are commonplace in the construction of the machine metaphor.
178
Cornelia Zelinsky-Wibbelt
realms, into mind and matter, in order to be able to legitimate the former as functioning independently of the latter. The continuous convention and desire to disembody the pure mechanism of the latest gadget by talking about its observable behaviour in terms of a mental operation is evidence of the ease with which common sense theories have been conquered by philosophical dualism. More recently, the machine metaphor is being promoted in the domain of computer technology, where the classical architecture of a computer is exploited to provide a model for the human mind as a pure information processing system. Consider the following ordinary language examples which reduce the functioning of the human mind to that of a computer programme: (4) (5) (6) (7) (8)
This news is very hard for me to process! The human mind is preprogrammed. Our memory has storage and retrieval limitations. Frequent mental processes become automatic routines. Consciousness is a feedback phenomenon. (Cf. Boyd 1993)
Beyond these manifestations in ordinary language the computer metaphor has put the ancient machine metaphor into a new guise in the philosophy of mind, which gave rise to the interdisciplinary field of cognitive science, in relating neuroscience to the newly emerging domain of computer science in order to interpret the relation between brain and mind computationally. Today the analogy between human and machine intelligence is one of the major topics in cognitive science, where linguists meet psychologists, philosophers and adherents of artificial intelligence. In this interdisciplinary collaboration, we may linguistically observe several levels of reducing the human mind. Let us consider some typical instantiations of the computer metaphor. These are common parlance in the development of the computational theory of mind in cognitive psychology. It is particular the design of human memory which seems to inspire psychologists in the creation of these expressions.2 (a) The mind is an information processing system which applies certain procedures with routines and subroutines. (b) The mind is a hybrid system, capable of parallel and serial processing. Within this system, a central processor controlls the peripheral processors. Human memory has storage and retrieval limitations (after Eysenck ————— 2
While the concept of the memory of the computer’s hard disk has initially been coined by analogy to our common-sense knowledge of human memory, nowadays the architecture of human memory is described on the model of the memory of a computer (cf. Neumann 1958; Wiener 1948: 301), without the concept in the domain of computation being well defined in the first place.
NLP: Minds, brains, and programmes
179
& Keane 2005). (c) Representations become (re-)activated in working memory which has a limited capacity store. Information is held on a stack, a last-in, first-out memory (after Garnham & Cowles, to appear 2009). (d) Automatic processes are fast, do not reduce the capacity for performing other tasks, suffer no capacity limitations, function rapidly and in parallel, but are very difficult to modify, whereas controlled processes are flexible and versatile but operate relatively slowly and in a serial fashion (after Eysenck & Keane 2005). (e) Two messages gain access in parallel to a sensory buffer. One of the inputs is then allowed through a filter on the basis of its physical characteristics, with the other input remaining in the buffer for later processing. This filter is necessary in order to prevent overloading of the limited-capacity mechanism beyond the filter (after Eysenck & Keane 2005). (f) Circuits strengthen and build up connections. Incoming data travel along connections strengthened by previous, similar input, with associative connections among them. The activation is spread further along an outgoing and incoming path (after Bartsch, to appear 2009). (g) A single digit appears on the shadowed or nonshadowed message. 8% of the digits on the nonshadowed message suggest very limited processing of that message (after Eysenck & Keane 2005). (h) Automatic processes operate in parallel and place no demands on capacity if there is a slope of zero in the function relating decision speed to the number of items in the memory set and the visual display (after Eysenck & Keane 2005). (i) “When a person you know has fallen in love, it’s almost as though someone new has emerged—a person who thinks in other ways, with altered goals and priorities. It’s almost as though a switch had been thrown and a different program has started to run.” (Minsky 2006). The statement in (a) propagates the dualistic disembodiment of the mind as a pure information processing system solely manipulating symbols without intentional reference to the outside environment. Examples (b) to (e) portray the human mind in terms of a biological computer. This is built on the model of the von-Neumann-architecture of a traditional computer disposing of a central processing unit with a limited storage capacity controlling several peripherally distributed information-processing components working sequentially (cf. Neumann 1951). Example (f) overcomes the simplistic analogy of the computer metaphor by modelling the human mind in terms of the brain metaphor of neural processing3. In this image the activities of the brain are —————
3
Initially, a new generation of computers had been developed which took the human brain as a model for the development of parallel distributed processing systems. This in turm gave rise to reconsired the human brain on the model of these
180
Cornelia Zelinsky-Wibbelt
distributed over a multitude of spatially distributed networks, the smallest ones being the neurons activating each other in parallel through their connections. Finally, examples (g) and (i) illustrate the ontological conflation of externalized, observable behaviour and internal mental qualities, which are caused by a sufficient amount of neuron firings in the brain, without any additional subjective experiencer being necessary to legitimate consciousness. We may summarize the different levels of reduction achieved in these examples by reference to their common commitment to ordinary language philosophy. Ordinary language philosophy abandoned Frege’s claim of absolute truth in terms of necessary and sufficient conditions unambiguously determining the extension of a term (cf. Frege 1892). By relying on ordinary language, the empirical observation of mental qualities became obsolete. Instead, the functioning of the mind was to be inferred from a corresponding behaviour (cf. Wittgenstein 1953; Quine 1960). As a corollary, we may never verify the truth of a theory of mind. Only in making reference to particular concepts by following the rules of ordinary language use may we evaluate the sense of the referring words in terms of their correspondence to our knowledge and experience. Language thereby becomes a medium of representing and probing the functioning of the human mind. The reduction of the human mind to a computer programme goes back to Alan Turing’s classic paper of 1950 (cf. Turing 1950; Hild & Stemmer 2007), which purports the theoretical possibility of building a machine capable of producing a behaviour which is indistinguishable from human intelligence: the Turing machine. With this Turing test, Alan Turing was convinced to have verified ‘artificial intelligence’, which was revolutionary at that time, since mental qualities, such as reasoning, problem-solving, and language performance, which had formerly distinguished human beings as unique, were now ascribed to a machine.4 Originally the assumption was that by writing the adequate programmes, humans could extend their own intelligence to that of their tools (cf. Weizenbaum 1984: 9). Within the newly emerging field of cognitive science this endeavour was then devoted to the discovery of human intelligence by way of simulating it in artificial systems. The Turing test thus gave birth to the discipline of artificial intelligence (AI) which in its strong version claims that the human mind is to its brain what the software of the computer is to its hardware (cf. Minsky 1990). Even more surprisingly, the concept of mind as an equivalent of the brain’s neural activities (cf. Penrose 1994; Grush & Churchland 1995), becomes disembodied and immaterialized by the coalition of strong AI with philosophical functionalism, for instance by Marvin Minsky (2005), one of its strongest defenders. As Searle put it in his critique of Strong AI, “the brain does not ————— 4
computers. On the other hand, the computer metaphor was a welcome alternative to the behaviourist “black box” metaphor of the human mind which denied any scientific account of unobservable mental phenomena.
NLP: Minds, brains, and programmes
181
really matter to the mind. Because the mind is a computer program, and because a program can be implemented on any hardware whatever” (cf. Searle 1992: 44). By contrast, the adherents of Weak AI use the computer as a tool which can support human beings most effectively if their intelligence is taken as a model as far as possible, but which cannot literally substitute their brain power. Whether the feasibility of Strong AI may be evaluated in principle, hinges on the possibility of a theoretical solution of the mind-body problem (cf. Johnson-Laird 1983: 475ff.). Although today computers and robots can achieve the most marvelous things, so far a substitute of the human mind is not foreseeable for the near future. From the relativist stance of ordinary language philosophy, introduced by Wittgenstein (1953) as intending to ban empirical treatment from philosophy, any theoretical treatment of philosophical questions is denied (cf. Bennett et al. 2007). Searle (2007), in contrast, groups the mind-body problem on a par with the problem of life as figuring among the very rare topics which may be granted a theoretical solution in philosophy. What became one of his most stimulating arguments against the equivalence between living and artificial systems within all branches of cognitive science was the argument which Searle (1980) developed in his paper on “Minds, Brains, and Programs”. Yet, one of the most frequent arguments brought forth in favour of Strong AI is the epistemological claim of the as yet incomplete decomposition of the human brain into the necessary subsystems (cf. Minsky 2005). As soon as this mereological problem will be solved, so it is claimed, the equivalence between brain and mind will be evident. In order to evaluate the contrasting philosophical implications involved in the positions of Strong and Weak AI, we have to consider the question of whether such a mechanistic decomposition of the human brain would be both necessary and sufficient in order to enable artificial systems to cope with the intelligent behaviour of living systems. We may start the evaluation of the necessity of this decomposition by consulting, in accordance with Searle, the achievements in the simulation of life. The simulation of life which intends to match the advanced achievements of living organisms is a topical objective in many technological fields. For our purposes, we may compare the objective of building natural language processing systems with the technological development of aircraft industry (cf. Hausser 1989). Throughout history human beings have studied the flight of birds, for instance the flapping of their wings, in order to learn how to fly. Just as experience and increasing theoretical knowledge about the functioning of airplanes led to a better understanding of the flight of birds, experience and growing theoretical knowledge in building natural language processing systems have resulted in a better understanding of the cognitive processes involved in human language comprehension and production. Nevertheless, while airplanes function basically on the same principles as birds, a sparrow and a jumbo jet work with radically different techniques. Just as an airplane ultimately is not evaluated by its resemblance to birds, but by its ability to fly,
182
Cornelia Zelinsky-Wibbelt
natural language processing systems need to be measured by their linguistic output instead of by their affinity to the human mind. From the perspective of ordinary language philosophy on language use as a medium of representation and evaluation, this excursion to modern aircraft technology in turn brings us back to the question about the linguistic legitimacy of the computer metaphor as a model for explaining the human mind. From a linguistic point of view, we may evaluate the feasibility of Strong AI in terms of how far we may push the analogy between human and machine intelligence through language use. After all, an investigation of the use of metaphors should first of all lie in the responsibility of the linguist. Thus, we may continue to investigate the claim of Strong AI about the decomposition of the human brain by considering the question of sufficiency from a linguistic point of view. This question burns down to the age-old philosophical discussion about whether the meanings of linguistic expressions can be verified in terms of ultimate truth conditions or essences. We will see that a linguistic consideration of the concept of mind in terms of essential properties is inevitably related to the Cartesian dualism of generative linguistics which explains linguistic knowledge as the competence which is engendered by the hard-wired and innate linguistic faculty in the left hemisphere of the human brain (cf. Chomsky 1986: 12f.). Within the philosophy of generativism, this distinction has in turn necessitated a further dualistic conception of the speakers’ stable and linguistic knowledge of a finite set of rules referred to as competence on the one hand, as enabling their infinite application of these rules in linguistic performance on the other. Chomsky’s mentalism culminates in the empirical paradox that the deeply unconscious mental representations of the speakers’ linguistic competence “underlying” their linguistic performance are not amenable to the empirical investigation of the linguist, but may only be accessed via introspection (cf. Chomsky 1957: 13, 91). Linguistic performance became discarded from linguistics proper as the speakers’ inconsistent and rule-violating linguistic behaviour to be investigated by pragmatics. In fact, this solipsistic scientific regress in favour of human essence as being established by the speakers’ inborn linguistic faculty which universally determines the principles and parameters of linguistic competence prevents a responsible consideration of language use.5 We will see within this paper that the defense of linguistic essences as pursued by linguistic Cartesianism does not provide a viable methodology for evaluating the implications of the computer metaphor. 2. Metaphors, analogy, and truth Metaphors are exploitations of already existing forms, in that they are semantic extensions from a more concrete, better known, often sensorily per—————
5
Of course, this solipsism of Cartesian linguistics is inherited from Descartes.
NLP: Minds, brains, and programmes
183
ceivable source domain, to a more abstract, weakly known, semantically more complex target domain. In this way metaphorical language use is a problem solving device by mediating between source and target domain through creating an analogy between the subject’s already existing knowledge and the object of learning. According to the theory of Mental Models metaphors are models which are created intentionally for achieving epistemic success in the target domain and thereby are theory-constitutive (Boyd 1993: 518ff., Hugh & Oshlag 1993: 583f., Mayer 1993: 568ff.). On account of their self-referential nature, speakers are capable of consciously evaluating their epistemic success in the use of metaphors, by mentally locating themselves on a higher level, where they have meta-knowledge at their disposal for evaluating their lower-level, object-language representations. They may thereby verify the mental representations of metaphors and falsify those which are incompatible with the experienced situation. Such incompatible mental representations typically result from a manipulation of reality which lacks any superordinate distance to itself. Non-manipulatory metaphors are created from an intentional need for a better representation of the target domain. In this way, speakers are responsible for constantly monitoring and revising their representations to make them more compatible with reality. By providing a model of reality, metaphors do in fact explain a central methodology of human reasoning and language. Metaphors establish the methodology by which speakers reason in terms of models, and gain knowledge by recursively transcending their target domain and eventually their object language representations. By evaluating their knowledge with respect to how well the corresponding expressions’ meanings fit into their knowledge, their experiences, and their intentions, speakers represent knowledge about knowledge. 3. Extending the computer metaphor to natural language processing From these initial considerations on the conceptualization of the human mind through the computer metaphor the following question arises: while the physical material and structure underlying a natural language processing system is very different from those of human speakers, cognitive linguists have to answer the question of whether and to what extent the speakers’ behaviour may provide a model for natural language processing systems. By focussing on the meaning of language, we will evaluate whether it is both possible and necessary for a computational system to know the ultimate nature of the meanings which humans associate with linguistic utterances. In order to perform this evaluation, a brief insight into the organization of natural language processing systems is provided for those readers who are not familiar with this domain.
184
Cornelia Zelinsky-Wibbelt
While linguists are concerned with writing grammars and lexicons in order to explain how natural language works, computational linguistics is concerned with the implementation of grammars and lexicons in computational formalisms. The objective of this implementation is to provide a natural language processing (NLP) system, as is schematized in figure 1 for illustration. A natural language processing system is a computational programme consisting of linguistic rules. If this system receives a linguistic input, it is capable of producing a linguistic output by applying the implemented rules. Such a system may be applied to specific purposes, such as machine translation, the building of spelling checkers or second language acquisition tools. The relevant question for cognitive linguistics is whether we may consider the linguistic behaviour of NLP systems as intelligent compatible with that of human speakers. During the 1980s, considerable progress was made in computational linguistics. Whereas previous-generation programming languages, such as FORTRAN and BASICS only allowed for a very inconvenient and untransparent programming style, and hence engendered fairly bad results, higher-order programming languages, such as LISP and PROLOG now enabled linguists to write grammars and lexicons directly in computer-implementable formalisms, whereby the quality of the output was considerably improved. Furthermore, during the last two decades the interdisciplinary status of computational linguistics has joined linguists, psychologists, philosophers, computer scientists as well as translation scientists, terminologists and special language researchers. The joint effort of this community has made obvious that a methodologically efficient implementation has to consider the way in which humans process natural language as far as possible, but no further than necessary (cf. Fass 1997; Hobbs 1991). 4.
The meaning of “meaning”
In order to evaluate the intelligence of natural language processing systems, the following question has to be considered: what does it actually mean to understand the meaning of an utterance? In order to pursue this objective, we have to compare how natural language processing is achieved by a computational system with how humans achieve this goal. 4.1. Bottom-up vs. top-down processing Consider the following sentence which a natural language processing system would receive as written input: (9) Fred asked the man on the boat.
NLP: Minds, brains, and programmes
185
Figure 1: Architecture of an NLP system Natural language processing systems generally start with segmenting the input string into words and punctuation marks and output a set of sentences. This serves as input to the parser which, by applying the phrase structure rules and the lexicon, would analyse each sentence into successively smaller parts of speech in a strictly bottom-up fashion. At the lowest level of analysis the lexical units become inserted and evaluated with respect to their syntactic and semantic agreement with the context. This analysis would produce two representations of example (9), as given in figures 2 and 3. A purely sentence-bound analysis would remain ambiguous as the system would be incapable of selecting the intended meaning. This would require pragmatic information mentioned in the larger context of the sentence. Either both men are interpreted as being on the boat, as represented in figure 3, or only the man is taken to be on the boat, as represented in figure 2.
186
Cornelia Zelinsky-Wibbelt
Figure 2: Fred asked the man. The man was on the boat It is doubtful whether speakers use their native language in this rule-based way. It rather seems that in this way they would get lost in the multitude of implicit meanings conveyed in natural language discourse. Very short reaction times in psycholinguistic experiments suggest that ever more complex utterances do not cause any significant delay in their processing. In fact, hearers deal with the most bizarre linguistic phenomena in fractions of milliseconds, indicating that they grasp the meanings of sentences in relation to the overall discourse context (cf. Strohner 1990). This is not achieved by stepwise explicit rule application but rather by implicit lateral thinking, which results in a representation of a text as a whole in terms of its multidimensional connections (cf. Neubert 1994; Beaugrande 1997). This meaning-based, holistic linguistic behaviour is obvious from psycholinguistic experiments in which listeners anticipate the meaning of a sentence even after the first word, due to top-down oriented inferences.
NLP: Minds, brains, and programmes
187
Figure 3: Fred asked the man. Both were on the boat If we apply these psycholinguistic observations to natural language processing, we come to realize that purely bottom-up oriented sequentially processing systems are insufficient tools for natural language processing. A detailed formal analysis of all explicit linguistic phenomena would have to be complemented by top-down oriented parallel processing. This is more likely to take care of implicit meanings lurking behind the many cases of vagueness, contradictoriness and ambiguity typical of the casual style of normal conversations. Such hybrid natural language processing systems composed of different subsystems should be capable of providing more satisfactory results of normal linguistic input (cf. Johnson-Laird 1988: 191; Kintsch 1998: 93ff.). This will be the issue of section 4.6 below. 4.2. Discourse context and relevant meaning In the preceding section we have observed that lexical meaning evolves in specific discourse contexts. According to Sperber and Wilson’s Relevance Theory of Communication speakers continuously negotiate the intended
188
Cornelia Zelinsky-Wibbelt
meaning by selecting the relevant information within the relevant discourse context and ignoring the irrelevant pieces of information (cf. Sperber & Wilson 1995, Winograd & Flores 1986: 49). In order to achieve the most relevant interpretation of an utterance listeners initially consider the immediately preceding context. Only if this proves insufficient, listeners may incrementally broaden the context. The structural ambiguity of the sentence Fred asked the man on the boat would not arise if the immediately relevant context were either Fred walked along the river or Fred entered the boat (cf. Forster 1981; Schoen 1988). In accordance with relevance theory of communication the ability to recognize the relevant relationships may be a universal disposition of human reasoning. Communication involves the comparison of interpretations. In order to agree in their interpretations speakers negotiate their mental representations with minimal cognitive effort and maximal cognitive benefit (cf. Sperber & Wilson 1995). This is the principle of relevance (cf. figure 4).
Figure 4: Communication and the principle of relevance The linguistic intelligence of humans thus crucially consists in grasping the relevant relations in very economical ways. This insight has given birth to the comparison of human chess players with automatic ones. Automatic systems, however, can only calculate a successful procedure according to explicit rules. Enabled by a large working memory, they may compute vast numbers of possible combinations, and thus may well be superior to human beings. Humans, on the other hand, may heuristically grasp a small number of lowlevel concepts as potentially representing the relevant information, such that the relations between them represent the top-level analogies to existing knowledge. From these analogies new symbols emerge (cf. Hofstadter 1996: 205). On this view, there is the as yet ineffable way of economical information processing speeding up human categorization. Categorization is achieved against the background of the speakers’ encyclopedic knowledge which, although representing a generalization from experienced information, is enor-
NLP: Minds, brains, and programmes
189
mous, both with respect to complexity and quantity. How is it that speakers manage to access and focus efficiently on those components of their knowledge which in the respective discourse situation figure as most relevant, such that new information, augmenting their existing knowledge, emerges effecttively? As yet, this economical behaviour remains by and large an enigma to psychologists and linguists as well as to computational linguists in particular. The answer to this question obviously would provide the key to human cognition and linguistic performance. Some preliminary explanations will be provided in the remaining sections. 4.3. Deterministic vs. non-monotonic linguistic processing As we have seen, the linguistic intelligence of humans crucially consists in the ability to process implicit information. This also comprises new information for which no automatic routines are available. In contrast to this flexible and creative behaviour, traditional natural language processing systems work in a strictly deterministic fashion, in that they can only process information which is compatible with the rules implemented in grammar and lexicon. Deterministically, each rule application follows from the preceding one according to a two-valued logic in a procedural fashion. Accordingly, the input sentence (10) Fred asked his database for advice. will be evaluated as false if the lexical rule of the verb ask constrains the verb to be followed by an noun phrase of which the noun is lexically described as HUMAN. What is at issue here is the notion of linguistic creativity. A natural language processing system with a stable grammar and a lexicon containing a finite set of rules is capable of generating infinitely many sentences. This mutation effect corresponds to the generativist notion of creativity (cf. e.g. Chomsky 1957, 2007). By contrast, from a cognitivist perspective, speakers are creative in continuously extending the rules in view of newly encountered information. This creativity is achieved through the construction of hypothetical examples by analogy to existing rule schemes (cf. Harnard 2007: 127ff.). Consider morphological blends such as infotainment, infomercial, which are created and immediately understood by analogy to lexical units already morphologically analyzed in the lexicon, such as entertainment, information and commercial, which in our case provide a model for new word formations. Humans, that is, can learn: speakers can evaluate new information and integrate it into their existing knowledge if the information is sufficiently compatible with it, as is the case with Fred asked his database for advice (Johnson-Laird 1983). Speakers can infer that a database is a human product which only contains knowledge of the human programmer who has thus extended his own intelligence to that of the database, which thereby may well
190
Cornelia Zelinsky-Wibbelt
follow the verb ask. That is, speakers are capable of relating the PRODUCT sense of database to its producer via a conversational implicature. If this inference is drawn sufficiently frequently, it may result in a metonymic lexicalization. But how do speakers infer these conversational implicatures in the first place? Prototype semantics claims that speakers reason non-monotonically, that is non-deterministically. Word senses in terms of necessary and sufficient conditions which in each discourse situation unambiguously determine the speakers’ reference do not explain the speakers’ flexibility and creativity. Speakers do not categorize the objects which they experience into concepts in an all-or-nothing fashion. Prototype semantics has provided evidence for the speakers’ flexible categorization of borderline cases, that is “bad” members, in the particular communicative situation. Although there are good reasons for it, there are no necessary and sufficient conditions for a CONTAINER to be referred to either by the nouns cup or mug (cf. Labov 1973). Cups may occur in a variety of shapes, with and without handle, they may be broad or narrow, so that, what is typically called a mug or a bowl, may well be referred to by the noun cup. The boundaries between concepts are negotiated between speakers in accordance with the experienced discourse situation. This is evidence for the fuzziness of conceptual boundaries. Accordingly, the assumption is that speakers draw default inferences about the preferred, normal interpretation which may be overridden and defeated by additional contextual information in the discourse. The ability to process information defeasibly is an essential feature of human intelligence: it is the ability by which assumptions made earlier in the discourse may be rejected by information processed later. Defeasible reasoning proceeds non-deterministically and can therefore cope with otherwise conflicting information (McDermott & Doyle 1980, Reiter 1980). Default inferences are drawn due to systematic expectations about language use. This has been introduced as Generalized Conversational Implicatures by Grice (1975). By frequent activation, these implicatures may feed back on the semantic structure of language in the process of lexicalization (cf. Levinson 2000; Tyler & Evans 2003). 4.4. The expert paradox Another characteristics of the human ability to process the relevant information is discussed as the expert paradox. Whereas humans are capable of accessing the relevant parts of their knowledge more rapidly with increasing amount of knowledge, artificial knowledge representation systems behave exactly in the opposite way (cf. Newell & Simon 1975). Each additional database entry increases the search space of the system. By analogy to human knowledge, processing time may be reduced to some degree by a hierarchical representation. This ability to taxonomize may be a universal disposition of human reasoning (cf. Johnson-Laird 1983). The hierarchical organization of libraries illustrates how search procedures may be improved
NLP: Minds, brains, and programmes
191
through an intelligent organization. This insight had been achieved even half a century ago, when the psychologist George Miller compared the hierarchical organization of human reasoning with a computer programme (cf. Miller et al. 1960). Within the hierarchy of concepts, speakers commonly neglect conceptual boundaries, in order to access basic level concepts. Concepts represented at this level are neatly graspable, in comparison to their more general superordinate and more specific subordinate concepts. For instance, if we hear someone talking about dogs, what comes immediately into mind is a typical instance of a dog, such as an Alsatian. This is what we experience most frequently, and what we can store in a fairly well delineated fashion in our semantic memory. By contrast, the hyperonym dog comprises too many different instances to be categorized homogeneously, whereas the hyponyms of an Alsatian, which, due to very specific instances, come in so low numbers that ordinary people have difficulties in discriminating them in their semantic memory. Thus speakers prefer basic level categories at a medium level of granularity, for which they may flexibly shift category boundaries up and down the hierarchy of concepts in order to avoid too coarse-grained or too fine-grained concepts, whenever they are referred to, but not available in their semantic memory. Applied to the speakers’ representation of metaphors and to our evaluation of the computer metaphor in particular, the psychological findings about basic level categorization bear very important implications. As metaphors are created through a similarity between source and target domain in order to gain knowledge about the latter, this analogy must likewise be achieved at a medium level of granularity (cf. Zelinsky-Wibbelt 1998: 316ff.; 2003: 22f.), where the meanings to be compared are neither too vague nor too specific, in order to achieve a successful reference to the target domain. Yet, the concept of “mind” is tremendously vague, in both scientific and common sense theories, as attested cross-linguistically by the very poor equivalences of the concepts of the corresponding terms (cf. Zelinsky-Wibbelt 2002b: 111; Deppert & Zelinsky-Wibbelt 2003: 15). Thus the question arises whether the vehicle of the von-Neumann computer is represented at the adequate level of granularity for speakers to felicitously construct a theory of mind. There is also flexibility in referring to different concepts of one and the same form. Depending on the frequency of usage which speakers have experienced, they have mentally stored different senses of a lexical unit in a hierarchical structure (cf. Simpson 1981). Thus the CONTAINER sense of cup may lexically dominate the CONTENT sense. This meaning dominance in turn explains why the CONTAINER sense may normally precede the CONTENT sense in lexical access. This hierarchically oriented bottom-up processing occurs if the context is neutral or lacking. However, if the context semantically related to the target word excludes competing senses right from the beginning, frequency effects are blocked and the contextually coherent sense is instantiated directly from knowledge about the whole text by top-down oriented processes. On the other hand, with a weakly biasing context, all senses be-
192
Cornelia Zelinsky-Wibbelt
come instantiated in parallel, that is, without conscious awareness, in a bottom-up oriented pre-lexical procedure (cf. Kintsch 1986; Simpson 1981: 134; Forster 1981: 204; Tanenhaus 1987: 217f.; Williams 1992: 196), such as the holistic CONTAINER – CONTENT image of cup. From this the contextually adequate sense is then selected to reach the level of consciousness by topdown oriented inferences (cf. Seidenberg 1982; Swinney 1979; Prather 1988). These findings illustrate how humans flexibly combine bottom-up and topdown oriented access procedures to lexical meaning. Today effectively organized expert systems may well be faster than human beings if large data structures are to be processed. In addition to the paradigmatic organization of knowledge, however, it is the organization of knowledge at the syntagmatic level, which a human expert is capable of making linguistically explicit more rapidly with increasing amount of knowledge. We will deal with this issue in the following section. 4.5. From implicit to explicit representations According to the opponents of strong AI human knowledge is typically subjective, i.e. implicit, as it is not fully conventionalized in many respects (cf. Schopmann & Shawky 1996: 71; Dreyfus 1972). The question arises whether we may predict, i.e. objectivize all explicit and implicit relations of human knowledge. A case in point are elliptical linguistic constructions created through ad hoc metonymies and metaphors, such as the following (Pustejovsky 1991: 425): (11) Let me finish the cigarette. The relation between the speakers’ knowledge stored in long-term memory and the information processed in working memory has given rise to the debate about how to draw the line between the conventional rules constituting the system of language and the inferences which are triggered by the situational and communicative context, between linguistic knowledge and world knowledge. This relation has become one of the key issues in which generative linguistic theories diverge from functional and cognitive linguistic theories. By rejecting arbitrary listings of senses in favour of grammatically motivated derivational rules the paradigm of the generative lexicon abandons redundancy on the one hand and purports to enable the creative interpretation of new and implicit senses on the other. In this way, a systematic treatment of polysemy pretends to solve many of the classic lexical-semantic problems. The overall idea is to associate polysemous lexical units with a basic representation in the lexicon in order to generate the potential of senses compositionally. This generation is achieved by “type coercion” resulting in “logical polysemy”. The generative concept of coercion is a deductive inference which relies on world knowledge implemented in the semantic structure
NLP: Minds, brains, and programmes
193
of nouns (cf. Pustejovsky 1991: 425). In (10) the noun cigarette does not refer to its lexicalized sense as a consumable product. Instead, it inherits its semantic representation from “true complement coercion”. The verb finish inherently predicates a BOUNDED event, an ACHIEVEMENT; thereby presupposing a potentially BOUNDED ACCOMPLISHMENT6 of the cigarette, from which it predicates the ACHIEVEMENT of the result. This information is coerced onto the NP argument in that the noun cigarette inherits the BOUNDED value of the ACCOMPLISHMENT from its own semantic structure, which we represent here in a simplified way: Cigarette [BOUNDED smoke (HUMAN AGENT, TABACCO OBJECT)] In this way, the general idea of the generative lexicon consists in the assumption of a limited number of indefeasible facts without specifying how these facts are assumed without overgeneration. This is left to pragmatics. From this view then, the notion of type coercion does not even touch on the complexity of lexical access and thus does not explain how speakers behave linguistically flexibly and creatively. In contrast to this non-redundant lexical representation, cognitive linguistics has put special emphasis on the rejection of the rule/list fallacy (cf. Langacker 1987: 29, 42). The representation of both rules and lists of items falling under a rule is claimed to follow from the speakers’ non-monotonic reasoning. As stated above, rules never apply in completely absolute terms, but are a matter of degree. Otherwise many phenomena, such as prototype effects in human categorization (cf. Rosch & Lloyd 1978), and language change could not be explained (cf. Bybee 2001). On this view, implicit representations cannot be made explicit by taking the semantic output as input to the pragmatic component. Instead, the reverse relation holds (cf. Levinson 2000). Universal pragmatic principles create the input to semantics and determine the induction of lexicalizing and grammaticalizing concepts. Taking a linguistic account of pragmatics seriously, Levinson proposes a third level of utterance type meanings in between sentence type and utterance token meanings (cf. Levinson 2000: 42). The structure, the content, and the processes of this intermediate pragmatic level of linguistic development become represented through the speakers’ nonmonotonic default reasoning. Speakers may perfectly process novel senses which are lexically impossible by relying on their knowledge about utterance type meanings (cf. Levinson 2000), as a more radical, i.e. less grammatically, but more encyclopedically based metonymy may illustrate: ————— 6
The verb to smoke, which would be a normal complement of the verb to finish in a corresponding non-elliptical construction, expresses such an ACCOMPLISHMENT, which is both EXTENDED through the time interval during which the consumption takes place, and BOUNDED by the ultimate consumption. The explicit designation of this concept would amount to the non-elliptical construction finish smoking the cigarette.
194
Cornelia Zelinsky-Wibbelt
(12) The beefsteak at table four is getting impatient. The verbal predicate of this sentence expresses an intentional activity which cannot be performed by a beefsteak: a beefsteak cannot get impatient. Our knowledge about the restaurant scenario, however, enables us to project the beefsteak, which is possessed by a customer, onto the customer herself, such that in this context the short-hand expression beefsteak is associated with the fully explicit one of “the customer who is eating the beefsteak”. Instead of hard-wiring the pragmatics in the lexicon, this account considers how humans are capable, for instance, of flexibly extending the identity of objects to that of their possessors. In contrast to the rigid mechanism of type coercion on the basis of introspectively provided semantic structures, humans are capable of defeasible reasoning by resolving conflicting information as far as possible. By defeasible reasoning humans may process incomplete information as involved in the ellipses of ad hoc meanings on the background of their knowledge. By the effort to communicate (cf. Bartlett 1932: 227; Grice 1975: 45ff.), co-operative listeners assume the most “impossible” utterances to be meaningful and try to endow them with sense as long as the opposite may not be proved. Introspection is thus clearly insufficient for achieving the relevant implicit representations, because introspection relies on the subjective intuition of the individual. Communication is the most efficient and natural way by which implicit representations can be made explicit (cf. Winograd & Flores 1986: 98f.). The subject’s intellectual growth consists in achieving coherent connections within their knowledge as well as in relation to the knowledge of other human beings. These connections result from the subject’s externalizing their knowledge in terms of observable behaviour in order to negotiate about this knowledge. In this way, implicit knowledge is continuously turned into explicit knowledge thereby fostering a decrease in subjectivity. Yet, the infinitude of communication prevents our knowledge from ever becoming completely explicit and objective. It seems unlikely that an artificial system might come into the possession of a communicative competence, such that it would be capable of this flexible and creative as well as infinite negotiation of implicit representations. Without this competence the allegedly semantic representations of artificial systems only dispose of syntactic relationships and therefore can never establish a true referential relationship to the external environment (cf. Searle 1992: 200; Johnson-Laird 1988: 54). The most compelling argument against artificial semantics has been brought forth by Searle in his effort on the relation between “Minds, Brains, and Programs” (1980). In this article he develops a thought experiment about a virtual cognitive system being locked in a socalled “Chinese room” without understanding any Chinese. In this room the putative cognitive system would be equipped with a huge number of Chinese symbols as well as with a brilliant computer programme for “communicating” in Chinese. Yet, as any relations considered as meaningful are purely formal
NLP: Minds, brains, and programmes
195
in nature, the alleged Chinese competence remains simulated in this experiment. We will elaborate on this issue in section 4.8. below. 4.6 Reference by mental models Another bottleneck in the acquisition of implicit knowledge which refutes artificial systems to possess and acquire meanings is the great amount of implicit knowledge that humans have about the relative sizes and shapes of an infinitude of all actual and prospective objects in their environment (cf. Hatim & Mason 1990), not to mention the theory- and culture-specific perspectives on their mental representations. In order to point out the difficulty of this endeavour, let us start with considering how children begin to acquire word meanings: by physically pointing to the object to which the particular expression refers, by ostensive reference (cf. Quine 1960). The child continues to transform this concrete activity of pointing into a mental representation of the perceived object (cf. Piaget 1970), i.e. it acquires the ability of object permanence by representing something independently of its presence. Within the acquisition of a mental representation, children run through a series of trial and error, in which their successive over- and under-extensions are corrected by their communicative partners, until they may eventually shape their concepts in accordance with their community. Yet, the acquisition of these object concepts is a relatively trivial objective. Let us, in a second step, consider how children eventually learn to embed these object concepts within the more abstract representations of complex situations. This complexity consists in the relations between all different kinds of objects and events. Accordingly, their representations are relational and relative, due to the infinitude of configurations between all actual and prospective objects. These representations can only be learned through schematic approximations to be enhanced and adapted to a new context. According to Piaget this mutual process of reference and categorization is highly interactive in that the child continuously assimilates new referential representations of external experiences under existing schemata, as long as the external information fits into the internal knowledge representation system. Otherwise the child has to resolve a conflict between knowledge and information. For this end, children also have to learn how to accommodate their schematic knowledge representations to their referential representations of their environment. In this way they reorganize their schemata in order to be capable of integrating newly experienced information. Humans are capable of inferring and evaluating their relational representations of the potential range of configurations through the construction of mental models. This construction comprises all actual and prospective objects and events. According to the theory of Mental Models speakers construe their semantic representations of an utterance in at least two ways: firstly, they build their propositional representations directly from explicitly uttered information. As this representation is often insufficient for evaluating the
196
Cornelia Zelinsky-Wibbelt
relevant meaning, speakers additionally build an analogous representation of the experienced situation (cf. Johnson-Laird 1983; Gentner 2003). Analogous representations create models of reality which do not primarily depend on the analogy between the objects, but on their configuration. Concrete pictures illustrate the relational analogy of models. A photograph normally represents a very fine-grained model of reality, in that the photographer preserves most of the properties of the perceived situation. By contrast, an abstract painting may be reduced to some basic relations, in order to engender the intended effect. In a similar vein, listeners seem to construe their mental models by reducing the analogous information to just the relevant relations, and by extending these only if they prove insufficient for evaluating their correspondence with the experienced situation. From their analogous representations listeners may draw a large variety of inferences about not explicitly mentioned information. An example of the insufficiency of a propositional representation would be the decoding of the sense of beefsteak in example (12) in the previous section. As we have shown, a mental model representing the spatial and temporal context in which this sentence is uttered enables the hearer to immediately infer the intended sense. Let us illustrate how mental models enable speakers to access implicit information with relational linguistic expressions. Speakers must represent the relational meanings which may be expressed by the German preposition über in a way which enables them to evaluate whether the corresponding situation translates into over or above. What is at issue here, is the translator’s achievement of the necessary degree of differentiation (cf. Hönig & Kußmaul 1991). Figures 7 to 12 demonstrate the necessity of mental models to provide the adequate degree of differentiation in order to select the correct surface expressions in English, that is, either over or above. (13)
Die Lampe hängt über dem Tisch. The lamp hangs over the table.
Figure 5: over(lamp, table)
Figure 6: over(A, B)
NLP: Minds, brains, and programmes
(14)
Der Hubschrauber schwebte über dem Haus. The helicopter was hovering over the house.
Figure 7: over(helicopter, house) (15)
197
Figure 8: over(A,B)
Der Picasso hängt über dem Miro. * The Picasso is over the Miro. The Picasso is above the Miro.
Figure 9: * over(Picasso, Miro)
Figure 10: over(A,B)
198
Cornelia Zelinsky-Wibbelt
(16)
Die Lampe hängt über dem Tisch. The lamp hangs over the table.
Figure 11: * over(lamp, table)
Figure 12: over(A,B)
Whereas the representations on the left-hand side have preserved some of the properties of the objects themselves as being relevant for capturing the necessary degree of differentiation, such as the SHAPE, the EXTENSION as well as the PART-WHOLE structure of the objects involved, the representations on the right-hand side have abstracted from all of these object properties, such that only the spatial relation between them is preserved. It is obvious that in order to achieve the necessary degree of differentiation for selecting the adequate English prepositions, the speakers’ construction of mental models would have to comprise some of the properties of the objects themselves. This degree of differentiation must enable them to infer that the helicopter may still be said to be hovering over the house according to the mental model assumed in figure 7, while the lamp may not be said to be hanging over the table on the basis of the mental model given in figure 11. Nor does the analogy represented in figure 12 provide the necessary degree of differentiation for rejecting the relation to be expressed by the preposition over. A concrete mental model as represented in figure 11 is needed for this evaluation. This dynamic and infinite processing of mental models is clearly not amenable to Levinson’s intermediate level of utterance-type meanings mediating between pragmatics and semantics, mentioned as a device for defeasible reasoning in the previous section (4.5.). Instead, the “flashlight metaphor” of cognitive linguistics would avoid the dualistic distinction between linguistic performance and competence (cf. Geeraerts 1993). According to this image, speakers construe linguistic meanings subjectively through their context-dependent interpretations, whereby sense becomes inherited from top to bottom, within the speakers’ continuity of reference (cf. Zelinsky-Wibbelt 2000: 34, 44, 128ff.; Zelinsky-Wibbelt 2002a: 45ff.; Boyd 1993: 505ff.; Pylyshyn 1993:
NLP: Minds, brains, and programmes
199
548ff.; Putnam 1977: 127; Beaugrande 1997: 78ff.)7. As a consequence, the validity of the computer metaphor can only be evaluated in terms of its contextually provided sense. This does not rely on truth or essence, but on the speakers’ relation between their conceptual knowledge representation and their referential representation within their mental model of the discourse situation (cf. Zelinsky-Wibbelt 2000: 123ff.; 2002a: 46f.). The deterministic processing device of the traditional von-Neumann computer definitely provides an insufficient model for explaining the speakers’ dynamic and infinite construction of mental models. Therefore the following section will explain how cognitive science tries to overcome this insufficiency through an alternative model. 4.7 The brain metaphor In order to explain the analogous representation of mental models by artificial systems, the original computer metaphor relying on the traditional architecture of a von-Neumann machine is overcome by the brain metaphor (Massaro 1991: 305). This metaphor does no longer rely on separate sub-systems which store and process the input strictly sequentially, but instead uses the image of a parallel-processing computer (McDermott 2001: 2). In terms of the image of a parallel distributed processing (PDP) system, the human brain is conceived of as connecting vast numbers of neurons through the transmission of activation potentials. This spread of activation through the network of neurons occurs largely in parallel (cf. Rogers & McClelland 2000; Neubert 1994: 416) and unconsciously, in contrast to conscious problem-solving activities enabled through sequential processing. Within this holistic methodology of connectionist systems a mental representation is modelled as being caused through sufficiently dense connections within the neural tissue of the brain. By the spread of activation, neurons within the network can acquire and lose activa—————
7
According to Putnam, semantics is typically a social science, in that the continuity of reference is achieved through the speakers’ social division of linguistic labour, in which the community of speakers successfully fixes the consistent reference of linguistic terms. Illustrated with respect to natural kind terms, such as water or zebras, Putnam has pointed out through his concept of metaphysical realism, that, although speakers may be successful in intending and agreeing on the very same substance in a possibly eternal chain of handing over the referent from speaker to speaker, the essence of meaning will always remain virtual, as an ultimate verification of the truth of scientific definitions will never be achieved. Much as in the psychological concept of a prototype, Putnam rejects the concept of intension in favour of the stereotype, “meanings just ain’t in the head” (Putnam 1977: 124). Applied to our objective of Strong AI’s functionalist view of the computational mind, this would imply, that an empirical verification of the metaphorical meaning will never be attained by the respective experts, although speakers’s may be very successful in constructing and communicating their naive theories in common sense discourse.
200
Cornelia Zelinsky-Wibbelt
tion potentials (cf. Hofstadter 1996: 205). Each transmission of energy between neurons is associated with an activation potential which can change as a function of strengthening and inhibition through the compatibility between the information of the transmitting and the receiving neuron. There is a threshold of connections above which the activation potentials reach the level of awareness. Not surprisingly, Minsky claims that the ability of connectionist systems to process new information by analogous representations is compatible to the human ability to learn (Minsky 2006). Above a certain threshold of connections within the network, parallel distributed processing systems can learn, much as outlined in the previous section with Piaget’s theory of learning, i.e. they are programmed in a way which enables them to generalize their knowledge incrementally whenever the processed information is sufficiently compatible with the knowledge of the system. The more general the knowledge, the more resistant it is to revision or decay. In this way, parallel distributed processing systems are claimed to provide a holistic methodology from which creative behaviour evolves and by which the infinitude of situations may be modelled (cf. Minsky 2005). Connectionist systems are assumed in particular to represent the attributes of objects and the relations between them and evaluate the correspondence of the resulting models to the experienced situation (Broadbent 1995, Rogers & McClelland 2000). The human ability to discriminate highly detailed and complex patterns is also attributed to the high-dimensional representation of neural networks. Thus real situations may be represented by analogous models. These multi-dimensional connections are claimed to be perfectly capable of coping with imperfect information. Within this materialist reduction, the mind is taken to categorize imperfect information in relation to prototypical knowledge representations by way of abductive “inference to its best understanding” (cf. Churchland 1995). Nevertheless, this claim cannot be maintained without explaining why a particular mental model is chosen over competing ones in a particular situation. Without the capacity by which humans evaluate their assumptions as being sufficiently similar with the environment and thus as being likely to function successfully in this, this mechanism would amount to an infinite regress and not establish an intentional referential representation (cf. JohnsonLaird 1983: 470). As yet, the adequate degree by which speakers differentiate between their mental models in order to choose the adequate one, as well as their subsequent translation into propositional representations is still a matter of research. It is evident from the examples given above that different degrees of abstraction are necessary for different linguistic purposes. While this associative reasoning is cheap for human intelligence, it is very expensive for the computer. Even if we assume that connectionist systems are capable of learning how to tune their mental model to each possible discourse situation, it remains doubtful whether this is a sufficient criterion for the analogy to the modelling of the human mind as a self-referential system which is capable of evaluating the representations of its own referential relationships. This issue
NLP: Minds, brains, and programmes
201
will be the topic of the last two sections. 4.8. Recursive reasoning and the first-person perspective The intelligence of human reasoning and linguistic processing crucially consists in the self-reflective nature of the human mind. This nature resides in the ability by which we evaluate the internal representations of our referential relationships to the outside world, such as our knowledge about the meaning of beefsteak in example (12). This ability allows us to behave consciously (cf. Churchland 1995; Biery 1996: 64; Johnson-Laird 1983) by locating ourselves on a mentally higher level from which we may conceive of ourselves as intentional subjects which distinguish themselves from others by their objectives and motivations. Recursive reasoning allows human beings to reason defeasibly by revising a representation achieved earlier in the discourse through additional evidence obtained later. For this end, speakers need to compare their referential representations with their knowledge representations and negotiate them with those of other conscious beings, normally by means of language. Not surprisingly, the ability to reason recursively is taken to be a criterion for the compatibility of human intelligence with the behaviour of connectionist systems (cf. Minsky 1990; Churchland 1995; Dennett 1991; Dennett 2007: 87ff.). As one of the strongest opponents of strong AI, Searle turns against this identification of recursive behaviour and human consciousness, because recursive behaviour on its own is far too narrowly constrained for granting any genuine consciousness to an artificial system (Searle 2007: 102ff.). Recursiveness consists in the self-referential principle of meta-representation. This principle consists in establishing a representation of a representation, much like a computer translates the higher-order programme of PROLOG into successively lower-order representations, until it eventually arrives at the machine code, as consisting only of the symbols “1” and “0” (Searle 1992: 212ff.). Meta-representations may be evaluated as mental representations if and only if these representations are conceived of consciously by somebody as representations of something (Searle 2007: 114ff.; Metzinger 2008). Our subjectivity is rooted in this self-reflective experience of mental representations, whereby subjectivity itself seems to be the key to human consciousness (cf. Searle 2007: 99). In residing in the intentional subject, consciousness is crucially dependent on attitudes and affective behaviour. Without this “first-person perspective” (cf. Metzinger 2008: 219) the neuro-psychological balance of a person would be lost. Any motivations and affections which one may be inclined to attribute to artificial systems are simulated (cf. Weizenbaum 1984: 269; Johnson-Laird 1983: 471ff.). An artificial device of meta-representation will always come to an end, where a physical representation is no longer possible (cf. JohnsonLaird 1983: 470), as is the case with the machine code of the computer. At
202
Cornelia Zelinsky-Wibbelt
this point, the mechanical imitation of physical systems would have to turn into the conscious awareness of natural meta-representations as occurring in the mind of human beings. According to Johnson-Laird consciousness consists exactly in the ability to infinitely embed mental models within mental models. Turning against objectivism, Searle criticizes the claim of the simulation of attitudes and emotions by artificial systems by reference to ordinary language philosophy’s claim of the criterion (cf. Ryle 1949; Wittgenstein 1953; Quine 1960) of language use as pointing to the existence of internal mental phenomena (cf. Searle 2007: 102ff.; Bennett & Hacker 2007: 41). Minsky illustrates this mistake most clearly: “if you can tell me that you have experienced sweetness then, somehow, that sensation has caused your mouth to move!” (Minsky 2005). Searle rejects Wittgenstein’s conflation of the human mind with the speakers’ observable behaviour. In the same way as a particular illness, e.g. complete paralysis, might prevent a patient suffering from pain to communicate this, most of human attitudes, although being definitely intentional (cf. Searle 1983), go more or less unnoticed, apart from being discussed or disputed about now and then, and yet, they have an undeniable and direct impact on the inferences which speakers draw (cf. Johnson-Laird 1993: 25). Thus a behaviouristic account of recursiveness as a criterion for consciousness in artificial systems is equally inadequate for modelling of the human brain as a self-referential system by relying on a category mistake. 4.9. The embodied mind In explaining subjective consciousness as an emergent property of neural representations, Searle discards the recursive homunculus of the machine. In his view, consciousness is a product of the functions of our brain and the theoretical possibility of identifying the human brain with its mind must be rejected (cf. Searle 1992: 14). He points out that the monistic position defended by materialism can only claim an objective physical correspondence of consciousness through performing a reduction which presupposes a dualistic view in the first place. Searle’s strongly emergentist commitment logically denies the human mind to exist independently of the brain. He substitutes the computational theory of mind with his biological naturalism, which annihilates the deeply rooted Cartesian world view (cf. Bennett 2007: 65ff.), according to which different physical faculties located in different regions of the brain engender different forms of intelligent behaviour (cf. Searle 2007: 106ff.; Minsky 2006; Chomsky 2007; Jackendoff 1997; Bennett 2007: 65ff.). In Searle’s view, this mechanistic decomposition of the mind into components is replaced by one holistic conscious field, being produced by large connections of neurons, and integrating as well as differentiating different perceptual processes which modify the conscious disposition of the brain (cf. Dennett 2007: 87ff.).
NLP: Minds, brains, and programmes
203
Damasio has provided abundant neuropsychological evidence for the embodied mind (cf. Damasio 1994: 89ff.; Gibbs 2007: 151ff.; Lakoff 1999). Whereas human consciousness is a joint product of both brain and body states, it achieves a balance by lower-order brain functions controlling the higher-order ones and vice versa (cf. Damasio 1994: 83ff.; Panksepp 2007: 176). The very basis of the biological nature of human beings is implemented in the neural function of emotions, by which the Darwinian principle of natural selection through survival and reproduction is involved on the one hand (cf. Darwin 1859), and which monitors human drives and instincts and thereby achieves a homeostatic control of the higher-order functions of the brain on the other. There is also abundant linguistic evidence about the embodiedness of the human mind. Speakers like to compare their mental states with their physical origin, which, again leads us to metaphors. In order to understand and communicate their subjective experiences, speakers transform their internal representations into external behaviour. This transformation surfaces linguistically, for instance, in what cognitive linguistics refers to as bodily determined image-schemata (cf. Lakoff 1980; Johnson 1987; Damasio 1994; Damasio 2003). Image schemata originate in general spatial orientations of the human body and are often claimed to have universal or cross-cultural validity. For instance, due to the normally upright position of the human body, the vertical dimension of space is socially evaluated in favour of its upwards directed pole as POSITIVE, in contrast to its downwards directed pole being evaluated as NEGATIVE (cf. Hampe 2005; Tyler & Evans 2003), as the following examples may illustrate. (17) (18) (19) (20)
My spirits are rising. My spirits are low. The chances of survival are very low. Living standards are falling due to the recession.
By continuously fostering the emergence of novel semantic values, a small number of image-schemata enables very efficient reasoning procedures, triggered by concrete and short linguistic constructions. Yet, by the exclusive reliance on their physical origin, mental representations remain insufficient for being evaluated as being adequate models of the perceived situation. This comparison can only be achieved through the experiencers’ contact with their environment. A linguistic utterance is ultimately endowed with sense by the speaker’s system-internal representation of system-external information with the objective of integrating the subjective behaviour into the objective environment (cf. Biery 1996). As we have pointed out in the previous section, the subjects’ externalizing their internal knowledge in terms of observable behaviour is an unnecessary and insufficient criterion of consciousness. Moreover, section 4.2. has outlined that the
204
Cornelia Zelinsky-Wibbelt
very reason for speakers to externalize their subjectivity is to negotiate about this and thereby to objectivize it in an infinitude of approximations. 4.10. The environmental metaphor The conclusion of the previous section implies that mental representations result from communicative interactions between living systems which thereby gain knowledge about their environment in order to change accordingly (cf. Barsalou 2005). There is abundant evidence from neuro-psychological dysfunctions that living systems must not be separated from their environment on account of the functional unity of the two systems. This insight has given rise to the view of distributed artificial intelligence, in which several cognitive systems are cooperating. A system is autonomous to the degree that it is able to create concepts without relying on predetermined rules. This allegedly creative behaviour is involved in the system’s ability to solve problems. In solving the problems which the environment imposes on them, autonomous agents become situated systems. Such a situated agent disposes of sensors by means of which it can contact its environment. Through communicating their concepts, autonomous systems seem to be capable of mutual problem-solving which in turn surfaces in a much more effective achievement of their goals (cf. Müller 1997). Yet, this cooperation does neither conform to real problem-solving nor is it compatible with the way in which human subjects socially negotiate their knowledge. The cooperation between humans is not merely purpose-oriented but is intrinsically destined to transcend the interactions between the individuals within the group. As we have pointed out, the continuous negotiation, revision, and objectification of mental models is necessitated through speakers recognizing the subjectivity of their tremendously incomplete and infinite knowledge representations. Yet, the subjects’ emerging mental representations will never become fully objective. For one thing, this is a corollary of the continuously and infinitely changing environment. For another, the cognizing subject continues to rely on qualitatively different levels of consciousness not directly and necessarily translatable into each other, such as desires, attitudes, beliefs and emotions often being far from cool reason. By contrast, several cooperating agents will always depend on the same syntactic representations as their manipulations of data are performed in terms of compatible algorithms, ultimately reducible to the same set of computations. Even if we consider one of the most recent versions of distributed artificial intelligence, such as egocentric qualitative spatial knowledge representations (cf. RoboCup 2003), we have to conclude that, although being definitely subjective at the outset, the respective systems cannot be successful without being tailoured to achieve translation-invariant behaviour. If confronted with huge amounts of complex data, this translation-invariant behaviour may well render artificial systems superior to the human mind. Yet, in recognizing the
NLP: Minds, brains, and programmes
205
relevance of data in relation to a particular purpose or problem and in making predictions about completely new domains, it is precisely the emergence of higher-order from lower-order functions of the human brain, i.e. the nontranslatability of feelings, emotions, beliefs, and attitudes into knowledge which makes human learning so efficient. The most complex physical activation potentials as engendering the systems’ representations in accordance with their environment will never come equal to the emergence of different qualities. In fact, the term “autonomous agent” is rather misleading, in that the system’s learning from scratch cannot literally be considered as such. The systems’ acquisition of concepts in terms of prototypical knowledge representation in no way implies that the system “knows” anything by which it could evaluate the environment, as for instance in taking voluntary decisions, conveying interest satisfaction and approval, granting encouragements and orders. The system does not dispose of voluntary actions. Its putative creativity results from the manipulation of huge data sets without having induced any theory. Autonomous agents thus are incapable of adapting themselves to an unlimited number of environments. Thus these systems are only “situated” within a quite limited number of domains. By contrast, the creativity of humans enables them to perform predictions about completely new data sets on the basis of their theories through abductive inference, an uncertain hypothesis likely to be confirmed. Hypotheses become accepted as being in conformity with the theory, be they scientific of common sense, through communication. For one thing it is doubtful whether the stimulus-response behaviour of artificial systems contacting their environment through sensors may be considered as intelligent communication which goes beyond pure information processing. For another, it seems completely unlikely, that this scenario will ever enable such an agent to become a subject because its ability of metarepresentation has confirmed its own knowledge representation to be at odds. Given this bottleneck, such a system will never be capable of making hypotheses about another agent’s internal representations which it intends to communicate about. Thus autonomous agents will continue to remain mechanistic objects programmed to interact in a finite manner. 6.
Conclusions
In this paper we have dealt with the implications of the computer metaphor, the presupposed equivalence between minds, brains, and programmes. We have rejected the assumption of a computational mind as working analogously to the software of the computer because we found this dualistic analogy to be trapped in a scientific solipsistic regress. This allows neither for a computational account of the self-reflective behaviour involved in human meta-representation nor for its subjectivity. The disembodiedness of the computational mind in functionalist views logically excludes a consideration of
206
Cornelia Zelinsky-Wibbelt
the continuous linguistic manifestation of the embodiedness of the human mind and thereby prevents an empirically valid linguistic investigation of language use. As artificial knowledge-representation was found to be devoid of meanings and theories, the essential conditions of linguistic intelligence to recognize relevant and implicit meanings and to hypothesize about completely new data may not be transferred to artificial intelligence. This does not mean that there is no sense in modelling the human mind computationally. To return to our analogy between the flight of birds and airplanes: the human mind may well be a model for natural language processing as long as we remain aware of the insight that a computational system will never know the ultimate nature of human images: the flight of birds is a finite model, but human images are essentially infinite. This insight rejects the tenets of Strong AI and confines artificial systems to the limits of Weak AI. Yet, the complexities from which human intelligent behaviour emerges become more clearly delineated by using the computer metaphor as a didactic model in linguistic discourse. The subsequent construction and evaluation of mental models enables us to contribute to the uncoverage of the enigma of human intelligence. The computer metaphor thus has both an epistemological function with respect to the structure and processing of the human mind8 and a didactic function with respect to the building of natural language processing systems. As long as the speakers’ recursive reasoning enables them to monitor these functions of the computer metaphor self-reflectively and responsibly, they are kept from an unwarranted transfer of meaning. The Turing Test was lacking any specific assumptions constraining linguistic performance as a direct index of intelligent behaviour (cf. Churchland 1995). Today the neural production of human intelligence is no longer a black-box phenomenon, but may be inferred from observing electrochemical processes, as for instance by magnetic resonance imaging. This observation of internal processes in particular enables cognitive science to specify the behaviour of intelligent systems in contexts which are far less restricted than that of a primitive stimulus-response behaviour. By contrast, the Turing Test makes no assumptions whatever about what intelligence actually is. Yet, the Turing Test is still influential in determining the computer metaphor in computational linguistics and computer science. In conformity with psycholinguistic findings about lexical access a computational model of the human mind can neither rely on the exclusively modular architecture of rule-based, serially processing systems nor can it be in favour of exclusively data-based, parallel processing systems as adequate means for natural language processing. If we continue to take the human mind as a model, we need to assume hybrid systems, composed of different subsystems, in order to provide more adequate and efficient simulations of different linguistic utterances (cf. Johnson-Laird 1988: 191; Pfeifer & Scheier —————
8
As this paper was intended to point out, the human mind can be elucidated within the bounds of the computer metaphor.
NLP: Minds, brains, and programmes
207
2001). Finally, an adequate use of natural language processing systems would consist in supporting speakers in intelligent ways. A mechanistic substitution would be of no help. This paper has not mentioned any of the fatal consequences which are likely to result from this irresponsible endeavour. The reader is free to imagine the outcome on the basis of the arguments given in our discussion. This discards fully automatic natural language processing systems in favour of interactive systems, in which the human user herself may control the use of the system. References Barsalou, Lawrence: “Situated conceptualization”, in: Cohen, Henri, Claire Lefebvre (eds.): Handbook of Categorization in Cognitive Science (2005), 619–49. Bartlett, Frederic C.: Remembering: A Study in Experimental and Social Psychology. Cambridge: Cambridge University Press 1932. Bartsch, Renate: “Concept formation, memory, and understanding”, in: ZelinskyWibbelt, Cornelia (ed.): Relations between Language and Memory. Saarbrücker Beiträge zur Sprach- und Translationswissenschaft. Frankfurt/Main etc.: Lang 2009. Beaugrande, Robert-Alain de: New Foundations for a Science of Text and Discourse: Cognition, Communication, and the Freedom of Access to Knowledge and Society. Norwood, New Jersey: Ablex Publishing Corporation 1997. Bennett, Maxwell, Daniel Dennett, Peter Hacker: “Philosophical Foundations of Neuroscience”, in: Bennett, Maxwell, Daniel Dennett, Peter Hacker, Peter, John R. Searle (eds.): Neuroscience and Philosophy. Brain, Mind, and Language. New York: Columbia University Press 2007, 16–48. Bieri, Peter: “Was macht das Bewußtsein zu einem Rätsel?”, in: Metzinger, Thomas (ed.): Bewußtsein. Beiträge aus der Gegenwartsphilosophie. Paderborn: Schöningh 1996, 61–77. Boyd, Richard: “Metaphor and theory change: What is ‘metaphor’ a metaphor for?”, in: Ortony, Andrew (ed): Metaphor and Thought. Cambridge: Cambridge University Press 1993, 481–532. Broadbent, David E.: “The Maltese Cross: A new simplistic model of memory”, in: Behavioral and Brain Sciences, 7(152) (1995), 55–94. Bybee, Joan: “Introductio”, in: Bybee, Joan, Paul J. Hopper (eds.): Frequency and the Emergence of Linguistic Structure. Amsterdam: John Benjamins 2001, 1–25. Chomsky, Noam: Syntactic Structures. New York: Harper and Row 1957. – “Chatting with Noam Chomsky”, in: Cohen, Henri, Brigitte Stemmer (eds.): Consciousness and Cognition. Fragments of Mind and Brain. Churchland, Patricia Smith: The Engine of Reason, The Seat of the Soul: A Philosophical Journey into the Brain. Cambridge: MIT Press 1995. Damasio, Antonio R.: Descartes’ Error. Emotion, Reason and the Human Brain. New York: Avon Books 1994. Damasio, Antonio R.: Looking for Spinoza/joy, sorrow, and the feeling brain. New York: Harcourt 2003. Darwin, Charles: The Origin of Species. London: John Murry 1859.
208
Cornelia Zelinsky-Wibbelt
Dennett, Daniel: Consciousness Explained. New York: Penguin 1991. – “Comment on Bennett and Hacker, Philosophical Foundations of Neuroscience”, in: Bennett, Maxwell, Daniel Dennett, Peter Hacker, John R. Searle (eds.): Neuroscience and Philosophy. Brain, Mind, and Language. New York. Columbia University Press 2007, 73–95 Deppert, Alex, Cornelia Zelinsky-Wibbelt: “Sprache über Sprache – Metaphern in Linguistik und Laienlinguistik”, in: Fachsprache. International Journal of LSP 25, Heft 1-2 (2003), 14–34. Dreyfus, Herbert: “From Micro-Worlds to Knowlege Representation: AI at an impasse”, in: Dreyfus, Herbert (ed.): What Computers Can’t Do. New York: Harper and Row 1972. Durkin, Kevin, Jocelyn Manning: “Polysemy and the subjective lexicon: Semantic relatedness and the salience of intraword senses”, in: Journal of Psycholinguistic Research 18(6) (1989), 577–612. Eysenck, Michael, Mark T. Keane: Cognitive Psychology. A Student’s Handbook. Hillsdale: Lawrence Erlbaum 2005. Fass, Dan F.: Processing Metonymy and Metaphor. London: Ablex Publishing Corporation 1997. Forster, Karl I.: “Frequency blocking and lexical access: One mental lexicon or two?”, in: Journal of Verbal Learning and Verbal Behaviour 20 (1981), 190–203. Frege, Gottlob: “Über Begriff und Gegenstand”, in: Vierteljahresschrift für wissenschaftliche Philosophie (1892), 192–205. Garnham, Allan and Cowles: “Mental models and noun phrase anaphora”, in: Zelinsky-Wibbelt, Cornelia (ed.): Relations between Language and Memory. Saarbrücker Beiträge zur Sprach- und Translationswissenschaft. Lang: Frankfurt/ Main etc. 2009 (to appear). Gentner, Dedre: “Why we’re so smart”, in: Gentner, Dedre, Susan Goldin-Meadow (eds.): Language in Mind. Advances in the Study of Language and Thought. Cambridge: MIT Press 2003, 195–235. Gibbs, Raimond W.: The Poetics of Mind. Cambridge: Cambridge University Press 1994. – “Where’s the missing body? A puzzle for cognitive science”, in: Cohen, Henri, Brigitte Stemmer (eds.): Consciousness and Cognition. Fragments of Mind and Brain. Amsterdam: Elsevier 2007, 149–60 Grice, H. Paul: “Logic and conversation”, in: Cole, Peter, J.L. Morgan (eds.): Syntax and Semantics 3: Speech Acts. New York: Academic Press 1975, 41–58. Grush, Rick, Patrica Smith Churchland: “Gaps in Penrose’s toilings”, in: Journal of Consciousness Studies 2 (1995), 10–29. Hampe, Beate: “When down is not bad and up is not good enough: A usage-based assessment of the plus-minus parameter in image-schema theory”, in: Cognitive Linguistics 16,1 (2005), 81–112. Harnard, Stevan: “Creativity: Method or magic?”, in: Cohen, Henri, Brigitte Stemmer (eds.): Consciousness and Cognition. Fragments of Mind and Brain. Amsterdam: Elsevier 2007, 127–38. Hatim, Basil, Ian Mason: Discourse and the Translator. London: Longman 1990. Hausser, Roland: Computation of Language. An Essay on Syntax, Semantics and Pragmatics in Natural Man–Machine Communication. Berlin–New York: Springer Verlag 1989. Hild, Manfred, Brigitte Stemmer: “Can evolution produce robouts?”, in: Cohen, Henri, Brigitte Stemmer (eds.): Consciousness and Cognition. Fragments of Mind and
NLP: Minds, brains, and programmes
209
Brain. Amsterdam: Elsevier 2007, 53–67. Hobbs, Jerry: Abductive models of language processing. Santa Cruz, Cal.: University of Santa Cruz 1991. Hönig, Hans G., Paul Kußmaul: Strategie der Übersetzung. Ein Lehr- und Arbeitsbuch. Tübingen: Narr Verlag 1991. Hofstadter, Douglas R.: “Analogy-making, fluid concepts, and brain mechanisms”, in: Clark, Andy, Peter J.R. Millican (eds.): Connectionism, Concepts, and Folk Psychology. The Legacy of Alan Turing. Vol. 2, chapter 11. Oxford: Clarendon Press 1996, 195–248. Hugh, G. Petrie, Rebecca S. Oshlag: “Metaphor and learning”, in: Ortony, Andrew (ed): Metaphor and Thought. Cambridge: Cambridge University Press 1993, 579– 609. Jackendoff, Ray: The Architecture of the Language Faculty. Cambridge, Mass.: MIT Press 1997. Johnson, Mark: The Body in the Mind: The Bodily Basis of Meaning, Imagination and Reason. Chicago–London: The University of Chicago Press 1987. Johnson-Laird, Philip N: Mental Models: Towards a Cognitive Science of Language, Inference and Consciousness. Cambridge: Cambridge University Press 1983. – The Computer and the Mind. Cambridge: Cambridge University Press 1988. – , Walter Kintsch: “The representation and the use of knowledge in discourse comprehension”. Technical Report 152, Institute of Cognitive Science University of Colorado 1986. Kintsch, Walter: Comprehension. A Paradigm for Cognition. Cambridge. Cambridge University Press 1998. Labov, William: “The boundaries of words and their meanings”, in: Bailey, CharlesJames, Roger W. Shuy (eds.): New Ways of Analyzing Variation in English. Washington, D.C.: Georgetown University Press 1973, 340–73. Lakoff, George: “The metaphorical structure of the human conceptual system”, in: Cognitive Science 4 (1980), 195–208. – , Mark Johnson: Philosophy in the Flesh. The Embodied Mind and its Challenge Western Thought. New York: Basic Books 1999. Langacker, Ronald: Foundations of Cognitive Grammar. Theoretical Prerequisites. Vol. 1. Stanford: Stanford University Press 1987. Levinson, Stephen C.: Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge, Mass.: The MIT-Press 2000. Massaro, D. W.: “Psychology as a cognitive science”, in: Psychological Science 5 (1991), 302–07. Mayer, Richard E.: “The instructive metaphor: Metaphoric aids to students’ understanding of science”, in: Ortony, Andrew (ed.): Metaphor and Thought. Cambridge: Cambridge. University Press 1993, 561–78 McClelland, James L., David E. Rumelhart: Parallel Distributed Processing. Explorations in the Microstructure of Cognition. Vol. 2: Psychological and Biolgical Models. Cambridge Mass.: The MIT Press 1986. McDermott, Drew: Mind and Mechanism. Cambridge: The MIT Press 2001. – , Jon Doyle: “Nonmonotonic logic”, in: Artificial Intelligence 1(13) (1980), 41–72. Metzinger, Thomas: “Empirical Perspectives from the self-model theory of subjectivity: a brief summary with examples”, in: Banerjee, Rahul, Bikas K. Chakrabarti (eds.): Models of Brain and Mind. Physical, Computational and Psychological Approaches. Amsterdam. Elsevier 2008, 215–45. Miller, George A., Eugene Galanter, Karl Pribram: Plans and the structure of behavior.
