Semblance and Signification (Iconicity in Language and Literature)

Semblance and Signification

Iconicity in Language and Literature A multidisciplinary book series which aims to provide evidence for the pervasive presence of iconicity as a cognitive process in all forms of verbal communication. Iconicity, i.e. form miming meaning and/or form miming form, is an inherently interdisciplinary phenomenon, involving linguistic and textual aspects and linking them to visual and acoustic features. The focus of the series is on the discovery of iconicity in all circumstances in which language is created, ranging from language acquisition, the development of Pidgins and Creoles, processes of language change, to translation and the more literary uses of language. For an overview of all books published in this series, please see http://benjamins.com/catalog/ill

Editors

Olga Fischer

University of Amsterdam

Christina Ljungberg University of Zurich

Volume 10 Semblance and Signification Edited by Pascal Michelucci, Olga Fischer and Christina Ljungberg

Semblance and Signification Edited by

Pascal Michelucci University of Toronto

Olga Fischer University of Amsterdam

Christina Ljungberg University of Zurich

John Benjamins Publishing Company Amsterdamâ•›/â•›Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data Symposium on Iconicity in Language and Literature (7th : 2009 : Victoria University (Toronto, Canada)) Semblance and signification / edited by Pascal Michelucci, Olga Fischer, Christina Ljungberg. p. cm. (Iconicity in Language and Literature, issn 1873-5037 ; v. 10) Papers given at the Seventh International Symposium on Iconicity in Language and Literature, hosted by Victoria University in the University of Toronto (Canada) and held on the Victoria College campus, 9-13 June, 2009. Includes bibliographical references and index. 1. Iconicity (Linguistics)--Congresses. I. Michelucci, Pascal. II. Fischer, Olga. III. Ljungberg, Christina, 1949P99.4.I26S96â•…â•… 2011 415--dc23 isbn 978 90 272 4346 1 (Hb ; alk. paper) isbn 978 90 272 8482 2 (Eb)

2011026794

© 2011 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

Table of contents Introduction Pascal Michelucci

xi

part i. Word forms, word formation, and meaning Toward a phonosemantic definition of iconic words Kimi Akita Iconic thinking and the contact-induced transfer of linguistic material: The case of Japanese, signed Japanese, and Japan sign language William J. Herlofsky

3

19

Ezra Pound among the Mawu: Ideophones and iconicity in Siwu Mark Dingemanse

39

Cognitive iconic grounding of reduplication in language Olga Fischer

55

Imagic iconicity in the Chinese language Zhuanglin Hu

83

Words in the mirror: Analysing the sensorimotor interface between phonetics and semantics in Italian Luca Nobile

101

part ii. General theoretical approaches Un mélange genevois: Tacit notions of iconicity in Ferdinand De Saussure’s Writings in General Linguistics Jui-Pi Chien How to put art and brain together Mark Changizi Image, diagram, and metaphor: Unmined resource and unresolved questions Vincent Colapietro

135 149

157

ïš© Semblance and Signification

part iii. Narrative grammatical structures The farmers sowed seeds and hopes: Element order in metaphorical phrases Yeshayahu Shen & Elad Kotzer Non-iconic chronology in English narrative texts Vyacheslav Yevseyev A burning world of war: How iconicity works in constructing the fictional world view in A Farewell to Arms Xinxin Zhao

175 191

211

part iv. Cognitive poetics Aesthetic qualities as structural resemblance: Divergence and perceptual forces in poetry Reuven Tsur

233

Mental space mapping in classical Chinese poetry: A cognitive approach Han-liang Chang

251

Iconicity in conceptual blending: Material anchors in William Morris’s News from Nowhere David Glyn Wilson

269

part v. Acoustic and visual iconicity Thematized iconicity and iconic devices in the modern novel: Some modes of interaction John J. White

291

Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy Gabriele Rippl

313

Words, like shells, are signs as well as things Heilna du Plooy

327

Unveiling creative subplots through the non-traditional application of diagrammatic iconicity: An analysis of Kingsley Amis’s The Green Man Andrew James

343



Table of contents ïš©ïš©

part vi. Intermedial iconicity The iconic indexicality of photography Piotr Sadowski

355

Unbinding the text: Intermedial iconicity in Peter Greenaway’s Prospero’s Books Christina Ljungberg

369

Argumentative, iconic, and indexical structures in Schubert’s Die schöne Müllerin Costantino Maeder

389

John Irving’s A Widow for One Year and Tod Williams’ The Door in the Floor as ‘(mult-)i-conic’ works of art Christine Schwanecke

405

Author index

423

Subject index

425

Preface and acknowledgements The essays in this volume form a careful selection of papers, most relevant for this series, given at the Seventh International Symposium on Iconicity in Language and Literature, hosted by Victoria University in the University of Toronto (Canada) and held on the Victoria College campus. It took place 9–13 June, 2009, during which forty papers were presented on a wide range of topics involving the notion of Â�iconicity in linguistics (phonology, morphology, and syntax), literary works, media studies, cinema, music, painting and architecture as well as featuring a workshop on Cognitive Poetics. The participants represented twenty different nations, including Austria, Belgium, Canada, China, France, Germany, Hungary, Ireland, Italy, Japan, Norway, Romania, Saudi Arabia, South Africa, Sweden, Switzerland, Taiwan, The Netherlands, The United Kingdom, and The United States. The editors of this volume would like to express their gratitude to those who helped make the seventh in this series of symposia, which have been organized every second years for the past fourteen years, such a fruitful and inspiring event. Our Â�special thanks go to Paul W. Gooch, President of Victoria University in the University of Toronto, and David Klausner, Vice-Dean for Interdisciplinary Affairs, University of Toronto, who supported the symposium from its very start and who let us use the dignified Victoria College as our conference venue; Paul Bouissac who was the local coorganizer together with Pascal Michelucci for their indefatigable and inspired Â�organization; Ms Ann Lewis, Executive Secretary, and her team of student volunteers who gave us so much of their time and energy; and Laura Grace Conlon, for her dedication to the attentive review of the manuscript in its early stages. We also would like to thank all those scholars who helped us to referee the papers, thus enabling us to produce a volume that meets the high Â�standards of the scientific community. We gratefully acknowledge the financial support of Victoria University (for a Social Sciences and Humanities Research Council Institutional Grant (SIG) Award), Victoria College (Principal Office), the Social Sciences and Humanities Research Council of Canada, , the Amsterdam Research Center for Language and Communication (ACLC) of the University of Amsterdam, and the Swiss Academy of Humanities and Social Sciences. 

P.M., O.F., and C. L.

Introduction Pascal Michelucci

Cognitive poeticians and cognitive linguists have observed that linguistic Â�expression makes use, to varying degrees but at times with striking effects, of special and creative applications of otherwise automatic cognitive structures and processes. Their collective approach to the use of aesthetic and expressive dimensions of language is in fact one that is more economical than positing a separate and independent Â�aesthetic cognitive module or capacity such as inspiration or taste, the semiotic basis for which is fuzzy at best. Iconicity, as a semiotically rich field, enjoys pride of place for the developing understanding of the mental basis for such creative uses of languages. Nowhere is the remotivation and overdetermination of Â�natural signs more resourcefully exploited than in literary, and generally artistic, Â�contexts. Â�Structures are put into unforeseen new uses, layered patterns of signification emerge in the dynamic integration of complementary modes, such as in ecphrasis or multimedia performance. Prevalent and extensive types of linguistic signs make use of reduplication, structural resemblances, ordering, mappings, (re)categorization. A cognitively inspired approach to iconicity therefore offers a unique vantage point from which to consider and reframe the notion of literariness and linguistic creativity thanks to a strongly grounded semiotic perspective. Thus the embodied properties of cognition, the semiotic affordances of linguistic and artistic signs, the view in favour of “nature harnessing” proposed by Mark Changizi (this volume), metaphorical and analogical reasoning, among many other concerns, echo through this volume with obvious prominence end recurrence. The articles assembled in Semblance and Signification interrogate the role of iconicity in the formation of concepts and communication, in their encoding and decoding, with attention to its necessary role as well as to new functions in Â�original configurations of iconic and non-iconic signs. Together, the articles explore the Â�relevance and workings of iconicity in a range of languages (English, Italian, Japanese, Chinese, Siwu, and Sign Language), of artistic media (literature, cinema, music, photography, and mixed-media works), and of issues at the juncture between iconicity and other key semiotic issues (motivation and its frontiers, Â�diagrammatic and metaphoric expressions, indexicality, multimodality). They also share a strong concern for the cognitive basis of iconicity and its role in a multimodal perspective. If iconicity has been looked at as a naturally driven relationship between the form of a sign and the mental representation of its referent,

 Pascal Michelucci

and has remained a key interest for the theory of signs since C.S. Peirce, dedicated investigations into the mental basis for the construction of some categories of signs have been relatively few. It is our hope that this volume contributes to narrowing the gap by drawing forcefully and convincingly from what Bernard Baars has nicknamed the “cognitive revolution” (1986), and by bringing to semiotic inquiry research from cognitive linguistics and cognitive poetics to illuminate iconic questioning. In Acts of Meaning, some twenty years ago, Jerome Bruner (1990) advocated for such a semiotic turn to the cognitive turn, to provide a foundation for what he called the “proper study of man”: There is no question that cognitive science has made a contribution to our understanding of how information is moved about and processed. Nor can there be much doubt on reflection that it has left largely unexplained and even somewhat obscured the very large issues that inspired the cognitive revolution in the first place. So let us return to the question of how to construct a mental science around the concept of meaning and the processes by which meanings are created and negotiated within a community. (1990: 10–11)

The question of cognition has since federated multiple disciplines around a more sophisticated understanding of information-processing activities, from the analysis of sensory and linguistic data, to applied and experimental research into Â�processing, through to the organisation of subjective intellectual, emotional, and aesthetic responses. Similarly, the papers gathered in Semblance and Signification draw from a range of disciplines and yield fruitful dialogue from their collaboration. We hope that the reader will find what we intended to offer – a multifaceted and inspiring contribution to a highly dynamic field.

References Baars, Bernard. 1986. The Cognitive Revolution in Psychology. New York NY: Guilford. Bruner, Jerome. 1990. Acts of Meaning. Cambridge, MA: Harvard University Press.

part i

Word forms, word formation, and meaning

Toward a phonosemantic definition of iconic words Kimi Akita*

University of California, Berkeley / University of Tokyo Most studies have tried to define inherently iconic words (mimetics, ideophones) in terms of their formal features but phonosemantic peculiarity, assumed without empirical consideration, is not evidently distinct from regular sound symbolism. Two experiments were conducted to probe the phonosemantic specificity of iconic words. Experiment 1 asked twenty native Japanese speakers to rate 140 novel words, half of which had a shape typical of Japanese iconic words: no systematic difference in consonantal or vocalic symbolism between the two types of stimuli was obtained. Experiment 2 asked twenty native Japanese speakers to judge the consonantal magnitude symbolism of 120 verbs with or without a typical iconic word shape presented in a referentially specific sentence. Verbs sharing a root and a morphophonological shape with an existent iconic word tended to yield sharper magnitude contrasts. Iconic words appear to have marked phonosemantic status, which is grounded on both their formal and referential markedness.

1.â•… Introduction This paper aims at experimentally clarifying the (phono)semantic definition of Â�sound-symbolic, iconic words in Japanese by focusing on both their form and Â�meaning. Like Basque, Cantonese, Korean, Zulu, etc., Japanese has a rich lexicon of inherently iconic words called ‘mimetics’, which basically correspond to ‘ideophones’ in Â�African

*╇ An earlier version of this paper was presented under the title of “The lexical basis of sound

symbolism of ‘sound-symbolic words’ in Japanese” at the ILL 7 Symposium. My sincere Â�gratitude goes to Olga Fischer and William Herlofsky, who offered detailed Â�comments, which led this paper to some fundamental improvement and development. I also thank Mark Â�Dingemanse, Hideki Kishimoto, Yo Matsumoto, and Lawrence Schourup for their constructive advice and suggestions. Remaining inadequacies are, of course, my own. This study is partly supported by Grant-in-Aid for JSPS Fellows to the author entitled “A crosslinguistic and developmental study of the syntax and semantics of sound-symbolic words: On the roles of lexical iconicity” (21·2238). Some statistical details of the present study can be found in Akita (2009: Chapter 4).



Kimi Akita

Â� linguistics and ‘expressives’ in South(east) Asian linguistics.1 As illustrated in (1) and (2), Japanese iconic words can depict not only auditory but also non-auditory Â�experiences, such as visual, textural, and emotional ones. (1) Auditory (i.e. onomatopoeic):2 a. Neko-ga nya^anyaa nai-te i-ta. (animate) cat-nom iw cry-conj be-pst ‘A cat was crying meow-meow.’ b. Huurin-ga riiN^-to nat-ta. (inanimate) wind.bell-nom iw-quot sound-pst ‘A wind-bell sounded ring.’ (2) Non-auditory: a. Hosi-ga ki^rakira kagayai-te i-ru. (visual) star-nom iw shine-conj be-npst ‘A star is shining in a twinkling manner.’ b. Ame-ga yan-de hoQ-to si-ta. (emotional) rain-nom stop-conj iw-quot do-pst ‘[I] got relieved that it stopped raining.’ The definition of iconic words has been one of the hardest problems in the study of this word class in both Japanese (Hamano 1998: 6–7; Tamori and Schourup 1999) and other languages (Fortune 1962; P. Newman 1968, 2001; Johnson 1976; Fordyce 1983; Childs 1994; Bartens 2000; Bodomo 2006; Dingemanse, 2011). It is largely agreed upon that iconic words are basically morphophonologically marked and Â�noncanonical (Samarin

.â•… The term ‘sound symbolism’ refers to a phenomenon in which the form and meaning of a word have a (partial) iconic, motivated relationship. Sound symbolism in a broad sense can be found at both morphophonological (e.g. reduplication, syllable structure, prosody) and segmental levels (e.g. vowel height, labiality, voicing of obstruents). Generally speaking, the former instantiates diagrammatical iconicity, whereas the latter instantiates imagic iconicity in Peirce’s (1932) terminology (cf. Haiman 1980:â•›515). In the psychological tradition, the imagic (or phonosemantic) type of sound symbolism, which is also an instance of phonetic/ phonological iconicity, is often called “phonetic symbolism” (Sapir 1929). Words with clear sound symbolism are generally called ‘sound-symbolic words’ or ‘iconic words’. They illustrate lexical iconicity. .â•… Abbreviations and symbols used in this paper are as follows: C = consonant; conj = conjunctive; iw = iconic word; N = moraic nasal (only for iconic and novel words); nom = nominative; pst = past; Q = first half of a geminate cluster (realized as an unpronounced obstruent, as in hoQ-to /hotto/ in (2b)); quot = quotative; top = topic; V = vowel; VV = long vowel; ^ = pitch fall (only for iconic and novel words); - (hyphen) = reduplicant boundary (only in templatic representations of words).



Toward a phonosemantic definition of iconic words

1971; Hinton et al. 1994: 9; Klamer 2002), as represented by their general preference for reduplicative forms (e.g. (1a), (2a)). However, there has been no successful attempt to define iconic words from a semantic or phonosemantic point of view. Often cited, Doke’s general definition of an iconic word covers its semantic aspect: “[a] vivid representation of an idea in sound”, where an idea is about “manner, colour, sound, smell, action, state or intensity” (1935:â•›118). Doke’s mention of ‘vividness’ may be compatible with the fairly distinct phonosemantic status given to iconic words by their native users (Tamori and Schourup 1999:â•›8). Nevertheless, this kind of impressionistic characterization is many removes from being empirical. The present paper approaches this problem from a pair of experiments with Japanese speakers, pointing out that referential specificity, in addition to peculiar morphophonology, is a general requirement for the special depictive function of iconic words. The organization of this paper is as follows. Section 2 will provide a brief survey of previous arguments about the phonosemantics of Japanese words, setting up the research question to be examined. Section 3 will introduce the previously identified formal condition of Japanese iconic words, which is formulated in terms of schematic morphophonological templates. Section 4 will report the first experiment, which used novel words whose referents were not fully specified, and show that the formal condition alone cannot define iconic words. Section 5 will report the second experiment, whose stimuli were referentially fully specified by concrete contexts. The results will lead us to conclude that iconic words do have special sound-symbolic semantics and that it is crucially based on their referential specificity as well as their morphophonological markedness.

2.â•… Japanese phonosemantics As extensively described by Hamano (1998), Japanese iconic words show more or less systematic phonosemantic correspondences with respect to their componential consonants and vowels. Since Sapir’s (1929) mal vs. mil experiment, it has been reported for many languages that the vowel /a/ tends to be associated with large referents, and the vowel /i/ with small referents (see also S. Newman 1933; Brown 1958; Ultan 1978). The most widely accepted account of the prevalence of vocalic magnitude symbolism seems to be an articulatory one, which focuses on the iconic relationship between the size of oral cavity in the articulation of these vowels and that of referents. However, acoustic or sensory-motor explanations have also been proposed (see Ohala 1994; Oda 2000; Masuda 2005). Moreover, some studies point out the tendency that /a/ is linked with dark, soft, and dull referents, whereas /i/ is linked with bright, hard, and sharp ones (Bentley and Varon 1933; Koriat and Levy 1977, among others). This widespread vocalic symbolism seems to be basically shared by Japanese iconic words. For example, pakuQ^ depicts a wide-open aperture like a mouth, whereas





Kimi Akita

pikuQ^ represents a small twitch of a body part. However, as suggested by the fact that these two words do not form a semantic minimal pair, (Modern) Japanese only has implicit vocalic symbolism (see Fordyce 1983; Childs 1994; and Dingemanse, 2011 for languages with explicit vocalic symbolism). Minimal semantic distinctions are instead mainly made by the voicedness of initial consonants (C1) (see Hamano 1998: 131–132; Haryu and Zhao 2007; Akita, to appear). Overall, iconic words with a voiced C1 depict large, strong, dark, and/or unpleasant referents, and those with a voiceless C1 small, weak, bright, and/or pleasant referents. For instance, za^razara and sa^rasara are minimally different from each other with respect to C1 voicedness. The former refers to a rough, coarse surface (e.g. of sandpaper), whereas the latter refers to a dry and smooth surface (e.g. of sand). The table below (Table 1) summarizes the general contrastive sound symbolism of consonants and vowels of Japanese iconic words, which will serve as a reference for the experimental observations in the present study. No description is available for the hardness symbolism of C1. Note that consonantal, but not vocalic, symbolism is said to be subject to positional effects: namely, sound-symbolic values are different between C1 and C2 (Hamano 1998). Table 1.╇ Sound symbolism of Japanese iconic words C1 Voiced Magnitude Brightness Hardness

large dark —

> <

V Voiceless

/a/

small bright —

large dark soft

/i/ > < <

small bright hard

What is especially worth paying attention to in the discussion of Japanese phonosemantics is the idea that sound symbolism exists explicitly or implicitly in the regular (‘non-iconic’) vocabulary as well. First, Hamano (1998:â•›190) does not fail to note that there are a few nouns with a voiced C1 that were deliberately made from their voiceless counterparts with the aim of sound-symbolically expressing the unpleasantness of their referents. Below follow some examples: (3) gara ‘chicken bones for stock’ (inedible part of chicken) < kara ‘husk, shell’ dama ‘lumps in cream sauce’ < tama ‘ball’ dori ‘inedible internal organs of chicken’ < tori ‘chicken, bird’ zama ‘sloppy appearance’ < sama Â�‘appearance’ (adapted from Hamano 1998:â•›190) Going one step further, some researchers suggest the ubiquity of covert sound symbolism in the language. Makino and Tsutsui (1986: 55–56) and Makino (2007),



Toward a phonosemantic definition of iconic words

who take an extreme standpoint, point out the existence of a phonosemantic Â�motivation in several pairs of function words. According to them, there is some stylistic difference between each pair in near synonymy in (4). The words in the left column (i.e. those with an alveolar nasal C1) have formal or polite tones (“more personal, subjective and speaker-oriented” tones in Makino and Tsutsui’s terminology), whereas the words in the right column (i.e. those with a velar plosive C1) have informal tones. (4)

-no -no -node -noni

vs. vs. vs. vs.

-ka (question marker) -koto (nominalizer) -kara ‘because’ -ke(re)do/-ga ‘although’ (adapted from Makino and Tsutsui 1986:â•›55–56)

Furthermore, Makino and Tsutsui posit some phonosemantic correspondences in adjectives. For instance, they remark that adjective roots ending in /si/ are associated with “human emotive psychological states” (e.g. kanasi- ‘sad’, sabisi- ‘lonely’, uresi‘happy’) (1986: 56). Likewise, the Japanese size adjectives ooki- ‘large’ and tiisa- ‘small’ are often regarded as instances of vocalic magnitude symbolism in the regular lexicon. Similar to these observations are the positive results of some earlier crosslinguistic phonosemantic investigations of actual lexical items, such as diminutives (Ultan 1978) and deictic/distal words (Tanz 1971; Woodworth 1991; Traunmüller 1996). Thus, it is suggested that words tend to be invented to sound-symbolically fit their referential meanings. In other words, sound symbolism is not a phenomenon unique to a particular word class (i.e. iconic words). This conclusion, however, can be a challenge to the above-mentioned assumption that iconic words are phonosemantically definable. In order to solve this paradox, we have to show that the sound symbolism of iconic words is clearer or more consistent than that of regular words. I will therefore explicate two conditions that qualify iconic words as phonosemantically special. The next section will summarize the morphophonological properties of iconic words, which constitute the first condition. A pair of experiments in Sections 4 and 5 will reveal the marked referentiality of iconic words as the second condition.

3.â•… Morphophonological condition of iconic words As discussed in Akita (2008;â•›2009: Chapter 4), Japanese iconic words have characteristic morphophonological properties, which seem to make a primary delimitation of the category. Specifically, almost all iconic words satisfy one of fifteen morphophonological templates, which are schematically represented with actual examples in (5). (For the sake of simplicity, the subscription for bimoraic roots of iconic words will be omitted in the rest of this paper.)





Kimi Akita

(5) Morphophonological templates for Japanese iconic words:3 a. Monomoraic (CV) root-based: CVQ(^) (e.g. niQ^ ‘grinning’, soQ ‘gentle’) CV(^)N(^) (e.g. do^N ‘bang’, tyoN^ ‘flipping’) CViQ^ (e.g. guiQ^ ‘jerking’, poiQ^ ‘tossing’) CV(^)V(^) (e.g. bu^u ‘oink’, pii^ ‘beep’) CV^V-CVV (e.g. bu^ubuu ‘complaining’, za^azaa ‘raining hard’) CVV-CVV (e.g. gyuugyuu ‘jam-packed’, noonoo ‘carefree’) CV^N-CVN (e.g. ku^NkuN ‘sniff-sniff ’, pyo^NpyoN ‘hopping’) CVN-CVN (e.g. paNpaN ‘fully inflated’, ruNruN ‘cheerful’) CV^i-CVi (e.g. su^isui ‘swimming/flying lightly’, tyo^ityoi ‘frequent’) b. Bimoraic (C1V1C2V2) root-based: C1V1^C2V2-C1V1C2V2 (e.g. me^ramera ‘blazing up’, ni^yaniya Â�‘smirking’) C1V1C2V2-C1V1C2V2 (e.g. berobero ‘dead drunk’, pasapasa ‘dried out’) C1V1C2V2Q^ (e.g. kuruQ^ ‘turning’, zokuQ^ ‘thrilled’) C1V1C2V2(^)N(^) (e.g. bata^N ‘slam’, koroN^ ‘rolling’) C1V1C2V2^ri (e.g. hyoro^ri ‘lanky’, pota^ri ‘dripping’) C1V1CC2V2^ri (e.g. doQki^ri ‘startled’, huNwa^ri ‘fluffy’) Akita reports that 23.12% (382) and 76.33% (1,261) of the 1,652 entries (supplemented) in Kakehi et al.’s (1996) dictionary of iconic words fit or are related to one of these mono- and bimoraic root-based templates, respectively. For example, niQ^, me^ramera, and huNwa^ri satisfy CVQ^, CV^CV-CVCV, and CVCCV^ri, respectively. Only nine (conventional) iconic words (.54%), such as kokekoQko^o ‘cocka-doodle-doo’, are judged as not related to any productive template. Moreover, it is argued that these templates are shared by a limited number of ‘semi-iconic’ words (e.g. siwasiwa ‘wrinkled’ (< siwa ‘wrinkle’), siNmi^ri ‘pensive’ (< sim- ‘soak’)), which derive from regular verbs, nouns, or adjectives. Importantly, the same study also reveals the psychological reality of the great coverage of the templates. It is reported that native Japanese speakers’ ratings of ‘mimeticity’ (i.e. the degree to which a word is iconic) of existent and novel words were crucially dependent on their satisfaction of the templates. Therefore, although the morphophonological condition is neither necessary nor sufficient in a strict sense, it plays an unarguably significant part in the definition of iconic words in the

.â•… Initial consonants are absent in some iconic words (e.g. o^NoN ‘weeping’, i^raira Â�‘irritated’). Based on the fact that Japanese iconic words as such are categorially unspecified – realized as (part of) an adverb, verb, adjective, or noun root – their approximate meanings are given in a participial (or onomatopoeic bare) form.



Toward a phonosemantic definition of iconic words

Japanese lexicon (or it is a strong ‘prototype condition’). Nevertheless, it will turn out below that this condition alone cannot totally define the word class.

4.╅ E  xperiment 1: The morphophonological condition as a non-sufficient condition In this section, I will show the limits of the morphophonological condition by �contrasting the basic sound-symbolic values of consonants and vowels of novel words satisfying and not satisfying the templates in (5) (see Akita (to appear) for a preliminary investigation). If the morphophonological condition is not the only condition for Japanese iconic words, its mere satisfaction should not directly yield special phonosemantics in question.

4.1â•… Method and prediction Twelve novel bimoraic roots, which exist in neither the iconic nor the regular lexicon of Japanese, were put in five existent morphophonological templates from (5b) (‘A templates’) and five novel templates (‘B templates’), which have no derivational relation to those in (5b). The CVCVQ^ template was not included because its geminate part /Q/ is difficult to perceive without a subsequent consonant (see note 2). The novel roots were created based on two factors: the voicedness of C1 (voiced: /g/, /z/, /b/ vs. voiceless: /k/, /s/, /p/) and the quality of V1 (/a/ vs. /i/).4 The second mora was fixed to /nu/. Table 2 lists all stimuli used. Table 2.╇ Stimulus words for Experiment 1 A. Existent templates for iconic words CVCVN^

CVCV-CVCV

CV^CV-CVCV

CVCV^ri

CVCCV^ri

C1 = voiced ganuN^ ginuN^ zanuN^ zinuN^ banuN^ binuN^

ganuganu ginuginu zanuzanu zinuzinu banubanu binubinu

ga^nuganu gi^nuginu za^nuzanu zi^nuzinu ba^nubanu bi^nubinu

ganu^ri ginu^ri zanu^ri zinu^ri banu^ri binu^ri

gaNnu^ri giNnu^ri zaNnu^ri ziNnu^ri baNnu^ri biNnu^ri (Continued)

.â•… In Japanese phonology, /a/ and /i/ stand in contrast with respect to height (/a/ = low vs. /i/ = high) and backness (/a/ = back vs. /i/ = front).



 Kimi Akita

Table 2.╇ Stimulus words for Experiment 1 (Continued) A. Existent templates for iconic words CVCVN^

CVCV-CVCV

CV^CV-CVCV

CVCV^ri

CVCCV^ri

C1 = voiceless kanuN^ kinuN^ sanuN^ sinuN^ panuN^ pinuN^

kanukanu kinukinu sanusanu sinusinu panupanu pinupinu

ka^nukanu ki^nukinu sa^nusanu si^nusinu pa^nupanu pi^nupinu

kanu^ri kinu^ri sanu^ri sinu^ri panu^ri pinu^ri

kaNnu^ri kiNnu^ri saNnu^ri siNnu^ri paNnu^ri piNnu^ri

B. Novel templates C1V1V1C2V2V2

C1V1^V1C2V2

C1V1^V1C2V2N

C1V1C2V2^NC1V1

C1V1^C2V2C2V2

C1 = voiced gaanuu giinuu zaanuu ziinuu baanuu biinuu

ga^anu gi^inu za^anu zi^inu ba^anu bi^inu

ga^anuN gi^inuN za^anuN zi^inuN ba^anuN bi^inuN

ganu^Nga ginu^Ngi zanu^Nza zinu^Nzi banu^Nba binu^Nbi

ga^nunu gi^nunu za^nunu zi^nunu ba^nunu bi^nunu

C1 = voiceless kaanuu kiinuu saanuu siinuu paanuu piinuu

ka^anu ki^inu sa^anu si^inu pa^anu pi^inu

ka^anuN ki^inuN sa^anuN si^inuN pa^anuN pi^inuN

kanu^Nka kinu^Nki sanu^Nsa sinu^Nsi panu^Npa pinu^Npi

ka^nunu ki^nunu sa^nunu si^nunu pa^nunu pi^nunu

For the confirmation of the importance of the morphophonological condition introduced in Section 3, twenty native speakers of Japanese were asked to judge the mimeticity of all the stimuli on a seven-point scale from 0 ‘not mimetic at all’ to 6 ‘very mimetic’. As a result, the words with an A template (e.g. ganuN^) were judged as more mimetic than those with a B template (e.g. gaanuu) (M’s = .61 and .35, respectively, recalculated to range from 0 to 1). The striking effect of the template condition received statistical support (F (1, 2396) = 741.87, p < .001, ɳ2 = .24).5 In the main experiment, native speakers of Japanese were asked to evaluate the randomly presented novel words recorded in a female voice in terms of the three

.╅ An eta square score, which ranges from 0 to 1, is an index of the effect size of a factor.



Toward a phonosemantic definition of iconic words 

best-explored semantic dimensions: namely, magnitude (11 females, 9 males; M age = 28.45), brightness (11 females, 6 males; M age = 28.59), and hardness (9 females, 11 males; M age = 28.15).6 The referent objects or actions (e.g. manner of walking) of the words were freely imagined. The evaluation scales ranged from 1 ‘very small/dark/soft’ to 7 ‘very large/bright/hard’. All three sub-experiments were preceded by a training phase with ten novel words. Our prediction was consistent across the three semantic dimensions. If the morphophonological condition alone cannot define iconic words, the A words (e.g. ganuN^) and the B words (e.g. gaanuu), which only differ in their morphophonology, should make no notable phonosemantic difference.

4.2â•… Results and discussion The prediction was borne out. In all three sub-experiments, it was not true that the A words yielded greater sound-symbolic contrasts than the B words. First of all, the present experiment replicated and supplemented the Â�phonosemantic correspondences introduced in Table 1. For all three dimensions, the voicedness of C1 (e.g. ganuN^ vs. kanuN^) yielded greater sound-symbolic contrasts than the Â�quality of V1 (e.g. ganuN^ vs. ginuN^). This result is consistent with the aforementioned fact that voicedness is the main feature in Japanese sound symbolism. Concretely, as shown in Table 3, association with size was evident in both consonants (voiced > voiceless; F (1, 299) = 288.11, p < .001, ɳ2 = .17) and vowels (/a/ > /i/; F (1, 299) = 97.22, p < .001, ɳ2 = .03). In the rest of this paper, results for each stimulus group will be presented in recalculated mean scores (between –1 and 1). Table 3.╇ The results of Experiment 1 (magnitude) Template C1

V1

Existent (A)

Novel (B)

Voiced

/a/ /i/ /a/ /i/

.19 .003 –.24 –.38

.23 .05 –.15 –.30

Voiceless

.â•… The sound pressure level of all stimuli with /i/ at V1, which were expected to be rated as smaller, was raised by 6 dB. If words with V1 /a/ are judged as larger even under this condition – as the case will turn out in Section 4.2 – it can be confirmed that magnitude ratings were conducted based on sound symbolism rather than the volume of the recording (Haryu and Zhao 2007).