210
Cornelia Zelinsky-Wibbelt
New York: Holt, Rinehart and Winston 1960. Minsky, Marvin: “Symbolic vs. Connectionist”, in: Winston, Patrick (ed.): Artificial Intelligence. Expanding Frontiers. Cambridge: MIT Press 1990. – “Interior Grounding, Reflection, and Self-Consciousness”, in: Brain, Mind and Society, Proceedings of an International Conference on Brain, Mind and Society. Graduate School of Information Sciences, Brain, Mind and Society, Tohoku University, Japan, September 2005. – The Emotion Machine. New York: Simon and Schuster 2006. Müller, B. S.: Konzeptbildung und Kommunikation bei Robotern. Problemskizze und Vorschläge zum experimentellen Einstieg. Arbeitspapiere der GMD, 1080. (ftp directory for electronic publications and software produced at the SET institute at GMD) 1997. Neubert, Albrecht: “Competence in translation: A complex skill, how to study and how to teach it”, in: Snell-Hornby, M., F. Pöchhacker, K. Kaindl (eds.): Translation Studies. An Interdiscipline. Amsterdam–Philadelphia: John Benjamins 1994, 411– 20. Neumann, J. von: The General and Logical Theory of Automata. Remarks on the Foundation of Mathematics. Oxford: Oxford University Press 1951. Newell, Allen, Herbert A. Simon: “Computer science as an empirical inquiry: Symbols and search”, in: Communications of the ACM 19(3) (1976), 113–26. Panksepp, Jaak: “Affective Neuroscience and the ancestral sources of human feelings”, in: Cohen, Henri, Brigitte Stemmer (eds.): Consciousness and Cognition. Fragments of Mind and Brain. London: Elsevier 2007, 173–88. Penrose, Roger: Shadows of the Mind: A Search for the Missing Science of Consciousness. Oxford: Oxford University Press 1994. Pfeifer, Rolf, C. Scheier: Understanding Intelligence. Cambridge: MIT Press 2001. Piaget, Jean: Genetic Epistemology. Translated by E. Duckworth. Columbia: Columbia University Press 1970. Prather, P.A., David A. Swinney: “Lexical processing and ambiguity resolution: An autonomous process in an interactive box”, in: Small, S.L., G.W. Cottrell, M.K. Tanenhaus (eds.): Lexical Ambiguity Resolution: Perspectives from Psycholinguistics, Neuropsychology, and Artificial Intelligence. San Mateo, California: Morgan Kaufmann Publishers 1988, 289–310. Pustejovsky, James: “The generative lexicon”, in: Computational Linguistics 17(4) (1991), 409–41. Putnam, Hilary H.: “Meaning and reference”, in: Schwarz, Stephen P. (ed.): Naming, Necessity, and Natural Kinds. Ithaca–London: Cornell University Press 1977 119– 32. Pylyshyn, Zenon: “Metaphorical imprecision”, in: Ortony, Andrew (ed): Metaphor and Thought. Cambridge: Cambridge University Press 1993, 481–532. Quine, William van Orman: Word and Object. Cambridge: Cambridge University Press 1960. Reiter, Raymond: “A logic for default reasoning”, in: Artificial Intelligence (1980), 81–132. Röfer, Thomas: Forschungsprojekt SPP RoboCup – Automatische Plan- und Intentionserkennung fremder mobiler Roboter in kooperierenden und konkurrierenden dynamischen Umgebungen. Antrag für die zweite Projektphase. 2003. (http://www.informatik.uni-bremen.de/spprobocup/) Rogers, Timothy T., James L. McClelland: Semantic Cognition. A Parallel Distributed Processing Approach. Cambridge: MIT Press 2004.
NLP: Minds, brains, and programmes
211
Rosch, Eleanor, Barbara B. Lloyd: Cognition and Categorization. Hillsdale, N.J.: Erlbaum 1978. Ryle, Gilbert: The Concept of Mind. New York: Penguin Books 1949. Schoen, Lawrence M. “Semantic flexibility and core meaning”, in: Journal of Psycholinguistic Research 17(2) (1988), 113–23. Schopmann, Joop, Aziz Shawky: “Remarks on the impact of connenctionism on our thinking about concepts”, in: Clark, Andy, Peter J.R. Millican (eds.): Connectionism, Concepts, and Folk Psychology. The Legacy of Alan Turing. Volume 2, chapter 3. Oxford: Clarendon Press 1996, 67–74. Searle, John R.: “Minds, brains, and programs”, in: Behavioral and Brain Sciences, 3 (1980), 417–24. – Intentionality. An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press 1983. – The Rediscovery of the Mind. Cambridge: MIT Press 1992. – “Putting consciousness back in the brain. Reply to Bennett and Hacker, Philosophical Foundations of Neuroscience”, in: Bennett, Maxwell, Daniel Dennett, Peter Hacker, John R. Searle (eds.): Neuroscience and Philosophy. Brain, Mind, and Language. New York: Columbia University Press 2007, 97–124. Seidenberg, Mark S., Michael K. Tanenhaus, James M. Leiman, Marie Bienkowski: “Automatic access of the meaning of ambiguous words in context: Some limitations on knowlege-based processing”, in: Cognitive Psychology 14 (1982), 489–537. Simpson, Gregory B.: “Meaning dominance and semantic context in the process of lexical ambiguity resolution”, in: Journal of Verbal Learning and Verbal Behavior 20 (1981), 120–36. Sperber, Dan, Deirdre Wilson, Deidre: Relevance: Communication and Cognition. Oxford: Blackwell 1995. Strohner, Hans: Textverstehen. Kognitive und Kommunikative Grundlagen der Sprachverarbeitung. Opladen: Westdeutscher Verlag 1990. Swinney, David A.: “Lexical access during sentence comprehension: (re)considerations of context effects”, in: Journal of Verbal Learning and Verbal Behaviour, 18 (1979), 645–59. Tanenhaus, Michael K., Margery M. Lucas: “Context effects in lexical processing”, in: Cognition 25 (1987), 213–34. Turing, Alan: “Computing machinery and intelligence”, in: Mind 59 (1950), 433–60. Tyler, Andrea, Vyvyan Evans: The Semantics of English Prepositions. Spatial Scenes, Embodied Meaning and Cognition. Cambridge: Cambridge University Press 2003. Weizenbaum, Joseph: Computer Power and Human Reason. London: Penguin Books 1984. Wiener, Norbert: Cybernetics: Control and communication in the animal and the machine. New York: Wiley 1948. Williams, John N.: “Processing polysemous words in context: Evidence for interrelated meanings”, in: Journal of Psychology 21(3) (1992), 193–218. Winograd, Terry, Fernando Flores: Understanding Computers and Cognition. A New Foundation for Design. New York: Addison Wesley 1986. Wittgenstein, Ludwig: Philosophical Investigations. Oxford: Basil Blackwell 1953. Zelinsky-Wibbelt, Cornelia: “Metaphors as meaningful choices in scientific discourse”, in: Schulze, Rainer (ed.): Making Meaningful Choices in English. On Dimensions, Perspectives, Methodology and Evidence. Tübingen: Narr 1998. – Discourse and the Continuity of Reference. Representing Mental Categorization.
212
– – –
Cornelia Zelinsky-Wibbelt Berlin–New York: Mouton de Gruyter 2000. “Lexical vagueness, referential variation and lexicalization”, in: Behrens, Leila, Dietmar Zaefferer: The Lexicon in Focus. Competition and Convergence in Current Lexicology. Frankfurt/Main etc.: Lang 2002a, 43–74. “Die anglo-amerikanische Wissenschaftssprache der Kognitiven Linguistik und ihre Übersetzbarkeit ins Deutsche”, in: Fachsprache. International Journal of LSP 24, Heft 3-4 (2002b), 107–19. “Discourse and lexicalization”, in: Zelinsky-Wibbelt, Cornelia (ed.): Text, Context, Concepts. Berlin–New York: Mouton de Gruyter 2003.
Neal R. Norrick (Saarbrücken, D)
Pear-shaped and pint-sized: Comparative compounds, similes and truth
0. Introduction 1. Some history 2. The corpus 3. Comparison, simile, comparative compound 4. Semantic analysis of corpus examples 5. Context, ambiguity and truth 6. Comparative compounds in botanical texts 7. Conclusions 0. Introduction In this paper, I investigate comparative compounds like pear-shaped and pintsized in particular and similes more generally as to their meaning relations and their truth values. Compounds like pear-shaped and pint-sized are linguistically interesting because of their apparently obvious comparative meaning. And comparisons are themselves interesting for many reasons, especially in cognitive linguistics, since comparison is at the basis of categorization in revealing prototypes, as Lakoff (1987) has shown; thus we use stock similes like blue as the sky because the sky provides a clear example of blue for comparison, and we say lightning-fast because lightning is a prototype for speed; see Norrick (1986) on stock similes. Moreover, many similes are at least implicitly scalar in placing something high or low on a scale; something is “slow as a snail” or “fast as lightning” or even “fast as greased lightning”. Of course, this often moves into hyperbole as well, but still it reflects movement along an implicit scale. However, compounds of the skyblue and lightning-fast type are not frequent enough to get comparable results from a computer search, even with corpora of several million words – in spite of claims by Marchand (1969), Lipka (1966) and others that this pattern is very productive. Old claims of productivity were purely impressionistic. Today with computer corpora, we can finally make substantive, testable claims about relative productivity in different types of texts (see Moon 1998). My computer searches led me from Noun-Adjective compounds like sky-blue to compounds with participle forms like cream-colored and mint-flavored, and finally to compounds with the
214
Neal R. Norrick
participles sized and shaped, which not only have implicit comparative meaning, but are also common enough to count as truly productive and provide data on relative frequencies. Their frequency follows at least in part from their filling noticeable gaps in vocabulary: thus kidney-shaped and heart-shaped are necessary, because there are no adjectives such as kidney-ish or hearty with the intended senses, and the learned terms reniform and cordate are not in general usage. Once I had collected a fairly large corpus of examples of compounds with sized and shaped in their real contexts, two observations stuck out: first, the semantics of these compounds is not so transparently comparative as I thought, and second, they are quite diverse. Beyond the literal and straightforwardly comparative egg-sized seed, we find hyperbolic examples like thimble-sized glasses for wine, and examples like heart-shaped, which is not strictly literal, since it orients itself toward some stereotype heart shape familiar from valentines and playing cards, but quite unlike a real-life human or animal heart. Further, we find examples like man-sized pump, which does not literally compare the size of the pump to the size of a man, but says instead that the pump has dimensions appropriate to use by a man – indeed by some prototype man: hence combinations like man-sized tissue and man-size sandwich, signifying extra strength and/or size; and finally compounds in constructions like pint-sized Peter Craven, which apparently does not compare Peter’s size directly with a pint measure, but rather with the position a pint occupies on the scale of liquid measures, namely on the bottom end below quart and gallon. Compounds like pint-sized are therefore not hyperbolic as such, since they associate their object with a position on some pertinent scale rather than claiming the object is the same size as a physical pint container. 1. Some history Historically shaped and sized are past or passive participles, but over time they have become increasingly grammaticalized as suffixal elements or special restricted forms of the corresponding verbs. Webster’s dictionary calls them “combining forms”. Parallel to their ongoing grammaticalization, they have tended to lose the dental suffix -t or -d in speech. And as speakers stopped pronouncing the dental suffix, they have increasingly also spelled the forms without the -ed. At the same time, their semantics has spread from original comparative meaning to the variations mentioned above. The history of the combining form shaped is particularly interesting. The oldest occurrence of the participle form of shape listed in the Oxford English Dictionary (OED) is from Ælfric in 1000 AD: wlitiz gesceapen roughly ‘nicely shaped’. The first apparently comparative use listed is from 1423: like to one hertschapin verily. By 1776 we find the conservative modern form with a hyphen and the -ed suffix: heart-shaped leaf, where the meaning is clearly comparative ‘shaped like a heart’. And this comparative reading is routinely
Pear-shaped and pint-sized
215
routinely identified as basic today by scholars working in word-formation, for example by Marchand (1969) and Adams (1973); see also Norrick (1987a). But parallel to examples with proper comparative meaning, we find the very common form shipshape with an at best derivatively comparative meaning like ‘arranged properly, as things on board a ship should be’. The earliest citation in the OED is from the year 1664: “for to make her shipshapen, as they call it”, where the addition of the phrase as they call it indicates that the phrase was probably already in general usage. By 1769 the old -en participle ending was missing in the OED citation “not rigged ship-shape”. Interestingly, the word misshapen retains the historical participle ending -en even today. If shipshape was common as early as 1769 without the participle ending and with a meaning not clearly comparative, it may have provided a model for both the morphological and semantic development of the array of forms we find with the element shape(d) today. If we accept the admittedly spotty evidence in the OED, both the adverb plus shape use in wlitiz gescheapen and the non-comparative meaning of shipshape pre-date the comparative function of shape(d) as an element suffixed to nouns, so we should not be surprised at the frequency of non-comparative examples. 2. The corpus Nevertheless, for present purposes I would like to focus on those compounds with what initially seems to be straightforwardly comparative meaning. For compounds with shaped this means discarding examples like: shipshape, oddshaped, well-shaped; for compounds with sized such examples as: large-sized, middle-sized and oversized; for colored such compounds as: many colored and bright-colored; for scented such compounds as: heavy scented and foul scented. Still we find in any large corpus borderline cases such as pointed shaped with the apparent meaning ‘shaped sort of like a point’, which hedge on the comparison they propose. The combination grayish colored with a meaning like ‘colored more or less gray’ seems even less motivated, since both gray and grayish are already color adjectives with no need for the addition of the form colored to modify a noun. Another even more common compound is life-size, which corresponds to the phrase as big as life – itself only idiomatically comparative. These borderline forms were not counted in the statistics reported below. With this definition of what I was looking for, I investigated the three best known computerized corpora to see what kind of sample I could generate. In the Brown corpus of one million words of written American English, I found eight occurrences of compounds with shaped in contexts like “a V-shaped inlet”, eight for compounds with sized in contexts like “an egg-sized seed”, nine for colored as in “sherbet-colored cottages”, two for scented as in “manure scented lawns”, and one instance of flavored, namely in “cherry-flavored No-Cal”. In the London-Oslo-Bergen Corpus of one million words of written
216
Neal R. Norrick
British texts, I found sixteen shaped compounds, five with sized, seven with colored, and none with either flavored or scented. The London-Lund Corpus of spoken British English produced six examples of shaped compounds, four compounds with colored, and one example each with sized and scented. This adds up to 30 total examples with shaped, 14 with sized, 20 with colored, 3 with scented, and one with flavored. I concluded that at least the top three compounds, those with shaped, sized and colored, are frequent enough to be considered truly productive, and to yield interesting results for semantic analysis. But the detailed semantic description of concrete examples with their actual contexts of occurrence must wait till after some discussion of notions like salience, prototype and motivation in comparison within recent cognitive linguistic frameworks. 3. Comparison, simile, comparative compound For purposes of exposition it will be convenient to designate the explicit statement of a comparison as a simile. The first term of a complete simile, that is the thing to be described, can be called the OBJECT of the comparison which the simile expresses. Thus the ax is the object of the comparison expressed in the simile The ax is sharp as a razor in example (1). (1) The ax OBJECT
is
sharp as TERTIUM
a razor. VEHICLE
In this same simile razor is generally called the VEHICLE. And we will do well to follow traditional terminology in labeling sharp the TERTIUM COMPARATIONIS, that is the third term of the comparison. Even when the simile sharp as a razor is collapsed into the compound razor-sharp, we will still call razor the vehicle and sharp the tertium of the comparison, as in (2). (2) The ax OBJECT
is
razor VEHICLE
sharp. TERTIUM
In similes we generally expect the tertium to be a highly salient property of the vehicle. Seen the other way around, the vehicle should illustrate the tertium as a prototype, either from our everyday experience or as what Max Black called a culturally associated commonplace (cf. Black 1962: 40). For instance, slyness is an associated commonplace for foxes, even for those of us with no first-hand knowledge of them. Because of this relationship between the vehicle and the tertium, we can generally recover the appropriate tertium even when it remains unstated, as in Sue’s like a giant. For much the same reason, we may use recurrent similes as a key to discovering the associated commonplaces in a given culture; compare Norrick (1987b).
Pear-shaped and pint-sized
217
Now shape, color and size are cognitively salient for all objects (along with context and function, it seems), but the only shapes routinely used in similes are clear simple ones like the crescent, wedge and pear; still some of the commonest ones like heart, kidney, star and diamond are based not on any real-world experience of the vehicles in question, either because the realworld object is too complex and/or amorphous like the heart, the diamond and the star, or too rarely encountered like the kidney and again the heart for most of us. Instead, our images of hearts, kidneys, stars and diamonds are based on stylized pictures such as those on playing cards. Once the heart, diamond and so on are stylized into standard shapes, they become prototypes for the linguistic community at large. And of course different communities will use different objects and vehicles for a single comparison: one cannot simply translate diamond, hook it up with the equivalent of shaped, and suppose that it will designate the form we mean: thus German uses Karo for ‘square standing on a point’ rather than the morphologically fine diamantenförmig ‘diamond-shaped’; and it takes the sickle as its prototype for the crescent shape, and so on for other such vehicles of comparison. In parallel fashion, colors, especially those like coral, fawn and dove, are learned more from crayon boxes and pictures than from observation of real coral, fawns and doves. We might imagine that explicit comparison in a simile or a compound is sometimes a necessary condition of description, given the difficulty of describing shapes digitally, and the practical impossibility of explaining colors, tastes and smells. We have a fairly large set of color adjectives like red, yellow, green, and blue which describe our perceptions without explicit comparison, though orange, violet, indigo and so on are comparative – at least in their origins. To talk about tastes and smells at all we must compare, hence the predominance of phrases on wine bottles like: “This wine offers bright cherry and currant aromas” with “crisp acidity and a modest touch of French oak”, as well as implicitly comparative words like earthy, spicy and metallic. Certainly this fact helps to explain the frequency of compounds with flavored and scented. With size the situation is somewhat different: Although we can measure dimensions quite accurately with conventional scales, it would often be inconvenient if not impossible to do so. When a speaker in the London-Lund Corpus describes a dog as roughly Alsatian-size, she is probably producing not only the most useful sort of information for her hearers, but also probably the only sort of information readily available to her. And of course the comparison of one type of dog to another more familiar one is common sense cognitive psychology. Nevertheless, when the vehicle of comparison comes from another species, it usually introduces connotative resonance beyond the actual scalar information it conveys. A simile identifying a dog as pony-sized or a human child as teddy-sized is necessarily tinged with affect vis-à-vis simple measurements. In fact, any time the vehicle for a comparison comes from a category without obvious relevance for the object, the incongruency may have a metaphorical feel to it and may draw attention to itself, as Ortony
218
Neal R. Norrick
(1979a and 1979b) convincingly demonstrates. Thus, as Ortony further argues, similes themselves are metaphorical to varying degrees. Further, a tertium with low salience for the vehicle chosen for a comparison will signal that the motivation for the vehicle lies in its connotations. Consequently, low salience in the tertium represents another source of metaphoricity in similes, as noted by Ortony. The description of an apartmenthouse as a “barracks-shaped building” in the London-Oslo-Bergen Corpus seems curiously inappropriate, because the function of a barracks has far higher salience than its shape. Consequently, we tend to hear barracks-shaped in this context as suggesting certain living conditions rather than as a description of the shape proper. Similarly, color names, especially in fashion and decorating contexts, seem chosen for their assonance and associations at least as much as for the precise digital shade represented by the vehicle. Examples in my corpus include: terra-cotta, bread, oyster, dove and tomato. These names become particularly suspect when one investigates the range of hues in any given real-life oyster, loaf of bread or tomato, let alone from one loaf of bread, oyster and tomato to the next. Again the linguistic community – or at least the designers and fashion-columnists – presumably accepts color swatches rather than real objects as standards of comparison, just as stylized hearts and kidneys must be promulgated and accepted. Given these prototype effects of similes and their tendency toward metaphoricity along two avenues, it should have become clear that we risk circularity when we invoke the traditional description of metaphors as elliptical similes, just as we fail to explain the semantics of comparative compounds like pear-shaped and pint-sized if we simply relate them to the similes shaped like a pear and sized like/the same size as a pint. 4. Semantic analysis of corpus examples First of all, there are straightforwardly literal comparative compounds with the suffixal elements sized, shaped, colored and scented. In the London-Lund Corpus of spoken British English we find the compound alsatian-size applied literally to dogs, as discussed above. A speaker maintains apparently literally that the tail of the red grouse is “always drawn as lyre-shaped”. The stones of Saint Paul’s Cathedral are described literally as “pale honey-coloured” without obvious affectation. And a speaker refers seemingly literally to “the thyme-scented Athenian Midsummer night envisaged by Shakespeare.” In an interesting variant, a newspaper writes of a man who would like to have been chancellor, “Now he is being tempted by Beeching-sized offers to leave politics and go into business.” This seems literal in comparing multiple offers to one particular offer, namely that made by Beeching, which presumably represents some sort of stereotype corporation for intended readers. As long as we realize that comparisons often evoke stereotypes and culturally determined commonplaces, we can classify as literal the shaped and colored compounds
Pear-shaped and pint-sized
219
discussed in the previous section as well. Then “the kidney-shaped toepiece” of a new shoe may be understood as ‘a toepiece shaped like the stereotype designer kidney’; and “the coral-coloured boucle dress” may be understood as ‘a dress colored like stereotype fashion coral’ and so on. Presumably with any comparison it is up to the audience to determine suitable prototypes, just as they must recover the tertium missing in some comparisons, as pointed out above. Comparisons often run to hyperbole (see Brdar-Szabó & Brdar, this volume, 385–429), and our comparative compounds are no exception, particularly those involving sized. Beyond the compound thimble-sized applied to wine glasses, as mentioned above, we find in the Brown Corpus such clearly hyperbolic examples as “a box-sized building on Main Street” and “the pocket-sized company . . . with $2,170 in sales”. This last is problematic, since pocket-sized should strictly literally mean ‘as big as a pocket’ rather than ‘of a size appropriate for a pocket’; but it illustrates hyperbole either way. This latter reading of ‘appropriate for’ is reminiscent of shipshape and other similar compounds like man-sized ascribed to a pump, as in the example discussed above. When toy stuffed bears are described as “leading a domestic life in Teddy-sized houses” in a passage from the London-Oslo-Bergen Corpus, teddy-sized is not hyperbolic, since the houses are in fact built to scale for the bears; but neither is the comparison literally comparative, since the houses are not said to be the same size as teddy bears; again, the compound receives an interpretation on which the houses are of a size appropriate for teddy bears. This ‘appropriate for’ interpretation is so broad that it can even stretch to cover the complex compound thumb and finger shaped used to describe the side of a wooden abacus in the spoken London-Lund Corpus. Clearly this nonce compound does not intend to describe a grip visually like a thumb and finger, but rather like the impression of these digits holding the abacus frame, as if pressed into malleable material. We might gloss the compound meaning as ‘appropriate for holding with the thumb and finger’. The scalar meaning associated with the construction pint-sized Peter Craven in the introductory section illustrates yet another meaning possibility beyond the literal. Pint-sized occurs twice in my combined corpus and it seems to me familiar enough to count as a lexical unit meaning simply ‘small’. But another example in the corpus develops the same basic metaphor with an original vehicle, namely sparrow-sized Virginia Gibson. Here sparrow-sized apparently does not relate the size of Ms. Gibson directly with that of a sparrow, but rather with the position a sparrow occupies on the size-scale for birds. On such a scale, a robin might represent the norm and a hummingbird the very smallest entry with the sparrow about midway between the two. In compounds like sparrow-sized, we have more than just a straightforward comparison; we have a position on one scale, here that of humans, compared to a position on a second scale, namely that of birds, perhaps in relation to some norm or prototype like the robin, as in the figure below.
220
Neal R. Norrick
Figure 1: “sparrow-sized Virginia Gibson” Again, such compounds are therefore not strictly hyperbolic, though they seem so initially, since they assign their object to a position on some scale S1 (here: human size) related to a position on another scale S2 (here: bird size) rather than claiming the object occupies a position on scale S2. Of course, this interrelation of literally incompatible scales also results in metaphor similar to the kind Ortony associated with salience imbalance, as discussed above. Nor is this dual scalar interpretation limited to compounds with sized. When water is called crystal-clear, the claim can be directly tested by comparing the passage of light through the two; but when a written description is called crystalclear, separate sorts of clarity are compared: the verbal clarity of the description on scale S1 correlates with the physical translucency of crystal on scale S2, and metaphoricity again enters the picture. Two interesting literary examples come from Golding’s Pincher Martin: In the first a character is described as having spider-length legs, where the person’s leg-length is certainly not literally compared to the length of a spider, nor even to the length of a spider’s legs; instead it seems the relation between the character’s legs and his body is compared to the relation between a spider’s legs and body. Even more complex is a superficially similar description of the same character’s spiderthin figure. Finally, let’s have a look at a profoundly metaphorical construction by contrast with the foregoing examples, which count as non-literal but not metaphorical in the traditional literary sense, namely: she thought, and lay still, thinking of how she had first met Graham in Tangier – the hot sun, the white roofs, the charming things he had said. Now it seemed like some twopence-coloured fairy-tale. The collocation of twopence-colored with fairy tale counts as synesthesia in traditional terms, since it transfers color from one area of perception into another area where it does not fit. But do we work from the literal color of a two-penny piece, thereby claiming the fairy tale possessed a particular copper tone, with whatever resonance that generates? Or do we start with a figurative
Pear-shaped and pint-sized
221
sense like ‘cheap’ for twopence, combining that first with the notion of color, then predicate that product of the fairy tale to get a sense like ‘cheaply colored fairy tale’? Of course, the simultaneous availability of multiple clashing interpretations is what gives real metaphor its texture. Happily, it is not my purpose here to work out finished interpretations for individual metaphors like this, but to analyze the basic types of shaped, sized, colored compounds as to their truth-values in natural contexts, and it is to the final step in this process that I would now like to turn. 5. Context, ambiguity and truth In the right context, some sized compounds can be taken as literal truth claims. A statement like: “These dogs are alsatian-sized” can be tested, either approximately by casual inspection, more systematically by standing the dogs in question up beside some actual alsatian or exactly with scales and yard sticks. This holds as well for: “These seeds are egg-sized”. We could easily imagine a conversation in which someone maintained either of these claims, and someone present said, “No way, just look” or something similar. Of course, in other contexts the speaker of either sentence will be heard as exaggerating, depending obviously on the size of the dogs and seeds at issue. Nevertheless, in some everyday contexts, claims involving sized compounds can clearly have a truth-value; though certain statements with sized compounds will never be construed as literal truth claims, for instance when the vehicle of the comparison is fanciful or mythical as in: “The goat had unicorn-sized horns” or “Igor was giant-sized”. Since we cannot find any unicorns or giants for purposes of comparison, this sort of statement can count only as impressionistic. A sentence like “The little girl ate her father’s man-size sandwich” could then be considered true or false, if man-size literally means ‘appropriate to a (prototypical) man’ rather than ‘the size of a man’. We should note that in the right context even compounds like sparrowsized can take on a literal interpretation, as with: a sparrow-sized bluebird or become ambiguous between scalar and hyperbolic interpretations, as with a sparrow-sized hawk. We can even coax pint-sized back out of its comfortable lexical entry for re-interpretation within a context of liquid measures, for instance: “Judy added a pint-sized jar of honey to the mixture”. Like any other comparative statement, constructions with sized compounds have a potential for ambiguity, which makes any determination of their truth-value context dependent. In this way, they duplicate the sort of ambiguity we saw associated with compounds like crystal-clear in the preceding section, and they display a parallel metaphoricity as well. Moreover, beyond this metaphoricity, the dual scalar interpretation outlined above for sized compounds is fraught with ambiguities. Even if Virginia Gibson is not being directly compared to a sparrow in our example “sparrow-
222
Neal R. Norrick
sized Virginia Gibson”, but rather being placed on the human scale of size in a position like that the sparrow occupies on the size-scale for birds, we have only an implicit scale to work with. Does the scale reach from the hummingbird all the way to the ostrich? Or only from the wren to the eagle? If we could decide this, we might determine just where the sparrow ranges, then see if Virginia in fact occupies the corresponding position on the human scale, though, of course, the human scale is itself problematic. Are we counting just adults? Just women? Just adult English women? It may make more sense to see Virginia in relation to some average human, just as the sparrow may be compared to the average or prototype bird, say the robin. Then the statement would be true just in case Virginia relates to the human norm as the sparrow relates to the robin. This sort of comparison might initially seem simpler with the compound pint-sized in our earlier example pint-sized Peter Craven, since the pint occupies a clear place on the scale of liquid measures, say as in Figure 2 below.
Figure 2: “pint-sized Peter Craven” But do we necessarily see pint on a scale from cup to barrel, or rather from dram on up to hogshead? If the quart counts as our standard, then the claim is that Peter is half as large as the normal human; but if pint is compared to gallon as the standard, then seemingly Peter must be only one-eighth the size of the largest human for the statement to count as true. Thus while most of our comparative compounds can make truth claims subject to argument or even testing, it seems in general that statements involving them are to be taken as rough guidelines rather than true or false. Certainly we would not expect a challenge to the judgment of Peter Craven as pint-sized. Even more so, statements about colors will generally pass as impressionistically true without argument, as when someone says: “Allen has forget-me-not-color eyes”. Presumably no one would demand to sit down with Allen and a bouquet of forget-me-nots for a careful comparison, though, again, someone might make a counter claim to the effect that Allen’s eyes are really skyblue. In parallel fashion, shaped compounds will usually be taken as impressionistic claims. In the London-Lund Corpus a student says that holding “seminars is difficult in horse shaped rooms”. Maybe he was heard as saying
Pear-shaped and pint-sized
223
the rooms were horseshoe-shaped, which is somewhat easier to accept; but if he really meant to compare the shape of rooms with the shape of some presumably prototype horse, then clearly his statement cannot be taken as making a serious truth claim. Again, however, we can imagine a context where students looking at microscopic organisms argue over the applicability of compounds like column-shaped, rod-shaped, wedge-shaped and so on. Clearly, claims about shapes can be mutually contradictory, and any predicate like heart-shaped will be contingently true or false of any given object. Finally, compounds with scented and flavored are intended to make truth claims in certain contexts. Lemon-scented Pledge must either have real lemon juice in it or some chemical with the same olfactory effect in order for the makers of Pledge to stay out of legal trouble. And rum-flavored toffee had better have either rum or some perceived equivalent for buyers to accept it. Still in everyday contexts, statements involving compounds with scented and flavored will be heard as reporting impressions rather than scientifically testable truths. In summary, some sized compounds are taken as making literal truth claims in appropriate contexts. Others can be construed as claims about relations between scales, but still true or false, once we agree on the scales to be used. However, since the scales will often not be truly compatible, some degree of metaphoricity obtains. Among the claims about scalar relations, some will involve positions on some whole scale, while others just relate their objects to some prototype. Similarly, shaped, colored, scented and flavored compounds can make testable truth claims under certain contextual conditions. Nevertheless, these compounds, like all comparisons, tend to be taken impressionistically. 6. Comparative compounds in botanical texts We turn now to an exploration of shaped and sized compounds in scientific writing. In scientific discourse with its emphasis on correct description, we expect to find comparative compounds with literal readings making testable truth claims. This means shaped and sized compounds will make unambiguous comparisons with the forms they identify rather than the loose analogy we saw in the phrase barracks-shaped building or the claim of parallel scales we found for the phrase sparrow-sized Virginia Gibson. The two botanical texts I investigated in detail both exhibited certain characteristics which helped ensure literal comparative meanings for sized and shaped compounds. First of all, the university textbook The Diversity of Green Plants avoids sized compounds entirely, using only numerical measurements, while the more popular field guide Trees of North America used only a single sized compound, namely tree-sized, which modified only the nouns species and member (of a family of plants). Second, both books pair their descriptions
224
Neal R. Norrick
with drawings, diagrams and photographs, which amounts to offering ostensive definitions for shaped compounds via accompanying figures. Third, the meanings of individual compounds are effectively circumscribed in both books due to the impressive array of distinctions drawn, for example between spindle-shaped, cigar-shaped, barrel-shaped and so on. Curiously, both books sometimes alternate between compounds like heart-shaped and its latinate equivalent cordate. In Trees of North America the whole set of latinate terms are introduced in the early pages along with appropriate diagrams; and in the more technical Diversity of Green Plants a glossary at the end of the book contains definitions like those below. Falcate: Sickle-shaped Reniform: Kidney-shaped Obviously the authors assume that shaped compounds are so well delineated and generally understood that they can serve alone as complete and sufficient definitions even for scientific terminology. Still the prototype effects discussed above hold even for such seemingly non-comparative terms as reniform and falcate. Here again, the prototype heart provides only a rough template for matching approximate contours; and, of course, heart-shaped leaves like those on Linden or Basswood trees (Tilia) are shaped only roughly like the prototype heart. Our classification of plants depends of necessity on our recognition of recurrent basic leaf shapes, and both ultimately rely on subjective judgments and comparisons with conventionalized standards. I found 39 shaped compounds in The Diversity of Green Plants, a text of some 110,000 words, which amounts to about 300 such compounds for every 1 million words of text. Trees of North America, a guide which aims to help the reader identify individual species and sub-species of trees, contains even more comparative compounds. In this text of only 52,250 words, I identified 56 shaped compounds, which amounts to more than one such compound for every 1,000 words, and corresponds to over a thousand shaped compounds per million words. Thus the field guide Trees of North America has over three times as many compounds with shaped as the textbook Diversity of Green Plants. Comparing these results just with those of the two general corpora of written English, we recall that the London-Oslo-Bergen Corpus of written British English yielded only 14 comparative shaped compounds in its one million words; and the Brown Corpus of American English, again with a million words, produced only eight shaped compounds. Averaging the results for these two written corpora, we arrive at 22 shaped compounds in two million words or 11 compounds per million. And this is only about one hundredth the number of shaped compounds per million in the guide Trees of North America and only one thirtieth the number in the textbook. Clearly, compounds with shaped are necessary to the scientific vocabulary of plant description, especially for the fine distinctions involved in taxonomy.
Pear-shaped and pint-sized
225
Thus we see that even the language (and thinking) of natural science is rife with comparative judgments. Some of the comparisons are fairly explicit, for instance those in relatively rare compounds like barrel-shaped, while others have become largely implicit, insofar as words like heart-shape are no longer analyzed into their components. Ultimately, technical terms in scientific texts like falcate rely on definition via everyday comparative words like sickleshaped. We learn and understand chordate as ‘heart-shaped’, and we envision this concept with the help of a culturally propagated model of a stylized heart. The apparent scientific accuracy of chordate masks the underlying fact that our highly vaunted concept of scientific truth ultimately comes down to subjective human judgments of similarity founded on basic-level concepts and culturally salient prototypes, as Lakoff and Johnson (1999) argue. 7. Conclusions To summarize my main conclusions, some size compounds are understood as making literal truth claims in appropriate contexts. Others can be construed as claims about scalar relations, but still true or false, once we agree on the scales to be used. Of these, some will involve positions on some whole scale, or even on a pair of related scales, while others just relate their objects to some prototype. Similarly, shape, color, scent and flavor compounds can make testable truth claims under certain contextual conditions. Nevertheless, these compounds, like all comparisons, tend to be interpreted impressionistically. Taken together, these results demonstrate ways in which both everyday language and the language of science rely on comparative judgments based on experientially grounded concepts and metaphorical understanding. References Adams, Valerie: An Introduction to Modern English Word-formation. London 1973. Bell, Peter R., Christopher L.F. Woddcock: The Diversity of Green Plants. London 1983. Black, Max: Models and Metaphors. Ithaca 1962. Brockman, C. Frank: Trees of North America. New York 1979. Lakoff, George: Women, Fire, and Dangerous Things. Chicago 1987. – , Mark Johnson: Philosophy in the Flesh. New York: Basic Books 1999. Lipka, Leonhard: Die Wortbildungstypen WATERPROOF und GRASS-GREEN und ihre Entsprechungen im Deutschen. Tübingen: dissertation 1966. Marchand, Hans: The Categories and Types of Present-day English Word-formation. Munich 1969. Moon, Rosamund: Fixed Expressions and Idioms in English: a Corpus-based Approach. Oxford 1998. Norrick, Neal R.: “Stock similes”, in: Journal of Literary Semantics 15/1 (1986), 39– 52.
226 –
Neal R. Norrick
“Semantic aspects of comparative Noun-Adjective compounds”, in: Neuere Forschungen zur Wortbildung und Historiographie der Linguistik, ed. by Brigitte Asbach-Schnitker, Johannes Roggenhofer. Tübingen, 1987a, 145–54. – “Der Vergleich im Mittelenglischen unter besonderer Berücksichtigung der Towneley Plays (Comparison in Middle English with special reference to the Towneley Plays)”, in: Neuphilologische Mitteilungen 88 (1987b), 256–67. Ortony, Andrew: “Beyond literal simile”, in: Psychological Review 86 (1979a), 161– 80. – “The role of similarity in similes and metaphors”, in: Metaphor & Thought, ed. by Andrew Ortony. Cambridge 1979b, 186–201.
Paul Georg Meyer (Aachen, D)
“Money is ruthlessly finding its own level”: Metaphor and metonymy in verb semantics
0. 1. 2. 3. 4.
Preliminaries Sweetser’s framework Analysing transitive events: different sources of metaphor The polysemy of find and stumble: a case study How can money find its level? A sample analysis
0. Preliminaries The purpose of this paper is the following: Through an analysis of pertinent examples, including the one quoted in the title, I will try to show how metaphor and metonymy account for polysemy in verb semantics. It is argued that metaphor and metonymy do not shift the meaning of a verb in arbitrary directions, but follow certain regular patterns that can be found in a large number of polysemous words (Meyer 1997). It is suggested, with all due caution, that work along the lines proposed here may eventually lead to constraining the notion of polysemy and coming to clearer decisions on an age-old bone of contention in semantic studies: the borderline between polysemy and homonymy, which has long been recognised as diachronically flexible, but as notoriously elusive (cf. Nerlich et al. 2003). Furthermore, I will try to show that questions of truth arise in a very specific manner in sentences containing metaphorical and metonymic readings of the verbs studied in this chapter. 1. Sweetser’s framework These patterns of polysemy are partly derived from Sweetser’s (1990) framework which was also designed to account for patterns of polysemy. Sweetser demonstrated the usefulness of her framework in the analysis of subordinating conjunctions and modal verbs. Sweetser’s account establishes a tripartite pattern of meaning extension by which lexical items with a concrete basic meaning develop cognitive and speech-act meanings. E.g., because serves as a conjunction signalling real-world1 physical causation of events, as in –––––––––– 1
The question of the reality of ‘reality’ is not raised in this paper. What counts is that certain entities are imagined and represented as belonging to a physical world of which the speaker and hearer are imagined to form part. Even when truth conditions
228
Paul Georg Meyer
(1a) Because your young brother wants to see a stupid quiz programme you have to look as well. |SJ26|P622 But it may also be used in a speech-act related sense, as in (1b) You’re not telling me that all schools are perfect except Waterloo, because I know better than that. |SB23|P71 The because in (1b) does not indicate a cause or reason for the proposition expressed by the matrix clause, but the reason why the speech act performed in uttering this proposition was made in the first place. From these senses, a third kind of causal relationship must be distinguished: one in which the because indicates neither a cause or reason for the truth of a proposition nor a reason for a speech act, but a reason for a certain line of reasoning followed in the speech act: (1c) Then come articles about the possible ways in which mountain ranges were built up, and magnetic methods of testing the theory of continental drift. These are particularly stimulating because little can be taken for granted in sciences at so complex and unsettled a stage. |SC14|P16–17 This may often be difficult to distinguish from the last-mentioned case, but the distinction between speech acts and the thoughts underlying them is a difficult one anyway. I am not actually adopting Sweetser’s model here, but I am using her ontology as a heuristic tool. The model is useful because (quite in accordance with the classical semiotic triangle, cf. Lyons 1968: 404) a distinction is made between physical, cognitive, and verbal entities and events, that is, between the facts that are the referents of linguistic signs, the concepts and thoughts (called the signifiés in Saussure 1965: 99), mediating between these and the language system, and the linguistic acts, performed by using images acoustiques, or signifiants (ibd.), referring to these. Not only the referents, but also the other angles of the classical triangle are regularly made the object of conceptualisations and verbalisations. Thus language does not only need signifiants to refer to facts and physical entities, but also to thoughts and verbal ––––––––––
2
are discussed, we need not be committed to any notion of ‘reality’. All we have to do is discuss hypothetically what the world would have to be like, i.e., what facts would have to be part of it for the sentence in question to be true. For brevity’s and simplicity’s sake, I will use terms such as fact or reality naïvely, ignoring such epistemological qualms. Most examples in this article are taken from the LOB corpus of written British English. The numbering refers to section, text, and sentence numbers within that corpus.
Metaphor and metonymy in verb semantics
229
acts. These signifiants are often gained by extending the signifié and thereby the referent of expressions primarily referring to physical entities – a process that leeds to polysemy. Sweetser describes such processes of meaning extension of lexical items as the lexicalisation or grammaticalisation of more general processes taking place in parole. Her data are thus another confirmation of the thesis that processes of language change begin in parole and then congeal into langue by way of lexicalisation and grammaticalisation. This particular langue-parole relationship is also relevant to the kinds of polysemy to be studied here. In particular, I’m going to use Sweetser’s ontology to explain processes of metonymy and metaphor, the most powerful mechanisms of such extensions. To do this, we must add to Sweetser’s model further dimensions. This has to do with the fact that most of the data to be looked at here involve transitive verbs. Sweetser points out the possibilities of representing abstract cognitive or communicative processes by creating metaphorical extensions of a meaning representing a physical process. 2. Analysing transitive events: different sources of metaphor Conceptualising an event in terms of a transitive verb, as in (2a) He took up a clothes brush. |SK01|P15 (2b) He opened the door and went in. |SK05|P60 means that a distinction can be made between the event proper (taking up, opening), its agent (“He”), and its object (clothes brush, door). That is, in this conceptualisation not only agent and object can be clearly distinguished, which is not very surprising given the prototypical situation of transitivity (Hopper & Thompson 1980). What is more, an object can be separated from the process proper. Now suppose we use the meaning of a transitive verb as a metaphor vehicle to describe a more abstract process. The abstraction may affect the object or the process separately3 or both the object and the process, as seems to be the rule: (3a) She had been in mental hospitals since 1944 and the council first took up her case in 1957. |SA13|P112 (3b) The Premier was opening a two-day debate in the Commons on the common market. |SA06|P92 Metonymic processes mostly affect the agent, and in the agent domain, another three-way distinction must be made: The prototypical agent is human or –––––––––– 3
Examples will be given below.
230
Paul Georg Meyer
at least animate, as in (2) above. The agent in (3a) is already more abstract, but may still be regarded as (collectively) human. Apart from that, we find metonymic non-animate physical agents. The process and object denoted may be physical (4a) or abstract (4b): (4a) The bright little river dashes along through a glen which opens the very heart of the mountain. |SE09|P86 (4b) This is a new venture in the port and this first cargo will open yet another type of trade to Shoreham Harbour. |SA40|P69 Metonymic agents may also be more abstract, and it does not really affect the type of process referred to whether the ‘agent’ is cognitive (5a) or verbal (5b) in nature: (5a) They can not beat the camera at its own work and they can not improve on the work of the great realistic painters before them, so they go back along the route of painting of the past in the hope of finding some side-track branching off, which will open up into a royal road to new achievements and exciting discoveries. |SF34|P63 (5b) The word of the Lord that opens chapter 9 is spoken of as a ‘burden’, since it starts with solemn words of judgment on peoples that surrounded the land of Israel. |SD11|P61 It does not seem inconceivable that a process involving a cognitive or verbal entity as metonymic agent could also have a physical object, as is shown by the expressions (6a) opening ceremony (6b) opening verses. That is, each of the three ingredients of a transitive situation, event, agent, and object can vary along three dimensions. If all combinations of different kinds of agents, processes and objects exist, this will yield 3*3*3=27 different theoretically possible ‘senses’ of such a verb. This kind of variation seems to account for a large number of polysemies found in verbs. Not all combinations of such features can be found in all polysemous transitive verbs, some of them may not be found at all. The following examples, though not necessarily all the examples used, come from my own studies in the vocabulary of academic texts (Meyer 1993; 1996; 1997). I had presumed that this genre would be most rewarding as a hunting-ground for such examples because it is in the academic process that an intermeshing of physical, cognitive, and verbal4 activities can be observed at its unadulterated best. ––––––––––
4
The verbal dimension will not be discussed any further in this article as it only plays a marginal role in the verbs discussed here (cf. Meyer 1997: 163f.)
Metaphor and metonymy in verb semantics
231
3. The polysemy of find and stumble: a case study The verb find will be the focus of interest. Let us look at some pertinent examples from everyday discourse first: (7a) The internal passengers need a native guide and the gods on their side to find the booking counter allotted to them. |SB09|P47 (7b) That is why John Cassavetes came to England to find someone who would take a risk on something new. |SC17|P80 Examples (7) are close to the prototype of finding: Persons are purposefully looking for a physical object (in (7b), a person) with a certain set of properties that would satisfy their intentions. Physical actions are necessary to perform that task. Thus, all the three components of the event are physical and no metaphor or metonymy is involved. The verb find in this sense primarily highlights the moment at which the object to be found enters into the finders’ field of attention, which is conceived of as simultaneous with the moment of the finder’s achievement (in the sense of Vendler 1967) of becoming aware that the object in question actually satisfies their intentions to an acceptable degree. But examples like (7) also profile the process (usually described as looking-for) leading to the moment of finding; this is shown in the fact that find can be used in some contexts which are not typical of achievements and that it can even be used in the imperative. One cannot normally order people to bring about achievements, only to perform actions normally leading to such achievements, which are then, however, normally classed as accomplishments (ibd.).5 A less prototypical situation is described in (8), where, as in the previous examples, all the components of the event are physical, but the object had not been especially looked for: (8) When Christmas came we had a mass of unauthorised departures. A pale-faced corporal reported one night that his entire barrack room was deserted. He had found a packet of cigarettes on his pillow with a message attached ‘happy Christmas, Corp’ – and signed by all the missing men. |SG23|P96–98 I consider accidental finding as in (8) less prototypical because most of the metaphorical extensions of the meaning of find seem to be derived from purposeful finding, that is, from finding after looking. The difference between accidental and purpose-driven finding, by the way, does not constitute a polysemy. Like many agentive achievement verbs, find is simply vague in this respect: –––––––––– 5
For more detail on the semantics of find see Meyer (1997: 143ff.).
232
Paul Georg Meyer
(9a) Food was short. Water-cress and palmetto leaves were poor sustenance for men on the march, and even the maize they had found at Cale was a mean substitute for the gold they had looked for SF25|P34–35 (9b) Captain ‘Paddy’ Atley of the east Yorks found the barge Ena grounded where Lemon Webb’s flotilla had lain, took her back with forty men, on the strength of five sailing holidays in Norfolk. |SF23|P68 The author in (9a) explicitly states that something else had been looked for, but this does not exclude that the men had also been looking for food. The text is simply not committed to a statement as to whether the maize had been specifically looked for or not. In (9b), the wider context might clarify whether the barge found had been looked for or not, but this does not change the lexical meaning of find.6 The next example comes close to an academic kind of finding: (9) Sent to gaol for two years, he has quixotically refused to clear himself by betraying his colleagues, and, when he gets out, finds his grandfather has been driven to suicide by a bunch of crooks. |SC16|P70 Just like in academic achievements, the object of find is not a physical object, but a cognitive entity, a bit of knowledge, encoded in a finite clause embedded as a complement of the verb. But there is a difference for which the missing that is significant: There are few examples of find + finite complement clause without a complementiser that in academic texts. There is in fact only one such example in the relevant section of the LOB corpus, and no typical academic process of coming to know is involved here: (10) In Paradise Lost he often seems half ashamed of the autocratic behaviour of his Father, because his role is to induce the subject angels to endure it; but when he is alone on the earth-visit which has been arranged for him we find he has merely the cold calculating pride which we would expect from his training. |SJ61|P50 All typical instances of academic finding with a finite complement involve the complementiser that: (12) We have applied equations of this form to discuss the motion of electrons under external electric and magnetic fields and have found that –––––––––– 6
The exact scope of intention is peculiarly complicated with find, but this will not concern us here.
Metaphor and metonymy in verb semantics
233
this description leads to results in excellent agreement with experiment when the fields are not too strong. |SJ05|P37 My interpretation of this syntactic difference is that find followed by a finite causal complement without that indicates a more instantaneous finding, one that does not involve elaborate academic procedure as is explicitly described in (12). Omission of the complementiser would thus iconically signal the instantaneousness of the finding-process. Does this mean that a Sweetser-type polysemy is involved here? Can the academic finding-process be regarded as more abstract than everyday instances of finding? First it must be noted that finding itself is always a fairly abstract achievement. If you simply stumble on or over something in a physical sense: (13) There was the long, sustained whisper of oiled metal moving, and then Ivor stumbled over my feet . |SP24|P12 you do not “find” it. Interestingly, however, stumble has developed a metaphorical meaning right along the Sweetserian lines: (14) The police will undoubtedly think that you stumbled upon some fresh evidence which made it certain that Ray had murdered his ex-wife. |SP04|P125 How does such an occurrence of stumble come to mean “find accidentally”? A physical event is taken as the metaphor vehicle for conceptualising a cognitive achievement. But in what way is stumbling like finding? It is only a small range of elements from the concept of stumbling, and not even the central ones, that enter into the metaphor. The dictionaries implicitly distinguish three phases of a stumbling-on event: 1. prior to stumbling proper, the experiencer / theme is moving along on a path 2. the achievement of stumbling proper consists in the experiencer’s hitting their foot against some (typically small and inconspicuous) physical object, as a result of which 3. the experiencer / theme starts to fall. None of this seems to bear any resemblance to a process of finding or discovery. Moving along a path is not even a preferred metaphor for the research process (cf. Meyer 1997: 225f.). When a travel metaphor is used as in the few examples of arrive at being used to describe academic achievements (only two in the LOB corpus), the context is evaluative, and the evaluation tends to be negative (ibd.: 221ff.). Thus, using the stumbling metaphor to describe an academic achievement almost amounts to a self-denigrating strategy.
234
Paul Georg Meyer
The stumbling event proper is not really represented in the metaphor tenor, it is only the surprising and accidental appearance of a certain phenomenon that suddenly demands attention that could be adduced as a common denominator. While it is true that as a result of stumbling, the experiencer usually becomes aware of the existence of the ‘stumbling-block’, this increase in awareness is not really highlighted in the usual use of stumble; on the contrary, a non-metaphorical ‘stumbling-block’7 is regarded as rather insignificant in descriptions of stumbling events: The ‘stumbling-block’ does not occur in an obligatory thematic role, but only as a fairly marginal prepositional adjunct that is not really governed by the verb even allowing a considerable and wellmotivated range of different prepositions (on, over, across, upon). As far as the result, the third phase of literal stumbling, is concerned, falling can in no way be related to a gain in knowledge. The common determinator is again rather abstract, in that the stumbling somehow disrupts the process of moving along in the same way, it is to be expected that if one stumbles upon a fact, this will somehow mess up one’s normal course of investigation. But that does not seem to be highlighted in that description. It is probably the topos of modesty that leads scholars to use stumble as a verb for coming to know8 and readers apply an R-based implicature (Horn 1984) to interpret it. If you stumble upon something, you certainly come to know about its existence. And you may wonder what it was. And then you find out what it was. That is, strictly speaking, the moment of stumbling on a fact is not identical to the moment of finding a fact. Finding a fact implies full consciousness of the identity of the fact found. This is in contrast to finding an object whose precise identity you may not be aware of: (15) Jones found the oldest extant papyrus fragments of the New Testament. does not entail that Jones could sincerely have said (15’) “I found the oldest extant papyrus fragments of the New Testament”. (15) covers a situation in which Jones is not necessarily aware of what he found; that is, it is possible that he himself could not have sincerely described his find in such words. Suppose it was only decades after his death that somebody discovered that the papyri displayed text fragments from the New Testament and that they were older than had hitherto been thought. So, although (15) is acceptable and could be true, the same statement rephrased as a report on finding a fact could be false: –––––––––– 7 8
Note that this use of the term ‘stumbling-block’ in a non-metaphorical sense is itself metaphorical! In the relevant section of the LOB corpus, it does not occur at all.
Metaphor and metonymy in verb semantics
235
(15’’) Jones found that the New Testament is actually much older than had hitherto been thought. Finding an object thus seems to be somewhat ‘less cognitive’ then finding a fact. That’s why it can also be ascribed to animals. To take another nice example: I remember taking part in a guided walk somewhere in Shropshire, looking for the site of an old riverbed. A dog had been running along with us for some time, disappearing occasionally on some trail and coming back again. One of the times the dog returned to us he was all wet. Somebody said: (16)
“He’s found the old riverbed”
(that we had been looking for all afternoon). The dog, of course, didn’t know he had found the old riverbed, but it is a cognitive achievement that we could ascribe to him. Now suppose we turn the object of this discovery into a fact: (16’) Jones found that the river Severn near Shrewsbury once flowed in a river-bed 500 yards away from the present one. (16’’) *Jones’ dog found that the river Severn near Shrewsbury once flowed in a river-bed 500 yards away from the present one. This is evidence that with find, the difference between different senses is triggered by different objects. A clear distinction can be made between physical objects, which may be found by anybody and even by animals, irrespective of their description, and facts, which can only be “found” by people who can appreciate the full sense of the description chosen for the fact. Returning to the case of stumble, the relevant difference is between events and not between objects. The ‘literal’ and the ‘metaphorical’ event of stumbling resemble each other very little, and we may clearly distinguish a physical event of stumbling in which the experiencer / theme’s body actually gets involved in a characteristic stumbling movement, and a cognitive event of stumbling on something. In (14) above, as we said, we find an achievement in which the stumbling is reinterpreted by way of an implicature as an accidental finding, a situation which is then used to describe processes of accidentally finding abstract cognitive objects. That is, stumbling over a physical object does not entail finding it, stumbling over an abstract object necessarily does. As soon as the objects become abstract it is no longer possible to distinguish a cognitive from a physical achievement. You cannot physically stumble over an abstract fact. But as soon as a cognitive achievement is involved you cannot stumble over a fact without finding it. Thus find is another adequate conceptualisation of the situation, although accidental finding is less prototypical than findingfollowing-looking-for. A contrast may thus be construed between (17a) I stumbled on some fascinating new data.
236
Paul Georg Meyer
and (17b) I found some fascinating new data. (b) seems to imply deliberate, goal-directed activity prior to the achievement of finding, whereas (a) highlights the accidental character of the achievement. It is, however, (a) which is marked, and it could be paraphrased using find (17’) I accidentally found some fascinating new data. That is, the extension of find includes the accidental case, the stumbling-on case, but not vice versa: Metaphorical stumbling-on is only a special case of finding. Note that the two only compete where cognitive achievements are concerned. Concerning physical objects stumbling on something may or may not be altogether different from finding something: (18a) On his first day in Klondyke, young Scrooge McDuck stumbled on a goldmine. (18b) On his first day in Klondyke, young Scrooge McDuck found a goldmine. Here the truth conditions overlap considerably; most probably the truth conditions of (a) are included in the truth conditions of (b). Compare this with: (19a) He stumbled over a pebble on the pavement and hurt himself. (19b) ?He found a pebble on the pavement and hurt himself. (20a) He stumbled over a pebble on the pavement, but didn’t pay further attention to it. (20b) ?He found a pebble on the pavement, but didn’t pay further attention to it. The (b) examples are distinctly odd, but even if they are acceptable, their truth conditions do not intersect with the truth conditions of the (a) examples. That is, if the process itself changes its character, metaphorisation of the meaning not only changes the truth conditions drastically; it also changes the relation between the truth conditions of paraphrases. Stumbling on a pebble is an altogether different process from stumbling on a goldmine, not only in terms of financial consequences: One is a physical, the other a cognitive achievement. Finding a pebble and finding a goldmine are more or less the same kind of achievement, because both are cognitive. Comparison of find and stumble on has thus provided us with evidence of a separation of event and object:
Metaphor and metonymy in verb semantics
237
Whereas the achievement of finding is always cognitive9, its object may be physical or cognitive. On the other hand, you may stumble physically on a physical object only, whereas cognitive stumbling, like finding, can have both a physical or an abstract object. What have we learnt about polysemy and metaphor so far? In a comparison of stumble on and find, we saw that the two are quasi-synonyms in their metaphorical meanings only. These metaphorical meanings seem to arise from the same kind of meaning shift from physical to cognitive, but closer inspection showed that in the case of stumble on, a real difference in the kinds of processes is created, with disjunct sets of truth conditions, whereas in the case of find the difference lies in the kind of object found. We have thus looked at metaphorisation in the domains of object and event. 4. How can money find its level? A sample analysis Let us now look at the agent domain. find occurs with a number of nonanimate pseudo-agents. I will discuss only two examples containing a phrasal use of this verb which has interesting ideological implications, too. The phrase find one’s own level occurs twice in the LOB corpus, and both occurrences are in somewhat ticklish and delicate socio-political contexts (immigration and housing): (22a) The reality of the world to-day is of unequal economic development, with the richer countries growing richer and the poorer being forced to export their unemployed. Within the commonwealth all other countries control immigration. The West Indian islands even discriminate against one another. The older dominions (especially Australia) discriminate against non-whites. The United States and Latin America have also recently tightened up their immigration controls. This is no reason for Britain to behave likewise, but it does raise the problem of what will happen if Britain remains the only uncontrolled area into which the spill-over can go. Is it right to assume that the volume of this spill-over should be allowed to find its own level without any attempt at planning? |SB15|P38–43 (22b) Thus, the demographic changes induced by the double decline in births and deaths are linked to an increasingly rapid change in the composition of the parish population. Money is ruthlessly finding its own level in housing, and as the wave of wealth sweeps from the ––––––––––
9
The thoroughly cognitive character of find is also emphasised by the fact that it is the one remaining major feature in its attitudinal reading such as in (21) ‘Do ye not find the whisky in London terrible?’ |SA19|P57 where its achievement character is completely lost. This reading will not be discussed in the present article, but it can ultimately be explained by the mechanisms described here (see Meyer 1997: 160ff.).
238
Paul Georg Meyer
Gran Vía to trickle away into insignificance in the poorer areas of Pez, so those who can not enter the economic swim are driven farther away from the centre of the city and their traditional parish. |SJ27|P24–25 Reconstructing the metaphorisation process leading to an interpretation of (22a) and (22b), it is necessary to distinguish several levels. Starting-point for the use of the phrase in both examples is the cliché (22c) Water finds its own level (22d) Water will always find its own level which is quoted s.v. find in some dictionaries. It seems to provide a common metaphor vehicle for speakers wanting to point out the working of some law of nature. First it should be pointed out that already the ‘literal’ meaning of this phrase is metaphorical. The conceptualisation of water as a pseudo-agent is not metonymic, but metaphorical: Water is likened to a living being that can find something. This is also emphasised by the use of the reflexive pronoun own which is rare with inanimate antecedents. The finding process proper is also far removed from prototypical finding: It neither implies having tried to find, nor is it accidental. Such uses of find are not uncommon; dictionaries list quite a number of them, with varying classifications: Similar phrases are find one’s way and find one’s place, which can be used with personal or metaphorical agents or find one’s mark, as in (23)
The knife found its mark
which is best described as an animistic metonymy by which the knifethrower’s intentions are transferred to the knife. The object of finding is also more abstract than prototypical objects of finding. A level is abstract even in the literal reading of the cliché. It is an imaginary mark that in the case of liquids may be measured by the distance of the surface from the bottom. But the level is not conceptually identical to the surface. This constitutes the first level of metaphorisation leading to an interpretation of (22a) and (22b), in a schematic representation: (24) ( and symbolise metaphorisation processes)
AGENT water
finds finds
LOCATION level
Metaphor and metonymy in verb semantics
239
The cliché seems to state a law of nature by which communicating columns of liquids, under the influence of gravitation, will tend to form surfaces at the same level relative to the centre of gravitation, depending on the shape of the available container and the quantity of the liquid. There’s nothing wrong with stating this law of physics; it becomes, however,problematic if laws of nature are used as metaphor vehicles in talking about social phenomena. This constitutes the second level of metaphorisation in our examples. The cliché is modified by using a second metaphor: Water is replaced by, that is, compared with, immigration in (a), and money in (b): (25)
AGENT water immigration/ money
finds finds
LOCATION level
finds
level
find and level in the last line are marked as metaphorised because their use is more abstract than in the level above. The water level, although an abstract entity, is conceptualised in a real space. A level of immigration or a money level is one step further away, involving yet another metaphorisation. The metaphor theme used is the well known “More is up” (Lakoff & Johnson 1980): The more there is of something, the higher the level. All these metaphor themes are well known, and problematic in themselves. In (21a) it is clear that the ‘Immigration is a flood’ metaphor theme structures large parts of the conceptual space constructed in the quoted text.10 Immigration is not only liable to “find its own level” like water, but it can also “spill over” and has a “volume”. ‘Money is a liquid’ is also a common metaphor theme that may structure considerable parts of the discourse about money. In the LOB corpus, money can flow (|SH27|P44), be frozen (|SA06|P125), poured (even down the drain, |SB13|P132) and directed into channels (|SE35|P4). In what way is immigration like a flood, and money like a liquid? Asking these questions reveals the ideological underpinnings of these metaphors. By conceptualising immigration as a flood,11 authors claim that immigration causes severe damage and should be controlled by all means, but also that –––––––––– 10 11
A phenomenon already observed with other metaphor themes in Lakoff & Johnson (1980). To do justice to the author of the passage quoted, it should be noted that the word flood is not used in this passage. But spill-over conjures up quite a similar image and is certainly a variation on the same theme. However, the flood motif can safely be assumed to be deeply entrenched in anti-immigration discourse and actually occurs in a different text fragment in the corpus:
240
Paul Georg Meyer
it is difficult to control so that tough measures are necessary. The metaphor also disguises the fact that immigration results from people’s decisions and consists of living human beings. You cannot argue with a flood, and a flood is not owed any compassion.12 The metaphor thus dehumanises immigrants. In terms of thematic roles, immigrants are merely a ‘theme’ that moves from one country to another. They are denied any qualifications for agent or experiencer status. In our passage, the use of the verb find does provide some indication that the “spill-over” mentioned is like a living creature. But this primarily applies to the first level of the metaphor (see above diagrams). And what is worse: Spill-over is not countable. Thus the tenor of the whole passage suggests that we are not dealing with human creatures, but rather with a malevolent monster that is not even individualisable, but an infinite amorphous mass of threatening living matter.13 By conceptualising money as a liquid, an even more complex process is involved. On the surface, money is conceptualised in terms of water. And in this case the agenthood triggered off by the use of find finds more resonance in the metaphor tenor. Money is primarily like water because both are fake agents. Their use in agent positions is both metonymic and metaphorical. Both can be used to get work done. This alone qualifies them for metonymic agent positions: If water can drive mills or wash clothes, money can buy things and manpower. But water is also like agents in that it can move, apparently without being driven by external force. And because it can move, it can find things. Money, on the other hand, cannot move on its own, but has to be moved by people. Thus, in the case of money finding a level, a metonymic step must be interpolated before its tertium comparationis to water can be established:
–––––––––– (26)The Polish ‘golden age’ was finally submerged by a flood of Italian musicians brought in by Sigismund 3. |SJ68|P54 12 13
Despite the knife finding its mark, our culture is not that animistic ... Star Trek connoisseurs will be acquainted with such creatures.
Metaphor and metonymy in verb semantics
241
(26) ( indicates a metonymy / metaphor, indicates an instantiation) AGENT finds LOCATION water finds level -------------------------------------------------------------------------------PEOPLE act using money Money acts (for people) Money finds level ============================================== Water finds level Money finds level Note that in the money metonymy, the people behind the money conveniently vanish as behind a smokescreen. We have seen that in the agent domain of metaphorised verbs, processes can become very complex. Several metaphors and metonymies may converge to apply verbs in new, more abstract senses. But the basic processes remain the same: – turning a physical process into an abstract cognitive one – turning a physical object into an abstract one – metonymically replacing an agent by an instrument – turning a physical agent into an abstract force. These processes can presumably be applied recursively, as it were. Whether they may actually yield an infinite number of senses, remains to be seen. References Hopper, Paul J., Sandra A. Thompson: “Transitivity in grammar and discourse”, Language 56, 2 (1980), 251–99. Horn, Laurence H.: “Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature”, in: Schiffrin, Deborah (ed.): Meaning, form and use in context: linguistic applications. Washington: Georgetown University Press (Georgetown University round table on languages and linguistics. 1984), XII, 1984, 11–42. Lakoff, George, Mark Johnson: Metaphors We Live by. Chicago: Chicago University Press 1980. Lyons, John: Introduction to theoretical linguistics. Cambridge: Cambridge University Press 1968.
242
Paul Georg Meyer
Meyer, Paul Georg: “Nicht fachgebundene Lexik in Wissenschaftstexten: Versuch einer Klassifikation und Einschätzung ihrer Funktionen”, in: Kalverkämper, Hartwig, Klaus-Dieter Baumann (eds.): Fachliche Textsorten: Komponenten – Relationen – Strategien. Tübingen: Narr (Forum für Fachsprachenforschung, 25) 1996, 175–92. – Coming to Know: Studies in the Lexical Semantics and Pragmatics of Academic English. Tübingen: Narr (Forum für Fachsprachenforschung, 35) 1997. – “Nicht-fachliches Vokabular im Fachtext: ein vernachlässigter Aspekt der Fachsprachenforschung”, in: Börner, Wolfgang, Klaus Vogel (eds.): Wortschatz und Fremdsprachenerwerb. Bochum: AKS-Verlag (Fremdsprachen in Lehre und Forschung, 14) 1993, 177–99. Nerlich, Brigitte, Zazie Todd, Vimala Herman, David D. Clarke (eds.): Polysemy: Flexible Patterns of Meaning in Mind and Language. Berlin–New York: Mouton de Gruyter 2003. Saussure, Ferdinand de : Cours de linguistique générale. Ed. By Bally, Charles & Sechehaye, Albert. Paris: Payot 1965. Sweetser, Eve: From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Sructure. Cambridge: Cambridge University Press (Cambridge studies in linguistics, 54) 1990. Vendler, Zeno: Linguistics in Philosophy. Ithaca: Cornell University Press 1967.
2. Metonymy, Synecdoche and Truth
Armin Burkhardt (Magdeburg, D)
Between poetry and economy Metonymy as a semantic principle*
0. 1. 2. 3. 4. 5. 6. 7.
Introduction Metonymy and rhetoric Metonymy and semantics Metonymy and pragmatics Metonymy and historical semantics Metonymy and semiotics Metonymy and truth Conclusion
0. Introduction The realm of the tropes is a world of colour and light. I would like to invite the reader on a trip into this richly coloured world. However, our path will not lead through the fertile regions on both sides of the equator, but entail a turning towards1 those tropes that were defined by Quintilian (IX, 1, 4) as linguistic transfer phenomena, which act as decoration for speech.2 One of the tropical phenomena that are meant here is metonymy. Since the days of antiquity it is known in rhetoric as a stylistic means of poetry. However, while its elder sister, metaphor, has always been in the focus of scientific and philosophical interest3 and while, indeed, the analysis of the metaphorical systems of literary language, everyday language and, more recently, ————— *
1
2
3
The German original of this chapter “Zwischen Poesie und Ökonomie. Die Metonymie als semantisches Prinzip” was first published in Zeitschrift für germanistische Linguistik 24 (1996), 175–94. For the present book it has been translated, updated, slightly revised and extended by some reflections on metonymy and truth. A first draft was translated by Jennifer Roberts. According to the origin of the rhetoric term from gr. τροπή ‘turning (towards), direction’ and τρόπος ‘turn’ (from τρέπειν ‘turn (back)’). More exactly as a “sermo a naturali et principali significatione translatus ad aliam ornandae orationis gratia” [“language transferred from its natural and principal meaning to another for the sake of embellishment”]. This is whitnessed by the impressing quantity of titles listed in the bibliographies of van Noppen, de Knop and Jongen (1985), van Noppen and Hols (1991) and de Knop, Dirven, Yu and Smieja (2005).
246
Armin Burkhardt
languages for special purposes, has, under the influence of Lakoff and Johnson’s (1980) cognitivist approach, assumed an almost paradigmatic status, the semantic structure, communicative impact and cognitive value of metonymy had attracted much less attention in research until 1995 when Panther and Radden’s volume appeared in which the ubiquity, cognitive importance and variety of metonymical phenomena was clearly exposed. Since then the trope which had formerly been widely neglected has been in the focus of numerous cognitivist studies (cf. e.g. Barcelona 2000, Panther & Thornburg 2003, Warren 2006). The present chapter is an attempt to derive a modern definition of metonymy from antique rhetoricians and to show the function and diversity of a trope which seems to be linguistically more important than Her Majesty, the metaphor, itself. 1. Metonymy and rhetorics (1) Ossian has ousted Homer from my heart (Goethe 2006: 107), a certain Werther wrote to his friend Wilhelm. This quote does not concern homoerotic relationships; rather the name of the author is used here to represent his work. Another example is when Faust says to Mephistopheles as an ultimatum: (2) Und das sag’ ich ihm kurz und gut, Wenn nicht das sueße junge Blut Heut’ Nacht in meinen Armen ruht; So sind wir um Mitternacht geschieden. (Goethe I, 14, 129, verse 263638) And this I say – few words are best – Unless that sweet young maiden lays Her head this night upon my breast, At midnight we’ve gone different ways. (Goethe 1941) This text excerpt is not an early example of vampirism; rather the “young blood” represents here the young woman (which is introduced in the English translation but absent in the German original!), through whose veins the young blood flows. (3) The herbs in between, Which I cut with ore by the still moonlight4 —————
4
My translation, A.B. The German original was quoted from Grimm (1962, 3: 1075).
Metonymy as a semantic principle
247
wrote Gryphius. This verse does not, as seems indicated, deal with metal ores, which of course would hardly be able to cut herbs; rather the name of the raw material represents the “knife” which is produced from the original metal as a result of various technical procedures. The three passages quoted contain examples of the trope of metonymy or “permutation of names” which, according to Lausberg (1973, § 565), consists in “replacing the real word [verbum proprium] by another word the proper meaning of which stands in a real relation to the topical meaning content which is originally meant”5. In books on rhetoric that use such definitions one can usually find lists of relations that may exist between the word which is substituted and the word which is substituting. According to the definitions of classical rhetoric, the substition “author for work” is considered a metonymy, the substitution of a part for the whole (blood for young woman), however, is considered a synecdoche. The classification for the relationship “material for product” (ore for knife), which appears in the Gryphius example, is traditionally considered controversial and, depending on the taste of the respective rhetorician, is made very differently. In general, the relationships “producer for product”, “author for work”, “owner for property”, “raw materials for endproduct”; “cause for effect”, “container for contents” are allotted to metonymy6, while the remaining relations of replacement “singular for plural” (for example, the German for the Germans), “plural for singular” (such as the pluralia auctoris and majestatis) and pars pro toto (roof for house) or totum pro parte (state for government) are catagorized as synecdoche. Despite this general tendency definitions have varied until today: Dubois et al. (1974, 1706) and Göttert (1991: 50) have ascribed the relations part/whole and raw material/object to synecdoche, while Plett (1979: 26772), Lausberg (1973: § 572) and Lakoff and Johnson (1980: 35) consider synecdoche as a special case of metonymy in a broader sense, whereas Seto only regards genus for species and species for genus, that is, taxonomic relations as synecdoche (see Nerlich, this volume, 299–321). This shows that a clear principle for the distinction between metonymy and synecdoche, which is more than 2,000 years old, is still missing. Eco (1985: 174) was therefore correct when he critically noted that traditional rhetoric has never satisfactorily explained why the substitutions genus/species and pars/totum should be synecdoches while all other types of substitution should be classified as metonymical. —————
5
6
My translation, A.B. The German original runs “[...], daß für das verbum proprium ein anderes Wort gesetzt wird, dessen eigentliche Bedeutung mit dem okkasionell gemeinten Bedeutungsinhalt in einer realen Beziehung [...] steht.” In the rhetoric tradition, at times some special subtypes of metonymy were called “metalepsis”. While for Donatus and other Latin grammarians metalepsis was a distance metonymy based on a chain of encyclopaedic associations within the same frame of experience, in Humanism and Baroque it was defined as a metonymy based on the relation of cause and effect by authors like Melanchthon, Vossius and many others. (Cf. Burkhardt 2001)
248
Armin Burkhardt
Metonymy, literally “renaming”, was defined by Quintilian as “nominis pro nomine positio”, or “the substitution of one name for another” (VIII, 6, 23) and is limited to using the creator in place of the creation and the owner in place of the property. Since then, it has been considered as a form of “replacing” or “permutation of names”. And the question about the point of such substitutions, as well as about the principle of division, has been faded away. Today there seems to be almost total agreement that metonymy is a “qualitative” word replacement or rather transfer, which is based on the relationship between the intended object and its associated aspects or elements. This is in contrast to synecdoche, which is based on the “quantitative” principles of set inclusion (cf. Lausberg 1973: § 572). Just as the problems of definition originated from the rhetoricians of Antiquity, their solutions are concealed in the literature of that time as well. Quintilian already ignored the clearer differentiations in Rhetorica ad Herennium, written in approximately 85 BC. In this work, which used to be attributed to Cicero7, metonymy appears under the name of “denominatio” (IV, 32) and is defined as a kind of directing manœuvre, which, loosely interpretated, consists in making comprehensible the actual intended object by using terms which denote related or “adjacent” (“res propinqui et finitimi”) objects instead. At the same time, synecdoche, under the title of “intellectio”, is restricted to the relationships singular for plural/plural for singular and pars pro toto/totum pro parte (IV, 33). The example sentences which are offered intimate that pars here is more readily understood as a subset than in the sense of a material component. In Rhetorica ad Herennium metonymy is defined not as a “replacement” but as the lexical displacement, which is made possible through the factual correlation of two elements of thought and which directs the attention to aspects of the objects referred to. In contrast, synecdoche is limited to the substitutive relationship between the main set and its subsets, i.e. set relationships. Therefore, the old argument concerning the categorization of the part/whole relationship can be put to rest as soon as it becomes clear that, as pars can mean ‘component’ as well as ‘subset’, it is but the effect of an equivocation. And it is only in the latter case that the pars/totum relationship can be viewed as quantitative and therefore as a case of synecdoche, while the relationship of a material component to the whole belongs to the set of metonymical relations.
————— 7
Apparently, Cicero himself conceives the terms hypallage (‘permutation’) and metonymia (‘change of denomination’) as synonyms: while the former is used by the rhetoricians, since for them the aspect of word interchange is in the focus, the latter is in use among the grammarians, the phenomenon in question is the transfer of a denomination (cf. Orator 27, 93).
Metonymy as a semantic principle
249
2. Metonymy and semantics Yet by the reason that it owes its name to the metonymical jump from the speech organ tongue, lingua, to the speech product language, the trope of metonymy is a particular challenge especially for linguistics. It can be sufficiently described only in the framework of semantic considerations. To do this, a suitable model for description must first be developed. Adopting two well-known terms from Roman Jakobson (1971: pp. 243) “contiguity” is defined as the neighbourly correlation of aspects and elements within a network of associations given by a joint frame of experience. “Similarity”, however, may be conceived as relations of set inclusion that are given by the intensional correspondence between hyperonyms and hyponyms. Word knowledge and word usage are always related to world knowledge and world experience. The knowledge of (content) words (such as nouns, verbs, adjectives and adverbs) and their meanings normally also includes the knowledge of the designated objects, events, properties and states of affairs. In this way, every word, in addition to the meaning of the core set of criteria which constitutes its meaning, i.e. the semantic features, simultaneously opens up a “field of association” (Bally 1940: 195–6) surrounding it. Bally´s conception of the “champs associatif” is comparable to the “frame” concept of recent semantics. However, as other authors, such as Minsky (1977) or van Dijk (1980), conceive of frames in a rather un-semantic sense as complexes of elements of stereotypical situations, I would like to found my ideas in this instance on Dressler and de Beaugrande’s (1981: 90) text-linguistic definition, according to which frames are “global patterns that contain commonsense knowledge about some central concept, e.g., ‘piggy banks’, ‘birthday parties’, etc.” Without wanting to raise the claim to be complete, the structure of such a frame-knowledge could be represented with a modified version of one of the schemes suggested by Eco (1985: 171–5), in which a vertical axis depicts synecdochical similarity and a horizontal axis metonymic contiguity (see p. 254). Because of the similiarity-relationship of the vertical axis, the person, in the sense of a generalizing synecdoche, can (or could) be substituted by mammal, living creature, mortal being or being, which in turn (conserving referential identity), e.g. when taking up again a forementioned element in the text, can (or could) replace one another in a particularizing or rather concretizing way in the opposite direction.8 In the same direction the singular forms could also replace the respective plural forms. ————— 8
Proceeding from the person (as an individual) alternative hierarchies of similarity would seem to be possible, e.g. individual, Westfalian, German, European, which also shows that many forms of explicit uptaking in texts are synedochical. (Implicit uptaking, by contrast, is intrinsically metonymical., both tropes, however, may serve as means of warranting textual coherence.)
250
Armin Burkhardt
The semantic contiguity relationship is represented by the horizontal axis. On this axis the metonymical selection can theoretically refer to every frameelement. According to our knowledge of the world or our cultural points of view, a person has certain characteristic components: a head, soul, face, ears, eyes, nose, mouth, tongue, heart, arms, legs, and many others, but he or she also has tonsils, a liver, an appendix, etc. He or she was produced from nature or his or her parents, consists of flesh, blood, cells, skin, and exists in order to live, think, feel and the like. In addition to this, he/she customarily wears clothing, under certain circumstances smokes cigarettes, perhaps eats cutlets, possibly has a car, and so on. Several characterics, especially the external parts of the body, belong to our human-stereotype and therefore are in the focus of our attention; others, such as the internal organs, are indeed in the same way obligatory, but usually only receive attention in the context of illness. Moreover, others, like the cultural characteristics that were mentioned, are completely optional. Because of this three-level attention-hierarchy metonymical substitutions originating from the core of the stereotype are quite common and easy to understand. For example, (4) (5) (6) (7) (8)
This land needs hard-working hands. That costs $5 a head. I´m glad to see a friendly face. They have to feed five hungry bellies. I’m all ears.
Metonymy as a semantic principle
(9) (10) (11) (12) (13) (14)
251
Two dark eyes were looking at me. There were 44 legs kicking on the playing field. His nature was problematic9. He’s an honest soul. His life came to an end.10 He deals with the thinking of early Heidegger.
In examples (4)–(14) characteristic parts, aspects or properties of human beings stand for the respective individuals themselves. According to the distance from the core of the frame some substitutions, however, are less frequent than others which seem to be more far-fetched.11 Features or components that lie outside the sphere of established stereotypes or are merely optional, generate metonymies that can only be interpreted or disambiguated in context: (15) The appendix is in room 12. (16) The steak at table 5 wants to pay. (17) Two cigarettes were walking through the city. (18) The black suits got into the taxi. In modern culture owning a certain car has become so strongly associated with the frame or stereotype of “person” that metonymical sentences such as the following may even be understood outside any concrete context, despite their ambiguity: (19) I am standing on the car park. (20) My mud-guard is broken.12 Moreover, the latter example in which the metonymy is expressed by the possessive pronoun proves that it is not limited to nouns at all. Adjectives and verbs may be used metonymically as well as is demonstrated by conventionalized wordings like ————— 9
10 11
12
Expressions such as the Puerto Rican for the baseball player Tony Ferndandez or Caesar’s adopted son for Brutus show that antonomasia – at least if the term does not refer to metaphors with proper names – is a special case of metonymy. Here life stands for the person whose organic function is life. Even if dysphemism, perhaps, on the whole is a domain of metaphor, in the lexica of invectives of the different languages a certain richness of those metaphors which are based on the pars-pro-toto use of the names of tabooed parts of the body can be noticed. As the respective words are not appropriate for use in written language, I ask for the reader’s comprehension, if I renounce giving examples here for he will know them anyway. “Are you full already?” a German bus driver asked his colleague at the central bus station on August 28, 1995, immediately before departing for the Rolling Stones concert in Wolfsburg.
252
Armin Burkhardt
(21) Apples are healthy. for ‘apples contain a lot of vitamins’ or ‘Eating apples makes or keeps people who eat them healthy’ or synaesthetical changes of perspective like (22) The cheese is smelling. for ‘the cheese is stinking’ or ‘the cheese brings about that it can be smelled’. th As was characteristic of 20 -century literature, which tried to explore the limits of language (cf. Henne 1996), Gottfried Benn, in his poem “Night café” (1986: 19) made use of an extreme form of metonymy: By just mentioning in an abbreviating way what is merely accidental he only intimates the intended referent and thereby focuses the designation in question on but one aspect of perception: Night Café 824: The love and life of the women. The cello takes a quick sip. The flute belches loudly for three bars: the lovely supper. The drum reads a detective book to the end. Green teeth, pimple on the face waves to an eye-shadow inflammation. Greasy hair speaks to the open mouth with tonsils faith hope charity around the neck. The young goitre is good to saddlenose. He pays for her three beers. Beard moss buys carnations to soften up double chin. B-Minor: Sonate 35. Two eyes shout out loud: Don´t spray Chopin´s blood in this room, so that the riffraff slouches around on it. Stop! Hey, Gigi! The door moves back: a woman. Dried-up like the desert. Canaanian brown. Chaste. Caved-in. A scent comes with her. Hardly a scent. It is just the sweet forward curving of the air against my brain. An obesity toddles along behind.13
—————
13
My translation, A.B.
Metonymy as a semantic principle
253
What first appears to be a sequence of personifications turns out to be a series of metonymies of the types “instrument for player” (“cello”, “flute”, “drum”), pars pro toto (“two eyes”) and accidentals for substance (“green teeth”, “pimple on the face”, “eye-shadow inflammation”, “greasy hair”, “open mouth with tonsils”, “young goitre”, “saddlenose”, “beard moss”, “double chin”, “obesity”, which imply a part-whole relationship and therefore could be considered as weaker forms of pars pro toto as well). Just the mere listing demonstrates that the last category of metonymy is dominating here. The metonyms are used to focus on negative characteristics and to identify their owners with them. In their book “Expressionismus” Vietta and Kemper (1965, pp. 62) have defined the numerous examples in Benn’s poem of replacing the designations of persons by those of the parts, attributes, and aspects belonging to them as being of the (even in rhetoric) earlier unheard of “(physical) indication for bearer” type of synecdoche14. Their reason for this definition is that these examples – with Lausberg – concern the displacement of the names of the intended objects within the sphere of intension. However, this holds for synecdoche but not for Benn’s expressions to which the definition refers, for indeed, they are metonymies, by means of which, as Vietta and Kemper write, the literary dissolution of the subject is achieved.15 With the help of frame theory the phenomenon of metonymy, in my opinion, is quite easy to explain: if a “frame” is nothing other than the customary schematic knowledge about a (proptotypical) object or state of affairs, metonymy can be defined as the replacement of the customary designation of an intended object of reference by the denomination of one of the factors or aspects usually (or incidentally) associated with it, that is to say, metonymies are focus-replacements whithin one and the same “frame” with the effect that the designation of the metonymically associated frame-characteristic is, in turn, associated with the intended object of reference.16 This also holds for the pars-pro-toto and the totum-pro-parte relationships so long as the relation to parts in the sense of material components is under consideration. By contrast, synecdoche in both its directions is limited intensionally to the hyponym/hyperonym relationship and extensionally to the subset/main set relation17 and in this respect may not, as in Plett (1979: 260–7), be understood as a “contiguity-trope”.18 With synecdoche, there is either uno pro pluribus as in the famous quote from Goethe (I, 42.2, 238), “The German should learn all languages, so that no stranger is uncomfortable to him at home, and he is at —————
14 15 16 17 18
Though examples like the hunchback of Notre Dame show that such uses are not as uncommon as one could be inclined to think. To be more exact, the main principle in Benn’s poem is what Warren (1999, 2006) calls “referential metononymy” which will be dealt with later. A frame-theoretical approach for the explanation of metonymy is also outlined by Nerlich & Clarke (1988: pp. 82). Cf. also the detailed explanations in Nerlich (1997). In the meantime this idea has been supported and elaborated by Seto (1999).
254
Armin Burkhardt
home anywhere else”19, or genus pro specie, in “Unfortunately the people trusted [in the princes] and laid down to rest. And thus Germany was betrayed like France”20 (1973: 139) from Büchner’s “The Hessian Courier”21 Consequently, synecdoche is the expression of a quantitative relation which is semantic in the narrow sense, while metonymy expresses a qualitative, encyclopaedic relation. While synecdoche is a result of set-inlusion, metonymy is the result of the affiliation, solidarity, or, with Jakobson, the “contiguity” within a certain “frame” or “script”.22 With synecdoches, either o n e object-set appears in the light of its parts or a part-set in the light of the whole set to which it belongs. In contrast, with metonymy o n e object is seen in the light of its parts, aspects or accidentals. I think, however, that the creative displacement by the substitutive use of a frame element is but one facet of metonymy. The other is economical shortening in speech, i.e. a (rather semantic) kind of ellipsis. Metonymy, therefore, “enables us to say things quicker, to shorten conceptual distances” (Nerlich, Clarke & Todd 1999: 362). It may be taken as a kind of puzzle – which it usually is in literature – or as an abbreviation device (cf. Warren 1999: 128; Nerlich & Clarke 2001)23 – which is its main use in everyday life. 3. Metonymy and pragmatics The following passage contains a good example of minimal linguistic discourse. It is taken from a locus classicus of German literature: We went to the window. The thunder was passing by and a wonderful rain was falling on the land, filling the warm air with the most refreshing fragrance. She stood there resting on her elbows, gazing deep into the country about us; she looked to the heavens, and at me, and I saw there were tears in her eyes; and she laid her hand on mine and said, ‘Klopstock!’. (Goethe 2006: 30–1)
————— 19 20
21
22
23
My translation, A.B. My translation, A.B. The German original runs: “Das Volk traute ihnen leider und legte sich zur Ruhe. – Und so ward Deutschland betrogen wie Frankreich.” It is interesting to note that the “official” English translation makes the original metonymy explicit: “Unfortunately the people trusted them [the princes] and laid down their arms. – And so the Germans were deceived like the French.” (1987: 238) The German original of this revolutionary pamphlet from 1834 was entitled “Der Hessische Landbote”. It was written by Georg Büchner and modified to a considerable extent by Friedrich Ludwig Weidig. By contrast to this metaphor is a transfer between different frames by which a focused object appears in the light of another (cf. Burkhardt 1987). On metonymy and contiguity see now Piersman & Geeraerts (2006). Nerlich & Clarke (2001) also show that metonymy is a means of semantic abbreviation while classical ellipsis is purely syntactic.
Metonymy as a semantic principle
255
The name Klopstock does not mean the illustrious German poet here but his work. Therefore, one could first think that it is a simple metonymy of the “author for work” type we are already familiar with. Certainly, this idea is not entirely wrong because it can be inferred from the immediately following sentence that the metonymy is understood exactly in this sense: Werther, the addressee of the short outcry, is immediately reminded of the “the glorious ode she had in mind” (ibid.: 31). Nevertheless, the metonymy has the character of an allusion at the same time, yet it communicates – in an indirect way – more than a mere reference to the work. In the second part of the next sentence one finds a metaphorical description of how Werther “was lost in the sensations that flooded me on hearing the name [password]”24 (ibid.). Klopstock does not just stand for the intended literary work, but is an invitation to remember and a coding for the feelings that are triggered through this work. Consequently the poet’s name which is uttered by Lotte, is correctly called a “password” in the text. It is in a way a spontaneous catchword: the most economical form of expression of the intended meaning. In poetic language therefore metonymy seems to serve as a short-cut by focusing on secondary characteristics within a given frame.25 Thus, the simple explanation in the sense of replacement, name exchange or displacement is not sufficient here. “When one speaks of a ‘Rembrandt’ or a ‘Bordeaux’”, Ullmann (1957: 243– 4) writes referring to Wellander, “these are partly metonymic uses rooted in contiguity (author – product, product – place of origin), partly shortenings of fuller phrases like ‘a Rembrandt picture, a painting by Rembrandt’ and ‘Bordeaux wine’ [...].” Since scholars used to concentrate on conceiving it as a trope, which serves the ornatus within the elocutio, metonymy (as well as syndecdoche) in rhetoric was viewed as a means of linguistic creativity, as an act of linguistic formation of poetic style. Jakobson (1971: 263), too, found fault with this. And it was the anonymous author of Rhetorica ad Herennium who already pointed out that “the use of metonymies of this kind is abundant not only amongst the poets and orators but also in everyday.” (IV, 32, 44) Though there it serves more profane purposes. In this way it is used, for example, in political speech. A typical case is the place name metonymy: —————
24
25
The German original runs “in dem Strome von Empfindungen, den sie in dieser Losung über mich ausgoß”; Losung here means something like ‘slogan’, ‘password’ or ‘watchword’, indicating that “Klopstock” is more than just a name here. Though, admittedly, this may also happen in more prosaic forms of speech. In England for example, if there is a crisis, emotional, financial or otherwise, the first thing people often say is “Let’s put the kettle on”, which means not only ‘Let’s make a cup of tea’, but is a metonymical shorthand for ‘Let’s sit down (, drink a cup of tea) and work out a solution to this problem together’. It encapsulates a whole social situation and a stereotypically English way of dealing with it. This example was pointed out to me by Brigitte Nerlich.
256
Armin Burkhardt
(23) Dr. Dregger (CDU/CSU): [...] Ladies and Gentlemen, we must prevent Aushwitz, prevent Hiroshima and prevent Dresden.26 Immediately, from encyclopaedic knowledge historical events from the time of World War Two are associated with these place names: the genocide of jews, the dropping of the first atomic bomb and the destruction of German cities. Dregger’s sentence is an extremely condensed and economical expression for “We must prevent methodical genocides, such as those which took place in Aushwitz between 1939–45, or were brought about by the atomic bomb in Hiroshima on August 6, 1945, and the bombing of Dresden on February 13/14, 1944, from becoming possible ever again.” In news headlines such as No Agreement Between Paris and London, the proper names are codings for the goverments residing in the capitals mentioned.27 In the same sense, proper name compounds such as Maastricht Treaty or Helsinki Conference can be understood as metonymies. Beyond the field of proper names, some of the political slogans can be interpreted as metonymically shortened forms which, serving communicative economy and requiring expert knowledge, just focus on single important aspects of the denoted object. For example, Location Germany28 or AmigoAffair. Trusting the knowledge of the recipient, only the important aspects of the denoted object are stressed lexically. Elsewhere (Burkhardt 2003: 352–69) I have called these types of slogans “cues” or “topic words”. Euphemisms, which are not untypical for political language in particular, can be synecdochical, metonymical or metaphorical. They are synecdochical when speakers avoid the appropriate word by using abstract generic terms which have less semantic features, such as device or system for ‘weapon’, flying aircraft for ‘missile’, or price adjustment for ‘price increase’. Euphemisms are metonymical when they exessively focus on secondary aspects within the relevant frames, such as calling the dismissal of workers as setting them free or the draining of poisonous chemicals into the high sea as dumping. And finally, euphemisms are metaphorical when they play down a danger by giving a harmless analogy, e.g. by speaking of an atomic volcano or of an industrial or wastedisposal park.29 —————
26
27
28 29
My translation, A.B. The German original runs: “Meine Damen und Herren, wir müssen Auschwitz verhindern, Horoshima verhindern und Dresden verhindern.” The example is taken from a parlamentary speech given on April 18, 1985, in the German Bundestag. The main topic of the debate was the so-called Strategic Defense Initiative (SDI). Metonymies of this type become particularly obvious in expressions like Finally! Bonn moves to Berlin [“Na endlich! Bonn zieht nach Berlin”, headline of BILDZeitung on Jan. 15, 1994]. For a detailed analysis of this catchword (“Standort Deutschland”) cf. Burkhardt (1996a: pp. 91). For a more detailed analysis of euphemism see Burkhardt (in this volume, 357– 74).
Metonymy as a semantic principle
257
However, even in everyday communication in the narrow sense, metonymy is mostly employed for the purpose of economical speech.30 When we still had chickens at my home, every evening one of the family members asked the question “Have you closed up the chickens yet?” Naturally, we were referring to the chicken-coup, to whose “frame” the existence of the chickens belongs.31 “‘Not everybody who is forgetful has Alzheimer’s’, Professor Dr. Siegfried Hoyer from Heidelberg points out”, neue braunschweiger (newspaper) announced several years ago.32 Here the name of the scientist who discovered the disease has become its metonymical shorthand. Utterances like I’d like to have another plate [with food] or The Christmas tree is burning [i.e. the candles on the tree’s twigs are burning], everyday requests like A [beer from the] Wolters [brewery], please!, Ring her up [and ask] what’s wrong, Please do fax us [a report on] your most terrible holiday experience!33 and everyday questions like When is the garbage coming tomorrow? or Has the cake rung already? must be understood in the same way. While in spoken language this kind of metonymy is a welcome method of saving time, energy, and words, it is regarded as extremely unwelcome in written language, which, because of the missing context, must follow the ideal of being explicit. Typical mistakes are referring to the described instead of referring to its description and that of referring to the object of study instead of referring to the respective concept or branch of knowledge. Put another way: it is a question of using object language in place of metalanguage. The following two examples from a small corpus of (German) student mistakes in grammar and style may illustrate this. One example concerns the handling of articles in German grammars: (24) The uses of articles are identical with those in other grammars [...]. And in a thesis on particles in different languages one could find: (25) In French linguistic research the interjections were as much disputed as in the other two languages. It is not meant in (24) that the u s e s of articles are identical with those, i.e. the uses, in other grammars, but rather that the d e s c r i p t i o n of the uses of articles in a particular grammar is identical with the d e s c r i p t i o n of the uses in other grammars. And (25) was meant to express that in French lingui—————
30 31
32 33
Cf. also Le Guern (1973: pp. 23) who interprets metonymy as an ellipsis as well, but neglects its poetic force at the same time. One could be inclined to take this process as a mere ellipsis of the basic component of the compound. However, as it is asked whether the respective person has alresady closed “the chickens” “die Hühner”, accusative plural] and not “the chicken” “den Hühner”, accusative singular], the wording tells against this. In no. 41 of Oct. 16, 1994. Earwitnessed on NDR 2, August 31, 1994.
258
Armin Burkhardt
stic research the c o n c e p t of interjection was as much disputed as in the l i n g u i s t i c s of languages such as English or German. In (24) the designation of the object of acting, in a way which is stylistically incorrect, stands for the act itself. In (25) the conceived objects are mentioned in place of the concept and the object of investigation in place of the field of research. In each case, however, of two obligatory elements of information one is mentioned at least. Probably, such metonymical lapses come about because the encyclopaedic connection between both elements of information is so narrow that one of them is simplified away by the speaker’s or writer’s mind. Such examples from everyday language show more clearly than literary examples that metonymy is often less a replacement than an elliptical shorthand: the reduction of syntactic contiguity under preservation and by taking advantage of semantic contiguity which is guaranteed by frame-identity. Through the metonymic focussing on a particularly important aspect of the intended subject the verbal expenditure as well as the semantic complexitiy of the “normal” expression is reduced.34 4. Metonymy and historical semantics Up to now we have dealt predominantly with metonymies which trace back to lexical replacements or shortened syntactic forms. Metonymies of this kind could be called “onomasiological” because they are created by the associative and at the same time focusing transfer between the designation of a certain referent and designations for its different aspects and elements within the same frame. Traditional rhetoric typically dealt only with this kind of metonymies. However, metonymy is not only frequent in the synchronic production of texts but is at the same time a semantic structuring principle, which has an effect on the development of the meanings of individual linguistic signs over time, that is diachronically. In contrast with the “onomasiological” type where the signifier is replaced, there is another manifestation of metonymy which consists in designating different frame-characteristics with the same word and, thus, gives rise to additional meanings. Metonymical derivations of this type I should like to describe as “semasiological” metonymies. “Semasiological” metonymy is a means of linguistic economy enabled by contiguity.35 As soon as it is recognized as a principle of semantic change lingui————— 34 35
Metonymies of the elliptic kind have been termed “referential” by Beatrice Warren (1999, 2006). Cf. also chap. 6 below. As it always includes a re-categorization conversion as a procedure of word formation generally takes profit from the principle of semasiological metonymy. As is shown by the following examples, however, conversion in semantic respect sometimes goes beyond the mere change of the (grammatical-ontological) category: das Essen [‘meal’] < essen [‘to eat’], a drink < to drink, a call< to call.
Metonymy as a semantic principle
259
stics holds in its hands the key for the explanation of the historical formation and development of polysemies – at least a large part of them. It is in fact a commonplace in historical semantics to say that metonymy is one of the major types of semantic change, alongside with metaphor, generalisation and specialisation. Ullmann (1957: pp. 232), among others, defined it as “transfer by contiguity between the senses” and refers to the example of the old French word burel, which was originally the diminutive from bure ‘coarse cloth’ and first meant the piece of furniture it used to cover, before it finally became – after also naming the office where the furniture stands – the expression for the staff who work in that office (vgl. ibid.: 233).36 On the basis of our common knowledge about frames and of the oscillation of reference between different aspects or elements of the same frame in communication practice word meanings may gradually shift.37 Such a change of meaning through what I call “semasiological” metonymy has occurred, for example, when we currently use the word handwriting for something written by hand as well or speak of a handwriting from the Middle Ages or when words such as church38 or school, which first designated just the relevant buildings or rooms, come to also describe the respective institutions and its members and representatives. The semantic development of the German word Klaue, which also shows that metaphorical and metonymical processes may be related to one another in the history of meaning, is comparable to the semantic history of handwriting. The ohg. word klawa, mhg. Klaue, first meant ‘claw, paw, originally the grasping, clenching’ (cf. Duden 1989: 842), but from Luther’s times on at latest in slovenly speech it was used as a metaphor for the human hand before it finally took over the additional meaning ‘bad handwriting’ by metonymical derivation (cf. Paul 2002, 536). Metonymical development of meanings pervades the entire dictionary. I just mention the examples railway [as the denomination of the rail tracks, the company and its employees], coffee [as the name of the well-known plant and the drink which is gained from its fruit], wardrobe [as the denomination of the place —————
36
37
38
This example is also discussed in Le Guern (1973: 93) who writes: “L’étoffe que désignait ce terme dans son sens primitif n’existe plus. [...] La métonymie a été lexicalisée à partir du moment où l’on a perdu l’habitude de recouvrir la table à écrire de cette étoffe particulière. Et c’est par rapport à ce nouveau sens propre qu’ont pu naître et se généraliser des emplois métonymiques qui désignent la pièce où se trouve la table à écrire, l’entreprise qui comprend un certain nombre de ces pièces (bureau d’études), l’équipe dirigeante d’une association (qui, lors des réunions, prend place autour de la table à écrire), etc.” Le Guern, too, describes metonymy as one of the principles of semantic change. The development of such new meanings may follow certain standard cognitive clusters or form historical “chains” which have been termed “serial metonymy” by Nerlich & Clarke (2001). In the case of Greek-Latin ecclesia the metonymical process of derivation went into the opposite direction. The denomination of the parish was transferred to the locality where the regular services used to take place.
260
Armin Burkhardt
where clothes are kept and for the clothing which is kept there] and phraseologisms like to run a red light [instead of ‘to pass a crossing with traffic lights one of which is shining red’], to be in one’s own four walls, to sign-post a direction or to file a petition.39 Even the term semantics itself has recently developed a new metonymical meaning: as the designation of the semiotic discipline which deals with meanings, semantics can now also be used as a synonym for meaning.40 “Not with words, not with language,” writes Botho Strauß, “but with the accented characteristcs, with the entire imposing behaviour of the voice can others communicate with me, a continuous phonetic making of an impression, which is subliminally as effective as the eyes and hands, as easy to mix as semantics or gestures [...].” (1992: 35)41 The cases that were dealt with until now are quite evident examples of the change of already existing lexical-semantic structures. However, in his famous attempt of deriving the origin of language from the original act of creating onomatopoetic interjections Herder (1772) has shown that linguistic signs arise from the mental process of combining object properties with phonetic forms associated with those objects to the effect that they become a “Merkwort” [retain word]: The sound of bleating perceived by a human soul as the distinguishing mark of the sheep became, by virtue of this reflection, the name of the sheep, [...]. He recognized the sheep by its bleating: This was a conceived sign through which the soul clearly remembered an idea – and what is that other than a word? And what is the entire human language other than a collection of such words? (1966: 117)
In the original imitation of sound the fixation of the object, which is already metonymical itself, via the external feature of the noise connected with it ————— 39
40 41
The language of sports in particular is full of metonymies, as is shown by the examples: He kicked the goal in minute 89, The corner flew high into the penalty area, She calmly converted her first matchball (cf. Burkhardt 2006). Other cases which are comparable to this are the common substitution of technical by technological and methodical by methodological. My translation, A.B. – Where the metonymical character of such semantic processes of differentiation has come to its attention at all, the explanations of diachronic semantics widely remained limited to such processes of generating polysemies. In the framework of my theory of the “semantic-pragmatic jumps” (Burkhardt 1991: 25–32) I have also dealt with cases of semantification of the typical contextual conditions. With Trier’s (1931) famous example of the semantic scattering of the mhg. lexemes kunst and list it could be shown that the semantification of a presumed rule of use ‘for intellectual and practical capacities in higher language use kunst, in normal to lower language use list instead’ must have lead to the consequence that, by the end of mhg. time, kunst had assumed he meaning ‘courtly intellectual and practical capacity’ while list developed the sense ‘lower, morally objectionable capacity or way of acting’. As they can be traced back to focus displacements within the respective frames, I now feel justified to consider context rule “jumps” as semasiological metonymies as well.
Metonymy as a semantic principle
261
turns over into the metonymical use of an external feature to name the referent. The first language is, therefore, “Imitation [...] of the sounding, acting, stirring nature! Taken from the interjections of all beings and animated by the interjections of human emotion!” (ibid.: 135). In this sense onomatopoetic expressions, such as chouchou, cuckoo, raffraff and the like, may be generally taken as the results of metonymical processes. The original semiotic representation of the sheep by imitating its bleating which is its typical sound expression is a – somewhat basic – metonymical process in the same way as the sound imitation of the noise of spitting, i.e. fie!, phew! or pooh!, is the common semiotic representation of the feeling of disgust which would provoke the spitting. In a similar way, the emergence of the German emotional interjection ach [‘alas!’] in its original historical meaning must be interpreted as a metonymical shift from an accompanying aspect to the designation of the object itself: a symptom (the expiring sigh) becomes the metonymical designation of the feeling causing it (and is then translated into the phonemes of language).42 In the field of comparative historical semantics the distinction between semasiological and onomasiological metonymies can be employed successfully as well. Semasiological metonymy, as a rule, is the basis of those “false friends” that can be traced back to a common origin. The German word Spektakel and its English, French and Italian relatives spectacle, spectacle and spettacolo, for example, take their historical starting-point in the Latin noun spectaculum, which in turn is derived from the verb spicere ‘see, watch’ and, in Latin already, has developed the meanings ‘sight, spectacle, scene’ which are related to one another by frame-contiguity. While the French language exactly maintains the first two Latin meanings and the Italian language, moreover, develops the meaning ‘theatre (as an art)’, the German language in a way likewise metonymically focuses on the noise and the sensation being connected with scenes or sights of the various kinds. In the English word spectacle the original meanings ‘sight’, ‘scene’ the German language has lost are preserved, it may, however, designate the means with which the scene or the sight can be viewed, namely the glasses (which in turn owe their designation to a metonymy of the material-for-product type). Historical processes of semantic differentiation between different languages, i.e. cases of “interlinguistic polysemy” (Nübold 1993: 203), may thus be explained as metonymical “jumps”.43 ————— 42 43
See also the more detailed explanation in Burkhardt (1998). Further examples of the interlingual metonymical unfolding of meaning are e.g. the development of Italian cavaliere ‘horse rider, knight’ via French chevalier ‘knight, nobleman’ to German Kavalier ‘man with polite and galant behaviour towards women’ and Spanish caballero ‘gentleman, man’, or the semantic history of cup which first meant ‘small porcelain bowl with a handle for tea, coffee etc.’ and ‘vessel given as a prize in competions’ and then by metonymical change of perspective became the denomination of the competition itself in which for winnig the
262
Armin Burkhardt
The historical operation of onomasiological metonymy, however, may be illustrated by the following comparison: while the German Auto(mobile) stresses the aspect of locomotion by one’s own force, the Italian macchina and the Russian mashina give special prominence to the presence of an engine serving this purpose. The English word car is derived from the Latin carrus and focusses on its being four-wheeled. The Spanish coche, in contrast, highlights its original carriage-like body and its purpose resembling that of a carriage as well.44 The French voiture, finally, owes its existence to the Latin vectura ‘freight, transport, conveyance’ and this way emphasizes the car’s character as a means of transportation. Today the words or heteronyms mentioned denominate exactly the same class of objects, i.e. have the same extension, but the historical motive of designation underlying present-day usage is different in each case. Here, too, metonymical designations are concerned, for every linguistic sign, as far as it is motivated, is a shorthand focusation. 5. Metonymy and semiotics Metonymy is not limited to the sphere of verbal language, but can be demonstrated also for the use of indexical, iconic and symbolic (or emblematic) signs. Foot and finger prints, for example, are manifestations of indexical metonymy. Iconic metonymy, in contrast, appears e.g. in what Eco calls the “semiotization of the referent” (1972: 81–5), i.e. in using objects as signs. If e.g., to pilfer an example from Eco, an Italian barber hangs up a soap dish in front of the entrance door of his shop, this is a metonymical icon which means ‘barber’. Like any other imitation, gestures, too, can be metonymical in ways that may differ in each culture: while the German gesture for ‘drink’ iconically imitates the hand holding a cup or mug, the Italian equivalent concentrates on representing the process of drinking. If, however, an object is used as a sign of its class, e.g. when exhibiting a vehicle as a specimen of a certain type of car in the shop-window of a car-dealer’s, we have a case of iconic synecdoche. Lakoff and Johnson (1980: 37), too, refer to a case of iconic metonymy when they call attention to the fact that THE FACE FOR THE PERSON is a typical metonymy in Western culture. A significant proof for this is the fact that “we look at a person’s face – rather than his posture or his movements – to get our basic information about what the person is like”, and therefore are more inclined to accept the photo of a person’s face as her portrait than a picture of his or her trunk. ————— 44
winner is awarded a trophy (e.g. participation in the European cups). For a more detailed account of the cup example cf. Dirven 1985: 101–4). Originally, it was – like the German word Kutsche (‘carrriage’) – a metonymical shorthand for the Hungarian koczi szekér ‘car from Kocs’ (cf. Paul 2002: 579).
Metonymy as a semantic principle
263
Symbols/emblems can be metonymical, too. In this way the cross became the symbol of the crucified and the crucified became the symbol of Christianity, i.e. of his followers. Transformed into a flag the cockade of the student’s associations became the symbol of the democratic national German state called for by them. Another example is taken from a novel by David Lodge: In her lecture on the industrial novel Robyn Pemrose, temporary lecturer in English literature at the (fictitious) University of Rummidge, talks about the “phallocentrism” of industrial capitalism and, among other things, utters the sentence: “The most commonplace metonymic index of industry – the factory chimney – is also metaphorically a phallic symbol.” (Lodge 1988: 78) The factory chimney as a more and more antiquated part of industrial society, comes to be its metonymical symbol, exactly because it has become antiquated. The (red) rose, by contrast, which has been a frequent topic in Eco’s writings, seems to be a metaphorical symbol, in the ecclesiastical tradition as well. It owes its existence as a symbol to the analogy between a rose’s attributes like beauty, redness, transitoriness and thorniness and human attributes like youth, blood, decay or death and pain (cf. Burkhardt 1991a: 85–9 and 1996a: 470–2, 476–7). The King, Queen or President of a country, however, can be regarded as the synecdochical symbol of its people. How accidental the metonymical relation in particular can be is illustrated by the fact that the fish came to be the symbol of the Christian faith simply because the acronym of the Greek equivalent of Jesus Christ, God’s Son, Saviour, if read letter by letter in one word, happens to sound like ichthys and then, in Greek, means ‘fish’. Metaphor, metonymy and synecdoche are what I call “semiotic modes” according to which icons, indices and symbols are construed (cf. Burkhardt 2000: 131). 6. Metonymy and truth While it seems to be rather clear that metonymies can be “wrong” in the sense that the alleged aspects they focus on do not belong to the objects they are referring to (if, e.g., someone says I read Hemingway over the weekend when, in fact, he had read a novel by Steinbeck or The buses are on strike when its not the bus but the taxi drivers who are downing tools45), the question which still has to be answered is whether sentences or rather propositions containing “correct” metonymies can be true or false at all. In the case of onomasiological metononymies the answer is quite easy because they are nothing but new meanings. And if such words with their metonymical new meanings are inserted into sentences their propositions should be able to be true or false like those of any other well-formed sentence. Why should the proposition of an —————
45
In cases of shared ignorance or successful guessing, however, such referentially defective utterances may still be pragmatically felicitous.
264
Armin Burkhardt
affirmation like Arsenal didn’t win the cup for they lost the cup final have no detectable truth value?46 With semasiological metonymies the question is more complicated. Let’s briefly discuss the following examples: (26) Dean Whitehead’s corner was headed in by Danny Collins. (http://soccernetauction.fantasyleague.com) (27) A second half charge leads Germany to victory and to a third place finish at the 2006 World Cup. Portugal gave a good effort but they were unable to overcome a German team determined to go out in style. (http://www.worldcupblog.org) (28) SOS – Save our souls. A proposition can only be true or false if it describes, with a certain degree of precision, the conditions under which it would be true. In example (26) corner is a metonymical expression for it does not mean the pitch corner, but is used as a shorthand for ‘ball which was kicked from the corner flying in front of the goal’. The meaning of the whole sentence therefore is ‘(I assert that) the ball which was kicked from the corner by Dean Whitehead flew in front of the goal and was then headed in by Danny Collins’. The proposition of this sentence is true if the ball in question was in fact kicked from the corner by Dean Whitehead, then flew in front of the goal and was finally headed in by Danny Collins. Thus, although it contains a metonymy it can be true or false. In (27) Germany and Portugal stand for the national football teams of the respective countries. The proposition containing these country names can thus be paraphrased as ‘a second half charge leads the German national football team to victory and to a third place finish in the 2006 World Cup. The Portuguese national football team made a good effort but they were unable to overcome a German team determined to go out in style’. There is no reason why such a proposition should be incapable of being true or false. (28) is an imperative and expresses a request in the proposition where soul stands for ‘person’. Save our souls therefore means ‘save us people who have souls’. Although it is obvious that the utterance cannot be true or false, this does not contradict the above thesis at all, because the (pragmatic) reason for this is simply that it is not an assertion but a request which, according to Austin (1962), can only be felicitous or not. An assertive sentence like (29) Our souls were saved by the German coastguard —————
46
At least as long as one accepts that sentences which are composed of words expressing their abstract and undefined conventional meanings can be true or false at all. But as their use has been successful over the years (sometimes even centuries) there is no reason to think that this generally is not the case. Therefore, Nietzsche’s nihilist position (1873/2006: 117) according to which metonymies (like metaphors) prevent sentences from being true (or false) must be rejected in this point.
Metonymy as a semantic principle
265
containing the same metonymy as (28), like (26) and (27) would certainly have a truth value.47 Examples (26)–(29) contain metonymies which are referentially successful48. And there is no reason why a proposition should not have the capacity to be true or false if the reference of its referring expressions is clear. Warren (1999; 2006) has introduced the distinction between “propositional” and “referential” metonymies. According to her opinion (30) I waved down a taxi (31) They went to the altar would be examples of the former and (32) Maria is a divine voice (33) I have been reading the man [Chomsky] for ages would be examples of the latter. For her propositional metonymies are based on contiguity and they always presuppose another proposition which would give or rather clarify the information intended. Referential metonymies, by contrast, are elliptical forms of underlying complete syntactic structures in which the explicit element is the complement of an implicit one (which is added in thought). While propositional metonymies can be true or false, referential ones tend to violate truth conditions, for “one cannot literally [...] read a man; nor can a woman be a voice [...]. Propositional metonymy, on the other hand, tends to be literally true since the validity of the consequent (implicitly conveyed notion) depends on the validity of the antecedent (explicit expression)” (ibid., 9). All four metonymies have in common that an element of thought is syntactically omitted, but remains present as a kind of encyclopaedic presupposition. Whereas, for example, (31) may be taken as a shorthand for ‘they went to the altar where in Christian culture customarily wedding ceremonies are performed and, therefore, got married’, (32) incompletely expresses the proposition ‘Maria is a person who has a divine voice’. Both sentences are, in a way elliptical, and the ellipsis is rendered possible by contiguity relations within the relevant frame. However, whereas (30) and (31), which each contain a propositional metonymy, can be true or false because going by taxi generally presupposes waving for it and going to the altar is traditionally part of being married, (32) and (33) which each contain a referential metonymy can be true or false because devine voices or books never occur without call————— 47
48
Though, if one believes in the existence of souls here is not a matter of shared knowledge but of religious belief. But (26) and (27) also have ontological implications which cannot be discussed here. At least in the sense that there is a cultural convention which warrants the shared belief in the existence of the objects referred to.
266
Armin Burkhardt
ing up the notion of a person having or writing them. In both cases the metonymy is the result of focussing on a central element of the frame in question. The only difference seems to be that sentences containing propositional metonymies are more complete in the sense that they literally express a proposition which is logically part of the proposition intended and therefore can be true or false itself. This, however, does not affect the idea that sentences containing metonymies express propositions which can be true or false. From a certain ideal of syntactic and semantic completeness metonymical sentences may be considered semantically defective because they represent one (the intended) frame element by another which is closely associated with it. From the pragmatic point of view, however, they may even turn out to be richer and, therefore, superior to more explicit versions. Imagine e.g. Lotte not saying “Klopstock” but something like “Remember the atmosphere expressed in Klopstock’s well-known ode. The present situation somehow seems to resemble it.” How unromantic that would be! However, as long as the (cognitive) propositions conveyed by metonymical sentences are referentially unscathed and can unequivocally be inferred, metonymies of any kind can certainly be part of true (or false) statements. 6. Conclusion It has been shown that traditional rhetoric has not distinguished the tropes clearly enough from one another. This holds for metonymy and synecdoche in particular. While synecdoche, when looking more carefully at it, reduces to the mere quantitative relations on the similiarity axis, metonymy proves to be the product of qualitative frame-contiguity. In this sense, literary metonymy can be understood as creative displacement supporting a certain poetic style and leading to interesting or even surprising associations. In every-day language and with the use of non-verbal signals, however, the economic aspect of focused shorthand is predominant. Metonymy is “name exchange” and (conceptual) ellipsis at the same time and from this it draws its creative force. Studies in meaning change led to the distinction between semasiological and onomasiological metonymies which has proved to be helpful for historical semantic and contrastive analysis. Metonymy, as this chapter (sic!) has tried to point out, is not simply a rhetorical trope, which it is as well, but an essential semantic principle of language and of other semiotic systems and of the cognitive processes they are connected with. As it must be explained as a special form of ellipsis which deliberately keeps cognitively present what semantically is left out, sentences containing it can certainly be true or false.
Metonymy as a semantic principle
267
References Adelung, Johann Christoph: Grammatisch-kritisches Wörterbuch der hochdeutschen Mundart mit beständiger Vergleichung der übrigen Mundarten, besonders aber der Oberdeutschen. Dritter Theil, M – Scr. Leipzig: Breitkopf 1798, zweyte vermehrte und verbesserte Ausgabe. Austin, John Langshaw: How to Do Things with Words. The William James Lectures delivered at Harvard University in 1955. Oxford: Clarendon Press 1962. Bally, Charles: “L’arbitraire du signe. Valeur et signification”, in: Le Français Moderne 8 (1940), 193–206. Barcelona, Antonio: Metaphor and Metonymy at the Crossroads. A Cognitive Perspective. Berlin–New York: Mouton de Gruyter 2000. Beaugrande, Robert-Alain de, Wolfgang Dressler: Introduction to Text Linguistics. Harlow, Essex: Longman 1981. Benn, Gottfried: Sämtliche Werke. Stuttgarter Ausgabe. In Verbindung mit Ilse Benn herausgegeben von Gerhard Schuster. Bd. 1. Stuttgart: Klett-Cotta 1986. Büchner, Georg: “Der Hessische Landbote”, in: Büchner, Georg: Werke und Briefe. Mit einem Nachwort von Fritz Bergemann. München 1973, 8. Aufl., 133–43. – “The Hessian Courier”, in: Büchner, Georg: The Complete Plays. Edited and introduced by Michael Patterson. London: Methuen 1987, 231–40. Burkhardt, Armin: “Wie die ‘wahre Welt’ endlich zur Metapher wurde. Zur Konstitution, Leistung und Typologie der Metapher”, in: Conceptus. Zeitschrift für Philosophie 21 (1987), Heft 52, 39–67. – “Vom Nutzen und Nachteil der Pragmatik für die diachrone Semantik”, in: Busse, Dietrich (ed.): Diachrone Semantik und Pragmatik. Untersuchungen zur Erklärung und Beschreibung des Sprachwandels. Tübingen 1991, 7–36. – “Die Semiotik des Umberto ‘von Baskerville’”, in: Burkhardt, Armin, Eberhard Rohse (eds.): Umberto Eco. Zwischen Literatur und Semiotik. Braunschweig: Ars & Scientia 1991a, 29–89. – “Können Wörter lügen?”, in: Universitas. Zeitschrift für interdisziplinäre Wissenschaft 47 (1992), 831–40. – “Geballte Zeichen. Das Symbol und seine Deutungen”, in: Zeitschrift für Semiotik 18/4 (1996a), 461–82. – “Politolinguistik. Versuch einer Ortsbestimmung”, in: Klein, Josef, Hajo Diekmannshenke (eds.): Sprachstrategien und Dialogblockaden. Linguistische und politikwissenschaftliche Studien zur politischen Kommunikation. Berlin–New York: de Gryuter 1996b. – “Interjektion”, in: Gerd Ueding (ed.): Historisches Wörterbuch der Rhetorik. Bd. 4: Hu – K. Tübingen: Max Niemeyer 1998, Sp. 202–04. – “Semiotik II. Philosophisch-linguistisch”, in: Krause, Gerhard, Gerhard Müller (eds.): Theologische Realenzyklopädie. Bd. 31. Berlin–New York: de Gruyter 2000, 116–34. – “Metalepsis”, in: Ueding, Gerd (ed.): Historisches Wörterbuch der Rhetorik. Bd. 5: L – Musi. Tübingen: Max Niemeyer 2001, 1087–99. – Das Parlament und seine Sprache. Studien zu Theorie und Geschichte parlamentarischer Kommunikation. Tübingen: Max Niemeyer 2003. – “Sprache und Fußball. Linguistische Annäherung an ein Massenphänomen”, in: Muttersprache. Vierteljahrsschrift für deutsche Sprache 116 (2006), 53–73. Cicero, Marcus Tullius: Orator. Lateinisch-deutsch. Ed. Bernhard Kytzler. München– Zürich 1988, 3., durchges. Aufl.
268 –
Armin Burkhardt
[Cicero] Ad C. Herennium. De Ratione Dicendi (Rhetorica ad Herennium) with an English translation by Harry Caplan. Cambridge, Mass.: Harvard University Press–London: William Heinemann 1981. Dijk, Teun A. van: Macrostructures. Hillsdale, N.J.: Erlbaum 1980. Dirven, René: “Metaphor as a basic means for extending the lexicon”, in: Paprotté‚ Wolf, René Dirven (eds.): The Ubiquity of Metaphor: Metaphor in Language and Thought. Amsterdam–Philadelphia: John Benjamins 1985, 85–120. Dubois, Jacques, Francis Edeline, Jean-Marie Klinkenberg, Philippe Minguet, François Pire, Hadelin Trinon: Allgemeine Rhetorik. Übersetzt und herausgegeben von Amin Schütz. München: Wilhelm Fink 1974 (= UTB128). Duden: Deutsches Universalwörterbuch. Herausgegeben und bearbeitet vom Wissenschaftlichen Rat und den Mitarbeitern der Dudenredaktion unter der Leitung von Günther Drosdowski. Mannheim–Wien–Zürich: Dudenverlag 1989, 2., völlig neu bearb. u. stark erw. Aufl. Eco, Umberto: Einführung in die Semiotik. München: Wilhelm Fink 1972 (= UTB 105), 5. Aufl. 1985. – Semiotik und Philosophie der Sprache. München: Wilhelm Fink 1985. Freiligrath, Ferdinand: Gesammelte Dichtungen. Bd. 1. Stuttgart: Göschen 1871. Goethe, Johann Wolfgang von: Herausgegeben im Auftrage der Großherzogin Sophie von Sachsen. I. Abtheilung. Bd. 14, 19 und 42.2. Weimar 1887–1918. – Faust. Translated by George Madison Priest. New York: Alfred A. Knopf 1941. – The Sorrows of Young Werther. Translated by Michael Hulse. London: Penguin 2006. Göttert, Karl-Heinz: Einführung in die Rhetorik. Grundbegriffe – Geschichte – Rezeption. Stuttgart 1991. Grimm, Jacob und Wilhelm: Deutsches Wörterbuch. Bd. 3. E–Forsche. Leipzig: Hirzel 1862. Henne, Helmut: “Wort und Wortschatz”, in: Duden. Grammatik der deutschen Gegenwartssprache. Herausgegeben und bearbeitet von Günther Drosdowski in Zusammenarbeit mit PeterEisenberg, Hermann Gelhaus, Helmut Henne, Horst Sitta und Hans Wellmann. Mannheim–Wien–Zürich: Dudenverlag 1995, 5., völlig neu bearb. u. erw. Aufl., 540–89. – Sprachliche Erkundung der Moderne. Mannheim–Leipzig–Wien–Zürich 1996: Dudenverlag. Herder, Johann Gottfried: “Essay on the Origin of Language”, in: Rousseau, JeanJacques: Essay on the Origin of Languages, Herder, Johann Gottfried: Essay on the Origin of Language. Translated, with afterwords, by John H. Moran and Alexander Gode. Introduction by Alexander Gode. New York 1966: Frederick Ungar, 85– 176. Jakobson, Roman: “Two aspects of language and two types of aphasic disturbances”, in: Jakobson, Roman: Selected Writings. Vol. II: Word and Language. The Hague–Paris: Mouton 1971, 239–66. Knop, Sabine de, René Dirven, Ning Yu, Birgit Smieja: Bibliography of metaphor and metonymy. Amsterdam–Philadelphia: John Benjamins 2005. (Available electronically at: http://www.benjamins.com/online/met.). Konerding, Klaus-Peter: Frames und lexikalisches Bedeutungswissen. Untersuchungen zur linguistischen Grundlegung einer Frametheorie und zu ihrer Anwendung in der Lexikographie. Tübingen: Max Niemeyer 1993. Lakoff, George, Mark Johnson: Metaphors We Live By. Chicago and London: University of Chicago Press 1980.
Metonymy as a semantic principle
269
Lausberg, Heinrich: Handbuch der literarischen Rhetorik. München: Max Hueber 1973, 2., durch einen Nachtrag vermehrte Aufl. Le Guern, Michel: Sémantique de la métaphore et la métonymie. Paris: Larousse 1973. Lodge, David: Nice Work. Harmondsworth/Middlesex: Penguin 1988. Minsky, Marvin: “Frame-System Theory”, in: Johnson-Laird, P., P.C. Wason (eds.): Thinking: Readings in Cognitive Science. Cambridge: Cambridge University 1977, 355–76. Nerlich, Brigitte: “Synecdoche. A trope, a whole trope, and nothing but a trope”, in this volume, 299–321. – , David D. Clarke: “A Dynamic Model of Semantic Change”, in: Journal of Literary Semantics 17 (1988), 73–90. – “Serial metonymy. A study of reference-based polysenisation”, in: Journal of Historical Pragmatics 2(2) (2001), 245–72. – , David D. Clarke, David D. Zazie Todd: “‘Mummy, I like being a sandwich’. Metonymy in language acquisition”, in: Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. Amsterdam–Philadelphia: John Benjamins 1999, 361–83. Nietzsche, Friedrich: On Truth and Lies in a Nonmoral Sense (1873). In: Keith Ansell Pearson and Duncan Large (eds.): The Nietzsche Reader. Oxford: Blackwell, 2006, 114–23. Noppen, Jean-Pierre van, S. de Knop, R. Jongen: Metaphor. A Bibliography of Post1970 Publications. Amsterdam–Philadelphia: John Benjamins 1985. – , Edith Hols: Metaphor II. A Classified Bibliography of Publications 1985 to 1900. Amsterdam–Philadelphia: John Benjamins 1990. Nübold, Peter: “Falsche Freunde im fachsprachlichen Lexikon”, in: Börner, Wolfgang, Klaus Vogel (eds.): Wortschatz und Fachsprachenerwerb. Bochum: AKS-Verlag 1993, 200–12. Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. Amsterdam–Philadelphia: John Benjamins 1999. – , Linda L. Thornburg (eds.): Metonymy and Pragmatic Inferencing. Amsterdam– Philadelphia: John Benjamins 2003 (Pragmatics and Beyond New Series 113). Paul, Hermann: Deutsches Wörterbuch. Bedeutungsgeschichte und Aufbau unseres Wortschatzes. 10., überarbeitete und erweiterte Aufl. von Helmut Henne, Georg Objartel und Heidrun Kämper. Tübingen: Max Niemeyer 2002. Peirsman, Yves, Dirk Geeraerts: “Metonymy as a prototypical category”, in: Cognitive Linguistics 17(3) (2006), 269–316. Plett, Heinrich F.: Textwissenschaft und Textanalyse. Semiotik, Linguistik, Rhetorik. Heidelberg: Quelle & Meyer 1979 (= UTB 328), 2., verb. Aufl. Quintilianus, Marcus Fabius: The Orator’s Education. Books 1–10. Edited and translated by Donald A. Russell. Cambridge, Mass.–London: Harvard University Press 2001. Seto, Ken-ichi: “Distinguishing Metonymy from Synecdoche”, in: Panther, KlausUwe, Günter Radden (eds.): Metonymy in Language and Thought. Amsterdam– Philadelphia: John Benjamins 1999, 91–120. Trier, Jost: Der deutsche Wortschatz im Sinnbezirk des Verstandes. Von den Anfängen bis zum Beginn des 13. Jahrhunderts. Heidelberg: Carl Winter 1931. Ullmann, Stephen: The Principles of Semantics.Oxford: Blackwell–Glasgow: Jackson nd & Son 1957, 2 edition. Vietta, Silvio, Hans-Georg Kemper: Expressionismus. München: Wilhelm Fink 1975 (= UTB 362).
270
Armin Burkhardt
Warren, Beatrice: “Aspects of referential metonymy”, in: Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. Amsterdam–Philadelphia: John Benjamins 1999, 121–35. – Referential Metonymy. Stockholm: Almqvist & Wiksell Internationa 2006.
Antonio Barcelona (Córdoba, E)
Metonymy in conceptualization, communication, language, and truth
0. 1. 2. 2.2 2.1 2.3 3. 4.
Introduction The concept of metonymy used in the present article The ubiquity of metonymy Metonymy and communication Metonymy in conceptualization Metonymy and linguistic structure Metonymy and truth Conclusions
0. Introduction The aim of this paper is to survey briefly the role of metonymy in thought, communication, language and truth, from a cognitive linguistic perspective. The bulk of the article (section 3) is devoted to reviewing the numerous areas, linguistic and nonlinguistic, in which metonymy can be claimed to play a major role (following Barcelona 2002b). The issue of the connection between metonymy and truth will be the topic of a brief section towards the end of the essay (section 4). Before addressing the central topics of the essay, it is important to outline my own conception of metonymy. This is the purpose of the following section. 1. The concept of metonymy used in the present article Cognitive linguistics regards metonymy as a cognitive mechanism (Lakoff & Johnson 1980, Lakoff 1987), not merely as a figure of speech, as claimed by traditional rhetoric, nor as a mere “contextual effect”, as claimed by relevance theorists (Papafragou 1996). It has only recently begun to be studied systematically by cognitive linguists (Goossens et al. 1995, Panther & Radden 1999, Barcelona 2000a). However, there is not a uniform notion of metonymy shared by all cognitive linguists. Some cognitive linguists (e.g. Croft 1993) require every metonymy to be connected to an act of reference, whereas for most others (e.g. Lakoff & Johnson 1980), metonymy is primarily referential, which implies that there can also be non-referential metonymies. Most ex-
272
Antonio Barcelona
plicit or implicit definitions of metonymy regard it as a mapping within the same domain. Some others (Langacker 1993, Kövecses & Radden 1998) draw attention to the fact that the source provides “mental access to” or “activates” the target. The term “metonymy” is used in cognitive linguistics to cover some very different phenomena, including “classical” or “prototypical” examples of linguistic referential metonymies for individuals, “clear” or “typical” instances of referential metonymies with non-individuals as targets, “clear” or “typical” non-referential metonymies, and relatively “peripheral” instances of metonymy. But all cognitive linguists agree in stressing the primarily conceptual nature of metonymy, which metonymy shares with metaphor. It would take a long article to discuss at length all the particular variants of this basic conceptual approach to metonymy that can be found in the cognitive linguistics literature, but this is not the aim of this chapter. Such a discussion can be found in Barcelona (2003a). Therefore, I will simply present and briefly discuss here my own cognitive-linguistic conception of metonymy. This is a broad notion capable of capturing what all of the abovementioned different phenomena have in common. I call this the schematic notion of met-onymy, which contains the necessary and sufficient conditions for met-onymicity: Metonymy is an asymmetrical mapping of a conceptual domain, the source, onto another domain, the target. Source and target are in the same functional domain and are linked by a pragmatic function, so that the target is mentally activated. (Adapted from Barcelona 2002b: 246) To this definition, we may add the specification that the mapping has to be asymmetrical (i.e. not a systematic matching of counterparts, as in metaphor, see Barcelona (2003a). Some comments on this definition may be necessary. As noted above, metonymy does not have to be referential. Some examples of non-referential metonymies will be provided below. The term mental activation means that the source is a reference-point providing mental access to the target (Langacker 1993, Kövecses & Radden 1998). Mapping refers to the fact that the source domain is connected to the target domain by imposing a perspective on it.1 Functional domain (Barcelona 2001, 2002a) is equivalent to Fillmore’s
—————
1
In the sentence Picasso is not easy to appreciate. PICASSO’S ARTISTIC WORK is a metonymic target, and its activation is carried out from the source Picasso, in his role as ARTIST, with the result that the hearer/reader is invited to conceptualize this artistic work primarily as the outcome of Picasso’s artistic genius – as an extension of his personality –, other aspects of this work being backgrounded (see Barcelona 2003a).
Metonymy and truth
273
(1985) concept of frame and to Lakoff’s (1987) concept of ICM (Idealized Cognitive Model).2 The expression pragmatic function refers to a fundamental property of metonymy, namely the fact that the source maps onto and activates the target by virtue of the experiential (hence pragmatic) link between the roles each of them performs in the same “functional domain”. This is why Fauconnier (1997: 11) regards metonymy as a “pragmatic function mapping”. A “pragmatic function” (Fauconnier 1994, 1997) is a strong built-in connection between roles in a frame or ICM (CAUSE–EFFECT, AUTHOR–WORK, AGENT– ACTION, etc.). If the link between the source and the target is a pragmatic link, then this link is conceptually contingent, hence in principle defeasible. Panther and Thornburg (2003b: 3, 7) point out this property to constrain Kövecses and Radden’s definition (1998), which simply characterizes metonymy as a cognitive process in which the source provides “mental access” (i.e. activates) the target, within the same “cognitive model” (i.e. within the same functional domain; see above). Panther and Thornburg note that not every activation of a target domain by a source domain within the same frame or ICM constitutes a metonymy. In the sentence The piano is in a bad mood, the subject noun phrase codes a real metonymic source for the metonymic target THE MUSICIAN PLAYING THE PIANO by virtue of the pragmatic function linking an INSTRUMENT to an AGENT within the ACTION frame; however, this activation is defeasible, since “the relationship between the piano and the piano player is contingent; the presupposition ‘There is a piano’ does not entail ‘There is a player’”. On the other hand, in the sentence The loss of my wallet put me in a bad mood the activation of the notion of NON-POSSESSION is not metonymic, since the relationship between LOSS and NON-POSSESSION within the POSSESSION frame is conceptually necessary; the proposition presupposed by the referring expression The loss “entails ‘I did not have my wallet for some time span beginning at time t’” (both citations taken from Panther & Thornburg 2003b: 4). In Barcelona (2003a), a set of additional specific definitions are proposed for the other general kinds of metonymy represented by each of the various different phenomena which are covered by the schematic definition. Some examples of the range of such phenomena covered by the definition are the following: ————— 2
In Barcelona (n.d.) and Barcelona (2002a) I have recently proposed that the cognitive domain mentioned in the definition should be a “functional cognitive domain” (i.e. a frame or ICM), and not just a taxonomic domain. In both papers, and in Barcelona (in press), I have also proposed that the mapping in metonymy is unidirectional and asymmetrical, whereas the one in metaphor is unidirectional and symmetrical. By “symmetrical” I mean that each source element has in its frame a structurally equivalent role to its counterpart in the target (e.g. in the LOVE IS A JOURNEY metaphor, the lovers have a role in the “romantic love frame” which is structurally equivalent to the role of the travelers in the “journey frame”).
274
Antonio Barcelona
(1) This book weighs two kilograms. (2) This book is highly instructive. (3) Belgrade did not sign the Paris agreement. (4) She’s just a pretty face. (Lakoff & Johnson 1980: 37) (5) He walked with drooping shoulders. He had lost his wife. In example (1), the whole domain BOOK can be argued to be mapped onto its sub-domain PHYSICAL OBJECT, which is thus mentally activated. In (2), the whole domain BOOK is mapped onto its sub-domain SEMANTIC CONTENT, which is thus mentally activated. Examples (1) and (2) would be “peripheral” or “purely schematic” instances of metonymy, whose target is a primary or near primary domain – in Langacker’s (1987: 165) sense of “primary domain”– included in the source domain. Example (3) is an instance of what I call “prototypical” metonymy (the classical instance of metonymy), since it is referential and since it has an individual (the government of Serbia and Montenegro) as target. Examples (4) and (5) are simply instances of what I call (simply) “typical” metonymies, since the noun phrases where they occur are not referential; furthermore, the target in (5) is not an individual, but a relation (an emotional state).3 The metonymic status of typical and prototypical metonymies is not controversial, because the target remains neatly distinct from the source, that is, it is not included in the source or it is a secondary domain (again, in Langacker’s sense of “secondary”) in it.4 Purely schematic, typical and prototypical metonymies constitute a continuum of metonymicity. Purely schematic metonymies are contextual semantic values often arising in the “literal” use of expressions, a fact which points to the artificiality of a strict, rigid literal-figurative distinction. Any of these general types of metonymy may become linguistically conventionalized, on the basis of a number of factors that cannot be discussed here (see Barcelona 2003a). As we can see from these examples, metonymy would be, under the above schematic definition, a very common, in fact omnipresent, phenomenon in a great many linguistic expressions. This broad conception of metonymy is, in fact, not exceptional in cognitive linguistics. A similar conception underlies, for instance, Langacker’s notion of “active zone” metonymies (Langacker ————— 3
4
What is conventionally believed to be a possible behavioral effect of sadness (walking with drooping shoulders) activates its cause (the emotion itself), so that an automatic inference is that the person exhibiting this bodily behavior was sad. In (4), the target (PERSON) is not a part of the source (FACE), but the other way around. In (5), it is not clear that an effect is a part of the source or viceversa. They might perhaps be regarded as two not mutually including parts or subdomains of the same generic frame, i.e. the CAUSATION FRAME. But an emotion and its effects can also be viewed as elements in another generic frame (the EMOTION FRAME, and more specifically, the SADNESS FRAME). From this latter perspective, the metonymic target in (5) would not be a part of the source either.
Metonymy and truth
275
1993, 1999). Examples (1) and (2) are “active zone” metonymies in which the “active zone” of the notion BOOK is different in each case. It is, thus, hardly surprising that the metonymic status of examples (1) and (2) should be controversial with many scholars. Therefore, virtually all the examples of metonymy I will present in this article will be either typical or prototypical, that is, they will be phenomena like (3), (4) or (5), and not phenomena like (1) or (2). The reason for this is that one of my two aims in this paper is that even if one assumes an uncontroversial, relatively constrained notion of metonymy, the latter still turns out to be pervasive in conceptualization, language and communication (the other goal is, as stated at the beginning, to discuss its connection with truth). Two final important properties of metonymy must be mentioned to round up this brief sketch of my notion of metonymy. Both of them have been convincingly put forward by Panther and Thornburg (2003b: 4–7): (i) Metonymy is not restricted to nominal metonymies (i.e. it is not just an NP phenomenon). Metonymies can also be predicational (as in example 5, where metonymy operates on the predicate walked with drooping shoulders), propositional (the whole proposition presupposed by a clause is a metonymic source for a different proposition),5 and illocutionary (in this case they connect a direct to an indirect speech act). An example of an illocutionary metonymy is provided by the illocutionary force of the utterance I would like you to close that window, which is that of a request (‘Please, close that window’). The request is activated metonymically (PRECONDITION OF ACTION [THE SPEAKER WANTS THE ADDRESSEE TO BEHAVE IN A CERTAIN WAY] FOR ACTION [THE SPEAKER REQUESTS THE ADDRESSEE TO BEHAVE IN THE 6 DESIRED MANNER]). ————— 5
6
According to Panther and Thornburg (2003b: 4), in propositional metonymies, referential nominal metonymies and predicational metonymies combine so that the whole proposition presupposed by the clause is metonymic. Their example is General Motors had to stop production, under the reading ‘General Motors stopped production’. The predicational metonymy is OBLIGATION TO ACT FOR ACTION, which accounts for the reading of had to stop production as ‘stopped production’. The referential nominal metonymy links General Motors to the executive officers of the company. The metonymic connection between this precondition and the speech act, i.e. the linguistic action of requesting, occurs in the conceptual frame or ICM for directive speech acts, which is a schematization or idealization of the basic experiences constituted by these acts (requesting, ordering, suggesting, etc.). Such a schematization, according to Thornburg and Panther (1997) stipulates two preconditions (the speaker wants the addressee to behave in a certain way; the addressee is capable to behave in that way – i.e. to follow a certain course of action or to abstain from it), a central or core part (the addressee is put under the obligation to observe the desired behavior), and a final part (the addressee behaves in the way the speakers wishes). Against the conceptual background of the directive speech act scenario, the indirect speech act performed by means of the utterance I would like you to close that window can, therefore, be argued to be guided or facilitated, among
276
Antonio Barcelona
ii) The link between a metonymic source and its target may vary in strength. This strength depends on how conceptually “close” source and target are to each other. When they are conceptually distant, the link is weak and the metonymic connection is also weak. So in Washington is insensitive to people’s needs, the metonymic connection between WASHINGTON and U.S. POLITICAL INSTITUTIONS is fairly strong. But in The Potomac is insensitive to people’s needs, if the NP The Potomac (which designates the river flowing through Washington) is intended as a reference to the political instititions, the metonymic connection between the concepts POTOMAC RIVER and U.S. POLITICAL INSTITUTIONS is more indirect, hence weaker. The relative strength of a metonymic link is one of the factors facilitating the conventionality of a metonymic expression. The metonymic reading of The Potomac would be unconventional because, in order to arrive at it, the listener would have to remember the connection of the river to Washington, and the connection of the latter to the political institutions; that is, it would require the activation of the metonymic chain POTOMAC RIVER → WASHINGTON → U.S. POLITICAL INSTITUTIONS. 2. The ubiquity of metonymy Given the limitation of space, I will only be able to discuss briefly the relevance of metonymy in some of the many areas of cognition, communication and linguistic structure in which it plays an important role (for a somewhat different survey, see Burkhardt 1996, or Radden 2005). 2.1. Metonymy in conceptualization The metonymic motivation of metaphor The ubiquity of metonymy can be assessed, first of all, by looking at the interaction between metaphor and metonymy. There are a large number of studies whose results point to the fact that the metaphorical extension of innumerable lexemes takes place only on the basis of certain specific aspects of their “literal” sense, and that the metaphorical sense only operates with respect to the same aspects. This, in my view, is equivalent to saying that both the metaphorical source and the metaphorical target are perspectivized metonymically. An example of these studies is Rudzka-Ostyn’s (1995) historical analysis of the metaphorical semantic extensions of English verbs of answering into other domains. The author pointed out that under a broad (i.e. not necessarily referential) notion of metonymy “any extension affected by abstraction, meta————— other factors, by the metonymic connection between one of the preconditions and the whole directive speech act scenario.