 Kimi Akita

Table 4 summarizes the results for brightness judgments. While voicedness again yielded striking rating contrasts (voiced < voiceless) (F (1, 254) = 354.06, p < .001, ɳ2 = .15), the vocalic feature only yielded moderate contrasts (/a/ < /i/ (F (1, 254) = 5.95, p < .05, ɳ2 = .002). Table 4.╇ The results of Experiment 1 (brightness) Template C1

V1

Existent (A)

Novel (B)

Voiced

/a/ /i/ /a/ /i/

–.26 –.25 .14 .11

–.25 –.16 .07 .15

Voiceless

Table 5 shows the mean hardness scores obtained. Voiced consonants (e.g. ganuN^) were associated with harder objects than voiceless ones (e.g. kanuN^) (F (1, 299) = 134.02, p < .001, ɳ2 = .09). Meanwhile, vowel quality did not present such a sound-symbolic contrast (F (1, 299) = .00, p = 1.00, ɳ2 = .00). These results not only make up for the blanks in Table 1, but also suggest the weakness of vocalic symbolism in this semantic dimension.7 Table 5.╇ The results of Experiment 1 (hardness) Template C1

V1

Existent (A)

Novel (B)

Voiced

/a/

.19

.21

Voiceless

/i/ /a/ /i/

.19 –.10 –.13

.22 –.07 –.05

.╅ The results here also suggest an interesting challenge concerning the relationships among sound-symbolic dimensions. According to Table 1, in magnitude symbolism and brightness symbolism, a voiced C1 behaves in parallel with /a/, and a voiceless C1 behaves in parallel with /i/. On the other hand, in hardness symbolism, one has to give up the parallels in order to retain both Table 1 (i.e. /a/ < /i/) and the present result for C1 symbolism (i.e. voiced > voiceless). This likely cross-dimensional mismatch suggests the need for investigation into the synesthetic difference between consonantal and vocalic symbolism.



Toward a phonosemantic definition of iconic words 

Next, more importantly, the existent vs. novel distinction in morphophonology (e.g. ganuN^ vs. gaanuu) did make a difference in magnitude ratings (F (1, 299) = 10.61, p < .01, ɳ2 = .005) and hardness ratings (F (1, 299) = 8.21, p < .01, ɳ2 = .002), but did not in brightness ratings (F (1, 254) = .60, p = .44, ɳ2 = .0003). However, these results do not directly support the idea that the morphophonological condition alone can be enough to guarantee privileged phonosemantics. Specifically, as can be seen in Tables 3 and 5, the rating differences between A and B were common to all A – B pairs (i.e. all rows of the tables): the B templates shifted in the positive direction not only the largeness/hardness ratings but also the smallness/softness ratings. This means that the novel templates (e.g. C1V1V1C2V2V2) themselves evoked larger and harder images than the existent templates of iconic words (e.g. CVCVN^). In statistical terms, this fact can be stated as the absence of between-factor interactions. Concerning the magnitude ratings, the morphophonological template condition yielded no significant interaction with the consonantal condition (F (1, 299) = 3.03, p = .08) or the vocalic condition (F (1, 299) = .001, p = .98). Similarly, concerning the hardness ratings, it showed a significant interaction neither with the consonantal condition (F (1, 299) = 1.42, p = .23) nor with the vocalic condition (F (1, 299) = .92, p = .34). Thus, we can conclude that the morphophonological condition is not sufficient for distinct phonosemantics, which is expected for iconic words.8

5.â•… Experiment 2: Referential condition of iconic words This section argues that, in the definition of iconic words in Japanese and perhaps in other languages as well, we must acknowledge a referential (or lexical-semantic) condition in addition to the self-evident morphophonological condition. This point will be made clear based on a phonosemantic experiment using novel words with fully specified referential contexts, which were critically absent in Experiment 1.

5.1╅ Method and prediction This experiment was a paper-based one, minimally designed based on the above �findings. Twenty native Japanese speakers (2 females, 18 males; M age = 19.65) were

.╅ There can be another interpretation for the current results. As Bentley and Varon (1933) and Brown (1958) remark, people tend to depend strongly on sound symbolism in understanding novel words without referential specification. If this was also true in our experiment, the absence of phonosemantic contrast between the A and B words might come from their full phonosemantic interpretation. Importantly, this alternative account is based on the referentiality of words, which I will discuss as a key notion in sound symbolism in Experiment 2.

 Kimi Akita

asked to make relative evaluations for each of thirty sets of four sentences in terms of the magnitude of their specific referents (e.g. the length of walking strides). A fivepoint scale from 1 ‘very small’ to 5 ‘very large’ was employed for the magnitude ratings for the 120 sentences. Four sentences in each set shared a sentence frame (e.g. ‘Someone [bent] an iron board’), only differing from one another with respect to their predicate forms. Two of them were predicated by a complex verb made up of an existent or novel word with an existent template for iconic words (i.e. A template) and the semantically skeletal dummy verb su- ‘do’ (e.g. zawazawa su- ‘rustle’, tekoteko su- ‘walk lightly’ (intended)), which is a productive iconic- verb form. The other two sentences were predicated by a verb consisting of an existent or novel root and a verbal suffix (e.g. zawa-tuk- ‘rustle’, teko-r- ‘walk lightly’ (intended)), which is a semi-productive verb form with a reduced iconic tone (for these verb forms, see Akita 2009: Chapter 6). These two types of predicates shared the sixty bimoraic roots listed with their ‘intended’ verbal meanings in Table 6, whose first row designates the morphophonological templates assigned to them. Each root with a voiced C1 (e.g. deko, zawa) was paired with its voiceless counterpart (e.g. teko, sawa). Novel roots were basically created on the basis of Hamano’s (1998) fine-grained descriptions of the phonosemantic system of Japanese. Table 6.╇ Stimulus roots for Experiment 2 CV^CV-CVCV CVCVQ^ Existent 1. zawa/sawa (rustle) 2. gira/kira (glitter) 3. doku/toku (glug) Novel

16. deko/teko (walk lightly) 17. gasyu/kasyu (crumple) 18. giru/kiru (twinkle)

CVCV(^)N(^) CVCV^ri

CVCCV^ri

4. zara/sara (feel rough) 5. boko/poko (beat) 6. doro/toro (drip)

7. bati/pati (slap) 8. guru/kuru (spin) 9. gata/kata (rattle)

10. bura/pura (ramble) 11. bata/pata (slam down) 12. bero/pero (lick)

13. gati/kati (be sturdy) 14. zito/sito (feel moist) 15. zaku/saku (chop)

19. gusi/kusi (stub) 20. zeku/seku (throb) 21. deru/teru (slurp)

22. guku/kuku (gulp) 23. deko/teko (bang) 24. bite/pite (shut)

25. beke/peke (break) 26. batyo/patyo (splash) 27. guro/kuro (go around)

28. buko/puko (expand) 29. buke/puke (swell) 30. geko/keko (be dejected)

The order of the thirty sets was randomized, but four sentences in each set were given in fixed order – voiced, template-satisfying (e.g. za^wazawa su-) > voiceless, template-satisfying (e.g. sa^wasawa su-) > voiced, template-free (e.g. zawa-tuk-) > voiceless, template-free (e.g. sawa-tuk-) – to facilitate the intended comparisons. Our prediction was opposite to the one we made for Experiment 1. Assuming that our sentence frames successfully assigned specific referential meanings to the Â�predicates,



Toward a phonosemantic definition of iconic words 

the voicedness-based rating contrasts (i.e. voiced > voiceless) should be sharper between the template-satisfying stimuli (e.g. za^wazawa su- vs. sa^wasawa  su-) than between the template-free ones (e.g. zawa-tuk- vs. sawa-tuk-).

5.2â•… Results and discussion The prediction was again borne out. As shown by the recalculated mean ratings in Table  7, the expected consonantal magnitude symbolism was obtained in both tÂ�emplate-satisfying and template-free sentences (F (1, 299) = 161.51, p < .001, ɳ2 = .03). Table 7.╇ The results of Experiment 2 Template C1

Root

Voiced

Existent Novel Existent Novel

Voiceless

Satisfying

Free

.65 .50 .43 .19

–.36 –.42 –.50 –.56

What is crucial for our discussion is that the template-satisfying stimuli (e.g. za^wazawa/sa^wasawa su-) were judged as by far larger than the template-free ones (e.g. zawa/sawa-tuk-) (F (1, 299) = 1769.14, p < .001, ɳ2 = .49). Moreover, a significant interaction was obtained between the consonantal and morphophonological Â�conditions (F (1, 299) = 26.99, p < .001). This means that the Â�phonosemantic contrasts between the voiced and voiceless consonants were greater in the Â�template-satisfying (e.g. za^wazawa su- vs. sa^wasawa su-) than in the Â�template-free predicates (e.g. zawa-tuk- vs. sawa-tuk-). Paired with Experiment 1, the present result means that Japanese iconic words must be defined both formally (i.e. Â�morphophonologically) and functionally (i.e. referentially). This conclusion is reinforced by the significant contribution of the third factor (i.e. the novelty of roots) (F (1, 299) = 48.23, p < .001, ɳ2 = .01) and its near-significant interaction with the other two (i.e. voicedness and morphophonological templates) (F (1, 299) = 3.40, p = .07). That is, the voicedness-based magnitude contrasts, which were sharper in the template-satisfying predicates, were still clearer in the predicates with an existent root (e.g. za^wazawa su- vs. sa^wasawa su-) than in those with a novel root (e.g. de^kodeko su- vs. te^koteko su-). This trend can be interpreted in favor of the proposed referential condition of iconic words. It is likely that the familiarity of roots helped the subjects access the intended referent scenes based on their Â�lexical knowledge. Consequently, it can be emphasized that referential specificity plays an Â�important role in the occurrence of special phonosemantic effects, which, in turn,

 Kimi Akita

together with peculiar morphophonology, give rise to the perceived categorial salience of iconic words. Our findings are further compatible with the exceptional semantics of some highly fixed semi-iconic adverbs. As is often true for languages abundant in iconic words, Japanese has a set of words that can be located around the boundary of the iconic word category (Tamori and Schourup 1999: 68–69; Childs 2001). Of particular relevance to the present discussion are semi-regular adverbs like (6).

(6) do^NdoN ‘steadily and rapidly’, do^sidosi ‘unreservedly’, meQki^ri ‘remark ably’, siQka^ri ‘tightly, surely’, suQka^ri ‘completely’, teQki^ri ‘(misunderstand) completely’, tyo^kutyoku ‘from time to time’

Despite the fact that these words satisfy the productive templates for iconic words (i.e. CV^N-CVN, CV^CV-CVCV, CVCCV^ri), they sound less iconic and are often conceived as regular adverbs (see Akita 2008; 2009: Chapter 2) for an experimental substantiation of this intuition). From the viewpoint of our referential condition, the ‘de-iconization’ in these examples is attributable to the abstraction of their referents. As can be known from their translations, these adverbs represent an abstract manner, degree, or frequency, which is obviously less specific than what the majority of iconic manner words represent, such as to^botobo ‘plodding’, niQko^ri ‘smiling’, and the examples given in earlier sections. Thus, we can now clearly point out a natural correlation between the specificity of referentiality and the degree of perceived iconicity of words.

6.â•… Conclusion This paper has empirically established the (phono)semantic definition of iconic words in Japanese, which have been treated as ‘sound-symbolic words’ on intuitive and impressionistic grounds. A pair of phonosemantic experiments clarified the existence and origin of their special (or ‘vivid’) sound-symbolic properties. It has been claimed that both of the lexical peculiarities (i.e. morphophonological and referential ones) of iconic words must be taken into account in their phonosemantic definition. It does not seem impractical to think that the two peculiar facets of iconic words are iconically linked with each other based on the natural correspondence between marked form and marked meaning (Klamer 2002). With its generality, the present fundamental improvement in the understanding of the semantics of iconic words seems to be extensible to other languages. Â�Furthermore, our attention to lexical referentiality – which is the basis for a semantic investigation, but which tends to be slighted in the study of ‘imagistic’ values of consonants and Â�vowels – will allow us to invite these apparently peripheral vocabulary items to general semantic theories.



Toward a phonosemantic definition of iconic words 

References Akita, K. 2008. Onsyootyoogo-no ‘hantyuuka-mondai’-e-no hitotu-no kotae: Tamori & Schourup (1999)-e-no ripurai (Defining the mimetic category in the Japanese lexicon: A reply to Tamori and Schourup (1999)). In Proceedings of the Eighth Annual Meeting of the Japanese Cognitive Linguistics Association, 428–438. Tokyo: Japanese Cognitive Linguistics Association. Akita, K. 2009. A Grammar of Sound-Symbolic Words in Japanese: Theoretical Approaches to Iconic and Lexical Properties of Mimetics. Ph.D. dissertation, Kobe University. 〈http:// www.lib.kobe-u.ac.jp/infolib/meta_pub/G0000003gakui_D1004724〉. Akita, K. To appear. Phonosemantic evidence for the mimetic stratum in the Japanese Â�lexicon. In Proceedings of the Thirty-Fourth Annual Meeting of the Berkeley Linguistics Society. Â�Berkeley, CA: Berkeley Linguistics Society. Bartens, A. 2000. Ideophones and Sound Symbolism in Atlantic Creoles [Suomalaisen Â�Tiedeakatemian Toimituksia. Sarja Humaniora, Nide 304]. Helsinki: Finnish Academy of Sciences and Letters. Bentley, M. & Varon, E.J. 1933. An accessory study of ‘phonetic symbolism’. American Journal of Psychology 45: 76–86. Bodomo, A.B. 2006. The structure of ideophones in African and Asian languages: The case of Dagaare and Cantonese. In Selected Proceedings of the 35th Annual Conference on Â�African Linguistics: African Languages and Linguistics in Broad Perspectives, J. Mugane, J.P. Â�Hutchinson & D.A. Worman (eds), 203–213. Somerville, MA: Cascadilla Proceedings Project. Brown, R.W. 1958. Words and Things: An Introduction to Language. New York, NY: The Free Press. Childs, G.T. 1994. African ideophones. In Sound Symbolism, L. Hinton, J. Nichols & J.J. Ohala (eds), 178–204. Cambridge: CUP. Childs, G.T. 2001. Research on ideophones, whither hence? The need for a social theory of Â�ideophones. In Ideophones [Typological Studies in Language 44], F.K.E. Voeltz & C. KilianHatz (eds), 63–74. Amsterdam: John Benjamins. Dingemanse, M. 2011. The Meaning and Use of Ideophones in Siwu. Ph.D. dissertation. Nijmegen: Max Planck Institute for Psycholinguistics/Radboud Universiteit. Doke, C.M. 1935. Bantu Linguistic Terminology. London: Longmans, Green, and Co. Fordyce, J.F. 1983. The ideophone as a phonosemantic class: The case of Yoruba. In Current Approaches to African Linguistics, Vol. 1, I.R. Dihoff (ed.), 263–278. Dordrecht: Foris. Fortune, G. 1962. Ideophones in Shona: An Inaugural Lecture Given in the University College of Rhodesia and Nyasaland on 28 April 1961. London: OUP. Haiman, J. 1980. The iconicity of grammar: Isomorphism and motivation. Language 56: 515–540. Hamano, S. 1998. The Sound-Symbolic System of Japanese [Studies in Japanese Linguistics 10]. Tokyo: Kurosio. Haryu, E. & Zhao, L. 2007. Yuuseion-to museion-o daisyoo-ni taioo-zukeru kankaku-no kigen: Giongo-rikai-no nit-tyuu-hikaku (Understanding the symbolic values of Â�Japanese Â�onomatopoeia: Comparison of Japanese and Chinese speakers). Shinrigaku kenkyu/Â� Japanese Psychological Research 78: 424–432. Hinton, L., Nichols, J. & Ohala, J.J. 1994. Introduction: Sound-symbolic processes. In Sound Symbolism, L. Hinton, J. Nichols & J.J. Ohala (eds), 1–12. Cambridge: CUP. Johnson, M.R. 1976. Toward a definition of the ideophone in Bantu. The Ohio State University Working Papers in Linguistics 21: 240–253.

 Kimi Akita Kakehi, H., Tamori, I. & Schourup, L. 1996. Dictionary of Iconic Expressions in Japanese. Berlin: Mouton de Gruyter. Klamer, M. 2002. Semantically motivated lexical patterns: A study of Dutch and Kambera expressives. Language 78: 258–286. Koriat, A. & Levy, I. 1977. The symbolic implications of vowels and of their orthographic Â�representations in two natural languages. Journal of Psycholinguistic Research 6: 93–103. Makino, S. 2007. An argument for non-arbitrary relationship between sound and meaning in Â�Japanese grammar. In Japanese Linguistics: European Chapter, Y. Ikegami, V. Â�Eschbach-Szabo & A. Wlodarczyk (eds), 239–250. Tokyo: Kurosio. Makino, S. & Tsutsui, M. 1986. Nihongo kihon bunpoo ziten (A dictionary of basic Japanese grammar). Tokyo: The Japan Times. Masuda, K. 2005. The physical basis for phonological iconicity. In Insistent Images [Iconicity in Language and Literature 5], E. Tabakowska, C. Ljungberg & O. Fischer (eds), 57–71. Amsterdam: John Benjamins. Newman, P. 1968. Ideophones from a syntactic point of view. Journal of West African Languages 5: 107–117. Newman, P. 2001. Are ideophones really as weird and extra-systematic as linguists make them out to be? In Ideophones [Typological Studies in Language 44], F.K.E. Voeltz & C. Â�Kilian-Hatz (eds), 251–258. Amsterdam: John Benjamins. Newman, S.S. 1933. Further experiments in phonetic symbolism. The American Journal of Â�Psychology 45: 53–75. Oda, H. 2000. An Embodied Semantic Mechanism for Mimetic Words in Japanese. Ph.D. Â�dissertation, Indiana University. Ohala, J.J. 1994. The frequency code underlies the sound-symbolic use of voice pitch. In Sound Symbolism, L. Hinton, J. Nichols & J.J. Ohala (eds), 325–347. Cambridge: CUP. Peirce, C.S. 1932. Collected Papers of Charles Sanders Peirce, Vol. 2. Cambridge, MA: Harvard University Press. Samarin, W.J. 1971. Survey of Bantu ideophones. African Language Studies 12: 130–168. Sapir, E. 1929. A study in phonetic symbolism. Journal of Experimental Psychology 12(3): 225–239. Tamori, I. & Schourup, L. 1999. Onomatope: Keitai-to imi (Onomatopoeia: Form and meaning) [Japanese-English Contrastive Studies 6]. Tokyo: Kurosio. Tanz, C. 1971. Sound symbolism in words relating to proximity and distance. Language and Speech 14: 266–276. Traunmüller, H. 1996. Sound symbolism in deictic words. Quarterly Progress and Status Report 37: 147–150. Stockholm: KTH Royal Institute of Technology. Ultan, R. 1978. Size-sound symbolism. In Universals of Human Language, Vol. 2: Phonology, J. Greenberg (ed.), 525–567. Stanford, CA: Stanford University Press. Woodworth, N.L. 1991. Sound symbolism in proximal and distal forms. Linguistics 29: 273–299.

Iconic thinking and the contact-induced transfer of linguistic material The case of Japanese, signed Japanese, and Japan Sign Language William J. Herlofsky

Nagoya Gakuin University Stated very simply iconic thinking is the ability to recognize similarities in different phenomena. This way of thinking can often lead to imitation and borrowing when languages come into contact, two important methods that languages have available to them for forming new words and enriching their lexicons. These methods, along with many others, are available to all languages, including sign languages. The present analysis describes how lexical borrowing and word-formation processes in Japan Sign Language (JSL) interact to expand JSL’s lexicon and grammar. The first portion of the analysis illustrates how the structures of words in spoken Japanese can be borrowed into JSL (and an interlanguage, Signed Japanese [SJ]) and then how this can influence the development and use of manual affixes for the transfer of meaning and syntactic relations. Can we imagine a golden age when all the words were young and…transparent…? If such an age existed, it was one of perfect harmony: things revealed their value in words, and words captured the most salient features of things. — Anatoly Liberman

1.â•… Introduction For most of the older languages of the world, the golden age of transparency – if it ever really existed – to which Anatoly Liberman (2005:â•›8) refers in the above quotation has, unfortunately, passed. It may not be too late, however, to observe some of this transparency in a number of the younger languages of the world. For example, for most sign languages, it may still be early enough in their evolution and development to call the present age the ‘golden age’. That is, in many young sign languages, the metaphors and other forms of iconicity originally instrumental in the formation of the signs are still somewhat transparent, although the iconicity might not be so apparent on first glance. For example, the four signs below are iconic in origin, but are not understandable without some explanation.

 William J. Herlofsky

The first three signs, (1a), (1b), and (1c), all have the same meaning, and all are originally (partially) iconic, but they are formed differently. The first two are Japan Sign Language (JSL) signs, while (1c) is from American Sign Language (ASL).

1a

1b

1c

Figure 1.

How can signs that are iconic look so different? The signs in (1a) and (1b) both mean NAME, and (1a) is used mainly in the western part of Japan, while (1b) is used mainly in the eastern part. The sign in (1a) is an iconic representation of a name card on a lapel or shirt pocket. The sign in (1b) is a bit more complicated. The sign originally was signed a bit differently, with the non-dominant left hand more in front of the body with the palm facing upwards, and the thumb of the dominant right hand was then pressed down on the upturned left palm, as if giving a thumb print. This is in fact what was originally iconic about the sign; it was originally an imitation of the movements of a person providing a thumb print to ‘sign’ his/her name. Many years ago in Japan, people signed their names using a special stamp, or if they did not have a stamp (or could not write their name), a thumb print was provided. Over the years, this sign has seen some phonetic reduction, and is now expressed in a more neutral position, as shown in (1b). The ASL sign in (1c) also means NAME, and the question is how can a sign that looks so different still be iconic and have the same meaning. The reasons are both cultural and historical. In the past, in the United States as well, there were many people who sometimes needed to sign their names, but for one reason or another, were not able to read or write. This group of people often included deaf people. In the United States, instead of a stamp or thumb print, people were often asked to draw an X where the signature should be. The sign in (1c), then, is intended to represent the X of the signature. In this way, then, both (1b) and (1c) are iconic representations, from different cultures and different countries with different histories, for ways of providing some sort of substitute for a name or signature. There can also be identical signs in different sign languages that are iconic in origin but very different in meaning. Sign (2), below, is a sign in both JSL and ASL, but it has a different meaning in each of these languages.



Iconic thinking and the contact-induced transfer of linguistic material 

Figure 2.

In ASL this sign means HISTORY, and is an initialized sign with the fingerspelling handshape for H, the first letter of HISTORY, and then a movement that represents, in a rather contracted manner, something like a movement from the past (from behind) to the present (signer’s front). In JSL this same sign means something like SEEMS or MAYBE, and originated with the tracing of something like a question mark (?) in the air after a doubtful or questionable statement, to indicate the doubt about the Â�statement. This sign has also been reduced, over time, to two little bumps in the air. Sign (2), then, is an example of how an (originally) iconic sign, with a similar appearance in two Â�different sign languages, can have completely different meanings. Sign languages, then, like spoken languages, can evolve and change in many ways, often causing a loss of transparency in originally iconic signs. Some of these changes are motivated by internal (i.e. cognitive) factors, while others are caused by external pressures. One common form of external pressure that can result in language change is language contact. The following analysis will be concerned mainly with iconicity and the transfer of linguistic material caused by the external pressure brought about by such language contact. The languages involved in the contact-induced transfer to be discussed in this paper are spoken Japanese and Japan Sign Language, the language of the deaf in Japan. Sign languages like JSL naturally arise when deaf people gather together in large enough numbers to form communities, and this situation is often said to first occur when schools for the deaf are established. In Japan, the first school for the deaf was

 William J. Herlofsky

established in Kyoto, in 1868, and its creation is usually considered to be the beginning of JSL in Japan. From this beginning over 140 years ago, what is now JSL has evolved through various stages, from home signs, to pidgins, to creoles, and over the generations, has arrived at the stage where it is now a somewhat unified and stable natural human language (see Kegl 1999; Lucas 2001; and Aronoff et al. 2004, 2005). This period of 140 years, however, is a short time for a language and its morphology, lexicon, and grammar to develop and evolve, and therefore JSL, like many minority languages in contact situations, has borrowed liberally from the closest majority language, spoken Japanese. JSL, then, like most other sign languages, is a minority language surrounded by a majority spoken language. In addition, most deaf Japanese children go through an oral education system, and most are raised by hearing parents (over 90% of deaf children are born to hearing parents), so they are often, to some extent, bilingual in both JSL and spoken Japanese. In such cases, in addition to the natural sign language of the deaf, and the spoken language of the majority hearing community, an interlanguage, something along the continuum between the signed and spoken language, often develops. In Japan, as well, a manual version of Japanese, Signed Japanese (SJ), coexists with JSL, and is often used by bilingual deaf people when communicating with hearing or �hard-of-hearing people. So although JSL is an independent language, different from spoken Japanese, spoken Japanese has had considerable influence on JSL, especially in its lexicon, while SJ employs the grammar of spoken Japanese with the signs of JSL. The present �analysis will consider how both internal processes and external pressures interact to affect the morphology and word-formation processes in both SJ and JSL to expand their lexicons.

2.â•… Sign language morphology and word-formation processes For sign language morphology, Aronoff et al. (2004:â•›21) offer a concise summary of two types of complex sign language morphology from previous crosslinguistic research: The study of sign languages from all over the world has made it clear that these languages constitute a morphological type: all well studied established sign languages are reported to have the same particular types of complex morphology. Two central sign language morphological constructions are verb agreement for person and number of subject and object in a semantically defined class of verbs (Engberg-Pedersen 1993; Meir 2002; Padden 1988); and a system of polymorphemic classifier constructions that combine nominal classifier handshapes with path shapes, manners of movement, and locations. (Emmorey 2003)

These two types of sign language morphology in JSL have also been discussed briefly in two previous research papers in the Iconicity in Language and Literature series



Iconic thinking and the contact-induced transfer of linguistic material 

Â� (Herlofsky 2005, 2007), and so the focus in this paper will be on a third type of sign language morphology that Aronoff et al. consider to be “another type of morphology that is more commonly found in spoken languages, including creole languages: sequential affixation that has arisen through grammaticalization” (2004:â•›21). This type of sequential affixation, which results from language contact, as in creole Â�languages, will be considered in Section 4, after a brief discussion of other types of word-formation induced by language contact. To summarize the conclusions of other research articles contained in Brentari’s (2001) cross-linguistic investigation of word-formation processes and foreign vocabulary in sign languages, there are three ways in which sign languages can create new vocabulary items: through the language internal word-formation processes (such as compounding and derivation) that all natural languages possess, through the borrowing of vocabulary items from other spoken languages, and through the borrowing of vocabulary items from other sign languages. (There is a fourth (more artificial) process in the case of many sign languages, a process that often involves components of the other three, that will not concern us here: creation, by decree, by some (often government sponsored) language planning/management committee.) The word-formation and borrowing processes noted in Brentari (2001) can be further elaborated, as discussed for spoken languages in Heine and Kuteva (2005:â•›2), into four types, in which what is transferred can be summarized as: a. Form, that is, sounds or combinations of sounds b. Meanings (including grammatical meanings or functions) or combinations of meanings c. Form-meaning units or combinations of form-meaning units d. Syntactic relations, that is, the order of meaningful elements It is only necessary to change the word ‘sounds’ in (a) to ‘signs’, or perhaps ‘manual/ non-manual signs’, to have a list of transfer types that is applicable to the study of sign languages. And although, as Heine and Kuteva (2005:â•›5) state, contact-induced change can be “a complex process that not infrequently extends over centuries, or even Â�millennia”, in some sign-language-contact situations, it can occur much more quickly, where the transfer or replication of linguistic material “starts out with gradually Â�changing use patterns, leading from minor to major patterns; at the same time, these patterns increasingly acquire properties of distinct categories, and eventually they may turn into conventionalized grammatical categories” (ibid.). This process is now Â�sometimes occurring rather rapidly in many sign languages throughout the world. For sign languages, which, in addition to being relatively young languages, have no writing systems, it is notoriously difficult to determine the etymology of their signs, let alone the development and change of signs and their use patterns. It is the purpose of the present analysis, however, to attempt to provide illustrations of some examples of

 William J. Herlofsky

the types of contact-induced linguistic transfer that Heine and Kuteva identify above. The languages to be discussed in the analysis that follows are spoken Japanese, JSL, and the interlanguage SJ, and this article will illustrate how Japanese has influenced both SJ and JSL to such an extent that a considerable amount of linguistic material has already been transferred. The basic transfer process begins with the ability to recognize similarities between different structures in different languages, and continues through changes in use Â�patterns that are concisely described in Heine and Kuteva as occurring when speakers (or signers) in a replica language R create a new use pattern or category in language R on the model of another language (M), where the outcome of the process is not an exact copy of what exists in M but rather a new structure that is shaped, first, by what is available in R, second, by universal constraints on conceptualization, third, by what speakers of R conceive as being pragmatically most appropriate in the situation in which language takes place, and fourth, by the length and intensity of contact and – accordingly – by the relative degree to which replication is grammaticalized.  (2005:â•›7)

In this paper, I refer to this ability to recognize similarities in different structures, and then modeling the R (replica) language on the M language, as being related to iconic thinking (see also Fischer 2007; Deacon 1997; and the final section of this paper), which I claim is a key ingredient to the contact-induced transfer of linguistic material. This means that the group using the R language recognizes that there is something in language M that fills a slot (either lexical form or grammatical pattern) that is not filled, or not filled adequately, in language R, and thus, over time, through extended use patterns, borrows this material for their own use in R The following section will provide illustrations of the four types of transfer listed above with data from SJ and JSL, and in Section 4, a more detailed analysis of the process described in the quotation above will be provided for various examples of manual affixes.

3.╅ Transfer of linguistic material 3.1╅ Form The transfer of form without meaning in JSL is most simply illustrated by the handshapes of the JSL manual syllabary. Sign languages have fingerspelling systems or manual syllabaries for representing the sounds/letters of the majority spoken/� written languages that surround them. In the JSL manual syllabary (that manually �represents the Japanese written syllabary), all of the vowel handshapes and some of the �consonant-vowel handshapes have been borrowed from ASL fingerspelling handshapes (see Herlofsky 2004, 2008 for details), and shown below are the JSL manual



Iconic thinking and the contact-induced transfer of linguistic material 

syllabary handshapes for the RA-RI-RU-RE-RO line of the Japanese written syllabary. The RA shape (3a), on the left, is borrowed from the ASL fingerspelling for R, while RI (3b) is written in the air in imitation of the written syllabary letter RI (These written syllabary letters appear over each of their respective handshapes in 3), and the other three handshapes on the right, RU-RE-RO (3c, 3d, 3e), imitate the shapes of the written syllabary letters with handshapes (without the movement included in 3b).

3a

3b

3c

3d

3e

Figure 3.

3.2â•… Meaning One way in which meaning alone can be transferred into a language is when, instead of borrowing a new word from the M language, only the meaning, using the forms available in the R language, is transferred and conveyed through the formation of a new word. For example, when a new word was needed to convey the meaning of someone who comes to help in someone’s (usually the elderly) home, the Japanese language borrowed the English words ‘home-helper’, and created the term hoomu-herupaa in Japanese, thereby borrowing both the meaning and pronunciation (form) from Â�English. JSL, on the other hand, created its own expression out of a combination of already existing signs, for HOME (4a) and HELPER (4b), without borrowing the sounds (form) of Japanese or English. (Although mouthing the Japanese words, or even voicing them, can accompany the signs, these and similar non-manual aspects of JSL/SJ will not be considered in this analysis.) The compound has thus borrowed the meaning from English, through Japanese, but has used its own preexisting forms (manual signs) in a new loan translation-like compound, as illustrated below (Figure 4).

4a Figure 4.

4b

 William J. Herlofsky

3.3â•… Form-meaning units The borrowing of form-meaning units, as in the borrowing of words like hoomuherupaa into Japanese from English, is one of the most common forms of borrowing, and, as suggested in Brentari (2001), can come in four types in sign languages, where the ‘forms’ borrowed are either written forms, gestural/emblem forms, or signs from another sign language. One of these types of form-meaning transfer, the imitation of a written form, can be illustrated by the following two different methods of Chinese character imitation. The first sign is a handshape imitation of the form of the Chinese character for NORTH (5a), while the second example consists of drawing the character for RIVER (5b) in the air.