Metonymy and truth
277
phoric or not, can be seen as involving a metonymic dissociation” (ibid.: 241). In the field of semiotics, Eco (1985) argues that metaphor depends on metonymy in that it abstracts a feature or features which the two domains have in common, i.e. the ground. Similar views were held by the behaviorist psychologist B.F. Skinner (1957). Dirven (1985) studied 24 senses of the English word cup, which developed over time from the original sense of a prototypical cup with its typical shape and function (drinking). In most cases the extensions due to metonymy and synecdoche preceded those due to metaphor (on the relation between synecdoche and metaphor, see Nerlich, this volume, 299–321). What is most important, though, is that, as Dirven (1985: 103) noticed, in all the metaphorical extensions (12 meanings out of the 24) “some characteristic features of the concept cup are in their entirety or partial aspects transferred to other domains”. All of the metaphorical extensions consisted of a transfer of some aspect of the cup, that is, they presupposed a metonymic understanding of the cup.7 For instance, in acorn cup, the domain of cups is mapped onto the domain of “natural formations”; only the “overall shape” sub-domain of prototypical cups is relevant for this metaphorical mapping. That is, other aspects or sub-domains of cups are not transferred to acorns: for instance, the fact that cups are conventionally used for drinking, or their material constitution. This choice of metaphorical source is imposed by the shape of acorns, which evokes that of a cup. The metonymic motivation of metaphor is discussed in detail in Barcelona (2000b), which includes a brief survey of the relevant literature, and in Radden (2000), which lists the main conceptual areas that give rise to metonymybased metaphors. The tendency for metaphor to be conceptually motivated by metonymy seems to be the rule, rather than the exception.. Even seeming exceptions to this rule can be shown to have a metonymic motivation. The examples of synesthetic metaphors, which, according to Taylor (1995: 139–40), have no metonymic motivation, can be shown, on the contrary, to be based on one or more metonymies: (6) That’s a loud color. The sense of loud encountered in this example is described by The Oxford nd English Dictionary (2 edition, 1989) like this: “4. Of colors, patterns, dress, etc.: Vulgarly obtrusive, flashy. Opposed to quiet.” The metaphor manifested in this example can be described as DEVIANT COLORS ARE DEVIANT SOUNDS. That is, a sound which violates an implicit social norm is mapped onto a color which violates another implicit social norm. The domain of deviant colors, like that of deviant sounds, is characterized by one typical “sub—————
7
But Dirven himself did not classify this selection of aspects as a metonymic operation.
278
Antonio Barcelona
domain”: the main effect caused on the perceivers by the percept, that is, by these (gaudy, eye-catching, often vulgar) colors. This effect is that of forcing themselves on their attention (note the term “obtrusive” in the O.E.D. definition of this sense of loud). Deviant colors are thus understood primarily as attention-getting deviant colors, and deviant sounds are understood primarily as attention-getting deviant sounds. This primary conceptualization of deviant colors on the basis of their main effect constitutes a “metonymic model” of the whole category. There also exists a metonymic model of deviant sounds with attention-getting deviant sounds as a metonymic center. The fact that a similar metonymic center is shared by deviant sounds and deviant colors makes the former an ideal metaphorical source for the latter. (For more details on this and other purported counter-examples, see Barcelona 2000b.) Cases like the mapping of loud sounds onto loud colors or the mapping of cups onto the cupulate involucre in which the acorn grows represent one of two major types of metonymic motivation of metaphor. In this type, both the metaphorical target and the metaphorical source are conceptualized metonymically from the same source (attention in deviant colors and sounds, shape in cups and acorns). The other major type consists of a generalization or decontextualization of a metonymy involving either the metaphorical target or the metaphorical source, or both. There is no space for a detailed discussion of this type, but a metaphor like MORE IS UP (manifested in The high cost of living, Skyrocketing prices, etc.) is based on the metonymic activation of the notion of quantity (MORE) by means of the activation of a level of verticality, due to their frequent experiential correlation.8 Metonymy-based prototypes Lakoff (1987, chap. 5) discusses and illustrates in some detail some of the main types of category models organized around a metonymy-based prototype: social stereotypes, typical examples, ideals, submodels, paragons, generators, and salient examples. These metonymic models, as he calls them, are innumerable and very frequent in cognition and language. I can only mention and briefly comment here on two of his examples. In metonymic models, the category prototype is a sub-category acting as a metonymic “reference point” (Langacker 1993) for the whole category. An example discussed by Lakoff is the metonymic stereotypical model of MOTHER, with the “housewife-mother” (i.e. the mother who is a full-time housewife) standing, as a prototype, for the whole category. Another example ————— 8
This frequent metonymic connection can be seen in the dialogue A: How much gas did you buy? B: I filled her up, in which the answer activates an amount via the mention of the level of height reached by the gasoline in the gas tank. This example, by the way, is metonymically very rich (Barcelona 2002b). Or when one says: A: Have you got enough coffee? B: Yes, my cup’s full (i.e. it has lots of coffee).
Metonymy and truth
279
is the BACHELOR stereotype (which highlights such properties of the stereotypical behavior of bachelors as dating a lot, frequenting single bars and being only interested in sexual conquest). This prototype for the category licenses such examples as (7) Mary’s husband is a real bachelor. This sentence is not a contradiction in the metonymic model. It would be one in the “standard”, “rigid” BACHELOR model, with necessary and sufficient conditions for membership (its prototype is human, male, adult, unmarried, and has reached, in the light of the social conventions prevailing at the time of utterance, a marriageable age). The very frequent occurrence of similar examples shows that metonymy has a fundamental function in categorization, namely, that of organizing an entire category in terms of one or a set of its subcategories. Metonymy-based prototype categorization guides reasoning, as the preceding example shows: what is a logical contradiction in terms of the standard BACHELOR model, ceases to be so in the metonymic model. Metonymy-based elements and links within categories The categorizing role of metonymy is not restricted to its role in the motivation of countless prototypes. It further interacts with propositional models (e.g. Fillmore’s frames; see Fillmore 1985) to provide some of the multiple conceptual links between a prototype and the other category members. Emotion categories have been studied in detail in this respect (see e.g. Lakoff & Kövecses in Lakoff [1987: 380–415]; Kövecses 1986, 1990, 1995; Barcelona 1986, 1992, 1995). The EFFECT-FOR-CAUSE metonymies arising on the basis of the folk models of the physiological and behavioral effects of emotion contribute a number of basic elements (properties, scales, end-points, entities, etc.) of the prototype scenarios for the various emotions. And they motivate the major metaphors that contribute most of the other elements and internal links of the cognitive categories for the various emotions in English and other languages (Barcelona 2000b). In many other categories, like marriage (Quinn 1987), metonymy also seems to provide a great many elements and links. Metonymy and conceptual integration The theory of conceptual integration or “blending” was put forward by Gilles Fauconnier and Mark Turner (see e.g. Turner & Fauconnier 1995, and, for a recent presentation of the theory, Fauconnier & Turner 2002). It is a general theory of conceptual mapping and of its linguistic manifestations geared to account for on-line meaning construction in discourse. It adds to the standard
280
Antonio Barcelona
source and target domains, which are called “input spaces”, and which are not necessarily restricted to two, another two types of spaces: a “generic” space with elements which are common to both input spaces and which facilitate their correlation, and a “blended” space, which takes elements from both inputs and includes new conceptual structure of its own. According to the proponents of the theory, the emergent structure in the blend is not necessarily predictable from the inputs. An example of conceptual integration is the conventional symbol of Death as the Grim Reaper (Turner & Fauconnier 2000): a skeleton holding a scythe and wearing a cowl. This symbol arises by blending many input spaces, including the following: a space with a human individual dying, a space with a prototypical human killer, and a space with reapers in the scenario of a harvest. In one of the input spaces, there is a metonymy connecting death with the priest’s cowl, via the Christian ritual for extreme unction, which requires the presence of a priest. This is then a “long distance” metonymy. One of the “optimality principles on integration networks” proposed by Turner and Fauconnier (2000) is the so-called “metonymy projection constraint”. This constraint reads: “When an element is projected from an input to the blend and a second element from that input is projected because of it metonymic links to the first, shorten the metonymic distance between them in the blend” (Turner & Fauconnier 2000: 139). In the Grim Reaper blend, the conceptual distance between Death and the cowl is very small, as the cowl is the attire of Death in the blend. Coulson and Oakley (2003) add many more similar examples. Ruiz de Mendoza (2000) further claims that metonymy is also often instrumental in the process of correlation leading to the creation of a generic space, a suggestion which I fully subscribe to. And, in my view, metonymy may also determine the selection and suppression of the conceptual structures to be projected from the inputs to the blend. 2.2. Metonymy and communication Metonymy, symbolism and iconic gestures Metonymy motivates directly or indirectly the meaning and the form of many types of symbols. Its motivating role of numerous linguistic symbols (not only words) will be reviewed later on. Its motivational role is important in paralinguistic and non-linguistic symbols. A metonymic mapping lies behind the creation, and also often behind the use and comprehension, of a very large number of these conventional symbols. This is so because in most cases a symbol is a particularly salient sub-domain or aspect of the domain, concept or object symbolized. An example might be the Cross as the symbol for Christianity (Lakoff & Johnson 1980: chap. 8). The cross was the instrument for torture on which Christ, the founder of Christianity, died after terrible suffering. Thus, the cross, a sub-domain in the cognitive domain of Christian-
Metonymy and truth
281
ity, provides mental access, i.e. activates, the entire domain. Even if the Cross had not been officially established by the Church as a symbol of Christianity, it would have been a natural symbol of Christ’s death – hence, indirectly, of Christianity –, anyway. Once a symbol becomes conventional, it functions as such even if its historical roots are no longer known by its interpreter. Non-Christian people who have no knowledge of the Gospel, but who have noticed the prominent place occupied by this symbol in Christian churches, ceremonies and inscriptions, still conventionally identify the Cross with Christianity, because the cross is a sub-domain in their limited “idealized cognitive model” of Christianity. That is, a metonymic motivation for the symbol is also at work in these cases. Iconicity often has a metonymic basis, too. Let us consider a simple iconic gesture, as in example (8) (8) When she saw the cake, she [GESTURE] The speaker interrupts his utterance right after the word she, and makes an iconic gesture consisting in joining his fingertips (as if picking up the cake) and carrying them near his mouth, which may or not be open. The gesture is intended to convey the information that the person in the story (she) picked up and ate the cake. In this gesture, a subset of a complex action sequence stands for the entire sequence. The sequence consists of these actions: picking up a piece of food with the fingertips, carrying it near the mouth, putting it into the mouth, chewing it, and finally swallowing it. Examples like (8) seem to show that a great number of instances of iconic gestures necessarily involve a metonymic mapping. Metonymy, pragmatic inferencing and discourse The role of metonymy in pragmatic inferencing will be given particular attention, as metonymy seems to lie at its very heart. As Panther and Thornburg (1998) claim, metonymies are “natural inference schemas”. I provide two detailed examples of metonymy’s role in the derivation of implicatures. In narrative and / or descriptive discourse it is quite common to mention or to allude only to certain aspects, or certain subsets, of the event sequence to be narrated or of the entities or the scenario to be described. Jane lives alone in a small cottage and prepares her breakfast every morning before dawn, at 5.30 a.m., and then goes to work. Today is a working day for Jane, but this morning no light comes through her kitchen window at 5.30 a.m. It is now 6.00 a.m. and her kitchen is still dark. A female neighbour of Jane’s who knows she lives alone and is aware of her morning habits makes the following casual remark to another female neighbour, who shares the same background information about Jane with the first:
282
Antonio Barcelona
(9) I got up at 5.30 a.m. and saw no lights coming through Jane’s kitchen window. I wonder whether she’s all right. Let us now study the implicatures arrived at by the listener in the comprehension of this utterance, and the metonymies guiding these inferences.9 The listener automatically perceives the connection between the speaker’s two sentences (I got up at 5.30 a.m. and saw no lights coming through Jane’s kitchen window and I wonder whether she’s all right), thanks to conversational implicature (a) (a) ‘she’s not been preparing her breakfast’, which prompts, among others,10 the following further implicatures: (b) ‘maybe she hasn’t got up’ and (c) ‘maybe she isn’t all right’. Implicature (a): ‘Jane has not been preparing her breakfast’. Metonymy: The mention of a salient aspect of the scenario (the fact that Jane has to turn on the light in her kitchen at 5.30 a.m. every day) metonymically evokes (activates) the whole breakfast-preparing scenario. We have here an instance of the metonymy PRECONDITION FOR WHOLE SCENARIO. Jane’s turning on the light in the kitchen is a precondition for Jane being able to prepare her breakfast; i.e., it is a precondition in the highly specific scenario JANE PREPARING HER BREAKFAST (the main part of the larger scenario for Jane’s morning habits), which includes the preconditions (turning on the kitchen light and doing this at 5.30 a.m.), the central part (preparing the breakfast and consuming it), and maybe a final part (tidying up the kitchen after breakfast). Now, by what we might call metonymic reasoning, the neighbour infers that if it is true that Jane did not satisfy the precondition, then it might also be true that the remaining actions included in the scenario did not take place. Negating the proposition that Jane turned on the light in her kitchen at 5.30 a.m. invites negating that she prepared her breakfast. ————— 9
10
An initial, very brief analysis of this text was presented in Barcelona (2002 b), and a fuller analysis was presented at the International Workshop The Interrelation between Cognitive and Discourse Approaches to Metaphor and Metonymy, Castelló de la Plana, Spain, April 3–4, 2003. Other possible inferences, which were analysed in my Castelló presentation, are ‘so maybe she hasn’t gone to work’, and ‘this is quite uncommon with her’.
Metonymy and truth
283
Since the other two inferences are factually and/or conceptually dependent on JANE PREPARING HER BREAKFAST, the negation of this proposition automatically brings about the negation of the propositions in these inferences (“maybe she hasn’t got up”, “maybe she isn’t all right”, etc.). Implicature (b): ‘Maybe she’s not got up’. Metonymy: A condition for being able to turn one’s kitchen light on and to prepare one’s breakfast is to rise from bed previously. Thus, the activation of JANE NOT TURNING ON HER KITCHEN LIGHTS activates JANE NOT GETTING UP (RESULT FOR CONDITION). Implicature (c): ‘Maybe she isn’t all right’. Metonymy: A condition for being able to get up is to be physically fit to do so. So the activation of the notion JANE GETTING UP causes the activation of its condition JANE BEING PHYSICALLY FIT (RESULT FOR PRECONDITION). And activation of the opposite RESULT (JANE NOT GETTING UP) causes activation of the opposite CONDITION (JANE NOT BEING PHYSICALLY FIT). The conceptual connection among the implicatures is presented below, next to the metonymies guiding them: ‘Jane has not been preparing her breakfast’ PRECONDITION (JANE (NOT) TURNING ON THE KITCHEN LIGHTS) FOR WHOLE SCENARIO (JANE (NOT) PREPARING HER BREAKFAST) (+ metonymic reasoning) ‘Maybe she hasn’t even got up’ RESULT (JANE NOT PREPARING HER BREAKFAST) FOR CONDITION (JANE NOT GETTING UP)
‘Maybe she isn’t all right’ RESULT (JANE NOT GETTING UP) FOR CONDITION (JANE NOT BEING PHYSICALLY FIT)
As can be seen, the inferencing work leading to the recognition of the pragmatic link between the two major sentences in the utterance is crucially guided by the chaining of one PRECONDITION FOR SCENARIO metonymy with two RESULT FOR CONDITION metonymies. In my view, metonymic connections within frames and domains regularly provide automatic pathways in pragmatic inferencing (see Barcelona 2003b). Another example that could be added is the following anecdote, which, apparently circulated in the thirties in Spain:
284
Antonio Barcelona
(10) Opposition M.P. (referring to the Premier): But what can we expect, after all, of a man who always uses silk underpants? The Premier (calmly standing up): “Oh, I would never have thought the Right Honorable’s wife could be so indiscreet!” In Barcelona (2003b) I have analysed the complex pattern of implicatures invited by this exchange, and I have found that all of them are motivated by metonymy. There is no space to discuss them all here. Let me just point out at least some of the implicatures and their metonymic basis: Implicatures: 1. The M.P.’s wife shares a secret with the Premier. 2. She knows that the secret consists in the fact that the Premier always uses silk underpants. 3. She has seen the Prime Minister undress. 4. She has had a sexual affair with the Premier, and is, thus, an adulteress (main inference). Metonymic basis: Inference 1 arises on the basis of the metonymy RESULT (being discreet/indiscreet) FOR CONDITION (knowing a secret). You can only be discreet/indiscreet if you know a secret. Inference 2 arises on the basis of the metonymy ENTITY (the propositional entity consisting of the fact that the Premier uses silk underpants) FOR ONE OF ITS CONVENTIONAL PROPERTIES (being secret). Information about people’s choice of underwear is supposed not to be disclosed to strangers. This metonymy identifies this information as the secret that the M.P.’s wife has imprudently disclosed. Inference 3 arises on the basis of the metonymy FACT (knowing the underwear used by a person) FOR ONE OF ITS CONVENTIONAL EXPLANATIONS (having seen the man undress). She has had access to this intimate information because she has seen the Prime Minister undress. The first part of inference 4 also arises on the basis of the metonymy FACT (seeing someone undress) FOR ONE OF ITS CONVENTIONAL EXPLANATIONS (having a sexual affair with that person). One of the explanations for the fact that a woman has seen a man undress is that she has had, or has been about to have, a sexual encounter with that man. The second part of this inference is due to the metonymy DEFINITION (a married woman having sex with a man other than her husband) FOR DEFINED (the behavioral category called ADULTERY). Spelling out the definitional properties of a category (of behavior, in this case) can automatically invoke the category. The role of metonymy in pragmatic inferencing and in discourse is generally recognized in cognitive linguistics. Lakoff’s discussion of the role of metonymy in certain conversational conventions of Ojibwa and of English, and in their intended meaning, was one of the earliest contributions in this
Metonymy and truth
285
direction (Lakoff 1987: 78–9). Other linguists have underlined the metonymic basis of the implicatures motivating metaphorical lexical extension in grammaticalization processes. An oft-cited example is the extension of the spatial meaning of the verb go to the domains of time, volition and prediction, in its grammaticalization as the semi-auxiliary be going to (Hopper & Traugott 1993; Heine, Claudi & Hünnemeyer 1991). Panther & Thornburg (1998), Thornburg & Panther (1997), and the papers in Panther & Thornburg (2003a), are important pieces of research into the metonymic basis of speech acts and other types of pragmatic inferencing, including so-called explicatures (Ruiz de Mendoza & Pérez Hernández, 2003). The relevance of metonymy in discourse understanding can be discovered in many other phenomena. Gibbs (1994: chap. 7) reviews some of them and argues for the psychological reality of the role of metonymy in these phenomena: conceptual tautologies, the mental reconstruction of texts previously read or heard, indirect speech acts, the interpretation and reasoning of eponymous verb phrases, euphemisms, film, drama and art conventions and techniques, th even literary styles, like that of most 19 century historians, which can be 11 defined as metonymic. 2.3. Metonymy and linguistic structure Metonymy and lexical meaning Metonymy-based conceptual metaphors, categories and pragmatic inferences are normally reflected in lexical and grammatical meaning. It is hardly necessary to argue in detail for the fundamental role of metonymy in lexical semantic extension and in lexical polysemy, as this role has been widely recognized for a long time by a number of historical semanticists (see e.g. Stern [1931], or Darmesteter [1932]; see Nerlich 1992, 1998). Some examples of metonymy-induced lexical meaning extension were provided in earlier sections of this article. The conceptual links in lexical polysemy chains are very often metonymic in nature. For instance, some of the nonprototypical senses of over studied by Claudia Brugman (1988) are motivated by a metonymic mapping, according to Taylor (1995: 127–30), who correctly notes that what Lakoff calls “image-schema transformations” in his reanalysis of Brugman’s data (Lakoff 1987: 440–4) are in fact metonymic mappings. An example: (11) (a) Sam walked over the hill (A specific variant of the prototypical sense of over; this variant sense profiles a complete path with an ascending part and a des—————
11
Because they often study a particular event as a model for the whole macrocosm of which they are just a part.
286
Antonio Barcelona
cending part and provides the link to the sense in the sentence be low.) (b) Sam lives over the hill (A non-prototypical sense, linked to the preceding sense via a metonymy of the type WHOLE [the complete path profiled by over in sentence (a)] FOR PART [end-point of the path].) Metonymy and phonology The two remaining sub-sections in this survey of the ubiquity of metonymy will be concerned with its role in the two basic traditional “levels” of linguistic form: phonology and grammar. In his excellent chapter on prototype effects in phonology, John Taylor (1995: 222–39), following Nathan (1986), suggests that on the basis of a phonological theory like Daniel Jones’s (1969),12 it is possible to describe a phoneme as a prototype category consisting of a chain of allophones organized around a central, prototypical allophone. An example would be the link between the central or prototypical allophone of the phoneme /t/ in English (which is articulated as a voiceless aspirated alveolar plosive), on the one hand, and the glottal stop, which lacks the alveolar closure and is a possible realization of /t/ in certain phonetic contexts. The link, according to Taylor, would be the glottalized t-allophone; this allophone, typical of the speech of southern England, occurs in syllable-final position in a stressed syllable and includes alveolar closure and glottalization. As pronounced by those speakers, this allophone is often perceived, in certain phonetic environments, as a mere glottal stop. A possible, very tentative explanation for the development and persistence of this partial allophone is as follows. Within the overarching cognitive category constituted by the phoneme /t/, the alveolar closure element in glottalized t mentally activates the central allophone, the plosive alveolar t. This is possible, in my view, because a part of this central allophone (namely, the alveolar closure present both in glottalized t and in plosive alveolar t) activates the whole of it when a glottal t is uttered. This metonymic connection, thus, highlights the substructure shared by both glottalized t and plosive alveolar t and facilitates recognition of the former as somehow similar to the latter, thus as a member of the same “family” (the phoneme /t/). Similarly, the glottal stop allophone activates, as a part, the glottalized t allophone, thus highlighting their similarity. This indirect link with glottal “t” ————— 12
Jones regarded phonemes as “a family of sounds consisting of an important sound of the language (generally the most frequently used member of that family) together with other related sounds which ‘take its place’ in particular soundsequences or under particular conditions of length or stress or intonation” (Jones 1969: 49).
Metonymy and truth
287
facilitates recognition of the glottal stop as a (peripheral) member of the t phoneme. The other chains linking directly or indirectly the other allophones to the central one could also be facilitated by metonymy, which probably plays the same linking role in other phoneme categories. In fact, Nathan (1996) suggests that conceptual connections akin to metaphor and metonymy may play an important role in the structure of phonological categories. Metonymy and grammar There are a very large number of grammatical processes and phenomena whose form and/or whose meaning are motivated or constrained by metonymy. Langacker (1999: 67) argues that though “usually regarded as a semantic phenomenon, metonymy turns out to be central and essential to grammar”, and that grammar is “a rich source for the investigation of metonymy. At the same time, a recognition of its prevalence and centrality is critical not just for describing grammar but for a realistic assessment of its basic nature”. I cannot agree more. Drawing on recent research in the field and on their own research, Ruiz de Mendoza and Pérez Hernández (2001) and Ruiz de Mendoza and Otal Campo (2002) survey the wide-ranging interaction between metonymy and grammar. These surveys show that metonymy “underlies and largely motivates” the “semantic import” of some grammatical choices and the “conventional value” of some grammatical constructions (though they also show that certain grammatical phenomena place constraints on the nature of some metonymic operations). Some of the areas where they claim metonymy has a motivating role are the use of “predicates” (in Simon Dik’s sense), noun-verb conversion, various other types of conversion and re-categorization, nominalization, valency extension and reduction, the functioning of certain “predications” (again in Simon Dik’s sense), modality, and anaphora. A few examples found in the published research of several linguists on some of these areas will illustrate the pervasiveness of metonymy as a motivating force in grammatical form and meaning. One of them is the ubiquity of “active zone” metonymies (see e.g. Langacker 1991: 189–201; 1993: 33–5), a type of WHOLE FOR PART metonymies required for the interpretation of clauses with certain relational predicates (e.g. verbs like begin, or “raising” sequences like be likely). The interpretation is that one or more of the elements of the clause are interpreted metonymically as a reference-point to its active zone, which may be one of its physical or abstract parts, another experientially associated entity, or another relationship in which it is involved. Take example (12), borrowed from Langacker (1993: 33):
288
Antonio Barcelona
(12) Zelda began a novel The verb begin implies some kind of activity, so the NP a novel stands metonymically for its “active zone” with respect to Zelda’s action of beginning the novel. That active zone is the unmentioned relationship “write/read a novel”, that is, an activity in which the metonymized direct object is involved. This example and the other numerous examples of active zones pointed out by Langacker in his writings show that even sentences which at first sight seem cognitively simple may imply one or more metonymic mappings. The metonymic nature of active zone phenomena is explicitly claimed by Langacker (1999: 62). The ubiquity of metonymy in grammatical structure is thus undeniable. The frequent (transient or permanent) conversion of English proper names into common nouns (as shown by their ability to take number morpheme, determiners, e.g. articles, and restrictive modifiers) is normally licensed by an underlying metonymy (Barcelona 2004). An example is the grammatical reclassification of the name Picasso as a common noun (John has five authentic Picassos), which is possible on the basis of the metonymy AUTHOR FOR WORK. Another, more complex instance is Mike is an Einstein, meaning that Mike is a great scientist, is motivated by the metonymy IDEAL MEMBER FOR CLASS; that is, Einstein, as a paragon (Lakoff 1987: 87) for the class of scientists can stand metonymically for the whole class. The generic metonymy EFFECT FOR CAUSE is responsible, according to Panther and Thornburg (2000), for the use of some stative predicates in dynamic constructions. In (13) (13) He asked her to be his wife. which is paraphrased by Panther and Thornburg as “He asked her to act in such a way so as to become his wife”, the relationship “be someone’s wife” is a stative relationship which is normally caused by a previous action or series of actions; that is, it is a “resultant state”, and as such it constitutes an excellent metonymic reference point for its implicit causal action. This metonymic connection licenses its use in a dynamic grammatical context like He asked her to ... Compare the oddness or unacceptability of using, in the same dynamic grammatical context, a non-resultant state, which cannot be a metonymic source for a causal action: (14) ?*He asked her to be tall. Dirven (1999: 275–87) shows that noun-verb conversion in English often depends on three major sets of metonymies: PATIENT, INSTRUMENT OR MANNER FOR ACTION; GOAL FOR MOTION; and CLASS MEMBERSHIP FOR DESCRIPTION (or, to use Dirven’s term, class membership for the whole “essive schema”). Some of Dirven’s examples:
Metonymy and truth
289
(15) He was angling. (INSTRUMENT FOR ACTION) The plane was forced to land in Cairo. (GOAL FOR MOTION) Mary nursed the sick soldiers. (CLASS MEMBERSHIP FOR DE13 SCRIPTION) In my view, the conversion of the noun nurse into the verb nurse is perhaps more accurately described as an instance of AGENT (NURSE) FOR TYPICAL ACTION (TO NURSE). Finally, two brief examples of metonymy-induced anaphor. Langacker (1999: chap. 7 and 9) claims that the only real factors constraining the application of anaphora are conceptual rather than syntactic in nature. Drawing in part on work by Van Hoek (1997), he pays special attention to instances in which the real antecedent is a metonymic target in the “dominion” of a reference point currently active in discourse. For instance, in (16) He speaks excellent French even though he’s never lived there. the antecedent of there is the metonymic target of French, namely France, as the country is evoked by the language spoken in it; this evocation constitutes a 14 PART FOR WHOLE metonymy. Ruiz de Mendoza and his associates (2000; Ruiz de Mendoza & Pérez Hernández; Ruiz de Mendoza & Otal Campo 2002) also discuss perceptively the role of metonymy in the identification of the antecedents of anaphorical pronouns in instances of conjoined predicates, like (17) In Goldfinger Sean Connery [= James Bond] saves the world from a nuclear disaster, but he [= James Bond] had real trouble achieving it. The metonymy at work here is ACTOR FOR ROLE. According to them, metonymic conceptual anaphors are governed by what they call the “Domain Availability Principle”: only what they call the “matrix domain” of a metonymic mapping is available for anaphoric reference; they claim that in this example, the ACTOR is the matrix domain because it includes the ROLE. Therefore, the anaphor he refers back to Sean Connery, the actor, and via metonymy, to the role, i.e. James Bond.
—————
13 14
That is, the agent in this example is identified via her class membership. A similar example offered by Langacker is I need to call the garage. They said my car is ready. They here refers anaphorically, not to the garage as a place, but to the people working in it. Thus the garage is a metonymic reference point for the people working in it. Conceptual anaphors are very frequent in conversation.
290
Antonio Barcelona
3. Metonymy and truth As a cognitive linguist I am less interested in the relationship between meaning and truth, at least in the traditional correspondence view of truth. Truth is, of course, very important for our daily lives, and a great many of our daily decisions and actions are based on what we perceive to be true (Lakoff & Johnson 1980: 160). However, in cognitive linguistics, the characterization of the conditions under which a given proposition would be “objectively” true are not regarded as the main goal of linguistic semantics. Meaning depends on conceptualization and understanding. Therefore, the characterization of the parameters that determine conceptualization and understanding constitute the real object of linguistic semantics, since these parameters also govern the meaning of linguistic expressions (Fillmore 1985, Lakoff & Johnson 1980: 166–84; Lakoff 1987: passim; Johnson 1987: 194–212; Lakoff & Johnson 1999: 94–129). These parameters are such things as entity structure, orientational structure, dimensions of experience, experiential gestalts, background, highlighting, interactional properties, prototypes, etc. (Lakoff & Johnson 1980: 178–79). The assessent of a proposition as being (relatively) true or false also depends on conceptualization and understanding: “We understand a statement as being true in a given situation when our understanding of the statement fits our understanding of the situation closely enough for our purposes” (Lakoff & Johnson 1980: 179; see also Lakoff & Johnson 1999: 106–18). Now metonymy, as we have seen in section 3, plays a major role in conceptualization (thanks to the existence of metonymic prototypes and of metonymic intracategorial links) and in understanding (thanks to its inferenceguiding power). Therefore, it must be a factor in our understanding-based perception of the relative truth or falsehood of a linguistic expression. In order to illustrate this claim, let us re-examine two of the examples presented in section 3. The first example is (7), repeated below: (7) Mary’s husband is a real bachelor. This example was presented as an instance of reasoning in terms of a metonymic prototype. Is this sentence true or false? Most people would consider it true if Mary’s husband does behave as a stereotypical bachelor, i.e., if, though married, he dates a lot of other women, frequents single bars and is only interested in sexual conquest. That is, the truth value of this sentence is assessed against the background of a metonymic model of bachelorhood, which is in turn based on social experience. The second example is (9), repeated below: (9) I got up at 5.30 a.m. and saw no lights coming through Jane’s kitchen window. I wonder whether she’s all right.
Metonymy and truth
291
For the sake of brevity, only one of the conversational implicatures of this utterance that were pointed out in section 3 is discussed below in terms of truth, namely conversational implicature (a) ‘she’s not been preparing her breakfast’. Both conventional and conversational implicatures are non-truth-functional inferences (in the technical sense of “truth-functional”), according to Allan (2001: 189–202), who also argues that both entailments and conventional implicatures are semantic (i.e. non-defeasible) inferences, whereas conversational implicatures are pragmatic (hence cancellable, defeasible) inferences. The notion of cancellability is closely connected to the notion of truth against the background of the understanding of a particular situation by the participants in a speech event. If a given proposition implicated by an utterance fits their understanding of that situation, the proposition will be regarded as (potentially) true, and if it does not fit their understanding of the situation it will be regarded as (potentially) false. Thus, again, truth is not necessarily connected to “objective” truth. Metonymy-triggered implicature (a) will, then, be deemed to be potentially true if it fits the understanding by the conversation partners, in terms of the specific experience-based JANE PREPARING HER BREAKFAST scenario, of the situation described by the utterance. Since this scenario stipulates that Jane turns on her kitchen light at 5.30 a.m. every morning to prepare her breakfast, if it is actually true that she failed to do so, 15 then it is reasonable to assume that it may be true that she did not prepare her breakfast. Given the PRECONDITION-SCENARIO connection between turning on the lights and the whole breakfast-preparing scenario, the preconditions constitute a natural metonymic reference point for the whole scenario. Metonymy, thus, triggers the implicature, given encyclopedic knowledge, which includes knowledge of the relevant cognitive model (the JANE PREPARING HER BREAKFAST SCENARIO), the cooperative principle, and grammatical knowledge (see Allan 2001: 192). However, if a metonymy-triggered implicature does not fit the understanding of the situation (i.e. if it contradicts it) then it will be deemed false and cancelled. Suppose that after delivering utterance (9) the speaker remembers that Jane had told her the previous evening that she would get up at 6.30 a.m. the following morning, because she had been allowed to start work one hour later than usual on that day. If the speaker delivers that information to her interlocutor after utterance (9), implicatures (a), (b) and (c) would be cancelled (as potentially false), or at least, suspended (until the interlocutor had evidence that Jane had started preparing her breakfast at about 6.30.)16 —————
15 16
The inference that it was Jane that failed to turn the lights is, in fact, a more basic implicature, whose metonymic basis has not been discussed above. The defeasibility of metonymy-based implicatures is due to the fact that metonymy is itself a “pragmatic function mapping”, so that its source-target link is contingent, not necessary (see section 1).
292
Antonio Barcelona
To sum up, metonymy, as a mechanism for conceptual activation, is in principle unaffected by truth value judgments. But since metonymy is one of the major factors in conceptualization, it may determine in part the cognitive context (e.g. by means of a metonymic model) on the basis of which a given sentence will be judged as true or false; on the other hand, metonymy is a major factor guiding pragmatic inferences, which may be confirmed or cancelled on the basis of later information. 4. Conclusions As I hope to have shown in this brief article, metonymy explains numerous facets of linguistic meaning and form, is more basic than metaphor, and at least as ubiquitous in language and thought. The study of metonymy is thus ideally suited to provide evidence of the existence of a continuum between cognition and all aspects of language. On the other hand, the inferences triggered by metonymy have been shown to be normally cancellable, since these inferences are pragmatic in nature. Therefore, their truth value is contingent on their consistency with the interpreter’s understanding (often also dependent on metonymy-based conceptualization) of the particular situation evoked by the utterance that invites them. References Allan, Keith: Natural Language Semantics. Oxford: Blackwell 2001. Barcelona, Antonio: “On the Concept of Depression in American English: A Cognitive Approach”, in: Revista Canaria de Estudios Ingleses, 12 (1986), 7–35. – “El lenguaje del amor romántico en inglés y en español”, in: Atlantis, XIV 1-2 (1992), 5–27. – “Metaphorical models of romantic love in Romeo and Juliet”, in: Journal of Pragmatics, 24-6 (1995), 667–89. – (ed.): Metaphor and Metonymy at the Crossroads. Cognitive Approaches. Berlin– New York: Mouton de Gruyter 2000b. – “On the plausibility of claiming a metonymic motivation for conceptual metaphor”, in: Barcelona, Antonio (ed.): Metaphor and Metonymy at the Crossroads. Cognitive Approaches. Berlin–New York: Mouton de Gruyter 2000b, 31–58. – “The difference between metaphor and metonymy: A question of asymmetry?” (Unpublished manuscript.). Paper presented at the Fourth Conference on Researching and Applying Metaphor. Tunis (Tunisia), 5–7 April 2001 (unpublished). – “Clarifying and applying the notions of metaphor and metonymy within cognitive linguistics: an update”, in: Dirven, René, Ralf Pörings (eds.): Metaphor and Metonymy in Comparison and Contrast. Berlin: Mouton de Gruyter 2002a, 207–77. – “On the ubiquity and multiple-level operation of metonymy”, in: LewandowskaTomaszczyk, Barbara, Kamila Turewicz (eds.): Cognitive Linguistics Today. (Lódz Studies in Language.). Frankfurt/Main: Peter Lang 2002b, 207–24.
Metonymy and truth –
293
“Metonymy in cognitive linguistics. An analysis and a few modest proposals”, in: Cuyckens, Hubert, Klaus-Uwe Panther, Thomas Berg (eds.): Motivation in Language: Studies in Honor of Günter Radden. Amsterdam: John Benjamins 2003a (in press). – “The case for a metonymic basis of pragmatic inferencing: Evidence from jokes and funny anecdotes”, in: Panther, Klaus-Uwe, Linda Thornburg (eds.): Metonymy and Pragmatic Inferencing. Amsterdam–Philadelphia: John Benjamins 2003b, 81–102. – “Metonymy behind grammar: The motivation of the seemingly ‘irregular’ grammatical behavior of English paragon names”, in: Radden, Günter, Klaus-Uwe Panther (eds.): Studies in Linguistic Motivation. Berlin–New York: Mouton de Gruyter 2004, 357–74. Brugman, Claudia: The Story of Over: Polysemy, Semantics and the Structure of the Lexicon. New York: Garland 1988. Burkhardt, Armin: “Zwischen Poesie und Ökonomie. Die Metonymie als semantisches Prinzip”, in: Zeitschrift für germanische Linguistik, 24 (1996), 175–94. Coulson, Seana, Todd Oakley: “Metonymy and conceptual blending”, in: Panther, Klaus-Uwe, Linda Thornburg (eds.): Metonymy and Pragmatic Inferencing. Amsterdam–Philadelphia: John Benjamins 2003, 51–79. Croft, William: “The role of domains in the interpretation of metaphors and metonymies”, in: Cognitive Linguistics, 4-4 (1993), 335–71. Darmester, Arsène: La vie des mots étudiée dans leurs significations. Paris: Librairie Delagrave 1932. Dirven, René: “Metaphor as a basic means of extending the lexicon”, in: Paprotté, Wolf, René Dirven (eds): The Ubiquity of Metaphor. Metaphor in Language and Thought. Amsterdam–Philadelphia: John Benjamins 1985, 85–120. – “Conversion as a conceptual metonymy of event schemata”, in: Panther, KlausUwe, Günter Radden (eds.): Metonymy in Language and Thought. Amsterdam– Philadelphia: John Benjamins 1999, 275–87. Eco, Umberto: “The semantics of metaphor”, in: Innis, R.E. (ed.): Semiotics: an introductory reader. Bloomington, Indiana: Indiana University Press 1985. Fauconnier, Gilles: Mental Spaces: Aspects of Meaning Construction in Natural Language. Cambridge, England: Cambridge University Press 1994. – Mappings in Thought and Language. Cambridge, England: Cambridge University Press 1997. – , Mark Turner: The Way We Think. Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books 2002. Fillmore, Charles: “Frames and the semantics of understanding”, in: Quaderni di Semantica VI (1985), 222–54. Gibbs, Raymond W, Jr.: The Poetics of Mind. Figurative Thought, Language, and Understanding. Cambridge, England: Cambridge University Press 1994. Goossens, Louis, Paul Pauwels, Brygida Rudzka-Ostyn, Anne-Marie SimonVanderbergen, Johan Vanparys: By Word of Mouth. Metaphor, Metonymy and Linguistic Action in a Cognitive Perspective. Amsterdam–Philadelphia: John Benjamins 1995. Heine, Bernd, Ulrike Claudi, Friederike Hünnemeyer: Grammaticalization. A Conceptual Framework. Chicago, etc.: University of Chicago Press 1991. Hopper Paul, Elizabeth Closs-Traugott: Grammaticalization. Cambridge: Cambridge University Press 1993. Jones, Daniel: An Outline of English Phonetics. Cambridge: A. Heffer and Sons, Ltd. th 1969 [1918], 9 ed.
294
Antonio Barcelona
Kövecses, Zoltán: Metaphors of Anger, Pride and Love. Amsterdam: John Benjamins 1986. – Emotion Concepts. New York, etc.: Springer-Verlag 1990. – “Anger: Its language, conceptualization and physiology in the light of crosscultural evidence”, in: Taylor, John, Robert E. MacLaury (eds.): Language and the Cognitive Construal of the World. Berlin–New York: Mouton de Gruyter 1995. Kövecses, Zoltán, Günter Radden: “Metonymy: Developing a cognitive linguistic view”, in: Cognitive Linguistics, 9.1 (1998), 37–77. Lakoff, George: Women, Fire and Dangerous Things. What Categories Reveal About the Mind. Chicago: University of Chicago Press 1987. – , Mark Johnson: Metaphors We Live by. Chicago: University of Chicago Press 1980. Langacker, Ronald W.: Foundations of Cognitive Grammar. Vol 1: Theoretical Prerequisites. Stanford: Stanford University Press 1987. – Concept, Image and Symbol. The Cognitive Basis of Grammar. Berlin–New York: Mouton de Gruyter 1991. – “Reference-point constructions”, in: Cognitive Linguistics 4 (1993), 1–38. – Grammar and conceptualization. Berlin–New York: Mouton de Gruyter 1999. Nathan, Geoffrey S.: “Phonemes as Mental Categories”, in: Proceedings of the Annual Meeting of the Berkeley Linguistics Society, 12 (1986), 212–23. Berkeley: Berkeley Linguistics Society. – “Steps towards a cognitive phonology”, in: Hurch, Bernhard, Richard Rhodes (eds.): Natural Phonology: The State of the Art. Berlin–New York: Mouton de Gruyter 1996, 107–20. Nerlich, Brigitte: Semantic Theories in Europe, 1830–1930: From etymology to contextuality. Amsterdam/Philadelphia: John Benjamins 1992. – “La métaphore et la métonymie: Aux sources rhétoriques des théories sémantiques modernes”, in: Sémiotiques 14 (1998), 143–70. Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. Amsterdam–Philadelphia: John Benjamins 1999. – , Linda Thornburg: “A cognitive approach to inferencing in conversation”, in: Journal of Pragmatics, 30 (1998), 755–69. – “The EFFECT FOR CAUSE metonymy in English grammar”, in: Barcelona, Antonio (ed.): Metaphor and Metonymy at the Crossroads. Cognitive Approaches. Berlin– New York: Mouton de Gruyter 2000, 215–31. – (eds.): Metonymy and Pragmatic Inferencing. Amsterdam–Philadelphia: John Benjamins 2003a(Pragmatics and Beyond New Series). – “Introduction: On the nature of conceptual metonymy”, in: Panther, Klaus-Uwe, Linda Thornburg (eds.): Metonymy and Pragmatic Inferencing. Amsterdam– Philadelphia: John Benjamins 2003b, 1–20. Papafragou, Anna: “On metonymy”, in: Lingua, 99, 4 (1996), 169–95. Quinn, Naomi: “Convergent evidence for a cultural model of American marriage”, in: Holland, Dorothy, Naomi Quinn (eds.): Cultural Models in Language and Thought. Cambridge: Cambridge University Press 1987, 55–8. Radden, Günter: “How metonymic are metaphors?”, in: Barcelona, Antonio (ed.): Metaphor and Metonymy at the Crossroads. Cognitive Approaches. Berlin–New York: Mouton de Gruyter 2000, 93–108. – “The ubiquity of metonymy”, in: Otal Campo, José Luis, Ignasi Navarro i Ferrando, Begoña Bellés Fortuño (eds.): Cognitive and Discourse Approaches to Metaphor and Metonymy. Castellón (Spain): Universitat Jaume I 2005, 11–28. Rudzka-Ostyn, Brygida: “Metaphor, schema, invariance. The case of verbs of answering”, in: Goossens, Louis, Paul Pauwels, Brygida Rudzka-Ostyn, Anne-Marie Si-
Metonymy and truth
295
mon-Vanderbergen, Johan Vanparys (eds.): By Word of Mouth. Metaphor, Metonymy and Linguistic Action in a Cognitive Perspective. Amsterdam– Philadelphia: John Benjamins 1995, 205–43. Ruiz de Mendoza, Francisco: “The role of mappings and domains in understanding metonymy”, in: Barcelona, Antonio (ed.): Metaphor and Metonymy at the Crossroads. Cognitive Approaches. Berlin–New York: Mouton de Gruyter 2000, 109– 32. – , José Luis Otal Campo: Metonymy, Grammar and Communication Albolote (Granada, Spain): Comares 2002 (Colección Estudios de Lengua Inglesa 7). – , Lorena Pérez Hernández: “Metonymy and the grammar: motivation, constraints and interaction”, in: Language and Communication 21.4 (2001): 321–57. – “Cognitive operations and pragmatic implication”, in: Panther, Klaus-Uwe, Linda Thornburg (eds.): Metonymy and Pragmatic Inferencing. Amsterdam–Philadelphia: John Benjamins 2003, 23–49. Skinner, B.F. : Verbal Behavior. New York: Appleton Crofts 1957. Stern, Gustaf: Meaning and Change of Meaning. Göteborg: Eladers boktryckeri Aktiebolag 1931. Taylor, John: Linguistic Categorization. Prototypes in Linguistic Theory. Oxford: Clarendon 1995 [1989]. Thornburg, Linda, Klaus-Uwe Panther: “Speech act metonymies”, in: Liebert, WolfAndreas, Gisela Redeker, Linda Waugh, Linda (eds.): Discourse and perspectives in cognitive linguistics. Amsterdam–Philadelphia: Benjamins 1997 (Current Issues in linguistic theory 151), 205–19. Turner, Mark, Gilles Fauconnier: “Conceptual integration and formal expression”, in: Metaphor and Symbolic Activity, 10 (1995), 183–204. – “Metaphor, metonymy and binding”, in: Barcelona, Antonio (ed.): Metaphor and Metonymy at the Crossroads. Cognitive Approaches. Berlin–New York: Mouton de Gruyter 2000, 133–45. Van Hoek, Karen: Anaphora and Conceptual Structure. Chicago–London: University of Chicago Press 1997.
Brigitte Nerlich (Nottingham, GB)
Synecdoche: A trope, a whole trope, and nothing but a trope?*
0. Introduction 1. The old rhetoric1 2. The figures of speech in historical semantics 3. Jakobson and beyond 4. The new rhetoric 5. Cognitive semantics and the figures of speech 6. Conclusion 0. Introduction Metaphor and metonymy have had two completely unrelated waves of fame th in the 20 century. The first was brought about by Roman Jakobson’s two papers entitled “Two aspects of language and the two types of aphasic disturbances” (1983[1956]; see now Dirven & Pörings, eds. 2002) and “The Metaphoric and Metonymic Poles” (1956). The second, more recent one, was brought about by Mark Johnson and George Lakoff’s book Metaphors We Live By (1980). Jakobson’s papers soon became classical texts of reference for French philosophers, semiologists and linguists interested in figures of speech. Metaphors We Live By became a standard text for those interested in what has become a new paradigm in linguistics, namely cognitive linguistics. In both contexts, though, synecdoche has been subordinated to metonymy and explored as one aspect of metonymy. A renewed interest in synecdoche itself has produced two counter-currants to these two waves of neglect. The first had its source in the work of the Groupe de Liège or Groupe µ (Jacques Dubois, Francis Edeline, Jean-Marie —————
*
1
I would like to thank the Trustees of the Leverhulme Trust for their continued support. Special thanks to all those who have helped me in the preparation of this article: Günter Radden, Ken-ichi Seto, Cynthia L. Hallen, Adolphe Nysenholc, Dirk Geeraerts, Bernard Meyer, and Beatrice Warran. And Sarah who inspired the title over a glass of wine. The title is based on a parody of the oath taken by witnesses in British courts of law who have to swear on the bible to tell ‘the truth, the whole truth and noting but the truth’. In this section I rely heavily on chapters 1 and 2 (vol. 1) of Meyer’s 1993 book on synecdoche.
298
Brigitte Nerlich
Klinkenberg, Philippe Minguet, François Pire, Hadelin Trinon), for whom synecdoche became the master-trope (cf. Groupe µ 1970). Gerard Genette (1970, 1972) and Tzvetan Todorov (1970, 1973, 1977) also contributed to this new counter-currant. Research according to the parameters set out by the Groupe µ continued in France and Belgium for about ten to fifteen years. A collection of French research into synecdoche can be found in the 1983 issue of Le Français Moderne, with contributions by Klinkenberg, Silingardi, Schmitz, Bouverot and Meyer. Bernard Meyer set himself the task of describing “avec minutie chacune des figures que le groupe µ nomme, dans sa Rhétorique générale, métasémèmes, et qui correspondent, en gros, au tropes des auteurs classiques” (Meyer 1983: 346).2 This project came to fruition in 1993 when Meyer published two volumes of work entitled Synecdoques. Etude d’une figure de rhétorique. In English and German speaking countries synecdoche has been reth evaluated at the end of the 20 century. Armin Burkhardt discusses the problems of demarcation between synecdoche and metonymy in his article “Zwischen Poesie und Ökonomie. Die Metonymie als semantisches Prinzip” (Burkhardt 1996; see also Burkhardt, this volume, 247–72). Ken-ichi Seto has devoted a seminal paper to “The Cognitive Triangle: The Relation between Metaphor, Metonymy and Synecdoche”, a paper that heralded a new assessment of synecdoche in cognitive semantics, to which we shall come back in section 5 (Seto 1995). Whereas, as we shall see, classical definitions of synecdoche distinguished between a part–whole and a genus–species type synecdoche, this novel assessment of synecdoche (cf. Seto 1999, 2003) sees metonymy is concerned with reference, whereas synecdoche is concerned with (extensional and intensional) sense. Synecdoche is defined as a categorical transfer phenomenon based on semantic inclusion as conceived by the speaker between a more comprehensive and a less comprehensive category. Hence, synecdoche deals with taxonomical relations only, not with part– whole relations. Based on this new view of synecdoche, Brigitte Nerlich and David Clarke (1999) have discussed the communicative function of synecdoche. Examples of synecdoches under this new definition are, according to Seto, stone for jewel, a genus for species synecdoche used by jewellers and Hoover for any type of vacuum cleaner, as an example of a species for genus synecdoche (for more examples, also of verbs, see Seto (1999). What is astonishing is that neither in their work on metaphor and metonymy nor in the more recent upsurge of interest in synecdoche, French and English scholars seem to be aware of each other’s work. As René Dirven has pointed out in a recent article on metaphor and metonymy, it is surprising that in “the latest very rich literature on metaphor one finds very few references to the epoch-making short paper by Roman Jakobson ‘The Metaphoric and ————— 2
“in every detail each one of the figures of speech which the Groupe µ in its General Rhetoric calls metasememes, and which correspond, roughly, to the tropes of the classical authors.”
Synecdoche
299
Metonymic Poles’ (1956)” (Dirven 1993: 2). Reading through Meyer’s volumes, I was astonished to find no reference to the cognitive tradition, apart from a brief note on Dan Sperber’s early article on a ‘cognitive rhetoric’ (Sperber 1975). In this paper I would like to achieve two goals: To summarise the past and present fate of synecdoche, and to make French linguists aware of the cognitive, and cognitive linguists outside France aware of the French rhetorical tradition, and last, but not least say something about synecdoche and truth. In the following sections I would therefore like to chart the rises and falls of synecdoche in the ‘old rhetoric’, in historical semantics, in Jakobson’s work, in the ‘new rhetoric’ as developed in France, and finally in the framework of cognitive semantics, currently being developed. 1. The old rhetoric3 Traditionally, synecdoche has been classified as one of the tropical figures of speech. Figures of speech themselves were split into two big groups: the tropes and other figures of speech such as construction, elocution and style. The number of tropes listed varies between thirty and two. Towards the end of what one can call the classical rhetorical tradition synecdoche is one figure amongst a minimalist group of four which was then further restricted to three and then two, namely metaphor and metonymy. The number of types of synecdoche also varied during that time, as well as synecdoche’s inclusion within or exclusion form other adjacent tropes. Neither the term trope nor the term synecdoche were as yet used by Aristotle, but in chapter 21 of his Poetics he distinguishes between four classes of ‘metaphors’, two types of which would later be included into the range of synecdoche: genus for species and species for genus. The Stoics developed the idea that there is a necessary relation between the poverty of languages and the need to make words shift from one meaning to another according to a certain number of logical or natural relations, relations that are also used in figurative speech (cf. Meyer 1993: 67). This idea would resurface at the beth ginning of the 19 century and be of seminal importance to historical semantics, by then also influenced by associationist psychology. It would be rediscovered yet again by cognitive semanticists after a period when structural linguistics dominated the semantic scene. Although the term ‘trope’ was still not mentioned in the Rhetorica ad Herennium, published before 82 and attributed (first to Cicero, then) to Quintus Cornificius, one can find in this treatise three figures of speech which were to become the cornerstones of classical rhetoric, namely translatio, which can be equated with metaphor, denominatio which can be regarded as ————— 3
In this section I rely heavily on chapters 1 and 2 (vol. 1) of Meyer’s (1993) book on synecdoche.
300
Brigitte Nerlich
metonymy, and intellectio which can be equated with synecdoche. Intellectio and conceptio were used synonymously with synecdoche in the Latin tradition.The most outstanding and influential author in that tradition was Marcus Fabius Quintilian who was the first to talk about changes in meaning brought about by tropes. He uses the term ‘trope’ in his Institutio Oratoria and defines the figures of speech (Book IX, 1, 4) as a way of speaking in which the natural or original meaning is transferred onto another, so as to adorn speech. He lists three double types of synecdoche (Book VIII, 6, 19): genus for species and vice versa (already encountered in Aristotle’s definition of ‘metaphors’), the part for the whole and vice versa, and the singular for the plural and vice versa (already mentioned by Cornificius and Cicero) (cf. Meyer 1993: 71). And he adds a forth category, later called metalepsis4, which consists in designating what precedes by what follows and what follows by what precedes (cf. ibid.: 72; Burkhardt 2001). It was from Quintilian’s treatise that classical rhetoric took its clues. Looking at the Latin tradition in general, synecdoche was defined on two levels: the locus a maiore ad minus which includes: genus pro specie, totum pro parte, pluralis pro singularis, materie pro opere; and the locus a minore ad maius which includes species pro genere, pars pro toto, singularis pro pluralis (cf. Lausberg 1963: §§ 192–201). As already mentioned above, during the evolution of rhetoric from Aristh totle to the 19 century, the extension of the figure of speech called synecdoche varied enormously. Furthermore, and interacting with these variations in extension, the original proliferation of figures of speech in classical rhetoric was gradually whittled down to four (metaphor, metonymy, synecdoche, irony), then three in what came to be known as rhétorique restreinte (cf. Genette 1970). The three surviving figures of speech were: metaphor, metonymy and synecdoche (later still reduced to two). In this process of reduction, the works of the Frenchmen Dumarsais, Bauzée and Fontanier were central (cf. Todorov 1977: 136). Following Beazée’s entries on the figures of speech for the Encyclopédie Méthodique, Pierre Fontanier distinguished clearly between metaphor, metonymy, and synecdoche in his Manuel classique pour l’étude des Tropes, published in 1821. It was followed by the Traité général des figures du discours autres que les tropes of 1827. Both the manual and the treatise have been reedited in 1968 by Gérard Genette under the title Les figures du discours. In his Manuel Fontanier defines the trope as a procedure whereby one changes the meaning of a word into another, or transports a word’s first meaning onto another meaning. He distinguishes between eight main types of synecdoche, namely the part for the whole and vice versa (types 1 and 2), the material for the object (type 3), the singular for the plural and vice versa (type 4), the genus for the species and vice versa (types 5 and 6), the abstract for the concrete (type ————— 4
Metalepsis is difficult to define and it is even more difficult to find good examples. The best explanation has been provided by Bill Long (2004).
Synecdoche
301
7) and the antonomasias (type 8) (Fontanier 1968: 87–97) (cf. Meyer 1993: 83). In this definition synecdoche has a very broad extension, which in the th course of the 20 century, would be reduced first to types 1, 2, 5 and 6, then to types 5 and 6, with 1 and 2 being included in metonymy. 2. The figures of speech in historical semantics Around the same time as Fontanier wrote the most classical of classical treatises on rhetoric, that is in about 1825, historical semantics was ‘born’. Christian Karl Reisig, a Classical philologist, included semasiology into the study of Latin and borrowed the figures of speech from rhetoric as procedures of semantic change (cf. Nerlich 1992: 37). Reisig wrote: Die Grundlage der Ideenentwicklung in den Wörtern ist die Gedankenassociation in der Gemeinschaft der Vorstellungen. [...] Es sind gewisse Ideenassociationen unter den menschlichen Vorstellungen vorzüglich gebräuchlich, welche mit gewissen Ausdrücken bezeichnet die Rhetorik sich angeeignet hat, welche aber in gewisser Hinsicht auch in die Bedeutungslehre gehören, nämlich die Synekdoche, die Metonymie und die Metapher. So weit diese sogenannten Figuren auf das Ästhetische hinzielen, gehören sie allerdings der Rhetorik an, auch insofern sich Einzelne derselben bedienen; wofern aber in einer besonderen Sprache nach diesen Redefiguren sich ein Redegebrauch gebildet hat, der dem Volke eigen ist, so gehören diese Figuren hierher. (Reisig 1890[1839]: 2)5
Reisig amalgamates here the ‘grammatical’ and the ‘rhetorical’ tradition, a fusion that would bring about ‘historical semantics’. Historical semantics began to flourish as a branch of historical comparative linguistics alongside historical phonetics, its more popular sister. It has to be stressed however that the rhetorical tradition had already begun to merge with the grammatical, especially the etymological one, well before Reisig (cf. Schmitz 1985). Examples are the works of Leibniz, Dumarsais (who wrote grammars as well as his famous treatise on tropes) and of Turgot. Gottfried Wilhelm Leibniz had written for example in his Nouveaux Essais sur l’Entendement humain (1765):
————— 5
“The basis for the development of ideas in words is the association of thoughts in the group of representations. [...] However, certain associations of ideas are preferred in human representation, and rhetoric has given them certain names, which however are to some extent also appropriate for semasiology, namely synecdoche, metonymy, and metaphor. Insofar as these so-called figures aim at something aesthetic, they certainly belong to rhetoric, even when individuals make use of them; but insofar as a usage based on these figures of speech has been established in a particular language, and this is particular to the nation, these figures should be dealt with here.” (All translations by Brigitte Nerlich, unless otherwise indicated.)
302
Brigitte Nerlich [C]e qui fait comprendre en même temps comment les metaphores, les synecdoques et les metonymies ont fait passer les mots d’une signification à l’autre, sans qu’on en puisse toujours suivre la piste. (Leibniz GP 5. 261)6
Anne Robert Jacques Turgot had written in his article on ‘Etymology’ for the Encyclopédie Méthodique (to which Dumarsais and, after his death, Beauzée had also contributed articles on grammar and on the figures of speech): Toutes sortes de tropes & de métaphores détournent la signification des mots; le sens figuré fait oublier peu à peu le sens propre, & devient quelquefois à son tour le fondement d’une nouvelle figure; en sorte qu’à la longue le mot ne conserve plus aucun rapport avec sa première signification. Pour retrouver la trace de ces changements entés les uns sur les autres, il faut connoître les fondements les plus ordinaires des tropes & des métaphores; il faut étudier les différents points de vûe sous lesquels les hommes ont envisagé les différents objets, les rapports, les analogies entre les idées, qui rendent les figures plus naturelles ou plus justes. (Turgot 1784: 22)7
In this quotation, as well as in that of Reisig, one can find the beginnings of a psychologically based historical semantics, which, according to Geeraerts (1995), has many things in common with modern cognitive semantics. Standing more fully in the French tradition inaugurated by Dumarsais, the French linguist Auguste de Chevallet enumerated four mechanisms of semantic change, namely synecdoche, metaphor, metonymy, and metalepsis. In 1853 defined synecdoche as follows: La synecdoque est un trope par lequel un mot, au lieu de l’objet ou du fait qu’il désignait primitivement, désigne un autre objet ou un autre fait en vertu de leur coexistence. On peut distinguer deux sortes de coexistences, l’une physique, qui consiste dans l’union essentielle des objets compris dans un même tout; l’autre catégorique, que nous imaginons dans les différentes classes d’objets ou de faits subordonnés les uns aux autres. (Ibid.: 201)8
————— 6
7
8
“We therefore understand at one and the same time how metaphors, synecdoches and metonymies have transferred words from one meaning to another, without us always being able to follow their exact route.” “All sorts of tropes & metaphors divert the meaning of words; the figurative meaning makes us gradually forget the literal meaning and sometimes becomes in turn the basis for a new figure of speech; in such a way that, in the long run, the words retains no link whatsoever with its first meaning. If we want to find the traces of these accumulated changes we have to know the most ordinary foundations of the figures of speech and of metaphors; we have to study the different points of view which we adopt to look at different things, the relation-ships, the analogies between the ideas and which make the figures more natural and more appropriate.” “The synecdoche is a trope through which a word, instead of designating the object or fact which it designated originally, designates another object or another fact by virtue of their coexistence. We can distinguish between two sorts of coexistence, one physical, which consists in the essential union of objects comprised in
Synecdoche
303
He distinguished between the synecdoche of the genus for the species (which can result in a restriction of meaning (cf. ibid.: 206), the synecdoche of the species for the genus which can result in an extension of meaning (cf. ibid.: 212), the synecdoche of the part for the whole, and of the material for the object. His synecdoches based on physical coexistence would later on be incorporated into metonymy, leaving for synecdoche the type based on coexistence in ‘categorical space’. But soon historical semanticists reduced Reisig’s three-way and Chevlet’s four-way distinctions of mechanisms of semantic change to the dichotomy of metaphor and metonymy. Based on it, they established the classical quartet of mechanisms of semantic change consisting in generalisation and specialisation, metaphor and metonymy (Paul 1880), of which the first duo focuses in fact on the results of semantic changes, the second on the procedures used to bring it about. In 1894 Robert Thomas wrote for example a treatise on the possibilities of semantic change in which he draws on Reisig, Paul, but also on Darmesteter’s (1886: 62ff.) ‘logical types’ of semantic change and establishes the following, rather plausible typology (cf. Kronasser 1968: 33). He distinguishes between semantic changes based on shifts inside the same conceptual sphere, namely specialisation (species pro genere) and generalisation (genus pro specie) and semantic changes based on transfers between conceputal spheres, namely metaphor (based on subjective correspondences) and metonymy (based objective correspondences). As one can see, metaphor and metonymy are used as terms to designate shifts in meaning across conceptual boundaries, based not yet on what later writers call similarity and contiguity but subjetivity and objectivity. Synecdoche, based, according to the modern definition (see sections 0 and 5) exclusively on taxonomical genus–species relations, could have been used as a cover-term for the mechanism that brings about specialisation and generalisation (see Nerlich 1992: 69). Léon Clédat, a French historical linguist and etymologist, who like his colleague, Arsène Darmesteter, used the classical theory of tropes to study semantic changes, would later make a similar distinction and postulate four logical procedures of meaning transfer: extension and restriction on the one hand; connection (synecdoche and metonymy) and comparison (metaphor) on the other (cf. Clédat 1912: ix). One year later, in 1913, the Danish semanticist, Kristoffer Nyrop, who wrote a massive Grammaire historique de la langue française (1913) with a fourth volume on semantics, devoted seven chapters to metonymy, and not one to synecdoche as such (Nyrop 1913; cf. Meyer 1993: 96). This tradition was carried on by Stephen Ullmann (1951) who established the most famous classification of semantic changes, which can be summarised as follows: Ullmann dintinguishes between the nature and the consequences of semantic change: Regarding the nature of semantic change he ————— one and the same whole; the other categorical, which we imagine to exist amongst the different classes of objects or facts subordinated under a category.”
304
Brigitte Nerlich
distinguishes between metaphor: a semantic change based on a similarity of senses, metonymy, a semantic change based on a contiguity of senses, folketymology, a change based on a similarity of names and ellipsis, a change based on a contiguity of names. In terms of consequences of semantic change he lists: widening of meaning, narrowing of meaning, amelioration of meaning and pejoration of meaning. Although Ullmann calls his classification a ‘functional’ one, it is based on strictly Saussurean, that is structural principles. As Geeraerts has pointed out in his book on ‘diachronic prototype semantics’, the distinction between sense transfer and name transfer is a consequence of the Saussurean distinction between signifiant and signifié, whereas the associative mechanisms of similarity and contiguity can be likened to the paradigmatic and syntagmatic relationships of Saussurean theory. (Geeraerts 1995: 82)
Geeraerts goes on to criticise Ullmann’s classification and to propose a new classification for the ‘current mechanisms of lexical change’ (that is, semasiological and onomasiological changes). In this new classification the traditional foursome, that is specialisation, generalisation, metonymy and metaphor, have a prominent place as independent referential semantic changes (cf. Geeraerts 1995: 92). It is astonishing to note that Jakobson fails to mention Ullmann’s classification in his seminal articles published a few years later, as it was perhaps in Ullmann’s book that the term contiguity came first into real prominence for semantic and rhetorical studies. 3. Jakobson and beyond Based on the synchronic linguistics advocated by Saussure and influenced by Freud, Jakobson took Saussure’s distinction between syntagmatic relations (in praesentia) and paradigmatic relations (in absentia) to define metaphor and metonymy respectively. Synecdoche, or rather one sub-type of the older definition of synecdoche (which comprised part–whole and genus–species relations), namely that based on the part–whole relation, is seen as being part of metonymy. Metaphor is based on the paradigmatic operation of selection or substitution, whereas metonymy is based on the syntagmatic operation of coordination or combination. Furthermore, metaphor is said to be based on similarity, whereas metonymy is based on contiguity. So, “whereas similarity exists as a paradigmatic relation that connects entities across different domains of experience, contiguity is a syntagmatic relationship that holds between entities in the same ‘chunk of experience’” (Geeraerts 1995: 94). I shall come back to the modern notion of ‘domain’, used to conceptualise metaphor and metonymy, later on. Let us first see what ‘syntagmatic’ could mean in this instance. When we say: I heard the trumpet we use a metonymy. What we ‘mean’ to say is We heard [the sound of] the
Synecdoche
305
trumpet. When the boss says to the waiter: Bring the bill to the hamburger, he means to say: Bring the bill to [the man eating or having eaten] the hamburger and so on. Is this the ‘syntagmatic’ relation we are looking for? In this case metonymy would be based on something like syntagmatic ellipsis. However, this does not work with all types of metonymies and neither does it work with part–whole ‘synecdoches’, such as there are several new faces in the team, whereby the manager wants to say there are several new [persons with] faces in the team (whereas in the previous sentences the omission of the sound of or the man eating seem entirely natural, persons with seems to be artificially added to the syntagmatic chain only to be omitted again). This means, we either have to say that ‘syntagmatic’ is used by Jakobson in a rather ‘metaphoric’ way, as designating any linguistic or real co-occurrence of entities9, or we can try to tease out the various meanings of ‘syntagmatic’, and clarify these meanings. This has recently been done by René Dirven (1993). Dirven distinguishes between three types of syntagm: the linguistic, linear syntagm, the non-linguistic, conjunctive or sociocultural syntagm and the inclusive syntagm based on a chain of inclusion, which exhibit a gradual increase in figurativeness (Dirven 1993: 10). One could argue that the traditional synecdoche based on the part–whole relationship could be included under the conjunctive, the ‘modern’ one based on set-inclusion or the genus– species relationship, under the inclusive syntagm. Dirven then continues to explore the two other crucial notions of similarity and contiguity which have had a revival in cognitive semantics since Lakoff and Johnson’s (1980) book. Reading through their book as well as Croft’s article (1993) on the role of domains in the interpretation of metaphors and metonymies, it becomes apparent that a clarification of the notion of domain is urgently needed, and that only such a clarification could give us a clear hint as to whether metonymy and/or synecdoche involve one or two domains, or to use Geeraerts’ expression one or two chunks of experience. From Dirven’s analysis it seems to follow that grosso modo metaphor involves the mapping of (contrasting rather than similar) domains, in the process of which these domains are superimposed onto each other and the target domain annihilates the source domain, whereas metonymy involves the linking of (adjacent) domains (or subdomains or parts of domains), in the process of which the domains are added onto each other and constructed as one domain. In metaphor it is important that we see something as something else, in metonymy it is important that we see a contiguity between domains. In metaphor conceptual distance is ex—————
9
As Warren has pointed out, Cooper (1986: 135–9) pinpoints the flaw of Jakobson’s argument, which consists in transferring contiguity from applying between referents of words (between those of, say bottle and champagne) to linear contiguity between words in a sentence. This means that, according to Warren, there are no metonymic relations between let’s and finish and the champagne in Let’s finish the champagne (cf. Warren 1995: 137–8).
306
Brigitte Nerlich
ploited, in metonymy conceptual closeness, and, as we shall see later: in synecdoche (in this case the genus–species type) the relation of inclusion inside conceptual domains is exploited. It is time to come back to Jakobson and his followers. Both Gerard Genette in France and Umberto Eco in Italy continued Jakobson to some extent, as the first provides a rationale for converting metonymy into synecdoche and vice versa (cf. Genette 1972: 27), and as the second, borrowing ideas about semiosis from Peirce, argues for a reduction of metaphor to metonymy on the grounds that language is a chain of semantic associations held together by contiguity (cf. Eco 1973; cf. Schofer & Rice 1977: 129). Genette’s work as author and as editor of Fontanier’s treatise on rhetoric can be seen as the link between the old, classical rhetoric and the new rhetoric which was inaugurated in France by the work of the Groupe µ. In this work synecdoche was to take its revenge on metaphor and metonymy. 4. The new rhetoric From around 1970 onwards the Groupe µ or Groupe de Liège (1970; 1977) argued that synecdoche was the basis for metaphor and metonymy and defined metaphor as a double synecdoche (Groupe µ 1970: chapt. 4, section 2.1). Following to some extent in the Latin tradition (where one distinguished between genus to species, species to genus, whole to part and part to whole synecdoches), the Groupe µ postulated various types of synecdoche. What they call synecdoches of the ‘referential mode’ could be said to exploit what is nowadays called ‘partonomies’; what they call synecdoches of the ‘semantic mode’ could be said to be based on ‘taxonomies’. As we shall see, only the latter are defined as synecdoches by modern linguists, such as Burkhardt and Seto for example, the former ones being attributed to metonymy. According to the Groupe µ, metaphor, metonymy and synecdoche are all metasememes, based, according to Minguet, on “an operation acting on the sememe of a word without altering the level of generality of the utterance. When the effect is taken for the cause, the producer for the product, the material for the object made of it, the number of elements [or semes, B.N.]10 ————— 10
Semes should not be confused with semantic features as they also include encyclopaedic ‘features’. This inclusion of encyclopaedic competence into semantics was criticised by Rwet (1975) and defended by Klinkenberg (1979; 1983). As Meyer points out, Klinkenberg “propose de distinguer, à l’intérieur d’un sémème, le noyau sémique ‘fourni par une compétence proprement linguistique’ et ‘les sèmes dits latéraux [fournis] par le savoir encyclopédique’ [...]. [...] Albert Henry, pour sa part, identifie sémème, concept et représentation des choses” (Meyer 1993:51). It seems that there are a number of overlapping features shared by this approach and the one advocated by modern cognitive semantics, but there is no room here to expand on this matter. For those interested in exploring this issue further, I would like to provide just two quotes as an incentive – and I shall not trans-
Synecdoche
307
deleted or added has no influence on the poetic ethos.” (1986: 550) However, synecdoche has a special status as “in the synecdochical process, the modified term implies a change in the degree of generality of the utterance” (ibid.). An even more special status has to be given to the generalizing (semantic) type of synecdoche. Take the utterance I saw a pachyderm. It may produce “a deviation in a context where the expected term would rather be elephant. Yet this statement is true since an elephant is a pachyderm. By contrast, the other three synecdochical modes always produce false assertions. Saying about pachyderms something that is only true for elephants would be erroneous; reduction only could make the meaning acceptable” (ibid.). We shall see later on why the intrinsic ‘truth’ of a generalising synecdoche of the semantic mode is important for its use in ordinary, rather than poetic, language (see Nerlich & Clarke 1999). The new rhetoric inaugurated by the Groupe µ had many enthusiastic adherents in France, but also a number of more critical followers, such as Henry, Le Guern, Meyer, to mention just a few central figures. They were all not quite as enthusiastic perhaps as Todorov who wrote in 1970 in defence of synecdoche: Tout comme dans les contes de fées ou dans Le Roi Lear où la troisième fille, longtemps méprisée, se révèle être la plus belle ou la plus intelligente, Synecdoque, qu’on a longtemps négligée – jusqu’à ignorer son existence – à cause de ses aînés, Métaphore et Métonymie, nous apparaît ajourd’hui comme la figure la plus centrale. (Todorov 1970: 30)11
In 1973, Michel Le Guern criticised the traditional view of synecdoche, as well as the one adopted by the Groupe µ, and returned to some extent to Jkobson. He devotes a whole chapter of his book Sémantique de la métaphore et de la métonymie to ‘the problem of synecdoche’ (chapt. 3). To cut a long story short, he includes, unlike the Groupe µ, the synecdoche of the part, the whole and the material into metonymy and argues that the synecdoches of genus and species are not tropes at all but belong to the normal functioning of language. This view was adopted by Le Guern’s disciple Marc Bonhomme —————
11
late them, as those who are interested in pursuing this matter should be able to read French. Meyer writes: “Le trope est la manifestation verbale d’une opération mentale largement indépendante du langage, qui, pour l’essentiel, se déroule en deçà et au-delà des mots, au niveau de ce que l’on nomme comunément la pensée et à partir de l’expérience.” (1993: 63) Geeraerts has stressed the importance of “lived experience” (1995: 186) for cognitive semantics, and he also stresses “the unmitigated acceptance of the importance of the psychological aspects of language, and the absence of any attempt to maintain a strict distinction between senses (deemed to belong to the only structurally relevant level of meaning), and (individual, psychological) encyclopaedic concepts” (ibid.: 183). “Just as in the fairy tales or in the King Lear, where the third daughter, long despised, is revealed as the most beautiful and intelligent, so Synecdoche, long neglected and even forgotten because of her elders Metaphor and Metonymy, seems to be nowadays the most central figure of speech.”
308
Brigitte Nerlich
who included the synecdoches of the part and whole amongst the metonymies in his book Linguistique de la métonymie (1987) (cf. Meyer 1993: 105–6). In 1979 Nobuo Sato also expressed doubt as to whether synecdoche was at all a figure of speech. He agrees with Le Guern in incorporating the part– whole synecdoches into metonymy, but he maintains that genus-species synecdoches are genuine figures of speech and sees in them, here following it seems Albert Henry (whose work Métaphore et métonymie he fails to mention), “les seuls mécanismes de la synecdoque proprement dite, irréductible à la métonymie”12 (Sato 1979: 120; cf. Meyer 1993: 109–10). This goes exactly against what Schofer and Rice, also working in the French tradition of new rhetoric, had argued for in their article “Metaphor, Metonymy and Synecdoche (Re)visited”, published in 1977. They had written: Synecdoche is characterized by a semantic or referential relationship of inclusion made possible by the fact that one of the signifieds is also a semantic feature of the other signified. Whereas we eliminated all but causal relationships from metonymy, synecdoche is expanded here to include not only the part for the whole but also the container for the contained. (Schofer & Rice 1977: 141)
However, this opinion can be regarded as a rather isolated deviation from the view that part–whole synecdoches are metonymies, a conception which was by then becoming generally accepted. In the 1980s the Belgian linguist Adolphe Nysenholc took up the gauntlet. He criticised Jakobson, Todorov and the Groupe µ, and drew inspiration from Henry and Sato, but also from an older source, Esnault’s 1925 book on metaphors. Nysenholc rejects the bipolar distinction between metaphor as based on similarity and metonymy (including synecdoche) as based on contiguity. Instead, he sees metaphor as based on similarity, synecdoche as based on similarity and contiguity, and metonymy as based on contiguity. Like Le Guern and other French writers, he regards the synecdoche of the part and whole as metonymy (cf. Nysenholc 1982: 121). Bringing together Jakobson’s distinction between similarity and contiguity, Esnault’s notions of intension and extension, and Henry’s distinction between champ sémique and champ associatif,13 he develops new definitions of metaphor, metonymy and synecdoche. He writes: En effet, il peut y avoir deux types de contiguïté: une contiguïté en compréhension entre caractères ou détails d’un même élément, autrement dit entre sèmes d’un même champ sémique. Et une contiguïté en extension, entre êtres d’un ensemble,
————— 12 13
“the only synecdochical mechanisms properly speaking, which cannot be reduced to metonymy”. Champ sémique cannot be translated by ‘semantic field’ because we are dealing here with the field of semantic and encyclopaedic features (sèmes) that makes up a word or concept or sémème. Champs associatif can more appropriately be translated as ‘semantic field’ in the traditional sense.
Synecdoche
309
entre espèces d’un genre, à savoir encore entre champs sémiques d’un seul champ associatif. (Nysenholc 1981: 311).14
Both metonymy and synecdoche are figures of contiguity, but whereas in the case of metonymy we focus on one aspect or characteristic trait of one class of ‘things’ (exploiting its intension), in using synecdoche we focus on the common characteristics of several classes (exploiting their extension) (cf. ibid. 312). After having clarified what he found implicitly contained in the treatises of his predecessors, Nysenholc proposes a novel distinction between what he calls horizontal and vertical synecdoche. In the case of a horizontal synecdoche we are dealing with the fusion of the contiguous species of one genus (related to each other by function or nature); in the case of a vertical synecdoche we are dealing with the fusion of a species and its genus (comparable to the generalising synecdoche of the semantic type in the classification of the Groupe µ). Examples for the vertical type of synecdoche are relatively easy to find (mortals for human beings and, in the other direction daily bread for daily food). Finding examples for the horizontal type is less easy. As a specialist of Charlie Chaplin’s films which he analyses ‘semiotically’, Nysenholc gives the example of Charlie using a carpenter’s file instead of a nail file to file his nails (1981: 123)15 . He argues that this type of synecdoche is based as much on contiguity as on similarity and that therefore the difference between synecdoche and metaphor is not a categorical one but rather a question of degree. All depends on the proximity of the species that are brought together and the culturally ratified classifications which are exploited (cf. ibid. 124). 5. Cognitive semantics and the figures of speech As we have seen, during the 1970s synecdoche became the central figure of speech in French ‘new rhetoric’, only to be allocated a more modest place in the research that was inspired by it. In English speaking countries, synecdoche would not lose its Cinderella-status until the 1990s. In 1980 Lakoff and Johnson published their seminal book Metaphors We Live By in which they studied synecdoche as one aspect of metonymy. Metaphor and metonymy, based, as in Jakobson’s work, on similarity and contiguity respectively, were once again the belles of the linguistic ball. Lakoff —————
14
15
“In fact there can be two types of contiguity: a contiguity in comprehension between characters or details of one and the same element, in other words between semes of one and the same field of semes. And a contiguity in extension between phenomena of one whole, between species of one genus, that is to say between fields of semes which belong to the same associative field”. Nysenholc studied the visual language of Chaplin-films in his thesis which he defended in 1976.
310
Brigitte Nerlich
and Johnson identified a whole range of regular patterns of metonymy, something also quite common in traditional diachronic semantics (Esnault for example distinguished 38 types), but they also listed a range of metaphorical patterns which, up to now had not been clearly recognised (cf. Geeraerts 1995: 95). The central theme of Metaphors We Live By was, as the title indicates, metaphor. Lakoff and Johnson demonstrated that conventional metaphors do not only permeate language, but that they are at the very basis of our thought. They showed that “whole domains of experience, not just linguistic expressions, are systematically conceptualized in terms of other domains of experience. These systematic cognitive mappings of one domain of experience, the source domain, onto another domain of experience, the target domain, are referred to as ‘conceptual metaphors’” (Radden 1992: 522). Lakoff and Johnson claimed that in metaphor we map domains onto domains whereas in metonymy we stay inside one domain of experience. This issue was taken up by William Croft and Dirven in 1993. Whereas Croft argued that metaphor is based on domain mapping and metonymy on a type of domain highlighting, Dirven, as we have seen, argued that we are dealing with (macro-) domain mapping combined with suppression of the source domain in the case of metaphor, with domain annexation or at least microdomain annexation in the case of metonymy. As to synecdoche, Croft only mentions the part–whole synecdoche which is usually “subsumed under metonymy” (Croft 1993: 350). However, the case of synecdoche lends itself to a reevaluation of the notion of domains. It seems that metonymy exploits the contiguity that exists between the features inside a referential (including partonomical) domain, associated with a concept; synecdoche exploits the horizontal and vertical contiguity between members of a taxonomy inside one taxonomical domain; whereas metaphor, finally, exploits the contrast and/or similarity between different semantic domains situated on similar hierarchical ranks in different taxonomical systems. This is a line of thought followed by Ken-ichi Seto (1995, 1999, 2003) Seto (1995) wishes to off-set Lakoff and Johnson’s claim that our conceptual system is largely metaphorical, by showing that metonymy and synecdoche are, or at least, may be, just as basic. He regards metaphor, metonymy and synecdoche as the three corners of what he calls the “cognitive triangle”, where metaphor is based on similarity, synecdoche on inclusion and metonymy on contiguity. Just as the critics of the Groupe µ in France, he excludes from synecdoche the classical part–whole relation in the physical world and subsumes it under metonymic contiguity, leaving only semantic inclusion to synecdoche, that is the relation inside or between categories. Synecdoche is based on taxonomic relations, whereas metonymy exploits partonomical (Tversky) or meronomic (Cruse) relations. Seto’s definition of metonymy runs therefore as follows:
Synecdoche
311
Metonymy is a real world relation based on contiguity. Contiguity has two kinds: spatial contiguity and temporal contiguity. Spatial contiguity has several major schemas and subschemas: container for content; whole for part, part for whole, etc. Temporal contiguity has several major schemas and subschemas: the preceding event for the following event; the following event for the preceding event; cause(r) for effect; effect for cause(r). (Seto 1995: 3–4)
Synecdoche on the other hand is defined as: Synecdoche is a category relation based on semantic inclusion. Semantic inclusion means the relation between a more comprehensive category and a less comprehensive category. Semantic inclusion has several major schemas and subschemas: genus for species; species for genus; type for token; token for type, etc. (Ibid.: 4)
And, I should add with Nysenholc: species for species. This view of metonymy and synecdoche is endorsed by Armin Burkhardt in an independently written article. Discussing the old question as to whether the exploitation of part–whole relations should be regarded as synecdoche or not, Burkhardt writes: Der alte Streit um die Zuordnung der Teil/Ganzes-Beziehung kann begraben werden, sobald einsichtig wird, daß er sich nur einer Äquivokation verdankt – kann doch pars sowohl ‘Bestandteil’ als auch ‘Teilmenge’ bedeuten. Nur im letzteren Falle darf die pars/totum-Beziehung als eine quantitative und damit synekdochische betrachtet werden, während die Beziehung des materiellen Bestandteils zum Ganzen in die Reihe metonymischer Relationen gehört. (Burkhardt 1996: 178)16
Metonymy is based on qualitative, synecdoche on quantitative relations, that is on set-inclusion. Metonymy is based on our world-knowledge about space and time, cause and effect, part and whole, whereas synecdoche is based on our taxonomic or categorical knowledge. Metonymy exploits our knowledge of how the world is, synecdoche of how it is ordered in our mind. Seto provides the following examples for synecdoche: (1) genus-for-species: mortal (→ a human being) man (‘a human being’ → a male) jeans (→ blue jeans [AmE]) drug (→ marijuana, cocaine) fish and chips (→ ? and potato chips) cf. French fry drink (→ drink alcohol → drink alcoholic beverages) (Seto 1995: 6)
————— 16
“The old quarrel about the whole-part relation can be forgotten as soon as we become aware of the fact that it is only based on an equivocation, for pars can mean ‘component’ as well as ‘subset’. But only in the latter case should we regard the pars/totum-relation as a quantitative and therefore synecdochical one, whereas the relation of a material component to the whole belongs to the set of metonymic relations.”
312
Brigitte Nerlich
(2) species-for-genus to earn one’s daily bread (→ food) walkman (Sony’s Walkman → any personal stereo) sellotape (BrE)/scotch tape (AmE) ( → any sticky thin clear tape) species (and many other brand-names) (ibid.: 7).
In an answer to a question “What’s the term for the generic use of a brand name to describe all similar products?” by a newspaper-reader, Bob Kyff writes in “Word watch” (confusing component with subset, but none the less in an enlightening and entertaining way): “Actually, she mailed me a Xerox of it, and I Scotch Taped it to the Frigidaire”. And he goes on to say: We’re using the part-for-whole type of synecdoche, for instance when we describe a smart person as a “brain,” a car as “Wheels,” or a corporate executive as a “suit.” (And if that executive sues us over generic use of brand names, a suit has brought suit.) Conversely, we’re using whole-for-part type of synecdoche when we call the Schenectady Blue Jays baseball team simply “Schenectady” or the police officer arresting us for trademark violation “the law.” Clearly, referring to an entire product category by a brand name falls into the first type of synecdoche. But “Synecdoche” is much too general a term for this specific practice, and much too hard to pronounce. So I propose a brand new term: “brandonym.” [...] I hereby define “brando-nym” as “a brand name commonly used to designate en entire category of similar products.” (Bob Kyff in The Hartford Courant, Wednesday, August 9, 1995.)17
Back to Seto’s examples (3) type-token: This jacket is our best-selling item. (→ This type of jacket) We both drive a Honda Accord. (→ two tokens of the same type of car) (Seto 1995: 7)
Looking for more examples, I found that what Nysenholc called the vertical type of synecdoche, and in particular the one that substitutes the genus for the species, seems to be the most common type of synecdoche. It is particularly common in any type of ‘official’ register, such as ‘police-speak’ (see Nerlich & Clarke 1999). Exploiting the fact that this type of synecdoche is always true, that it is easier and safer to use a generic term rather than a specific one, one can hear utterances such as I apprehended a person/vehicle on the premises. Synecdoche is used here “as an instrument to avoid introducing too many and especially too specific referents” (Dirven 1993: 22). What is said is therefore true, but it is also vague and therefore not easily disputable. These uses of synecdoche in police-speak may strike us as somewhat odd, but they —————
17
I would like to thank my sister Anita for kindly sending me this article.
Synecdoche
313
are generally not perceived as ‘figures of speech’. We are dealing hear rather with stylistic variations. Other examples are even more innocuous, such as: TV (→ colour TV) fridge (→ electric refrigerator) bag (→ a paper/plastic/hand/shopping bag).
After having discussed these examples, Seto too comes to the conclusion that this type of synecdoche is at one and the same time the most common and the least ‘tropical’ – and I would add the most trivially ‘true’. As Le Guern and Klinkenberg have pointed out, they seem to belong to the normal functioning of ordinary language (Klinkenberg 1983: 291). Seto writes: In advanced countries a TV means a color TV and a fridge means an electric refrigerator. That this is possible is, I suppose, due to the mechanism of the genusfor-species synecdoche. This kind of synecdoche may be related to the so-called maxim of quantity. If this is possible, it seems reasonable to suppose that the genus-for-species synecdoche is not a special figure of speech, but a very general semantic mechanism which governs a wide range of language use including the use of pronouns and pronominal expressions.18 (Seto 1995: 7)
Lausberg had already pointed out in his Elemente der literarischen Rhetorik that the genus-for-species type of synecdoche is especially frequent “um im Kontext die Wiederholung eines die Art bezeichnenden Wortes zu vermeiden (variatio: § 86), wobei der bestimmte Artikel als rückverweisendes Bekanntschafts-Signal (§ 180) auftritt”19 (Lausberg 1963: § 194). Synecdoche can therefore be regarded as one of the mechanisms used to establish coherence in text, and could or should be studied by text-linguists and those studying the phenomenon of coherence in text. An example, randomly chosen from a Sunday colour-supplement, demonstrates this mechanism nicely. In an article on North Korea, we can read: By the Arch of Triumph tiny girls in white ballet tutus went cartwheeling across lay-bys, bigger girls practised gravity-defying gymnastics routines; never had I seen youngsters so supple and fit. Here, 10,000 kids assembled to march with silver spears; there, 20,000 headed for the 120,000-seat Kim Il Sing sports stadium to rehearse a series of dazzling visual tricks with multi-coloured shields. When, finally, we tore ourselves away, the gymnasts kept waving until we had turned the corner and were out of sight. (The Observer, Life magazine, 7 January 1996, 10)
————— 18
19
To include pronouns under the heading of synecdoche might be going too far, as pronouns have a purely grammatical status, acquire meaning in discourse, after which they return to their grammatical status. They don’t change or vary in meaning. (Nysenholc, personal communication). “so as to avoid in a certain context the repetition of a word that designates a species (variatio: § 86), with the definite article serving as an cataphoric signal of acquaintance”.
314
Brigitte Nerlich
Having lost its most tropical part, the part–whole synecdoche, to metonymy, synecdoche thus seems to have lost at one and the same time its status as a trope and become a general semantic mechanism (which could be studied in the context of semantic relations, such as synonymy and hyponymy). The tropical nature of the genus-for-species synecdoche can certainly be revived, as demonstrated in the Queneau-example quoted by the Groupe µ (1970: 103), but still, our much reduced synecdoche is certainly the least tropical of our three master-tropes: metaphor, metonymy and synecdoche. To close this section, I would like to come back to some of my own work: In 1992 David Clarke and I tried to establish what we called a ‘metasemantic expert system’ with metaphor and metonymy as the main procedures of linguistic innovation, including, like so many others,20 synecdoche in our treatment of metonymy (cf. Nerlich & Clarke 1992). In the light of the old and the new texts on synecdoche, which I have analysed, it seems to be necessary to review this demotion of synecdoche to an ancillary role and put it into its right place, next to metaphor and metonymy. In 1992 we wrote: The trick of being innovative and at the same time understandable is to use words in a novel way the meaning of which is self-evident. But there are only two main ways of going about that: using words for the near neighbours of things you mean (metonymy) or using words for the look-alikes (resemblars) of what you mean (metaphor). (Nerlich/Clarke 1992: 137)21
We should perhaps now add: there are only three ways of going about that: using words for the near neighbours of the things your mean (metonymy), using words for the taxonomically related things your mean (synecdoche), or using words for the look-alikes (resemblars) of what you mean (metaphor) (which, to be effective, should neither be closely linked in our world-knowledge or situational knowledge nor closely related in our categorical knowledge). On the other hand, it seems however that synecdoche, reduced as it is, is the least innovative of these procedures, and could therefore be regarded, ————— 20
21
Keller (1995) has argued that metaphor and metonymy are the two basic processes of lexical innovation. Keller shows clearly that there are three basic types of signs (symptoms, icons, and symbols), three procedures by which we can interpret signs (by causal, associative and rule-governed inference), and two procedures of semantic innovation: metonymy, the linguistic counterpart to symptoms, and metaphors, the linguistic counterpart to icons. But where does this leave synecdoche? Keller studies synecdoches under the rubriques of symptoms and metonymy (ibid.: 121). However, he also says that language serves three purposes: categorisation, representation and communication (ibid.: 57), which, to some extent, can be correlated with his three procedures of linguistic interpretation: causal, associative, and rule-governed inference. In the light of what has been said above it seems plausible to claim that the first type of inference should be split into two sub-types: synecdochical or taxonomic inference and metonymic or causal (including partonomic) inference. However, metonymy and metaphor would still be the most basic types of semantic innovation it seems (cf. main text). For a very similar assessment of metaphor and metonymy, cf. Warren (1995).
Synecdoche
315
not as a mechanism of semantic innovation, but as a mechanism of semantic variation rather than innovation or change. It would therefore be right to maintain metaphor and metonymy as the two basic mechanisms of semantic change.22 But what about truth? Unlike in the case of metaphor, truth seems to be a rather trivial question in the context of studying synecdoche in its reduced modern form. Utterances which contain synecdoches are trivially true as the referent always falls under the genus and the species designated by the synecdochally used word. When the coach attendant says Put your towels in the receptacles provided, he uses a genus for species synecdoche and the utterance is always true, as all bins are receptacles. When I say, It’s hard to earn one’s daily bread by writing articles about synecdoche, this utterance (based on a species for genus synecdoche or particularising synecdoche) is true only if what I earn is both the (protoypical) species, bread, and the genus – other food stuffs and beyond. However, this might change if and when people start to live in outer space and have to go without bread and eat protein paste instead. In this case, prototypicality comes into play in the use of synecdoche and the truth value attributed to utterances containing synecdoches. This is also the case for some of the type-token examples given by Seto. 6. Conclusion Looking back at the older and newer definitions of synecdoche one can only agree with Bernard Meyer when he writes that “La catégorie de la synecdoque apparaît donc comme une classe rhétorique d’extension flottante, une nébuleuse de figures variant autour d'un noyau stable.”23 (Meyer 1993: 85) First synecdoche was part of metaphor, then a whole trope in itself with a set of members whose number fluctuated over time, until, more recently, part of it was amalgamated with metonymy, namely the part–whole type of synecdoche. For a very long time the kernel of synecdoche consisted of two types of synecdoches: the part–whole one and the genus–species one, with the part– whole one being the epicentre, as a quick look at some reference dictionaries will confirm. Only recently has this kernel been broken up and one part of it been annexed by metonymy. In some cases the genus–species synecdoche was forgotten altogether in the process of reducing all tropes to two, namely metaphor or metonymy, and sometimes the genus–species part of the kernel was preserved to redefine synecdoche as a third member in a triplet of essential tropes, namely metaphor, metonymy and synecdoche. ————— 22 23
As to whether irony should be added as a third member is a matter for another article! “The category of synecdoche appears therefore to be a rhetorical class of fluctuating extension, a nebula of figures varying around a stable kernel.”
316
Brigitte Nerlich
Looking back at the history of synecdoche, its splendour and misery, one has to conclude that this category of trope exhibits many features commonly attributed to prototypical categories. As Geeraerts has pointed out: First, prototypical categories cannot be defined by means of a single set of criterial (necessary and sufficient) attributes [...]. Second, prototypical categories exhibit a family resemblance structure, or more generally, their semantic structure takes the form of a radial set of clustered and overlapping meanings [...]. Third, prototypical categories exhibit degrees of category membership; not every member is equally representative for a category [...]. And fourth, prototypical categories are blurred at the edges [...]. (Geeraerts 1992: 184)
This definition of prototypical categories, used in cognitive semantics, could, it seems, be used equally successfully in ‘cognitive rhetoric’. In this type of rhetoric it would seem that not only the hypercategory of ‘tropes’ is a prototypical category with boundaries and memberships that fluctuate, but that its most prototypical members, namely metaphor, metonymy, and synecdoche also exhibit all signs of being prototypical categories. They, too, exhibit shifting and fluctuating boundaries and category memberships. However, despite this fluctuation and despite its humble Cinderella status in much of rhetoric and cognitive linguistics, synecdoche might be more important than we think, especially when it comes to ‚truth’. On the conceptual level it reflects and exploits the order in our categories, on the linguistic or structural level it exploits stable and orderly semantic relations, and on the communicational level it brings order into texts and social relations (see Nerlich & Clarke 1999: 216). And what is truth but a reflection of what we think of as order and regularity? Truth might therefore have more in common with synecdoche than with any other trope.
Synecdoche
317
References Bonhomme, Marc: Linguistique de la métonymie. Bern: Peter Lang 1987. Burkhardt, Armin: “Zwischen Poesie und Ökonomie. Die Metonymie als semantisches Prinzip”, in: Zeitschrift für germanistische Linguistik 24 (1996), 175–94. – “Metalepsis”, in: Ueding, Gerd (ed.): Historisches Wörterbuch der Rhetorik. Bd. 5: L–Musi. Tübingen: Max Niemeyer Verlag 2001, 1087–99. Chevallet, Auguste de: Origine et formation de la langue française. 2 parties en 3 vols. Paris: Dumoulin Imprimérie Impériale 1853–7. Clédat, Léon: Dictionnaire étymologique de la langue française. Paris 1912: Hachette. Cooper, David: Metaphor. Oxford: Blackwell 1986. Croft, William: “The Role of Domains in the Interpretation of Metaphors and Metonymies”, in: Cognitive Linguistics 4 (1993), 335–70. Cruse, D.A.: Lexical Semantics. New York: Cambridge University Press 1986. Darmesteter, Arsène: The Life of Words as Symbols of Ideas. London: Kegan Paul, Trench 1886. Dirven, René: “Metonymy and Metaphor. Different mental strategies of conceptualization”, in: Leuvense Bijdragen 82 (1993), 1–28. – , Ralph Pörings (eds.): Metaphor and Metonymy in Comparison and Contrast. Belin–New York: Mouton de Gruyter 2002. Eco, Umberto: “Sémantique de la métaphore”, in: Tel Quel 55 (1973), 25–46. Esnault, Gaston: L’imagination populaire: Métaphores occidentales. Paris 1925. Fontanier, Pierre: Les Figures du discours. Paris: Flammarion 1958. Geeraerts, Dirk: “Prototypicality Effects in Diachronic Semantics: A round-up”, in: Kellermann, Günter, Michael D. Morrissey (eds.): Diachrony within Synchrony: Language History and Cognition. Papers from the International Symposium at the University of Duisburg, 26–28 March 1990. Frankfurt/Main etc.: Peter Lang 1992, 183–204. – Diachronic Prototype Semantics. A Contribution to Historical Lexicology. To be published by Oxford University Press 1995. Genette, Gérard: “La Rhétorique restreinte”, in: Communications 16 (1970). Reprinted in Genette (1972: 21–40). – Figures III. Paris: Eds. du Seuil 1972. Groupe µ [Jacques Dubois, Francis Edeline, Jean-Marie Klinkenberg, Philippe Minguet, François Pire, Hadelin Trinon]: Rhetoric générale. Paris: Larousse 1970. – Rhétorique de la Poésie. Brussels: Ed. Complexe 1977. Henry, Albert: Métonymie et métaphore. Paris: Klincksieck 1971. Jakobson, Roman: “The Metaphoric and Metonymic Poles”, in: Fundamentals of Language. Ed. by Roman Jakobson and Morris Halle. s’Gravenhage: Mouton & Co. 1956, 76–82. – “Two Aspects of Language and Two Types of Aphasic Disturbances”, in: Language in Literature. Ed. by Krystyna Pomorska and Stephen Rudi. Cambridge, Mass.–London: The Belknap Press of Harvard University Press 1983 [1956], 95– 120. Keller, Rudi: Zeichentheorie: Zu einer Theorie semiotischen Wissens. Tübingen– Basel: Francke Verlag 1995. Kellermann, Günter, Michael D. Morrissey (eds.): Diachrony within Synchrony: Language History and Cognition. Papers from the International Symposium at the University of Duisburg, 26–28 March 1990. Frankfurt/Main etc.: Peter Lang 1992.
318
Brigitte Nerlich
Klinkenberg, Jean-Marie: “Le rôle du composant encyclopédique dans une rhétorique linguistique”, in: IIe Congrès de l’Association Internationale de Sémiotique. Vienna 1979. – “Problèmes de la synecdoque. Du sémantique à l’encyclopédique”, in: Le Français Moderne 51 (1983), 289–99. Kronasser, Heinz: Handbuch der Semasiologie; Kurze Einführung in die Geschichte, Problematik und Terminologie der Bedeutungslehre. Heidelberg: Winter 21968 [1952]. Lakoff, George, Mark Johnson: Metaphors We Live By. Chicago–London: University of Chicago Press 1980. Lausberg, Heinrich: Elemente der Literarischen Rhetorik. Eine Einführung für Studierende der klassischen, romanischen, englischen und deutschen Philologie. München: Hueber 1963, 4., durchgesehene Aufl. Le Guern, Michel: Sémantique de la métaphore et de la métonymie. Paris: Larousse 1973. Leibniz, Gottfried Wilhelm: Die philosophischen Schriften von Gottfried Wilhelm Leibniz. Hrsg. von C.I. Gerhardt. Vierter Band. Berlin: Weidmann. (= GP) 1880. Long, Bill: “Metalepsis”, 2004, article available at http://www.drbilllong.com/MoreWords/Metalepsis.html (accessed 1 July, 2008). Meyer, Bernard: “La synecdoque d’abstraction”, in: Le Français Moderne 51 (1983), 346–60. – Synecdoques: Etude d’une figure de rhétorique. Tome I. Paris: Eds. L’Harmattan 1993. M[inguet], P[hilippe]: “Metonymy and Synecdoche”, in: Encyclopedic Dictionary of Semiotics, ed. by Thomas A. Sebeok, vol. 2. Berlin–New York–Amsterdam: Mouton de Gruyter 1986, 2nd ed., revised and updated, 550–1. Mounin, Georges: “Théorie et terminologie à propos des concepts de contiguïté et similarité”, in: La Linguistique 21 (1985), 37–45. Nerlich, Brigitte: Semantic Theories in Europe 1830–1930. From Etymology to Contextuality. Amsterdam: John Benjamins 1992. – , David D. Clarke: “Outline of a Model for Semantic Change”, in: Kellermann, Günter, Michael D. Morrissey (eds.): Diachrony within Synchrony: Language History and Cognition. Papers from the International Symposium at the University of Duisburg, 26–28 March 1990. Frankfurt/Main etc: Peter Lang 1992, 25–142. – , David D. Clarke,: “Synecdoche as a cognitive and communicative strategy”, in: Blank, Andreas, Peter Koch (eds.): Historical Semantics and Cognition. Berlin– New York: Mouton de Gruyter 1999, 197–214. Nyrop, Kristoffer: Sémantique. Vol. IV of his Grammaire historique de la langue française. Copenhagen: Gyldendalske Boghandel Nordisk Forlag 1913. Nysenholc, Adolphe: “Métonymie, synecdoche, métaphore: Analyse du corpus chaplinien et théorie”, in: Semiotica 34-3/4 (1981), 311–41. – “La métaphore”, in: Linguistics in Belgium 5. Ed. by M[arc] Dominicy. Brussels: Didier Hatier 1982, 119–28. Paul, Hermann: Prinzipien der Sprachgeschichte. Halle/Saale: Max Niemeyer Verlag 1880. Radden, Günter: “The Cognitive Approach to Natural Language”, in: Thirty Years of Linguistic Evolution. Studies in Honour of René Dirven on the Occasion of his Sixtieth Birthday. Ed. by Martin Pütz. Philadelphia–Amsterdam: John Benjamins 1992, 513–41. Reisig, Christian Karl: Vorlesungen über lateinische Sprachwissenschaft. Zweiter Bd. Lateinische Semasiologie oder Bedeutungslehre. Neu bearbeitet von Ferdinand Heerdegen. Berlin: Calvary 1890[1839].
Synecdoche
319
Ruwet, Nicolas: “Synecdoques et métonymies”, in: Poétique 23 (1975), 371–88. Sato, Nobuo: “Synecdoque, un trope suspect”, in: Rhétorique, Sémiotiques, 1-2, ed. by Groupe µ. Paris 1979: U.G.E., 116–27. Schmitz, H. Walter: “Die durchgängige Tropikalisierung der Sprache. Über einen Aspekt von ‘Zeichen im Wandel’”, in: Historiographia Semioticae: Studien zur Rekonstruktion der Theorie und Geschichte der Semiotik. Hrsg. von Klaus D. Dutz & Peter Schmitter. Münster: MAkS Publikationen 1985, 241–70. Schofer, Peter, Donald Rice: “Metaphor, Metonymy, and Synecdoche Revis(it)ed”, in: Semiotica 21-1/2 (1977), 121–49. Seto, Ken-ichi: “On the Cognitive Triangle: The Relation between Metaphor, Metonymy and Synecdoche.” Unpublished manuscript, 1995. – “Distinguishing Metonymy from Synecdoche”, in: Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. Amsterdam–Philadelphia: John Benjamins 1999, 91–120. – “Metonymic Polysemy and its Place in Meaning Extension”, in: Nerlich, Brigitte, Zazie Todd, Vimah Herman, David D. Clarke (eds.): Polysemy: Flexible Patterns of Meaning in Mind and Language. Berlin–New York: Mouton de Gruyter 2003, 195–216. Sperber, Dan: “Rudiments de rhétorique cognitive”, in: Poétique 23 (1975), 389–415. Thomas, Robert: “Ueber die Möglichkeiten des Bedeutungswandels. I”, in: Bayrische Blätter für das Gymnasialschulwesen 30 (1894), 705–32; 32 (1896), 1–27. Todorov, Tzvetan: “Synecdoques”, in: Communications 16 (1970), 25–35. – Théories du symbole. Paris: Eds. du Seuil 1977. Tsohatzidis, S.L. (ed.): Meanings and Prototypes. Studies in Linguistic Catgorization. London: Routledge 1990. Turgot, Anne Robert Jacques, baron de l’Aulne: “Etymologie”, in: Encyclopédie méthodique: Grammaire et littérature, éditée par Beauzée et Marmontel. Vol. II. Pris–Liège: Panckoucke et Plomteux 1784. [Reprint of the article which appeared in vol. VI of the Encyclopédie ou Dictionnaire raisonné des Sciences, des Arts et des Métiers, 1756: 98a–111.] Tversky, Barbara: “Where Partonomies and Taxonomies Meet”, in: Tsohatzidis, S.L. (ed.): Meanings and Prototypes. Studies in Linguistic Catgorization. London: Routledge 1990, 334–44. Ullmann, Stephen: The Principles of Semantics. Oxford: Basil Blackwell 1951; Glasgow: Jackson, Son and Co. 1957. Warren, Beatrice: “Distinguishing between Metaphor and Metonymy”, in: Studies in Anglistics. Ed. by B. Warren & G. Melchers. Stockholm: Almqvist & Wiksell 1995, 137–49.
3. Other Tropes and Truth
Wolfgang Braungart (Bielefeld, D)
Eironeia urbana*
Irony is a ‘sad trope’ which, through its purported vicinity to lie and deception, very quickly arouses suspicion. Friedrich Schlegel’s concept of urban irony, however, situates the ‘truth’ or essence of this trope within sociable communication which requires for those who engage in irony absolute freedom but also makes them responsible for their actions. – The article tries to outline this romantic concept and shows how and in which contexts it has developed since Roman rhetoric. “When a philosopher is told to sacrifice to the Graces, it means in effect: Create irony and achieve urbanity.”1 When Ernst Behler, the great connoisseur and researcher of the theory and history of irony2, died on 16 September 1997, his family announced his death in the Frankfurter Allgemeine Zeitung st with this quotation. It is the 431 fragment from Friedrich and August Wilhelm Schlegel’s journal Athenaeum, one of those somewhat cryptic, forced fragments that are so characteristic of early Romantic discourse in general. Actually, it is not a fragment, but rather an aphorism which is meant to sound witty in a forced way. The Athenaeum fragments are always – among many other things – not only forms, but also self-stylizations of romantic thinking at the same time. As literary-philosophical utterances they are, therefore, acts of both social position-finding and position-claiming, presented in the mode of an outbidding in permanence. Nevertheless, the choice of this 431st fragment for an obituary, introduces a note of existential seriousness into the ingenious literary-philosophical and aesthetic-social game, which the romantic conception of irony at first sight ————— *
1
2
Translated from the German by Armin Burkhardt and Brigitte Nerlich. – I thank Armin Burkhardt (Magdeburg) for his long patience and persistent encouragement to outline, in written form, an idea which had once been fleetingly thrown into a friendly-informal discussion. I thank Heike Pfaff for her assistance in finding the material. I thank Jan Andres and Lothar van Laak (Bielefeld), Manfred Koch (Tübingen), and Claus-Artur Scheier (Braunschweig) for their friendly criticism. See also Scheier (1992; on romantic irony in particular, see pp. 318; 1990). “Opfre den Grazien heißt, wenn es einem Philosophen gesagt wird, so viel als: Schaffe dir Ironie und bilde dich zur Urbanität.” (Quoted from Schlegel [1967: 251], in the following quoted as KA. [Our translation, A.B./B.N.] See Behler (1972, 1990, 1997). With his contribution to Historisches Wörterbuch der Rhetorik (1998), ed. by Gert Ueding, Behler has provided the best recent overview of irony research from Antiquity to present-day literary theory and philosophy. My small study, too, owes many suggestions to the research of Behler. For romantic irony see the still fundamental book by Strohschneider-Kohrs (1977).
324
Wolfgang Braungart
does not necessarily seem to imply. At least one does not associate such an existential seriousness with it at first sight. For the philosopher irony and urbanity are, according to this fragment, goals of life and education. They are what he or she is obliged to sacrifice to the Muses, who demand it. Correct philosophical thinking is and needs self-education with regard to irony and urbanity. In this context, irony is not only a rhetorical figure of simulatio and/or dissimulatio, a figure of improperness, as claimed by so many attempts up to the present day to define irony (however much detailed explanations may vary), but, like urbanity, it is also a real form of communication, i.e. a way of life, though an exclusive one. But what, now, is urbanity?3 Although this term had already been connected with irony in Antiquity, it never really played an important role in its explanation, as far as I can see, not even in the rhetoric of early modernity, and this for good reasons, as will be explained later on. In principle, this still applies to present-day theories of irony, because thinkers tended to attribute to modern, romantic irony a transcendental-philosophical dignity, which does not necessarily go with urbanity. Again I quote Friedrich Schlegel: “urbanity is”, as defined only a little later in his 438th fragment, “the wit of harmonious universality, and this is the one and all of historical philosophy and Plato’s highest music. The humaniora are the gymnastics of this art and science.”4 Again, this is not easily understood. And it illustrates what I have just called a form of writing which is witty in a forced way. The referential framework is a new reading of Plato in the philosophy and literature around 1800 (“harmonious universality”, “one and all”), which certainly does not have to be achieved with naive seriousness, but with urbanity, i.e. with “wit”: mobile, mundane, ingenious, reflexive and discursive.5 “Harmonious universality”, “highest music” – for Schlegel, urbanity is obviously also an aesthetic and social category. The paramount importance of Platonic philosophy for literature and philosophy around 1800, including Friedrich Schlegel, and in particular the importance of his “Symposion”, is well known and has often been proven by research. Plato’s “Symposion” with its conception of eros offers various possibilities of reference for romantic philosophy: It is a narrative, discursive, sociable and mythologically intuitive philosophy, combining poetry and philosophy, and thus intuition and concept, myth and abstraction. Plato’s conception of eros can be used syncretistically; it can for example be amalgamated ————— 3
4
5
Sociological, architectural etc. aspects of urbanity will not be considered here; for a more comprehensive conceptual history of this concept cf. F. Pröfener’s article “Urbanität” (2001, see esp. col. 351). Cf. also Lütteken, Pott & Zelle (2002). “Urbanität ist der Witz der harmonischen Universalität, und diese ist das Eins und Alles der historischen Philosophie und Platos höchste Musik. Die Humaniora sind die Gymnastik dieser Kunst und Wissenschaft.” (Schlegel, KA 2: 253) [our translation A.B./B.N.] Platonic philosophy is of greatest importance e.g. for Hölderlin: see Gaier (1993); as for the reception of Plato’s “Symposion” cf. Matuschek (2002).
Eironeia urbana
325
with late 18th-century Spinozism. The jerky, mobile, restless “great Daimon Eros”6 itself has a romantic potential; it provides, as it were, a mythological framework for Friedrich Schlegel’s own conception of poetry and irony. “Wit”, finally, is one of the key terms in 18th-century discourse and is normally used above all to refer to ingenious and open-minded Frenchmen and Frenchwomen. Without urbanity and wit, according to Schlegel, historical philosophy would just be classical studies. Thus, the study of the humaniora is no end in itself. The humaniora prepare the mind for “historical philosophy” and Plato; they convey mobility to the mind. From this ‘urban’ point of view, which is also a romantic one, “art and science” coincide. Urbanity and wit are, therefore, both conditions and manifestations of irony. Friedrich Schlegel would probably agree in calling the 438th fragment itself ironical. It is a small philosophical piece of art, because it “makes one think a lot” (Kant, Kritik der Urteilskraft) without ever being conceptually exhaustive. This fundamental and infinite reflexivity which Kant attributes to works of art for systematic reasons, is for Schlegel the genuine mode in which modernity presents itself.7 The poetry of modernity is necessarily ironical, since modernity is ironical, i.e. infinitely reflexive, infinitely procedural and infinitely dependent on communication. “Among the arts, romantic poetry is”, as the famous 116th Athenaeum fragment states, “what wit is to philosophy, and what sociability, behaviour, friendliness and love are to life.”8 It “is still in the process of emerging; indeed, it is, by its very essence, eternally becoming and never complete.”9 As a consequence, ironical reflexive poetry in general becomes the genuine expression and symbol of modernity. One may have doubts as to whether it is justified for modernity to claim reflexivity and infinite processuality so exclusively. Or is this rather an expression of the attempt by modernity to ennoble itself, a modernity which, by defining itself as reflexive and procedural through and through postulates the relativity of all stable meaning horizons as the only stable meaning horizon? Therein lies its performative self-contradiction: If nothing is certain, if everything always remains “in the balance” – to use another feature by which Schlegel defines irony –, then at least this is certain: that nothing is certain at all. This at least we can rely on. ————— 6 7
8
9
Plato, “Symposion”, 202b–e, 203 a–e. – thanks to Karl-Heinz Ott for a nice discussion on Plato in autumn 2002! For this see the following new investigations: Götze (2001, for Schlegel cf. esp. pp. 195) and Schmidt (2001). Both studies discuss the early romantic poetics of irony in the context of modernism and/or modernity. “Die romantische Poesie ist unter den Künsten was der Witz der Philosophie, und die Gesellschaft, Umgang, Freundlichkeit und Liebe im Leben ist.” [Our translation A.B./B.N.] Sie “ist noch im Werden; ja das ist ihr eigentliches Wesen, daß sie ewig nur werden, nie vollendet sein kann.” (Schlegel, KA, vol. 2: 183) [our translation A.B./B.N.].
326
Wolfgang Braungart
More important in my opinion, however, is how this definition of poetry can be used to establish a relation between irony and urbanity. Schlegel’s attempt to define irony through urbanity can be linked to the emergence of the concept of world literature, for which mainly Herder and in particular Goethe, the greatest and most ironical ‘romantic’ (cf. Faust), are responsible.10 Urbanity is a communicative category. It is the view, or attitude, which actually makes it possible to become open-minded about the modern ‘circulation’ of thoughts and ideas and submit to them. Therefore, urbanity, in Schlegel, is not only a category of modernity, but also a social virtue. Urbanity unmistakably requires the subject’s relation to the world and society and transcends the bare arbitrariness of subjectivity that Hegel and Kierkegaard criticised in the romantic conception of irony, particularly in Friedrich Schlegel’s version. Urban irony could not thrive at all if it were to renounce the world. On the contrary: urban irony is the only adequate attitude that one can adopt in modernity. The 116th Athenaeum fragment also alludes to the aspect of sociability. Just as romantic, i.e. modern poetry, is the pinnacle of the arts, as it is eternally unfinished and never concluded and therefore per se absolute, so “wit” is the pinnacle of philosophy as a discursive practice – and not as a system! And thus life finds its genuine expression in “sociability, behaviour, friendliness and love”11, i.e. in successful sociability. Sociability is the fundamental social and utopian category of the Enlightenment and the period around 1800.12 This was crucially anticipated in Plato’s ‘Symposion’ which declared that there is no access to the heighest being without the great interpreter and mediator eros. Thus, a hermeneutic concept is formulated in nuce, too: There is no understanding without love! For Schlegel romantic poetry therefore becomes the analogue of life and the symbol of its success. If one remembers the Berlin ‘salons’, one can quite imagine what Schlegel may mean by this.13 They are the social and mundane context in which Schlegel’s wit, irony and urbanity can thrive. This sudden emphasis on living life successfully is quite astonishing, as it goes beyond a debate about transcendental-poetic and philosophical reflexivity and the tendentially deconstructive play that characterises the ironical art of modernity which takes liberties left, right and centre. This should provide food for thought, especially with regard to the ironic Schlegel. In contrast to ————— 10 11 12
13
See Koch (2002). th Schlegel, 116 Athenaeum fragment. For a brief, but excellent introduction see Stollberg-Rilinger (2000, esp. chapt. 5: “Ein Jahrhundert der Geselligkeit” [a century of sociability], 114–45); in addition Peters (1999, esp. the conclusion chapter: VIII. “Poetische Geselligkeitsentwürfe um 1800”, 3. “Die Welt als Chaos – gesellige Liebe als poetisches Ordnungskonzept bei Friedrich Schlegel” “Poetic conceptions of sociability around 1800”, 3. “The world as chaos – social love as poetic concept of order in Friedrich Schlegel”: 305–20]. See Seibert (1993).
Eironeia urbana
327
the maelstrom of relativity and fragmentation into which irony seems inexorably to draw, this is the true regulative idea. Urban irony as the real characteristic of modern poetry is essentially social and ethical and has, at the same time, its own ethos. This approach to social irony was, however, not continued in the reception of romantic art philosophy and thus of romantic irony, as suggested by Schlegel, although a lot of work has been done on the aesthetic and poetological meaning of sociability in Schlegel and the importance of sociability in the discourse about art around 1800.14 In modern deconstructive literary theory and philosophy, Friedrich Schlegel has been promoted to the position of the great Nestor of a conception of art based on the infinite postponement of sense, the dissolution of sense in the autoreferential play 0f signs, the incomprehensibility of literary texts – as if nothing was more important than to talk about oneself.15 The modern champion of irony has converted the insight that it is impossible to achieve reliable, unquestionable meaning, into an attitude of somebody who, allegedly, has deeper insights, who pretends to know what being, things, life and language really are. In his essay on irony Martin Walser was not unduly polemical when he wrote that Schlegel’s irony – as a “refined method of being aloof”16 – has generally been victorious in modern concepts of literary irony, and leads directly to Thomas Mann. Martin Walser claims that this is actually a bourgeois kind of irony, which one must be able to afford in the first instance. Schlegel’s Lyceum-formula of 1797, according to which irony is “the freest of all licenses, which one can use to go beyond oneself”17, is criticized by Walser, for it has no longer anything in common with “Fichtean intellectual freedom”, which Fichte had to struggle for through arduous intellectual work. This romantic “freedom is empty. It is without any definition”.18 As the entire conception of modernism depends on the modern literary and philosophical – but not on the linguistic19 – concepts of irony (and vice versa), a critical revision of this concept of irony directly entails a revision of a certain conception of modernism. For a poetics and an aesthetics of irony as an expression of a specifically modern understanding of the world, making sense ————— 14 15
16 17 18 19
See Kurz (1996: 91–113). As for the aesthetic relation between literature and sociability now see also my own attempt in Braungart (2003: 1–18). A text which has already become almost classical: de Man (1988). As for the connecting lines between Friedrich Schlegel and Paul de Man see Choi (1993: 181–205). The anthology gives a good idea of the discussions around Paul de Man. Cf. also Schumacher (2000). “feine Methode des Über-allem-Stehens” (Walser 1981: 82) [our translation, A.B./B.N.]. “die freieste aller Lizenzen, denn durch sie setzt man sich über sich selbst hinweg” (Schlegel, KA, vol. 2: 160) [our translation, A.B./B.N.]. “Freiheit ist leer. Es fehlt ihr jede Bestimmung” (Walser 1981: 80–1) [our translation, A.B./B.N.]. For an excellent overview see Lapp (1997).
328
Wolfgang Braungart
in a comprehensive way that does not immediately relativise itself has become impossible. This is matched by the specifically modern type of melancholy which makes suffering from the world a positive obligation.20 With its refusal to write in ways that suggest serenity or friendliness, modern German literature has, on the whole, readily submitted to this obligation. Exceptions – Jean Paul, Mörike – confirm the rule. In order to affirm the uniqueness of this ‘modern irony’, irony research as part of literary studies likes to emphasize that the pre-modern conception of irony is predominantly a rhetorical one. That is, it restricted itself above all to studying single linguistic figures in rhetorical and also effect-oriented oratory. It is argued that this rhetorical concept of irony, put forward by Cicero and Quintilian in particular, stayed essentially the same through to early modernity. Only modernity, it is said, developed a comprehensive concept of irony. This may be the case, but it should not be overlooked that, on the one hand, rhetoric, even in early modernity21, has always been a pragmatic theory of linguistic action within which the definitions of the respective Tropes and Figures of the Ornatus were inherently related to the fundamental criterion of Aptum, i.e. appropriateness to the situation. In Cicero and Quintilian this does not mean a mere pragmatic appropriateness for reasons of persuasion, interest or power. To that extent the ironical figure of speech, too, is the expression of an attitude which is expressed in the attribute of urbanity. The rhetoric of Cicero and Quintilian also implies an ethics: The good speaker (perfectus orator) must be a ‘vir bonus bene dicendi peritus’. Such a ‘vir bonus’ has ‘urbanitas’ as one of his attributes. His ‘virtus’ aims at the ‘common affair’, at the ‘res publica’, the ‘community’.22 With a concept of poetry that is rhetorical in principle like that of the Early Modern Age, the competition between art of speech and poetry always becomes visible where the Ornatus of the poetic work of art seems to stress a certain self-right, as, for instance, in mannerist poetry.23 On the other hand the most important rhetorical treatises of Antiquity always refer to Socratic irony which is regarded as a productive performance of life. Socrates’ ‘ironical speech’ is all about achieving something new, that means: insight. The relation of power itself, which is produced by Socrates’ way of asking, becomes a means for the production of insight, i.e. for educa————— 20 21 22
23
See Brückner (1996), Heidbrink (1994), Bohrer (1995). For a classical treatment, see Barner (1970). Introductions are: Ueding & Steinbrink (1994), as for the ideal of the “Vir bonus” cf. esp. 86–9; Ottmers (1996); Knape (2000). Müller (1995) also gives a good overview of the literary and formal history of irony. On understanding mannerism as intended and recognizable formal artistics see Zymner (1995). – One could, however, also interpret this formal artistics as an act of aesthetic-social position-finding and defence, particularly in the literary system of the Baroque. I once tried to point this out using the example of Jean Paul, see Braungart (2000: 307–22).
Eironeia urbana
329
tional purposes (and can naturally be problematised24). This is, however, how it legitimises itself. Rendering irony positive in this way reveals itself in the fact that irony is no longer defined by contrasting it with urbanity, as was still the case in Aristophanes, but by seeing urbanity as its actual expression and form of realization. The most important proof for the existence of such a positive view of irony in Antiquity and for an awareness of the constitutive connection between irony and urbanity can be found in the 8th book of Quintilian’s Institutio oratoria, which also discusses the Tropes in the context of the representation of Ornatus. However, this source has long been overlooked and has been overshadowed by Cicero’s De Oratore. Quintilian regards irony as a subtype of ‘allegory’, “in which the opposite is expressed”25. Certainly, this definition has been criticized again and again. Irony, as meaning the opposite of what is expressed, can only be regarded as a borderline case, however. All definitions of linguistic irony since Antiquity, nevertheless, move around this core: the linguistic expression does not correspond to the actual intention; furthermore, participants in communication must be aware of this fact, otherwise we are simply dealing with a lie. This is quite remarkable already: irony actually is a linguistic practice through which one reveals onself to others more clearly than by using a nonironical expression. Irony calls attention to the speaker’s intention; this is, in fact, its central focus. For who would hit on the idea of asking for the intention of the speaker specifically in the case of a non-ironical expression? To be recognizable thus also means: one must be able to deal with it, to prepare oneself for it. According to Quintilian, the speaker can make his or her intention recognizable e.g. by the “tone, in which it [the ironical expression] is spoken”. Irony can also be recognized “according to the person in question or the nature of the object”. Therefore, irony is context-dependent from the beginning. Otherwise it does not come into existence. This is certainly a necessary, but not an exclusive criterion. Ironical speech, writing, and understanding, therefore, is always situated speech, writing, and understanding. It is speech, writing, and understanding in social communicative performance. That is the reason why irony cannot really be described according to the model of authenticity and non-authenticity. The ‘urban’ ironist is not a “metaphysic”.26 There is no ‘authentic’ alternative to —————
24 25 26
In the ethical context irony is defined by Aristotle in his Nikomachean Ethics; cf. Niehues-Pröbsting (1987). Quintilian (1995: 240–1). See Rorty (1989: 75): “By contrast, ironists do not see the search for a final vocabulary as (even in part) a way of getting something distinct from this vocabulary right. They do not take the point of discursive thought to be knowing, in any sense that can be explicated by notions like ‘reality’, ‘real essence’, ‘objective point of view’, and ‘the correspondence of language of reality’. They do not think its point is to find a vocabulary which accurately represents something, a transparent me-
330
Wolfgang Braungart
the fundamental performativity of irony, to its context sensitivity and to the situatedness of its production and understanding.27 Thus, (here, as in general) one cannot think radically enough about performativeness. This is what makes irony so interesting for aesthetics. For aesthetics – knowing that we can never escape from our historical hermeneutic horizon – applies in principle: hic Rhodos, hic salta. Aesthetic experience is always situated experience.28 ‘Urban’ irony, therefore, cannot be construed as a figure of substitutio. It does not substitute anything. It is necessary in order to make possible and to support a certain, more amicable form of social behaviour. This means that it is predicated on social and hermeneutic premises such as good-naturedness, willingness, and confidence, without which there would be no human understanding. This is why the category of urbanity is so significant. It refers to the ethos of ironical communication. The “truth of the trope ‘irony’” can only be determined from the point of view of this ethos. Without this ethos irony’s performativity would be merely arbitrary – without any prospect of recognition and communicative progress. “In addition to this”, Quintilian continues, “allegory”, and as a subtype, irony, “is used to express the unpleasant truth but without using harsh words and with an air of ‘urban’ politeness”. Irony thus does not take back anything, and it does not conceal anything, but it facilitates social communication nevertheless. ‘Urban’ irony takes the sting out of communication. It should be noted that early modern rhetoric does not discuss irony, if I see it correctly, as an ‘urban’ one any more, but only as a rhetorical instrument of power. Johann Matthaeus Meyfart writes in his Teutsche Rhetorica oder Redekunst (1634): “Ironia occurs when the orator means the direct opposite of what he says. It is, therefore, a trope of scorn and derision; its intricacy lies in the fact that the orator uses it frequently not only to suggest something else, but also means the opposite”.29 By contrast, Cicero and Quin—————
27
28
29
dium.” In my opinion, this is formulated much too brusquely, for the premises of communication, which even the ironist accepts, do not perhaps have a metaphysical status any longer, but they nevertheless do have the status of high commitment. To illustrate this an anecdote from everyday life: When our elder daughter, Sarah, was about three years old, she had a little tic. She was annoyed when her socks were not pulled up completely smoothly when she slipped into her shoes. Once she sat crying on the stairs, refused to come with us, just stared at her feet, and said: “You did very well. The socks are wonderfully creased. You have made a marvelous job again.” Without intending to take up the problem discussed again and again in language acquisition research of irony, children’s ability to use irony and their linguistic competence: to this form of linguistic self assertion and criticism of parental competence and authority – blame by praise – there is no “actual” ‘literal’ alternative, or at best one that would have meant something else and then, perhaps, might have provoked sanctions. With regard to this see now the excellent scholarly work of Lothar van Laak (2003). Cf. also Küpper & Menke (eds.) (2003), esp. the contributions of Wolfgang Iser on the situatedness of aesthetics (ibid.: 176–202), and Hans Ulrich Gumbrecht, on epiphany (ibid.: 203–22). “DIe Ironia ist / wenn der Redener durch das Wort / welches er geredet gantz das Gegentheil verstehet. Ist demnach ein Hohn- unnd Spott-Tropus / und beruhet sei-
Eironeia urbana
331
tilian, who were still the two most important authorities for early modern rhetoric (e.g. for Julius Caesar Scaliger, Gerhard Johannes Voss), regard ‘urban’ irony as a civilized and civilising form of communication. This is exactly why it is an urban form of communication. Irony is not natural.30 The opposite of “urbanitas” is “rusticitas”, rural simplicity, which can also mean awkwardness and even crudeness.31 (Taking this into account one may also suggest that Schlegel defined irony and urbanity in the way he did because he wanted to assert his belief in the modern world and to comment critically on the 18thcentury cult of naturalness.) Rural life in its daily struggle with nature does not need irony. To grow potatoes and breed pigs is a completely unironical business. “Eironeia urbana”, however, is a matter of communication, of the linguistic shaping of life, of creating communicative spaces for interpretation and meaning, and not of straight speaking and straight acting. The connection between irony and urbanity bathes the city and its specific ways of life in a positive light.32 The urban environment (‘Lebenswelt’), characterised for Quintilian, amongst other things, by irony, can be neither equated to a hotbed of vice nor a utopian image of the Heavenly Jerusalem.33 For the history of the meaning of the city, however, these two models are fundamental. For the ironical, ‘urban’ way of life there is therefore no great semantic apriori. In the acts of urban communication everything must be negotiated with regard to the Polis and/or the Res Publica. Obviously, the value and meaning of the Res Publica will always be tacitly recognized and presupposed here. This is particularly crucial for Latin rhetoric. Here the orator proves himself to be really a “vir bonus”, and the suitability of rhetorical speech must be evaluated accordingly. It should be noted that the utopian text par excellence, Thomas Mores’ Utopia of 1516, is on the one hand a highly ironical text and has become, because of its “polyvalence”, the foundation text of a genre or writing34, but it refrains on the other hand from making too many allusions and references to the larger motif of the Heavenly Jerusalem. Literary utopias are primarily drafts for an ideal urban life. With the semantic disambiguation of utopia all irony and with it urbanity disappear from utopian discourse, most clearly in Campanella’s Civitas Solis (written in 1602/22, printed in 1623).35 Campanella’s utopian life in the city is not ‘urban’ in the sense of Quintilian or Cicero. —————
30 31 32 33 34 35
ne Zierligkeit darinn / so offt der Redener solchen gebrauchet / er nicht nur ein anderes andeutet / sondern auch gar das Widerspiel verstehe.” (Meyfart, ed., Trunz 1977: 112). See Blumenberg (1998: 215–8). Pröfener (2001: Sp. 351). Cf. Blumenberg (1998). There is an extensive literature on the great topic of ‘city and literature’; see in particular Klotz (1969), Kugler (1986), Smuda (1992). Cf. Voßkamp (1982, vol. 2: 183–96). See Braungart (1989), Rahmsdorf (1999).
332
Wolfgang Braungart
About 150 years before Quintilian, Cicero had already argued along similar lines: In the second book of his principal work on rhetoric, De Oratore, one can read: “Fine wit always displays irony [urbana etiam dissimulatio est], in which what one says differs from what one thinks, not in [...] the sense that one says the opposite [... ], but insofar as the whole style of speech is marked by playful seriousness, while thinking in a way different from what one says.”36 This has echoes of the Aristotelian definition of irony (“fine wit”).37 In this sense, urbanity is both a characteristic and a condition of irony. In order to understand it, it is not sufficient to understand only the language in which it is expressed. One must also be able to move around rather competently in the social and cultural world which is peculiar to its use One must know the rules of the game of social and cultural life. This also means that the social and cultural situation must be formable and formed in such a way that one can really play in it and simulate seriousness. ‘Urban’ irony is not a figure of the either– or, but of freedom. Irony can only be achieved in a situation that is not characterised by immediate need. As for Quintilian and Cicero, the “total intention” must “shine through”, irony can also cause “laughter”, and wit and humor, which must be recognized as such as well, are related speech acts. In the third book of De Oratore Cicero briefly comes back to irony again, when he talks about speech as pretence (“dissimulatio”), which he certainly defines as irony (“dissimulatio” is a possible Greek translation of “eironeia”): “Moreover pretending has to be mentioned, which means something other than what it says, and which, as it were, maliciously creeps into the human heart; it works in a very advantageous [pleasant] way, as one does not state it ————— 36
37
“Von feinem Witz zeugt auch die Ironie [Urbana etiam dissimulatio est], bei der man anders redet, als man denkt, nicht in dem [...] Sinne, daß man das Gegenteil sagt [...], sondern in gespieltem Ernst des ganzen Stils der Rede, wobei man anders denkt, als man redet.” (Cicero 1981: 382–3) [my italics, W.B.]. [Our translation, A.B./B.N.] Aristotele (2001 4th book, chapt. 13): Aristoteles distinguishes irony from pretence, which he associates more with lying. “The ironist, who makes himself smaller, seems to be of a finer kind: for he does not seem to be like this because of the profit he gains, but to avoid arrogance. He mostly denies himself the achievement of large honours, as Sokrates used to do as well. [... ] Who uses irony with measure, however, and in matters which are not too blatant and well-known, appears as kind. Imagination is opposed to truthfulness. For that is the worse error.” [“Der Ironische, der sich geringer macht, scheint eine feinere Art zu haben: denn er scheint nicht wegen des Gewinnes so zu sein, sondern um die Anmaßung zu meiden. Am liebsten verleugnet er, was große Ehre macht, wie es auch Sokrates zu tun pflegte. [...] Wer aber die Ironie mit Maß anwendet und in nicht gar zu handgreiflichen und bekannten Dingen, erscheint als liebenswürdig. Der Wahrhaftigkeit ist vor allem die Einbildung entgegengesetzt. Denn sie ist der schlimmere Fehler.”] (Ibid.: 179.) [Our translation, A.B./B.N.]
Eironeia urbana
333
with intense commitment, but in a chatty tone.” (“quae est periucunda, cum orationis non contentione, sed sermone tractatur”).38 Obviously, the ethical quality of simulatio/dissimulatio, to which irony is considered to belong as well, is solely dependent on its responsible use in the sense of “urbanitas” and with regard to the Res Publica. If one loses sight of this relation, irony as an ‘urban’ form of communication and way of life, in which everything is decided, becomes suspicious, too. Urbanitas is, e.g. in Augustinus, anything but worth achieving.39 It is later rehabilitated in a certain way – not expressis verbis, but only in the practice of the courtier – in the context of the courtly culture of the Early Modern Age. Baldassare Castiglione’s Book of the Courtier (Il libro del Cortegiano, 1528), which is fundamental for the formulation and conceptualization of early modern courtly culture, does not speak of “eironeia urbana” explicitly, but nevertheless treats in detail the art of speaking ironically and of ironical discourse. He regards it as “quite a nice kind of joking”, which seems to be “very appropriate for persons of high standing”.40 Referring explicitly to Plato, Xenophon and Cicero, Castiglione adds to their “ideas [... ] of the perfect state, the perfect prince and the perfect speaker” that of the “perfect courtier”41 who must be able to master ironical speech and the art of courtly behaviour. But now the Polis has become a court, which is, as all literary court criticism has known for a long time, a snake pit.42 Speech and communication are, after all, not determined by the axioms of ‘urban’ life and responsibility for the “common cause”, but by what is pragmatically sensible and useful for the elegant and socially adroit communication of the courtier who, under the permanent pressure of self-assertion, is in principle oriented towards the social centre. Those who do not master this type of communication, become socially ridiculous and are perceived as tedious and boring. The “cortegiano”, and later the “honnête homme” and the “gentleman”, are all idealizations of a social practice at court, which can, however, look very different in reality. It was only in the culture of the Rococo, with its specific aesthetic ideals of grace, delicateness, sociability, and amusement, that courtly communication opened up in such a way as to approximate to modern urbanity. In literature one can observe this well in Wieland.43 Thus in Quintilian and Cicero “eironeia urbana” is defined in a very distinct sense which is based on its relation to the Polis. It thereby refers to the central element of Greek and Roman ethics: something is good, sensible and reason————— 38 39 40 41 42 43
Cicero (1981: 574–5). [Our translation, A.B./B.N., italics W.B.] Cf. Pröfener (2001: column 351). “eine recht hübsche Art von Scherzen”, die “als sehr angemessen für hochstehende Personen [erscheint]” (Castiglione 1986: 200–1; cf. also pp. 164, 210, 211). “Ideen [...] des vollkommenen Staates, des vollkommenen Fürsten und des vollkommenen Redners” und “des vollkommenen Hofmanns” (ibid.: 8). Fundamental: Kiesel (1979). Thanks to Lothar van Laak for reminding me!
334
Wolfgang Braungart
able with regard to the Other and with regard to the Polis. To that extent the rhetorical category of aptum is also a social and ethical category. To put it crudely: Christianity became historically more and more successful, once the Polis was no longer an entity that could be experienced and that had an effect on communication, and this was (beyond the loss of common values and their meanings) because the Roman Empire had over-expanded and become too fragmented. The ethics of the Polis was replaced by a Christian ethics focused on the individual, which makes it the responsibility of every particular person to do the right thing according to his or her own internal moral law.44 Further, communication inside the Polis is now replaced by a communication which is grounded in individual ethics. Therefore it does not seem to be a mere coincidence that Friedrich Schlegel refers to “eironia urbana” around 1800, i.e. at the beginning of modernity. As the Christian internalization of ethics begins to detach itself from the theological metaphysical horizon that justifies it, “urbanitas” as the communicative practice of an ethics of ‘urban’-sociable life, is rediscovered – in the modern social environments of Jena, Berlin, London and Paris. In this context one should interpret Friedrich Schlegel’s ‘urban’ irony (“eironia urbana”) not only in a transcendental-poetic but also in a more fundamentally communicative-pragmatic way. Modernism is inextricably tied to sociable ‘urban’ discourse with its specific forms. It is the social communicative consequence of the Enlightenment conception of reason as basically social, conversational and processual.45 ‘Urban’ irony thus becomes a form of a modern communicative and sociable “art of life”.46 In a way, the theory of irony tried to theoretically integrate the danger of abusing ironical speech forms (which would imply abandoning its urbanity), stressing again and again that irony must be understood as such. But what does this exactly mean? If we feel that an ironical act of communication could be misunderstood, and that we want to avoid this, we make the irony explicit and thereby also dissolve it. Then we say e.g.: “But I meant this ironically.” With the dissolution of irony we regain urbanity. In the historiography of literary irony Shakespeare’s tragedy Julius Caesar, created shortly before the turn to the 17th century, is rarely missing. The great speech by Caesar’s ally Marcus Antonius during the central part of the drama causes the reversal of the plot. It is considered rightly as a prime example of the demagogical potential of ironical speech – a type of speech characterised by the fact that it is not part of a republican, ‘urban’ ideal, but merely a —————
44
45 46
The process, as stressed occasionally, starts at the beginning of the establishment of the Christian time order, esp. in the context of Stoic philosophy, but only gains currency in the context of the Christian theory of conscience. Basic: Kittsteiner (1992). See my investigations regarding Lessing’s “ring parable” in Braungart (2003). See Schmid (1998: 375–81): “technique of dealing with contradictions: art of irony” [“Technik des Umgangs mit Widersprüchen: Kunst der Ironie”].