5a

5b

Figure 5.

Another type of form-meaning unit transfer is the borrowing of a gesture/ emblem from a spoken M language. The Japanese gestural emblems for ‘boyfriend’ and ‘girlfriend’ appear below, and were borrowed into JSL to mean MALE (6a) and FEMALE (6b).

6a

6b

Figure 6.

A third type of form-meaning unit transfer is the case of a sign borrowed from another sign language. For example, the sign for I-LOVE-YOU in ASL (Figure 7) has become an international greeting in many sign languages, including JSL.



Iconic thinking and the contact-induced transfer of linguistic material 

Figure 7.

And finally, the fourth type of form-meaning transfer involves signs that are formed with the initial sound of a word being expressed by fingerspelling, and then the semantic portion of the sign being expressed by the motion of the hand, as in the sign for EUROPE (Figure 8) below, where the handshape (form) is the fingerspelling for ‘e’, and the motion expresses metaphorically the concept (meaning) of a large expanse of land. In this way, then, the handshape (the borrowed portion) is combined with the movement (the semantic portion) to form the new JSL sign.

Figure 8.

 William J. Herlofsky

3.4â•… Syntactic relations Syntactic relations, or the order of elements in JSL, are most visibly influenced by the grammar and syntactic order of Japanese, and by extension, SJ. In the examples below, sentence (Figure 9) is SJ and follows the Japanese word order, and means YET (9a) MARRIED (9b) NOT (9c) (‘I’m not married yet’), while the second Â�sentence Â�(Figure  10) is an example of more natural JSL order with YET following the verb (‘MARRIED (10a) YET (10b)’), which carries the same negative meaning ‘I’m not married yet’), but the negation is not expressed lexically (i.e. by a separate form) but by non-manual markers. Notice the facial expression and the increased distance between the hands (representing the goal, and where the person is at present), making the sentence mean something like ‘I’m not married YET!’

9a

9b

9c

Figure 9.

10a

10b

Figure 10.

What should be especially interesting to a linguist about the above example, other than the clear illustration that SJ (with the sequence borrowed from Japanese) and JSL have different grammars, is that JSL seems to be able to express negation without an explicit negative form. It is perhaps the iconic nature of the JSL sign for YET, graphically illustrating that the one hand has not reached the other yet, that includes the negative meaning. It is this type of possible combination of meaning (YET + NOT) in one visual sign that results in what is often referred to as the ‘simultaneous’ nature of



Iconic thinking and the contact-induced transfer of linguistic material 

sign language signs, as opposed to the ‘sequential’ nature of spoken/written language words. The sequential/simultaneous contrast is even clearer in the following examples. Sentence (Figure 11) is SJ, and follows the Japanese sequential word order, and means ENTIRELY (11a), SEE (11b), NOT (11c), REASON (11d) NOT-SAY (11e), ‘(I’m) not saying (that I) don’t watch any (movies)’.

11a

11b

11c

11d

11e

Figure 11.

The JSL sentence in Figure 12, which has the same meaning, is much shorter.

12a

12b

12c

Figure 12.

The signs in Figure 12 that do not appear in Figure 11 are (12b) and (12c), and these are signs that are a bit more complicated (and therefore simultaneous) than those in (11). The signs in Figure 12 are SEE (12a), ZERO (12b), and DIFFER (12c), and might be translated directly as ‘It’s not (that I) see zero (movies)’. Again, no overt negative sign is used, but in fact two near-negatives are used to form a positive meaning. The sign DIFFER in (12c) means something like ‘NOT X’, while the ZERO sign in (12b) also has a negative meaning. In fact, according to Aronoff et al. (2004), it is common for sign languages to use a sign meaning ZERO to form negative morphemes. The following section provides a more detailed analysis of sentences like these and the transfer of syntactic relations and meaning.

 William J. Herlofsky

4.â•… The transfer of meaning and syntactic relations Although lexical borrowing can be rather sudden, in contrast, grammatical Â�borrowing and/or grammatical ‘pattern borrowing’ usually take a long period of time. For interlanguages like SJ, however, grammatical pattern ‘borrowing’ can resemble lexical Â�borrowing, and can therefore occur over a relatively short period of time, as is also the case with substratum influence. That is, as noted in the discussion of borrowed use patterns referred to by Heine and Kuteva (2005:â•›7) above, the replica language (SJ) can create (relatively quickly, since the people doing the creating are bilingual) a new ‘use pattern’ based on the model language (Japanese) by recruiting SJ (or JSL) forms to imitate the forms and use patterns of Japanese. Aronoff et al. (2005:â•›305–6) consider such morphological/grammatical borrowing to be of little interest, stating that replica languages (‘mesolects’ in their terminology, and the model languages are ‘acrolects’) “that mirror the grammar of the lexifier (or acrolect) language are clearly uninteresting for any discussion of morphological properties of young contact languages.” It is not so clear to me why this kind of borrowing or ‘mirroring’ (iconic thinking) should be of little interest. It seems to me that there is a need to record the development of any young language, but especially young sign languages that have no writing systems that can preserve a record of their changes and development. In fact, just a page after the above quote, Aronoff et al. (2005:â•›307) lament the fact that “detailed histories of grammatical development of most sign languages are unknown.” Since documentation of any changes and developments in sign languages may at some point in the future prove valuable, and also since it is always possible that even these types of imitations of the model language’s morphology/grammar might become an integral part of the fully developed sign language in the future, even these patterns that “mirror the grammar” of the spoken language, should be of interest. At least it should be interesting to see (and record) how sign languages (especially their interlanguage forms) mirror the majority spoken language, and if these methods of imitation are similar to those of the spoken-language mirroring processes. The discussion that follows will focus on this type of process. As mentioned above, this mirroring process involves what I refer to as iconic thinking, and begins with the recognition of similarities in different entities. This same process has been given different names by different researchers. For example, Lyons (1977:â•›76) defines the notion of language-related creativity as “the language user’s ability to extend the system by means of motivated, but unpredictable, principles of abstraction and comparison.” Aikhenvald (2007:â•›57–8), in a discussion of the productivity of word-formation processes, and after noting this definition by Lyons, states that creativity “in the application of a rule or a process by analogy may be indicative of its productivity.” Again, it is this utilization of extension of systems, by “motivated” principles of “abstraction”, “comparison”, and “analogy”, that I refer to as iconic thinking, and which can lead to the transfer of linguistic material.



Iconic thinking and the contact-induced transfer of linguistic material 

This type of iconic thinking will be illustrated with the following examples, which  will consist of an analysis that will first examine the transfer of meaning and  syntactic relations by examining combinations of the following four signs: SLEEPY/TO-SLEEP (13a), NOT/DON’T-HAVE (13b), DIFFICULT/CAN’T (13c), and FINISHED/-ED (13d). (The pair of meanings for the sign in 13d indicates that it can function as the free-form verb FINISHED, as well as the fact that it has been grammaticalized to function as the past tense-like bound form -ED.)

13a

13b

13c

13d

Figure 13.

It should be noted that these signs are free forms that can be used just as they are as words in sentences, or even can form one-sign sentences themselves. These signs can also be combined with other signs, as shown below. In the examples that follow, the four signs above may still be free morphemes, but might also be interpreted as stems and affixes (see Aronoff et al. 2004 for a discussion of a similar phenomenon in ASL).

14a

14b

SLEEPY/SLEEP

NOT

(‘not sleepy’, ‘don’t sleep’) Figure 14.

 William J. Herlofsky

15a

15b

SLEEPY/SLEEP

CAN’T (‘can’t sleep’)

Figure 15.

16a

16b

SLEEPY/SLEEP

DID

(‘was sleepy’, ‘slept’) Figure 16.

These are all fairly close to being direct translations of Japanese – in Japanese, the later portion of the word would be considered an affix – and therefore can be considered to be more SJ-like than JSL-like, although two-sign combinations like these are also common in JSL. A problem arises, however, when there are longer sign combinations that directly imitate the affix sequences typical of a synthetic language like Japanese. For example, in Japanese, it is common to combine the negative, potential, and past affixes into a single word. For example, while ‘I sleep’ would be ‘nemuru’, ‘I didn’t sleep’ would be ‘nemuranakatta’, where the negative (-nakat-) (17b) and past (-ta) (17c) affixes are added. In SJ, these affixes are directly reflected in the sign combinations.



Iconic thinking and the contact-induced transfer of linguistic material 

17a

17b

17c

Figure╇ 17.

This can also be seen in SJ sentences, like the one below, which is a direct translation of the Japanese sentence, YESTERDAY (‘kinoo’) (18a), HARDLY (‘nakanaka’) (18b), SLEEP (‘nemur-’) (18c), CAN’T, (‘-e-nakat-’) (18d), FINISHED/-ED (‘-ta’) (18e), or ‘I could hardly sleep yesterday.’

18a

18b

18c

18d

18e

Figure╇ 18.

There is a problem here, however, because although this SJ is understandable to deaf people who are bilingual in Japanese and SJ, it might be difficult to understand for monolingual deaf Japanese. For them, the following JSL sentence, with the same meaning and some of the same signs, would be easier to understand. That is, though the affixes of Japanese can be converted into manual affixes in SJ, and may eventually make it into JSL, this kind of borrowed manual affix combination in JSL is still somewhat marked. The sentence in Figure 19, which directly translated is something like ‘YESTERDAY SLEEP HARDLY’, differs from (18) in a number of ways. One difference is that the sign HARDLY (19c) follows the verb stem SLEEP (19b), while in the SJ (Figure 18), HARDLY (18b) precedes the verb stem SLEEP (18c). This is in fact similar to what occurs in the JSL (Figure 10), where YET (10b) follows the verb stem MARRY

 William J. Herlofsky

(10a), while in the SJ (Figure 9) YET (9a) precedes MARRY (9b), and the 11/12 pair, where ZERO follows the verb stem in the JSL (Figure 12), but precedes it in the SJ (Figure 11). This suggests that in addition to SJ verbs mirroring Japanese verb Â�morphology, in SJ, portions of JSL verb-stem following signs are used to function as verb-preceding adverbs. That is, it appears that bound forms in JSL are free form adverbials in SJ.

19a

19b

19c

Figure 19.

It should also be noted that the negative potential JSL sign CAN’T combines what would be two bound morphemes in Japanese, -e- (‘can’) and -nakat- (‘not’), and so is not really a direct manual translation of Japanese. There are separate JSL signs for CAN and NOT, but these are not used in a sentence like (Figure 18) even in SJ. In this way, then, SJ is already showing signs of a hybrid nature, not really being just a simple manual representation of Japanese, but something more complicated.

5.â•… Conclusion Although spoken Japanese and JSL are different languages, from what has been Â�discussed and illustrated in the previous sections, it should also be apparent that they are interconnected in ways that might be better described as a continuum, in which some kind of idealized version of spoken Japanese is at one extreme, and an idealized version of JSL at the other. A more extreme view of ‘real’ and ‘idealized’ languages, is provided, for example, by Joseph (2002:â•›14), who states that, “given that no two of us share the same language except in highly idealized terms, the very existence of “a language” as a supra-individual thing is a myth”. SJ is somewhere in between. Perhaps this inter-relatedness can be better understood by considering Grosjean’s description



Iconic thinking and the contact-induced transfer of linguistic material 

of what goes on within a bilingual individual according to his activation model for bilinguals: [F]irst, bilinguals have two language networks which are both independent and interconnected, and second, in the monolingual mode, one network is strongly activated while the other is weakly activated, whereas, in the bilingual mode, both networks are activated but one – that of the base language – more so than the other. (2008:â•›4)

For SJ, it would appear that the grammar portion of Japanese is activated while at the same time the lexical portion of JSL is also activated, a relationship which, again, might be better described with a continuum metaphor. The decision of what language to use, JSL or SJ, might also be related to what Kirkpatrick (2007:â•›11) describes as an identity-communication continuum, where at one end is the communicative function, where “being intelligible and getting your meaning across is the most important aspect”, while the identity end is favored when signifying identity with a group is most important. So someone wanting to identify with the hearing majority culture in Japan might use Japanese or SJ even if these languages might not communicate well. Even the lexical/functional distinction that has been discussed, the distinction between free and bound morphemes, really represents a continuum of fuzzy boundaries between these categories (see Muysken 2008 for further discussion). And last but not least, the distinctions among hearing and hard-of-hearing and deaf, is another continuum where boundaries are difficult to determine (see Brueggemann 2009). Perhaps Larsen-Freeman and Cameron’s definition of complexity theory can aid in getting a handle on the complex, interwoven phenomena we are dealing with: they define the scope of complexity theory as “the study of complex, dynamic, non- linear, self-organizing, open, emergent, sometimes chaotic, and adaptive systems” (2008:â•›4). That seems to me to also be a good description of what we are dealing with when we study sign languages and their uses. While realizing that there are still problems with trying to analyze and describe such complex processes, a few generalizations about the results of what we have called ‘iconic thinking’ can be made from the discussion in the previous sections. First, for sign languages, along with the borrowing of lexical items, there is also a kind of ‘borrowing’ of grammatical patterns that exists somewhere along the shifting continuum between spoken and signed languages. And this language-shift-like pattern-borrowing process (see Thomason & Kaufman 1988), in contrast to the grammaticalization process, can occur relatively quickly, perhaps almost as quickly as the loan-word process. And it should also be added that this process, though perhaps not technically grammaticalization as such, is important in the development of sign languages, and therefore recognizing and documenting this process is important.

 William J. Herlofsky

It has been claimed that the phenomena noted above are the result of iconic Â�thinking, and factors that motivate the process of iconic thinking and the transfer of linguistic material can be summarized as follows: 1. Recognize some kind of similarity or equivalence relation between some aspects (x) of the two languages (M and R) in contact 2. Find something in R, Ry, that can be used as Rx 3. Use Ry to function as Rx in R as Mx functions in M We have seen that free morphemes, like NOT and DIFFICULT and DID, acting as Ry’s in JSL, were recruited to function as affixes (Rx’s), -NOT, -CAN’T-, -ED, respectively, in SJ (R), and perhaps also in JSL (R), on the model of the Japanese (M) template verb-stem+potential+negative+tense (Mxs). The present analysis has thus attempted to illustrate how language contact and iconic thinking can influence word formation and even affixation, and that what is borrowed is sometimes the order of the affixes. The claim that borrowed affixation is iconically motivated seems to have been overlooked in much of the sign language research, since Sandler and Lillo-Martin state, citing data from Aronoff et al. (2005), that for sign languages “the individual words from which the base and affix arise are iconically motivated, but their combination is not” (2006:â•›506). In most of the affixal signs presented in this paper, there is both sign-internal imagic iconicity for individual signs, and sign-external diagrammatic iconicity (iconic thinking and imitation of Japanese affix order) for the more complex polymorphemic signs. Although the interwoven aspects of different (both signed and spoken) languages, and the complications of the many multilayered continuums make an analysis of these complex phenomena difficult, the iconic aspects of the lexicon and the grammar also make some aspects of the analysis more transparent. In a way, then, to some extent, it may still be the golden age for Japan Sign Language, when many of the signs are still young and transparent, revealing the salient features of things they represent in their structure. Not quite perfect harmony, but harmony nonetheless.

Note There is a less productive (and somewhat rare) manual prefix of JSL origin meaning Â�something like PARENT, that, in contrast to the affixes discussed above, is never used as a free Â�morpheme. First recall the JSL signs for MALE (6a) and FEMALE (6b) discussed in the paper. These signs can be used in combination with this native manual prefix to form the signs for FATHER and MOTHER. (This is somewhat similar to Aikhenvald’s (2003:â•›197–8) description of Â�similar Â�kinship terms in Tariana, but in Tariana, it is the word ‘parent’ that is the base, and the Â�gender distinctions are suffixes.). This manual prefix (the touching of the cheek with the Â�forefinger) is illustrated below. The touching of the cheek plus MAN means FATHER (ia), and the Â�cheek-prefix plus FEMALE (ib) means MOTHER.



Iconic thinking and the contact-induced transfer of linguistic material 

ia

ib

Acknowledgement Once again I would like to thank the Deaf group kusa-no-ne (Grassroots) for permission to use illustrations from their book Akusesu! Rou-sha-no-shuwa (Access! Deaf Sign Language), Akaishi-shoten, Tokyo, 1998.

References Aikhenvald, A. 2003. A Grammar of Tariana, From Northwest Amazonia. Cambridge: CUP. Aikhenvald, A. 2007. Typological Distinctions in Word-Formation. In Language Typology and Syntactic Descriptions. T. Shopen (ed.), 1–65. Cambridge: CUP. Aronoff, M., Meir, I., Padden, C. & Sandler, W. 2004. Morphological universals and the sign Â�language type. In Yearbook of Morphology, G. Booij & J. van Marle (eds.), 19–39. Â�Dordrecht: Kluwer. Aronoff, M., Meir, I. & Sandler, W. 2005. The paradox of sign language morphology. Language 81: 301–344. Brentari, D. (ed.). 2001. Foreign Vocabulary in Sign Languages: A Cross-Linguistic Investigation of Word Formation. Mahwah NJ: Lawrence Erlbaum Associates. Brinton, L. & Traugott, E. 2005. Lexicalization and Language Change. Cambridge: CUP. Brueggemann, B. 2009. Deaf Subjects: Between Identities and Places. New York NY: New York University Press. Deacon, T. 1997. The Symbolic Species: The Co-Evolution of Language and the Brain. New York NY: Norton. Emmorey, K. 2003. Perspectives on Classifier Constructions in Sign Languages. Mahwah NJ: Â�Lawrence Erlbaum Associates. Engberg-Pedersen, E. 1993. Space in Danish Sign Language: The Semantics and Morphosyntax of the Use of Space in a Visual Language. Hamburg: Signum. Fischer, O. 2007. Morphosyntactic Change: Functional and Formal Perspectives. Oxford: OUP. Grosjean, F. 2008. Studying Bilinguals. Oxford: OUP. Heine, B. & Kuteva, T. 2005. Language Contact and Grammatical Change. Cambridge: CUP. Herlofsky, W. 2004. A brief history of the manual alphabet fingerspelling from Europe to Japan. Nagoya Gakuin University Foreign Language Education Journal 34: 1–8.

 William J. Herlofsky Herlofsky, W. 2005. Nihonshuwa – nihongotaioushuwa – nihonngo: genngo no okeru sesshoku o tsuujite ten o rikai suru tame (Japan Sign Language – Signed Japanese – Japanese: Understanding Contact-Induced Transfer of Linguistic Material). Paper Presented at The 31st Japanese Sign Language Conference, Chiba, Japan. Herlofsky, W. 2005. Now you see it, now you don’t: Imagic diagrams in the spatial mapping of signed (JSL) discourse. In Outside-In—Inside-Out. [Iconicity in Language and Literature 4], C. Maeder, O. Fischer & W. Herlofsky (eds), 323–345. Amsterdam: John Benjamins. Herlofsky, W. 2007. Iconic thumbs, pinkies and pointers: The grammaticalization of animateentity handshapes in Japan Sign Language. In Insistent Images. [Iconicity in Language and Literature 5], E. Tabakowska, C. Ljungberg & O. Fischer (eds), 37–53. Amsterdam: John Benjamins. Herlofsky, W. 2008. Cultural identity and the production and understanding of sign language signs: Identifying, interpreting, and preserving the cultural information in iconic signs. In Identity in Text Interpretation and Everyday Life, M. Amano (ed.). 89–98. Nagoya: Nagoya University Press. Joseph, J. 2002. From Whitney to Chomsky: Essays in the History of American Linguistics, Â�Amsterdam: John Benjamins. Kegl, J., Sangha, A. & Coppola, M. 1999. Creation through contact: Sign language emergence and sign language change in Nicaragua. In Language Contact and Language Change, M. DeGraff (ed.), 179–237. Cambridge MA: The MIT Press. Kirkpatrick, A. 2007. World Englishes: Implications for International Communication and English Language Teaching. Cambridge: CUP. Larsen-Freeman, D. & Cameron, L. 2008. Complex Systems and Applied Linguistics. Oxford: OUP. Liberman, A. 2005. Word Origins… and How we Know Them: Etymology for Everyone. Oxford: OUP. Lucas, C. 2001. The Sociolinguistics of Sign Languages. Cambridge: CUP. Lyons, J. 1977. Semantics, Vol. II. Cambridge: CUP. Meir, I. 2002. A cross-modality perspective on verb agreement. Natural Language and Linguistic Theory 20: 413–450. Muysken, P. 2008. Functional Categories. Cambridge: CUP. Padden, C. 1988. Interaction of Morphology and Syntax in American Sign Language [Outstanding Dissertations in Linguistics]. New York NY: Garland Press. Sandler W. & Lillo-Martin, D. 2005. Sign Language and Linguistic Universals. Cambridge: CUP. Thomason, S. & Kaufman, T. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley CA: University of California Press.

Ezra Pound among the Mawu Ideophones and iconicity in Siwu Mark Dingemanse

Max Planck Institute for Psycholinguistics, Nijmegen The Mawu people of eastern Ghana make common use of ideophones: marked words that depict sensory imagery. Ideophones have been described as “poetry in ordinary language,” yet the shadow of Lévy-Bruhl, who assigned such words to the realm of primitivity, has loomed large over linguistics and literary theory alike. The poet Ezra Pound is a case in point: while his fascination with Chinese characters spawned the ideogrammic method, the mimicry and gestures of the “primitive languages in Africa” were never more than a mere curiosity to him. This paper imagines Pound transposed into the linguaculture of the Mawu. What would have struck him about their ways of ‘charging language’ with imagery? I juxtapose Pound’s views of the poetic image with an analysis of how different layers of iconicity in ideophones combine to depict sensory imagery. This exercise illuminates aspects of what one might call ‘the ideophonic method’.

1.â•… Introduction1 In the early essay Vorticism (1914), the poet Ezra Pound, one of the central figures of the Modernist movement, writes about his attempts to verbalize an experience he had in a station of the metro: Three years ago in Paris I got out of a ‘metro’ train at La Concorde, and saw suddenly a beautiful face, and then another and another, and then a beautiful child’s face, and then another beautiful woman, and I tried all that day to find words for what this had meant to me, and I could not find any words that seemed to me worthy, or as lovely as that sudden emotion. (Pound 1914:â•›465)

.â•… For countless sparkling ideophonic exchanges I thank the Mawu people, especially Ɔɖimɛ Kanairo, Ruben and Ella Owiafe, Rev. A.Y. Wurapa, and Stephen Addae. I am grateful to Nick Enfield, Gunter Senft, and John Haiman: Olga Fischer for helpful comments on an earlier version; and to Filiocht for introducing me to Ezra Pound. This work was funded by the Max Planck Society for the Advancement of Science. It derives from doctoral work reported in more detail in Dingemanse (2011).

 Mark Dingemanse

This search for words is typical of Pound; in fact, the unrelenting quest for words, for ways to “present […] an intellectual and emotional complex in an instance of time” (1914:â•›464), is a leading thread through all of his work from early Imagism to the latest Cantos, and can be seen as the catalyst for his innovative views on poetry and poetry writing. A useful summary of his thoughts about this process of Â�image-making is provided in the ABC of Reading. There, he describes three main methods of Â�“charging language”: melopoeia: “inducing emotional correlations by the sound and rhythm of speech”; phanopoeia: “throwing the object onto the visual imagination”; and Â�logopoeia: “using the intellectual and emotional associations of words and word groups” (Pound 1934:â•›36–7). The first two of these Pound sees as primary, but he emphasizes their Â�constant interplay in poetry and devotes a good deal of the ABC of Reading to Â�anthologizing work from a range of different traditions in which these techniques can be seen to be at play. Let us return to Vorticism. In that essay, Pound describes how he first composed a thirty-line poem about the metro experience but destroyed it because it was “of second intensity” (1914:â•›467). He wrote a poem half that length six months later; and finally he formed the following haiku-like poem: The apparition of these faces in the crowd: Petals, on a wet, black bough.

(1914:â•›467)

Of all his attempts, he explains, the final version is closest to the ideal of achieving the “most intense” form of expression. This process of successive condensation is Â�succinctly expressed in Pound’s dictum “Dichten = condensare” (1934:â•›36), and this is where ideophones come in, as words that, precisely, condense a rich experience into one word. Consider the following ideophone poem in Gbaya, composed by the Â�Cameroonian artist Dogobadomo: (1)

Ŋgalaŋ Pindoŋ Sĕlĕle Dawa iŋa só te. A leap Whoosh Silence How the monkey knows trees.

(from Noss 1989:â•›33)

The first three lines of this poem are simply three ideophones. It launches into action with ŋgalan, “the image of an animal’s agile leap from the ground to the trunk of a tree”, then shifts viewpoint with pindoŋ, “the image of the swaying whoosh of a leafy branch that dips suddenly as the weight of a fairly large creature pushes it downward” (ibid.: 34). After these successive action shots, a sudden silence descends, evoked by the ideophone sélélé. Finally we zoom out to the full scene with the exclamation “How



Ezra Pound among the Mawu 

the monkey knows trees!”, which, Noss tells us, can be seen as an Â�allusion to the Gbaya proverb “One doesn’t teach the monkey how to climb trees” (ibid.: 34).2 The way in which ideophones provide this poem with cinematic qualities Â�resonates with Pound’s notion of ‘luminous detail’. “The artist,” he writes, “seeks out the luminous detail and presents it. He does not comment.” (Pound 1973 [1912]: 22; see also Basbøll 2008). The focus is on presentation rather than commentary, on depiction rather than description. Indeed, for Dogobadomo, creator of the above poem and native speaker of Gbaya, ideophones are “shortcuts”, “words that tell lots of things” (Noss 1989:â•›34). Siwu speakers offer similar metalinguistic reflections. For instance, Ruben Owiafe, my senior consultant, explains that they “illuminate things” and “make people see how things are”. They are, in Poundian terms, charged with meaning to the utmost possible degree.

2.â•… Ideophones Ideophones are marked words that depict sensory imagery. They are found abundantly in African, Asian, and Amerindian languages; as a distinct class of words they are rare in Indo-European languages (Voeltz & Kilian-Hatz 2001). Their use has been summarized eloquently by Fortune: With them one is in a special realm of spoken art. There is a roundness, a complete shape, not so vividly conveyed by more complex constructions, more formal expressions. They attempt to be a vivid re-presentation or re-creation of an event in sound […]. Always they try to capture the freshness of an event and to express it of themselves with nothing to dull or cloud the evocation.  (Fortune 1962:â•›6)

The similarity between ideophonic and poetic language is easy to see. Yet the shadow of Lévy-Bruhl (1910), who assigned mimesis in language to the realm of primitivity, has loomed large over linguistics and literary theory alike. Ezra Pound is a case in point: while his fascination with Chinese writing spawned the ideogrammic method, the mimicry and gestures of the ‘primitive languages in Africa’ would never be more than a mere curiosity to him. This paper imagines Pound transposed into the culture of the Mawu people of Â�eastern Ghana. What would have struck him about their ways of charging Â�language with imagery? I juxtapose Pound’s perpetual interest in the exact qualities of Â�perceptions with

.â•… A fuller analysis of this poem (and several others) can be found in Noss’ (1989) article. If there is anything that corresponds to Karl Bühler’s (1934:â•›201) tentative notion of a Tonfilm, this is it.

 Mark Dingemanse

the ways in which ideophones suggest perceptual meanings. The goal of this analysis is to sharpen our views of iconicity in ideophones and thereby to shed light on an important everyday poetic device in the world’s languages, one that we might call (by analogy to Pound’s ideogrammic method), ‘the ideophonic method’.

2.1â•… The Mawu people and their language The Mawu people of eastern Ghana live in a total of eight villages scattered about in the mountains north of Hohoe. They number about 12,000, although there is also a Â�sizable diaspora in Ghana’s major cities and abroad. Their land (Kawu in Siwu) is divided into Akpafu (north-west) and Lolobi (north-east), corresponding to a Â�dialectal Â�division. Most Mawu are peasant farmers growing upland rice, cocoa, corn, cassava, yam, Â�plantain, and other crops on small plots of land. The men sometimes have day jobs in neighbouring towns, while the women sell in the markets what remains from the produce after their family has been fed. Siwu, the language spoken by the Mawu people, is a Kwa (Niger-Congo) language of the Ghana-Togo Mountain group. Siwu has twenty consonants and seven oral Â�vowels with five nasalized counterparts. The language is tonal, with tone performing both lexical and grammatical functions. Siwu has an active noun class system with singular and plural classes expressed by prefixes and agreement marked on the verb. Serial verb constructions are common. Verb paradigms express mainly aspect features; tense is less important. The examples in this paper are drawn from an extensive corpus of data representing the Akpafu dialect of Siwu as spoken in Mempeasem.

2.2â•… Ideophones in Siwu To get a feel for the kind of words we are talking about, consider the examples in (2) below. The clunky English glosses are provided out of necessity, but one should really pronounce the words to appreciate the rich bouquet of meanings they embody in their singular form.3 (2)

kpebebee kpɔtɔrɔ-kpɔtɔrɔ tsarara ɣiii

the rigid posture of a muscular person the jerky walk of a turtle the fluidity of a liquid substance (e.g. palm oil) the roaring feeling of vertigo

.â•… I use a slightly extended version of the Siwu orthography. Most characters represent their IPA equivalents; /ts/ and /dz/ are palatal affricates, /kp/ and /gb/ are digraphs for the doubly ⁀ and /ɣ/ is a pharyngeal approximant (Ford & articulated labiovelar consonants [kp⁀ ] and [gb] Iddah 1973: §0.1.3).



Ezra Pound among the Mawu 

These are words that belong to the large class of ideophones in Siwu. Ideophones, in Siwu as well as in other languages, set themselves apart from ordinary words in Â�several ways: they are longer on average than nouns and verbs;4 they have Â�deviant word Â�structures (featuring for example long vowels as in saaa ‘cool sensation’ and wùrùfùù ‘fluffy texture’, and bisyllabic and trisyllabic reduplicated roots as in Â�tsɔ̀kwɛ-tsɔ̀kwɛ Â�‘sawing movement’ and nyɛmɛrɛ-nyɛmɛrɛ ‘slithering movement’); they are often only loosely integrated in the sentence, if at all; and finally, in actual use, they are more often than not Â�emphasized, and they easily undergo expressive word formation processes like lengthening and reduplication (for example, saa may be greatly lengthened and tsɔ̀kwɛ-tsɔ̀kwɛ may be indefinitely repeated). The effect of this is that ideophones are quite unlike ordinary words, which do their work Â�without attracting much attention themselves; instead, ideophones stand out as speech heard in a special way. Their material Â�properties (articulatory gestures, sound) are foregrounded, opening up the way for iconic mappings between sound and sense. Ideophones make minute distinctions between different kinds of sensory experiences. Consider the following four Siwu ideophones having to do with silence: (3) kananana quietness, lacking noise or sound nɔmɛ, ɔmãgɛ̃ lo kanana ‘the town is quiet today’ kpoo stillness, nocturnal silence a rɛ kpoo? ‘did you have a sound sleep?’ gbigbinigbi speechless, having no retort ǹdo lòkarɛ ɔ itɔ̃me aɣɛ àta ànyɔ mɛ gbigbinigbi ‘I’m asking you a Â�question and now you’re standing looking at me dumbfounded!’ mùnùmùnù inability to speak for physical or emotional reasons ɔ̀si mùnùmùnù ‘he just sits dumb’ The meanings of these four ideophones are quite different. Kanana is perhaps the most general, often heard in requests for silence or to refer to the silence of the town when everyone is on the farm. Kpoo is also quite common, being part of the Â�morning greeting formula a rɛ kpoo? ‘did you have a sound sleep?’ (Dingemanse 2009). It has a positive connotation of noctural silence and sound sleep, and like kanana it can be used in requests for silence. The other two ideophones carry negative valence; gbigbinigbi evokes being at a loss for words, while mùnùmùnù evokes a physical or emotional inability to speak. To borrow a phrasing of Pound’s, translating these four

.╅ A variance analysis of the word length (counted in phonemes) of 305 verbs, 799 nouns, and 290 ideophones shows that the differences between these three groups are statistically significant: F (2, 1390) = 270.56, p <.0001.