Eironeia urbana
335
means for gaining rhetorical power. The fictitious audience in the piece itself is not able to see through the ironical speech strategy and so to possibly dissociate itself from it. It is the real spectator/reader alone who is able to recognize this tragic irony (as is also the case for the prime example of “tragic irony”, i.e. Sophocle’s King Oedipus). The free, ‘urban’, republican basis is missing in the Roman society depicted in Shakespeare’s piece – as demonstrated by the play – and in Shakespeare’s society as well. Roman society is seen here as being controlled by demagogical speeches and persuaded by archaic acts of violence which culminate in the killing of Caesar and the exhibition of his corpse. And it is finally corrected and reconciled only by new violence: by the self-sacrifice of Brutus.47 Contrasted with this instrumentalization of irony, or rather its “abuse” as a mere instrument of power, as demonstrated by Shakespeare, the “eironeia urbana” emerges as a deeply humane, civilising virtue in a social and sociable game of communication based on the freedom to engage in ironical “simulation”48 and the freedom of the addressee to take position with regard to it. In this context freedom also means: to be able to free oneself from being fixed on the act of communication. Let us quote again Friedrich Schlegel: “Irony is a permanent parecbasis.”49 Now, for social-ethical and systematic reasons it should not be permanent. Everyone knows this: one cannot stand the permanent ironist for a long period of time. Those for whom everything is theatre appear to us to be ‘merely’ engaged in theatrical play. Life is performative. I.e.: one must be able to play theatre, if one does not want to be just at the mercy of the performative execution of life. Therefore, one must also be free to engage in parecbasis. This is also constitutive of “eironeia urbana”. And of this it reminds us. In a recently published study of irony this trope was defined as simulated insincerity, as second-order simulation.50 This definition should be accepted. It avoids the problems entailed by the model of improperness. What one is able to say “improperly”, one should be able to say “properly” as well. But to say something in a different way, also means: to say something else. “Eironeia urbana” is not a trope which can be “true” or “false”, “proper” or “improper”. There is no “lie in the extra-moral sense” (Nietzsche – and against him!). No linguistic figure, or linguistic trope is true or false in itself. Truth and morality are problems we have to address as individual responsible persons. Yet we can certainly abuse irony in a morally problematic way: see Mark Anthony. However, we can also make a great effort to use “eironeia urbana”. Then we must follow the moral and social rules of communication without which ‘urban’ communication is not possible and cannot succeed. With the “eironeia ————— 47 48 49 50
I have worked out this interpretation in more detail and with reference to René Girard’s literary anthropological theory of sacrifice (Braungart 1998). Cf. Lapp (1997). Schlegel KA, vol. 18, Abt. II, 668. Cf. Lapp (1997).
336
Wolfgang Braungart
urbana” a widely reviled trope on the verge of lie becomes a trope of our social-communicative freedom and responsibility. References Barner, Wilfried: Barockrhetorik. Untersuchungen zu ihren geschichtlichen Grundlagen. Tübingen 1970. Behler, Ernst: Klassische Ironie, Romantische Ironie, Tragische Ironie. Zum Ursprung dieser Begriffe. Darmstadt 1972. – Irony and the Discourse of Modernity. Seattle etc. 1990. – Ironie und literarische Moderne. Paderborn–Munich–Vienna–Zurich. – “Ironie”, in: Historisches Wörterbuch der Rhetorik. Ed. by Gert Ueding. Bd. 4. Tübingen 1998, 599–624. Blumenberg, Hans: “Urbanität”, in: Blumenberg, Hans: Begriffe in Geschichten. Frankfurt/Main 1998, 215–8. Bohrer, Karlheinz: Der Abschied. Theorie der Trauer: Baudelaire, Goethe, Nietzsche, Benjamin. Frankfurt/Main 1995. Braungart, Wolfgang: Die Kunst der Utopie. Vom Späthumanismus zur frühen Aufklärung. Stuttgart 1989. – “Freundschaft, Ironie und Opfer in Shakespeares Tragödie ‘Julius Caesar’”, in: van Ingen, Ferdinand, Christian Juranek (eds.): Ars et Amicitia. Beiträge zum Thema Freundschaft in Geschichte, Kunst und Literatur. Festschrift für Martin Bircher zum 60. Geburtstag am 3. Juni 1998. Amsterdam 1998, 99–114. – “Manierismus als Selbstbehauptung: Jean Paul”, in: Braungart, Wolfgang (ed.): Manier und Manierismus. Tübingen 2000, 307–22. – “Prolegomena zu einer Ästhetik der Geselligkeit (Lessing, Mörike)”, in: Euphorion 97 (2003), 1–18. Brückner, Pascal: Ich leide, also bin ich. Die Krankheit der Moderne. Eine Streitschrift. Weinheim/Berlin 1996 [French orig. Paris 1995]. Castiglione, Baldesar: Das Buch vom Hofmann. Translated and explained by Fritz Baumgart. With an epilogue by Roger Willemsen. Munich 1986. Choi, Moon-gyoo: “Frühromantische Dekonstruktion und dekonstruktive Frühromantik: Paul de Man und Friedrich Schlegel”, in: Bohrer, Karl Heinz (ed.): Ästhetik und Rhetorik. Lektüren zu Paul de Man. Frankfurt/Main 1993, 181–205. Cicero, Marcus Tullius: De Oratore. Über den Redner. Lateinisch/Deutsch. Translated and ed. by Harald Merklin. Stuttgart 1981, 2nd ed. Gaier, Ulrich: Hölderlin. Tübingen–Basel 1993. Götze, Martin: Ironie und absolute Darstellung. Philosophie und Poetik in der Frühromantik. Paderborn etc. 2001. Heidbrink, Ludger: Melancholie und Moderne. Zur Kritik der historischen Verzweiflung. Munich 1994. Kiesel, Helmuth: ‘Bei Hof, bei Höll’. Untersuchungen zur literarischen Hofkritik von Sebastian Brant bis Friedrich Schiller. Tübingen 1979. Kittsteiner, Heinz D.: Die Entstehung des modernen Gewissens. Darmstadt 1992. Klotz, Volker: Die erzählte Stadt. Ein Sujet als Herausforderung des Romans von Lesage bis Döblin. Munich 1969. Knape, Joachim: Allgemeine Rhetorik. Stationen der Theoriegeschichte. Stuttgart 2000. Koch, Manfred: Weimaraner Weltbewohner. Zur Genese von Goethes Begriff ‘Weltliteratur’. Tübingen 2002.
Eironeia urbana
337
Küpper, Joachim, Christoph Menke (eds.): Dimensionen ästhetischer Erfahrung. Frankfurt/Main 2003. Kugler, Hartmut: Die Vorstellung der Stadt in der Literatur des deutschen Mittelalters. Munich–Zurich 1986. Kurz, Gerhard: “Das Ganze und das Teil. Zur Bedeutung der Geselligkeit in der ästhetischen Diskussion um 1800”, in: Jamme, Christoph (ed.), unter Mitwirkung von Frank Völkel: Kunst und Geschichte im Zeitalter Hegels. Hamburg 1996, 91–113. Laak, Lothar van: Hermeneutik der Sinnlichkeit. Historisch-systematische Studien zur Literatur des 17. und 18. Jahrhunderts. Tübingen 2003. Lapp, Edgar: Linguistik der Ironie. Tübingen 1997, 2nd ed. Lütteken, Laurenz, Ute Pott, Carsten Zelle (eds.): Urbanität als Aufklärung. Karl William Ramler und die Kultur des 18. Jahrhunderts. Göttingen 2002. Man, Paul de: Allegorien des Lesens. Frankfurt/Main 1988 [Orig. Allegories of Reading]. Matuschek, Stefan (ed.): Wo das philosophische Gespräch ganz in Dichtung übergeht. Platons ‘Symposion’ und seine Wirkung in der Renaissance, Romantik und Moderne. Heidelberg 2002. Meyfart, Johann Matthäus: Rhetorica oder Redekunst. Ed. by Erich Trunz. Tübingen 1977 (Deutsche Neudrucke. Reihe: Barock 25). Müller, Marika: Die Ironie. Kulturgeschichte und Textgestalt. Würzburg 1995. Niehues-Pröbsting, Heinrich: Überredung zur Einsicht. Der Zusammenhang von Philosophie und Rhetorik bei Platon und in der Phänomenologie. Frankfurt/Main 1987. Ottmers, Clemens: Rhetorik. Stuttgart–Weimar 1996. Peters, Emanuel: Geselligkeiten. Literatur, Gruppenbildung und kultureller Wandel im 18. Jahrhundert. Tübingen 1999. Pröfener, F.: “Urbanität”, in: Ritter, Joachim, Karlfried Gründer, Gottfried Gabriel (eds.): Historisches Wörterbuch der Philosophie. Bd. 11. Darmstadt, 2001, col. 351–4. Quintilianus, Marcus Fabius: De Institutione Oratoria Libri XII. Ausbildung des Redners. Zwölf Bücher. Ed. and translated by Helmut Rahn. Zweiter Teil. Buch VII– XII. Darmstadt 1995, 3rd ed. Rahmsdorf, Sabine: Stadt und Architektur in der literarischen Utopie der frühen Neuzeit. Heidelberg 1999. Rorty, Richard: Contingency, irony, and Solidarity. Cambridge etc. 1989. Scheier, Claus Artur: “Klassische und existentielle Ironie: Platon und Kierkegaard”, in: Philosophisches Jahrbuch 97 (1990), 238–50. – “Philosophische Tendenzen in der deutschen Frühromantik”, in: Abhandlungen der Braunschweigischen Wissenschaftlichen Gesellschaft XLIII (1992), 303–32. Schlegel, Friedrich: Friedrich-Schlegel-Ausgabe. Ed. by Ernst Behler unter Mitwirkung von Jean-Jacques Anstett und Hans Eichner. 35 Vols., vol. 2: Friedrich Schlegel: Charakteristiken und Kritiken I (1796–1801). Ed. and with an introduction by Hans Eichner. Erste Abteilung. Kritische Neuausgabe. Munich etc. 1967, 251. [Quoted as KA]. Schmid, Wilhelm: Philosophie der Lebenskunst. Eine Grundlegung. Frankfurt/Main 1998. Schmidt, Benjamin Marius: Denker ohne Gott und Vater. Schiller, Schlegel und der Entwurf der Modernität. Stuttgart–Weimar 2001. Schumacher, Eckhard: Die Ironie der Unverständlichkeit. Johann Georg Hamann, Friedrich Schlegel, Jacques Derrida, Paul de Man. Frankfurt/Main 2000. Seibert, Peter: Der literarische Salon. Literatur und Geselligkeit zwischen Aufklärung und Vormärz. Stuttgart–Weimar 1993. Smuda, Manfred (ed.): Die Großstadt als ‘Text’. Munich 1992.
338
Wolfgang Braungart
Stollberg-Rilinger, Barbara: Europa im Jahrhundert der Aufklärung. Stuttgart 2000. Strohschneider-Kohrs, Ingrid: Die romantische Ironie in Theorie und Gestaltung. Tübingen 1977, 2nd ed. Ueding, Gert, Bernd Steinbrink: Grundriß der Rhetorik. Geschichte – Technik – Methode. Stuttgart–Weimar 1994, 3rd ed. Voßkamp, Wilhelm: “Thomas Morus’ ‘Utopia’: Zur Konstituierung eines gattungsgeschichtlichen Prototyps”, in: Voßkamp, Wilhelm (ed.): Utopieforschung. Interdisziplinäre Studien zur neuzeitlichen Utopie. 3 vols. Vol. 2. Stuttgart 1982, 183–96. Walser, Martin: Selbstbewußtsein und Ironie. Frankfurter Vorlesungen. Frankfurt/Main 1981. Zymner, Rüdiger: Manierismus. Zur poetischen Artistik bei Johann Fischart, Jean Paul und Arno Schmidt. Paderborn etc. 1995.
Herbert L. Colston (Wisconsin-Parkside, USA)
Irony, analogy and truth
0. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14 15. 16.
Introduction Tropes and truth Pragmatic versus semantic truth Truth mingling That’s so true The paradox of tropical truth Truth ownership Master of my domain Having one’s cake and eating it too Praising the messenger Cognitive dissonance Meaning endorphins End run around Truth flirtation Social truth Tropes and truth: the case of ironic analogies Conclusion
0. Introduction The notion of truth is intimately and intricately connected with the family of figurative and other forms of indirect language commonly referred to as tropes. This relationship is also a great deal more than just the relatively simple partnership commonly observed between counterparts (e.g., keys & locks, poisons & antidotes, seals & impressions) – as might be advocated by the view that the figurativeness of tropes obviates or nullifies their semantic truthfulness (Cutler 1976; Grice 1975, 1978; Searle 1979). Indeed, although the well-documented yet not often noted fact about meaning – that semantic opposites are often closer together in meaning than other seemingly closely related or, indeed even widely considered “synonymous”, pairs of term – can be applied in one sense to the relationship between tropes and truth, a great deal more complexity holds, however, in this relationship. The following chapter will first provide a general overview of the intricacies of the relationship between tropes and truth. It will then highlight the relationship between truth and a particular type of “tropical” language – ironic analogies. 1. Tropes and truth
340
Herbert L. Colston
In perhaps a dozen related but separable ways to be considered here, tropes and truth can be said to bear an intimate relationship. This review is not offered as an exhaustive list, nor are the connections presented in a meaningfully rank- or otherwise ordered pattern. Rather, the following are simply presented as a variety of potential tropical truth linkages that might be representative of how the two entities are related. 2. Pragmatic versus semantic truth The first connection between tropes and truth was alluded to in the opening paragraph. One might argue that the figurativeness and hence, semantic nonveridicality, of common tropes like metaphor and irony, makes them inherently untrue in the sense of typical truth-functioning in semantics. If a speaker says, for instance, that a man is a beanpole or that a disastrous event is lovely, there exist no possible worlds in which such statements have positive truth functions because men by their nature cannot be beanpoles, and without considering extenuating premises, disastrous events are not lovely. So in this sense, one can argue that tropes are not true statements, or that the relationship between tropes and truth is one of counterfactuality. However, on a pragmatic level tropes are certainly true statements if one considers that a speaker very likely intends to communicate what is metaphorically, ironically, or otherwise figuratively comprehended from such language. Whatever metaphorical meaning is conveyed by he is a beanpole, given the conversational context, interlocutor common ground, metaphorical meaning underpinning, etc. (i.e. that the human referent is thin, tall, lanky, etc.), is indeed very much a truth about the world provided that the referent genuinely bears those physical characteristics. The speaker of this metaphor is thus very much uttering a truthful expression. Moreover, a preponderance of empirical evidence suggests that hearers/ readers need not and frequently simply do not, recover the “literal” senses of such expressions when they’re comprehended as metaphors or irony (Gibbs 1994b). Although in some instances, some component of the veridical sense of such expressions can play a limited role in the final comprehended meaning, particularly so for irony where often a contrast effect can be observed between the referent situation and a veridical semantic sense of the utterance (Colston 2002). But in nearly all instances of metaphoric and ironic language, the equivalent to what would likely be comprehended by the utterances if used veridically (i.e., were a speaker while attempting to identify a tall narrow stake-like object found in a dark garden shed, upon bringing it out into the light say, “he is a beanpole”) is not recovered when that same surface form is used metaphorically or otherwise figuratively. Indeed, even in seemingly consistent situations of a “literal” usage of a statement, one is hard-pressed to identify the objective, default, universal,
Irony, analogy and truth
341
semantic underpinnings that all such usages would share. Notions such as tall, thin, narrow, useable for staking up bean plants, etc., although perhaps commonly comprehended in usages of beanpole, are neither necessarily present nor universal. One can always identify contexts in which such notions are not at play even though the usage is still more literal than metaphorical (e.g., a speaker could refer to a flat brick wall upon which ivy is growing as a beanpole). 3. Truth mingling There is also a sense in which notions of both semantic and pragmatic truth are at play in the use of tropes, rather than what might seem the suggestion from the above discussion that they’re necessarily separate. As mentioned above with irony in particular, both semantic and more pragmatic intentional expressions of meaning are in operation when a speaker utters, for instance, nice weather to comment upon what would generally be perceived as bad weather. A variety of studies has shown that the comprehension of such utterances is influenced by both the more veridical sense of nice, as well as the intended ironic negative meaning (see Colston 2000b for a review). Use of such utterances, which typically, although not universally, takes the form of a veridically positive comment made about a negative referent, sets up a contrast effect where the negative situation must be viewed in the context of positive commentary. This juxtaposition causes the referent situation to frequently be seen as more negative than had it been judged with neutral or no commentary, such that its perceived valence shifts toward the negative. In this fashion, the final comprehension is the result of a blend and contrast of meanings that can be framed by semantic or pragmatic truth structures. The same mingling process might be argued for other tropes where semantic and pragmatic meanings are consistent, but in these cases the mingling is not one of contrast but rather corroboration, symbiosis or even synergy. Consider an idiom applying to a situation both literally and figuratively (e.g., saying he’s skating on thin ice, about an ice fisherman who has wandered onto a thin patch of ice on a lake). With these mixtures, as is also the case with many double entendres, an interesting emergent third meaning is often enabled. This meaning can be borderline ironic to the extent that two potentially very different meanings can arise from the same verbatim phrase, and yet both still apply to the contextual situation at hand (e.g., an upholstery shop that specializes in repairing damaged antique furniture being named, The Recovery Room). Or the emergent meaning may be something else. For instance, an utterance can be seen as a lame pun if the different meanings are only weakly related. Or an artful pragmatic manipulation can take place where social meanings are deftly and differently accomplished (e.g., one interlocutor is manipulated into miscomprehension where a second is given witness to this insult), or other emergent meanings.
342
Herbert L. Colston
Another kind of semantic/pragmatic meaning mixture holds when metaphorical meanings are not semantically exclusive of more veridical meanings (e.g., saying this weather is garbage). In these cases, the use of a semantic vehicle that already has a great deal of flexibility (depending upon the context, anything can be literally garbage) may enable a different meaning blending than semantic vehicles that are more limited (e.g., saying a man is a tortoise). Whether such differential mixability of meanings is a meaningful difference between kinds of metaphors is an open question, but it nevertheless is another way in which semantic and pragmatic senses, or truths, can mix. 4. That’s so true A very important sense in which tropes and truth are related concerns the notion of degree of truthfulness. Aside from the notion of fuzzy logic which allows a gradation between truth and falsity, truth can also vary in its own depth or density, as well as breadth or reach, and tropes are undeniably relevant to this notion (Gibbs 1989; Honeck 1997; Lakoff 1988). Let’s first consider the depth or density of truth. One well known characteristic of many kinds and instances of tropes is that they are rich in meaning. They’ve been popularly called meaning-in-concentrate, nuggets of wisdom, pearls of wisdom, and other terms that all allude to this quality. In this fashion, tropes can thus be claimed to have a greater potential concentration of truth than some comparable instances of more veridical language. Next consider breadth or reach of tropical truth. Tropes have also widely been argued to have very far reaching, broadly encompassing truthfulness, and have been so described with terms such as ringing true, being true everywhere, containing universal truths, etc. On this view, tropes can be argued to have greater range of truth than comparable veridical language. But the question must be addressed as to how this is possible. How can a particular kind of language have more concentrated and wide ranging truth than other comparatively lengthy language, with no seeming difference in the number of propositions, idea units, concepts or other measures used to assess the amount of information? The answer lies in two possibilities, meaning means and meaning cooccurence. Although varying widely in their form and theoretical base, a number of mechanisms have been proposed and evaluated to explain the motivation and groundedness of the meanings of different tropes (see below). Without diverting into an evaluation or comparison of these different mechanisms, and without going in great detail into any one or set of them, suffice it for now to say that many if not most such mechanisms supply powerful explanations for the richness of tropical meaning. And this holds even if one only considers a given mechanism in isolation. Although it is very likely that not all of the mechanisms are completely valid, a number of them nevertheless probably are, and if one considers the convergence of mechanisms that
Irony, analogy and truth
343
would hold due to the sheer number of potentially valid accounts, the evidence for the enhanced truthfulness of tropes becomes clear. Let me briefly mention three of these mechanisms for illustration. One is conceptual metaphor, or the idea that much linguistic metaphorical meaning is motivated by conceptual metaphorical meaning (Lakoff & Johnson 1980). In essence, we talk metaphorically in systematic ways because our concepts are organized accordingly. As just one example, people cross-culturally often talk about the concept of UNDERSTANDING in terms of the vehicle SEEING (e.g., I see your point, That’s not clear to me, Your argument is murky, but they’ll see what you mean, etc.). According to conceptual metaphor theory, such a systematic pattern in talk is due to a corresponding systematic structure in cognition – understanding is conceptualized in terms of vision. Such an account can explain the truth density of metaphorical tropes because terms that make use of such metaphorical patterns could be argued to tap into these deep underlying conceptual alignments and thus seem more deeply or richly true. Another mechanism that can help explain tropical truth density is embodiment. On this view, much tropical and other language cognition is enabled because of the embodied nature of language, or that one source of linguistic meaning is grounded in our bodily experiences of the world. As one example, trope terms that use the metaphorical word stand (e.g., stands to reason, I won’t stand for this, etc.) are particularly meaning rich because the word stand invokes the embodied experience of exerting skeletal or dynamic muscular force to combat external pressures such as gravity (Gibbs 1994a). Thus the terms are rich because they tap into meanings that are motivated by our embodied experience, and thus seem somehow truer than comparable non-embodied terms (I won’t tolerate this). The last mechanism I’ll mention that can explain tropical truth density is an emerging idea in cognitive linguistics and in my own and others’ research in psycholinguistics and language cognition, that a great deal of overlap exists in language and other arenas of mental activity (e.g., sensation, perception, cognition, etc.). As such, the richness of tropical language can be argued to stem from this intimate connection among levels and arenas of cognitive functioning. My own work that has demonstrated the functioning of a contrast effect – an all pervasive perceptually-based reactionary judgment resulting from biasing contexts – in people’s comprehension and interpretation of verbal irony is just one example. Negative situations that are referred to in an utterance of verbal irony are made to look worse via the contrast effect produced in ironically labeling the situations positively, and comprehension of the utterance is affected accordingly. A variety of other language/perception/ sensation/etc., connections are also being proposed and evaluated (see Boroditsky 2000, 2001; Boroditsky & Ramscar 2002; Gentner, Imai & Boroditsky 2002; Richardson, Spivey, Barsalou & McRae 2003). The other general answer to the question of how tropes have greater depth and breadth of truth is meaning coincidence. Many tropes might be lent a
344
Herbert L. Colston
broader reach of truth because the different levels of meaning of a trope are in alignment with one another. If one remark can thus be “true” at a number of different levels, then this coincidence bolsters the apparent range of truthfulness of the trope. As just one example, consider the following proverb, while two dogs fight over a bone, a third walks away with it. Such a remark is true at both salient levels of meaning – that having to due with the behavior of canines and that concerning people’s actions. The multiplicity of such applicability can make a remark such as this seem particularly insightful, wise and broadly “true”, for it seems to convey a widespread fact about how the world operates with a seemingly narrow topic focus. 5. The paradox of tropical truth The preceding discussion may be revealing a paradox in the relationship between tropes and truth. On one level, tropes have the characteristic of enhanced truthfulness, in terms of both depth and breadth, in that their comprehension can tap into deeply embedded meaning motivation mechanisms and their applicability can spread across multiple layers of fact about the world and, thus they can seem deeply and broadly true. But at the same time, tropes achieve this enhanced truthfulness with a greater disconnection with their referent topics on the seemingly central function of semantic meaning. Put simply, tropes are more true, by not being true. How is this possible? How can something be more true and less true at the same time? My contention is that such a paradox is only apparent on the surface. A deeper look at the notions of language, truth and meaning clarifies the picture. Tropes make use of a wide variety of embodied, sensory, perceptual, cognitive, linguistic, semiotic, social, cultural and other mechanisms in their conveyance of meaning. In doing so they are powerful tools for affecting feeling, belief, comprehension, knowledge, attitude, intimacy, ingratiation and a host of other facets of a shared conscious life. They are thus quite adept at ringing true for people. One need only recall the subjective experience when newly encountering a well formed metaphor or ironic statement to see this (I’ll unabashedly steal two examples heard from the same witty tropicalfactory colleague, 1) that a newly proposed idea, …needs to marinade for a while, and 2) a retort to a comment at an academic senate meeting that a proposed administrative policy would require far too much communication, negotiation and cooperation between academic departments, … oh no, we can’t have that!). At the same time, tropes do not nicely fit with a more traditional view of meaning conveyance based on a truth-functional philosophical and semantic notion. That this is not a paradox I think can be seen from the view that such a traditional framework of language, truth, and meaning is simply far too limited for the job it faces. It adopts a greatly overextended metaphor of symbolic processing and computational and mathematical rigidity in an attempt to ex-
Irony, analogy and truth
345
plain what is naturally a much more dynamic, flexible and multidetermined system. That it would thus struggle with the vast array, complexity and frequency of tropical and indeed other forms of figurative and indirect language is thus no surprise. 6. Truth ownership Another connection between tropes and truth lies in the notion of truth ownership, or, the question of where the source of some piece of knowledge lies – in an individual speaker or in the broader world. The more a truth is centered in the outside world rather than in a single person/speaker, the broader and stronger that truth is. One might argue that the inherent nonveridicality of tropes enhances their truthfulness because it shifts the ownership of a truth away from the speaker. But, the question then is how does the nonveridicality of tropes shift meaning ownership away from speakers. The primary answer is that tropes, relative to nonveridical language, place more of the comprehension work on the shoulders of the hearer and less upon the speaker or language. Hearers must derive the meaning of a trope more from contextual, background, inference and memory sources. Successful comprehension is thus more dependent upon information sources outside of the speaker, and thus the truthfulness of a trope is more broadly distributed. I’ll return to this issue in the final section on ironic analogies, but for now let me just illustrate this truth externalization process by deconstructing a real instance of comprehension of a proverb. In a recent discussion I was having with a friend, we were complaining about a colleague’s research in Psychology. Our issue was that the colleague’s main claim was so broad that it could essentially fit with any phenomenon in this particular area of study, and thus wasn’t really falsifiable – a hallmark of good science. Our complaint mainly lay with the fact that despite this limitation of the research, it had achieved a limited but not inconsequential degree of success. This success had been cited by a number of audience members of our discussion, to contest our complaints, who questioned how this research could be influential if it has this flaw. To remark upon this, I uttered the following, If you can build a house, you can sell a house. The consequence of this statement was that it essentially got the entire audience of overhearers, including the skeptics, to understand how such a seemingly weak scientific account could be accepted. The point for present purposes, is that the persuasiveness of such a proverb can be traced to the generalizeability of truth that is achieved by the proverbial form. Had I uttered a veridical paraphrase of the meaning of the proverb (which itself is a difficult thing to do, again because much of the planned meaning of the proverb arises in the hearer’s comprehension process and is thus hard to explicitly state by a speaker), the support for the truthfulness of that meaning would have been based solely upon my credibility. But by putting the support instead on the analogous, vivid, concrete, and, most
346
Herbert L. Colston
importantly, external, proverbial vehicle topic, I as speaker become merely the usher to such a truthfulness, and I strengthen that truthfulness accordingly. Essentially, by diminishing personal ownership, truth is lent a more universal quality. 7. Master of my domain The next connection between tropes and truth in some ways runs counter to the connection discussed in the preceding section. Speakers can also enhance the truthfulness of meaning they convey in tropical language, while retaining personal ownership of that truth. The trick here is that the speaker enhances their credibility by demonstrating mastery over the referent topic. Most interestingly, the success of this demonstration of mastery is traceable back to the use of the tropical language (Colston & Connelly 2004). The mechanism works essentially as follows: All else being equal, tropical language, particularly if novel and creative, is often somewhat more sophisticated and complex than comparable veridical language. A speaker using such tropical language would thus have to be generally skillful. A skillful speaker is likely to be intelligent and observant about the world. Thus, the truthfulness of what this speaker says is likely to be enhanced, again all else held equal, by speaking figuratively. Put succinctly, an expert on something is more likely to be correct than a nonexpert. 8. Having one’s cake and eating it too I mentioned above that the mastery demonstration and source externalization mechanisms of truth enhancement in tropical language seem in opposition. In practice, however, if they are both at play, they usually compliment one another. A speaker who is able, through the deft use of a clever and indirect language form, to leverage a hearer into making a seemingly independent, deep truth realization, is in fact demonstrating a keen mastery of communication. 9. Praising the messenger Indeed, when a speaker is able to make hearers aware of such a blend of the truth enhancing mechanisms of externalization and mastery demonstration, he can catalyze truth enhancement. Such skill in meaning delivery can result in both ingratiation, in that the hearer may come to realize the roundabout compliment delivered by the speaker’s use of the indirect language, as well as galvanization, in that the hearers’ beliefs can become steadfast. The speaker has to have enough faith in the hearer’s ability to correctly comprehend the
Irony, analogy and truth
347
indirect tropical utterance in order to attempt using it. And the reaction to such a compliment, as well as to the recognition of such adroit communication skill, is typically one of increased admiration and trust. This attitude can enhance the hearers’ adherence to the expressed meaning – hearers are more apt to find and hold truth in speakers they like. 10. Cognitive dissonance The next connection between tropes and truth lies in a long-standing and powerful mechanism from Social Psychology that has an impact on the value people place on anything they’ve had to do, including any language/information they’ve had to contemplate. This mechanism, cognitive dissonance, proposes first that people seek to align their actions with their beliefs. If there is any discordance between these two, for instance a person is an advocate of voting but doesn’t vote in an election, then a form of cognitive/emotional tension – cognitive dissonance, increases. Secondly, the mechanism proposes that people seek to minimize their cognitive dissonance, because having high dissonance is uncomfortable. Next, the mechanism acknowledges that people are highly susceptible to influences on their behavior that run counter to what might be popularly termed willpower. In essence, a great deal of our behavior is brought about by influences beyond our control or even awareness. The mechanism then claims that when people are in a high state of dissonance, they must lower it by adjusting either their behavior or beliefs. Lastly, the mechanism purports that in most instances, it is easier to change beliefs than behavior, so people will do so accordingly and, often, without even knowing they’ve done so. As to the relationship between tropes and truth, cognitive dissonance would arise when a person has had to engage in any cognitive work in an encounter with some kind of tropical language. Given the kinds of pragmatic meaning ramifications discussed thus far, that arise from an encounter with tropical language, a person’s successful comprehension and interpretation of rich tropical language, although not necessarily involving different kinds of processes, can often involve more cognitive work than is the case for relatively bland, more veridical talk1. A mild state of heightened cognitive dissonance can thus arise when comprehending tropical language. People commonly believe that it is wasteful to engage in work that is not necessary for some reason, purpose, goal or whatever. So, people might come to view the truthfulness behind the meaning of a tropical utterance to be greater, in order to justify the work they expended to derive that richer and more far-reaching meaning. ————— 1
In either kind of language, however, comprehension still involves work that in a sense is involuntary – comprehension processes are so automatized that they’ll almost always occur once triggered.
348
Herbert L. Colston
Indeed, this cognitive dissonance mechanism is a very powerful tool of communication, found far beyond the boundaries of tropical language, and often used by people, even unknowingly, for persuasive purposes. It lies behind the general tendency for people to believe anything they see in printed form, for people to believe things they’ve been told, and for people to come to believe things they themselves have said or written, if only for the justification of the work involved in such comprehension/production, by both speaker/ writers and hearer/readers. 11. Meaning endorphins Related to this process of cognitive dissonance is the sense of accomplishment and emotional satisfaction people often get in the comprehension/interpretation of tropical language. These sorts of pleasurable feelings can become associated with the meaning(s) derived from tropical language, such that the truthfulness of those meanings is heightened in the minds of the hearers. Essentially, many hearers have a positive subjective experience when they comprehend a trope. Whether due to the enhanced truthfulness mechanisms already discussed, which can make a hearer feel as if they’ve just encountered something “better”, compared to a veridical language comprehension experience. Or perhaps related to a sense of discovery that can arise because of the more objective quality of the meaning, again due to lessened speaker truth “ownership”. Or even if due to the complementary feelings one might have, either from the ingratiation process or just due to feeling clever or smart for having gotten the tropes’s figurative meaning. In all these cases people simply feel good at having comprehended a trope. Such a positive feeling, when accompanied by a revision to one’s cognitive world representation, can make that new knowledge seem somehow better, or more truthful. 12. End run around Yet another way in which tropes can connect with truth concerns the unique kinds of meanings, or truths, that are gotten at only by certain tropical comprehension and/or pragmatic mechanisms. In other words, there are some kinds of truths that only tropes can get at. Much has been said about the frequent difficulty in literally paraphrasing certain tropical meanings, which is certainly related to this idea. But I’d like to offer a different illustration. Consider certain instances where a person goes to very unusual and extreme lengths to make some communicative point. The classic anecdote where then Soviet leader Nikita Kruchev banged his shoe on a table is one example. Another is an oft used cinemagraphic technique of a person in a group of people gaining the conversational floor with some unusual attention getting trick (e.g., the character Quinn scraping his fingernails
Irony, analogy and truth
349
across a blackboard to get attention at a boisterous town meeting in the popular American film Jaws). When a speaker does something like this, he or she is breaking set and doing something that just isn’t often done as an aspect of communication. This deviance first grabs people’s attention because of its distinctiveness, but it can also be communicative in its audacity. It can emphasize importance, disrupt cloistered, self-important or herd-like thinking, and reveal truths that would not otherwise be seen. The main point is that it is often necessary to go to the extreme roundabout lengths used in these actions to make those points. Standard talk alone cannot easily achieve them. My point is that some tropes may operate in this way. In order to bring about a certain kind of truth revelation, a speaker might have to do something rather extreme to draw attention to something, to make something distinctive, to get people to see things differently, etc. As an example, consider the trope of hyperbole (see Brdar-Szabó & Brdar, in this volume, 385–429). A colleague of mine recently used this trope in a meeting. Some faculty members in my department were meeting with members of the University’s administration, and were trying to get the administrators to realize how burdensome one of our courses had become because of its large enrollment. In the past the administration had repeatedly taken no notice of this issue. The faculty member finally said, There are nine-hundred thousand people taking this course!. This exaggeration had the effect of getting the administrators to realize the problem with the course’s enrollment, and a dialog finally ensued as to how to remedy the situation. This utterance made use of the standard mechanism supplied by hyperbole – drawing attention to a discrepancy between desires or expectations and reality, by inflating that discrepancy (Colston & Keller 1998). All else being equal, large things are more noticeable than small things, so making a discrepancy large can draw attention to it. Veridical speech has a more difficult time achieving this kind of attention-getting, and thus hyperbole exists because it points out such a truth rather readily. Similar claims can be made for many kinds of tropes, including metaphor, irony and others. 13. Truth flirtation The penultimate connection between tropes and truth that I’ll discuss concerns a process I’m referring to as truth flirtation. I chose this term purposefully because of the familiarity to the process of romantic flirtation. In the latter process, people will often build intimacy by the dual mechanisms of prowess display and indirectness. The prowess display is a means of making the other person attracted, and the indirectness is used as a means of enticement. So a person is made to desire something, but the accessibility of that something is left ambiguous, enough so to leave hope. This can serve to increase the person’s curiosity and make their desire even stronger.
350
Herbert L. Colston
With truth flirtation, a speaker will bring an interlocutor closer to, or more intimate with, some meaning or truth, in much the same way that a person in romantic flirtation brings another person closer to him or herself in an emotional or sexual sense. The tropical language serves as a form of both prowess display and indirectness. The prowess display occurs via the language going to what may be greater than necessary lengths for communicative purposes, as well as through the truth burgeoning mechanisms discussed earlier. The indirectness is achieved by the ambiguity of meaning. Tropical language can thus increase an interlocutor’s curiosity and attention toward some target information or truth. It is also interesting to note that one of the ready means of achieving romantic flirtation, is through the truth flirtation of tropical language. Here, the “truth” that is being made enticing is the knowledge of the person doing the flirting. There is also another sense in which truth is related to flirtation through tropes. For the same reason that a personal relationship that has a great possibility of being very successful, lasting a long time, making both partners happy, etc., is best served by starting out delicately (a too early revelation of very strong feelings by either partner can kill the relationship quickly), so can a great understanding of widespread truths be hindered by a too early overexposure to that knowledge. So tropical language can serve the purpose of indoctrinating a person into some new knowledge or truth, without overwhelming the person. 14. Social truth The last connection between tropes and truth that I’ll consider is a form of social knowledge or truth that often gets exchanged when interlocutors use tropical language. Related to the mastery display mechanism discussed previously, a speaker who uses tropical language can often reveal a greater amount of truth about themselves. Such speakers often seem more “real” than people who use more veridical language. Speakers of tropical language can seem more observant, insightful, sensitive, deep, experienced, worldy, etc., in that they know enough about the world and are skilled enough in their language skills that they can leverage a great deal of that knowledge in the mind of the hearer. Skilled users of tropical language are often seen as more interesting people because they have this quality.
Irony, analogy and truth
351
15. Tropes and truth: the case of ironic analogies In the preceding sections I make a large number of claims about the relationship between tropes and truth. In this section, I will provide some validation of some of these claims, via consideration of a series of empirical studies undertaken to investigate the lesser-studied trope of ironic analogies. To begin, first consider an example of such an ironic analogy. In ironic analogies (originally called “rebuttal analogies” by Whaley & Holloway 1996, and Colston & Gibbs 1998), a target proposition, behavior, idea or the like, is typically compared with an ironic base, such as: (1) “Calling Chilies2 just another steakhouse is like saying the Great Wall of China is just a fence” (actual [although not necessarily verbatim] radio commercial) In a series of papers, these kinds of analogies were first identified as being tropical for their use of irony in their base terms (Colston & Gibbs 1998). Ironic analogies were then created and empirically evaluated for their argumentativeness and degree of social attack (Colston 1999). Three kinds of ironic analogies were created, those with highly ironic, somewhat ironic and nonironic bases, for example: You are in your Political Science class watching a debate between another student and the teacher. The student says that the U.S. should double the defense budget in order to intimidate North Korea. The teacher responds to this by saying, (Highly Ironic Base) “Doubling the U.S. defense budget to intimidate North Korea is like using a chainsaw to file your nails” (Somewhat Ironic Base) “Doubling the U.S. defense budget to intimidate North Korea is like using a hack saw to file your nails” (Nonironic Base) “Doubling the U.S. defense budget to intimidate North Korea is like using a nail file to file your nails” Experimental participants rated these types of ironic analogies as statistically significantly different from one another on both their argumentativeness and degree of social attack, with more irony producing more of these pragmatic ————— 2
Name of a popular chain of restaurants in America.
352
Herbert L. Colston
effects. Irony was thus identified as the causal mechanism in these statements, enabling speakers’ control over the degree of expression of argument and social attack, given that speakers craft the bases in ironic analogies to align with the typically quoted target propositions. A third study then evaluated two possible alternative mechanisms to the pragmatic functioning of ironic analogies – “absurdity comparison”, in which a speaker can merely analogically pair another person’s proposition with something absurd to achieve argument and social attack, and “argumentative convention”, in which a speaker is just seen as being argumentative and attacking because of the characteristics of the discourse in which ironic analogies are typically used (Colston 2000a). This study revealed that, although analogies with absurd bases (e.g., Doubling the U.S. defense budget to intimidate North Korea is like using ketchup to wax your car) could explain a portion of the functioning of ironic analogies, it did not enable prediction of the full degree or subtlety of such functioning. Rather, the degree of irony in the base term of the analogies provided full prediction of the extent to which a speaker was viewed as being argumentative and socially attacking. This series of studies empirically documents several of the mechanisms that unite the concepts of truth and tropical language. Ironic analogies were shown to employ the truth externalization mechanism discussed previously in that the analogical rather than simple modification form, the lack of qualifiers, and the detailed analogical structure-mapping all contributed to an objectification of the claims made by the statements (Colston 1999). The use of analogical form is also a straightforward use of the truth alignment or meaning coincidence mechanism in that the internal structure of a target domain is highlighted via the direct comparison with a similarly structured base domain. The trope also utilizes mastery demonstration in that the more complicated ironic bases serve to convey the strongest rebuttal messages. And, although not explicitly measured, much previous research has documented the relative difficulty in comprehending analogical forms, so one could readily argue that part of the persuasiveness of ironic analogies lies in the cognitive dissonance that must be lowered – to justify the work required to garner the pragmatic ramifications of the statements. Lastly, these tropes nicely point out the “end run around” nature of tropical language in that the relatively complex system of coupling another person’s proposition with an invented base term in order to highlight the ironic nature in that target proposition is probably only possible with such a roundabout and extreme method. As the empirical data from the studies revealed, using other comparative but less “tropical” statements simply cannot do what ironic analogies do. 16. Conclusion This article has attempted to reveal the complex relationship between tropes and varieties of notions of truth. Although not all of the claims have empirical
Irony, analogy and truth
353
support to date, most of them do and the remaining ones seem highly plausible. One overarching principle might be considered, however, in the consideration of the validity of the claims made here, and that concerns the mere presence of tropical language as a form of communication. If there is any viability to the philosophical notions of a search for meaning, a cooperative principle, relevance and others, then the simple existence of tropical language must support a connection between tropes and truth. Unless one would argue that all tropes are outright attempts to create falsity, or to unnecessarily place layers of obstacles around a core sense of communicated truth, then some such relationship(s) must hold. True, tropes do introduce degrees of complexity that might require more deconstruction to unpackage an expressed truth, but they do this for warranted purposes, many of which result in expressions of additional kinds of truth that themselves are achieved with remarkable strength and efficiency. References Boroditsky, L.: “Metaphoric structuring: Understanding time through spatial metaphors”, in: Cognition, 75 (2000), 1–27. – “Does language shape thought? Mandarin and English speakers conceptions of time”, in: Cognitive Psychology, 43 (2001), 1–22. Boroditsky, L., M. Ramscar: “The roles of body and mind in abstract thought”, in: Psychological Science, 13(2) (2002), 185–9. Colston, H. L.: “The pragmatic functions of rebuttal analogy”, in: Metaphor and Symbol, 14(4) (1999), 259–80. – “Comprehending speaker intent in rebuttal analogy use: The role of irony mapping, absurdity comparison and argumentative convention”, in: Language and Speech, 43(4) (2000a), 337–54. – “On necessary conditions for verbal irony comprehension”, in: Pragmatics & Cognition, 8(2) (2000b), 277–324. – “Contrast and assimilation in verbal irony”, in: Journal of Pragmatics, 34 (2002), 111–42. – R.W. Gibbs: “Analogy and irony: Rebuttal to ‘Rebuttal Analogy’”, in: Metaphor and Symbol, 13(1) (1998), 69–75. – , C. Connelly: Figurative language and composture display. Poster session presented at the meeting of the Midwestern Psychological Association, Chicago, IL. May 2004. – , S.B. Keller: “You’ll never believe this: Irony and hyperbole in expressing surprise”, in: Journal of Psycholinguistic Research, 27(4) (1998), 499–513. Cutler, A.: “Beyond parsing and lexical look-up: An enriched description of auditory sentence comprehension“, in: Wales, R., E. Walker (eds.): New Approaches to Language Mechanisms. Amsterdam: North-Holland 1976, 133–50. Gentner, D., M. Imai, L. Boroditsky: “As time goes by: Evidence for two systems in processing space/time metaphors”, in: Language and Cognitive Processes, 17(5) (2002), 537–65. Gibbs, R. W.: “Taking a stand on the meanings of stand: Bodily experience as motivation for polysemy”, in: Journal of Semantics: An International Journal for the In-
354
Herbert L. Colston
terdisciplinary Study of the Semantics of Natural Language, 11(4) (1994a), 231– 51. – The Poetics of Mind: Figurative Thought, Language and Understanding. Cambridge: Cambridge University Press 1994b. – , N.P. Nayak, C. Cutting: “How to kick the bucket and not decompose: Analyzability and idiom processing”, in: Journal of Memory and Language, 28(5) (1989), 576–93. Honeck, R.: A Proverb in Mind: The Cognitive Science of Proverbial Wit and Wisdom. Mahway, NJ: Lawrence Erlbaum Associates 1997. Lakoff, G., M. Johnson: Metaphors We Live by. Chicago: University of Chicago Press. 1980. – “Cognitive semantics”, in: Eco, U., M. Santambrogio, P. Violi (eds.): Meaning and Mental Representations. Bloomington: Indiana University Press 1988, 119–54. Grice, H. P.: “Logic and conversation”, in: Cole, P., J. Morgan (eds.): Syntax and Semantics 3: Speech acts. New York: Academic 1975, 41–58. – “Some further notes on logic and conversation”, in: Cole, P. (Ed.): Syntax and Semantics 9: Pragmatics. New York: Academic 1978, 113–28. Richardson, D. C., M.J. Spivey, L.W. Barsalou, K. McRae: “Spatial representations activated during real–time comprehension of verbs”, in: Cognitive Science, 27(5) (2003), 767–80. Searle, J.R.: “Literal meaning”, in: Searle, J.R. (ed.): Expression and Meaning. Cambridge University Press 1979, 117–36. Whaley, B. B., R.L. Holloway: “‘Rebuttal’ analogy: A theoretical note”, in: Metaphor and Symbolic Activity, 11(2) (1996), 161–7.
Armin Burkhardt (Magdeburg, D)
Euphemism and truth*
0. 1. 2. 3. 4. 5. 6.
Introduction Subfunctions Subtypes Euphemism and semantic change Euphemism and speech acts Euphemism and truth Conclusion
0. Introduction Since ancient times language has not only been used as a means of information and communication, but also as an instrument of disinformation and wilful misunderstanding. One of the favourite means of misleading or diverting people’s thoughts by linguistic expression is euphemism. In the Old Greek tradition euphemism was linked to magical thinking, i.e. the belief that fatal consequences could be avoided by substituting an alleged taboo word by a word or periphrasis which was considered to have no such magical force. “The eupheme was originally a word or phrase used in place of a religious word or phrase that should not be spoken aloud; etymologically, the eupheme is the opposite of the blaspheme (evil-speaking).” (http://en.wikipedia.org/ wiki/Euphemism) But euphemisms are not limited to the religious sphere. By changing the implied conception of the addressee from a secretly monitoring supernatural power to the human target of persuasion it soon also became a means of political propaganda. In Antiquity and in the Middle Ages it was known by rhetoricians as a linguistic phenomenon, but categorised differently: While Aristotle (Rhet. III, 2, 6–14) considered them a subtype of metaphor, Cicero used the terms verecundia (‘reservedness, discretion’) and verba tecta (‘covered words’), Quintilian (VIII, 6, 57) briefly mentioned them as special forms of allegory1, and Bede (1975: 163), e.g., called them Charientismos ————— *
1
I would like to thank Brigitte Nerlich and Pedro J. Chamizo Domínguez for their linguistic corrections and for some helpful comments and suggestions. I am also very grateful to the Faculty of Letters of the University of Málaga for a grant that helped me a great deal to finish this chapter. “haec si quis ignorat quibus Graeci nominibus appellent, σαρκασµον, αστεισµον, αντιφρασιν, παροιµιαν dici sciat.” (‘if somebody does not know, how the
356
Armin Burkhardt
(‘good behaviour, decency’). During the English Renaissance euphemism was used in the sense of “prognostication of good” (H. Peacham), i.e. as the prophecy of a good end. Among German Baroque rhetoricians, however, euphemism was used in its present-day sense: Vossius (1630, IV, ix, 186–929) defined it as a figura res odiosas ingratasque designans nominibus non ingratis (‘figure that designates odious things or things said with reluctance by pleasant words’), Fabricius, however, reverted to the antique view of euphemism by calling it a metaphor which conveys “a milder concept”2 than the object itself (cf. 1724: 188). And these are main senses in which euphemism nowadays is widely understood. Dubois et al. (1974: 227) defined it as an expression “which erases semes which seem disturbing or superfluous from an assertion counting as objective and substitutes new semes for them”3. And even in an ordinary dictionary such as Longman’s Dictionary of Contemporary English (2001: 466) euphemism is explained as “a polite word or expression that you use instead of a more direct one to avoid shocking or upsetting someone”. The questions which have to be answered in this chapter are, therefore, not what euphemism means or how the concept of euphemism must be defined, but how euphemisms work and whether they can be true or false. As euphemisms appear in different linguistic forms and may serve different communicative purposes, a number of subclasses may be distinguished. In the following sections I will discuss different subfunctions (1), distinguish different (linguistic) subtypes (2), discuss their role in semantic change (3), compare them to speech acts (4), and then try to answer the question of euphemism’s possible truth and falsity (5). 1. Subfunctions Euphemisms, generally speaking, are instances of presenting an unwelcome or contested reality in a linguistically pleasant or less controversial way. They are means to present bad things in good terms and must be distinguished from words with offensive connotations which have been termed dysphemisms and orthophemisms which are considered to be the neutral or standard expressions (cf. Allan & Burridge 2006: 31–40). With regard to their function in discourse one can distinguish between two subclasses of euphemism, the “veiling” (“verhüllend”) and a “concealing” (“verschleiernd”) subclass (Luchtenberg 1985: 24). Veiling euphemisms are means of indirectly expressing what ought not to be uttered expressis verbis. They are used in the realm of religious or social taboo, i.e. in the context of divine or evil powers on the one hand and death, disease, excretion and sexu–––––––––– 2 3
Greeks call these expressions, he should know that they are called sarcastic, urban, euphemistic-contrary and proverbial allegory’; my translation, A.B.). “einen gelindern concept”. “er tilgt aus einer für objektiv geltenden Aussage Seme, die störend oder überflüssig erscheinen und substituiert ihnen neue Seme”.
Euphemism and truth
357
ality on the other (see e.g. Chamizo Domínguez 2004: 2005; Chamizo Domínguez & Sánchez Benedito 2000). Therefore, a magical and a social subtype of veiling euphemisms must be distinguished. The idea of using euphemisms in the religious or magical sphere is to avoid the explicit use of a tabooed holy name such as God by calling him Lord or He, for instance, or to evade mentioning the names of evil powers such as the devil, demons or dangerous animals, for fear of provoking them or their appearance, by speaking of cloven foot, Eumenides (‘the kindly ones’) or der Braune (‘the brown one’ referring to a bear) instead. Social euphemisms, however, are used to show regard for the feelings or imagination of others by glossing over unpleasant or indecent features of the objects referred to. This is why people tend to speak of the deceased in the presence of relatives of a dead person who is also said to have passed away, to repose or to sleep the last sleep and why modern funerary workers refer to the corpse by terms like the loved or the dearly departed.4 For the same reason people usually do no longer speak of cripples, but of the handicapped or disabled or refer to cancer as the big C. A similar motivation can be detected when people, at least in social situations where politeness is required, prefer to say that they go to the bathroom, make love to or sleep with others and refer to their private parts, instead of using more common words which are referentially identical but considered too indecent or even vulgar. While magical euphemisms are caused by religious or profane fear and superstition, social euphemisms can be taken as expressions of shame, tactfulness and compassion and, therefore, as forms of moral behaviour. In both cases, however, no idea of deception or lie can be detected.5 Concealing euphemisms, by contrast, are means of deliberate hiding. They are used intentionally by speakers to deceive or manipulate others, especially in politics, in the military and in advertising (cf. Dietl 1996: 1).6 Euphemisms of this kind may be exemplified by minus growth for ‘recession’, mission for ‘military attack’, neutralize for ‘kill’, pacify for ‘occupy and suppress by military force’, pre-owned for ‘used, second-hand’, full-figured for ‘fat’, body odour for ‘smell’, or senior citizens for ‘old people’ who certainly do not fly second class, but coach. All these expressions serve the purpose of blocking unwelcome thoughts and are designed to partially mislead the recipients’ world perception. Often such euphemisms are also used to calm the speaker’s own conscience and hide uncomfortable aspects of reality to himself. —————
4
5 6
For an analysis of the euphemistic metaphorization of death see Crespo Fernández (2006) who also shows that, in primitive societies, the names for death were avoided for fear of their magical powers as well (ibid.: 103). For a detailed discussion of the concept and history of (linguistic) taboo and for more examples see Zöllner (1997: 29–89; Allan & Burridge 2006: 1–28). As political and military euphemisms are taken to be particularly dangerous forms of manipulation which contradict the idea of democracy they are object of constant language criticism (see e.g. Leinfellner 1971; Zöllner 1997: 317–90; Bohlen 1994).
358
Armin Burkhardt
All this does not mean, however, that the intended deception must always be successful. Especially if such a word or word usage becomes the object of public discussion its euphemistic impact may be diminished. There may even be a “secret coalition between the parties concerned”7 (Rada 2001: 85), a tacit understanding which implies that the true reference of a euphemistic expression, say, golden age in place of old age or vertically challenged for short, is or, at least, could be recognised by almost all speakers. This is typical for the vocabulary of “political correctness”.8 But even if the euphemistic character of an expression is obvious to any language user, there will still be a slight psychological effect which makes it easier to talk about or deal with the unwelcome object or state of affairs referred to. 2. Subtypes From the point of view of semantics one can distinguish between syntactic euphemisms in which the whole sentence is in a way infected and lexical euphemisms in which the palliative meaning is carried by just one word or phrase. Syntactic euphemisms are whole sentences which are formulated, by means of words or phrases with no or relatively few negative connotations, in a way which warrants that unpleasant truths are hidden or at least played down. They are incomplete descriptions of states of affairs which hide unwelcome information from their addressee. Imagine the coach of a football team which has just lost a decisive match and whose relegation thereby has become almost unavoidable saying in the press conference after the match: (1) It didn’t turn out well, and now one must be prepared for a new situation. Here, the whole sentence or rather, both sentences embellish an unfavourable reality: Instead of admitting explicitly that the match was lost, the coach refers to some impersonal “it” and predicates it a lack of goodness with regard to the results. And then he refers to an abstract “one” (replacing the concrete we) predicating it that it ‘is forced to face hitherto unexpected circumstances’ instead of explicitly telling the public and their supporters that, by losing the match, his team has virtually been relegated. The most infamous examples of syntactic euphemisms may be found in politics. Some of them have assumed an almost proverbial status as shown by the following examples: ————— 7 8
My translation, A.B. For a detailed description of the American PC-movement and its linguistic consequences see Zöllner (1997: 211–316; Allan & Burridge 2006: 90–111).
Euphemism and truth
359
(2) In any case the present market condition gives rise to discussions. (3) At the moment we live through a severe military strain in the East.9 (2) was pronounced by the president of the German Bundesbank in 1970 in view of an economic crisis, and (3) was exclaimed by Goebbels in his Berlin Sportpalast speech a few days after the battle of Stalingrad. In both cases the real facts are played down and belittled with regard to duration and extent . In both instances the euphemistic force is carried by the whole sentence. Among syntactic euphemisms there is a subtype which is often used in diplomacy and may be called “referential vagueness”. It is used in the contexts of alleged guilt and blame and consists in avoiding to mention the author of a certain fact or deed described in the predicative part of the sentence, when, indeed, this very fact is known to almost every participant in the relevant discourse. Let us consider the following examples: (4) Some people will never learn to behave. (5) Our government has to get ahead of the climate-change issue. [...] We need people in Washington who understand that.10 (6) “We understand the concerns of some countries about the proliferation of weapons of mass destruction,” said Kong Quan, the foreign ministry spokesman [of China].11 (4) might have been said by a well-educated person who had been insulted by his addressee and wants to show his intellectual superiority without escalating the confrontation. (5) was said indirectly of the Bush administration by Hillary Clinton during the presidential race without mentioning it. (6) was expressed with regard to the US policy towards North Korea including the plan to intercept ships and aircraft belonging to the Asian “rogue state” which are suspected of carrying equipment for use in nuclear and chemical weapons. All three sentences are examples of making reproaches without indicating their target which is obvious to almost everybody. Indirect or vague reference, combined with other euphemistic expressions (“(not) behave” for ‘blatant misconduct’, “understand” for ‘to be intelligent enough’ and “concerns” for ‘fear and worries’) and morally suggestive presuppositions (‘one should have learned to behave’, ‘the Bush administration does not get ahead of the climate-change issue, and there are no people in the government who understand that’, and ‘the US overreact with regard to the proliferation of weapons of mass destruction’), creates euphemistic readings. Another common way of concealing the author of an unfavourable fact is the use of the passive voice in place of the active: —————
9
10 11
The German original reads: “Wir durchleben im Osten augenblicklich eine schwere militärische Belastung.” See http://www.wtop.com/?nid=213&sid=977788. http://www.derechos.org/nizkor/iraq/doc/chn1.html (05sep03).
360
Armin Burkhardt
(7) At least 5,000 civilians may have been killed during the invasion of Iraq, an independent research group has claimed. (http://www.guardian.co.uk/Iraq/Story/0,2763,976392,00.html) Not to mention the author(s) of a deed means to pass over in silence the question of guilt. Therefore, this grammatical form of euphemism is often used for self-exculpatory reasons. In making use of the passive voice the intentions of the politician in front of the TV-cameras do not differ in structure from those of “the small boy who explains how the window has ‘got broken’” (Ayto 1994: 2). Lexical euphemisms, however, are based on word meanings or rather the choice or creation of words and can be either abstracting or positivising. Abstracting euphemisms consist in replacing the normal or more appropriate designation of an unfavourable object or state of affairs by an abstract hyperonym in which the negative features are absent. In this way, mass murder was reduced to special treatment (the original Nazi term was Sonderbehandlung), a military attack may be called an action, modern weapons can be referred to as systems, a price increase is always announced as a price adjustment, the toilet may be called facilities, and poverty is trivialized by calling poor people economically disadvantaged. The killing of civilians who are often referred to as soft targets by the military will be regretted by using the term collateral damage, while killed soldiers are usually counted as casualties. Positivising euphemisms are those in which negative semantic features are cancelled and replaced by positive ones. According to this principle, a recession is called minus or negative growth, being bombarded or gunned by one’s own troops is publicly admitted by speaking of friendly fire, old people are referred to as senior citizens, looting may be justified by calling it redistribution of property, aggressors or terrorists may call themselves freedom fighters, and a nuclear accident or even a nuclear catastrophe, such as the Chernobyl nuclear power plant disaster of 1986, may be played down by referring to it as a simple malfunction (thereby substituting the feature ‘of inestimable extent’ by the feature ‘of limited and calculable size’). Occupation will often be called pacification, while the colony or the occupied country itself is named a protectorate. Unpleasant matters may even be euphemised by their opposite, e.g. war can be called peace, torture love, and blatant lies can be called truth. This violation of semantics underlying what he calls “doublethink”12 was the main ————— 12
“Doublethink means the power of holding two contradictory beliefs in one’s mind simultaneously, and accepting both of them” (Orwell 1973: 171): “To know and not to know, to be conscious of complete truthfulness while telling carefully constructed lies, to hold simultaneously two opinions which cancelled out, knowing them to be contradictory and believing in both of them; to use logic against logic, to repudiate morality while laying claim to it, to believe that democracy was impossible and that the Party was the guardian of democracy; to forget whatever it was necessary to forget, then to draw it back into memory again at the moment
Euphemism and truth
361
principle of “Newspeak” in Orwell’s terrifying vision of Nineteen EightyFour (1973). It is a method of hiding non-opportune semantic features in order to make them sink into oblivion and thereby “narrow the range of thought” (ibid.: 45). These are extreme (and fortunately rare) cases of positivising euphemism, but somehow generals who call their soldiers peace keepers or societies who give the name of Ministry of Defence to what was formerly called Ministry of War come quite close to this extreme. Some positivising euphemisms are metaphorical because a comparison between a “focus” and a “frame” underlies them implicitly. Others are metonymical, as they just mention a part or (incidental) feature instead of the whole. Ethnic cleansing for ‘genocide’, fratricide for ‘two atomic bombs exploding close to one another thereby reciprocally diminishing their effectiveness’, human shields for ‘hostages that are kept in or around combat targets to deter an enemy from attacking’, and surgical strike for ‘missile attack on military targets which is so precise that it allegedly excludes civilian victims, i.e. collateral damage’ are examples of the former, while to make someone redundant (as there is no more work to do for him) or to give someone the sack (i.e. the bag containing his tools) for ‘to dismiss’, and to neutralize for ‘to kill’ (as killed soldiers cannot take sides any more) are examples of the latter.13 According to the metonymical principle toilets are often called lavatories (as you can also wash your hands there), and camps which held a high concentration of people to be killed or at least exploited were called concentration camps by their criminal inventors. Allan and Burridge (2006: 49–53) pointed out that a euphemism is often “linked with the speaker’s point of view, dysphemism with some other view” and described this as “an us versus them situation.” Particularly in political language, which is often taken up uncritically by journalists, a euphemism will often correspond to a dysphemism which is used to refer to quite the same objects, kinds of actions or groups of persons on the enemy’s side: In the Iraq of 2004, American liberators/invaders fought Iraqi insurgents/freedomfighters. [...] military heroes disengage from the enemy or make a tactical withdrawal; the enemy retreats. The troops cause collateral damage while they are getting the job done; the enemy commits terrorist acts against civilians. In both cases the orthophemism is civilian deaths (that result from military action). (Ibid.: 51)
––––––––––
13
when it was needed, and then promptly to forget it again: and above all, to apply the same process to the process itself. That was the ultimate subtlety: consciously to induce inconsciousness, and then, once again, to become unconscious of the act of hypnosis you had just performed. Even to understand the word ‘doublethink’ involved the use of ‘doublethink’.” (Ibid.: 31–2) For examples of the conceptual metonymies THE SENTIMENTAL EFFECTS OF DEATH STAND FOR DEATH and THE PSYCHOLOGICAL EFFECTS OF DEATH STAND FOR DEATH see Crespo Fernández (2006: 110).
362
Armin Burkhardt
Dysphemisms as well as euphemisms are semantic means of evaluation and, therefore, not intended to reflect a given reality in a psychologically neutral way. 3. Euphemism and semantic change In his “Principles of the History of Language”, first published in 1880, the Neogrammarian Hermann Paul deals with euphemism in his chapter on semantic change. He discusses it briefly, along with hyperbole, dysphemism, litotes, pejoration and irony, under the headline of “modification of meaning”. He writes: A phenomenon which stands in opposition to popular folk coarseness is e u p h e m i s m insofar as it consists in avoiding a taboo expression because of a sense of shame and substituting it by another one which alludes to it. It is very likely that this one will also become offensive soon. Cf. expressions like der Hintern [bottom, backside], die Scham [private parts, pudenda], sein Wasser abschlagen [to relieve oneself], Abtritt [lavatory, toilet], lat. coitus, etc. Similar to the sense of shame religious or superstitious inhibition triggers descriptive expressions, cf. Gottseibeiuns [devil, Old Nick].14
Paul correctly mentions the sense of shame and religious or superstitious avoidance as the standard motives for the use of euphemistic expressions, but does not make any reference to the deliberate use of euphemisms in politics and public discourse. These matters have been dealt with above. More important here is that Paul also describes euphemism as one of the sources of semantic change and that he, by mentioning that a euphemistic expression may be felt to be offensive after a while, also hints at what was to be called the “euphemism treadmill” a century later (see below). As Paul’s and many other examples of lexical euphemism show, euphemistic expressions, i.e. words and idioms, may be conventionalised and lexicalised, so that they may become the standard expressions and, therefore, will no longer be felt to be euphemistic. The German krank (‘ill’) e.g. originally meant ‘weak’, but then, from the 14th century on, was used euphemistically to describe an ill person. One hundred years later it has lost its euphemistic character and has become the standard expression replacing the older siech (a ————— 14
My translation, A.B. The original German passage, which, for unknown reasons, is missing in the English translation (Paul 1888), runs: “Der volkstümlichen Derbheit gegenüber steht der E u p h e m i s m u s , insoweit er darin besteht, dass aus Schamgefühl der eigentliche Ausdruck vermieden und durch einen andeutenden ersetzt wird. Sehr leicht wird dann auch dieser wieder anstössig. Vgl. Ausdrücke wie der Hintere, die Scham, sein Wasser abschlagen, Abtritt, lat. coitus, etc. Ähnlich wie das Schamgefühl ist religiöse oder abergläubische Scheu die Veranlassung zu umschreibenden Ausdrücken, vgl. Gottseibeiuns.” (Paul 1975: 102)
Euphemism and truth
363
relative of the English sick).15 This shows that as long as the corresponding taboo or discrimination prevails, another euphemism will be found or created by the speakers to replace the expression which is no longer felt to be euphemistic, and so on. The very moment a euphemism is commonly accepted, its former meaning fades and the search for a new euphemistic expression begins. Such euphemisations may occur several times throughout language history with regard to the same referent. Or, in other words, under the pressure of social taboos former orthophemisms become dysphemisms which are replaced by euphemisms which become orthophemisms and then dysphemisms which must be replaced by euphemisms and so on. “The result is a constant turnover of vocabulary for words denoting taboo concepts” (Allan & Burridge 2006: 243). This explains why political correctness can never be successful over a longer period of time (as long as the old taboos, prejudices and superstitions survive). The most famous example for this are the expressions referring to black Americans who were first called negroes (Spanish negro, derived from the Latin adjective niger ‘black’), then colored or black people and, more recently, African-Americans or people of color. Another impressive example is given by Leinfellner (1971: 130) who quotes (from W. Safire’s “The New Language of Politics” (1968)) the following sarcastic remark of a female slum dweller: “I used to think I was poor. Then they told me I wasn’t poor, I was needy. They told me it was self-defeating to think of myself as needy, I was deprived. Then they told me underprivileged was overused, I was disadvantaged. I still don’t have a dime. But I have a great vocabulary.” The progression lame > crippled > handicapped > disabled > differently-abled is another well-known example. Such a process of constant euphemisation has been called “the euphemism treadmill” by Steven Pinker (1994)16: “People invent new ‘polite’ words to refer to emotionally laden or distasteful things, but the euphemism becomes tainted by association and the new one that must be ————— 15
16
This example is taken from Fritz (2005: 106–7) who briefly describes the role of euphemism in language change. For English examples see e.g. Zöllner (1997: 121– 4). Allan and Burridge (2006: 243) prefer the notion of “X-phemistic recycling” as dysphemisms and orthophemisms are involved in the euphemisation process as well which, for them, resembles more a cycle than a (tread)mill. The same basic idea is expressed in Chamizo Domínguez (2007) who also offers a sociopsychological reason for de-euphemisation: “Therefore, a euphemism is a linguistic device which allows us to name what is considered unmentionable by a given society or human group. But as the taboo continues to exist although we do not refer to it by using the term which literally names it and which is itself considered forbidden, the euphemistic term tends to end up being contaminated by the taboo object, so to speak. And, from this very moment, the euphemism starts to cease being so and, ipso facto, starts to be felt by the speakers as an inappropriate word. As a result, euphemistic terms cease to be ambiguous, become lexicalised, their euphemistic sense becomes their salient (or literal) meaning and, eventually, they become dysphemisms.” (Chamizo Domínguez 2007: 11; also Chamizo Domínguez 2004: 46–7)
364
Armin Burkhardt
found acquires its own negative connotations.” And he rightly infers from this that in this process “concepts, not words, are in charge: give a concept a new name, and the name becomes colored by the concept; the concept does not become freshened by the name.” Connotations, therefore, one might be inclined to think, are rather cognitive than semantic phenomena. And if they remain unaltered any such normative fiddling about with language, as wellmeaning as it may be, must be a Sisyphean task: “We will know we have achieved equality and mutual respect when names for minorities stay put.” (Pinker 1994) 4. Euphemism and speech acts In his famous chapter “Logic and conversation” (1975) Grice uses the sentence (8) He was a little intoxicated (of a man who is known to have broken up all the furniture), along with cases of irony, metaphor and hyperbole, as an example in which the first of his conversational maxims, that of Quality: “Do not say what you believe to be false”, is flouted. The example sentence, for him, is a case of meiosis, i.e. a negative hyperbole which has also been called minutio in the rhetoric tradition.17 (8) in the indicated context means ‘the person referred to was drunk only to a small extent’ and thus obviously is in contradiction with the context of shared knowledge in which it is uttered. The speaker might also have said: (9) He was tipsy. Both utterances are obvious violations of the first maxim of Quality. So, given that the general Cooperative Principle is still considered to be adhered to by the speaker, they must be re-interpreted in their opposite sense as ‘the person referred to was legless and did not know what he was doing’. (8) and (9), therefore, are examples of indirect propositions which can be detected by rational re-interpretation (cf. Burkhardt 1986: pp. 157; 1990). Pragmatically, i.e. with regard to their illocution, the utterances have still to be considered as assertions, but may serve the purposes of joking, downplaying or appeasement, and their perlocutions must be determined correspondingly. What, now, is the difference between meiosis and euphemism? Only a very fine line separates the two. They are both strategies used for avoiding an adequate description of the facts, but while in meiosis this is done openly and with the strong belief of the partial untruth being perceived by the addressee, ————— 17
British people are commonly held to have a tendency towards using understatements of this kind.