 Mark Dingemanse

ideophones simply by ‘silence’ would be “on a par with translating rhinoceros, fox and giraffe indifferently by ‘quadruped’ or ‘animal’” (1969:â•›21). The point being that ideophones fulfil much the same role as the images in Pound’s work, namely to evoke different species of experiences.

3.â•… Iconicity: Relations between form and meaning in ideophones One of the recurring points of agreement in the literature on ideophones is that they are pervasively iconic, or sound-symbolic, phonosemantic, mimetic. All of these roughly interchangeable terms have the same basic import, namely that ideophones are not just arbitrary linguistic signs but that there is some perceived resemblance between their form and meaning. Iconicity in language relies on the natural fact that linguistic signs are not passive labels paired with abstract meanings, but that their ‘sensory properties’ afford certain possibilities for suggesting meaning (Bühler 1934:â•›195–216; Marks 1978; Jakobson and Waugh 1979; Tsur 1992). How does one go about depicting sensory events in words? The question can be reframed as follows: what are the properties of words and of Â�perceived events such that there can be iconic mappings between the two? There are several. First of all, since speech is sound, it can be used to mime non-speech sounds. But speech is a lot more than the auditory signal: since it is bodies that do the talking, there is also a rich internal structure in the form of articulatory gestures, and this internal structure can be harnessed to depict certain aspects of perceived events. More generally, since speech is itself a kind of sensory experience, it shares with other sensory experiences the basic suprasensory attributes (Marks 1978:â•›51) of duration, intensity, and quality, and this affords us with several possibilities as we will see below. In the discussion that follows I build on the Peircean sense of iconicity, defined as a perceived resemblance between form and meaning. There has been much discussion of iconicity at the level of grammar in recent years (Jakobson 1965; Haiman 1980; Newmeyer 1992; Haspelmath 2008, among others), but I restrict myself to iconicity at the lexical level and specifically in ideophones. First, however, a few caveats that I will explain by means of a further analogy with Pound’s work.

3.1â•… A cautionary tale Over the years, Pound developed a fascination with the poetic affordances of Â�logographic writing systems, especially Chinese. This fascination originated with his Â�discovery of a theory of the Chinese character by Ernest Fenollosa (1936), who argued that Chinese writing reflects etymology (‘true sense’) in a way that phonetic writing does not. In Pound’s idealist view of etymology (Li 1986), this rendered the Chinese Â�character vastly superior to Western phonetic script in terms of picture-making. Soon



Ezra Pound among the Mawu 

enough however, scholarly studies of logographic writing systems showed that Â�Chinese characters are semantic-phonetic compounds rather than transparent pictures, and Pound’s idyllic conception of Chinese characters as evocative ideograms was severely and justly criticized (Kennedy 1958; cf. also DeFrancis 1984). Now, ideophones are often too readily characterized as imitative words in a way that is quite reminiscent of Pound’s overeager iconization of Chinese writing. In reality, the picture is much more complex. First of all because ideophones are never exclusively iconic (Nuckolls 1992). Like all linguistic signs, they mix Peircean modes of signification (for example, they are symbolic in that they are subject to conventionalization so that their interpretation is partly socially mediated; and they are indexical to the extent that they ‘point to’ perceptions, inviting the listener to see for themselves). Secondly, because there are several types of iconicity at play in ideophones, and simple imitation is but one of them (and a minor one at that). Thirdly, because not all ideophones are transparently iconic. This last point, which is sometimes too easily glossed over by those who investigate sound-symbolism in ideophones, bears underscoring. Not all ideophones show transparent form-meaning mappings, for there are limits to the iconic representational powers of speech. That speech can mimic non-linguistic sound is trivial, and that it can depict aspectual structure is also widely attested (see Sections 3.2 and 3.3.1 respectively), but beyond that we quickly run out of possibilities. Thus, while the aspectual structure of a sawing movement can be depicted by reduplication, it is less clear how one would go about depicting a colour in speech. And yet we do find ideophones for colours – not just in Siwu, but in many ideophonic languages. In fact, cross-linguistically, the meanings of ideophones frequently extend over all sensory modalities and also include inner feelings and sensations, thus encompassing many percepts that seem difficult or impossible to map iconically in speech (take for instance Siwu kpɔkɔtɔɔ ‘nervous feeling’ or ɣɛkpɛtɛɛ ‘frail’).5 Anticipating the discussion of relative iconicity in Section 3.3.2 we may note that even though colour as such may be difficult to depict in speech, there are ways to depict relations between colours by means of relations in sign forms. This does indeed happen, but it is of course highly unlikely that this strategy could scale to accommodate all possibly relevant dimensions of the varieties of sensory experience. This is what Karl Bühler meant when he wrote that sound-symbolism could never form a “coherent representational field” in language (Bühler 1934:â•›203, my translation). To reiterate, there are limits to the iconic representational powers of speech, and there is reason to be careful in ascribing

.â•… To be clear, these words are members of the form class of ideophones, a class that is defined (for Siwu) in terms of semantic and structural properties (summarized in Section 2.2) rather than in terms of iconicity.

 Mark Dingemanse

Â� iconicity to ideophones. (I will come back to the implications of this for the class of ideophones as a whole in Section 4). With these caveats out of the way, there are still significant regularities in form-meaning mappings to be accounted for in Siwu ideophones, and it is to these Â�regularities that we will now turn.

3.2â•… Imagic iconicity Imagic iconicity in ideophones pertains when the sound of the word mimics a sound in the real world. Examples from Siwu include ideophones evoking sounds of collision and explosion (gbíim̀ ‘explosion [boom]’, kpà ‘dry impact [bang]’, tuu ‘dull impact [thud]’), water-related sounds (kpɔtsɔkpɔtsɔ ‘water bubbling’, tòlontòlontòlon ‘water dripping’), and other ecologically-relevant sounds (korrrò karrrà ‘gnawing bones’, gadàm ‘stamping’, kìrìrì-kìrìrì ‘scratching of claws’). Semiotically, this is the simplest type of mapping; it is probably also the most familiar, as the existence of onomatopoeia is usually granted even by the staunchest proponents of the fundamental arbitrariness of the sign (e.g. Müller 1899; Â�Newmeyer 1992). The simplicity of the mapping brings with it inherent limitations. Firstly, and most obviously, only sound can be represented directly in the modality of speech (imagic iconicity in sign language is therefore another thing altogether); secondly, words employing imagic iconicity are never perfectly faithful copies of the sound mimicked; as conventional linguistic signs they are embedded in the language system and largely dependent on the affordances of its phonemic inventory (this is the relative arbitrariness that, as Saussure famously pointed out, holds even for onomatopoeia). Numerically, imagic iconicity is of minor importance in the Siwu corpus; only about 8% of the inventory can be classified as directly imitative of sound. (I should note that there are also a handful of nouns of onomatopoeic origin like kpɔ̀ɔkpɔ̀ ‘duck’, ì-fokofoko ‘lung’, and kɛsɛkɛsɛ ‘brush-tailed porcupine’, the latter named after the Â�warning rattle it produces with its tail.) Clearly, most Siwu ideophones do not directly imitate sound. But not only are onomatopoeic ideophones relatively rare in Siwu, they also deviate from other ideophones in several ways. For example, they cannot be used as verbs and do not appear in some of the most common ideophonic Â�constructions; and unlike most other ideophones, they do appear in the quotative construction. It seems that this broad distinction between onomatopoetic ideophones and ideophones depicting other types of sensory events reflects a cross-Â�linguistic tendency.6

.╅ Cf. for example the basic distinction made in Japanese linguistics between giongo (soundmiming ideophones) and gitaigo (event-miming ideophones) (Akita 2009). Kilian-Hatz (2001) notes the same broad distinction comparing Baka (an Adamawa language of Cameroon) and Kxoe (a Central Khoisan language from Namibia).



Ezra Pound among the Mawu 

It is worth noting that not all ideophones for acoustic phenomena are Â�necessarily of the imagic iconic type. Take for instance the ideophone kpɛnɛnɛ ‘shrill, piercing [of sound, especially voice]’. Instead of imitating a particular type of sound, as gbíim̀ or tòlontòlon do, kpɛnɛnɛ picks out a psychoacoustic property (roughly, brightness). We get a better feel for its meaning when we consider that it has an antonym wɔrɔrɔ ‘low, hoarse [of sound, esp. voice]’. The relevant psychoacoustic property seems to be brightness; the correspondence of front vowel /ɛ/ and back vowel /ɔ/ with Â�different shades of brightness has been well documented (Marks 1978:â•›76–80). This, then, brings us from imagic iconicity – sound mimicking sound – to more complex types of mappings.

3.3â•… Diagrammatic iconicity In diagrammatic iconicity, a relation between forms bears a resemblance to a relation between meanings. Diagrammatic iconicity brings with it a wider range of possible mappings because the full range of sensory attributes of speech – including acoustic dimensions, articulatory gestures and temporal unfolding – can be used to suggest meaning. At the same time, in exchanging the direct mapping of sound onto sound for more abstract mappings, we enter the realm of the weakly iconic (Lyons 1977:â•›103), where the meaning of signs cannot be deduced solely on the basis of their form, even though knowing the meaning we can see some resemblance of form and meaning. Two types of diagrammatic iconicity are found in Siwu ideophones. The first, Gestalt iconicity, holds for individual words, while the other, relative iconicity, plays out in the relations between multiple words and multiple meanings.

3.3.1â•… Gestalt iconicity Gestalt iconicity involves a resemblance between word structure and (spatio-Â� temporal) structure of the perceived event; or in Bühler’s terminology, words that are “Gestalt-faithful” (Bühler 1934:â•›208, my translation). It is a type of Â�diagrammatic iconicity in that a relation between forms (the parts of the word) has a resemblance to a relation between meanings (in this case parts of the perceived event, that is its aspecto-temporal unfolding). Gestalt iconicity is most apparent in the word structures of some Siwu Â�ideophones. Fully reduplicated ideophones for example predominantly evoke Â�perceptions of iterated or distributed events, like mũnyɛ̃mũnyɛ̃ ‘sparkling light’, sàsàsàsàsà ‘pulsatile release of urine’, biribiri ‘sowed at close intervals’, gidigidi ‘running energetically’, tsɔ̀kwɛtsɔ̀kwɛ ‘irregular sawing motion’, nyɛnɛnɛ ‘shivering’ and kpɔrɔkpɔrɔ ‘drizzling rain’. Monosyllabic ideophones on the other hand often evoke perceptions of unitary events, like gbÃ�im̀ ‘explosion’, dzâ ‘sudden appearance’, kpiɛ ‘careful step of an antelope’, pɔ ‘frog hop’, wãĩ ‘bright flash’. Words of the latter sort cannot be reduplicated, though they

 Mark Dingemanse

may sometimes be repeated.7 Monosyllables with an extra-long vowel (marked here by three vowel characters) often depict unitary events that are also durative, like saaa ‘cool sensation’, ɣààà ‘flowing quietly without obstruction’, kããã ‘looking attentively’. What is common to these examples is that the form of the word betrays Â�something of the perceived event structure. Thus in mũnyɛ̃mũnyɛ̃ ‘twinkling of light’ the recurring of the base mũnyɛ̃ (which cannot occur on its own) corresponds to recurring elements in our perception of the visual pattern; in dzâ ‘sudden appearance’ the unitary form of the ideophone corresponds to the punctual nature of the perceived event; and in ɣààà ‘flowing quietly without obstruction’ the extra-long vowel corresponds to the durative nature of the perceived event. These regularities appear to be robust for percepts with salient aspecto-temporal structure, such as movement, visual phenomena, and sound.8 We have to ask –mindful of our cautionary tale– whether we are not Â�over-eager in ascribing Gestalt iconicity to these examples. Though it is difficult to offer Â�conclusive proof at this point, there are two arguments that can be made. First, it would be rather surprising if the distribution of forms over meanings were reversed, as in *dzâ ‘twinkling of light’ and *mũnyɛ̃mũnyɛ̃ ‘sudden appearance’, or *pɔ ­‘shivering’ and *nyɛnɛnɛ ‘frog hop’. That is, the distribution of forms over meanings at least in the subset Â�considered seems too skewed to be arbitrary. I leave this as an empirical question, which can be tested along the lines of Klamer’s (2002) study of semantically Â�motivated Â�lexical Â�patterns. Secondly, there is a suggestive piece of evidence in the form of the hand Â�gestures that accompany ideophones (Diffloth 1972:â•›441 Â�Dingemanse 2011: Â�217–222). For example, in explaining mũnyɛ̃mũnyɛ̃ ‘sparkling of light’, speakers make flashing gestures in perfect synchrony with the reduplicated base *mũnyɛ̃. And in explaining bɔgɔbɔgɔ ‘resilient’, my consultant Ruben Owiafi

.â•… Repetition and reduplication are distinct in Siwu ideophones. In repetition, there are pauses between the repeated tokens (marked here by a comma). Thus, the frog hops ‘pɔ, pɔ, pɔ’ and not pɔpɔpɔ, and the antelope may walk ‘kpiɛ, kpiɛ’ but not kpiɛkpiɛ (which happens to be another ideophone – ‘sizzling hot’). .â•… There is furthermore the possibility that some of the ideophones that do not appear to be transparently iconic in fact reflect perceived event structure. Ideophones might be taken to project a certain cognitive construal of events (cf. Nuckolls’ (1995) analogy with Â�cinematic techniques). I hesitate to mention this possibility because circularity is never far away – I might claim that Siwu ɣèèè ‘animals swarming in great numbers’ entails a holistic construal of the swarming event as opposed to a construal that focuses on the Â�pluractionality of the event (in which case we would expect reduplication), chalking it up as another example of Gestalt iconicity, but this would sidestep what is in fact an empirical question. Gesture Â�provides Â�suggestive evidence – ɣèèè comes with a single left to right flowing hand gesture – but the issue merits further research and could in fact be a fruitful locus for cross-linguistic Â�investigation. For example, does the Japanese ideophone uja-uja ‘many small things gathered together and moving’ (Gomi 1989:â•›24), which is reduplicated unlike Siwu ɣèèè, entail a Â�construal of a swarming event that focuses more on the plurality?



Ezra Pound among the Mawu 

repeatedly bends an imaginary cane, the bending Â�gestures perfectly aligned with the repetitions of *bɔgɔ. Ideophone and gesture work together to depict imagery of sparkling lights and a resilient cane, respectively.9 To summarize, Gestalt iconicity is a weak iconic relationship between form and meaning that is most clear in ideophones evocative of events with salient aspectotemporal structure.

3.3.2â•… Relative iconicity Relative iconicity is about mapping a relation between forms onto a relation between meanings. Like Gestalt iconicity, it is a type of diagrammatic iconicity; but unlike Gestalt iconicity, which focuses on the internal structure of signs and meanings, Â�relative iconicity concerns a relation between multiple lexical signs bearing a resemblance to the relation between multiple meanings. consider, again, kpɛnɛnɛ ‘high, shrill voice’ and wɔrɔro ‘low, hoarse voice’. In these two signs, the relation between the front vowel /ɛ/ and the back vowel /ɔ/ bears a resemblance to the relation between their objects, a sound high in brightness versus a sound low in brightness (on brightness as a psychoacoustic attribute of sound, see Marks 1978). There are quite a number of constellations of ideophones showing relative Â�iconicity. The members of such constellations differ minimally from each other in both form and meaning. This can be seen as a form of templatic morphology in that there is a stable segmental skeleton (the template) and a variable part in which members of a sound-symbolic series appear. In (4) for example, the vowel slots in the template pVmbVlVV can be occupied by /ɔ/, /u/, or /i/, resulting in different shades of the basic meaning ‘protruding (of the belly)’. The stable consonantal skeleton holds together the three ideophones, signifying that they belong to the same domain. (4) pimbilii pumbuluu pɔmbɔlɔɔ

‘protruding (of the belly)’ (/i/ is smallest, /ɔ/ is largest)

Or take the contrast between foforo-fo ‘lightweight (e.g. as a slim person)’ and fɛfɛrɛ-fɛ ‘very lightweight (e.g. as a feather)’, exemplified in (5) and (6) below. The different magnitudes are represented by the relation between front vowel /ɛ/ and the back vowel /o/. (5) ɔ-sò ɔ̀rãrã, ɔ̀so ɔ̀bra foforofo 3sg:pst-lose weight, reason 3sg-make idph.light ‘she lost weight, so now she is foforofo [light]

.╅ It remains to be shown, of course, that this is more than mere synchronization of motor processes.

 Mark Dingemanse

(6) kÀkosɛÃ� ba fɛfɛrɛfɛ paper.piece have idph.very.light ‘The piece of paper is fɛfɛrɛfɛ [very light]’ It is not only vowels that change. In the pair tsratsratsra ‘fast and light walk’ and dzràdzràdzrà ‘fast and heavy walk’ the differences in the weight of the moving figure are represented by the relation between the voiceless palatal affricate /ts/ and its voiced counterpart /dz/ as well as by the difference in tonal melody (high versus low). Relative iconicity is attested in ideophone inventories all over the world. Â�Westermann (1927) was the first to outline several recurrent relative iconic mappings in the ideophone systems of a number of West-African languages, for example high tones, light vowels and voiceless consonants evoking smallness, clearness, and speed, versus low tones, dark vowels and voiced consonants evoking large size, dullness, and slowness. Since then, mappings like these (especially vowel contrasts) have become one of the most widely researched topics in sound symbolism; useful reviews are Marks (1978) and Hinton, Nichols and Ohala (1994). Many of the cases of relative iconicity in Siwu ideophones have to do with magnitude symbolism and fit the frequency code (Ohala 1994). But there are also cases like saaa ‘cool sensation (e.g. ginger)’ and suuu ‘burning sensation (e.g. pepper)’, in which the different vowel qualities map onto different sensory qualities instead of a scalar attribute like intensity. In such cases there seems to be no reason other than convention why /a/ should go with ‘cool’ and /u/ with ‘burning’ (cf. Diffloth 1994), but still the resemblance between the words marks them as being about the same kind of meaning. In all, about one-fifth of the Siwu ideophone inventory partakes in constellations of relative iconicity.10 Like primary and Gestalt iconicity, relative iconicity is visible in the Siwu ideophone corpus, but it is by no means a feature of all ideophones.

4.â•… Beyond lexical iconicity We have seen a number of different form-meaning mappings in Siwu ideophones. None of them extends over all ideophones in the corpus. Have ideophone enthusiasts (native speakers as well as linguists) simply been over-eager in iconizing ideophones, or is there another reason that the form of ideophones is so often identified with their meaning? Speakers often self-report that ideophones feel like what they evoke; that words like nyɛnɛnɛnɛ ‘shivering’, vɛlɛvɛlɛ ‘feeling dizzy’ and fututu ‘purely white’

.â•… The majority of these constellations has only two members, like foforo and fɛfɛrɛ ‘light’/‘featherweight’ described above; the maximum is 4, the average is 2.4.



Ezra Pound among the Mawu 

are Â�particularly good representations of their meaning in ways that words like ti ‘to shake’, ɔ̀nyɛ̃ Â�‘sickness’ or yuɛ ‘be green’ are not. For nyɛnɛnɛnɛ we can point to the similarity in the repeated movement and the repeated syllable and call it Gestalt iconicity. Yet such an account seems less plausible for vɛlɛvɛlɛ, which represents an inner sensation with no necessary outward signs, and even less for fututu, which represents a colour. What then makes speakers of Siwu feel these words are good evocations of sensory imagery? Their depictive nature does. Depiction, rather than iconicity, is what invites people to treat the ideophone as a performance of sensory imagery. An analogy may help to explain this point. Consider the category of objects called paintings. Â�Paintings vary quite widely in the degree to which they are iconic (i.e. show a Â�perceived Â�resemblance to what they depict). And yet there is a distinct Â�interpretive frame we bring to all of them: we tend to view them as depictions rather than read them as texts (Gombrich 2002[1960]; Walton 1973).11 In a similar way, we may think of ideophones as setting up a depictive interpretive frame, inviting the Â�listener onto the scene and invoking images of being there. Their status aparte in the stream of speech can be seen as the frame of a painting, their sound as the canvas at which we are invited to ‘look’ in such a way that we make believe we are Â�actually Â�experiencing the scene depicted. In that sense, the German term Lautmalerei (painting in sound) in the early literature on ideophones was remarkably apt – not because ideophones are imitative (a view that doesn’t suffice for paintings nor for ideophones) but because they invite the listener to savour them as depictions. If we want to invoke iconicity here at all, we should call it coerced iconicity. The depictive nature of the ideophone coerces us into treating the word as an adequate rendition of the depicted event. There is some suggestive evidence from experiments carried out by Kimi Akita (Akita 2009; this volume) to support this point. Akita investigated sound symbolic interpretations of ideophones and non-ideophones. In one experiment, speakers of Japanese rated novel words for size of the imagined referent. Consonant voicing, vowel quality and morphophonological similarity to ideophones were systematically varied in the novel words. In a follow-up experiment, the novel words were placed in a Â�sentential frame that specified part of their meaning. The results suggest that subjects were more inclined to ‘see’ magnitude symbolism in novel words that sounded like ideophones than in novel words that sounded like ordinary words; and the effect was

.â•… Many of the oppositions stressing the special status of ideophones – commentative vs. presentive (Kunene 1965), prosaic vs. expressive (Diffloth 1972), discursive vs. performative (Nuckolls 1995), descriptive vs. mimetic (Güldemann 2008), to sample a few – point to this crucial difference in mode of representation.

 Mark Dingemanse

strongest for ideophone-like words presented in an ideophonic construction. In terms of the present discussion, people were most inclined to see form-meaning mappings in novel words when they were presented as depictions. Although this experimental finding speaks to existing words only indirectly, it does underline the importance of framing ideophones as depictions.

5.â•… Concluding remarks “The problem of the word,” wrote Pound, “cannot be exhausted in a single lifetime” (1973:â•›321). All through his life Pound displayed a deep interest in how different Â�languages and ecologies may afford different types of poetics. This interest formed the backdrop to his fascination with Chinese ideographic writing and his wonder over exotic languages. He was obviously impressed by Lévy-Bruhl’s Les Fonctions mentales dans les sociétés inférieures, and his work contains several allusions to Lévy-Bruhl’s characterizations of non-Western language structures. The ABC of Reading mentions “primitive languages in Africa […] that are still bound up with mimicry and gesture” (1934:â•›21). This is an allusion to Lévy-Bruhl’s (1910:â•›183–6) summary of a Â�description of Ewe ideophones by the eminent linguist Diedrich Westermann (1907:â•›83–5, 129–30). It is the closest Pound has ever been to ideophones, and it is unfortunate (though probably inevitable) that the meeting had to take place on the unholy ground of scientific racism and cultural evolutionism. I will not rehearse the Â�arguments against these views. My point in this paper has rather been to provide a Â�characterization of these ill-understood words that resonates with Pound ideas about the poetic word and that is more fertile from the point of view of the study of semiotics and verbal art. I have shown how different types of iconicity allow ideophones to move beyond the imitation of singular events towards perceptual analogies and generalizations of event structure; and I have argued that even though not all Â�ideophones are transparently iconic, they do set up a depictive interpretive frame, inviting listeners to ‘be there’. Ideophones are akin to Pound’s luminous details in that they focus on presentation rather than commentary. Evocative and defiant, they remind one of a memorable passage from Pound’s interpretation of the Â�Confucian writings: Intangible and abstruse the bright silk of the sunlight Pours down in manifest splendor, You can neither stroke the precise word with your hand Nor shut it down under a box-lid.

(Pound 1969:â•›133)



Ezra Pound among the Mawu 

References Akita, K. 2009. A Grammar of Sound-Symbolic Words in Japanese: Theoretical Approaches to Iconic and Lexical Properties of Japanese Mimetics. Ph.D. dissertation, Kobe University. Basbøll, T. 2008. Decision and desire: A ‘rain-sparkling crystogram’ in Nabokov’s The Defence. Jacket 36. 〈http://jacketmagazine.com/36/basboll-decision.shtml〉 Bühler, K. 1934. Sprachtheorie: Die Darstellungsfunktion der Sprache. Jena: G. Fischer. DeFrancis, J. 1984. The Chinese Language: Fact and Fantasy. Honolulu HI: University of Hawaii Press. Diffloth, G. 1972. Notes on expressive meaning. Chicago Linguistic Society 8: 440–447. Diffloth, G. 1994. i: big, a: small. In Sound Symbolism, L. Hinton, J. Nichols & J.J. Ohala (eds), 107–114. Cambridge: CUP. Dingemanse, M. 2009. Ideophones in unexpected places. In Proceedings of the Second Â�Conference on Language Documentation and Linguistic Theory, P.K. Austin, O. Bond, M. Charette, D. Nathan & P. Sells (eds), 83–97. London: School of Oriental and African Studies. Dingemanse, M. 2011. The Meaning and Use of Ideophones in Siwu. Ph.D. dissertation. Â�Nijmegen: Max Planck Institute for Psycholinguistics/Radboud University. Fenollosa, E.F. 1936. The Chinese Written Character As a Medium for Poetry. London: S. Nott. Ford, K. & Iddah, R.K. 1973. A Grammar of Siwu. Ms, Legon, Ghana. Fortune, G. 1962. Ideophones in Shona: An Inaugural Lecture Given in the University College of Rhodesia and Nyasaland on 28 April 1961. London: OUP. Gombrich, E.H. 2002. Art & Illusion: A Study in the Psychology of Pictorial Representation, 6th edn. London: Phaidon. Gomi, T. 1989. An Illustrated Dictionary of Japanese Onomatopoeic Expressions. Transl. by J. Turrent. Tokyo: Japan Times. Güldemann, T. 2008. Quotative Indexes in African Languages: A Synchronic and Diachronic Â�Survey. Berlin: Mouton de Gruyter. Haiman, J. 1980. The iconicity of grammar: Isomorphism and motivation. Language 56: 515–540. Haspelmath, M. 2008. Frequency vs. iconicity in explaining grammatical asymmetries. Â�Cognitive Linguistics 19: 1–33. Hinton, L., Nichols, J. & Ohala, J.J. (eds). 1994. Sound Symbolism. Cambridge: CUP. Jakobson, R. 1965. Quest for the essence of language. Diogenes 13(51): 21–37. Jakobson, R. & Waugh, L.R. 1979. The Sound Shape of Language. Bloomington IN: Indiana Â�University Press. Kennedy, G. 1958. Fenollosa, Pound, and the Chinese character. Yale Literary Magazine 126:â•›24–36. Kilian-Hatz, C. 2001. Universality and diversity: Ideophones from Baka and Kxoe. In Ideophones [Typological Studies in Language 44], F.K. Erhard Voeltz & C. Kilian-Hatz (eds), 155–163. Amsterdam: John Benjamins. Klamer, M. 2002. Semantically motivated lexical patterns: A study of Dutch and Kambera expressives. Language 78: 258–286. Kunene, D.P. 1965. The ideophone in Southern Sotho. Journal of African Languages 4: 19–39. Lévy-Bruhl, L. 1910. Les Fonctions mentales dans les sociétés inférieures. Paris: F. Alcan. Li, V.P.H. 1986. Philology and power: Ezra Pound and the regulation of language. Boundary 2 15(1–2): 187–210. Lyons, J. 1977. Semantics, Vol. 1. Cambridge: CUP.

 Mark Dingemanse Marks, L.E. 1978. The Unity of the Senses: Interrelations among the Modalities. [Academic Press Series in Cognition and Perception]. New York NY: Academic Press. Müller, F.M. 1899. The Science of Language: Founded on Lectures Delivered at the Royal Â�Institution in 1861 and 1863, 2nd ed., 2 vols. London: Longmans, Green, and Co. Newmeyer, F.J. 1992. Iconicity and generative grammar. Language 68: 756–796. Noss, P.A. 1989. The ideophone poems of Dogobadomo. Crosscurrent 23(3–4): 33–43. (Special issue Ancestors’ Beads, Philip A. Noss (ed.)). Nuckolls, J.B. 1992. Sound symbolic involvement. Journal of Linguistic Anthropology 2: 51–80. Nuckolls, J.B. 1995. Quechua texts of perception. Semiotica 103: 145–169. Ohala, J.J. 1994. The frequency code underlies the sound symbolic use of voice pitch. In Sound Symbolism, L. Hinton, J. Nichols & J.J. Ohala (eds), 325–347. Cambridge: CUP. Pound, E. 1914. Vorticism. Fortnightly Review 96: 461–471. Pound, E. 1934. ABC of Reading. London: Routledge. Pound, E. 1969. Confucius: The Great Digest, the Unwobbling Pivot, and the Analects. New York NY: New Directions. Pound, E. 1973. Selected Prose, 1909–1965. William Cookson (ed.). New York NY: New Directions. Tsur, R. 1992. What Makes Sound Patterns Expressive? The Poetic Mode of Speech Perception. Durham NC: Duke University Press. Voeltz, F.K.E. & Kilian-Hatz, C. (eds). 2001. Ideophones [Typological Studies in Language 44]. Amsterdam: John Benjamins. Walton, K.L. 1973. Pictures and make-believe. The Philosophical Review 82: 283–319. Westermann, D.H. 1907. Grammatik der Ewe-Sprache. Berlin: Dietrich Reimer. Westermann, D.H. 1927. Laut, Ton und Sinn in Westafrikanischen Sudanssprachen. In Â�Festschrift Meinhof: Sprachwissenschaftliche und andere Studien, C. Meinhof (ed.), 315–328. Hamburg: L. Friederischsen.

Cognitive iconic grounding of reduplication in language Olga Fischer

University of Amsterdam/Vienna My aim in this paper will be to find out to what extent opaque reduplications could also be said to be (or to have been) motivated (i.e. forming iconic signs), with the further aim of exploring the possibility of a common source for all reduplicated forms. I will show, by referring to the way repetition is used in signed languages, by looking at the various functions the prefix ge- has in Germanic languages, which resembles reduplication in terms of its semantics, and by taking common pathways of semantic change into consideration, that such a common source may be said to exist.

1.â•…Introduction Many reduplicated forms that occur in languages around the world and especially in creoles or historically younger languages are cognitively transparent or iconic, but there are also instances of reduplication where iconic motivation is not, or much less clearly, evident. In this paper I will consider reduplications which are generally considered to be iconically opaque and compare them to others which are seen as transparent. Transparent reduplications are motivated through the repetition of a morpheme expressing a cognitive increase of some sort. They generally show semantic enhancement of the dominant semantic feature of a reduplicated noun, verb or adjective/ adverb (prepositions and pronouns are less often involved). This feature may also be of a more pragmatic nature, that is, reduplication may strengthen a connotative rather than denotative meaning, due to the context in which it appears. My aim in this paper will be to find out how far opaque reduplications could also be said to be (or to have been) motivated, with the further aim of exploring the possibility of a common source for all reduplicated forms.1 .â•… William Herlofsky (p.c.) describes this transparent relation as one of “a natural resemblance between the mental representation of the form of a sign and the mental representation of the object or concept it refers to”. It will be clear from the discussion below that it is indeed not the form and the object/concept themselves that provide the analogy or iconic grounding,

 Olga Fischer

Reduplication does not only occur in spoken languages. It is a very common Â�phenomenon in signed languages in the form of the repetition of manual gestures; it will, therefore, be useful to compare the uses to which it is put there with the way Â�reduplication is used in spoken languages. I will work from the assumption that both are grounded cognitively in the same way since both concern human minds and Â�mental linguistic abilities. In my investigation, I will also briefly address the question of why, in general, reduplication seems to be so much more common in creoles and in languages without a written standard and/or a long documented history – these two aspects often go together (but see Note 17 below) Â�– than in well-established standardized languages.2 The paper is structured as follows: I will first globally discuss the various forms of repetition that have been found, since reduplication is in essence a form of repetition. I will be mostly interested in reduplication proper, that is in reduplicated forms that have been lexicalized (Section 2). Next (Section 3), I will look at those forms of reduplication which are generally considered to be transparent as far as their function or meaning is concerned. Section 4 will then be devoted to opaque forms of reduplication, wherein I will categorize various functions. In Section 5, I discuss the reasons why most linguists deny transparency in the latter case and investigate the extent to which opaque forms could still be linked to an iconic source. To come to a deeper understanding as to why reduplication is employed with such opaque functions, I will first of all make use of evidence provided by repeated gestures from signed languages (Section 5.1). Secondly, I will consider the various functions of the prefix ge- as they are found in a number of Germanic languages (Section 5.2). Since ge- shares some of its functions with reduplicated forms, it would be of interest to find out whether the other functions of geÂ�-, which are not related to the idea of increase, might also occur among opaque reduplicated forms. If this is the case, it could, of course, be ascribed to sheer coincidence, but it might also tell us something about a deeper cognitive source for both the ge- forms and the reduplicated forms. Finally, in Section 5.3, I will consider cases in which transparency may get lost, arguing that processes such as grammaticalization and lexicalization may obscure the transparency of the original reduplicated form, and its original function. Section 6 provides some brief conclusions.

but the way these are present in our mind. This gives a form the space to enter into new relations by metaphorical as well as metonymic means. How this works becomes clear from Figures 2 and 3 in Section 5 below. .â•… The term ‘standardized’ as used here, refers to languages that have been codified as part of a national identity providing a set of norms for standard use in writing. This codification is supported by dictionaries and (prescriptive) grammar, and is part of the educational system learned by children at an early age. The conventions of such a written standard differ in many ways from those used in the spoken standard, and it is well-known that the written variety can influence, as a prestige form of language, the way people speak (cf. Crystal 1987:â•›179).



Cognitive iconic grounding of reduplication in language 

2.â•…Forms of repetition including reduplication The repetition of lexical material is found on various levels: on the level of (i) the clause or phrase, (ii) the word or morpheme, and (iii) on the sub-morphemic level. We normally only speak of reduplication when the item in question has been lexicalized. It seems clear, though, that this lexicalized repetition developed out of an earlier, looser usage since the functions it serves are more or less the same as that of more general repetition. i.

The examples of (looser) repetition all indicate an increase of some sort, which can be read as intensity or urgency, or boredom/tiredness as a result of something being repeated: Come in, come in! (‘do come in!’, aspect of urgency) All right, all right (urgency, ‘stop saying this’, but also boredom depending on the intonation) Out, out, brief candle (intensity of emotion)  (Macbeth, Act 5, v, 19)

ii. At the word level, the repetition of a word can still be loose, in which case it may be found in all languages. However, it can also be fully lexicalized; not surprisingly, these cases are more rare. Lexicalization is indicated in written texts by means of a hyphen or by writing the two words together as one; when spoken, the difference between lexicalized and non-lexicalized forms is usually indicated by an intonational clue or a difference in stress pattern. Some languages only have incidental lexicalized forms (these may be relics of an earlier, more productive rule), while in others this type of reduplication is productive, playing an active role in word formation. Examples of lexicalized types will be discussed in the next section together with their categorization. Italian bella bella (intensification: ‘very beautiful’, with emotional overtones) ‘A sálad salad’ (intensification: a ‘real’ (genuine) salad, not e.g. a tuna salad; cf. Ghomeshi et al. 2004) He’s a wonderful, wonderful person (intensification, but exaggerated)

iii. Reduplication at the sub-morphemic level or partial (syllable) reduplication is always lexicalized and usually the result of phonetic reduction, often accompanied by semantic bleaching through grammaticalization, but it is also possible that the reduplicated form does not derive from an originally full morpheme. An Â�example of grammaticalized, reduced reduplication is the perfect in Â�Indo-European. Â�Remnants of this kind can be found in Gothic (infinitive fahan ‘seize’ – perfect faÃ�fah; Old English hatan – heht (< *hehet)); in Greek (λυω ‘to loosen’– λελυκαi); and in Latin (cano ‘to sing’ – cecini). In Tepecano (an Uto-Aztecan language) reduced reduplication is found to express the plural: nov > nonov ‘hand(s)’, dur > dudur ‘ant(s)’ (Haugen 2005), and in Saramaccan it usually expresses result or an

 Olga Fischer

iv.

v.

vi.

vii.

Â� instrument: taitai > tatai ‘rope’ (tai ‘to bind’), waiwai > wawai ‘fan’ (wai ‘blow’) (Bakker & Parkvall 2005). Next to these forms of repetition and reduplication on different linguistic Â�levels, there are also a number of subsidiary reduplications which are Â�different from the above-mentioned reduplications, because, although a syllable is repeated, the syllables themselves are not based on an existing word-stem, as far as can be Â�ascertained. These concern (iv) ‘Ablaut reduplication’, (v) ‘onomatopoeic reduplication’, (vi) reduplication in babytalk/motherese, and (vii) hypocoristic use. These categories may indicate the pervasive use of a rhythmical form of repetition in language, which is presumably inherent to most of us (consider the importance of rhyme, metre, and rhythm, which are all repetitive patterns). However, apart from the baby-talk and the hypocoristic varieties, these categories will play no further role in the present discussion. Ablaut reduplication is very common: Dutch bim bam (sound of church bells); English knick-knack, see-saw. It is interesting that in all these cases (and this seems to be true for most languages that show this phenomenon) a back vowel follows a front vowel; the back vowel appears to function as a kind of closure. Onomatopoeic reduplication occurs in many languages, and onomatopoeia is often present in (iv) above as well as (vi) below. It is especially prevalent in bird names, as in English cuckoo, Dutch koekoek; and to represent animal noises, as in Dutch kwek kwek (for ducks), bè-bè (for sheep), and for other noises (made by cars etc.) as in Dutch toet-toet, vroem-vroem. They can also be more fully lexicalized into verbs and nouns: Greek murmurizo, English murmur; Greek gargara ‘gurgle’. Reduplication in baby talk or motherese can be found in almost universal words like papa, mama, and in English words such as wee-wee, or German lu-lu to refer to the important interest in toilet-training with respect to babies. It is likely that reduplication here is a (motherese) mimicry of ‘babbling’, and in that sense it could be seen as onomatopoeic in nature (cf. Aitchison 1994:â•›31). Hypocoristic terms are often used to refer to small children, which relates this category to (vi) above. It can be seen in names such as Mimi, Fifi, Didi etc. This is often combined with high, front vowels, which are frequent in diminutive affixes, possibly to emphasize the small child. Fónagy (2001:â•›27; 347–349) notes that there is an iconic, articulatory link between high front vowels (and/or palatalized consonants) and smallness, lightness and tenderness (qualities typically associated with children or our relation to children). It is interesting to observe in this respect that when there is a choice of a phonotactic base, the first legitimate string may be discarded in favour of the second when the second has a high front vowel, or an extra vowel such as [i] may be added, as has happened in



Cognitive iconic grounding of reduplication in language 

the Cape Â�Verdian Creole names ‘Bibina’ from ‘Balbina’ and Totói from ‘António’ (cf. Baptista 2003:â•›177–178).

3.â•…Transparent use of lexicalized reduplication at the word level Reduplication is considered to be transparent when it expresses ‘more of the same’, that is an increase or enhancement of some sort. ‘More of the same’ can either be seen as repetition on a vertical axis, which could be graphically represented as: ≡, in which case it conveys intensity or augmentation. Or it can be interpreted as repetition on a horizontal axis: – – –, in which case it expresses iteration or continuity/duration. The specific function of reduplication also depends on the part of speech with which it combines; and within each part of speech, it further depends on what subordinate class they belong to.3 Thus with a count noun, reduplication generally expresses plurality (often a scattered type of plurality, dispersion), and with a mass noun it usually conveys augmentation in the form of growth or intensification. Similarly, with a punctual or telic verb, iteration is usually conveyed, while a reduplicated atelic verb is more likely to express duration or an interrupted activity. Adjectives can be divided into gradable and non-gradable types. With the former, reduplication leads to intensification, while with the latter plurality or iteration is expressed. The various possible categories are illustrated in (a) to (d) below. a. Iteration and reciprocity is found with telic verbs, while iteration in the form of plurality is common with count nouns. It can also be found with adjectives (especially colour terms) when they are perceived as a ‘Gestalt’.4 Instances are taken from Afrikaans (cf. Conradie 2003; den Besten, Luijks & Roberge 2003) and various creoles (cf. Kouwenberg and La Charité 2001): Afrikaans: klip-klip, lit. ‘pebble-pebble’ (a game involving the repeated handling of a klip ‘pebble’) Afrikaans: rye-rye tente ‘rows and rows of tents’ (i.e. manifold rows of tents)

.â•… Bybee et al. (1994:â•›166ff.) consider ‘iteration’ to be the iconic basis for reduplication, while the sense of ‘continuative’ is seen as a further development on a cline. I believe that both meanings are there from the beginning and not a diachronic development, and that the particular meaning involved depends on the subcategory of noun, verb, and adjective. .â•… Lass (1994:â•›165) mentions some interesting examples of reduplication with plural marking in the verb for ‘do’ in Gothic. He notes that reduplicated forms only occur in the dual and the plural but not in the singular, giving the first person forms: ida ‘I did’, idedu ‘we (both) did’, idedum ‘we (all) did’.

 Olga Fischer

Berbice Dutch: kopu-kopu ‘to buy sth., each by different persons’ (i.e. several acts of buying) Jamaican Creole: yala-yala ‘yellow-yellow’ (yellow-spotted, but see also Section 4)

b. Intensification/augmentation or growth occurs typically in gradable adjectives and mass nouns, but may also be found with atelic verbs: Afrikaans: die vinnige speletjie vir slim-slim mense, ‘the fast game for clever-clever people’ (very clever people) Jamaican Creole: trash-trash ‘a large quantity of trash’ Jamaican Creole: laaf-laaf ‘to laugh a lot’ (but cf. laafi-laafi in Section 4 below)

c. When a stative (intransitive) verb or adjective is reduplicated, it may also indicate repetition, but then in the form of discontinuity or interruption, i.e. the activity occurs by leaps and bounds: Afrikaans: Die stoet beweeg staan-staan, ‘The procession moves stand-stand’ (i.e. the procession is frequently halted) Jamaican creole: luk-luk ‘to look now and again’ Berbice Dutch: šiki-šiki ‘now and then, or regularly, ill’, ‘sickly’

d. With atelic intransitive activity verbs, continuity or durative aspect may be expressed: Afrikaans: Hulle sing swem-swem. ‘They sing swim-swim’ (They sing while swimming) Jamaican Creole: luk-luk ‘to keep looking’ (and cf. (c) above)

The examples show that reduplication is a living, creative phenomenon, which can be used in different situations, with different semantic-pragmatic functions, often depending heavily on the context. Kouwenberg and La Charité (2001:â•›66; 2003:â•›8) Â�provide useful overviews of the various functions of transparent reduplication in Â�English-based Creoles.

4.╅Types of lexicalized reduplication generally considered non-transparent In the first category given below (1), the function of reduplication is again dependent to some extent on the part of speech and the subclassification within the part of speech. The other categories (2) to (7) are more idiosyncratic and mostly occur with only one part of speech. 1. A sense of approximation in terms of similarity, of hesitation and attenuation can be found with inherently durative (stative) and with atelic verbs, with �inherently



Cognitive iconic grounding of reduplication in language 

stative adjectives, and with abstract and non-countable nouns. The examples below are again from Conradie (2003) and Kouwenberg and La Charité (2001), unless otherwise indicated: Afrikaans: Desperaat voel-voel sy weer aan die diamantring, ‘Desperately feel-feel she again on the diamond-ring’ (Once again she desperately fingers the diamond ring) Jamaican Creole: yala ‘yellow’ > yala-yala ‘yellowish’ (cf. Section 3)5 Jamaican Creole: laaf ‘laugh’ > laafi-laafi ‘inclined to laughter’ Sranan: ferfi ‘to paint’ > ferfi-ferfi ‘to paint a bit’ Ndjuka: lepi ‘(to be) ripe’ > lepi-lepi ‘(to be) not quite ripe’ (Keenan & Polinsky 1998:â•›571) Malagasy: ló ‘rotten’ > lòló ‘somewhat rotten’

2. Reduplication is also found to indicate a change in lexical category, where in other languages, such as English, conversion or (affixal) derivation is usually employed. Mostly this involves verbs turned into nouns and verbs used as adjectives. The expressive function of the reduplication is one of result or resulting state, or instruments connected with the activity expressed by the verb. The examples below and the descriptions of the reduplicative functions are taken from Kouwenberg and La Charité (2001:â•›75, 76). Note that Kouwenberg and La Charité’s ‘cause’, ‘cognate object’, and ‘abstract noun’ functions, can also be categorized as ‘result’, while a bloodsucker (chupa-chupa), such as a leech, can be interpreted as an instrument for sucking blood. Mo-mói is a little more complicated; more context is needed to decide whether it could be result or instrument. Jamaican Creole: kriep ‘to scrape’ > kriep-kriep ‘scrapings’ (Verb>Noun, Result) Jamaican Creole: ich ‘to itch’ > ich-ich ‘dry rash’ (Cause) Ndjuka: mói ‘to be beautiful’ > mo-mói ‘pretty thing, gift’ (External argument) Papiamentu: chupa ‘to suck’ > chupa-chupa ‘bloodsucker’ (External argument) Ndjuka: fon ‘to beat’ > fon-fon ‘beating’ (Cognate object) Papiamentu: tembla ‘to shiver’ > tembla-tembla ‘shivers’ (Abstract noun) Saramaccan: nai ‘to sew’ > nai-nai ‘needle’ (Instrument) Jamaican Creole: brok ‘to break’ > brok-brok ‘broken’ (Verb>Adjective, Result)

.â•… Gooden (2003:â•›95) shows that forms such as yala-yala, which can mean both ‘very yellow’ and ‘yellowish’, are distinguished in Jamaican Creole by pitch differences. When used as an intensifier, the first syllable in the reduplication is given extra prominence by using higher pitch, whole in the approximative sense the reduplication has “a level tone melody”. Â�Something similar seems to be the case in Mauritian Creole [MC]. Baker (2003:â•›217) concludes that Â�“reduplication with augmentative effect in MC requires, in each case, that the first Â�occurrence of the reduplicated morpheme is stressed”, while “reduplication of verbs and postposed Â�adjectives with attenuative effect in MC both have the same stress pattern: first occurrence of morpheme unstressed, second occurrence stressed”.

 Olga Fischer

3. Reduplication occurs in Afrikaans in names given to children’s games, but seems rare elsewhere. Kuipers (1967) gives one example from Squamish (a language spoken in British Columbia), and Grant (2003) notes some in Mindanao Chabacano (a variety of Philippine Creole Spanish). The reduplication here is possibly related to ‘repetition’ (cf. den Besten et al. 2003:â•›276) or ‘affectivity’ (cf. baby-talk). Interesting in this respect is the fact that in Dutch such names for children’s play are usually in the diminutive, expressed by the suffix -(t)je (which may also be related to baby-talk). In fact the first example from Afrikaans below contains reduplication as well as a diminutive (the suffix -ie). Grant (2003) has noted that the Mindanao Chabacano instances may be related to affectiveness, but also to attenuation because they can be extended with the Philippine derived derivational prefix tampa-, which means ‘pretend’. Plurality may also be involved, at least in Afrikaans: Botha (quoted in den Besten et al. 2003:â•›276) noted that tol-tol (a game in which a top is used) can only be used when there are two agents present (Kom ons speel tol-tol ‘Come let’s play with our tops’, and not *Hy speel tol-tol ‘He plays with his top). Some examples (taken from Conradie 2003; den Besten et al. 2003; Kuipers 1967; Grant 2003) are: Afrikaans: huisie-huisie ‘house[DIM]-house[DIM]’, to play house (cf. Dutch: vader-en-moedertje ‘father and mother+[DIM)’) Afrikaans: aan-aan ‘on-on’, the game of tag (cf. Dutch: krijgertje, related to krijgen ‘to get’) Mindanao Chabacano: pÀdre-pádre ‘to play at fathers’ Mindanao Chabacano: kÀrro-kárro ‘a toy cart’ Squamisch: kai-kai ‘to play hide and seek’

4. In Section 2 type (iii), reference was made to the use of partial reduplication to express perfective aspect in certain Indo-European languages. This was widespread in Sanskrit, Gothic and Greek (see examples in the relevant sections). Perfectivity can be related semantically to the notion of result, as found in (2) above. The partial form of reduplication may be the result of phonetic reduction via a process of grammaticalization (cf. Bybee, Perkins & Pagliuca 1994:â•›167). 5. Sometimes the reduplicated form carries a pejorative meaning. Keenan and Â�Polinsky (1998:â•›571) note for Malagasy, an Austronesian language that “[a]pplied to nouns… reduplication often has a pejorative effect”, as in their Â�example Â�latábatra ‘table’ > latÀbatábatra ‘something that only approximates a table’. Troyer (2008:â•›9) notes for Luwanga (a Bantu language) that reduplication of a verb “often adds a pejorative connotation, depending on the context” and that “[s]imilar meanings given by reduplication have been documented in a closely related Bantu language, Lusaamia”, e.g. ol-ir-a ‘tell’> ol-ir-a ol-ir-a ‘to tell around’, i.e. to tell tales or to gossip. Another example is from Kamtok: ben ‘bend’, benben ‘crooked’ (from Todd 1990).



Cognitive iconic grounding of reduplication in language 

6. Reduplication is also found occasionally to distinguish between lexical items which have become homonyms in creoles due to the loss of consonant clusters and marked consonants (which still distinguished them in the lexifier language): Krio: was ‘wash’, was-was ‘wasp’; san ‘sun’, san-san ‘sand’ Tok Pisin: pis ‘fish’, pispis ‘to urinate’; sip ‘ship’, sipsip ‘sheep’

(Todd 1990) (Aitchison 1994)

7. Hypocoristic usage is clearly not transparent in that it cannot be related in any sense to ‘increase’ (for examples, see Section 2 (vi) above).

5.â•…Is there iconicity in the non-transparent cases? The majority of linguists whose work I have consulted are of the opinion, if indeed they discuss iconicity in relation to reduplication, that iconicity plays no role in the cases discussed in Section 4. Kouwenberg (2003:â•›5) writes “that reduplication can be used with a range of non-iconic readings”, citing the following as non-iconic: approximative, diminutive, pejorative and category changing (deverbal adjectives and nouns) (similarly, Bakker 2003:â•›79; Baker 2003:â•›217; Baptista 2003:â•›179). The reason for the absence of iconicity most often mentioned is the fact that, if there is iconicity, it seems to be used for completely opposite purposes. Thus, for Bakker and Parkvall (2005:â•›517) attenuating reduplication is non-iconic, and they also suggest that reduplicated past participles must be substrate-derived, thereby dismissing the possibility of an internal, iconic source. More generally they remark (2005:â•›518): “Yet, the fact that the various kinds of reduplicative processes manifested in creoles are so diverse in meaning, coupled with the obvious substratal source for some of them, makes this [i.e. an independent, iconic development] less likely”. Dressler et al. (2005:â•›467), call it paradoxical that reduplication may have “hypocoristic, diminutive and attenuative meanings […] when plurality, distribution, iterativity, extension, intensification are considered to be iconic meanings of reduplication”.6 One can indeed observe the following totally contrasting functions between the examples given in Section 3 and those in Section 4: continuity vs. discontinuity augmentation vs. diminution/attenuation intensification vs. hesitation/approximation continuity vs. result (or durative vs perfective)

.â•… It is interesting to note that Kouwenberg and La Charité (2005) differ from Kouwenberg and La Charité (2001, 2003), in that in the later article they suggest that the approximative and diminutive functions, which are very common in creoles, are iconic after all, via semantic extension. I assume that the other common opaque functions they recognized in (2001, 2003) remain opaque for them since they do not discuss them in 2005.

 Olga Fischer

The question thus automatically arises of whether there can be a common source when the employment of iconicity leads to such opposite usages. In what follows I would like to argue that this is indeed possible provided one takes a number of aspects concerning the working of iconicity into consideration.7 Concerning a common iconic source, I believe that the principle ‘more of the same’ – referred to by, among others, Kouwenberg and La Charité (2001) – is inadequate because it is too much related to the literal idea of plurality and iteration. First of all, it is clear, as already briefly mentioned above, that ‘more’ can refer to quantity on a horizontal as well as a vertical plane; in the latter case it indicates an increase in quality rather than quantity since the increase is not a consecutive one in time or place (i.e. iteration or plurality/dispersal) but a simultaneous one at one point in time or in one place (i.e. intensification, augmentation). Secondly, the iconic notion of ‘increase’ is a more abstract, diagrammatic type of iconicity since the reduplicated form in itself does not resemble the object it refers to in the outside world in some direct way. It is the doubling of the stem, not the stem itself that leads to the recognition of a similarity. It differs from the direct type of iconicity, imagic iconicity, in other words, in that it does not capture in its phonetic form a similarity with the object referred to. To make this diagrammatic/imagic difference more visual, compare the following two pairs:

a. tree + frog

b. tree + frog

Figure 1.

The first pattern (a) reflects the isomorphism of a combination of two of the same things; there is no iconic (=imagic) relation between a single and a single

.â•… The existence of different functions is no doubt also related to suprasegmental features involved in reduplication (cf. Note 5 above). It is not surprising in itself that tone or stress influences meaning, leading to the repeated form being variously interpreted as ‘boredom’, ‘intensity’ etc. (see Section 2). Overall, reduplication, especially of the non-grammaticalized or non-lexicalized type is highly sensitive to the pragmatic context, of which intonation forms a part. However, such variation need not deter us from finding a common source as long as the various functions for which reduplication is used can be explained from both the phonetic and the suprasegmental form in which it first occurs. The difficulty of representing suprasegmental information in the written mode may be one of the reasons why it seems less common in written, standardized languages. See further, Note 17 below.



Cognitive iconic grounding of reduplication in language 

phonetic sign [tri:] ‘tree’. In (b), this sameness does not exist, and therefore the iconic notion of increase is not present: (b) could be compared to the formation of a compound. If the morpheme ‘frog’ would occur with every stem, then it could again be a diagrammatic combination of ‘tree’ + generalized morpheme, Â�indicating plural, but this is much less transparent and has to be learned (this is the case in languages where there is one arbitrary morpheme to indicate plural, such as -s in English). (a), on the other hand, even though diagrammatic and therefore more abstract, can be understood without knowledge of the particular, arbitrary, Â�language conventions. From this, it is clear that the isomorphism (and thus the ‘increase’) only works if the stem itself repeated, not when some other morpheme, such as English -s, is added to the stem. Because the iconicity in Figure 1(a) is diagrammatic and not imagic, the meaning of the reduplicated form depends fully on the meaning of the single element it is based on. Not surprisingly, therefore, the eventual meaning of the reduplication is the result of a combination of the particular meaning of the verb, noun, adjective, etc. in question (a meaning that arises from both its semantic-pragmatic and its grammatical meaning/function) and the general notion of doubling or ‘increase’. In other words, the reduplication may represent a variety of meanings. When the second morpheme has been replaced by a generalized conventional one, we will need a different morpheme for each specific function, e.g. in English -s for plural, -ish for adjectives to indicate attenuation, be -ing to indicate aspect with verbs, etc. Reduplication as we find it in babytalk or motherese, on the other hand, can be seen as an instance of imagic iconicity; that is, it constitutes a different iconic root, which reflects the onomatopoeic imitation of the actual CV-CV syllabic babbling sounds made by children. Even more important when considering possible iconic roots, is the fact that the more concrete iconic uses can also shift metaphorically and metonymically and thus produce new functions of usage. Metaphorical and metonymic shifts are the two most common and natural paths of semantic change in language. Both are the result of, or based on, the deeper cognitive principle of analogy, which involves both iconic and indexical thinking (Anttila 2003), and which is an essential mechanism in both language acquisition and (semantic) change (Tomasello 2003; Fischer 2007). If metaphorical and metonymic shifts are common in all language use, and grounded cognitively, then forms of reduplication shifted by these means can still be considered to be related to a common iconic source, whether of the imagic or the diagrammatic type. Thus, the concrete sense of dispersal or the spread of objects over a larger plane, is metaphorically connected with the idea of something being spread out or scattered in a mental sense, and hence with an event that is diminished or attenuated, at the

 Olga Fischer

same time suggesting a lack of control. Both of these ideas (diminished effect, lack of control) can easily be seen as negative in certain contexts, as deserving contempt, by means of indexical thinking. Similarly, the iteration of an activity is metaphorically linked to intensity especially when one act follows hard upon the other. Indexically, it is related to the eventual result of the activity, to the product of the iteration, leading to the use of reduplication to indicate perfective or completive aspect. An indexical link also makes it understandable that the reduplicated form can indicate the instrument with which the activity is carried out. Further examples of such shifts could be the use of reduplication in children’s games because these games are indexically related to what children typically do and thus to children in general, just like reduplication in motherese may be a mimicking of babies’ babbling. Reduplication may then also be shifted elsewhere analogically, for instance as an effective means of distinguishing homonyms when one of the existing functions of reduplication happens to apply. It is of interest (but more investigation is necessary) that in the examples given in Section 4 (6), the reduplicated forms either suggest plural (sipsip ‘sheep’ – sheep usually appear to us in flocks, compared to sip ‘ship’; sansan ‘sand’, a mass noun consisting of grains of sand’, while san ‘sun’ is bound to be singular; waswas ‘wasp’, these usually appear in droves, unlike the idea of was ‘wash’), or continuity (pispis ‘to pee’ versus pis ‘fish’), or a combination of functions (both pispis and waswas could be seen as onomatopoeic and perhaps also pejorative). The question that needs to be raised in these homonymic cases is why the choice was made this way and not the other way round: why not sansan for ‘sun’ and san for ‘sand’ etc.? Regier (1998) lists three root nodes for reduplication. Each root then has further developments by means of semantic extension: Reduplication Sound Meaning

Baby

small

attenuation

affection

contempt

Repetition

continuation

Plural

intensity

spread-scatter

completion lack of control non-uniformity

Figure 2.╇ Three root nodes for reduplication, from Regier (1998:╛888)

I would suggest two roots instead of three, with the diagrammatic iconic root ‘increase in quantity’ subdivided into vertical and horizontal increase as indicated



Cognitive iconic grounding of reduplication in language 

Â� earlier. The imagic ‘baby’ root is initially separate but has joined the diagrammatic one because it shares meaning extensions or associations with this root, that is via diminution and possibly contempt (cf. the meaning of ‘childish’, ‘babyish’): Reduplication (ICONIC) Imagic: baby babbling

Diagrammatic: increase in quantity

Horizontal increase (across space)

Vertical increase (in same space)

repetition/iterative aspect plural distribution /spread repetition /habitual via semantic shift: habitual > bored > pejorative spread > attenuation/ diminution spread > continuation/durative spread > scattered > lack of control scattered > pejorative > contempt distribution > reciprocal repetition > children’s games? repetition > stuttering > contempt

intensity augmentation increase in tone via semantic shift: completion/result perfective aspect instrument intensity > emotional intensity > expressive

baby talk/motherese via semantic shift: small > diminution small > affection affection/diminution > hypocoristic babbling = stuttering > contempt baby = young > naïve > contempt child > children’s games

Figure 3.╇ The diagrammatic and the imagic iconic root of reduplication

The common aspects or associative links between the two roots have been put in small caps in Figure 3. It shows that the ‘baby’ root is less important, but it may (have) function(ed) as a re-inforcer for these aspects in the main, ‘increase’ root. The idea of contempt may also have been fed by another more general association some (mainly European, cf. Note 9) language users have with reduplication an sich, namely that it represents a primitive form of repetition (which speakers generally avoid, just as they avoid echo words, cf. Aitchison 1994),8 thus providing metonymically an overall negative feeling of naïvity and hence contempt.9

.╅ Aitchison (1994:╛32) notes that repetition sounds like stuttering, which is hardly ever seen positively. .╅ The negative feeling associated with reduplication evinces a form of language ideology prevalent among speakers of Standard Average European. Nuckolls (1995:╛145) describes with

 Olga Fischer

In order to show that these may indeed be common pathways converting a concrete sense of ‘increase’ in a metaphorical or metonymic way from an iconic source, it pays to look at two other linguistic areas, which have certain traits or semantic aspects in common with the reduplications found here, but which are also, in different ways, independent of the above cases of reduplication, either because they are from a different language channel (signed rather than spoken language) or because they make use of a conventional arbitrary means (an affix) to form new words.

5.1â•… E  vidence for cognitive connections between transparent and opaque reduplication from sign language The first type of evidence I want to discuss comes from signed languages. In order to form new words, sign language frequently makes use of the modulation of the sign itself rather than adding extra morphemes (derivational affixes) or using conversion, as is the rule in most spoken languages (with the exception of isolating languages).10 Such modulation could be compared to Ablaut in spoken languages, where also the stem of a morpheme is modified. A sign can be modulated in a variety of ways (cf. Klima & Bellugi 1979:â•›245ff.): i. ii iii. iv.

by repeating the sign, i.e. the hand movement by making the hand movement less intensive by not fully completing the hand movement by the speed with which the hand movement is executed

reference to the use of ideophones (which include reduplications) that these are “widely used by people whose languages have not been written down”, and that they “create a type of concreteness and imageability that is functionally analogous to visual modes of expression”. She argues that the development of literacy has led to a discursive mode that is detached and impersonal, “insulated from affective connotations and so more congenial to the communication of abstract logic” (ibid.: 146). In Nuckolls (2003), three more factors are given that explain the “ideophonically impoverished” (132) nature of European languages, namely urbanization, westernization and globalization. .â•… Pfau and Steinbach (2005:â•›581), however, show that conversion can be used in German Sign Language but its use is restricted to special cases. They discuss the case of zero plurals. Normally the plural is indicated by reduplication but when a sign is already inherently complex (i.e. reduplicated for other reasons) then it cannot be pluralized by repetition, and a zero plural gets used instead. Johnston and Schembri (2007:â•›139) note a similar constraint in Australian Sign Language. Also in spoken languages, we often find a more general constraint against the reduplication of longer, bisyllabic words. Gooden (2003:â•›96) notes this with respect to Jamaican Creole, and Haugen (2005:â•›318, 322) shows that the usual pattern of reduplication in Yaqui and other Uto-Aztecan languages is “light syllable reduplication”, where only the first syllable of the base is reduplicated, e.g. chike ‘to comb one’s hair’ becomes chichike and not *chikechike (when used as an habitual).



Cognitive iconic grounding of reduplication in language 

v. by the place where the signing takes place vi. by combinations of (i) to (v) The hand movement itself or the repetition (reduplication) of the movement can also be done in a variety of ways (cf. Klima & Bellugi 1979:â•›256ff.; Deuchar 1984:â•›87ff.; Sandler 1990; Pfau & Steinbach 2005, Wilbur 2005; Johnston & Schembri 2007:â•›126ff.): a. slow repetition ( this expresses continuity in atelic verbs; or iterative aspect in punctual verbs) (cf. Supalla & Newport 1978:â•›112) b. fast repetition in combination with punctual verbs (this expresses habitual aspect) (cf. Norton-Warren 1978:â•›135) c. lax hand movement and reduction of the repetition (conveying approximation) d. repetition of a sign for a verb in a contained manner (i.e. the repeated movement is smaller than that of the first movement (this leads to nominalization, e.g. to sweep > broom, to sit > chair, to compare > comparison, to read> reading), indicating instruments as well as abstract nouns) (cf. Brentari 2002:â•›48–50; Â�Johnston & Schembri 2007:â•›126). Instead of reduplication conversion may also be used (cf. Johnston & Schembri 2007:â•›127, and Note 10 above) e. repetition is sometimes used to express plural by means of a replacement of the movement in space in order to express a scattered plural (e.g. sign for car > car car ‘cars here and there’) The interesting fact that emerges from looking at signed languages is that repetition of the sign is a very common means of word formation, presumably because it is an easier and more natural (= iconic) means than the use of an extra morpheme (an affix), which is common in many spoken languages.11 More importantly, in connection with iconicity is the fact that the different types of repetition correspond to the meanings expressed by reduplication in spoken languages, whereby (c) and (d) are particularly interesting because they represent the opaque cases of reduplication, mentioned in Section 4. Pfau and Steinbach (2005:â•›569) note that in German Sign Language reduplication is common for aspectual, plural and reciprocal marking, but that in other signed languages it is also used for the opaque type of noun-verb conversion. Wilbur (2005:â•›597) has found that typically a long movement is used for a verb (e.g. ‘to sit’) while similar but shorter repeated movements are used to indicate ‘chair’; in a similar way ‘reading’ may be derived from ‘to read’.