Euphemism and truth
365
in euphemism this is done secretly and with the hope that the inadequacy will remain undetected or at least pass unnoticed. Meiosis, though it does not involve the re-interpretation of the indicated illocution, only requires a partial re-interpretation of the proposition. Therefore, it is a form of indirect speech act (of the propositionally indirect kind; see Burkhardt 1986: 390–5) which euphemism, by contrast, is not. Like meiosis, euphemisms do not operate on the illocutionary but on the perlocutionary level: one cannot sensibly say I hereby euphemise that... as one cannot say I hereby lie that ... either. By contrast, however, in everyday linguistic practice euphemistic utterances do not require any re-interpretations of their propositions at all. Such clarifications would even be regarded as counter-productive by their users as they would ruin the intended perlocutionary effect. If euphemistic utterances were meant to create in the hearer/reader a certain interpretation they could be taken as indirect speech acts as well. But as they are actually used to avoid such reinterpretations they must be taken as weak forms of lying and, therefore, as direct assertions which unfortunately are partly untrue. One exception to this rule are abstracting euphemisms which do not flout the maxims of Quality but the first maxims of Quantity (“Make your contribution as informative as is required (for the current purposes of the exchange)”) and Manner (“Avoid obscurity of expression”) instead. An utterance like (10) Tony Blair conceded last night that western intervention in Iraq had been a disaster. (http://politics.guardian.co.uk/iraq/story/0,,1951267,00.html) may be called a euphemism because the British Prime Minister used the hyperonym intervention instead of the hyponym invasion which would have been more appropriate to describe what happened in the Gulf region since March 20, 2003. By using the hyperonym Blair’s utterance is less informative than it could be and, therefore, violates the first maxim of Quantity. As it may also be said to be more obscure than necessary, it can be taken as a violation of the first maxim of Manner as well. The same applies for syntactic euphemisms such as (1) – (6) in which the requirements of informativeness and clearness are flouted. As rule of thumb, therefore, one may summarize that while positivising euphemisms are first and foremost violations of the maxims of Quality, syntactic and abstracting euphemisms flout the (first) maxims of Quantity and Manner. 5. Euphemism and truth In a radical sense no sentence, or proposition, can ever be true or false. For as it does not mirror a state of affairs, i.e. a section of reality in all its aspects and components like a picture, it will always be limited to focusing, by its word-
366
Armin Burkhardt
ing, on those aspects and components which are considered to be particularly important by the speaker. From this follows, on the other hand, that every sentence ignores the greater part of the reality it refers to. In this sense all sentences are partial lies (though they may be partly true at the same time). To clarify this let us consider the following examples: (11) As Peter’s father, an unskilled worker, was fired without notice by his boss in his office at 10.45 a.m. last Friday, the Williams family has to live on £500 unemployment benefit now. (12) As Peter’s father was fired without notice by his boss last Friday, the Williams family has to live on unemployment benefit now. (13) As Peter’s father was fired, the Williams family has to live on a very small income now. (14) As Peter’s father lost his job, the Williams family is poor. (15) As Peter’s father was made redundant, the Williams family is economically disadvantaged. Sentences (11) – (15) may all be true (or false), but they differ with regard to the concreteness and detailedness of the given information. However, while (11) – (14), with decreasing amount of information, are still descriptive, (15) will be considered to be euphemistic as it contains expressions (make redundant, economically disadvantaged) which shed a positive light on what will generally be taken to be an unfortunate state.18 Make redundant, instead of fire (or dismiss) switches the focus from the authoritarian decision of notice by the employer to the alleged lack of necessity of a certain job and presents the dismissal as a consequence of economic developments and the employer as irresponsible. And economically disadvantaged substitutes the social misery implied in the concept of poverty by a vague indication to abstract “disadvantages” which are the consequences of poverty. In diverting the attention from authoritarian decision to economic necessity and from misery to the lack of advantages, (15) does not describe the state of affairs its main lexical components refer to in the expected and straightforward way, but gives it a positive or benevolent reading. The euphemistic utterance does not convey the expected information in the merely descriptive way. And, therefore, because of this subjective and slightly deviating element in them, such utterances cannot be considered to be just true, though they are not completely false either. In his article “Politics and the English Language” George Orwell writes: In our time, political speech and writing are largely the defence of the indefensible. Things like the continuance of British rule in India, the Russian purges and deportations, the dropping of the atom bombs on Japan, can indeed be defended, but
—————
18
As soon as a euphemism is recognised as such by the public, however, it becomes useless and will be replaced by a new one. Therefore, one may expect such a process for the words redundancy and make redundant which, in a period of massdisoccupation, evoke horror in everybody who hears them for the very near future.
Euphemism and truth
367
only by arguments which are too brutal for most people to face, and which do not square with the professed aims of political parties. Thus political language has to consist largely of euphemism, question-begging and sheer cloudy vagueness. Defenceless villages are bombarded from the air, the inhabitants driven out into the countryside, the cattle machine-gunned, the huts set on fire with incendiary bullets: this is called pacification. Millions of peasants are robbed of their farms and sent trudging along the roads with no more than they can carry: this is called transfer of population or rectification of frontiers. People are imprisoned for years without trial, or shot in the back of the neck or sent to die of scurvy in Arctic lumber camps: this is called elimination of unreliable elements. Such phraseology is needed if one wants to name things without calling up mental pictures of them. Consider for instance some comfortable English professor defending Russian totalitarianism. He cannot say outright, ‘I believe in killing off your opponents when you can get good results by doing so’. Probably, therefore, he will say something like this: ‘While freely conceding that the Soviet régime exhibits certain features which the humanitarian may be inclined to deplore, we must, I think, agree that a certain curtailment of the right to political opposition is an unavoidable concomitant of transitional periods, and that the rigours which the Russian people have been called upon to undergo have been amply justified in the sphere of concrete achievement’. The inflated style is itself a kind of euphemism. A mass of Latin Words falls upon the facts like soft snow, blurring the outlines and covering up all the details. The great enemy of clear language is insincerity. When there is a gap between one’s real and one’s declared aims, one turns as it were instinctively to long words and exhausted idioms, like a cuttlefish squirting out ink. (Orwell 1967: 153–4)
For Orwell, as demonstrated also by his cynical conception of totalitarian “Newspeak”, euphemism is a form of insincerity and, therefore, a hiding (part of) one’s thoughts. As a manifestation of unclear language it is unacceptable, at least in the field of politics and mass media communication. If euphemisms are insincere they must be considered a form of lying. And in order to be lies they must be assertions which can be true or false. As it is the defining characteristic of euphemistic utterances to conceal at least part of the relevant information, they can never be true in the strict sense. However, they may not be considered deliberate lies either because part of the information conveyed by them will always be correct. They are Janus-headed, a cross between telling the truth and lying. This why Leinfellner (1971: 42) defined them as “partial lies”. Euphemistic sentences like (16) It didn’t turn out well, and now one must be prepared for a new situation. (17) In any case the present market condition gives rise to discussions. (18) At the moment we live through a severe military strain in the east. are true, but (as violations of the first maxim of Quantity) they do not give the complete information which is required or expected or even obvious:
368
Armin Burkhardt
(16’) We lost the match, and we must look forward to being relegated. (17’) In any case our economy is in a recession. (18’) Our army is in the retreat. In the case of all these examples the truth conditions of their propositions are fulfilled. But while the propositions of (16) – (18) remain, in a way, defective, (16’) – (18’) will at least be considered to be ‘truer’, for they frankly describe the relevant facts and give the full information which is known to the speakers of (16) – (18) and expected (with regard to its degree of precision) by their listeners. The same applies for sentences containing lexical euphemisms: (19) The U.S. Army calls for contractor Peak Systems to design and fabricate a light-based immobilisation system/deterrent device and integrate it with an unmanned aerial system. (Cf. http://www.livescience. com/scienceoffiction/070227_beam_system.html) This example sentence may be true as the truth conditions of its proposition seem to be fulfilled but there is no doubt, that a light-based immobilisation system must be classified as a weapon because it will be an instrument to exert violence against others. By calling it system or device the writer focuses on its technical aspects while ignoring its aggressive character. The information given, therefore, is partly more abstract than necessary and does not include reference to a relevant component. Thus, it can be true only on an abstract level, though one could also say that it is epistemologically true but pragmatically false or rather: inadequate. With positivising euphemisms, which violate the maxim of Quality, it is a quite different matter. As their positive connotations are part of the respective proposition they must also be part of its truth conditions. E.g. the invented sentence (19) Freedom-fighters tried to pacify the protectorate denotatively expresses that partisans tried to conquer a certain region (which would be true if, in fact, they tried to do so). It contains three positivising lexical euphemisms which, by describing the corresponding state of affairs in a way which is more positive than it deserves, make the proposition partly untrue. This means that utterances containing positivising euphemisms per definitionem cannot be taken as assertions that tell nothing but plain truth. Euphemistic utterances, be they syntactically or lexically euphemistic, for the most part are semantically incomplete with regard to the state of affairs they describe. In this general form this applies for all sentences, for it has been shown above that radical completeness can never be achieved. Therefore, another criterion must be added: euphemistic utterances are incomplete with regard to shared expectations. This means that, in contrast to scientific state-
Euphemism and truth
369
ments19, the truth or falsity of a euphemistic proposition can only be judged from the point of view of shared knowledge or common consent with regard to the completeness of the relevant information. Euphemisms, therefore, can be true (or false) only against the background of a certain norm of expectation. However, one must distinguish between the truth or falsity of abstracting and positivising euphemisms. While the former are as true as their noneuphemistic equivalents, but only less concrete or informative with regard to the required degree of veracity, the latter are always partly false because they counterfactually add a positive feature to a deplorable propositional content. As utterances that are partly true and partly false or vague and misleading euphemisms cannot only be assessed from an epistemological point of view20, but must first of all be examined from a moral standpoint (which will never be free of ideological implications). This is the reason why they are often object of public discussions or chosen as examples of the worst word of the year (“Unwort des Jahres”), for instance. That e.g. ethnic cleansing21 was chosen as the worst word of the year 1992 in Germany was founded on the ethical argument that the metaphor of “cleansing” must be considered as immoral with regard to human beings, for it compares human beings to dirty clothes which need washing and conceals the true referent of the expression which is the murder and expulsion of people for racist reasons. 6. Conclusion Euphemisms are expressions that are used to hide unwelcome associations. A “veiling” and a “concealing” subfunction must be distinguished. While the former are used in the realm of taboo and may have a moral foundation, the latter must be taken as ways of deliberate deception in public discourse. Euphemisms have different forms of appearance: syntactic and lexical. The lexical forms are either abstracting or positivising. Via conventionalisation the use of both may lead to semantic change and thereby provoke the emergence of new euphemisms. This process may cause a historical chain of euphemisations which Pinker has called “euphemism treadmill”. Though euphemisms are violations of the Gricean maxims of Quality on the one hand and of the maxims Quantity and Manner on the other, they may not be considered as —————
19
20
21
According to the respective theory of truth scientific statements should either correspond to verifiable facts or be consistent with a theoretical framework which is commonly accepted. Epistemological problems cannot be dealt with in more detail here. For my own account of truth see Burkhardt (1985). This would be an illusion anyway for in the political sphere “the opinion as to which designation appears as neutral and adequate is dependent on one’s own political or ideological point of view” [“ist die Auffassung, welche Benennung als neutral und angemessen erscheint, abhängig vom eigenen politischen Standpunkt, von der eigenen Ideologie”] (Forster 2006: 201; my translation, A.B.) I.e. the German equivalent ”ethnische Säuberung”.
370
Armin Burkhardt
indirect speech acts, for their purpose is to block rather than to provoke reinterpretations in their recipients. By contrast, all euphemisms are forms of partial lying in the sense that they do not conform to the standards of expected explicitness. The positivising ones, in particular, are even more mendacious because they add a positive valuation to a deplorable fact. In public discourse, especially in politics, euphemisms are a strategic means of persuasion and at the same time a psychological rampart which an evil-doer or the teller of bad news puts round himself . There is always the danger that euphemisms secretly become an impediment to thinking. If euphemistic words or utterances are successfully introduced into the common use of language, displacing hitherto competing linguistic signs, and if the lexical extenuations can be successfully introduced into the speakers’ lexicon as ordinary designations so that, in extreme cases, even the affected people themselves begin to speak of their missions, their golden age or economic disadvantages or tell others that they were made redundant, euphemism has come to cover our perception like a veil of mist, and it becomes difficult, even for their inventors, to see the object or state of affairs referred to in a less positive light. “Many an evil deed would have been left undone”, Kainz (1972: 386–7) writes, “if language had forced the perpetrator to look in its face by placing at his disposal only the plain direct designation”22. Euphemisms, therefore, may not only be understood as a means of “intentional deception” (Heringer 1990: 56) but also as elements of unintentional selfdelusion (cf. also Dieckmann 1964: 108). References Allan, Keith, Kate Burridge: Forbidden Words. Taboo and the Censoring of Language. Cambridge–New York–Melbourne–Madrid–Cape Town–Singapore–São Paulo: Cambridge University Press 2006. Ayto, John: Euphemisms. Over 3,000 Ways to Avoid Being Rude or Giving Offense. London: Bloomsbury 1994. Beda Venerabilis: De Schematibus et tropis II: De tropis. In: Corpus Christianorum. Series Latina 123A, 1 (1975), 151–71. Bohlen, Andreas: Die sanfte Offensive. Untersuchungen zur Verwendung politischer Euphemismen in britischen und amerikanischen Printmedien bei der Berichterstattung über den Golfkrieg im Spannungsfeld zwischen Verwendung und Mißbrauch der Sprache. Frankfurt/Main–Berlin–Bern–New York–Paris–Wien: Peter Lang 1994. Burkhardt, Armin: “Wittgensteins Wahrheit oder ‘Kann man die Wirkung der Erfahrung auf unser System von Annahmen leugnen?’ (ÜG § 134)”, in: R.H. Chisholm, J.C. Marek, J.T. Blackmore, A. Hübner (eds.): Philosophie des Geistes – Philosophie des Psychologie. Akten des 9. Internationalen Wittgenstein Symposiums, 19. bis 26. August 1984, Kirchberg/Wechsel (Österreich). Wien: Hölder-PichlerTempsky 1985 (= Schriftenreihe der Wittgenstein Gesellschaft 11), 491–5. – Soziale Akte, Sprechakte und Textillokutionen. A. Reinachs Rechtsphilosophie und die moderne Linguistik. Tübingen: Max Niemeyer Verlag 1986.
—————
22
My translation, A.B.; the last noun phrase in the German original is “die unverblümte Direktbezeichnung”.
Euphemism and truth –
371
“Speech act theory – the decline of a paradigm”, in: Armin Burkhardt (ed.): Speech Acts, Meaning and Intentions. Critical Approaches to the Philosophy of J.R. Searle. Berlin–New York: W. de Gruyter 1990, 91–128. – “Können Wörter lügen?”, in: Universitas. Zeitschrift für interdisziplinäre Wissenschaft 47 (1992), 831–40. – “Deutsche Sprachgeschichte und politische Geschichte”, in: Werner Besch, Anne Betten, Oskar Reichmann, Stefan Sonderegger (eds.): Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung. Berlin– New York: de Gruyter 1998, 2. vollständig neu bearbeitete u. erw. Aufl., 98–122. Chamizo Domínguez, Pedro J.: “La función social y cognitiva del eufemismo y del disfemismo”, in: Panacea, vol. V, Núm. 15 (2004), 45–51. – “Some theses on euphemisms and dysphemisms”, in: Studia Anglica Resoviensia 25 (2005), 9–16. – “Linguistic Interdiction: Forbidden words and the censoring of language”, in: Language Sciences 2007 (submitted). – , F. Sánchez Benedito: Lo que nunca se aprendió en clase: eufemismos y disfemismos en el lenguaje erótico inglés. Foreword by Keith Allan. Granada: Comares 2000. Crespo Fernández, Eliecer: “The Language of Death: Euphemism and Conceptual Metaphorization in Victorian Obituaries”, in: SKY Journal of Linguistics 19 (2006), 101–30. Dieckmann, Walther: Information oder Überredung. Zum Wortgebrauch der politischen Werbung in Deutschland seit der französischen Revolution. Marburg: Elwert 1964. Dietl, Cora: “Euphemismus”, in: Ueding, Gerd (ed.): Historisches Wörterbuch der Rhetorik. Vol. 3: Eup – Hör. Tübingen: Max Niemeyer Verlag 1996, col. 1–10. Dubois, Jacques, Francis Edeline, Jean-Marie Klinkenberg, Philippe Minguet, François Pire, Hadelin Trinon: Allgemeine Rhetorik. Übersetzt und herausgegeben von Armin Schütz. München: W. Fink 1974. Fabricius, Johann Andreas: Philosophische Oratorie. Das ist: vernuenftige Anleitung zur gelehrten und galanten Beredsamkeit. Leipzig 1724 (Reprint Kronberg 1974). Forster, Iris: “Lexikalische Verführer – euphemistischer Wortschatz und Wortgebrauch in der politischen Sprache”, in: Kilian, Jörg (ed.): Sprache und Politik. Deutsch im demokratischen Staat. Mannheim–Leipzig–Wien–Zürich: Dudenverlag 2006, 195–209. Fritz, Gerd: Einführung in die historische Semantik. Tübingen: Niemeyer 2005. Grice, H. Paul: “Logic and conversation”, in: Cole, Peter, Jerry L. Morgan (eds.): Speech Acts. (Syntax and Semantics. Vol. 3). New York–San Francisco–London 1975: 41–58. Heringer, Hans Jürgen: “Ich gebe Ihnen mein Ehrenwort.” Politik – Sprache – Moral. München: Beck 1990. Holder, R.W.: A Dictionary of Euphemisms. Oxford–New York: Oxford University Press 1995. Kainz, Friedrich: Über die Sprachverführung des Denkens. Berlin: Duncker & Humblot 1972. Leinfellner, Elisabeth: Der Euphemismus in der politischen Sprache. Berlin: Duncker & Humblot 1971. Longman Dictionary of Contemporary English. With New Words Supplement. Munich: Langenscheidt-Longman 2001.
372
Armin Burkhardt
Luchtenberg, Sigrid: Euphemismen im heutigen Deutsch. Mit einem Beitrag zu Deutsch als Fremdsprache. Frankfurt/Main: Lang 1985. Orwell, George: “Politics and the English language”, in: Orwell, George: Inside The Whale and Other Essays. Harmondsworth, Middlesex: Penguin Books 1967, 143– 57. – Nineteen Eighty-Four. A Novel. Harmondsworth, Middlesex: Penguin Books 1973. Paul, Hermann: Principles of the History of Language. Translated from 2nd edition of the original by H.A. Strong. London: Sonnenschein & Co. 1888. – Prinzipien der Sprachgeschichte. Tübingen: Niemeyer 1975, 9., unveränderte Auflage. Pinker, Steven: “The Game of the Name”, in: The New York Times, April 3, 1994 (http://pinker.wjh.harvard.edu/articles/media/1994_04_03_newyorktimes.pdf). Rada, Roberta: Tabus und Euphemismen in der deutschen Gegenwartssprache. Mit besonderer Berücksichtigung der Eigenschaften von Euphemismen. Budapest: Akadémia Kaidó 2001. Quintilianus, Marcus Fabius: The Orator’s Education. Books 1–10. Edited and translated by Donald A. Russell. Cambridge, Mass.–London: Harvard University Press 2001. Vossius, Gerhard Johann: Commentariorum rhetoricorum sive oratoriarum institutionum libri VI. Leiden 1630 (Reprint Kronberg 1974). Zöllner, Nicole: Der Euphemismus im alltäglichen und politischen Sprachgebrauch des Englischen. Frankfurt/Main–Berlin–Bern–New York–Paris–Wien: Peter Lang 1997. http://en.wikipedia.org/wiki/Euphemism
Kenneth Holmqvist (Lund, S) / Jarosław Płuciennik (Łódź, PL)
Princess Antonomasia and the Truth: Two Types of Metonymic Relations
0. 1. 2. 3. 4.
Introduction Metonymic antonomasia Metaphorical antonomasia Quixotism of culture Conclusion
0. Introduction Antonomasia is generally regarded as a minor figure of speech and therefore not treated as a princess among the figures of speech. However, in this chapter, we would like to give antonomasia a different place in the tropical ranks and elevate her to the status of Princess, following the advice of Cervantes who wrote a chapter on Princess Antonomasia in his well-known novel Don Quixote (Cervantes 1981, vol. 2, chapt. 38). We will try to show that antonomasia might be rhetorically more important than has hitherto been assumed. In many literary and rhetorical dictionaries antonomasia is regarded as a rhetorical figure similar in some respects to metonymy. It is based on the substitution of regular nouns by proper names and vice versa, as indicated by its etymology: in Greek anti means ‘instead’ and onomazein means ‘to name’; in Latin, antonomasia is called pronominatio. It is therefore based on two different kinds of semantic mechanisms. One mechanism is involved when we use a regular noun instead of a proper name, e.g. “the Thunderer” instead of “Zeus”. This can explain why antonomasia is regarded as a kind of metonymy: a name for one function or aspect of an object is regarded as a typical representation of the object. However, the second mechanism is involved when a proper name replaces a regular noun, e.g. “Don Quixote” instead of “an idealist”. This mechanism is quite different from metonymical relations: a name for a typical representative of a set is used in order to refer to all representatives of the set. Most definitions of antonomasia are based on the classical substitution theory, which also pervaded definitions of metaphor for a long time. According to this theory there is always a proper meaning of a word and an improper meaning of a word, that is, to give an example for metaphor, you replace the proper word fish by the improper word arrow in A brilliant arrow shot through the water. In the case of antonomasia, according to this theory, there
374
Kenneth Holmqvist / Jarosław Pluciennik
is always a proper name for a phenomenon and a less proper name for it. Tzvetan Todorov (1967), for example, regards antonomasia as an anomaly just like hyperbole, irony, metaphor and metonymy. In his view, we can talk of anomaly in these figures and the semantic mechanism refers to a relationship between a sign and a reference. But why do people use a less proper name for a more proper name? What kind of aims do speakers have when they use either kind of antonomasia in certain communicational, social and cultural contexts? 1. Metonymic antonomasia In the first kind of antonomasia (antonomasia1), which is similar to periphrasis, a proper name is substituted by an epithet. This kind of antonomasia was often used in ancient times when talking about mythical and epical characters. For instance, Zeus and Achilles were frequently described by using the epithets the Thunderer and the swift-footed respectively. This usage of antonomasia can be explained if we take into account the formulaic character of oral culture (Parry 1971). In every oral culture a bard must have prefabricated material available in order to remember particular elements in his stories. Using prefabricated material is a much wider cultural phenomenon, however. It is found in many religions, where different variations of a particular god are presented in a social imagery function. For instance, in Polish Catholicism, Gromniczna (adjective derived from the Polish name for thunder meaning literally ‘related in some respect to thunders’) is a name for one Madonna used for her function as protection from thunders. So, one of the main functions of antonomasia1 could be mnemonic. Antonomasia1 could thus be used in order to remember better the subject of a story and the main function of a character in that story. A metonymic relationship is involved here: the property of an individual (see Barcelona 2002; see also Barcelona, this volume, 273–97) that best characterizes an individual defines the individual by pointing out a main role in a story. It may also work the other way around, as is the case in many religions, where a defining property can create an individual and one god multiplies in many avatars. Attributing a particular property to an individual stabilizes both the meaning and the ontology of that individual. One can argue that this metonymic function of antonomasia1 explains the stabilizing social character of prefabricated material. The formulas used in artistic performances stabilize social thinking about characters from mythology or religion. It is striking that, when for instance one god has several different names, an antonomastic epithet also stabilizes the perspective from which that particular god is viewed. Changing the perspective can cause a further process of idolizing the particular property of an individual. The idolizing feature of antonomasia1 can be illustrated by an example from Don Quixote by Cervantes where the main character calls his horse
Princess Antonomasia and the Truth
375
“Rosinante”. This is a meaningful name: rocin means ‘jade’, and antes means ‘before’. So Rosinante is a horse which is in a much better condition now than before and can be called therefore now “the first steed in the world” This example of antonomasia1 is a very interesting case because it can also be used as antonomasia2 in which a typical example of a set can represent all the examples of the set: a car in Graham Green’s novel Monsignor Quixote is called “Rosinante”. It is not a horse, but the condition of the car is very similar to that of its prototype from Don Quixote. The name of the horse is transferred to the car because they share both the ‘real’, and the imagined attribute of Don Quixote’s horse, namely that of being old and worn out and that of serving as a very good vehicle. The function of antonomasia1 can also be understood by taking into account its relationship to reference. Usually, antonomasia1 gives us an expression which has exactly the same reference as the proper name. If a Duke is called “the most mighty lord”, this is a periphrasis and underlines his one hyperbolized feature, but the referent of this name is identical with that of the title Duke or that person’s proper name. 2. Metaphorical antonomasia In antonomasia2 we find a different mechanism at work. Now we use a proper name in order to describe a very different kind of reference. The reason for this is that antonomasia2 highlights analogical features shared by the paragon and the instance of the type. Antonomasia2 was described by Lakoff in a chapter on Idealized Cognitive Models (1987: 87–8) in terms of paragons (together with social stereotypes, typical examples, ideals and generators). He pointed out that “we also comprehend categories in terms of individual members who represent either an ideal or its opposite.” As a kind of metonymic semantic mechanism, antonomasia2 (as paragons) is also mentioned in Barcelona (2002). Paragons have a very important role in society: they mark an area of ambition and imitation (cf. Girard 1965). In an antonomasia2 such as Don Quixote, we can observe a mechanism common in all antonomasias2: namely, their “quixotism”. It is striking that we can say that Madame Bovary is a quixote of the 19th century, or a female quixote. What is common to both these literary characters is their fascination with culture and their fondness of imitating paragons. Imitating paragons is based on a process of cultic identification, a process explored, for example by Jauss (1965, cf. also chapter on imitation in Don Quixot: Melberg 1992). In ancient times, objects of identification usually had a cultic or ritual status. It has however become clear that, not only in ancient times but in many different historical epochs, this type of cultic identification has played an important role in popular culture. Superman and characters in computer games as well as various religious heroes provide good examples.
376
Kenneth Holmqvist / Jarosław Pluciennik
Imitation and identification make use of human mimetic skills in order to disseminate cultural representations. The result is repetition, which is merely another name for imitation (cf. Melberg 1992). However, repetition is not the same as identity. As Dan Sperber has pointed out: What human communication achieves in general is merely some degree of resemblance between the communicator’s and the audience’s thoughts. Strict replication, if it exists at all, should be viewed as just a limiting case of maximal resemblance, rather than as the norm of communication […]. A process of communication is basically one of transformation. The degree of transformation may vary between two extremes: duplication and total loss of information. Only those representations that are repeatedly communicated and minimally transformed in the process will end up belonging to the culture. (Sperber 1996: 83)
The process of communication described by Sperber and Wilson (1986) can be compared with the process of “becoming infected”. This communicative infection process has anthropological foundations, as culture depends on the epidemic spread of representations (cf. Sperber 1996). In this spreading process, mimesis, the basic human instinct to imitate, plays a crucial role. Antonomasia2 can be regarded as a case of mimesis of this type and so can its most famous examples. Don Quixote and Madame Bovary are the best examples of fictional characters that are infected by books and literary imaginative worlds. Joseph Conrad (1919: 186) once said that Don Quixote’s name builds on an allusion to the book-born madness of the knight. The personal identity of both characters is based on fiction. This is a part of the entire process of enculturation. Culture, according to one of the proponents of the simulation theory, Stephen Turner, is not inherited or innate, but acquired through experience. Culture is a set of symbolic roles and ritualized actions. Games of make-believe form a part of enculturation by taking attitudes of others that refer to stereotyped roles and symbols (Turner 2000: 112). Culture nowadays does not consist only of literature, however. In addition to the very literary antonomasias2, such as bovarism and quixotism, one can also find comparisons with paragons from other spheres of culture: Lyotard is a pope of postmodernism (Płuciennik 2000: 133), Bush is no Demosthenes (Enos 1996: 445); and we can buy the Cadillac of vacuum cleaners (Lakoff 1987: 87). Antonomasia2 can be regarded as the product of the imitative nature of human beings. We simply need paragons and prototypes. It is possible that they are a means for us to identify with something more sublime than we are ourselves. This desire is described in a prototypical way in Don Quixote, in Chapter 38 of the second volume. Here we find a story built around a character called Princess Antonomasia. The story is invented in order to mock Don Quixote and his enthusiasm for books on knight-errantry. The countess Trifaldi (which is another example for an antonomasia1 and means ‘having a three-striped gown’) tells a story about Princess Antonomasia. The Countess was supposed to guard the Princess and her decency, but instead she was
Princess Antonomasia and the Truth
377
seduced by a poet who read a lot of poetry to her. The words of his poetry seemed like pearls to her and his voice was like syrup. She finally decided that poets should be banished from all states, as Plato had postulated. In reply, the poet sang: Come Death, so subtly veiled that I Thy coming know not, how or when, Lest it should give me life again To find how sweet it is to die. – and other verses of the same sort, such as enchant when sung and fascinate when written. And then, when they condescend to compose a sort of verse that was at that time in vogue in Kandy, which they call seguidillas! Then it is that hearts leap and laughter breaks forth, and the body grows restless and all the senses turn to quicksilver (mercury). And so I say, sirs, that these troubadours richly deserve to be banished to the isles of the lizards. Though it is not they that are in fault, but the simpletons that extol them, and the fools that believe in them; and had I been the faithful duenna I should have been, his stale conceits would have never moved me, nor should I have been taken in by such phrases as ‘in death I live,’ ‘in ice I burn,’ ‘in flames I shiver,’ ‘hopeless I hope,’ ‘I go and stay,’ and paradoxes of that sort which their writings are full of. And then when they promise the Phoenix of Arabia, the crown of Ariadne, the horses of the Sun, the pearls of the South, the gold of Tibar, and the balsam of Panchaia! Then it is they give a loose to their pens, for it costs them little to make promises they have no intention or power of fulfilling.
The paradoxical way in which the Countess speaks, as described above, can be regarded as a case of sublime discourse, a type of discourse that employs rhetorical figures in order to evoke unimaginable imaginings and enthusiastic and ecstatic emotions. Indeed, if we look at the history of reflection on the sublime, we see many traces of an antimimetic theory of the sublime (cf. Holmqvist & Płuciennik 2003). This antimimetic tendency is already present in Pseudo-Longinus’ treatise. In Chapter 36, he develops an account of nature and the function of language. In works of art (craft) we look for likeness, but in language, as in nature, we look for something transcending human nature. Language is creative as is nature and it has to reveal what transcends human existence. Poetry of the sublime is an evocation of the supernatural. The antimimetic aspects of the theory of the sublime are also apparent in Burke’s A Philosophical Enquiry into the Origin of our Ideas of the Sublime and the Beautiful from 1756/57, when he writes: […] poetry, taken in its most general sense, cannot with strict propriety be called an art of imitation […]. Nothing is an imitation further than as it resembles some other things and words undoubtedly have no sort of resemblance to the ideas for which they stand. (Burke 1958: 172–3)
The same applies for identifications of other phenomena: it is always the case even if we treat somebody with irony, as in Your Romeo phoned (Enos 1996: 445). Culture provides us with all sorts of prototypes and paragons. These
378
Kenneth Holmqvist / Jarosław Pluciennik
paragons have cognitive features, not only linguistic ones. This becomes clear when we examine a pictorial antonomasia with a political allusion: one elected official in Poland (Andrzej Lepper) has been depicted on a magazine cover as Hitler with the little moustache which is characteristic of the Nazi leader. An caption below the photo said: “Heil Lepper!”. Thus antonomasia2 is based on a cognitive mechanism of seeing something as something else, which is also the basis for metaphor. This way of seeing something as something else can find expression in a picture, but it can also be presupposed as a mental image: if I address somebody as if he/she was Hitler this is the same as calling that person “Hitler”. A crucial problem for the phenomenon in question is the relationship between presuppositions and cognitive models or, broadly speaking, a relationship between presuppositions and extra-textual phenomena: other texts, clichés, cognitive schemas, social stereotypes, frames, scripts etc. All these phenomena can be regarded, following de Beaugrande & Dressler (1990: 127–8), as phenomena of intertextuality (cf. Płuciennik 1995)1. 3. Quixotism of culture As a product of the cultural instinct to imitate others, antonomasia also has another feature: it is historically situated and context dependent. In a book on sentimentalism and quixotism, Wendy Motooka writes: Focusing on Tom Jones, I show that Fielding, like his contemporaries, associates quixotism with specific political and intellectual conflicts – women’s equality, empirical method, social diversity, Jacobitism – and that he responds to these conflicts by embracing sentimentalism. He complicates quixotism, however, by representing it not just as malady of readers, but as an affliction of authors as well. Fielding tries to mock quixotism as feminine, irrational and peculiar, and to applaud rational authority as masculine, cerebral and general. Yet because he approaches this task already committed to moral empiricism – the empirical study of invisible things – he finds himself unable to represent his own authority as nonquixotic. (Motooka 1998: 28)
In this passage, Quixotism is associated with attributes such as feminine, irrational and peculiar. But in a renowned book about the rise of the novel, Ian Watt (1957: 97) writes that Don Quixote is a paradigm of rash nobility and blindness of knight-errant idealism. In the Romantic period, Don Quixote functioned as the paragon of a romantic hero. According to Motooka, “Locke uses the trope of quixotism to smear his political opponents with the taint of ————— 1
But we can follow Lakoff (1987) as well and talk here of a relation between presuppositions and Idealized Cognitive Models (ICMs) (we describe this connection in Holmqvist & Pluciennik 1996). Lakoff’s concept of ICMs covers frames as postulated by Fillmore, schemata as postulated by Rumelhart, scripts as postulated Shank and Abelson, frames as postulated by Minsky, and mental spaces as postulated Fouconnier and Putnam’s stereotypes (cf. Holmqvist & Pluciennik 2002a).
Princess Antonomasia and the Truth
379
enthusiasm, tyranny and violence” (1998: 41). It was quite common in the 18th century to see Don Quixote as enthusiastic or even mad. On the other hand Conrad writes in Nostromo: There is a curse of futility upon our character: Don Quixote and Sancho Panza, chivalry and materialism, high-sounding sentiments and a supine morality, violent efforts for an idea and a sullen acquiescence in every form of corruption. We convulsed a continent for our independence only to become the passive prey of a democratic parody, the helpless victims of scoundrels and cut-throats, our institutions a mockery, our laws a farce – – a Guzman Bento our master! And we have sunk so low that when a man like you has awakened our conscience, a stupid barbarian of a Montero – – Great Heavens! a Montero! – – becomes a deadly danger, and an ignorant, boastful Indio, like Barrios, is our defender. (Conrad 1924: 171)
It is obvious that quixotism means different things to different people at different times. In a contemporary book about sentimentalism we can read about Locke: “If a quixote is one who attempts to universalize his or her own peculiar way of thinking, then Locke is more quixotic than Tristram Shandy” (Motooka 1998: 27). One can argue that antonomastic names function as a resource for inculturation (a concept similar to enculturation, see below) of newly encountered phenomena. If we refer to Leonard Cohen as the Lord Byron of rock music (Enos 1996: 445), we treat a popular singer as a famous romantic poet elevating him and popular songs to a higher level of culture. We should also be aware that eponymy, which widens the meaning of a proper name into a common noun (sandwich, ampere, hertz) is not the same as antonomasia, because eponymy lacks this common feature which builds an analogical link between two referents. 4. Conclusion Going back to our initial question, namely why do people use a less proper name for a more proper name, we could propose the following explanation: Antonomasia is essentially humanistic and anti-computational. It highlights the anti-computational nature of language and shows that language might instead be understood as a kind of game of make-believe or as a simulation tool. A computer will not find all the references to the word king, if we call him His Majesty. Both kinds of antonomasia discussed in this chapter require a special kind of historical and contextual knowledge to be understood or indeed ‘deciphered’. Antonomasia can perhaps best be regarded as a cryptographic tool and culture as a kind of cryptographic game of make-believe. It is a game played by those who know against those who do not know. To find the ‘true’ meaning of an antonomasiologically encrypted reference we have to go beyond the linguistically given and engage in the culture in which it is used.
380
Kenneth Holmqvist / Jarosław Pluciennik
During this process of communication understood as a game of makebelieve, we refer to some kind of reality that presupposes the existence of ‘the truth’. In order to refer to the truth we use cognitive models. In the frame of those cognitive models we employ very fuzzy references to imagined realities. A crucial question here is the problem of similarity. Resemblance of the imagined truth to the real truth might be quite accidental and unessential like the similarity of windmills to giants in the case of Don Quixote or the similarity of a mustache on a picture of a politician to a mustache of another historical person, but our search for a match between words and worlds make us seek a motivation even if it does not exist. Establishing a resemblance of reality as a basis for identification might be casual, subjective and even based on illusion or delusion. The mechanisms described by antonomasia2, that is the mechanisms of referring to imagined realities, reveal the fuzzy character of our conceptualization. Our conceptualization processes are not always logical. This fuzzy character may be a scandal to some people who might want to avoid it, especially in the case of some discourses, such as scientific ones, but it seems to be important to keep it in mind when analyzing human thinking and human language in relation to products of human cultures, where the power of imagination might be more important then reality. We should remember that perhaps, as Don Quixote said, “Facts are the enemy of truth!” References Barcelona, Antonio: “On the ubiquity and multiple-level operation metonymy”, in: Lewandowska-Tomaszczyk, Barbara, Kamila Turewicz (eds): Cognitive Linguistics Today. Frankfurt/Main–Berlin–Bern–New York–Paris–Wien: Peter Lang 2002, 207–24. Beaugrande, Robert-Alain de, Wolfgang Ulrich Dressler: Wstep do lingwistyki tekstu. Transl. A. Szwedek. Warszawa: PWN 1990. Burke, Edmund: A Philosophical Enquiry into the Origin of our Ideas of the Sublime and Beautiful. Edited with an Introduction and Notes by J. T. Boulton. London: Routledge & Kegan Paul 1958. Cervantes Saavedra, Miguel de: Don Quixote: the John Ormsby translation, revised, backgrounds and sources, criticism by Joseph R. Jones and Kenneth Douglas (eds). New York: Norton 1981. Conrad Joseph: A Personal Record. Garden City, NJ: Doubleday, Page & Company 1919. – Nostromo: a Tale of the Seaboard. Garden City, NY: Doubleday, Page & Company 1924. Enos, Theresa (ed.): Encyclopedia of Rhetoric and Composition. Communication from Ancient Times to the Information Age. New York–London: Garland Publishing 1996. Girard, René: Deceit, Desire, and the Novel; Self and Other in Literary Structure. Translated by Yvonne Freccero. Baltimore: Johns Hopkins Press 1965. Green, Graham: Monsignor Quixote. Middlesex–New York: Penguin Books 1982.
Princess Antonomasia and the Truth
381
Holmqvist, Kenneth, Jarosław Płuciennik: “Conceptualised deviations from expected normality. A semantic comparison between lexical items ending in -ful and -less”, in: Nordic Journal of Linguistics 19 (1996), 3–33. – “Appearance Markers”, in: Lewandowska-Tomaszczyk, Barbara, Kamila Turewicz (eds): Cognitive Linguistics Today. Frankfurt/Main–Berlin–Bern–New York–Paris–Wien: Peter Lang 2002a, 655–70. – “Image and imagination in the theory of the sublime”, in: Literary Texts and the Arts. Interdisciplinary Perspectives, ed. by C. Federici. Frankfurt/Main–Berlin– Bern–New York–Paris–Wien: Peter Lang 2003, 15–27. Jauss, Hans Robert: “Levels of identification of hero and audience” (translated by Benjamin and Helga Bennett), in: New Literary History 5 (2) (1974), 284–317. Lakoff, George: Women, Fire, and Dangerous Things. What Categories Reveal about the Mind. Chicago: The University of Chicago Press 1987. Longinus: On Great Writing (On the Sublime). Translated by G. M. A. Grube. Indianapolis–Cambridge: Hacket Publishing Company 1991. Melberg, Arne: Mimesis. En repetition. Stockholm/Stehag: Brutus Östlings Bokförlag Symposion 1992. Motooka, Wendy: The Age of Reasons. Quixotism, Sentimentalism and Political Economy in Eighteenth-Century Britain. London–New York: Routledge 1998. Parry, Milmann: The Making of Homeric Verse: The Collected Papers of Milman Parry. Ed. Adam Parry. Oxford: Clarendon Press 1971. Płuciennik, Jarosław: “Presupozycje, intertekstualnosc i cos ponadto…”, in: Bloecki, Wlodzimierz, Wojciech Tomasik (eds.): Poetyka bez granic. Warszawa: Wydawnictwo IBL PAN 1995, 120–33. – Retoryka wznioslosci w dziele literackim. Kraków: Universitas 2000. Sperber, Dan: “The epidemiology of beliefs”, in: Fraser, Colin, George Gaskell (eds.): The Social Psychological Study of Widespread Beliefs. Oxford: Clarendon Press 1990, 25–44. – Explaining Culture. A Naturalistic Approach. Oxford: Basil Blackwell 1997. – , Deirdre Wilson: Relevance. Communication and Cognition. Oxford: Basil Blackwell 1986. Todorov, Tzvetan: Littérature et signification. Paris: Larousse 1967. Turner, Stephen: “Imitation or the internalization of norms: Is twentieth-century social theory based on the wrong choice?”, in: Kögler Hans Herbert, Karsten R. Stueber (eds), Empathy and Agency. The Problem of Understanding in the Human Sciences. Oxford: Westview Press 2000, 103–18. Watt, Ian P.: The Rise of the Novel; Studies in Defoe, Richardson, and Fielding. Berkeley: University of California Press 1957.
Rita Brdar-Szabó (Budapest, H) / Mario Brdar (Osijek, HR)
“Mummy, I love you like a thousand ladybirds”: Reflections on the emergence of hyperbolic effects and the truth of hyperboles*
0. 1. 1.1 1.2 1.3 1.4 1.5 2. 2.1 2.2 2.3 2.4 2.5 3.
Introduction Identifying hyperbole Hyperbole: a trope, a figure, or…? Attempts at defining and typologizing hyperboles How pure is pure hyperbole? Beyond the confines of the classical rhetoric: Some more types of conceptually dependent hyperbolic expressions Hyperbolic effects: a tentative inventory A cognitive account of hyperbolic expressions Hyperboles as over-reachers: The role of mental spaces, scales, domains and mappings Large numbers: A case study on the role of metonymy and metaphors Computer gurus and media talibs: another case study on metaphors as inputs for scalar models Hyperboles as loud liers Case study on Shir HaShirim (Song of Songs) Concluding remarks
0. Introduction The present paper is concerned with the type of expressions traditionally called hyperbole, and popularly known as exaggeration or overstatement. By way of an informal illustration, consider the various types of such expressions in the following monologue from Shakespeare’s Romeo and Juliet. In this ————— *
The first part of the title is the translation from Croatian of an utterance by our nephew, Josip, recorded in spontaneous conversation with his mother at 4;7,10: Mama, volim te kao tisuću bubamara Mummy love-1SG you like thousand ladybirds This example is discussed in some detail in 2.3. below. It is obvious that the first part of the title is at the same time an allusion to Nerlich, Clarke and Todd (1999), one of the trail-blazing investigations into early acquisition of figurative language.
384
Rita Brdar-Szabó / Mario Brdar
scene, swarming with hyperboles, Romeo comments on Juliet’s beauty after first encountering her: (1) O she doth teach the torches to burn bright! It seems she hangs upon the cheek of night As a rich juwel in an Ethiop’s ear; Beauty too rich for use, for earth too dear. So shows a snowy dove trooping with crows, As yonder lady o’er her fellows shows. The measure done, I’ll watch her place of stand, And, touching hers, make blessèd my rude hand. Did my heart love till now? Forswear it, sight, For I ne’er saw true beauty till this night. (Shakespeare, Romeo and Juliet, I.v) It does not take long to detect in this relatively short passage a surprisingly high number of exaggerated expressions which (i) cannot possibly be taken to describe the common-sense world as we know it, or (ii) they happen at worst to be expressions describing extremely unusual states of affairs, or at least statements that may be literally true, but just as well not so, such as she doth teach the torches to burn bright, beauty too rich for use, for earth too dear, So shows a snowy dove trooping with crows, as yonder lady o’er her fellows shows, I ne’er saw true beauty till this night. This concentrated collection of hyperboles aptly illustrates some of the difficulties that a systematic study of this phenomenon poses. At the same time we see how hyperbole interacts with other figures (metaphors, personification, simile, etc.). The main goal we set ourselves below is to attempt to contribute towards answering two related questions concerning hyperbole. The first of these two questions is whether hyperbole can be identified and/or defined as a trope or a figure. The essence of the other question is already intimated in the very title of the present volume: are hyperbolic expressions to be considered as just another case of flouting one or more of Gricean maxims (notably the maxims of quality and quantity) in rhetoric? The answers, it seems to us, in both cases cannot but be extremely complex and are far from being unequivocal. In fact, our position is that these answers cannot be of the yes OR no, but only of the yes-AND-no type. Although such ambiguous answers may seem paradoxical at first blush, and even be considered as an open admission of the failure of a scientific attempt to describe or explain the phenomenon at hand because we are in the end no clearer about it than we were before, we are convinced, and hope to be able to demonstrate below, that we are really no worse off for such answers, as the picture we get in the end is far more realistic and certainly in keeping with observed linguistic facts. We intend to construct a coherent picture of hyperbole by picking up some loose ends of traditional rhetoric and integrating them with the findings of some most recent strands of cognitive linguistic research such as Mental space theory, as developed by Fauconnier
The truth of hyperboles
385
and Turner, and Michael Israel’s work on semantic scalarity or scalar reasoning1, as well as with work on conceptual metaphor and metonymy theory.2 As for the first question, we show that the range of phenomena that have been traditionally considered hyperboles, as well as a number of apparently related phenomena we adduce as examples, do not exhibit any sort of formal common denominator. At best, there obtains a sort of family resemblance type relationship between some of them, while the rest must be recognized as a ragtag group of heterogeneous constructions. The phenomena perceived as hyperboles span a number of levels, from the lexicon to syntax to phonology. What is more, many examples of hyperbole can be shown to instantiate or at least involve some other tropes, such as metaphors, metonymies, antonomasia, similes, etc. One may then become inclined to try to identify hyperboles only in terms of their pragmatic effect. However, it turns out that this approach leads into comparably thorny thickets because what is traditionally considered hyperbole may exhibit different, more or less related, pragmatic effects. Still, we think that this is the most sensible approach. One of our main goals below will be to demonstrate that all these various pragmatic effects that hyperboles may exhibit rest on scalar models constrained by two, possibly three, types of mappings and inferences obtaining within networks of mental spaces (and that metaphors, metonymies, etc., underlying these expressions contribute significantly towards establishing these intricate networks). As for the second question, it is almost trivial to state that hyperbolic expressions are not true. However, we claim that, just like metaphors and metonymies, and other similar phenomena based on conceptual mapping processes, hyperbolic expressions can, perhaps occasionally in an apparently round-about way, come quite close to (revealing) the true nature of things, to the degree that the latter can be subjectively assessed at all. As some of our examples, and the case studies in the final part of the present paper show, hyperbolic expressions often come handy as mental and linguistic shortcuts in cases where a non-hyperbolic expression would require a lot more linguistic material and time to convey roughly the same idea, and for that reason discourse participants who otherwise might be loath to incur somewhat higher processing costs nevertheless make recourse to a figurative expression. In some cases the unfolding of the discourse and the chain of inferences thrives on simultaneous activation of literal and figurative meanings of hyperbolic expressions or their ingredients, which shows that hyperboles may ultimately create a discourse subworld of their own in which the issue of truthfulness is treated in a way that is very different from the way predominant in our everyday common sense world. In sum, they are truthful to the degree that they ————— 1 2
Thanks are due to Frank Brisard, with whom we discussed scalar models and who first made us aware of Michael Israel’s ground-breaking work. Cf. Gibbs & Steen (eds.) 1999; Barcelona (ed.) 2000, and Panther & Radden (eds.) 1999, for state-of-the-art cognitive accounts of metaphor and metonymy theory.
386
Rita Brdar-Szabó / Mario Brdar
help us understand and more easily cope with our everyday common sense world. The paper is structured as follows. In order to shed more light on the nature of hyperbole we start out in Part 1 by considering a number of widely known attempts at formal definition as well as a range of constructions used to exemplify the phenomenon, ranging from those mentioned in classical sources to those discussed in some present-day studies. We also provide a tentative inventory of pragmatic effects of hyperbolic expressions and then raise the question of their veridicality against the background of these various pragmatic effects. In Part 2 we take a look at various cognitive mechanisms, i.e. types of mappings that contribute towards bringing about apparent hyperbolic effects and discuss the issue of their truthfulness. Expositions of our theoretical positions are interspersed in this part with several brief case studies on which we test our approach to hyperbole by trying to work out some aspects of the cognitive processes involved. Finally, our conclusions and some prospects for further research are offered in Part 3. Our main languages of demonstration will be English and German, but throughout the paper we draw data from a variety of other languages, too. 1.
Identifying hyperbole
1.1 Hyperbole: a trope, a figure, or…? As Turner (1998) points out, the Classical Greek word σχηµα (schema) was used as a technical term of Greek rhetoric to refer to a conventional pairing of a form and a meaning, i.e. of a linguistic form and a conceptual pattern. In her discussion of over two and a half thousand years of rhetorical tradition, Fahnestock (1999) points out that the nature of rhetorical devices has been confused and uneven, but their origin is the notion of pairing: The goal of a compendium of figures was [...] to define the formal means for achieving certain cognitive or persuasive functions. One or the other arm of this form-function connection could pivot [...] but the central link should still hold. (Fahnestock 1999)
Unfortunately, schema, (and its Latin equivalent, figura, even more so) in practice often designated just the formal half of a form–meaning pair, because rhetoricians focussed exclusively on the formal side, i.e. linguistic patterns associated with a certain figure. However, the opposite is also attested: in defining some figures rhetoricians relied on conceptual patterns only. Consequently, some figures were often defined as having to do with abstract conceptual patterns but not so clearly with linguistic patterns, since their products are expressible in many ways. On the other hand, some other figures were defined as linguistic forms only. This fundamental divorce of formal proper-
The truth of hyperboles
387
ties from conceptual patterns rears its ugly head in most treatments of hyperbole too. Before we take a closer look at hyperbole, we must note that in the centuries-old rhetorical tradition this divorce is also reflected in the typology of figures; its most interesting facet in the context of our present intentions being the distinction between figures proper and tropes. The distinction between tropes and figures is seen, in the mainstream works of the present-day rhetoric, as paralleling the distinction between paradigmatic and syntagmatic relationships. Tropes are single words or expressions that are used as semantic “deviations”, i.e. they are used figuratively instead of verba propria and stand in a paradigmatic contrast to them. For that reason a trope can frequently be summed up by a single word. Further subclassification of tropes is based on the nature of the semantic relationship between figurative vehicles and their targets, i.e. verba propria. In other words, conceptual criteria are more important in such a subclassification than formal ones. Figures proper are, on the other hand, devices that normally exhibit some syntagmatic “deviation”. It is only understandable that they should generally be more tractable to formal definitions and subclassifications, and that rhetoricians are less precise about or interested in their pragmatic effects. On the basis of such a general characterization (although it is far from being one that enjoys a wide consensus), we might expect hyperbole to be considered a trope rather than a figure proper. We cannot afford here to provide a full historical account of the research tradition on hyperbole, but the general picture one gets from inspecting principal classifications of figures and statements on hyperbole within them is confusing in this respect. Hyperbole is sometimes treated as a trope, sometimes as a figure proper, or as belonging to a special subtype of figures, called figures of sentences (or, as they are also called, particularly in other languages, figures of thought), while there are authors who see it as belonging to all these types. These possibilities will be discussed presently in connection with various definitions, but a more general question must be raised first. Despite the fact that handbooks of rhetoric mention the use of hyperboles in combination with different other tropes no substantial effort has been made to analyze and explain the nature of this interaction in detail, and particularly to check whether what is left after all such cases of interaction with other tropes have been detracted deserves to be treated as a trope on its own. The view seems to be prevalent that “pure” hyperbole is possible, i.e. that it exists as a trope independent of other tropes, but that it often appears in combination with other tropes. Modifying this position slightly, we arrive at the following central question: Can hyperbole indeed appear in isolation, i.e. without the support of other tropes, or does it rather emerge as a secondary effect of some other tropes? As a first approximation to an answer to this question, as well as to the question whether hyperbole is a trope or not, we take a look at various
388
Rita Brdar-Szabó / Mario Brdar
widely available rhetorical classificatory systems, and the examples supporting them. 1.2 Attempts at defining and typologizing hyperboles Our aim in the course of this review of definitions of hyperbole and its various forms of manifestation is certainly not to provide a full account of the research history. The paragraphs that follow are far from being an exhaustive account on this topic in at least two respects. Firstly, we do not venture on an analysis of available theories of hyperbole in the sense of pointing out their advantages or drawbacks. Secondly, we do not yet aim at establishing an inventory of linguistic phenomena leading to hyperbolic effects. Rather, this section of our paper is selective as our main point is to show a tendency emerging in the course of the centuries-old tradition. Not surprisingly, we begin by looking at the definitions of hyperbole and its subtypes in classical sources. One of the first classificatory attempts is by Demetrius, who considers hyperbole to be a figure and distinguishes three types: one is based on a simile (A is like B with respect to some property predicated to both), another on gradation, i.e. a comparison (A is AdjCOMP than B), while the third type is instantiated by expressions denoting something impossible because the entities and/or properties brought thus into relationship are mutually incongruent (cf. Naschert 1998). It seems that the first two types cannot but be realized by syntactic groups such as phrases or whole sentences, and there is no single word or a group of words with a figurative meaning that can simply replace them. This means that these first two types should not count as instances of a trope, as defined above. On the other hand, we note that the third type is defined in quite a different way, no attempt is made to circumscribe it by specifying the form(s) it may take; it just relies on the conceptual pattern of incongruity/impossibility bringing about a certain effect. In the Roman rhetorical tradition the term hyperbole is translated as superlatio. It is claimed in Rhetorica ad Herennium that: Superlatio est oratio superans veritatem alicuius augendi minuendive causa.
or, in other words hyperbole is a presentation model which surpasses the truth in order to make something appear larger or to attenuate it. Cicero adheres very closely to this definition, but Quintilian insists on hyperbole being “decens veri superiectio.” In contrast to Demetrius, he distinguishes as many as five types: expressions stating more than is actually true (aut enim plus facto dicimus, 8,6,68), and possibly denoting something which is, strictly speaking, impossible (actually referred to as gradual intensification or gradation), intensification by means of similarity (similitudo), comparison-based hyperbole
The truth of hyperboles
389
(comparatio), highlighting through certain properties or signs (signa), and metaphor-based hyperbole (translatio). Except for the first type, the rest can be easily approached from the formal side. This first type, again described only in conceptual terms, may then be considered pure hyperbole. It may be illustrated by the following examples: (2) a. Forty thousand brothers Could not, with all their quantity of love, Make up my sum (Shakespeare, Hamlet V.i) b. Vomens frustis esculentis gremium suum et totum tribunal implevit. (Cicero, Phil. II, 25, 63) ‘He vomited undigested remains of food all over himself and the whole court’ c. “Das war das erste vernünftige Wort, das ich von euch höre,” sagte der Sterbende. ‘“That was the first sensible word that I have heard from you”, the dying man said.’ Quintilian (8,6,70), however, allows for the possibility of hyperboles of varying size interacting with each other and thus combining into a hyperbolic utterance. This clearly shows that for him hyperbole can be not just a trope but also a figure of sentence or a figure of thought. Medieval definitions of hyperbole are basically repetitions or slight rewordings of Rhetorica ad Herennium and Quintilian. Hyperbole is thus for Beda Venerabilis “dictio fidem excedens augendi minuendive causa” (cf. Naschert 1998). The same holds for later periods, from the Renaissance to the Enlightenment. Thus in Tudor England Richard Sherry (1550), who treats hyperbole as a figure of sentences, just repeats Quintilian when he says that (hyperbole) “saythe more then the truthe is indeede.” Thomas Wilson (1551) is no different, “There is a figure of Rhetorick, called Hiperbole, that is to say, when a thyng is spoken beyond measure uncredible, and yet is not so largely ment.” A rare example of an attempt at modification is Abraham Fraunce’s treatise The Arcadian Rhetorike (1588). He does not recognize hyperbole as an independent figure but considers it to be “hyperbolicall metaphor”, which is, together with allegory and metaphor proper, a subtype of translatio. In other words, metaphorical hyperbole is seen as a kind of trope. Henry Peacham, on the other hand, in The Garden of Eloquence (1589), in another attempt at modification, considers hyperbole a figure of thought and distinguishes only two types. There is a simple hyperbole, which he illustrates by the following example: (3) Streames of tears gushed out of her eyes, and the greatness of her grief rent her heart in sunder.
390
Rita Brdar-Szabó / Mario Brdar
and a “compared hyperbole”, which appears in a number of subtypes. Cf. some of his examples: (4) a. to call a good man Saint, to call a drunkard a swine b. slower than the snaile, swifter than the swallow c. more unkind than Timon, more cruel then a Scithian, more vigilant than the watchman. One could continue such a review of treatments of hyperbole by stretching it into time periods closer to our time,3 and also extend it to other countries and linguistic communities, but the picture one would get would probably be very similar. What we witness as going on here is a constant attempt to come to grips with formal aspects of hyperbole by identifying one of its manifestations as a reference point, or the central/pure type in terms of its conceptual and pragmatic effect, and then establishing a typology of phenomena that can be isolated formally. In essence, this is, if we may use a metaphor, a process of constantly applying an analytical chisel at what is believed to be a monolith of hyperbole to remove an outward layer and thus uncover its true form, but the result has always been, somewhat disappointingly, a mass of chips lying around the monolith. The monolith now only shows cracks on its surface but is otherwise underlyingly as amorphous as ever. The mass of chips is equally without any recognized intrinsic order, and on closer inspection they turn out to exhibit heterogeneous properties linking them to other stones lying in the nearby field. That this trend is continuing into the present is shown by an ever-growing inventory of constructions treated as hyperboles of one or the other type. We illustrate this point below in sections 1.3. and 1.4. on some examples culled from various sources: some handbooks on rhetoric, a study of language in advertising (Römer), a study on idioms in German, and from a standard grammar handbook of English by Quirk et al. (1985). In the end we even consider some phonological aspects. 1.3 How pure is pure hyperbole? So-called number hyperboles, illustrated in the first part of the title of the present paper, are often cited as examples of the pure type. (But ironically enough, by identifying it as a special subtype it is already set off from the rest of pure hyperboles as well as from combined ones.) Cf. the following examples from Hungarian, German, French and Croatian, respectively: —————
3
Cf. among others Dietz (1999: 220ff.), who, following Quintilian, distinguishes between hyperboles proper on the one hand, and comparative and metaphorical hyperboles on the other.
The truth of hyperboles
391
(5) a. Már százszor mondtam neked, hogy… ‘I have told you a hundred times that…’ b. Ich habe dir schon tausendmal gesagt, dass... ‘I have told you a thousand times that…’ c. Je vous l’ai dit mille fois. ‘I have told you that a thousand times’ (6) Nisam te vidio sto godina. ‘I haven’t seen you for a hundred years.’ Such hyperbolic expressions containing large numbers were occasionally claimed to be cases of number metaphors (cf. Wells 1918). As shown above, most classifications of hyperbole acknowledge the fact that it can be combined with other tropes, mainly with metaphor. If we tried to translate the basic idea of this suggestion into the terminology of the present-day cognitive metaphor theory, we might assume that such large numbers as a source domain suggest excessive amounts or degrees of something in the target domain. What this target domain may be becomes slightly more palpable when we consider the fact that hyperboles can for Blair (1842) be descriptive or emotive, as discussed in 1.5. below. The latter hyperbole expresses the speaker’s or some other participant’s internal reaction or emotion towards a certain entity or an event. Large numbers thus correspond to the intensity of emotive reactions, i.e. the extent to which the speaker admires something, or is irritated by something (as in the above set of examples), etc. In other words, large numbers can be seen as partially structuring the domain of emotive experiences. This is, however, not necessarily the only cognitive analysis available that would make number hyperboles appear less “pure” because they are conceptually dependent on some basic processes. We could also assume that metonymy is at work here, too. However, it may strike us as odd that metonymy has never been mentioned as a possible companion to hyperbole. It will be shown below that there are many cases of metonymically motivated hyperbolic expressions. What we would like to claim in the present context is that metonymic mappings are essential in all cases of number hyperboles, and this means that the case for pure hyperboles is even shakier. In this section we will restrict our argumentation to data from just a couple of domains, but we will show in 2.2. below that without metonymic mappings in the broadest sense of the term, including both metonymy proper and synecdoche, it would be quite difficult to explain why it is numbers, referentially the most precise of all word classes, that should appear as prime example of hyperbole. What is more, we dare postulate that the role of metonymy may have been seriously downplayed, if not completely ignored, in the whole history of research on hyperbole. To our knowledge, the only work that discusses at some length which are used hyperbolically is Feyaerts’ (1999) study of the role of metonymy in the conceptualization of stupidity in German idiomatic expressions, while Jäkel (1999: 215) just notes in passing, that hyperbole and met-
392
Rita Brdar-Szabó / Mario Brdar
onymy may interact in naming patterns, where a salient quality of the original name-bearer is expressed hyperbolically. In the examples that follow, numbers, some larger, and some smaller ones, are used to talk about distances: (7) a. Ta se vatra vidi na deset kilometara. ‘That fire can be seen from ten kilometres’ b. Promašio je gol za bar 30 metara. ‘He missed the goal by at least 30 metres’. In (8), a large number is used to talk about age: (8) a. Our teacher is a hundred years old. b. This computer is fifty years old. Utterances like (7–8) can of course be literally true, but let us assume that they are meant as hyperboles, i.e. to indicate great distance or age. To substantiate the claim that examples like (5–8) are based on conceptual metonymy, we first point to the phenomenon of meaning inclusion as discussed by Burkhardt (1996). Numerical elements which are hyperbolically used can then be regarded as a synecdoche because the meaning of large numbers is included in the meaning of general multal forms such as much or many. Concretely, the meaning of Hungarian hyperbolic frequency adverbials like százszor ‘a hundred times’, ezerszer ‘a thousand times’, milliószor ‘a million times’ etc. is included in the meaning of sok ‘much/many’. In other words, they are interpreted as a MEMBER-OF-A-CATEGORY-FOR-CATEGORY metonymy, i.e. one specific large or small number stands for the whole class of large or small numbers, respectively. Alternatively, if we operate with the notion of scales, as we intend to do in Part 2, they could be reinterpreted as instantiations of the part-for-whole metonymy involving a numerical scale and its upper or lower end, i.e. UPPER-END-OF-A-SCALE-FOR-WHOLESCALE or LOWER-END-OF-A-SCALE-FOR-WHOLE-SCALE metonymy (cf. Radden & Kövecses 1999: 32 for the former). An additional piece of evidence squaring with the view that these numerical expressions are based on metonymic mappings, in fact some data that independently support the assumption of this particular variant of metonymy comes from certain basically temporal adverbials in various languages which explicitly mention either just the beginning or the end of the period or of an activity/process but are normally used to denote an extended period of time or a stage of the event of unspecified length, which can be further metonymically extended to nouns that do by themselves imply some sort of procedural activity. Cf. the following German and Croatian examples: (9) a. Anfang Mai, Ende Juni ‘beginning of May, the end of June’
The truth of hyperboles
393
b. am Anfang/Ende des Spiels ‘at the beginning of the match/at the end of the match’ c. am Anfang/Ende des Heftes/Buchs ‘at the beginning/end of the notebook/book’ (10) a. početkom svibnja, krajem lipnja ‘beginning of May, the end of June’ b. na početku/kraju utakmice ‘at the beginning of the match/at the end of the match’ c. na početku/kraju sveska/knjige/romana ‘at the beginning/end of the notebook/book/novel’. Similarly, some names of holidays, i.e. labels for single days such as Christmas, or Weihnachten, Fasching, and Ostern in German, or Božić or Duhovi in Croatian, may be used to refer to shorter or longer periods of time, the holiday itself plus the following day, or the period of time around that particular day, or between that day and another holiday. Note that some of these look to be or actually are morphologically in the plural (Weihnachten, Ostern, in German, and the Croatian expression Duhovi). Cf. the following expressions: (11) a. zu/an Weihnachten ‘at Christmas’ b. o Božiću, za Božić ‘at Christmas’. Now that we have provided some evidence that metonymy may function as the base of some hyperboles and thus called into question the “purity” of number hyperboles, we turn to another prominent candidate for pure hyperbole – elative superlatives (superlativi elativi). This type of hyperbolic expressions may be illustrated by the following set of examples: (12) a. b. c. (13) a.
We pay the highest interest rates on insured ratings. I shall do it with the greatest pleasure. The letter was written in the kindest terms. Das strahlendste Weiß meines Lebens ‘the most gleaming white in my life’ b. Gestern war das herrlichste Wetter. ‘The weather was most agreeable yesterday’ c. Der Betrieb arbeitet mit den modernsten Maschinen. ‘The plant uses most modern machines’ d. mit den herzlichsten Grüßen ‘best wishes’.
Here the superlative form of the adjective is not used in a relative but in an absolute sense, i.e. the noun it modifies is not compared with anything but is considered as possessing the quality in a very high degree. If we agree that the
394
Rita Brdar-Szabó / Mario Brdar
UPPER-END-OF-A-SCALE-FOR-WHOLE-SCALE metonymy is operative in the
case of numerical expressions above, then there is every indication that such superlatives might be handled in a similar way, i.e. by means of a combination of related conceptual metonymies. The extreme point on the scale stands here for the whole region of the upper part of the scale. Another class of hyperboles that might be considered as belonging to the “pure” type could be utterances containing negative or positive scalar items such as never, nothing, always, etc. Provided that they are not to be taken in their literal meaning, they are understood as giving rise to a hyperbolic interpretation only if understood as cases of metonymic mapping of the type we have just discussed.4 Note that there are some other metonymy-based hyperbolic expressions that have hitherto gone largely unnoticed as such. In one of these, exemplified by some German and Hungarian utterances: (14)
a. Das ganze Stadion starrt auf dich. ‘The whole stadium is staring at you’ b. Az egész udvar/játszótér téged néz. ‘The whole yard/playground is looking at you’
a location is used to refer metonymically to a number of people normally associated with the place via a salient event/activity taking place at the location. The locative noun is typically modified by an expression of the ‘whole’ type. In the other type, i.e. metonymy-based hyperbole (frequently found with brand names in advertising, but also with ordinary nouns in colloquial language), we witness the MEMBER-OF-A-CATEGORY-FOR-CATEGORY metonymy: (15)
a. Kodak. A fényképezőgép Kodak. THE camera’ b. Outside the US, soccer is THE sport. (Abbott 1999: 3)
In both examples in (15) the definite article is used emphatically and this phonological clue prompts both metonymic mapping and the subsequent hyperbolic reading. By taking a closer look at these two most promising candidates for “pure” hyperbole we succeeded in achieving two related goals. First, we further reduced “pure” hyperbole’s territory because it was revealed that they do not come unsupported by other tropes, viz. conceptual processes. Specifically, we have indicated that they build on metaphors and metonymies. Secondly, and as a consequence, the rest of the expressions that might be suggested as —————
4
Cf. also Ruiz de Mendoza and Peña Cervel (2002), who show how metaphoric mappings interact with scalar concepts in producing hyperbolic readings.
The truth of hyperboles
395
“pure” hyperboles, is such a heterogeneous mass that they defy any attempt at generalizing their formal properties, i.e. they must be treated as isolated individual cases that exhibit a more or less ordinary type of language but appear as hyperboles only in a broader context, and must therefore be at best treated as figures of sentences. It is telling that even some recent typologies based on classical sources, such as Ueding and Steinbrink (1986: 271), inadvertently admit that some of these instances of the pure type are in fact based on some other tropes, specifically (16)
Du hast ein Herz aus Stein in deiner Brust. ‘You have a heart of stone in your breast.’
is said to belong to the core type, Quintilian’s gradatio, but is a hyperbolic metaphor. As an interim conclusion, we might point out that against the evidence adduced above the case for unsupported “pure” hyperbole seems to rest on very shaky grounds indeed. No reliable all-round definition of hyperbole can be formulated which would be based on an isomorphic relation between linguistic form and a corresponding meaning shift. What seems to remain constant across different traditionally defined subtypes of hyperbole is just the hyperbolic effect which can be brought about by a great variety of linguistic means and interplay of conceptual processes. We conclude this section by listing first some more examples of hyperbolic metaphor (17) as well as some examples of hyperbolic expressions based on other tropes/figures, such as similes (18) and comparisons (19): (17)
(18) (19)
a. One of the biggest gatherings of royalty, religious chiefs, political leaders and military top brass ever seen in Britain poured into Westminster Abbey as a sea of black. b. A carpet of flowers gleamed on the grassy banks of Windsor Castle yesterday. c. Forget our children’s children; how about our children’s diaper rash? Like, now. It’s three o’clock in the morning for me, too, you know. So you can run a country with your eyes closed, but you can’t remember to boil the kettle? In short, Mr. Blair must decide: should he go to work in a clean suit, or stay at home with a Nile delta of vomit moistening his shoulder blades? d. Die Schulden erdrückten ihn fast. ‘Debts almost crushed him.’ Az arca úgy ragyogott mint a Nap. ‘His face was shining like the sun.’ Fényesebb mint a Nap ‘shinier than the Sun’.
396
Rita Brdar-Szabó / Mario Brdar
1.4 Beyond the confines of the classical rhetoric: Some more types of con ceptually dependent hyperbolic expressions In addition to hyperboles based on metaphors and metonymies, as well as those based on comparisons and similes, there are numerous other expressions and constructions that serve the same function. It is not very surprising that idioms of various structural types (and not just idiomatic similes) are very frequently recognized as exhibiting hyperbolic nature. Cf. the following examples of resultative idiomatic constructions expressing the idea of talking (and/or asking) too much in different languages: (20) a. talk somebody’s head/leg off b. talk the hind leg off a donkey c. X can talk a blue streak (21) jemandem ein Loch/Löcher in den Bauch reden/fragen ‘talk somebody a hole/holes in the stomach’ (22) lyukat beszél a hasába ‘talk somebody a hole in the stomach’. The first English hyperbolic expression in (20)a. is structurally realized as transitive phrasal verb, i.e. the verb, otherwise inherently intransitive, is followed by a noun phrase denoting a body part and particle. The second example is a variation on this theme as it contains a ditransitive prepositional verb and a complement. It is interesting to point out that the placement of the particle before the object NP is blocked in conventionalized hyperbolic expressions, although this is normally possible with very many phrasal verbs (cf. Quirk et al. 1985: 1155). Cf. the following contrast: (23) a. They turned the light on. vs b. They have called the strike off. vs (24) a. I was crying my eyes out. vs b. He was laughing his head off. vs
They turned on the light. They have called off the strike. *I was crying out my eyes. *He was laughing off his head.
Many other structures, from syntactic constructions to morphologically complex one-word constructions may be used as hyperboles. In the transition area between phraseology and syntax we find some recurrent collocations consisting of adverbs as the first element and a verb or an adjective as the second. Consider first some cases of modifications of verbs by maximizer adverbs in English that have become conventionalized and almost cliché-like: (25)
a. They fully appreciate/understand our difficulties. b. We utterly deplore his methods. c. I can perfectly see your reasons.