.â•… Another reason may be that signed languages are generally young and/or show less Â�historical development because they are not normally, like spoken languages, transmitted Â�directly from generation to generation. Hence they show a smaller amount of ongoing Â�conventionalization due to economy and erosion. In this sense they may reflect creoles, which are also seen as younger languages (see Section 5.3).

 Olga Fischer

The fact that quite a few of these uses can be seen to be linked to non-transparent cases of reduplication in spoken language suggests that there may indeed be a cognitive link between all these uses. In this respect it is very interesting that Nuckolls (1995:â•›145) compares ideophones to “visual modes of expression”. What the ideophonic languages and signed languages have in common is their ‘visuality’, they are both immediate forms of communication; they lack the “detached truth” that is Â�common to literate standards because they communicate “by imitation or iconicity, rather than by arbitrariness” (ibid.: 147) (cf. Note 9 above).

5.2â•… E  vidence for cognitive connections between transparent and opaque reduplication from the use of the ge-prefix in Germanic languages The second piece of evidence comes from the field of morphology involving the use of an affix that combines some of the same functions shown by reduplicated forms. The affix itself, ga- in Gothic, which developed into ge- in Old English, Modern Dutch and German is arbitrary in form (its etymology is still disputed). This makes it all the more interesting that it covers almost the same functions as a reduplicated form. Again the question arises whether there may be a common cognitive source behind both these linguistic means. The Dutch/German prefix ge- is used: i. generally in past participles (result, perfective) ii. in count nouns expressing result (often derived from past participles): gezegde/ Gesagte ‘what has been said, a saying’, gewoonte/Gebrauch ‘what has been done before, a custom’, gestalte/Gestalt ‘appearance’, gesprek/Gespräch ‘a talk’, gezicht/ Gesicht ‘face’, geboorte/Geburt ‘birth’/in German also ‘child’; gevolg/Gefolge ‘result’ iii. in collectives formed from nouns: gebergte/Gebirge ‘mountains’, gezusters/ Geschwister ‘sisters’/’brother(s) and sister(s)’, getij/Gezeit ‘tides’, gestoelte/Gestühl ‘chairs’, gesternte/Gestirn ‘stars’, gesteente/Gestein ‘stones’; from adjectives: gezamenlijk/Gesamt ‘together’; and from verbs: gewas/Gewächs ‘vegetation, plants’, Getreide ‘corn’, gebak/Gebäck ‘pieces of cake’, gedrieën, gevieren ‘with three/four people together’12 iv. repetition: getijde ‘recurring time’, geschreeuw/Geschrei ‘cries, acts of crying’, gezeur ‘repeated bothering’ v. in abstract nouns derived from verbs: gezang/Gesang ‘singing’, gezicht ‘sight’, gehoor/Gehör ‘hearing’ (in Dutch also ‘audience’), gezag/Gewalt ‘authority’ (from Dutch zeggen ‘to say’), gewicht ‘weight, importance’, geweld/Gewalt ‘force’

.â•… Ladham et al. (2003:â•›170) note that reduplicated numerals, which often have a distributive sense in creoles, can also have a “closed class” interpretation, i.e. they are seen as a collective, in Gulf of Guinea Creoles.



Cognitive iconic grounding of reduplication in language 

vi. in instrumental nouns: gezant/Gesandte(r) ‘messenger’ (from zenden/senden ‘to send’), gewicht/Gewicht ‘a weight’ (piece of metal used in weighing), getouw ‘loom’(from touwen ‘to work on’), Gebläse ‘blow-instrument, compressor’, Gefährte ‘companion’ vii. in adjectives expressing a result: genegen/geneigt ‘inclined’(from nijgen/neigen), geschikt ‘apt’, geletterd ‘lettered’, ‘educated’, geleerd/gelehrt ‘learned’, geruit/gewürfelt ‘checkered’, gelakt ‘varnished’ viii. with a pejorative sense: gepeupel ‘mob’, gespuis/Gesindel ‘rabble, scum’, gedoe/ Getue ‘a to-do’, geklets/Geschwätz ‘twaddle’, gelazer ‘bother’, gezeur/Getratsch ‘drivel, bothersome behaviour’ It is of interest that the Dutch Etymologisch Woordenboek writes: “ge-: oorspronkelijke betekenis wrsch. ‘met, samen’” (ge- original meaning probably ‘with, together’), indicating some form of plurality. Plurality, one of the transparent functions of reduplication is clearly present in (iii) above, and another transparent function, ‘repetition’ can be found in (iv). The opaque functions are even more prominent among the ge- words. Result is present in (ii) and (vii), and to some extent also in (v), instruments derived from verbs can be found in (vi), and pejorative meaning in (viii). Another interesting cognitive link between an affix and reduplication concerns the diminutive affix used in Dutch words to refer to children’s games and the reduplication used in Afrikaans. However, this is more likely accidental as I have found no further evidence in other languages. Such a link would constitute the use of both reduplication and diminutives in baby-talk, or the imitation of adults of baby-talk.

5.3â•… Why transparency may be lost: Grammaticalization and lexicalization In the two previous subsections, I have tried to show that the range of functions expressed by reduplication in spoken languages resembles a similar range of functions represented by one and the same form (repetition of hand movement and the prefix ge-) in signed languages and Germanic languages respectively, suggesting that this points to a common iconic source for all of these forms/functions. It is also possible to approach the idea of a common source from another, diachronic, direction, and ask the question why it is that the transparency of the reduplicated sign is quite clear in some cases and not in others. I would therefore like to take a look at a number of factors that may obscure or obliterate transparency. In the reduplication literature, iconicity is often related to productivity, the idea being that a linguistic sign which is easy to use and easy to understand – a motivated transparent sign – is more likely to be productive. It becomes as it were rule-like. Thus, on the contrary, those forms of reduplication that are frequent and regular are also seen as more likely to be iconic, while those that are rare and/or irregular are not (these are often attributed to substratum influence, see for example Parkvall 2003:â•›29–30; Singler 2003:â•›158). Kouwenberg and La Charité (2003:â•›10) write: “it is clear that semantic and

 Olga Fischer

categorial changes – which, in many cases, have to be determined on a language-specific basis – as well as formal idiosyncracies and limited productivity in CC [Caribbean Creole] reduplications are typical only of reduplications which are non-iconic”.13 This lack of productivity and rule-like behaviour does not necessarily mean, however, that there cannot be an iconic source; it may well be due to the process of lexicalization. When a reduplicated item becomes lexicalized (becomes a single word and/ or acquires a more idiosyncratic meaning), it begins to fill a niche in the lexicon. Â�Productive, transparent reduplications often still depend to some extent on their context (including the suprasegmental level). We have seen above (Section 3(a), and Section 4(1)) that some reduplicated forms indeed have more than one meaning, for example Jamaican yala-yala can be interpreted sometimes as ‘spots of yellow’ and sometimes as ‘yellowish’. When an item lexicalizes, it semanticizes (i.e. it becomes less dependent on context), which may cause one of the (still more) pragmatic Â�meanings to get lost depending on the needs of the lexicon and/or the system of grammar. If the Â�system of grammar has an affix available or develops one to express one of these senses, or if another word becomes available (through contact or internal word Â�formation), then the reduplicated form can be restricted to the second, remaining sense. For instance, in the case of a reduplicated form such as ‘sweep-sweep’ for ‘broom’, if there is already a word like ‘broom’ from another source in the lexicon, then there is no need for a reduplicated form derived from ‘sweep’. In other words, within the lexicon, the pattern is not ‘across the board’, therefore it is less easy to spot. Furthermore, the lexicon is generally seen as a collection of loose items without much internal patterning, especially in cases where no morphological rules are involved. We may find a rule governing the use of derivational affixes, like English un-, re-, -ness, -ment etc. We may also find a rule operating in ‘conversion’, such as a zero-affix being ‘added’. Still, it is more unusual for linguists to see a rule operating when there is no common/zero affix. This is the case in reduplication, because the reduplicated form involves each time an individual, hence a different morpheme, even though it may express the same function in each new formation. In other words, although reduplication is iconic, it does not obey the more abstract iconic principle of isomorphism. There is no overall unitary relation between one form and one meaning; instead, as I have shown in Figure 1, the relation depends on a diagram of two forms, and its function may be different in each case. The meaning of the reduplication thus varies because it depends on the type of lexical element (the subcategory of noun, verb etc.) that is reduplicated. This may lead to ambiguity and confusion. To have one single affix (e.g. the English -(e)s morpheme to express plural, or Old English ge- to

.â•… Kouwenberg and La Charité (2003:â•›8–9) characterize the transparent reduplications given above in Section 3 as productive, while the opaque instances in Section 4 are not productive.



Cognitive iconic grounding of reduplication in language 

express perfective aspect) with the same basic meaning all the time, is probably more economic, and more transparent even though the affix itself may be arbitrary. The semantic relation between a verb and reduplicated result/instrument noun (e.g. between ‘sweep’ and ‘sweep-sweep’=‘broom’) is likewise difficult to capture as a ‘pattern’, especially when compared to most other languages where these two forms usually constitute two totally different lexical items. Hence there is a very strong tendency to consider these reduplications as isolates, as not part of a common pattern either. This is especially true when the base-form from which the reduplication was formed disappears, or when the reduplicated form, now lexicalized, undergoes phonetic or semantic erosion/change. In both cases the transparency disappears because the form can no longer be seen as motivated by the base. Another diachronic development that may obscure transparency concerns grammaticalization. When reduplication is used to express a function-like completive aspect (as in the Indo-European perfect), then the reduplicated form within the verbal system may begin to grammaticalize, because aspect is a general function which is typical of verbs and may be expressed on a great many verbs (when they are not already inherently perfective). The reduplicated form then becomes a highly regular, wide-spread pattern (i.e. productive) with only one function. Thus, although the reduplicated verb itself uses a different form, the function is always the same. This makes it different from lexicalization, where both form and function were different. This may then have two results, both of which have an enormous impact on transparency. The reduplicated form will become phonetically reduced when it grammaticalizes, that is when it no longer resembles its base. It becomes an arbitrary affix. This makes a second step possible: in grammaticalization processes there is usually an ‘increase in paradigmaticity’. This means that only one of the possible reduced affixes remains, usually the most frequent one. In other words, one half of the reduplication, reduced to an affix, representing the most frequent sign may replace the less frequent affixes by means of analogical extension, so that we end up with a single, opaque perfective aspect marker that comes to be used with all items in the paradigm. In the end it may even become a general marker of the past participle, so that it can even be used with verbs that were already inherently perfective.14

.â•… Kulikov (2005) notes that most of the reduplication in the oldest Indo-European has become non-transparent and reduced. He only remarks on iconicity and more complete reduplication in the intensive meaning (441). It could well be the case that an intensive function is much less likely to grammaticalize to an inflection because it has a less frequent occurrence. I know of no languages with an intensifying inflectional affix in either the verbal or the nominal system. Where such an affix exists, it is usually derivational (e.g. a-, be-, for- in Old English, per- in Latin) or expressed by a particle.

 Olga Fischer

One aspect that should be taken into consideration in the grammaticalization or lexicalization of reduplicated forms, is that the process itself (the way it materializes) depends on the overall synchronic system of the language and its history. The introduction of loanwords or loan-affixes may obviate the need for lexicalization; in that case, the reduplicated forms may simply get lost. Whether or not reduplicated items come to be used lexically or grammatically as functional elements depends to a large extent on the type of language; it depends on whether we are dealing with an inflecting, agglutinative, polysynthetic, or isolating language, and on the possible syllable structures available in a language. In the case of isolating languages, it is quite possible that reduplicated forms do not grammaticalize in the same way as in inflecting languages, that is to say that full reduplication may remain formally, but this form need not be transparent any longer because functionally it may have grammaticalized.15 Such a different pathway of grammaticalization may have happened in Gungbe. Aboh (2007) describes an interesting case of reduplication in Gungbe, an isolating language belonging to the African Gbe group of languages. He proposes that the reduplication of certain adjectives “serves a syntactic purpose and has no semantic effect” (2007:12). He notes that Gungbe attributive adjectives are morphosyntactically distinguished from predicative ones in that only the attributive adjectives reduplicate. The only exception is attributive adjectives that are non-gradable, such as colour terms; they lack reduplication. In addition, reduplication is also found in nominalizations to express result and in verbs to express the progressive (durative aspect). With intransitive verbs, reduplication is even obligatory. It is interesting to observe that all these three functions can be related to iconicity as discussed above. Attributive adjectives, unlike predicative ones, express a more permanent state, a result, a more inherent feature of the noun it accompanies (cf. Bolinger 1952, 1967; Vincent 1988; Fischer 2001).16 To express this permanency or result, reduplication could then be used (as probably happened in Gungbe), except in non-gradable adjectives because they are inherently permanent. Interestingly, Klima and Bellugi (1979:â•›252) observe for signed languages that signs referring to inherent characteristics (e.g. ‘ugly’, ‘tall’) do not undergo modulation (i.e. repetition) while signs meaning ‘quiet’ or ‘rough’ may undergo modulation, whereby they come to have more Â�inherent

.╅ In an isolating language, the use of affixes that are regularly used in non-isolating languages to express modulations on the verb, noun, etc. is restricted. We have seen above that a similar situation exists in signed languages, which use reduplication rather than affixes. Bergman and Dahl (1994:╛41; 415) show that the functions of affixes and reduplication are quite similar, and that inflection-less languages correlate with the presence of ideophonic morphology. .╅ In discourse the attributive adjective usually expresses old information, the predicative one, new information.



Cognitive iconic grounding of reduplication in language 

meaning: ‘taciturn’ and ‘cruel’ respectively. Singler (2003:â•›156) furthermore notes that in Vernacular Liberian English, “a small number of common adjectives […] are regularly doubled, with the doubling conveying little or no emphasis”, and that this doubling is “[o]verwhelmingly […] limited to instances where these adjectives occur attributively, as opposed to predicatively”. He observes the same pattern in other Â�African languages such as Klao and Ewe. Since all or most adjectives can be used attributively as well as predicatively, and since in many other languages no formal distinction exists between attributive and predicative adjectives, the difference could be considered to be semantically meaningless. It is possible that with further grammaticalization, even non-gradable adjectives may begin to reduplicate when used attributively, thus more completely destroying earlier transparency. Something similar seems to be the case with reduplication found in nominalizations (gerunds) and verbs in Gungbe. The original transparency seems to have disappeared in cases where reduplication is blocked for purely syntactic or rhythmical reasons. Aboh notes, for instance, that reduplication with verbs is blocked when some other element (i.e. the direct object with transitive verbs, or an adverbial phrase with intransitives) precedes the verb. The blocking may have started with a prospective aspect marker (Aboh shows [2007:â•›4], that the aspect marker ná always blocks reduplication with both verbs and gerunds) because the meaning of the prospective marker clashes with durative aspect, but it then seems to have spread to other lexical phrases occurring in the same position. Aboh (2007:â•›3) writes, “This position must be filled by an overt phrase otherwise verb reduplication is obligatory”. He concludes therefore that “the reduplicated part of the verb represents an INFL morphology” (2007:â•›5). Here reduplication is no longer semantically motivated, as it must have been in the beginning, but probably the result of a rhythmical pattern that developed and/or the need to fill an otherwise empty position. A further question arises from the above discussion about the role of the lexicon and the grammar: why is reduplication especially common in creoles (cf. Bakker & Parkvall 2005:â•›511, 514), and not in pidgins (cf. Mühlhäusler 2003:â•›239), nor in most languages with a long history of standardization and a written standard?17 In other

.â•… Two very notable exceptions (Herlofsky p.c.) are Japanese and Korean (cf. Sohn 1999), where reduplication is rife even in formal writing. Why this should be so is as yet unclear but one reason may be that Japanese (and Korean to a somewhat lesser extent, see Sohn 1999:â•›161) has a rather simple syllable structure (CV), which is one of the parameters (next to number of vowels and consonants, and the use of tone and stress) that greatly reduce the number of “contrasting syllabic possibilities” (Maddieson 1984:â•›21) which can be used to form words. Maddieson (ibid.: 22) also notes that of the above parameters, syllable structure is the best predictor for the number of permitted syllable types. To this can be added that stress and tone are more difficult to represent in writing. Reduplication, which can be represented in writing,

 Olga Fischer

words, reduplication seems to be especially common when the lexicon and the grammar are being formed, and less so when they are fully formed (as in languages with a written standard, cf. Note 2), or when they are still very primitive or even absent (as in pidgins). Why? Pidgins make do with only few words, fewer phonemes/phoneme clusters, and no word formation to speak of; instead they use polysemy or rather general circumscriptions, like ‘grass belong head’ for ‘hair’ in Tok Pisin. Only when a pidgin becomes a native language, we see a development towards a larger lexicon, more grammar and a more complex phoneme system. It is interesting too, to observe that young creoles are often (still) isolating languages, and if they remain isolating in their further development, reduplication may well remain in its full form. Present-day standard inflectional languages (i.e. originally inflectional languages with a long history and a written standard, such as the languages derived from Indo-European) on the whole lack reduplication Â�possibly because of the language type to which they belong, but perhaps also due to a long Â�process of grammaticalization and lexicalization by which they will have developed an adequate system of word formation and grammatical rules and morphemes through long Â�conventionalization and standardization. Conventions are learned by each new generation as abstract arbitrary forms, and they are enhanced by Â�education and a standard writing style, so that generally there is no need for transparent new words formed by reduplication. In other words, any formerly reduplicated items tend to disappear out of sight (through linguistic loss/reduction and loss of Â�transparency). In these languages, reduplication, or rather repetition, is restricted to the spoken Â�language where these forms may occur for expressive reasons. Whether these reduplicated forms may be the start of a new cycle depends on the further Â�development of the language in question. Another reason for these different developments may be that in the standardization of the European languages (in particular the development of a written language), reduplication comes to be seen as naïve, so that at the moment reduced and arbitrary forms became available, they were preferred and became unlikely to be replaced by new reduplications. Reduplication proper is rare in most western European written languages, even though the repetition of words or phrases is quite common in the spoken varieties, but they do not normally reach a lexicalized stage.18

may help solve this problem (i.e. widen the possibilities for word-building) to some extent, but clearly more research is necessary to find out whether such a relation between (simple) syllable structure and the use of reduplication exists. The fact that both Japanese and Korean use tone to some extent may also be influential (see also below). .â•… This sense of naïvety is probably largely culturally determined, which in turn may be determined by language typology (see the discussion above, and Note 9). Aboh (p.c.) writes



Cognitive iconic grounding of reduplication in language 

There may even be an iconic motivation, but of a more abstract diagrammatic kind, behind the above development of doing away with reduplication (as also noted above). Arbitrary grammaticalized affixes have an advantage in that there is only one pattern and one uniform morpheme. In the reduplicated form, the duplicate resembles the stem so that the external forms of the pattern are all different (since each stem has its own linguistic shape), i.e. there is no isomorphism between the linguistic form (which is different with each base form) and its general function. In the case of a grammatical morpheme, the function is always carried by exactly the same affix attached to the root, which is more abstractly iconic but also completely regular and general. Possibly, this solution is also motivated by economy. Tone, especially lexicalized tone, may also be a factor in the use or non-use of reduplication. Reduplication in creoles often depends for its interpretation on intonation (tone)19 (cf. Note 5 above) and the immediate context. This is also true for repetitions on the phrase or word level in spoken English. For instance, the sense of boredom conveyed by ‘alright, alright’ (cf. Section 2) is mainly present in the intonation and the context; it could also be used as an intensifier in a positive context. Reduplication may therefore simply be too ambiguous for written communication, when the concrete situational context and the intonation (not to mention, facial and other gestures!) are missing. The situation may be different, however, for tone languages where tone is no longer pragmatic but semanticized, part of the lexical system. If speakers of these languages are already used to tone playing a crucial role in the interpretation of signs, they will be more aware of the fact that the interpretation of a reduplicative form may also depend on tone, even in cases (as in written script) where they cannot hear it. Because reduplication is no longer merely expressive in these languages, but a necessary part of the system, its use cannot be avoided in written communication. The absence of reduplication and the use of arbitrary elements to express the functions of reduplication in more formal written language may also account for the fact that the transparent reduplicated or repeated words are often seen as more expressive, more affective, more emotional, when they are used. This would explain their frequent use in hypocorisms and ideophones, and also the pejorative shades of meaning that they may carry (cf. Bergman & Dahl 1994:â•›415; Sohn 1999:â•›102; Bakker 2003:â•›75ff.;

that reduplication is fully accepted in the poetic, ritual, and court language of the Niger Congo languages, which are all very old. Similarly, Sohn (1999:â•›96ff.) notes their use in Korean literary works. It may well be the case that the conventional written form of inflecting languages is mostly to blame for this largely western phenomenon/attitude. .â•… Tone and intonation, although traditionally seen as different (the former operating on the lexical level, the latter on the sentence level), in fact may be seen as belonging to one and the same category typologically, organized on a cline, cf. Zerbian, Downing & Kügler (2009).

 Olga Fischer

Grant 2003:â•›205; Dingemanse, this volume).20 It is also notable that ideophones are mostly used in phrases that are extraneous to the clause; they occur as adverbs and particles more than as central parts of speech such as nouns or verbs (Sohn 1999:â•›102; Dingemanse, this volume). There may also be a relation here with reduplication as found in baby-talk, which in itself is more charged with emotions. However, this is possibly not the only reason why reduplication comes across as more emotional and expressive. I think it is important to remember that many, if not most, of the classic rhetorical devices also use repetition (Müller 2001:â•›311ff.). These have become to a large extent conventional, but they too have started off as expressive devices to strengthen the speech of the orator, to persuade his audience not only in terms of logos (reasoning) and ethos, but also, and most importantly, in terms of pathos. Repetition works like stress: when a person is emotional, he raises his voice, and repeats himself, using the same phrases, and arguments again and again. This also explains why both intensity (or vertical quantity), and repetition via horizontal spread, are expressed by the same means. The link with spoken discourse may also explain why in sign language reduplication continues to play an important role, since the only mode available here is the ‘spoken’ mode. ‘Written’ (i.e. visualized) sign language is resorted to in academic work, but since deaf speakers of signed languages are usually able to read the regular written mode of the language used by regular speakers, a written signed language is slow to develop. It also has to be taken into account that signed language is not normally acquired via a natural generational process, so that conventionalization has less of a chance. To sum up, reduplication of a transparent kind is more likely to appear in young languages (creoles) and in spoken discourse, while it may have also have been better preserved, but with some loss of transparency, in isolating languages and tone languages, and more specifically in isolating languages with a simple syllabic structure.

6.â•…Concluding remarks I have tried to show that it is possible to find a common cognitive source for reduplicated forms provided one takes into account, when interpreting the form as ‘increase’ of some sort: (i) the role of context (in the widest possible sense, including language typology) and intonation, (ii) analogical thinking leading to metaphorical and metonymic shifts, (iii) the diachronic pathways that reduplicated forms may travel

.â•… Conradie (2003) remarks that he found most of his examples of reduplication in Â�Afrikaans in spoken sources or dialogues in novels. For some notable exceptions, see Note 17.



Cognitive iconic grounding of reduplication in language 

through afterwards by means of lexicalization and grammaticalization. The only form of reduplication that does not quite fit is the one found in baby-talk. This may be due to the fact that children start off with a simple CV structure, with a small inventory of sounds, and the tendency towards sound harmony, which all leads to the production of words like ‘papa’ and ‘mama’. But I have also shown that even this could become linked to the emotive feelings that reduplication (cf. the first branch in Figure 3) also conveys so that in fact the two sources mingle or have mingled imperceptibly at some point, making metonymic and metaphorical shifts between them also more likely.

Acknowledgements I am grateful to Enoch Aboh, Mark Dingemanse, and William Herlofsky for discussing this paper with me and for their many valuable comments and suggestions, which I hope I have made good use of in spite of the fact that we may not always agree on some broader theoretical points.

References Aboh, E. 2007. A ‘mini’ relative clause analysis for reduplicated attributive adjectives. In Â�Linguistics in the Netherlands 23, B. Los & M. van Koppen (eds), 1–13. Amsterdam: John Benjamins. Aitchison, J. 1994. ‘Say, say it again Sam’: The treatment of repetition in linguistics. Swiss Papers in English Language and Literature 7: 15–34. Anttila, R. 2003. Analogy: The warp and woof of cognition. In The Handbook of Historical Â�Linguistics, B.D. Joseph & R.D. Janda (eds), 425–440. Oxford: Blackwell. Baker, P. 2003, Reduplication in Mauritian Creole, with notes on reduplication in Reunion Â�Creole. In S. Kouwenberg (ed.), 211–218. Bakker, P. 2003. Reduplication in Saramaccan. In S. Kouwenberg (ed.), 73–82. Bakker, P. & Parkvall, M. 2005. Reduplication in Pidgins and Creoles. In Hurch (ed.), 511–531. Baptista, M. 2003. Reduplication in Cape Verdean Creole. In S. Kouwenberg (ed.), 177–184. Bergman, B. & Dahl, Ö. 1994. Ideophones in sign language? The place of reduplication in the tense-aspect system of Swedish Sign Language. In Tense, Aspect and Action. Theoretical Contributions to Language Typology, C. Bache, H. Basbøll & C.E. Lindberg (eds), 397–422. Berlin: Mouton de Gruyter. Bolinger, D.L. 1952. Linear modification. PMLA 67: 1117–1144. Bolinger, D.L. 1967. Adjectives in English: Attribution and predication. Lingua 18: 1–34. Brentari, D. 2002. Modality differences in sign language phonology and morphophonemics. In Meier et al. (eds), 35–64. Bybee, J.L., Perkins, R. & Pagliuca, W. 1994. The Evolution of Grammar. Tense, Aspect, and Modality in the Languages of the World. Chicago IL: University of Chicago Press. Conradie, J. 2003. The iconicity of Afrikaans reduplication. In From Sign to Signing [Iconicity in Language and Literature 3], W.G. Müller & O. Fischer (eds), 203–224. Amsterdam: John Benjamins.

 Olga Fischer Crystal, D. 1987. The Cambridge Encyclopedia of Language. Cambridge: CUP. den Besten, H., Luijks, C. & Roberge, P. 2003. Reduplication in Afrikaans. In S. Kouwenberg (ed.), 271–288. Dressler, W.U., Dziubalska-Kołaczyk, K, Gagarina, N. & Kilani-Schoch, M. 2005. Reduplication in child language. In Hurch (ed.), 455–474. Deuchar, M. 1984. British Sign Language. London: Routledge & Kegan Paul. Etymologisch woordenboek van het Nederlands, M. Philippa, F. Debrabandere & A. Quak (eds), Vol. 2, 2005. Amsterdam: Amsterdam University Press. Fischer, O. 2001. The position of the adjective in (Old) English from an iconic perspective. In O. Fischer & M. Nänny (eds), 249–276. Fischer, O. 2007. Morphosyntactic Change. Functional and Formal Perspectives. Oxford: OUP. Fischer, O. & Nänny, M. (eds). 2001. The Motivated Sign. [Iconicity in Language and Literature 2]. Amsterdam: John Benjamins. Fónagy, I. 2001. Languages within Language. An Evolutive Approach. Amsterdam: John Benjamins. Ghomeshi, J., Jackendoff, R., Rosen, N. & Russell, K. 2004. Contrastive focus reduplication in English (the salad-salad paper). Natural Language & Linguistic Theory 22: 307–357. Gooden, S. 2003. Reduplication in Jamaican Creole: Semantic functions and prosodic constraints. In S. Kouwenberg (ed.), 93–103. Grant, A. 2003. Reduplication in Mindanao Chabacano. In Kouwenberg (ed.), 203–210. Haugen, J.D. 2005. Reduplicative allomorphy and language prehistory in Uto-Aztecan. In Hurch (ed.), 317–349. Hurch, B. (ed.) 2005. Studies on Reduplication. Berlin: Mouton de Gruyter. Johnston, T. & Schembri, A. 2007. Australian Sign Language (Auslan): An Introduction to Sign Language Linguistics. Cambridge: CUP. Keenan, E.L. & Polinsky, M. 1998. Malagasy (Austronesian). In The Handbook of Morphology, A. Spencer & A.M. Zwicky (eds), 563–623. Oxford: Blackwell. Klima, E.S. & Bellugi, U. 1979. The Signs of Language. Cambridge MA: Harvard University Press. Kouwenberg, S. (ed.) 2003. Twice as Meaningful. Reduplication in Pidgins, Creoles and other Contact Languages. London: Battlebridge. Kouwenberg, S. & La Charité, D. 2001. The iconic interpretations of reduplication: Issues in the study of reduplication in Caribbean Creole languages. European Journal of English Studies 5: 59–80. Kouwenberg, S. & La Charité, D. 2003. The meanings of ‘more of the same’. Iconicity in reduplication and the evidence for substrate transfer in the genesis of Caribbean Creole languages. In S. Kouwenberg (ed.), 7–18. Kouwenberg, S. & La Charité, D. 2005. Less is more: Evidence from diminutive reduplication in Caribbean Creole languages. In B. Hurch (ed.), 533–545. Kuipers, A.H. 1967. The Squamish Language. Grammar, Texts, Dictionary. The Hague: Mouton (see also The World Atlas of Language Structures Online, 〈 http://wals.info/feature/27 〉). Kulikov, L. 2005. Reduplication in the Vedic verb: Indo-European inheritance, analogy, and iconicity. In B. Hurch (ed.), 431–454. Ladhams, J., Hagemeijer, T., Maurer, P. & Post, M. 2003. Reduplication in the Gulf of Guinea Creole languages. In S. Kouwenberg (ed.), 165–176. Lass, R. 1994. Old English. A Historical Linguistic Companion. Cambridge: CUP. Maddieson, I. 1984. Patterns of Sounds. Cambridge: CUP.



Cognitive iconic grounding of reduplication in language 

Meier, R.P., Cormier, K. & Quinto-Pozos, D. (eds). 2002. Modality and Structure in Signed and Spoken Languages. Cambridge: CUP. Mühlhäusler, P. 2003. A note on reduplication in Pitkern-Norfolk. In S. Kouwenberg (ed.), 239–243. Müller, W.G. 2001. Iconicity and rhetoric. A note on the iconic force of rhetorical figures in Shakespeare. In Fischer & Nänny (eds), 306–322. Norton-Warren, K. 1978. Aspect marking in American Sign Language. In P. Siple (ed.), 133–159. Nuckolls, J.B. 1995. Quechua texts of perception. Semiotica 103: 145–169. Nuckolls, J.B. 2003. To be or to be not ideophonically impoverished. In SALSA XI: Proceedings of the Eleventh Annual Symposium about Language and Society, Austin, W.F. Chiang et al. (eds), 131–142. Austin TX: Texas Linguistic Forum 47. Parkvall, M. 2003. Reduplication in the Atlantic Creoles. In S. Kouwenberg (ed.), 19–36. Pfau, R. & Steinbach, M. 2005. Backward and sideward reduplication in German Sign Language. In Hurch (ed.), 569–594. Regier, T. 1998. Reduplication and the arbitrariness of the sign. In Proceedings of the Twentieth Annual Conference of the Cognitive Science Society, M.A. Gernsbacher & S.J. Derry (eds), 887–892. Mahwah NJ: Lawrence Erlbaum Associates. Sandler, W. 1990. Temporal aspects and ASL phonology. In Theoretical Issues in Sign Language Research, Vol. I: Linguistics, S.D. Fischer & P. Siple (eds), 7–35. Chicago IL: University of Chicago Press. Singler, J.V. 2003. Reduplication in Vernacular Liberian English. In S. Kouwenberg (ed.), 155–160. Siple, P. (ed.). 1978. Understanding Language through Sign Language Research. New York: Academic Press. Sohn, H.M. 1999. The Korean Language. Cambridge: CUP. Supalla, T. & Newport, E.L. 1978. How many seats in a chair? The derivation of nouns and verbs in American Sign Language. In P. Siple (ed.), 91–132. Todd, L. 1990. Pidgins and Creoles, 2nd edn. London: Routledge. Tomasello, M. 2003. Constructing a Language. A Usage-based Theory of Language Acquisition. Cambridge MA: Harvard University Press. Troyer, M. 2008. Reduplication in the Luwanga verb. 〈http://web.mit.edu/mltroyer/www/ files/ Reduplication.pdf〉. Vincent, N. 1988. La posizione dell’aggettivo in italiano. In Tema-Rema in Italiano. Symposium, H. Stammerjohann (ed.), 181–195. Tübingen: Gunter Narr. Wilbur, R. 2005. A reanalysis of reduplication in American Sign Language. In Hurch (ed.), 595–623. Zerbian, S., Downing, L.J. & Kügler, F. 2009. Introduction: Tone and intonation from a typological perspective. Lingua 119: 817–826.

Imagic iconicity in the Chinese language Zhuanglin Hu Peking University

The nature of arbitrariness and iconicity has been a heated topic of debate among Â�Chinese linguists and semioticians in the past twenty years. Due to the fact that modern Chinese linguistic scholarship has been strongly influenced by Saussure’s Course in General Linguistics, the dominant view in China was once marked by arbitrariness. With the development of cognitive linguistics, the scale has now tilted toward iconicity. Most of the papers in favour of arbitrariness have focused on Saussure’s view of the sign as the dualistic relation between “signifier” and “signified”, and almost remained the same in content in their argument. The present paper will deal with iconicity studies in the Chinese language in China as more and more articles have been emerging in this field and helped deepen our understanding of the ontology and phylogeny of the Chinese language. It will offer a brief review of concepts and views held by Chinese philosophers and etymologists, past and present, in order to elucidate the background to the debate.

1.â•… Introduction The nature of arbitrariness and iconicity has been a heated topic of debate among Â�Chinese linguists and semioticians for the past twenty years. Due to the fact that modern Chinese linguistic scholarship has been strongly influenced by Saussure’s Course in General Linguistics (2001), the dominant view in China was once marked by arbitrariness. Yet, with the development of cognitive linguistics, the scale has tilted toward iconicity today. Since most of the papers in favour of arbitrariness have focused on Saussure’s view of the sign as a dualistic relation between ‘signifier’ and ‘signified’, that is, ‘sound-image’ and ‘concept’, and have not so much criticized the role of iconicity at the syntactic level and the textual level (which is usually of a diagrammatic type), this paper, to save space and time, will not deal with sequential or temporal iconicity, iconicity of distance, of quantity, of space, of markedness iconicity, etc., but will confine itself to imagic iconicity, that is, sound iconicity and graphic iconicity. In doing so, I will start with a brief review of concepts and views held by the Chinese philosophers and etymologists, past and present.

 Zhuanglin Hu

2.â•… A historical-theoretical overview In contrast to the principle of arbitrariness that has constituted mainstream linguistics in the twentieth century both in and outside China with regard to the relation between the sound pattern and the concept referred to, the principle of iconicity has always been more highly valued and studied in the history of Chinese philosophy and linguistic scholarship (Xu Guozhang 1988:â•›2–10).

2.1â•… The Pre-Qin period Iconicity has always been the chief mode for the Chinese people in sign formation. Figure 1 is a photo of part of a stone inscription dating back to the New Stone Age (5800 BC to 4700 BC). It is an image of the ‘God of Sun’ in the mind of the ancient Chinese, with the figure in the middle imitating a human being, twenty-three lines around a circle over the head representing the sun and sunlight, and several small circles around the waist and legs resembling the planets around the sun (Hu 2009b: 1). The ‘image-based analogy’ of signs was found in Yi Jing (The Book of Changes), which was said to be completed in the Zhou dynasty (1066 BC–256 BC). The Book of Changes mainly discussed the meaning of Bā GuÀ (‘the eight diagrams’), consisting of an arrangement of single and divided lines in eight groups of three lines each. Besides this, it gave us examples of interpreting a character by its similarity in sound with another character such as, 乾 (qián)1 means 健 (jiÀn, healthy), 坤 (kūn) means 顺 (shùn, act in submission to). This can be illustrated by comparing the similarities in the VC structures for each pair. Admittedly, there is a sense of speculation in this book, yet it is quoted and valued for its contribution to the study of signs and the relevant term ‘analogy and inference.’ (Zhang Xiaoguang 2003:â•›32–35; Zhang Xu 2007:â•›28–29). The Book of Changes also writes: “Baoxi was the king of the country. He looked above observing the astronomy of the heaven and bent his body observing the law of the earth; he observed the features of birds and animals in harmony with the earth. He started from those objects which were near his body and extended further to those in the distance […] so as to describe the states of myriads of objects” (Baidu 2009a: Xici II. In this sense, these words sound like the views of embodiment and experientialism discussed by Lakoff and his colleagues today. After this classic text, we come to Gongsun Long’s (?320 BC–?250 BC) well-known thesis: “the white horse is not horse” (Baidu 2009b). This ancient philosopher argued that we should make a difference between the generic name ‘horse’ and the Â�specific

.╅ The italicized roman letters represent the Chinese Pinyin system.



Imagic iconicity in the Chinese language 

name ‘the white horse’, that is to say, naming is not decoded arbitrarily. Accountability is based on different perspectives.

Figure 1.╇ The God of Sun

The ‘sound with meaning’ phenomena were also reported at this period. When asked by a king about the meaning of 政 (zhèng, governing), Confucius explained that it means 正 (zhèng, upright). From this, we can see that the two words 政 and 正 share the same sound zhèng, share the same form of 正, and consequently are close to each other in meaning.

2.2â•… The Han period Although the concept of Liù Shū (Six Scripts) first appeared in the book Zhōu Lĭ (The Zhou Ritual) in the pre-Qin period, it was first annotated by Zhen Xuan of the East

 Zhuanglin Hu

Han period. It was a book about the creation of Chinese characters. The six scripts are as follows (Baidu 2009c; Liang Dongmei 1998:â•›73–75): 1.â•… xiÀngxÃ�ng (象形, pictographic character): e.g. 山 (shān, mountain), 日 (rì, sun), 月 (yuè, moon). 2.â•… huìyì (会意, associative compound character): e.g. the character 信 (xìn, belief) suggests the speech (言) of man (人); 林 (lÃ�n, forest) suggests the repeated occurrences of 木 (mù, tree/wood), meaning more than one tree. 3.â•… zhuănzhù (转注, mutually explanatory character): e.g. 老 (lăo, old age) and 考 (kăo, long life, aged) have similar meaning as shown by the similarity in their pronunciation and writing. 4.â•… chŭshì (处事, self-explanatory): e.g. 上 (shÀng, on); 下 (xiÀ, under). Their meanings are expressed by the position of 人 (rén, person) in relation to the line. 5.â•… jiăjiè (假借, phonetic loan character): e.g. 求 (qiú, for); 求 (qiú, entreat). The second meaning ‘entreat’ is borrowed from the first one. 6.â•… xiéshēng/xÃ�ngshēng (谐声/形声, picto-phonetic): e.g. 江 (jiāng, river), 河 (hé, river) Each character has one element expressing the (generic) meaning of ‘water’, another element expressing the sound. Since then it has been agreed that ‘mutually explanatory characters’ were derived from the carving of signs on tortoise shells and other materials, ‘associate compound characters’ helped understand the original meaning of some characters, ‘mutually explanatory characters’ revealed how ancient Chinese created synonyms, ‘phonetic loan characters’ showed the creation of one character from another character, and ‘picto-phonetic characters’ formed a small part at the very beginning (20% in late jiágúwén – inscription on tortoise shells), but became widely used in forming new words (around 80% of the Chinese characters). As far as written Chinese is Â�concerned, the pictographic character forms the basis of the other five categories (Baidu 2009c). This tradition of Liù Shū or the Six Scripts was later extensively studied in Xu Shen’s Shuō Wén Jiě Zì (Interpreting Words and Characters), written in the East Han Period. Xu Shen used the term Yì shēng (亦声, a form where the sound expresses meaning iconically). As we know, the Chinese picto-phonetic characters usually consist of two elements, one indicating meaning and the other sound. For instance, the character 江 (jiāng, river) has two elements, the left element carries the meaning of ‘water’, the right element carries the sound ending in -āng. Xu Shen’s yì shēng is to point out that sometimes the sound element still carries meaning. The character 酣 (hān, drink to one’s heart’s content) is a case in point, its left element suggesting ‘wine’ and its right element indicating both the sound and the meaning of 甘 (gān, sweet and pleasant). In all, the yi-sound characters amount to 950 characters, ten percent of the total characters reported by Xu Shen’s dictionary (Huang Yuhong 1995:â•›28).



Imagic iconicity in the Chinese language 

According to Duan Yuchai’s Shuō Wén Jiě Zi Zhù (Annotations of Interpreting Words and Â�Characters), Xu’s theoretical thrust can be summarized as ‘sound comes from meaning, meaning is Â�realized in sound, and sound leads to the form/graph. If the learners want to recognize characters, they should observe the form so as to know its sound, and observe the sound so as to get its meaning’. This shows that form, sound, and meaning are closely connected (Ying Xuefeng 2006a: 1–6). Somewhat later than Xu Shen, Liu Xi’s book ShÃ� MÃ�ng (The Interpretation of Names) became known to the academic world (2009d). This is a book discussing the meaning of a character from the perspective of sound. Liu Xi was right in explaining that the meaning of the character 浍 (huì, a ditch) derives partly from the left element’s meaning of water, and partly from the sound element 会 (huì), which means a place where furrows meet. Although the book has been criticized for its subjectivity, it has also been praised for its description of the pronunciation of Chinese characters during the Han dynasty, as well as of the pronunciation of various dialects. The same is true for Liu Xie’s (?465–?520) Wén Xīn Diāo Lóng (The Literary Mind and Dragon Carving (Liu 2007)). Although it is a treatise on literary criticism, the book touched upon the view that pictographic writing co-exists with nature, and that heaven and earth are concepts formed by the human cognition of the objective world. According to Xu Guozhang (1988:â•›8), Liu’s linguistic theory goes like this: The objective world ↓ Human conceptual processing (human perception) ↓ Language (speech is the reflection of mind) ↓ Writing (writing comes from speech) In Liu’s work, one can also find the categorization of signs into iconic signs, echoic signs, and emotive signs.

2.3â•… The Song period During the Song Dynasty (960 AD–1279 AD), Wang Shengmei was reported in Shen Kuo’s book (Shen 2007) Mèngxī BÃ� Tán (Dream Pool Essays) to have developed ‘右文 说’ (yòu Wén Shuō, ‘On the Right Element Principle’). Before Wang, it had been widely accepted that the left element of a character gives meaning and the right element the sound. Wang challenged this view by arguing that the meaning of a Chinese character also comes from its right element, that is to say, many Chinese characters express their generic meaning from their left elements (or radicals) and specific meaning as well as their pronunciation from their right elements. For instance, a cluster of objects/ concepts related to ‘wood’ has as its left element 木 (mù, wood); a character related to

 Zhuanglin Hu

‘hill’ has as its left element 山 (shān, hill), etc. When it comes to the meaning of the right element, this can be illustrated by 戋 (jian, small), so a small quantity of water is called 浅 (qiăn, shallow), a small piece of gold is called 钱 (qián, money), the disappearing sun, and therefore the remnant, is called 残 (cán, incomplete), a small shell is called 贱 (jiÀn, cheap), a small piece of paper is called 笺 (jiān). All these characters get their specific meaning from the right element 戋, which ends with the sound – an (Ying Xuefeng 2006b). Next to Wang Shengmei, Dai Tong was also known for his ‘Principle of Decoding Meaning from Sound’. Dai said that “the written language derives from the sound”, “The existence of the sound results in the form of written language, thus meaning and sound co-exist and do not derive from the written language” (from Dang Huaixing 1992:â•›61–68). “The written language is the icon of the sound; and the sound is the expression of human spirit. With this spirit we have the sound […] without the Â�language the sound is not to be realized” (from ibid.). Dai also explained the phenomenon of a character that can take several forms. This is a matter of the transference of one sound to another by different language users when the sounds are close to each other. Dai was also correct in talking about the phenomenon of phonetic loans, which is only a matter of sound borrowing and has no relevance to its form. This especially occurs in expressing functional words when it is difficult to find icons for them. The Song period was also marked by Zhen Qiao’s Liù Shū Lvč (Sketch of the Six Scripts). In this book, Zhen was acknowledged for his two great achievements. Â�Theoretically he confirmed the importance of Six Scripts in etymological scholarship, including the role of pictographic characters and rules concerning the development of Chinese characters, features of self-explanatory characters, pictographic characters, compound characters, and picto-phonetic characters. Practically, Zhen was not confined to Xu Shen’s framework. He worked out concrete rules, suggested new explanations and revised some of Xu Shen’s explanatory errors (Zhang Biao 1997:â•›92–97).

2.4â•… The Qing period The etymological research in the Qing period (1644–1911 AD) was represented by scholars such as Gu Yanwu, Jiang Yong, Duan Yucai, Wang Niansun, Kong Â�Guangshen, Jiang Yougao, etc. Generally speaking, these scholars managed to develop quite a few practical analytical methods, such as exploring the meaning of loan characters from their sounds, determining the literal meaning of characters so as to explain their extended meanings, determining the word meaning from sentences, etc. (Wiki Encyclopedia 2009). Among the above scholars, Wang Niansun pointed out that the meaning of two-character words is based on their sounds. For instance, Duan Yucai noticed that the character 犹 (yóu, irresolute) and the character 豫 (yu,in advance/beforehand) have many duplicated forms, such as 犹豫 (yóuyú), 犹与 (yóuyú), 尤豫 (yóuyú), all



Imagic iconicity in the Chinese language 

Â� expressing the meaning of hesitancy (Wang Niansun 1985). With regard to the origin of language, Zhang Taiyan clearly mentioned that [l]anguage by no means comes out of void. The reason for us to call a horse ma, an ox niú, is not done by free will or presumption. […] Why do we call a magpie juè? It makes sounds like juèjuè. Why do we call a sparrow què? Its sound is quèquè. Why do we call a crow yā? It makes sounds like yāyā. Why do we call a wild goose yÀn? It makes sounds like Àn Àn. All this shows that they are called according to the sounds they make. (Zhang Taiyan 1996)

Here Zhang did not agree with the view of arbitrariness but sided with the view of iconicity.

2.5â•… From 1949 to the present For this period, I would like to mention two points. First, the scholar who did most of the research concerning ‘音近义通’ (yīn jìn yì tōng ‘meaning can be decoded from the closeness of sounds’) is Wang Li, who put forward the term ‘同源词’ (tóng yuán cÃ�, words with the same etymology) for this phenomenon in the Chinese language. For instance, a young dog is called 狗 (gŏu, puppy), a young bear or tiger is written as 豿 (gŏu) but keeps the same pronunciation, and a young horse is written as 驹 (colt) but sounds like jū. Based on this principle, Wang Li compiled the Dictionary of Characters with the Same Etymology (Lin Yan 2006:â•›219–224). Second, quite a few scholars have started to question and challenge the principle of arbitrariness in the past twenty years. The supporters of arbitrariness in China have often quoted the view of Xunzi (?–238 BC), a pre-Qin philosopher because of his principle of ‘约定俗成’ (yuēdìng sùchéng, sanctioned by popular usage), that is, rules are based upon the arbitrary practice of the language users. However, this long-Â� standing theoretical view was reinterpreted by the late Guozhang Xu (1988:â•›9) and Wang (2006:â•›1–7). Both Xu and Wang argued that Xunzi’s quoted wording has been misread. The first two characters, yuēdìng (sanctioning), involve the problem of ‘how to sanction’ and the last two characters súchéng (popular usage) involve ‘what we mean by popular usage’. Here, sanctioning and popular usage presuppose the role of convention, agreed upon by language users and society, and consequently social accountability, because only on the basis of those representations that are closely related to the world can language users make their first choices and then reach consensus. If every one attempts to speak in his or her own way, that is, arbitrarily, mutual understanding and communication will break down. In contrast, any language user would try their best to associate the concept with the relevant image, especially the sound image at the very beginning. At this point, Yin Wang (2006:â•›2) further noticed that there should be three Â�sentences in all with regard to the quotes from Xunzi’s saying, whereas the Â�arbitrariness

 Zhuanglin Hu

scholars only quote from the first two sentences and retain the four characters. These three sentences are as follows: “(1) As there is no fixed appropriateness for naming, it has to resort to consensus, so what is sanctioned by popular usage is appropriate, and what differs from popular usage is inappropriate. (2) Naming is by no means factual. It becomes factual only when it is sanctioned, that is, what is sanctioned by popular usage is factual. (3) Naming can be good. Good naming is simple and easy to Â�understand” (Wang 2006:â•›2). By ignoring the third sentence, some people have denied the underlying principle of motivation or accountability. On the one hand, Xunzi saw the existence of the phenomenon of being ‘sanctioned by popular usage’ in language; on the other hand, he emphasized the importance of social accountability so that the Â�language users can communicate with each other through an appropriate name. In addition to theoretical exploration, many papers in China deal with practical Â�analysis. It is true that it would be difficult for people, both arbitrariness supporters and Â�iconicity supporters, to find out what happened in the early days when our forefathers started to create language or signs, but we could still get some cues from analyzing the language we are using today, especially the Chinese language. Section 3 will deal with this Â�matter, the crux of which is to prove the feasibility of imagic iconicity.

3.â•… Imagic iconicity 3.1â•… Picto-phonetic iconicity In his discussion of signs forming a natural connection between sound image and concept, Saussure acknowledged the existence of onomatopoeia and exclamation in language, but he did not delve deeper into the matter. Quite a few Chinese researchers, however, have provided a growing body of evidence to support such a natural connection.

3.1.1â•… Onomatopoeic iconicity Onomatopoeic iconicity can be found in great numbers in the Chinese language. Apart from examples such as 汪汪 (wāng wāng) of a dog, 咪咪 (mīmī) of a cat, 嗡 嗡 (wēngwēng) of a bee, 哞哞 (mōumōu) of a cow, etc., one can find many more Â�Chinese characters representing natural sounds (Lin Yan 2006:â•›219–224; Dong Weiguang 1997:â•›101–108). For instance, for sounds produced by things hitting one another, Â�plosive consonants are often used, such as 嘣 (bēng, the sound of Â�palpitations or Â�explosions), 噔 (dēng, heavy footsteps on the stairs), 滴 (dÃ�, dripping), 咚 (dōng, the sound of beating a drum, knocking at a door), 哒 (dá, clatter of horses’ hoofs), 嘎 (gă, a loud, high-pitched sound), 咣 (guāng, bang), 啪 (pā, the sound of Â�clapping and Â�slapping), 嘭 (pēng, the sound of something falling heavily on the ground or Â�striking against something else), 嗵 (tōng, the sound of striking against something or Â�heart-beating), etc. Obviously, these sounds reveal their meaning already.



Imagic iconicity in the Chinese language 

One can also find the compounding of onomatopoeic characters to express the Â�complexity of sounds, either by reduplication or by the use of similar sounds, such as: 吧 吧 (bābā, smacking one’s lips), 吧嗒 (bādā, smacking one’s lips), 吧唧 (bājī, the sound of squelching barefoot through the mud), 嚓嚓 (cācā, the sound of rubbing), 嘀嗒 (dÃ�dá, the sound of dripping), 叮咚 (dīngdōng, the sound of tingling), 叮当 (dīngdāng, the sound of jingling or cluttering), 咕哝 (gūnōng, the sound of murmuring), 咕咚 (gūdōng, the sound of a heavy thing falling down), 轰隆 (hōnglōng, the sound of rumbling or rolling), 呼呼 (hūhu, the sound of breathing), 呼拉 (hūlā, the sound of flapping), 咔嚓 (kāchā. the sound of cracking), 噼啪 (pīpā, the sound of crackling), 扑哧 (pūchā, the sound of snorting or fizzing), 扑拉 (pūla, the sound of flapping or spreading wings), 扑通 (pūtōng, the sound of something heavy dropping into the water or to the ground), 飒飒 (sÀsÀ, sough/ rustle), 瑟瑟 (sèsè, rustling of the wind), 刷拉 (shuālā, a swishing sound), 唰唰/刷刷 (shuāshuā, rustling), 咝咝 (sīsī, whistling), 飕飕 (sōusōu, soughing/rustling of wind), 簌â•›簌 (sùsù, rustle), 索索 (suŏsuŏ, a rustling sound), 嗡嗡 (wēngwēn, buzz/hum), etc.

3.1.2â•… Phonaesthetic iconicity Saussure did not say much about the phonaesthetic iconicity of language. In fact, this category plays a very important role in the Chinese language. As far as vowel sounds are concerned, it is well recognized in Chinese that the degree of vocalic openness reflects the meaning of many characters. If the mouth is wide open, such as with [a:], it expresses something which is wide, or large, or high, or loud, especially when one is in high spirits, as shown in the pairs of characters given in Table 1. Against this, one can also find those sounds bearing the vowels [i:] and [e:], expressing the basic meaning of smallness, lowness, baseness, etc., when one wants to express the opposite meaning. We might call them the Yang (masculine) sound and the Yin (feminine) sound respectively. The sounds with [o:] and [u:] express the middle ground (Gu Zhengkun 1995:â•›87–108; Du Wenli 1996:â•›60–65; Yan Junrong 2000:â•›51–54). Table 1.╇ Contrast of openness between Yang sound and Yin sound concerning Â�naturalness Yang (Masculine) sound

Yin (Feminine) sound

Yang (Masculine) sound

Yin (Feminine) sound

长 (cháng, long) 开 (kāi, open) 上 (shÀng, above) 笑 (xiÀo, laugh)

短 (duăn, short) 关 (guān, close) 下 (xiÀ, below) 哭 (kū, cry)

大 (dÀ, large) 高 (gāo, high) 响 (xiăng, loud) 深 (shēn, deep)

小 (xiăo, small) 矮 (ăi, low)/低 (dī, low) 轻 (qīng, low) 浅 (qiăn, shallow)

Obviously, this kind of distinction between two opposite meanings is based on naturalness. It is psychologically natural and physiologically feasible for people to adjust the opening of their mouths.

 Zhuanglin Hu

Then, what about those concepts which are abstract and which are difficult to describe physically? There are good reasons for us to say that our forefathers could still apply this principle of picto-phonetic iconicity to the description of mental ideas or abstract ideas. They used vowels with wide opening to mean those concepts which are true, good, primary, male, strong, active, and those sounds with narrow opening to mean the opposite. Examples are provided in Table 2. Table 2.╇ Contrast of openness between Yang sound and Yin sound concerning �abstract ideas Yang (Masculine) sound

Yin (Feminine) sound

Yang (Masculine) sound

Yin (Feminine) sound

爱 (Ài, love) 白 (bái, white) 粗 (cū, coarse) 东 (dōng, east) 好 (hăo, good) 刚 (gāng, hard) 公 (gōng, public) 来 (lái, come) 近 (jìn, near) 男 (nán, man) 巧 (qiăo, skillful) 头 (tóu, head) 勇 (yŏng, brave) 有 (yŏu, have)

恨 (hèn, hate) 黑 (hēi, black) 细 (xì, small) 西 (xī, west) 坏 (huài, bad) 柔 (róu, soft) 私 (sī, personal) 回 (huÃ�, return) 远 (yuăn, distance) 女 (nǚ, woman) 拙 (zhuó, awkward) 尾 (wěi, tail) 怯 (qiè, timid) 无 (wú, nil)

褒 (bāo, praise) 饱 (băo, full) 旦 (dÀn, morning) 动 (dòng, stirring) 夫 (fū, husband) 忠 (zhōng, loyal) 善 (shÀn, good) 进 (jìn, advance) 良 (liáng, fine) 强 (qiáng, strong) 同 (tóng, same) 雄 (xióng, male) 友 (yŏu, friend) 真 (zhēn, true)

贬(biăn, depreciate) 饿 (è, hungry) 夕 (xī, evening) 静 (jìng, tranquil) 妻 (qī, wife) 奸 (jiān, wicked) 恶 (è, evil) 退 (tuī, retreat) 莠 (yòu, vicious) 弱 (ruò, weak) 异 (yì, different) 雌 (cí, female) 敌 (dÃ�, enemy) 伪 (wěi, false)

It has also been noticed that sounds with an -ang ending usually express the meaning of something which is strong, wide, primary, long, fragrant, etc., such as 昂 (áng, in high spirits), 盎 (āng, abundant), 棒 (bÀng, excellent), 昌 (chāng, prosperous), 长 (cháng, long), 常(cháng, constant), 敞 (chăng, wide open), 畅 (chÀng, unimpeded), 创 (chuÀng, start), 芳 (fāng, fragrant), 放 (fÀng, open), 刚 (gāng, strong), 纲 (gāng, the key part of something), 况 (kuÀng, expansive), 沆 (hÀng, a vast expanse of water), 皇 (huáng, emperor), 光 (guāng, sunlight), 广(guăng, broad), 奖 (jiăng, reward), 芒 (máng, bright), 强 (qiáng, strong), 壮 (zhuÀng, healthy), 洋 (yáng, vast), 扬 (yáng, loud), 糖  (táng, sugar), 堂 (táng, the big room of a house), etc. In opposition to this tendency, those sounds ending in -ing usually express the meaning of smallness, narrowness, lightness, coldness, or slenderness, for instance, 兵 (soldier), 冰 (bīng, ice), 轻 (qīng, lightness), 钉 (dīng, nail), 叮 (dīng, sting), 茎 (jīng, narrow and long stem), 零 (lÃ�ng, withered), 囹 (lÃ�ng, jail), 泠 (lÃ�ng, cool), 冥 (mÃ�ng, dark), 暝 (mÃ�ng, sunset), 瞑 (mÃ�ng, close), 轻 (qīng, of light weight), 倾 (qīng, bend), 顷 (qīng, a little while), 罄  (qìng, use up), etc. (cf. Dong Weiguang 1997:â•›101–108; Ying Xuefeng 2006a: 1–6).



Imagic iconicity in the Chinese language 

There is another factor with which language users can express differences in expressing the degree of evaluative meaning, that is, the choice of tone. There are four tones in modern Chinese, that is Tone 1: high and level tone marked by the sign: ˉ ; Tone 2: rising tone, marked by: ′ ; Tone 3: falling-rising tone, marked by: ˇ ; and Tone 4: falling tone, marked by:’. Compare the tones used in each pair in connection with the difference in meaning of the two characters concerned. Usually they fall in the order of tone 4, tone 1, tone 2, tone 3. Table 3.╇ Variation of tones to express different degrees of evaluative meaning 放 (fÀng, let go) 4

收 (shōu, restrain) 1

教 (jiÀo, teach) 4

学 (xué, learn) 2

宽 (kuān, broad) 1 入 (rù, come into) 4 师 (shī, teacher) 1 问 (wèn, ask) 4 硬 (yìng, stiff) 4 正 (zhèn, obverse) 1

窄 (zhăi, narrow) 3 出 (chū, go out) 1 徒 (tú, apprentice) 2 答 (dá, answer) 2 软 (rŭan, soft) 3 反 (făn, reverse) 3

热 (rè, hot) 4 生 (shēng, be born) 1 山 (shān, hill) 1 攻 (gōng, attack) 1 真 (zhēn, true) 1

冷 (lěng, cold) 3 死 (sĭ, die) 3 谷 (gŭ, valley) 3 守 (shŏu, defend) 3 伪 (wěi, false) 3

Examples in Table 3 have shown that characters with positive meaning are usually expressed by the choice of tone 4 and tone 1, whereas those with negative or relatively secondary meaning are expressed by the choice of tone 2 and tone 3. The choice of consonants also reveal what sort of meaning is expressed by some of the characters. Dong Weiguang (1997), Lin Yan (2006:â•›219–224), Ying Xuefeng (2006a: 1–6) report on the role of consonants in expressing some specific meaning. i.â•…When appearing in the initial position, the Chinese consonant [x] (IPA: Â�voiceless alveolar-palatal fricate) is sometimes related to speaking in a laughing Â�manner as one tends to spread one’s mouth in doing so, such as 嘻 (xī, giggling), 喜 (xĭ. delighted), 戏 (xì, play), 笑 (xiÀo, smile), 欣 (xīn, happy), 噱 (xuú, laugh), 喧 (xuān, speak loudly). ii.â•…Some characters beginning with the Chinese consonant [h] (IPA: velar fricative) sometimes represent the meaning of loud cry, such as, 哈 (hā, the sound of laughing), 咳/嗨 (hāi, to call somebody), 喊 (hăn, a loud call), 鼾 (hān, snoring), 吭  (háng, heartily), 号 (háo, to yell), 嚎 (háo, to cry aloud), 喝 (hē, to shout loudly), 呵 (hē, the sound of laughing), 嗬 (hē, to express surprise), 嘿 (hēi, to exclaim), 哼(hēng, to groan), 哄 (hōng, to roar), 轰 (hōng, to bang like thunder

 Zhuanglin Hu

or shelling), 吼 (hŏu, to howl), 呼 (hū, to cry out), 哗 (huā, to sound like flowing water, clamour), 话 (huÀ, to talk), 欢 (huān, loud cheer), 唤 (huÀn, to call out), 诙 (huī, to crack jokes). iii.â•…Some characters beginning with the Chinese liquid consonant [w] have negative meanings or express non-existence, such as 完 (wán, to finish), 无 (wú, nothing), 勿 (wú, do not), 毋 (wú, not), 亡 (wáng, die), 忘 (wÀng, forget), 罔 (wŭng, no), 惘 (wăng, to feel frustrated), 伪 (wěi, false), 未 (wèi, not). We could only find the following w-characters with the positive meaning of ‘many’. They are: 万 (wÀn, a very great number), 旺 (wÀng, prosperous), and 伟 (wěi, great). â•›iv.â•…Some characters beginning with the Chinese nasal consonant [m] suggest obscurity, such as 霾 (mÀi, haze), 盲 (máng, blind), 茫 (máng, in the dark), 瞀 (mÀo, dizzy), 昧 (mèi, to have hazy notions about), 蒙 (méng, hazy), 梦 (mèng, fancy), 懵 (měng, muddled), 朦 (méng, obscure), 迷 (mÃ�, to be confused), 谜 (mì, riddle), 眯 (mì, to slightly shut one’s eyes), 冥 (mÃ�ng, dark), 暝 (mÃ�ng, dark), 暮 (mù, dusk). This correlation gives us the impression that one tends to close one’s lips when one is not sure about something. Sometimes, this set of words also suggests non-existence or smallness, such as 靡 (mĭ, not have), 渺 (miăo, tiny), 藐 (miăo, small), 蔑 (miè, to despise), 泯 (měn, to vanish), 末 (mò, end), 莫 (mò, not), 没 (mò, till the end), 殁 (mò, to die), 墓 (mù, tomb). â•›v.â•…Some characters beginning with the Chinese consonant [f] suggest negation, such as 罚 (fá, to punish), 乏 (fá, to lack), 反 (făn, to reverse), 犯 (fÀn, to violate), 非 (fēi, not), 匪 (fěi, not), 诽 (fěi, to speak ill of somebody), 费 (fèi, to spend), 废 (fèi, to stop), 菲(féi, poor/humble), 分 (fēn, to divide), 坟 (fén, mound), 焚 (fén, to destroy by fire), 否 (fŏu, not), 弗 (fú, not). In contrast, one can only find a few characters that support the opposite. They are 繁 (fān, in large numbers), 肥 (féi, fat), 丰 (fēng, abudant), 富(fù, rich), 福 (fú, good fortune).