The situation with adverbs modifying adjectives is very similar:
The truth of hyperboles
(26)
397
a. We are painfully aware of the problem. b. To be brutally/perfectly frank,…
Consider also some novel combinations in which adverbs seem to be metaphorically motivated. The second combination in (27)b. is subsequently subjected to a metalinguistic reflection by the speaker: (27)
a. The delivery itself is no longer open to debate; the days when Prince Philip went off to play squash while the Queen was giving birth seem jurassically distant, and Mr. Blair’s duties on May 24th are unambiguous. b. And there, I think, one of the biggest points of significance is actually the power of air power that Steve was talking about, in the sense that if air power is quantumly better than – to make it into an adverb, which I don’t think it is – if it’s a real quantum improvement in air power in terms of accuracy and the ability to achieve aims, and it lowers the threshold price for going to war for the United States because it means less collateral damage, fewer civilians killed, fewer Americans in the line of fire.
Among the phrasal syntactic constructions that may have a hyperbolic function we find some genitive/possessive constructions of a reduplicative kind, where the head noun is repeated in the genitive plural. Cf. the following translational equivalents: (28)
a. Shir Hashirim (Hebrew) song song-GEN-PL b. Canticum canticorum (Latin) c. Il canto dei cantici (Italian) d. El Cantar de los Cantares (Spanish) e. Énekek éneke (Hungarian) song-PL song-POSS f. sang der sange (Middle High German).
The Croatian counterpart makes use of the spatial preposition nad ‘over’: (29)
Pjesma nad song over
pjesmama songs-LOC-PL.
These are, in fact, cases of dead hyperboles that have a referential or naming function, which is absent with most other structures. A similar hyperbolic effect, but without a comparable referential role, arises in appositive genitive constructions of the following type (cf. Austin 1980), but it appears that a metaphorical mapping is involved in creating this effect:
398
Rita Brdar-Szabó / Mario Brdar
(30)
a. b. c. d.
this great flabby water-melon of a show a Methuselah of a pear tree a Sierra Nevada of eggs a fool of a man.
Corresponding constructions are reported in French (Bennett 1976) and in Dutch (Paardekooper 1956), and can probably all be traced back to Latin constructions like: (31)
monstrum mulieris.
Broadening our perspective to clausal syntax, we notice that some indirect conditional clauses also fall within the scope of hyperbolic constructions: (32) (33)
a. If he can do it, I’ll eat my hat. b. Ha ez igaz, akkor megeszem a sapkámat. (Hungarian) ‘If this is true, I’ll eat my hat’. Ha ő az igazgató, akkor én vagyok a Robert Redford/a dalai láma. (Hungarian) ‘If he is the director, then I am Robert Redford/the dalai lama’.
More or less the same could be said of some so-called clauses of preferences: (34)
I’d rather die than do it.
Hyperbolic effects can be observed in the realm of word-formation, too. Many processes, from prefixation and suffixation to compounding to some minor word-formation processes, could be adduced as examples, but we will restrict ourselves to just some examples of affixation and compounding. Suffixes for augmentatives are naturally suited to hyperbolic use. Cf. some examples from Croatian: (35)
a. Makni te nožurde! move these leg-AUG-PL b. Ta kravetina me prevarila. that cow-AUG me cheated.
The second of these of course operates on a metaphorical pejorative expression where the noun denoting an animal is used for a female person. There is a whole range of prefixes of degree or size in many languages that can be attached to nominal, verbal and adjectival bases to express the idea of something extreme, e.g.: (36)
a. hypersensitive, hyperactive b. overreact, overplay
The truth of hyperboles
399
c. arch-hypocrite, arch-enemy d. megarich, megabucks. Finally, we consider several types of compounds. In one type of adjective compounds, the first element can function as an intensifier. It can be a noun, as in (37)a. and b., or an -ing participial form, as in (37)c. and d.: (37)
a. stone-
b.
cold deaf dead blind tired
dogcheap c.
d.
boiling stifling baking piping steaming soaking sopping dripping.
-hot
-wet
Römer (1968: 95ff.), discussing verbal means of amelioration in advertising, adduces some examples of compounding that may be cases of euphemism or hyperbole, or of both at the same time. Thus a florist’s shop, Blumenladen in German, may be called Blumenstudio. A similar tendency, with an even more outspoken tendency towards the hyperbolic, may be observed with some compounds denoting people: (38)
health-guru, computer guru, media mogul, movie mogul, press baron, property tycoon, shipping magnate.
These compounds that tend to assume naming or referential function as they gradually lexicalize will be discussed in more detail in one of the case studies in Part 2.3. below.
400
Rita Brdar-Szabó / Mario Brdar
1.5 Hyperbolic effects: a tentative inventory We have seen so far that hyperbole, unlike some other master tropes, cannot become a conventionalised way of referring to a concept to the exclusion of other ways of referring to it. In other words, its primary function is not referential, i.e. naming, but it rather serves rhetorical functions. If there is no hope of providing a catalogue of formal defining features of hyperbole, as we have seen in sections 1.2–4, one might expect that definitional efforts should rather go in the other direction, i.e. one might then become inclined to try to identify hyperboles only in terms of their pragmatic effect, i.e. functions that they have in discourse. However, it turns out that this approach leads into comparably thorny thickets because – unlike some other tropes or figures – hyperbole is not a one-track phenomenon. It is as versatile concerning its rhetoric/ pragmatic effects as it is difficult to capture in terms of its formal instantiations. However, we are convinced that it is a sensible approach to try to arrive at a tentative inventory of possible pragmatic effects, which will be the purpose of the present section. In the next section our main goal will be to demonstrate that all these various pragmatic effects that hyperboles may exhibit rest on just a handful of types of mappings and inferences obtaining within networks of mental spaces. Probably one of the most prominent functions of hyperbole is intensification, i.e. to express the assertion that an entity possesses some property to a very high degree, that the circumstances and/or effects of an event are extreme, etc. Regardless of whether what is exaggerated are properties of an entity or the reaction of the speaker or some other participant, the general propensity to intensify through exaggeration seems to be one of natural human drives, as pointed out by Quintilian himself. According to Stempel (1983: 92f.), this fondness for hyperbole is a reflection of the ever-present wish to claim attention and recognition in the community. As pointed out by Gibbs (1994: 393), hyperbole contrasts with litotes with respect to its philosophical attitude. While hyperbole expresses optimism and idealism, litotes is associated with pessimism and cynicism. Blair (1842) speaks of two basic types of hyperbole. The descriptive hyperbole amplifies the externally observable appearance of an entity, while the emotive hyperbole expresses the speaker’s or some other participant’s internal reaction or emotion towards a certain entity or an event. In other words, the former has to do with visual impressions (or some other perceptions), while the latter is based on internal subjective processing and eventual evaluation of these impressions. This basic pragmatic function of hyperbole, to modulate the intensity of an “objective” state of affairs, or one’s reactions to it, exhibits a number of variants with blurred boundaries. Whether it will be one or the other depends, among other factors, on the type of discourse. The exaggeration of the “objective” properties can have heuristic/didactic aims (docere, perspicuitas). Overstating certain crucial facts about the subject
The truth of hyperboles
401
matter makes them stand out more sharply from the background and thus easier to notice or teach. In the language of advertisments, it is clearly the former aspect that is most prominent, i.e. the “objective” properties of an entity are exaggerated for an obvious reason. In certain historic periods, it has become a convention to overuse hyperbolic expressions of both types in certain literary genres, e.g. to help produce parody on a more general level (cf. Jónácsik 2001, Naschert 1998). As is well known, parodies are considered to be imitations that keep the style and techniques characteristic of well-known earlier texts and replace their content by more commonplace ones. But parodies can borrow some content elements, e.g. the characters and/or the plot, but (adapt the style. In the following text, a present-day rendering of Beowulf, in fact one in the set of three parallel renderings collected by Ben Greenman that appeared in the New Yorker,5 the humorous effect is in part due to breaking the temporal frame, i.e. bridging, or more precisely, obliterating the gap in time between the original epic and the present-day by importing into the framework of the model some technical devices and famous personalities from the present day, which results in incongruity. Concretely, these are the microwave oven and Elizabeth Taylor and her garbage can. Notice the distinctive role of hyperboles in actually introducing these alienating elements. The text plays on the reader’s encyclopaedic knowledge concerning Elizabeth Taylor’s customs, i.e. the supposed contents of her garbage can, and claims that Grendel’s lair can even surpass the virtually unbelievable quantitiy attributed to Elizabeth Taylor: (39)
Beowulf: A New Monologue Translation, by David Letterman Whew! Is it cold! Anybody feeling the cold? [Applause] I’ll tell you, ladies and gentlemen, it was so cold today that Grende had to pop them Danes into the microwave before he devoured them. [Laughter, applause] You ever pop any Danes in the microwave yourself, Paul? [Laughter, applause] Oh, by the way, I want to congratulate King Hrothgar. He’s doing a great job over there in the palace. Apparently, he posted a sign on the front door of the place, “Closed for Mass Burials.” [Laughter, applause] Anybody been down to the moors? I haven’t been myself, but I hear that Grendel’s lair has more bones in it than Elizabeth Taylor’s garbage can. [Laughter, applause]
————— 5
Beowulf: New Prose Translation, May 29, 2002, 67
402
Rita Brdar-Szabó / Mario Brdar
You know, a week ago I wouldn’t have given Hrothgar’s troubles to a monkey on a rock. But I think things are looking up for him. They say in the papers that there’s a new guy, Beowulf, coming to help the King with the monster. I just have one piece of advice for Mr. Beowulf: Run! Run for your life! [Laughter, applause] Some hyperbolic expressions that focus not so much on the properties of their topic but rather on the experience of the speaker (or on the experience imputed by the speaker to a character) can function as indications of surprise or may even signal irony, as shown by a number of recent psycholinguistic studies (cf. Kreuz & Roberts 1995, Colston 1997, Gibbs 1998, Colston & Keller 1998). Thus, the Croatian example (6) might be construed as a way of conveying surprise in addition to intensification. An overlooked function of hyperbole is to express annoyance or irritation, or just disapproval, on the part of the speaker concerning some state of affairs in the immediate environment that is also mentioned in the hyperbolic expression. This function of hyperbole is most obvious in an example such as (5) above. Hyperboles are also relatively common in their social function as politeness formulae. Very frequently these are dead hyperboles, and are found in both formal style (e.g. business letters) and in informal, written or spoken, style. Hyperbole can be protective of the speaker, not because attack is the best defence, but because in some situations it is better to exaggerate in flattering the interlocutor. 2.
A cognitive account of hyperbolic expressions
2.1 Hyperboles as over-reachers: The role of mental spaces, scales, domains and mappings If we take a look at cross-linguistic counterparts for the term hyperbole, we find that they are in many languages motivated by a spatial metaphor. The meaning of the original Greek term is actually ‘throwing above’ (hyperbolē ← hyper ‘over, above’ + bolē ← ‘throw’). Quntilian’s term superiectio is a calqued translation using the corresponding stems and the same imagery. The same spatial metaphor is attested in English and German terms, over-reacher and Überbietung, respectively, although these are not as current as their loanword counterparts. In the Hungarian counterpart költıi túlzás ‘poetic exaggeration’, the second element is a nominalization derived from the basis túl, which means ‘over’, ‘over the border, over the edge of something’. What this systematic cross-linguistic correspondence suggests is an underlying conceptualization in terms of a vertical scale.
The truth of hyperboles
403
The degree of exhibiting a property, the intensity of performing an action, etc. may be presented as points on a scale with values that are judged by the common sense to be the average, normal or default ones being located in the centre of the middle section, while the points on the sections corresponding to lower or higher degree are felt to refer to situations that are decreasingly likely or possible. The limiting case on the lower end of the scale is of course in many cases zero occurrence or zero quantity, while the upper limit is far less clearly set. The scale concept has recently been made use of in linguistics in a number of ways. The sort of scalar model that seems to be necessary for a descriptively and theoretically adequate account of expressions exhibiting hyperbolic effects is in our opinion a combination of two recent approaches to scales. Discussing various types of non-branching hierarchies, i.e. certain sets of incompatibles, Cruse (1986: 192) takes scales to underlie ordered sets of lexical items whose constituent units “relate to different values of some variable underlying property.” Unlike Lehrer (1974: 29) and Lyons (1977: 289) who distinguish between ranks, whose members are non-gradable, and scales, whose members are gradable, Cruse splits ranks into two types of sets, and depending on the type of underlying scale distinguishes between three types of such sets. Lexical units that operate over a discontinuous scale that varies in discrete jumps are rank-terms. These rank-terms are non-gradable. Lexical items in sets operating over continuous scales may be gradable or nongradable. The former terms are called grade-terms, the latter ones are degreeterms. The set of cardinal number names, more precisely the subset of natural integers, are a prime example of rank-terms, but there is a whole host of sets of lexical items based on these numbers, such as the set of ordinal numbers (first, second, third,…), names for fractions (half, third,…), names for geometrical shapes (triangle, square, pentagon,...) etc. If we take a closer look at natural integers, we notice that the property of numerosity, similar to military rank, increases by discrete jumps because no set of X items outnumbers any other set of X items. Cruse illustrates this property by the oddness of: (40)
?There are only just twelve students in our group; in yours there are nearly thirteen.
Another relevant feature is that sets of number names combine features of measure terms and of helical chains. In English there are two sets of terms resembling helical chains, which are continuously re-cycled, one inside the other, so to speak, as higher and higher numbers are named. These are the ‘units’ – one, two, three, four,… nine, and the “tens” twenty, thirty, fourty, ... ninety. In addition to these helix-like sets, there is a set resembling measure terms: hundred, thousand, million, billion, etc. Interestingly, it is these measure-like number terms that are most likely to be used in hyperbolic expressions, as will be shown below in the case study in 2.2.
404
Rita Brdar-Szabó / Mario Brdar
Degree-terms may be illustrated by sets such as terms used in examination assessments, or some sets of terms exhibiting a temporal sequence, such as stages in a human’s life, ands some measure terms: (41)
a. fail, pass, credit, distinction b. baby, child, adolescent, adult c. second, minute, hour, day,…
Measure terms like those in (41)c. are different from the rest in that the intervals between measure terms exhibit geometrical rather than linear progression. Grade-terms are most adjectives since they are normally gradable, unlike degree-terms: (42)
a. freezing, (cold), cool, warm, (hot, scorching) b. atrocious, (bad), indifferent, average, fair, (good), excellent c. minuscule, tiny, (small), (big), huge, gigantic.
The boundaries between members of such sets are frequently quite vague. It is apparent that expressions that exhibit hyperbolic effects, as illustrated in 1.4. above, belong to all the three types of expressions which Cruse calls scales. What is more, we have listed some types of expressions that cannot be easily assigned to any such established set of incompatibles. It is also evident that this concept of scale is too static and hardly furnished with any predictive power that seems to be necessary to derive the appropriate kind of interpretation of hyperbolic expressions. However, an important insight on a most general level that can be gained from Cruse’s discussion is that sets based on scales can be combined with and/or embedded into each other. In the scalar model, developed by Israel (1996, 1997, 1998, and 2001) primarily for polarity sensitive items, forms which denote an element within a scalar ordering, such as: (43) (44)
a. b. a. b.
Margo didn’t sleep a wink before her big test. Margo didn’t sleep much before her big test. Belinda won scads of money at the Black-jack tables. Belinda won a little bit of money at the Black-jack tables.
are called polarity items. Such items are ‘scalar operators’ whose profiled content is construed against the background of an ordered set of alternatives and which are interpreted within the information structure provided by a more general scalar model. Such a general scalar model is for Israel a structured set of propositions ordered along one or more parameters in a way that supports inferencing. An obvious point of convergence between Cruse’s scale concept and Israel’s scalar model is that a scalar model can also be simple, consisting of
The truth of hyperboles
405
one variable and values ordered on a single scale, or be complex, exhibiting many variables and scales. However, an important point of difference between the two scale concepts is that Israel’s model supports pragmatic entailments, or inferencing, as already pointed out by Fauconnier (1975). Israel (2001: 304) shows that inferencing in a scalar model is defined relative to the propositional function on which it is built. For an affirmative function like: (45)
a. This building can withstand an earthquake that would register as 5 on the Richter scale. b. Bill can eat 10 hamburgers.
The direction of entailments is from the value specified to lower values on the scale, such that the building can withstand any earthquake of lesser intensity, and that Bill is capable of eating any number of hamburgers between 1 and 10. If the proposition is, however, negative, as in: (46)
a. This building cannot withstand an earthquake that would register as 5 on the Richter scale. b. Bill cannot eat 10 hamburgers.
the inferences now run in the opposite direction, i.e. from the value specified upwards on the scale, entailing that any earthquake of greater magnitude would also be fatally destructive, and that we can be sure that Bill cannot eat more than 11 or more hamburgers. Propositions within a scalar model may be understood with respect to some implicit scalar norm. A given lexical item can denote either a high or a low scalar value relative to this scalar norm. Propositions that pragmatically entail the scalar norm because they say something that goes beyond some default expectations are claimed by Israel to be rhetorically emphatic, while propositions that are entailed by the scalar norm are rhetorically attenuating and weakly informative. This is schematically presented in the following figure (Israel 1996: 628):
Figure 1. Polarity sensitive items arranged according to their quantitative and informative values (Israel 1996: 628)
406
Rita Brdar-Szabó / Mario Brdar
This model seems to work perfectly for some hyperbolic expressions. Cf. the following example, originally from Fillmore et al. (1988: 520) for which a complex scalar model with coordinated scales must be assumed: (47)
You’d never get a poor man to wash a car for $2 in bad times, let alone a rich man to wax a truck for $1 in prosperous times.
As Israel (2001: 322f.) shows, five scalar parameters (cf. (49)) are involved in this complex proposition, schematically represented as in (48): (48) (49)
M would do W on O for R under C a. M = individuals, ranked in terms of their financial need b. W = types of work, ranked in terms of strenuousness c. O = object to work on, ranked in terms of size d. R = rewards for work, ranked in terms of financial value e. C = general circumstances, ranked in terms of the availability of work.
The conjunction let alone that joins the two parts of the proposition requires that all five scales should be coordinated in such a way that the contrast between the two conjuncts is maximized. In other words, the value for each of the five variables in the first conjunct must be ranked lower than the value for its counterpart in the other conjoin, e.g. a poor man is ranked lower than a rich man on the scale of prosperity, washing is lower in terms of strenuousness than waxing, etc. Now, more or less the same complex coordination of scales can be found in simple monoclausal sentences, clearly exhibiting hyperbolic expressions. Cf. the following examples from Israel (2001: 323): (50)
a. Stella can calculate an eigen vector in the blink of an eye. b. Dim can’t even count to two in less than an hour.
Both examples follow the schema Person X can(not) solve problem Y in Ztime. Stella is extremely bright because she can solve a difficult task in an extremely short period of time, unlike Dim, who is extremely dull and cannot solve even the simplest problem in an extremely long span of time. Both sentences express amazement about certain extreme properties of the people in question. Isreal points out the amazing speed and ease with which such utterances are interpreted, although they are remarkably complex in involving three scales that must be coordinated. Of course, one of the requirements for the expression of wonder to arise is that the polarity of the items must be the right, i.e. we would not be that surprised if Stella could not calculate the vector. But it is even more important for such examples that the three scales, the one involving problem-solvers (ranked from most competent to least competent), one involving problems (ranked from the least to the most difficult), and
The truth of hyperboles
407
the one involving time spans (ranked from longest to shortest possible), be properly coordinated and integrated. To emphasize the extreme position of any given element on a scale within the complex model, we must associate it with elements from the opposite ends of the other scales within the model. For example, to emphasize that Stella is extremely bright and therefore ranked quite high on the scale involving the likelihood of her solving any problem, we assert a proposition in which low values are assigned on other scales, i.e. the particular problem Stella faces is ranked extremely low on the scale involving the likelihood of its being solved, and equally the amount of time she needs is ranked extremely low with respect to the likelihood of its being sufficient for the problem of this magnitude. The effect of amazement or surprise, listed as one of possible effects of hyperboles in 1.5. above, is thus naturally explained as resulting from this sharp contrast between the values for various scales. The cognitive mechanism that underlies the coordination of scales is actually not elaborated by Israel, but we might assume that a network of mental spaces is involved, exhibiting intricate mappings, as modelled by Fauconnier and Turner. Mental spaces are small conceptual packets constructed on-line, as we think and talk, for purposes of local understanding and action. They are partial assemblies containing elements, structured by frames and cognitive models, and they are interconnected. Mental spaces can be used generally to model dynamical mappings in thought and language, and are created through conceptual integration and modified as thought and discourse unfold. Various variables, i.e. entities and properties, as well as the scales related to them, together with values for these variables and default values, can all be seen as input spaces that are linked by cross-space counterpart connections. A generic space, which maps onto each of the inputs, contains what the inputs have in common. In yet another space, a blended one, we witness a selective projection of structure from input spaces, but there is also the emergent structure not found in any of the inputs. We surmise that the coordination of scales is the result of such a conceptual integration or blending across spaces containing individual scales. As pointed above, Israel makes his scalar model do without explicit recourse to mental spaces and blending. It seems that constructions like (47) or (50) can be accounted for regardless of whether we make use of mental spaces or not. However, it will be shown in a case study in 2.5. that the conceptual work done by mental spaces seems to be prerequisite for the proper alignment of scales prior to coordination. It is quite obvious that his framework also holds great promise for the treatment of hyperbole’s sibling, litotes. Taking another look at Figure 1, we realize that what Israel calls emphatic positive or negative polarity items most of the time function as hyperboles, while understating items are in fact best candidates for litotes. This situation is, after all, schematically indicated by different regions they occupy on the scale, the latter expressions are considerably closer to the default value, i.e. to the norm.
408
Rita Brdar-Szabó / Mario Brdar
2.2 Large numbers: A case study on the role of metonymy and metaphors In 1.4. above we briefly discussed the so-called number hyperboles as putative examples of the pure type. To the Hungarian, German, French and Croatian illustrations above, we now add some English examples: (51)
a. Today you can buy hundreds of flavours of ice-cream. b. Thank you a thousand times, you’re an angel. c. But Dog knew thousand and one delaying devices which could keep a man incommunicado for a couple of day or more. d. Millions of words have been written about the introduction of the euro.
The phenomenon seems to be cross-linguistically wide-spread. There is always a numerical element, a real or a fictive one, that belongs to the realm of what is in the folk number model perceived to be a large one, and is normally a round one if real, containing number one and a number of zeroes. This probably serves as one of the most obvious clues of the intended non-literal meaning because a large round number of entities or events in a set or a series is conspicuously less common in real life than others, and obviously less easily accessible to immediate precise verification. Another clue is the nominal use of such large numbers in the plural. A plural large number seems to be more easily recognized as an exaggeration than a singular form. One of the differences between linguistic and cultural communities seems to concern the choice of the preferred number or numbers; it is most of the time a hundred or a thousand, but there are other possibilities as well (but see Jansen & Pollmann 2001, who claim that the roundness effect of numbers seems to be constant cross-culturally). It is also quite possible that different large numbers operate as hyperbolic devices for specific pragmatic purposes in different domains within one language, e.g. that one number is used to phrase a reproach for something not done in the linguistic domain, but another number may perhaps be used in the domain of visual perception to indicate a high frequency of the event, etc. This possibility must remain for the time being a putative hypothesis awaiting empirical cross-linguistic and crosscultural testing. The number in question, however, need not be one of the ‘tens’. Cf. the following example from Marlowe: (52)
Mighty lord, Three hundred thousand men in armour clad, Upon their prancing steeds, disdainfully With wanton paces trampling on the ground. Five hundred thousand footmen threatening shot, Shaking their swords, their spears and iron bills, [...]. (Tambourlaine IV,1, 20–5)
The truth of hyperboles
409
Of course, it is generally the case that with less than perfectly round numbers, the smaller these numbers are, the more likely the whole situation, i.e. the likelihood of the expression being recognized as an intended hyperbole significantly decreases. In addition to real numbers, there are in such hyperbolic expressions also some lexical items that are only fictive numbers, i.e. in their phonological makeup they resemble a number but there is no corresponding item in any mathematical domain containing numbers: (53)
a. We were attacked by zillions of mosquitoes. b. For the umpteenth time, don’t do it!
The inference made here is that the numbers in question must be at least on the order of million and thirteen, respectively, because of their phonological overlap with million, billion, trillion, etc. and with ordinal numbers between thirteenth and nineteenth. In both cases there are at least some clues as to the magnitude of the number and the situation thus designated. Another possibility is to take just the suffix used to derive larger round numbers between 20 and 90 (zwanzig ‘20’, vierzig ‘40’, fünfzig ‘50’, etc.), where the suffix –zig in a sense corresponds to the 0 digit and the base is a number between 2 and 9 (occasionally slightly modified), as in the following German example: (54)
Ich hab’ dir zig mal gesagt, dass ... ‘I have told you many times that …’.
The expressions in question can, however, be even more schematic, i.e. be just symbols for unknown values, as in the following Croatian example: (55)
Iks puta sam ti rekao. ‘I told you x times’.
We might surmise that in such cases the most widely known symbol for the unknown value, X, evokes the concept of distance because what is distant is usually unknown or at least less well known (e.g. distant countries, continents, planets). If this is correct, and we thus have a sort of metaphorical mapping (MENTAL DISTANCE IS PHYSICAL DISTANCE), then we seem to have a natural explanation why such schematic expressions can acquire more precise meanings leading to a hyperbolic effect. If the number is an unknown variable it is conceptually distant and must occupy a point or an area on a scale of intensity where larger numbers are found. It has often been noted in traditional rhetorical accounts that in the case of number hyperboles the process of exaggeration can go into two directions. In all the examples above we witness the upwards exaggeration, i.e. the number of entities or the occurrences of events, etc. has been purposefully inflated.
410
Rita Brdar-Szabó / Mario Brdar
The opposite, traditionally called meiosis, is also frequently attested, mostly in colloquial speech, and it is also found as a number hyperbole, chiefly with zero or numbers approaching zero. Cf. the following example: (56)
There is no one in the queue.
which might have been uttered by someone who expects a considerable number of people waiting in the queue, but then finds only a couple of people there. Some German idiomatic expressions that use fictive numbers apparently close to zero belong here: (57)
Null Komma Josef ‘zero point Joseph, i.e. nothing’.
Similarly, the idiomatic expression in (58), exhibiting a similar structure except that the number after the decimal point is another zero: (58)
in Null Komma nichts ‘in no time’
is used in German to refer to an extremely short period of time. As shown above, most authors agree that hyperboles can be combined with other tropes, mainly with metaphor. We have suggested above that it may be sometimes the case even with what is considered pure hyperbole. A metaphorical layer seems to be necessary in order to work out the intended meaning in examples like (54), which means that the ground for the so-called pure hyperbole is further eroded. One of the problems with the scalar model, as far as its utility for hyperboles is concerned, is that the region on the scale that is entailed is too large in that entailments in the case of emphatic items go down to the norm and possibly even include it, while hyperboles normally imply values significantly over or below the expected norm or the default value. In other words, the scalar model, if it is to be applicable to all hyperbolic expressions, should have some in-built mechanism that would stop inferences well before we approach default values. A second, more serious problem is that Israel’s model (1996: 629–30) recognizes only two positions on a scale. Most polarity sensitive items clearly encode a scalar semantics by designating some relative or absolute position within an ordering of elements along some gradable dimension of semantic space, i.e. by having certain quantitative value (q-value). In theory, we might expect any scale to have an infinite number of distinct q-values, as numerical scales do. However, as Israel shows: […] languages seem to be quite stingy about lexicalizing such distinction. For the purposes of polarity items we only need to recognize two: high q-value and low q-
The truth of hyperboles
411
value, both of which are understood relative to contextual norms associated with a given dimension. (1996: 629)
So what happens here is that the granularity of scales is effectively diminished. What is more, certain types of scales involving rank-terms and degreeterms are so defaced when used in a hyperbolic sense that they cease to involve these two types of terms. In other words these two types of scales are converted into the third type and then the superfluous q-values are eliminated. How this happens remains largely unclear in this model. In some cases, the presence of an intensifier like very is claimed “to narrow the denotation of a scalar predicate to the high end of the scale” (Israel 2002: 429). But we are left in the vague in other cases where no such intensifiers are present. The scalar model, developed by Israel, is however claimed to belong among the most ubiquitous and basic conceptual tools, almost on a par with metaphor, metonymy, conceptual integration, etc. If it were indeed that basic, then we should not expect it to exhibit a submodel that would constrain it, i.e. inhibit certain inferences. Nevertheless, something of the sort is apparently needed if the model is to account for hyperboles. So either we must look for some other cognitive mechanisms that could better account for hyperboles, or the otherwise attractive model must be allowed to interact with other cognitive mechanisms in required ways. As we have already indicated on large numbers in 1.3. above, it is possible to conceive of these hyperbolic expressions as being motivated by the partfor-whole metonymy involving a numerical scale and its upper or lower end, i.e. UPPER-END-OF-A-SCALE-FOR-WHOLE-SCALE or LOWER-END-OF-ASCALE-FOR-WHOLE-SCALE metonymy. Note that the scale itself may have innumerably many points but that it is only the extreme points that are used in hyperbolic expressions. Now, of course, we know that in most languages spoken by communities whose number systems are of the decimal type, there are two sets of terms resembling helical chains, and a set resembling measure terms. We have seen that it is precisely these measure-like number terms that are most likely to be used in hyperbolic expressions. So metonymy seems to be the mechanism that singles out these measure-like, and some other, number terms placed somewhere within the scale (note that number names form in principle, if not in practice, an endless series) and promotes them to end points, practically obliterating all the other numbers between them and the default value. At the same time, the ground seems to be destroyed for all kinds of bizarre entailments that might have arisen if all the other numbers had remained in play. In other words, metonymy virtually constrains the inferencing process and thus makes possible the emphatic hyperbolic interpretation.
412
Rita Brdar-Szabó / Mario Brdar
2.3 Computer gurus and media talibs: another case study on metaphors as inputs for scalar models Another drawback of relying solely on the scalar model as it is envisaged by Israel when we try to account for hyperboles is that the range of formal devices exhibiting hyperbolic effects is much broader than the range of data that can be handled by the scalar model. For many nouns and adjectives, the problem is that their “meanings do not saliently contrast with an ordered set of alternatives” (Israel 2001: 310), and it would be difficult to conceive of them as occupying a certain place on a scale in the first place. Of course, true intensifiers enforce a scalar reading even where it is normally not available (Israel 2002: 429), as for example in very married or very pregnant. But we are left with a large mass of other examples without any such overt intensifiers, for which no explanation is offered. Israel’s own examples, the adjective red and the noun oyster in not spend a red cent, and not worth an oyster, respectively, seem to intimate that some other types of basic cognitive mechanisms are occasionally prerequisite for the scalar model to kick in. In the case of these two expressions (and numerous other, similar expressions) the lexical item in question is of course inherently non-gradable, non-scalar, but if we assume a prior metaphorical mapping whereby, e.g. oyster is converted to gradable expressions and comes to mean something like a ‘generically worthless object’, then it may be assigned a certain position on a scale. We show below that the cognitive mechanism in question could be a metaphorical mapping between two fairly disparate domains. The scalar model of a sort comes into being in the blended space by bringing into correspondence the two input mental spaces. We illustrate this claim on some English and Hungarian collocations of the type N + N, comparable to the German construction exemplified by Blumenstudio (as discussed by Römer 1968: 95ff.), the only difference being that all the heads, i.e. the right-hand elements are personal nouns. The following English examples from Brdar-Szabó and Brdar (2000) may serve as an illustration of this construction type in general: (59)
a. The Mail group is presided over by Lord Rothermere, the last of the hereditary press barons. b. But it is also a brutal and backward region dominated by a group of strong-armed plantation bosses known as caciques. c. Hiaasen saves his sharpest scorn for the sugar-cane industry magnates [...].
There is also a recently attested Hungarian collocation with tálib: (60)
média tálib. media Taliban
The truth of hyperboles
413
The items under discussion are labels for people characterized by possessing an extreme degree of power/influence/knowledge in certain areas of human activities. It will be noted that most of these are marked as belonging to slang or informal kind of language and that they are often derogatory, pejorative or, at best, humorous. Their use, however, very frequently results in a sort of hyperbolic effect. These personal nouns belong to a relatively compact semantic group which may be labelled as ‘potentate’, i.e. a person that may be described as having extreme influence or power due to his/her position (e.g. N + lord/ baron/magnate/mogul/tycoon). The semantic contribution of heads seems to be bleached down to a general meaning like [(UNSYMPATHETIC) PERSON WITH AN EXTREME DEGREE OF POWER IN AREA X]. It is of course possible to establish some further subgroups of such personal nouns in English, e.g. nouns denoting people that are extremely knowledgeable in certain areas (investment guru, health guru, sports pundit, foreign policy pundit, etc.), or nouns denoting enthusiasts (e.g. jazz freak, film buff, etc.). In this case study we restrict ourselves to combinations with guru and talib. What these two nouns have in common is that they belong to the general domain of religion. In Hinduism, the former word is used to refer to a preceptor giving personal religious instruction, whereas the second word originally denotes religious disciples in Islamic countries, but has come to be associated by the recently overturned oppressive Afghanistan regime, whose members were characterized by extreme conservatism and intolerance of anything Western and/or un-Islamic, which resulted in tight control in all areas of human life. Part of the folk model of society, and also part of the scientific model, is that religion belongs to social superstructure. This means that the whole domain of religious life, including people vocationally associated with it, generally enjoys high esteem and is considered prestigious in comparison to other domains of life. The right-hand elements in these collocations, guru, and talib, respectively, represent this religious domain as one of the inputs in metaphorical mapping. The other input mental space comes from a variety of domains, but none of these is as high in terms of esteem as the domain of religion in the above folk model. There are of course other nouns that can be used as heads to replace guru and talib, e.g. someone may be called an investment adviser in everyday language, etc. Such combinations are apparently more congruent because both collocates come from mental spaces from the same domain, or at least from the same matrix domain. However, these congruent but ordinary constituents are replaced by terms from a different domain considered to be higher in terms of its social prestige. In the classical cognitive approach to metaphor, one could say that elements of the domain of religious life are mapped onto various other domains of human life and partially structure them. Simplifying things a little, one
414
Rita Brdar-Szabó / Mario Brdar
could also say that a socially more prestigious, higher domain structures a lower one. This suggests that these sets of two domains form a sort of minimal or rudimentary scale, or still better, a continuous region in a larger scale underlying the hierarchical structure of the folk model of social life. Two otherwise non-gradable concepts are thus converted into gradable ones. Adopting a view allowing for multiple source domains in metaphorical mappings, as the proponents of the conceptual integration theory do, we could then stipulate that this scale is either part of the emergent structure in the blended space, or that this folk model with its scale feeds into an additional input space, and that the blended space contains just the relevant section of the scale. Ordinary right-hand elements are replaced in the blend by their counterparts from the hierarchically higher domain which have a higher qvalue. The inferences that arise from this in the blended space are, first, that we have an extreme degree of knowledge or intolerance accompanied by a wish for oppression and control, respectively, and secondly, probably due to the incongruence of the two domains, that the speaker wishes to attach at least a hint of evaluation, usually resulting in a humorous or derogatory interpretation. Returning to the issue of conversion of non-gradable terms into gradable ones, we also note that hyperbolic expressions may combine large numbers metonymy with a metaphorical conversion of a term from non-gradable to gradable thus implying a scale. This possibility is exemplified by the expression in the first part of the title of our paper. Thousand and ladybird, the latter is taken here to mean an ‘extremely valuable object’ after a metaphorical mapping, reinforce each other in making quite clear the intensifying function of the hyperbole. The reason for this reinforcement may be the fact that the second scale is, so to say, private, not conventionalized. As pointed out in footnote 1, the utterance was produced by a child at an age when hyperboles are not even properly understood, let alone productively used by average children (cf. Winner et al. 1987, who report that errors in interpretation of hyperbole persist at least through age 10). However occasional such expressions may be in child language, the fact is that a modified expression was recorded (more precisely half-elicited) by the second author during a telephone conversation, when the child was asked “How much do you like your mum?” he replied by further reinforcing the expression, “Like a thousand heaps of ladybirds.” All this means that metaphors and metonymies may both prepare the ground for the scalar model and later constrain its workings in order to obtain correct results, i.e. the set of pragmatically appropriate inferences. Pending further research on the scalar model, and its interaction with other cognitive mechanisms, we are inclined to assume that it is basically correct in involving ordered sets of items and scales with default values or norm, as well as in its prediction of entailments, but also that it may just trigger the inferencing process which may be steered and constrained by metonymic and metaphorical mappings.
The truth of hyperboles
415
2.4 Hyperboles as loud liers It hardly needs to be stated that hyperbolic expressions as such cannot be true or false. As the whole first part of our paper amply shows, it is hardly possible to formally identify and define hyperbole as a trope. Certain constructions can be recognized as bringing about hyperbolic effects only within a broader context. In other words, it is more realistic to talk about hyperbolic utterances than about hyperbolic expressions. Then, of course, the question arises whether these hyperbolic utterances are true or false. Most laymen and linguists will readily agree that such hyperbolic statements simply cannot be taken to literally mean what they state. But this is about all the consensus there is about hyperbolic truth, and the search for answers to the above question can go in various directions. If we strictly apply Grice’s (1975) truth maxim to hyperbolic statements, i.e. the expectation that declarative utterances should be literally true, we must conclude that hyperboles are just like errors. Propositions containing errors cannot be taken as they stand to be true propositions. However, hyperboles are almost never interpreted as errors, at least not by adult speakers. In other words, there is a remarkable contradiction here between this obvious “untruth” of hyperboles on the one hand, and between the fact that they are not perception errors/mistakes either, on the other. We might try with the belief maxim, i.e. the expectation that speakers ordinarily believe the literal content of what they say. Utterances violating this maxim are then intentionally false. Lies violate both the truth maxim and the belief maxim. We have just stated that hyperbolic statements violate the truth maxim. But they are at the same time in fact clearly meant to be different from the plain God’s truth, i.e. they are intentionally “false”. This means that they should qualify as lies in truth conditional terms. But on encountering hyperbolic utterances, we do not react to them in the same way that we react to lies. On closer inspection of the differences between lies and hyperbolic statements we find that at least three essential ingredients of prototypical lies are in fact missing. There is first of all no intention to deceive. Hyperbolic expressions assume the form they have not because the speaker wants to deceive but rather because he/she wants to communicate something. Secondly, by lying speakers normally achieve some material or social gain. By uttering a hyperbolic statement, the speaker also achieves a gain, but it is of a different kind. It is a rhet-orical gain. Hyperbole may make it possible for him or her to claim and keep the addressee(s) attention (cf. 1.5), or to express certain concepts in a way that is less costly in the long run. Hyperboles may be cost-effective in pragmatic terms. This may sound surprising at first blush, because, of course, exaggerating things in one’s face may be rude, e.g. telling someone that he took ages to complete a relatively simple task, but stating the facts bluntly without exaggeration or understatement can be just as rude. On the other hand, like all forms of non-literal lan-
416
Rita Brdar-Szabó / Mario Brdar
guage, hyperbolic utterances can be used as less direct speech acts, and therefore make one’s utterance less face-threatening. Last but not least, occasional humorous effects may help defuse potential tensions. Hyperbolic statements may also help to reduce communicative costs. Most obviously, they do two things for the price of one. In addition to stating something they express the speaker’s attitude (surprise, amazement, irritation, etc.). There is another communicative advantage to hyperbole that actually goes unnoticed most of the time. It is usually assumed that when speakers use hyperbole they exaggerate something, a property or event that they have a very clear idea of. Speaking in terms of a scalar model, we tend to think that when using a hyperbole the speaker has a clear idea of the norm for a given scale, and the actual value exhibited, and then simply inflates the difference between the two. But the ubiquity of metaphor and metonymy teaches us that speakers may occasionally have less than clear concepts in their minds, and the tropes are used to negotiate or set off the relative lack of the clarity of the concept. What is more, we know very well that natural languages cannot be expected to lexify all the possible distinctions that can be made on the conceptual level. Metaphors, metonymies and other tropes can again provide a useful handle on such concepts. Hyperbole, too, can function in both these ways and assume a sort of referential function, that is to say, exaggeration (in either direction) may be an indication that the speaker cannot or does not want to commit herself or himself to the exact expression. In this respect, hyperbole is very similar to more referential tropes like metaphor and metonymy. This idea seems to be hinted at already in Priestley (1777). To the extent that they occasionally have this sort of hidden referential function, hyperboles cannot be treated as not being true, just like metaphors and metonymies are true in this respect, because they may enable us to see things which we had not seen before or to see them in a different light. Secondly, hyperbole is very different from a prototypical lie because most of these gains may be shared with interlocutors. At best, as far as pragmatic gains are concerned, hyperboles may be likened to a degree with white lies, because both have a protective function extending to both interlocutors. Thirdly, hyperboles tend to be clearly marked as being different from literally true statements. Puttenham, who coined the English term “overreacher” for hyperbole in The Arte of English Poesie (1589), also used the term “loud lier”. The idea was to indicate that hyperbole is different from ordinary lies in that it conspicuously departs from what is the truth. This marking can assume a number of forms which often combine to make the non-literal intention quite clear, ranging from prosodic phenomena to grammatical and lexical ones. However, a caveat is in order here. Hyperboles need not always be recognizable as such. Hyperbolic statements may denote states of affairs that are realistic, those that are possible but not very likely, and those that are clearly impossible. In this last case, if they refer to patently impossible and absurd situations, they are hard to confuse. If the state of affairs is realistic, and even
The truth of hyperboles
417
if it is possible but not very likely, hyperbole may be confused for something else. As Gibbs (1994: 391) points out, it may be confused with overstatement. One and the same utterance can be meant and/or understood as hyperbolically intended or as simply overstating the truth. What is more, one and the same utterance may in a different situation be perceived as a hyperbole or as an understatement, or even as factual, i.e. truthful. By way of illustration consider the following passage from Marlowe’s Tamburlaine (I. ii.): (61)
I hold the Fates bound fast in iron chains, And with my hand turn Fortune’s wheel about? And sooner shall the sun fall from his sphere Than Tamburlaine be slain or overcome. Draw forth thy sword, thou mighty man-at-arms, Intending but to raze my charmed skin, And Jove himself will stretch his and from heaven To ward the blow, and shield me safe from harm.
In this passage, swarming with hyperbolic statements, we witness a sort of metamorphosis brought about partly through hyperbolic statements whereby the speaker gets raised to the status of a mythic Titan. As a result, hyperbolic statements themselves come to refer to a newly created reality system and the world they describe is thus “true”. In sum, hyperboles are “true” because literalized within the discourse. On a more general level, we must realize that hyperboles cannot but be “true”, given the nature and workings of scalar models. Once we enter the non-literal mode through invoking a scalar model, whatever is predicated of an entity on a scale can no longer be false, regardless of its possibly even being patently absurd in a discourse world dominated by the everyday literal language. If by a linguistic convention the choice of expressions available to express a given concept is practically limited to one or two terms, these must be true in their own right. To use a parallel from the domain of colour terms, we are well aware of the fact that the colour spectrum is carved up differently by different languages. Among the more interesting divergences between languages we find that a single colour term in one language can cover the area for which another language may use two terms. Thus Hungarian has piros and vörös for what the English see as red and Germans as rot. Translating both Hungarian terms by one term in English or German is of course never seen as an instance of mistranslation because the obvious loss of distinctions made in the source language is necessarily dictated by the conventions of the target languages. Similarly, if the scalar model works in such a way that all the terms except for those with extreme q-values are eliminated, the only lexical means left of expressing an idea cannot but be “true” within the model. Last but not least, one and the same utterance may be seen in different contexts (discourse types) by different people as hyperbolic or as literal.
418
Rita Brdar-Szabó / Mario Brdar
These various possibilities will be exemplified on our last case study in 2.5. below. It has been noted that for hyperboles, like understatement and sarcasm, speakers use characteristic intonation (Cutler 1974, Egan 1980). Hyperbolic utterances are typically accompanied by heavy stress and slow speaking rate. On a most general structural level, however, we note a certain degree of iconicity in most hyperbolic utterances. The exaggeration in terms of denotation is mirrored by exaggeration on the level of expression or form. In other words, such utterances are also inflated and tend to be longer in terms of their phonological substance, in terms of the lexical material and/or grammatical constructions involved (e.g. hyperbolic statements containing superlatives become longer in English because of the superlative suffix or the superlative adverbial marker most). A lot of these lexical items and constructions would be absent if it were not for the hyperbolic nature of the utterance. At the same time, we may note a kind of conceptual iconicity at work. Outstanding (or extreme) values ascribed to certain entities on scales involving relevant properties are iconically mirrored by the outstanding (literally out-standing, i.e. standing outside the range of the normal) or extreme status of concepts presented by means of hyperboles, as discussed above. As for grammatical ways of marking non-literalness, we may mention two such devices that have been discussed above. First, note that in English very frequently large numbers and some scalar terms indicating quantities (figurative partitive expressions), if in the singular and used hyperbolically, exhibit indefinite articles.6 This indefinite article is a marker of non-referentiality: the values in question seem to be actually projected from irrealis mental spaces. More or less the same thing happens in counterfactual conditionals and clauses of preference, where we have hypothetical forms of verbs warning us of the irrealis. Lexical ways of indicating non-literalness are very numerous, most importantly the use of scalar terms. To indicate the non-literalness, and perhaps even to amplify the strength of a hyperbole, speakers of English may use the adverb literally. Corresponding adverbs are wide-spread cross-linguistically (buchstäblich in German, doslovno ‘till-letter-ADV’ in Croatian, szószerint ‘word-according’ in Hungarian, etc.). 2.5 Case study on Shir HaShirim (Song of Songs) This case study deals with a passage from Shir HaShirim (Song of Songs). Some basic background information seems to be necessary for a proper understanding of the topic of the analysis. Shir HaShirim is part of the five Megillos, the five holy scrolls that are read at various religious festivals of the ————— 6
Cf. the very first part of the title of the present paper.
The truth of hyperboles
419
Jewish year. This particular scroll is read at Passover, the celebration of the Exodus of the Israelites from Egypt. This text is considered by Jewish believers as Holy of Holies because its subject is the highest form of love possible, the selfless love between God and Israel. For Christian believers it is also a holy text that belongs to the third section of scriptural canon, known as Writings. According to Ohly (1958), Christian exegesis interprets the text – focusing in different periods on different issues – as an expression of the love between God and the church, between Christ and the individual soul of the believer, and between God and the Virgin Mary, respectively, whereas the level of human love can be excluded or included, as suggested in the contemporary interpretation by Balázs Barsi, a Hungarian Franciscan monk. Despite the supreme holiness of the Song of Songs for Jewish and Christian believers, there are nevertheless – as stressed in the preface of the ArtScroll edition (ibid., xi) – “countless generations of readers” […] who “had seen in it the opposite of holiness”, and “the cynical canards and misinterpretations of the ignorant and unbelieving” are also mentioned. We are of the opinion that the identification of some instances of hyperbolic language in this text is closely connected to the basic interpretation possibilities of the text as a whole. As Roman Catholic believers, we are aware that we are “a wild olive shoot” that “had been grafted in among the others and now share in the nourishing sap from the olive root”, as St. Paul puts it in Romans,11: 17, and therefore recognize the basic distinction as holding between the positions of believers and non-believers to be somewhat of a simplification. Believers can be the whole community of Isrealites, the mystic body of Christ (in St Paul’s interpretation of the Church), or any part of it as individual human beings who see the text in question as part of revelation. On the other hand, the nonbeliever, sceptic or ignorant does not accept the divinely inspired nature of this text but considers it as the product of exclusively human efforts. Applying Fauconnier’s mental spaces model and conceptual integration theory as an analytical tool for the description of meaning construction on the part of the recipient, we can assume in the case of believers the setting up of a believe-space relative to the base of experiential reality, while in the latter case, a desire-space is being set up relative to experiential reality space. Let us now be more precise about the space construction processes. The attitudes and expectations of different recipient groups build up different mental space configurations. Instead of a space of experiential reality we observe here either a believe-space which entails the confidence and acceptance of the revealed transcendental reality or a desire-space which entails human longing for absolute never ending love as mirrored in myth. If recipients set up a belief- or a desire-space, they assume in both cases that the text in question is the result of a felicitous speech act. Although the actual analysis focuses on just two lines of Shir HaShirim, we now provide a larger portion of the text crucial to our analysis (8:6–7). The interlinear literal English translation of this passage from Biblical Hebrew original provided in the ArtScroll edition reads as follows:
420
Rita Brdar-Szabó / Mario Brdar
(62) ‘set me as a seal upon your heart, as a seal upon your arm’ ‘that love is strong as death’ ‘jealousy is hard as the grave’ ‘its flashes are flashes of fire which is a flame of God’ ‘Many waters cannot extinguish the love’ ‘and rivers cannot drown it’ ‘if a man would give all the substance of his house in exchange for love’ ‘he would be laughed to scorn’. The main communicative aim of this passage is summarized on in the ArtScroll edition (ibid., 196) as follows: “The flow of the verses is that nothing can extinguish true love, nor can love be bought with even great treasure.” As we cannot go into the details of exegesis here, we will be concerned in our analysis with the complex interplay of scalar reasoning, conceptual blending, metaphor and metonymy as well as with how the position of believers versus non-believers influences the identification of hyperbole as such. The focus of the analysis will be on the following line: (63)
Many waters cannot extinguish the love.
The Latin translation (Vulgata) reads: (64)
Aquae multae non poterunt extinguere caritatem.
In Middle High German (Williram) these two lines are translated as: (65)
Hévigiv uvázzer ne móhton írléskan dîe minna (cf. Ohly, ed., 1998: 1223).
Sankt Trudperter Hohes Lied renders them as follows: (66)
Diu manegen wazzer mahten niht erleschen die minne mînere gemahlen (cf. Ohly, ed., 1998: 290).
Also compare the Modern High German translation: (67)
Die Ströme Wassers konnten die Liebe zu meiner Geliebten nicht auslöschen (cf. Ohly, ed., 1998: 291).
We might as well add that the hyperbolic effects can be strengthened; see e.g. the Hungarian translation (published by the St Stephen Society) where instead of “rivers”, the expression “sea-amount-like” is used, which seems to confer another metaphorical level on water:
The truth of hyperboles
(68)
Tengernyi víz sem olthatja sea-amount water not-even extinguish-can
421
el a szerelmet PREF DEF love.
At first look, and particularly when taken outside the context, most of these sound like incoherent statements. On closer inspection, however, most of them turn out to be good candidates for hyperbole. The twist in this part of the text is that it does not say “Many waters cannot extinguish the fire” but rather “Many waters cannot extinguish the love.” If the context is enlarged by adding “Rivers cannot drown/wash away/quench this fire”, as actually happens in the next line, then the likelihood of hyperbolic reading definitely increases. It appears that the meaning construction process utilizes a complex integrated conceptual network relying on metaphorical projections and a folk theoretical model. In this enlarged context, it becomes obvious that four elements are involved in this network, two pairs of elements that are mutually related by four links. First of all, we have water and fire as one pair, and then some constraining or destructive influence or power and love. The two pairs are obviously related by metaphorical mappings. Water and fire, belonging to more concrete domains of physical reality, serve to structure the more abstract and less-easy-to-grasp domain of love. Fire corresponds to love, and water corresponds to the unknown constraining or destructive influence or power. The folk theoretic model that tightens and motivates the blend is based on a simple observation from the realm of experiential reality and is also strengthened by cultural socialization. Its essence is: Water extinguishes fire. The whole (folk theoretic) story is however much more complicated than it may appear at first blush because an integral part of it is a scalar model with two quantifiable variables. The probability of extinguishing the fire by water is calculated by computing the quantity values of the two variables. Scalar reasoning stems from the observation that the more water available and the smaller the intensity of fire, the more likely it is that water will extinguish the fire. The two lines that are of central concern to us, however, make explicit reference to only two of these four elements, water and love, while fire and the unknown power are left implicit. There is no doubt that the two lines are the result of an intricate conceptual integration. The utterance “Many waters cannot extinguish the love” is the result of conceptual blending whereas input space 1 contains the two elements “water” and “fire” and their relation characterized by the possibility of extinguishing the latter by the former. Input space 2 contains “love” and an unspecified cause which could possibly diminish or destroy it. The generic space includes an abstract schema of an entity, state or event and an unspecified force which could possibly bring about a decrease in intensity or put an end to a state-of-affairs. The integrated space contains the elements “water” and “love” standing in a relation such that the latter can potentially be “extinguished” by the former. The components of this integrated structure are analogically matched to their counterparts in the input
422
Rita Brdar-Szabó / Mario Brdar
and generic spaces in the following ways: “water” to unspecified cause, and “extinguish” to the possibility of causing decrease in the intensity or to the possibility of destroying the state of love. Returning now to our main problem here, namely whether the utterance can be considered hyperbolic or not, and whether it is true or not, the answers are paradoxical: it is simultaneously both. Unless we have the knowledge of the special circumstances under which the described state of affairs obtains as well as the knowledge of the wider context of the utterance, we are not in the position to decide whether any hyperbolic effects emerge or not. First, note that the generically quantified description “many waters” is matched by a generic constraining or destructive power which can be interpreted either as a very serious obstacle or as many different destructive causes, specified by different values in theological exegesis. The inference is then that the love in question must be great and stable if it can resist such a powerful antagonist. Of course, the inferencing can go into the other direction, depending on what we know. If we know that the love in question is divine and therefore great and stable, then we might conclude that its antagonist must be out of the ordinary, too, concerning its magnitude and power (although it cannot prevail). Note also that both concepts in question, God’s love and its adversary, are beyond the grasp of rational human thought and that they are both metaphors. Fire is of course a metaphor for God’s love, but the term God’s love itself is no less metaphorical. If one of the two is hyperbolically intensified, namely waters standing for the adversary cause, it is first an instance of “very loud lying” because literally waters cannot possibly extinguish love. The presence of the intensifier prompts us to look for a metaphorical reading. But once we have entered the network of conceptual integration and waters are interpreted metaphorically, the intensification is no longer a case of lying: the constraining or destructive power is extreme. And if we are not able to conceptualize it satisfactorily in a non-metaphorical way, then we are probably not able to express the magnitude in no other way but by a hyperbolic expression. At the same time, recalling that the scalar model brings together q-values of these two entities that function as variables, we realize that the hyperbolic intensification must also apply to God’s love, and that the utterance is then eventually true, and therefore in the long run not to be considered a hyperbole at all. Secondly, the emergence of hyperbolic effects is connected to the desire space set up in the meaning construction process by non-believers, who however at the same time assume the felicitousness of the speech act. The motivation for the putatively almost natural hyperbolic interpretation is based on folk theoretic and cultural models which teach us that absolute, never ending love without boundaries or constraints is not possible between mortals in experiential reality, given the fact that death definitively puts an end to it. The truth of hyperbole boils down in this case to the expression of the strength of the experienced emotion. On the other hand, in the believe space set up in the interpretation process by recipients who have faith and confidence in the revealed
The truth of hyperboles
423
transcendental content there is less room for hyperbolic reading because the utterance is meant and interpreted as literally true, in the absolute sense, on at least one level. As our last point here, we would like to add that hyperboles can be identified as such only in optimally large contexts. In our case a full and appropriate analysis should involve all intertextual occurrences of the images of water and fire in the Bible which can be modelled as different input spaces in a complex integrated conceptual network. 3. Concluding remarks To sum up our discussion, we can say that in seeking to answer two related questions concerning hyperbole, whether it can be identified and/or defined as a trope or a figure, and whether hyperbolic expressions are to be considered as just another case of flouting one or more of Gricean maxims, the data and arguments presented in this paper seem to warrant two cautiously formulated conclusions. Firstly, pure hyperbole is hardly possible as such, as hyperbolic effects are a basically contextually induced and pragmatically embedded phenomenon. It has been shown furthermore that hyperbolic effects can be identified as such only under full consideration of the expectations and mental states/constructs of the recipient and the communicative aims of the producer of the utterance, which also means that the problem of truth of hyperboles should be approached from the perspective of the basic interpretation possibilities of a text as a whole. References Abbott, Barbara: “Support for a unique theory of definiteness”, in: Matthews, Tanya, Devon Strolovich, Devon (eds.): Proceedings from Semantics and Linguistic Theory IX. Ithaca, NY: Cornell University 1999, 1–15. Austin, Frances O.: “‘A crescent-shaped jewel of an island’, appositive nouns in phrases separated by of”, in: English Studies 61 (1980), 357–66. Barcelona, Antonio (ed.): Metaphor and Metonymy at the Crossroads. A Cognitive Perspective. (Topics in English Linguistics 30). Berlin–New York: Mouton de Gruyter 2000. Barsi, Balázs: A szeretet misztériuma. Elmélkedések az Énekek énekérıl. Budapest: Efo Kiadó 1996. Bennett, W. A.: “One type of expressive Noun Phrase in French”, in: Nottingham Linguistic Circular 5.1 (1976), 20–1. Blair, Hugh: Lectures on Rhetoric and Belles Lettres. Halifax 1842. Brdar, Mario: “How pure is the pure hyperbole? The role of metonymic mappings in the construction of some hyperbolic effects”, in: Kučanda, Dubravko, Mario Brdar, Boris Berić (eds.): Teaching English for Life. Studies to Honour Prof. Elvira
424
Rita Brdar-Szabó / Mario Brdar
Petrović on the Occasion of Her 70th Birthday. Osijek: Filozofski fakultet 2004, 373–85. Brdar-Szabó, Rita, Mario Brdar: “Collocational potential in language contact”, in: Dakowska, Maria (ed.): English in the Modern World. Festschrift for Hartmut Breitkreuz on the Occasion of His Sixtieth Birthday. (Foreign Language Studies 5). Frankfurt/Main–Berlin–Bern–Bruxelles–New York–Oxford–Wien: Peter Lang 2000, 201–12. – “Scalar models in a cognitive approach to hyperbolic expressions: With a little help from metonymy”, in: Cap, Piotr (ed.): Pragmatics Today. (Studies in Language 12). Frankfurt/Main–Berlin–Bern–Bruxelles–New York–Oxford–Wien: Peter Lang 2005, 75–94. Burkhardt, Armin: “Zwischen Poesie und Ökonomie. Die Metonymie als semantisches Prinzip”, in: Zeitschrift für germanistische Linguistik 24.2 (1996), 175–94. Cicero, Marcus Tullius: Orator. Lateinisch-deutsch. Ed. Bernhard Kytzler. München– Zürich 1988. Colston, Herbert L.: “‘I’ve never seen anything like it’: Overstatement, understatement, and irony”, in: Metaphor and Symbol 12.1 (1997), 43–58. – , Shauna B. Keller: “‘You’ll never believe this’: Irony and hyperbole in expressing surprise”, in: Journal of Psycholinguistic Research 27.4 (1998), 499–513. Cruse, D. Alan: Lexical Semantics. Cambridge: Cambridge University Press 1986. Cutler, Ann: “On saying what you mean without meaning what you say”, in: CLS 10 (1974), 117–27. Dietz, Hans-Ulrich: Rhetorik in der Phraseologie. Zur Bedeutung rhetorischer Stilelemente im idiomatischen Wortschatz des Deutschen. (Reihe Germanistische Linguistik 205). Tübingen: Max Niemeyer Verlag 1999. Egan, Owen: “Intonation and meaning”, in: Journal of Psycholinguistic Research 9 (1980), 23–39. Fahnestock, Jeanne: Rhetorical Figures in Science. New York: Oxford University Press 1999. Fauconnier, Gilles: “Pragmatic scales and logical structures”, in: Linguistic Inquiry 6.3 (1975), 353–75. – Mental Spaces. Aspects of Meaning Construction in Natural Language. Cambridge: Cambridge University Press 1994. – Mappings in Thought and Language. Cambridge: Cambridge University Press 1997. – , Mark Turner: “Conceptual integration networks”, in: Cognitive Science 22.2 (1998), 133–87. Feyaerts, Kurt: “Metonymic hierarchies: The conceptualization of stupidity in German idiomatic expressions”, in: Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. (Human Cognitive Processing 4). Amsterdam– Philadelphia: John Benjamins 1999, 309–32. Fillmore, Charles J., Paul Kay, Mary Catherine O’Connor: “Regularity and idiomaticity in grammatical constructions: The case of let alone”, in: Language 64.3 (1988), 501–38.
The truth of hyperboles
425
Fraunce, Abraham: The Arcadian Rhetorike. Ed. by Ethel Seaton. Oxford 1588 (reprint 1950). Gibbs, Raymond W., Jr.: The Poetics of Mind: Figurative Thought, Language, and Understanding. Cambridge: Cambridge University Press 1994. – “Irony in talk among friends”, in: Metaphor and Symbol 15.1-2 (1998), 5–27. – , Gerard J. Steen (eds.): Metaphor in Cognitive Linguistics. Selected Papers from th the 5 International Cognitive Linguistics Conference, Amsterdam, July 1997. (Amsterdam Studies in the Theory and History of Linguistic Sience. Series IV. Current Issues in Linguistic Theory 175). Amsterdam–Philadelphia: John Benjamins 1999. Grice, H. Paul: “Logic and conversation” in: Cole, Peter, Jerry L. Morgan (eds.): Speech Acts. (Syntax and Semantics 3). San Francisco–New York–London: Academic Press 1975, 41–58. Holmqvist, Kenneth, Jarosław Płuciennik: “Virtual scanning in hyperbolas”, in: Lewandowska-Tomaszczyk, Barbara, Alina Kwiatkowska (eds.): Imagery in Language. Festschrift in Honour of Professor Ronald W. Langacker. (Łódź Studies in Language 10). Frankfurt/Main–Berlin–Bern–Bruxelles–New York–Oxford–Wien: Peter Lang 2004, 777–85. Israel, Michael: “Polarity sensitivity as lexical semantics”, in: Linguistics and Philosophy 19.6 (1996), 619–66. – “The scalar model of polarity sensitivity”, in: Forget, Danielle, Paul Hirschbühler, France Martineau, María Luisa Rivero (eds.): Negation and Polarity: Syntax and Semantics. Amsterdam–Philadelphia: John Benjamins 1997, 209–29. – The Rhetoric of Grammar: Scalar Reasoning and Polarity Sensitivity. Ph.D. Dissertation, U.C. San Diego 1998. – “Minimizers, maximizers and the rhetoric of scalar reasoning”, in: Journal of Semantics 18.4 (2001), 297–331. – “Literally speaking”, in: Journal of Pragmatics 34.4 (2002), 423–32. Jäkel, Olaf: “Metonymy in onomastics”, in: Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. (Human Cognitive Processing 4). Amsterdam–Philadelphia: John Benjamins 1999, 211–29. Jansen, Carel J.M., [Thijs] M.M.W. Pollmann: “On round numbers: Pragmatic aspects of numerical expressions”, in: Journal of Quantitative Linguistics 8.3 (2001), 187– 201. Jónácsik, László: Poetik und Liebe: Studien zum liebeslyrischen Paradigmenwechsel, zur Petrarca- und zur Petrarkismus-Rezeption im ‘Raaber Liederbuch’. Zugleich ein Beitrag zur Geschichte protestantischer ‘Renaissancelyrik’ in Österreich. (Mikrokosmos. Beiträge zur Literaturwissenschaft und Bedeutungsforschung 48). Frankfurt/Main–Berlin–Bern–Bruxelles–New York–Oxford–Wien 1996: Peter Lang. – “Zur Parodie der ‘Hohen Liebeslyrik’ im Deutschland des 17. Jahrhunderts: Die Gattung ‘Schönheitsbeschreibung’”, in: Nagy, Márta, László Jónácsik (eds.): ‘swer sînen vriunt behaltet, daz is lobelîch.’ Festschrift für András Vizkelety zum 70. Geburtstag. Piliscsaba–Budapest 2001, 309–21.
426
Rita Brdar-Szabó / Mario Brdar
Kreuz, Roger J., Richard M. Roberts: “Two cues for verbal irony: Hyperbole and the ironic tone of voice”, in: Metaphor and Symbolic Activity 10.1 (1995), 21–31. Lausberg, Heinrich: Elemente der literarischen Rhetorik. Eine Einführung für Studierende der klassischen, romanischen, englischen und deutschen Philologie. Ismaning: Max Hueber Verlag 1990a. – Handbuch der literarischen Rhetorik. Eine Grundlegung der Literaturwissenschaft. Stuttgart: Franz Steiner Verlag 1990b. Lehrer, Adrienne: Semantic Fields and Lexical Structure. Amsterdam: North-Holland 1974. Lyons, John: Semantics. 2 vols. Cambridge: Cambridge University Press 1977. McCarthy, Michael, Ronald Carter: “‘There’s millions of them’: Hyperbole in everyday conversation”, in: Journal of Pragmatics 36 (2004), 149–84. Naschert, G.: “Hyperbel”, in: Ueding, Gert (ed.): Historisches Wörterbuch der Rhetorik. Band 4: Hu–K. Tübingen: Max Niemeyer Verlag 1998, 115–22. Nerlich, Brigitte, David D., Zazie Todd: “‘Mummy, I like being a sandwich’: Metonymy in language acquisition.” In: Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. (Human Cognitive Processing 4). Amsterdam–Philadelphia: John Benjamins 1999, 361–83. Norrick, Neal R.: “Hyperbole, extreme case formulation”, in: Journal of Pragmatics 36.9 (2004), 1727–39. Ohly, Friedrich: Hohelied-Studien. Grundzüge einer Geschichte der Hoheliedauslegung des Abendlandes bis um 1200. Wiesbaden: Franz Steiner Verlag 1958. – (ed.): Das St. Trudperter Hohelied. Eine Lehre der liebenden Gotteserkenntnis. (Bibliothek des Mittelalters 2). Frankfurt/Main: Deutscher Klassiker Verlag 1998. Paardekooper, P. C.: “Een Schat Van Een Kind”, in: Nieuwe Taalgids 49, 1956, 93–9. Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. (Human Cognitive Processing 4). Amsterdam–Philadelphia: John Benjamins 1999. Peacham, Henry: The Garden of Eloquence, Conteining the Most Excellent Ornaments, Exornations, Lightes, Flowers, and Formes of Speach, Commonly Called the Figures of Rhetorike. London 1589. Plett, Heinrich F.: Einführung in die rhetorische Textanalyse. Hamburg: Helmut Buske Verlag, 1991, 8. Aufl. Priestley, Joseph: A Course of Lectures on Oratory and Criticism. Ed. by V. M. Bevilacqua and R. Murphy. 1965 [1777]. Puttenham, George: The Arte of English Poesie. Ed. by Gladys Doidge Willcock and Alice Walker. Cambridge, 1936 [1589]. Quintilianus, Marcus Fabius: Ausbildung des Redners. Zwölf Bücher. Herausgegeben und übersetzt von Helmut Rahn. 2 Bände. Darmstadt 1988. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, Jan Svartvik: A Comprehensive Grammar of the English Language. London: Longman 1985. Radden, Günter, Zoltán Kövecses: “Towards a theory of metonymy”, in: Panther, Klaus-Uwe, Günter Radden (eds.): Metonymy in Language and Thought. (Human Cognitive Processing 4). Amsterdam–Philadelphia: John Benjamins 1999, 17–59. Rhetorica ad Herennium. Herausgegeben und übersetzt von Theodor Nüßlein. Darmstadt 1994.
The truth of hyperboles
427
Römer, Ruth: Die Sprache der Anzeigenwerbung. (Sprache der Gegenwart 4). Düsseldorf: Pädagogischer Verlag Schwann 1968. Ruiz de Mendoza Ibáñez, Francisco José: “The role of mappings and domains in understanding metonymy”, in: Barcelona, Antonio (ed.): Metaphor and Metonymy at the Crossroads. A Cognitive Perspective. (Topics in English Linguistics 30). Berlin–New York: Mouton de Gruyter 2000, 109–32. – , Sandra Peña Cervel: “Cognitive operations and projection spaces”, in: Jezikoslovlje 3.1-2 (2002), 131–57. Sherry, Richard: A Treatise of Schemes and Tropes. London 1550. Shir HaShirim. Song of Songs. An Allegorical Translation Based upon Rashi with a Commentary Anthologized from Talmudic, Midrashic and Rabbinic Sources. (The ArtScroll Tanach Series). Brooklyn: Mesorah Publications 1979. Stanivuković, Goran: “Hiperbola u kontekstu engleske renesansne književnosti”, in: Benčić, Živa, Dunja Fališevac, ur. Tropi i figure. Zagreb: Zavod za znanost o književnosti Filozofskog fakulteta Sveučilišta u Zagrebu 1995, 221–51. Stempel, Wolf-Dieter: “‘Ich vergesse alles.’ Bemerkungen zur Hyperbolik in der Alltagsrhetorik”, in: Faust, Manfred (ed.): Allgemeine Sprachwissenschaft, Sprachtypologie und Textlinguistik. Festschrift für Peter Hartmann. Tübingen 1983, 87–98. Turner, Mark: “Figure”, in: Katz, Albert N., Cristina Cacciari, Raymond W. Gibbs, Mark Turner: Figurative Language and Thought. New York–Oxford: Oxford University Press 1998, 44–87. Ueding, Gert, Bernd Steinbrink: Grundriß der Rhetorik: Geschichte, Technik, Methode. Stuttgart: Metzler 1986. Wells, Henry: Poetic Imagery: Illustrated from Elizabethan Literature. New York 1918. Wilson, Thomas: The Rule of Reason, Conteyning the Arte of Logique, set forth in English. London 1551. Winner, Ellen, Gail Windmueller, Elizabeth Rosenblatt, Laurie Bosco, Edella Best, Howard Gardner: “Making sense of literal and nonliteral falsehood”, in: Metaphor and Symbolic Activity 2.1 (1987), 13–32.