3.2â•… Pictographic iconicity The discussion so far has already showed that picto-phonetic iconicity has played a very important role in the development of the Chinese language. This is all the more true when it comes to the role of pictographic iconicity in transforming the spoken form to the written form. To start with, we need first to take a look at the following two pictures:

Figure 2.╇ Hill



Imagic iconicity in the Chinese language 

Figure 3.╇ Bow

Figure 2 shows that there is a likeness between the Chinese character “山” (shān, hill) and a hill with its peaks. The same is true for Figure 3 with the character 弓 (gōng, bow) and the objects which can shoot an arrow. Based on this sort of iconic representation, one can account for more Chinese characters by their shape, such as 日 (rì, sun), 月(yuè, moon), 天 (tiān, sky), 人 (rén, man), 女 (nv, woman), 大 (dÀ, big), 火 (huŏ, fire), 水 (shuĭ, water), 目 (mù, eye), 田 (tián, farmland), 竹 (zhú, bamboo), 羽  (yú, feather), 雨 (yŭ, rain), etc. These basic signs are then widely used as an element carrying the generic meaning of a compound character. Note the following sets of examples: ╇ i.â•…Characters such as 号 (háo, to yell), 叼 (diāo, to hold in the mouth), 叫 (jiÀo, to cry), 吐 (tù, to vomit), 吓 (xiÀ, to threaten), 唬 (hŭ, to intimidate), 呼 (hŭ, to exhale), 吸 (xī, to inhale), 咬 (yăo, to bite), 吞 (tūn, to swallow), all have 口 (kŏu, mouth) as one of their elements, or radical, that is, they are all related to the activity of the mouth, whether human or non-human. ii.â•…Characters such as 松 (sōng, pine tree), 杉 (shān, fir tree), 杨 (yáng, poplar tree), 柳 (liŭ, willow tree), 李 (lĭ, plum tree), 床 (chuáng, bed), etc. all have a generic element 木 to show that their meaning is related to a plant or wood. In the same vein, characters such as 椅 (yĭ, chair), 桌 (zhuō, table), 柜 (guì, cupboard), 械 (xiè, tool), 棍 (gùn, rod), 棒 (bÀng, stick), 扛 (káng, to carry on the shoulder), 抬 (tái, to raise), 朽 (xiŭ, rotten), 枯 (kū, withered), 荣 (róng, to flourish), etc. are related to a piece of furniture or tool made of wood, or the state of the plant or wood. iii.â•…Characters such as 闩 (shuān, door bolt), 闭 (bì, close), 闯 (chuăng, dash), 闱  (wéi, a side gate of an imperial palace), 间 (jiān, room), 闷 (mèn, to shut oneself indoors), etc. have the generic element 门 (mén, door), the sign embracing as it were another character, and thus all these signs express the meanings related to ‘door’. Based on this understanding of the relation between meaning and shape, Chinese characters can be classified into 189 radicals (or so) and serve as the generic elements of characters.

3.3â•… Picto-phonetic-graphic iconicity As mentioned before, the traditional view of Chinese linguistic scholarship held that a compound character in Chinese has two elements, the left element expressing

 Zhuanglin Hu

Â� meaning and the right element sound. However, Section 2.3 has showed that it was Wang Shengmei of the Song period who first systematically reported that the right element not only shows the character’s pronunciation, but also its meaning, the specific meaning. Just like Wang’s classic example of “戋” (jiān, small), “浅” (qiăn, shallow), “钱” (qián, money), “残” (cán, incomplete), “贱” (jiÀn, cheap), and 笺 (jiān, small piece of paper), getting their specific meaning of ‘smallness’ from the right element “ 戋”, we can also test it by analyzing another character 茎 (jīng, narrow and long stem). The upper part only gives the generic meaning of ‘plant’ or ‘grass’ while the specific meaning of ‘narrow and long stem’ is to be decoded from the right or lower element. For instance, 胫 (jìng, narrow and long small leg), 径 (jìng, narrow and low path), 颈 (jìng, neck), 经 (jīng, narrow and long thread), 泾 (jīng, narrow and long stream) (Gu Zhengkun 1995). More examples could be found in the following groups: ╇ i.â•… 叉 (chā, separate), 汊 (chÀ, tributary), 杈 (chÀ, trunk), 衩 (chÀ, underpants). ii.â•…垂 (chuÃ�, to hang down), 捶(chuÃ�, to beat with a stick or fist), 棰 (chuÃ�, stick), 锤 (chuÃ�, to beat with a hammer). iii.â•…低 (dī, low), 诋 (dĭ, to slander), 底 (dĭ, bottom), 柢 (dĭ, root of a tree), 骶 (dĭ, lower part of the tail bone). â•›iv.â•… 框 (kuāng, door frame), 眶 (kuÀng, socket). â•›v.â•… 胧 (lóng, dim), 聋 (lóng, deaf), 曨 (lóng, dusky), 籠 (lóng, general/sweeping).

4.â•… Concluding remarks To bring my paper to a close, I would like to summarize the following points. First, from the discussion in Section  2 and Section  3, it has become clear that imagic iconicity, whether a sound image or a visual image, plays a very important role in the development of the Chinese language. To be more exact, Chinese scholars, ancient and contemporary, have always highly valued the study of iconicity. Second, Saussure should be praised for his prediction of the birth of a new discipline, semiology, and to have initiated the study of language as a system of signs. However, limited by the period he worked in and his research goal, he restricted himself mainly to the study of the relationship between concept and sound pattern (Saussure 2001:â•›67). In contrast, most linguists today have acknowledged that there are three levels of language – semantics, lexico-grammar, and phonology/graphology. Language as a system of signs should be studied at all three levels. In this regard, Saussure’s principle of arbitrariness or sound pattern-concept duality can exert only a very limited role. It cannot help us very much as it would be difficult for us to prove or disprove the occurrence of arbitrariness, in spite of the fact that we can find more examples that show the role of iconicity at various levels.



Imagic iconicity in the Chinese language 

Third, Saussure talked about his first principle, the fact that the linguistic sign is arbitrary, in different ways (2001:â•›66–69). On one occasion, he said: “No one disputes the fact that linguistic signs are arbitrary” (2001:â•›68); on another, he also stated that the word ‘arbitrary’ “must not be taken to imply that a signal depends on the free choice of the speaker” (2001:â•›68). If choice is not free, then there must be some principle or guidelines to justify the choices of language users. Immediately following this view, Saussure argued that “the individual has no power to alter a sign in any respect once it has become established in a linguistic community” (2001:â•›68). Clearly, Saussure refers to conventionality but he did not bother to further explain how the sign was established; he simply declared that “the signal is unmotivated” (2001:â•›69). Consequently, Saussure’s arbitrariness equals unmotivatedness (Hu 2009a: 93–102). This is against the cognitive processes of human beings. Without motivation or social justification, how can one defend or explain one’s proposal of signs? When a sign is established, it must be presupposed that one proposed signifier is better than another proposed signifier, otherwise it could never be a sign. It is true that we do not know what happened at the very beginning among our forefathers. One thing is clear, however: with iconicity, we can try our best to explore and discover some cues, but with arbitrariness one just sits there, one does not have to do anything, and one can never get the problem solved. At this point, Saussure did not give much attention to non-European languages, such as the Chinese language. It gives the impression that Saussure’s theory is somewhat eurocentric, as he did not take into account the languages in the world in general, and the Chinese language in particular. Furthermore, Saussure said that “onomatopoeic and exclamatory words are rather marginal phenomena […]” (2001:â•›69). He made no mention of phonaesthetic iconicity and did not see that such words are by no means marginal in the Chinese language, to say nothing of its pictographic characters.

References Baidu Encyclopedia (百度百科). 2009a. 周易系辞下 (The Book of Changes：Xici II). . Baidu Encyclopedia (百度百科). 2009b. 公孙龙子 (Gongsun Long Zi). . Baidu Encyclopedia (百度百科). 2009c. 六书 (The Six Scripts). . Baidu Encyclopedia (百度百科). 2009d. 刘熙 (Liu Xi). From . Dang, Huaixing (党怀兴). 1992. 《六书故》“因声以求义论” (On the theory of meaning from sound in The Origin of the Six Scripts). Journal of Sh’an Xi Normal University 21(1): 61–68. Dong, Weiguang (董为光). 1997. 话说 “音义初始” (On the initiation of sound and meaning). 语言研究 (Language Studies) 1: 101–108. Du, Wenli (杜文礼). 1996. 语言的象似性探微 (An exploration of language iconicity). 四川外 语学院学报 (Journal of the Sichuan Foreign Languages Institute) 1: 60–65.

 Zhuanglin Hu Duan, Yuecai (段玉裁). 《说文解字注》(Annotations of Words and Characters), 经韵楼原刻 本. Shanghai Classics Publication House. Gu, Zhengkun (辜正坤). 1995. 人类语言音义同构现象与人类文化模式 (The structural identity between sound and meaning of languages in connection with patterns of human cultures). 北京大学学报 (哲学社会科学版) (Journal of Peking University, Philosophy and Social Sciences edition) 5: 87–108. Hu, Zhuanglin (胡壮麟). 2009a. 对语言象似性和任意性之争的反思 (Reflections on the argument over iconicity and arbitrariness. 北京大学学报：哲学社会科学版 (Journal of Peking University, Philosophy and Social Sciences edition) 46(2): 93–102. Hu, Zhuanglin (胡壮麟). 2009b. Words in Honour of the First Issue of Chinese Semiotic Studies, Chinese Semiotic Studies 1: 1. Huang, Yuhong (黄宇鸿). 1995. 试论《说文》中的 “声兼义” 现象 (A tentative explanation of ‘sound plus meaning’ in Shuo Wen). 广西师范大学学报 (哲学社 会科学版) (Journal of Guangxi Normal University, philosophy and social sciences edition) 31(1): 26–31. Liang, Dongmei (梁冬梅). 1998. 右文说探源 (Exploring the right-element theory). 贵州文史 丛刊 (Collected Works of Literature and History of Guizhou) 2: 73–75. Lin, Yan (林艳). 2006. 语词音义联系及联系机制初探 (A tentative exploration of word-sound connections and their mechanisms). 西南民族大学学报 (人文社科 版) (Journal of Southwest Nationalities University, Human Studies Edition) 1: 219–224. Liu, Xie (刘勰). 2007. 文心雕龙 (The Literary Mind and the Dragon Carving). 华 文出版社 (Huawen Press). Shen, Kuo (沈括). 2007. 梦溪笔谈 (Dream Pool Essays). 湖北辞书出版社 (Hubei Press of Classics). de Saussure, F. 2001. Course in General Linguistics, R. Harris (transl.). London: Gerald Â�Duckworth & Co. Ltd/Foreign Language Teaching and Research Press. Wang, Niansun (王念孙). 1985. 广雅疏证 (Annotations of Guangya). Jiangsu Classics Â�Publication Press. Wang, Yin (王寅). 2006. 荀子论语言的体验性认知辩证观 语言哲学再思考: 语言 的体验性 (The cognitive dialectic view of Xunzi’s language experientialism – Rethinking of language philosophy: experientialism in language). 外语学刊 (Journal of Foreign Languages) 5: 1–7. Wiki Encyclopedia. 2009. Chinese Philology. 〈http://www.cn./wiki〉 (30 January, 2009). 语言哲学探索之一 Â�(Problems Xu, Guozhang (许国璋). 1988. 语言符号的任意性问题 of arbitrariness in linguistic signs). 外语教学与研究 (Foreign Language Teaching and Research) 3: 2–10. Yan, Junrong (延俊荣). 2000. 汉语语音与语言意义象似性例举 (Illustrations of Chinese sound-meaning iconicity). 解放军外国语学院学 (Journal of PLA Institute of Foreign Â�Languages) 23 (5): 51–54. Ying, Xuefeng (应学风). 2006a. 语音象似性与右文说 (Phonetic iconicity and the right-Â� element theory). 内蒙古农业大学 (哲学版) 学报 (Journal of Neimenggu Agriculture Â�University, Philosophy Edition) 2: 1–6. Ying, Xuefeng (应学风). 2006b. 国内象似性研究动态 (Trends in iconicity studies in China). 江西语言学会年会论文 (A paper presented at the Annual Conference of Jiangsu Linguistics Association). Zhang, Biao (张标). 1997. 论郑樵的《六书略》(On Zhen Qiao’s Sketch of The Six Scripts). 古 汉语研究 (Studies in Classic Chinese) 2: 92–97.



Imagic iconicity in the Chinese language 

Zhang, Taiyan (章太炎). 1996.《国故论衡·语言缘起说》(On Chinese Classics: the origin of language). Chen Pingyua (ed.). 中国现代学术经典·章太炎卷 (Contemporary Classics of China: Zhang Taiyan). Hebei Education Press. Zhang, Xiaoguang (张晓光). 2003. 《周易》中类比推论思想 (Analogical thoughts in The Book of Changes). 社会科学期刊 (Journal of Social Sciences) 5: 32–35. Zhang, Xu (张旭). 2007.《易经》中 “取象比类” 的句法意义及其影响 (The syntactic meaning of iconicity and analogy in The Book of Changes and its influences). 太原师范学院学 报 (社会科学版) (Journal of Taiyuan Normal Institute, Social Sciences Edition) 6(5): 28–29.

Words in the mirror Analysing the sensorimotor interface between phonetics and semantics in Italian Luca Nobile

Université de Bourgogne This paper proposes a theoretical, methodological and descriptive framework for combining a gestural-mimetic theory of the sign with the principle of arbitrariness. First, it draws attention to the history of the question, discusses its theoretical core and presents a method to analyse the relationship between the differential system of phono-articulatory signifiers and the differential system of logical-semantic signifieds. Next, this method is tested on the Italian grammatical monosyllables, focusing on a complete description of the words formed from one phoneme (‘monophonemes’), of the words distinguishing grammatical persons and of the adverbial pairs. The main result is that phono-articulatory oppositions (such as [open : closed] and [front : back]) can be systematically described as images of logical-semantic oppositions (such as {complex : simple} and {positive : negative}, respectively). In Section 3, a number of conclusions are drawn, particularly about the difference between linguistic ‘values’ and psychological ‘concepts’.

1.â•… Theoretical and methodological aspects 1.1â•… Echo-mirror neurons and arbitrariness The discovery of mirror neurons by Giacomo Rizzolatti’s team in the mid-1990s revealed a physiological, automatic connection between the perception of others’ actions and the execution of one’s own. Indeed, mirror neurons fire both when one performs an action and when one sees or hears someone else executing it. As they were specifically found in the Broca’s area, the well-known zone of the left Â�hemisphere involved in the production of language, they were presented as evidence of a strong neurophysiological and evolutionary contiguity between gestural mimicry and Â�language articulation (Rizzolatti & Arbib 1998). In particular, there is evidence that hearing the speech sounds of others evokes motor potentials in the muscles required to articulate them (Fadiga et al. 2002), and that

 Luca Nobile

the wideness of the articulatory gestures is influenced by the wideness of Â�simultaneous hand gestures, whether one’s own or that of others (Gentilucci & Corballis 2006; see Figure 1). 50 mm

Lips aperture of /ba/

Apple Cherry

35

20 1160 Hz

0

ms

250

500

F1 of /ba/

1030

Apple Cherry

900

Figure 1.╇ Grasp observation influences vowel aperture (Gentilucci & Corballis 2006). A Â�speaker articulating the syllable /ba/ while grasping an object or observing someone grasping an object (on the right) articulates a more open syllable if the object is bigger, a more closed one if the object is smaller (on the left, top). This modifies the acoustic formant F1 (a distinctive feature of the vowel quality) strongly if the speaker is a child, weakly if he is an adult (left, bottom)

It has thus been hypothesized that such an ‘echo-mirror neuron system’ could explain not only the ability to imitate speech sounds but also the capability to Â�understand their meanings. Indeed, the original ‘meaning’ of speech sounds seems to have been the motor pattern they share with simultaneous, analogous hand gestures; therefore their semantics and phonetics should originally be coded together: In other words, neurons appeared which coded phonetics simultaneously with semantics. In this way, heard speech produces not only a tendency to imitate the sound but also an understanding of the accompanying body-action gestures.  (Rizzolatti & Craighero 2007:â•›783)

Such a naturalistic explanation of the link between phonetics and semantics is traditionally refused by linguists. For example, in his paper tellingly entitled Â�“Language beyond our grasp” (2004), James Hurford criticizes Rizzolatti’s and Arbib’s “Language within our grasp” (1998). Citing Saussure, he merely draws attention to the fact that the word ‘apple’ is not similar to the image of an apple. This traditional argument is considered sufficient to conclude that the relationship between signifier and signified is arbitrary and that, consequently, the mirror neuron hypothesis will not explain the efficiency with which children acquire semantic distinctions by means of phonetic ones.



Words in the mirror 

Hurford’s paper reflects a widespread position among linguists, one for which Sylvain Auroux (2007) argues with particular clarity. Introducing a recent number of Histoire Epistémologie Langage entitled “Le naturalisme linguistique et ses désordres”, he maintains that “arbitrariness is not a fact, but the theoretical principle at the origin of history and culture”. The Enlightenment would introduce it as a foundation of both human freedom and the humanities, while its denial by 19th-century linguistic naturalism would prepare the arrival of Nazism. So, one of the most authoritative European theorists of linguistics presents the relationship between arbitrariness and naturalism as a war between civilization and barbarism. The aim of the present paper is to propose a theoretical, methodological, and descriptive framework for combining the gestural-mimetic principle implied by the mirror system hypothesis and the principle of arbitrariness of the sign required to explain the diversity of languages and the freedom of human consciousness. Moreover, this paper offers some evidence to prove Hurford’s prediction wrong concerning language acquisition and it shows how the mirror system hypothesis could explain children’s rapidity in acquiring semantic distinctions by means of phonetic ones. First, the paper briefly draws attention to the history of the question, discusses its theoretical core and presents a method to analyze analogies between phonological and semantic structures. In Section 2, this method will be applied to Italian Â�grammatical monosyllables, focusing on a complete description of the words formed from one Â�phoneme (‘monophonemes’), of the words distinguishing grammatical persons and of the adverbial pairs. In Section 3, a number of conclusions will be drawn, particularly about the difference between linguistic ‘values’ and psychological ‘concepts’.

1.2â•… An ancient question The relationship between sound and meaning is one of the original problems in the Western philosophical tradition. Indeed it constitutes the topic of one of the Â�dialogues by Plato, the first philosopher whose whole corpus tradition has given to us. In the Cratylus, the character of Socrates discusses Hermogenes’ and Â�Cratylus’ thesis on the relationship between names and things. Hermogenes is a sophist who Â�considers it a contractual and arbitrary relationship (384c–385e), Cratylus is a Â�Heraclitean who considers the relationship natural and iconic (422d–27d). Although tradition Â�attributes Cratylus’ ideas to Plato, some recent interpretations have placed him at an equidistant position. His distance from Cratylus would be a distance from the sacred, oral, poetic, traditional knowledge and it allowed him to base the secular philosophical thought on prose (Fresina 1991:â•›75–110; and cf. Genette 1976:â•›11–37). So the question of arbitrariness seems to be linked to the origin of abstract thought, as based on the practice of writing. One may ask if the current technological change infuses new relevance into the issue. As it is, since Plato, the debate has continued through the entire tradition.

 Luca Nobile

Hermogenes’ thesis is reformulated by Aristotle in the first pages of De interpretatione (16a). Thomas picks it up from there and puts it in his Summa Â�theologica (PII-II Q85 A1), making it the official doctrine of Christian Aristotelianism. So one can find it in Dante’s De vulgari eloquentia (ca. 1305; I, 3), in Arnauld’s and Nicole’s Logique of Port-Royal (1662:â•›98; 1683:â•›58), in the cartesian Discours Â�physique de la parole by Gérard de Cordemoy (1668:â•›23), and finally in John Locke’s Essay Â�Concerning Human Understanding (1690:â•›III, 2), which problematizes it in a Â�decisive manner. Franz Bopp implicitly accepts Hermogenes’ idea in the preface of his Â�Vergleichende Grammatik (1833:â•›iii) and makes it one of the foundations of new comparative linguistics. Cratylus’ thesis is reformulated by Epicurus in his Epistula ad Herodotum (75–76). It further emerges in different forms in the Stoa, as Augustine says in De dialectica (VI) and as one can read in Noctes atticae (X, 4), where Aulus Gellius reports the ideas of Latin grammarian Nigidius Figulus. The indirect nature of these statements suggests, however, that the doctrine does not pass the filter of Christianity. It is not by accident that in the Middle Ages it rather characterized the Jewish tradition, Â�beginning with the anonymous founder of the Kabbalah, the Sepher YezirÀ (‘The Book of Formation’; circa 6th century AD), up until the classic by Abraham Abulafia, Or ha-Śekel (‘The light of intellect’ VIII, 5; circa 1280). Only with the rediscovery of Epicurus, between the editio princeps of Diogenes Laertius’ Vitae philosophorum (1533; X, 75–76) and Pierre Gassendi’s Animadversiones (1649:â•›54–55 and 700–706), did cratylism Â�reappear in lay clothing. First, in the fourteenth chapter of John Wallis’ Grammatica linguae Â�anglicanae (1653), next at the opening of Gottfried Wilhelm Leibniz’s Brevis Â�designatio… ex indicio linguarum (1710:â•›2) as well as in his famous answer to Locke’s Essay (1765:â•›III, 2), then assuming an important place in Giambattista Vico’s Scienza Nuova (1744:â•›92 and 181), and finally providing a foundation to French materialist linguistics, with the Traité de la formation méchanique des langues by Charles de Brosses (1765:â•›I, 195–294). Hence in various ways, including Etienne Bonnot de Condillac’s Grammaire (1775:â•›18–22) and Dieudonné Thiébault’s Grammaire philosophique (1802:â•›I,â•›14–18), the theory joins Wilhelm von Humboldt, the political godfather of Bopp. Humboldt presents a well-known reformulation of Cratylus’ theory in his Verschiedenheit des menschlichen Sprachbaues (1836:â•›81–106).

1.3â•… Two notions of arbitrariness Even if we can call Hermogenes-like positions ‘conventionalist’ or ‘arbitrarist’ and Cratylus-like positions ‘naturalist’ or ‘iconicist’, it should be pointed out that, inevitably, each theory also covertly includes its opposite, which the other one exhibits in full light. Indeed, if Hermogenes’ theory (in Aristotle’s formulation) can claim that speech sounds are conventionally associated with ideas, it is only because ideas are considered natural products of perception: the same for all people, independent from



Words in the mirror 

languages ; this represents Hermogenes’ naturalism. Conversely, if Cratylus’ theory (in Epicurus’ formulation) can claim that sounds are naturally associated with ideas, it is only because ideas are considered relatively arbitrary products of perception, different for people in accordance with their language; and this represents Cratylus’ arbitrarism. In both cases, one must account for a certain adherence and non-adherence of language to reality. If the idea perfectly fits the object, then the sound cannot fit the idea, given that different languages use different sounds to signify the same thing. But if the idea does not fit the observed object perfectly, then the sound can fit the idea, given that different languages will use different sounds to signify different ideas of things. Behind the opposition between ‘conventional’ and ‘natural’ then lies an opposition between two different theories of meaning, which are also two different notions of arbitrariness. As is well-known, both concepts are present in Cours de linguistique générale (see Figure 2): an Aristotelian Saussure, with introductory intention and pedagogical caution, says that the association between sound and concept is arbitrary (1922:â•›100–102); an Epicurean Saussure, with more daring investigative ambition and the intention of studying in depth, says that, even if paradoxical when compared to the above (1922:â•›159), arbitrariness concerns the segmentation of both semantic and phonetic fields from which signifiers and signifieds emerge (1922:â•›155–169). A Arbor B Concept Image acoustique

Sign e Sign t

Sign e Sign t

Sign e Sign t

Figure 2.╇ Representations of two types of arbitrariness in Saussure’s Cours. The Aristotelianlike arbitrariness (on the left) concerns the relationship between the sound and the concept within each word. The Epicurean-like arbitrariness (on the right) concerns the relationship between the cuts the sign produces simultaneously, on the one hand, in the psychological Â�continuum of the concepts ‘A’ and, on the other hand, in the physiological continuum of the sounds ‘B’. In the latter perspective, one cannot say that the relationship between signifier and signified is arbitrary, because neither of them exists independently from their reciprocal Â�relationship

As Emile Benveniste (1966) underlined, Saussure’s formulation of Epicurean Â� arbitrariness destroys the foundation of Aristotelian arbitrariness. Indeed the Â�concept of langue as a system of values requires the relationship between signifiers and Â�signifieds to be necessary. The langue does not work with individual ‘sounds’

 Luca Nobile

arbitrarily Â�associated with individual ‘concepts’, but with differential Â�systems of Â�‘signifiers’ structured by their coupling with differential systems of ‘signifieds’. Â�Signifiers and signifieds are fields of sounds and concepts the boundaries of which are defined by their oppositions to other fields. Even if each individual Â�association between sounds and concepts seems arbitrary, the way langue differentiates it from other associations is not. It is this differentiation – rather than its acoustical or Â�psychological contents – which characterizes the language faculty and defines the object of linguistics. Thus, in the Cours, the signifier and signified are presented as ‘aspects’ of the same ‘value’ (1922:â•›158–165), and the theory of value is enunciated as the theory of a systemic link between phonetics and semantics: A linguistic system is a series of differences in sounds combined with a series of differences in ideas, but this putting together of a certain number of acoustic signs with corresponding cuts made in the mass of thought gives rise to a system of values; and it is this system which constitutes the effective link between the phonic and psychic elements on the interior of each sign.  [Saussure 1922:â•›166; translation by Paul J. Thibault]

1.4â•… Two notions of iconicity Starting from Saussure’s perspective, Roman Jakobson accepted Charles Sanders Peirce’s theory of iconicity. After reminding us that Peirce’s triad of symbol, index and icon refers to three semiotic dimensions simultaneously present in each sign, Â�Jakobson (1965 :â•›28) focused on Peirce’s distinction between two subtypes of icons: images and diagrams. While images depict things, diagrams depict relationships between things. For example, two rectangles represent the relationship between the production of steel in the US and the USSR: the rectangles are not similar to the Â�production of steel, but the ratio between their sizes is similar to the ratio between the two productions. Jakobson (1949, 1965) has repeatedly shown that language employs not only imagic iconicity (onomatopoeia, synaesthesia and the like) but also this more abstract kind of iconicity in which the signifier tends to appear as a diagram of the signified. Canonical examples: the time sequence pictured in the syntactic order of Caesar’s phrase veni, vidi, vici (1965:â•›27); the largest amount shown by the increased length of plurals and superlatives (1965:â•›30); the smallness of objects depicted by acute sounds and closed articulations (1965:â•›34; 1980:â•›224–230, with Linda Waugh). The concept of diagram was subsequently developed by several scholars Â�including Talmy Givón (1985 and 1995), Linda Waugh (1993), Olga Fischer (1999 and 2001, the latter with Max Nänny), Marianne Kilani-Schoch and Wolfgang Dressler (2005:â•›39–57), and it seems today to constitute the prevailing sense of the term Â�iconicity. Â�Nevertheless, research into imagic iconicity was also renewed in several directions, for example by Ivan Fónagy (1983 and 2001), Maurice Toussaint (1983), Kawada Junzo (1988),



Words in the mirror 

� Margareth Magnus (2001), Didier Bottineau (2002), Philippe Monneret (2003 and 2004), Dennis Philps (2006), and Federico Albano Leoni (2009). Sometimes scholars opposed imagic and diagrammatic iconicity as if they excluded each other. This is not our purpose.

1.5â•… Working hypothesis This paper aims to propose a framework for combining naturalism and arbitrariness. To do so, it problematizes the notion of arbitrariness: it accepts the Epicurean variant as reformulated by Saussure (arbitrary segmentation of phonetic and Â�semantic Â�continua) and rejects the Aristotelian-Thomistic variant (arbitrary relationship between sound and meaning). In other words, it assumes that both phonetic and semantic fields are internally differentiated in a relatively arbitrary way with respect to reality (as Â�evidenced by the diversity of languages), but it denies that within every language the relationship between the phonetic differentiations and the semantic Â�differentiations is arbitrary. In particular, this paper hypothesizes that, if one limits the analysis to a single Â�language, the relationship between the differential system of signifiers and the differential Â�system of signifieds can be described systematically as a diagrammatic relationship. This Â�diagrammatic relationship, however, is more Â�concrete than usual: even if it is purely differential, it clearly shows some ‘imagic’ properties. Olga Fischer (1999:â•›347) already pointed out this “other type of diagrammatic Â�iconicity”, comparing some of Max Nänny’s examples with John Haiman’s. Below, I will try to distinguish these two types of diagrams, which result from successive levels of analysis. I will use the term ‘iconic diagram’ for the more abstract and accessible pattern, and ‘imagic diagram’ (with an intentional oxymoron) for the more concrete and hidden one, the analysis of which depends on the former.

1.6╅ Methodological remarks Accepting a neo-Saussurean perspective (see Saussure 2002; Rastier 2005; Bouquet 2008) means aiming to achieve a systematic comparison between the whole system of phonetic differences and the whole system of semantic differences of a language, before starting comparisons with words in other languages. Indeed, the same sound (or the same meaning) can have different values in different languages, depending on the internal differential system of each. While the whole system of semantic differences does not have a formal �organization allowing a systematic description, the whole system of the phonetic �differences does: it is the phonological system with its combinatory rules. Thus, if one intends to �analyze systematically the signifier-signified relationship, one can base the analysis on the �phonological organization of the lexis. Besides, words are naturally organized with respect to their size. Shorter and �phonologically simpler words are on average more frequent, less numerous and

 Luca Nobile

� semantically more generic than the longer ones. Moreover, in shorter words the �relationship between signifiers and signifieds can be analyzed more easily and precisely because their linguistic value depends on a smaller number of distinctive features. So, if one intends to achieve the analysis of a whole differential system, one has to start with monosyllabic words. This method is not only epistemologically efficient but also �ontologically justifiable: it locates the starting point of the analysis at the space-time origin of the neuro-cognitive experience, that is the perpetuity of the now-here, and then it progresses towards the temporal duration of the syllables. For these reasons the initial corpus of our research is constituted by all the grammatical lexemes of Italian formed from one, two or three phonemes, excluding interjections, onomatopoeias and any other special type of word (see Table 1). These two hundred forms alone represent about 43% of all tokens in an average utterance (data extracted from De Mauro et al. 1993). Table 1.╇ The corpus includes Italian words formed from one, two or three phonemes Monophonemes V C

i e è ha ho o uh v’ t’ d’ s’ l’ l’ c’ c’ gl’ ch’ ch’ m’ n’

Biphonemes VV VC CV

io ai hai il in ed al ad han od un po’ fa fa fa’ fo fu vi vi ve ve va va’ vo ti te te tè tu di dì di’ da dà da’ do sì si se sé se sa so su su li lì le le le la la là lo lo re ci ci ce ce ciò già giù sci scià gli gli gnu chi che che che mi me me ma ne né no

Triphonemes CVV

CCV

CVC

pio… poi bei bèo... bai boa… bui bue bua fio fai fui via via vai voi tuo... dia Dio dèi… dei dai dài due duo zio… zoo sii sia sèi sei sai suo... lei lui rio… rio… reo… ciao cioè scio... scia coi cui mio... mai nei nèo… noi più può pro fra tre tra sfo... sta sta’ sto qui qua gru per pan bel ben bar far fan val van tal dir del dal dar dan don ser san san son son sub sud sul lor can col con gas men mal mar nel non



Words in the mirror 

This paper describes the entirety of only three subsystems of this corpus: the words formed from one vowel, the words distinguishing grammatical Â�persons, and the adverbs. In order to study the relationship between signifiers and Â�signifieds in these subsystems, we will spatialize them by means of a ‘phonosemantic Â�diagrammer’ (see Figure 3). A phonosemantic diagrammer is a schema Â�representing the Â�phonological system of a Â�language within the oral cavity of a speaker placed within a speech act. This allows the articulatory differences among the words with their logical-Â�semantic Â�differences to be compared taking the point of view of the Â�speaker’s body. For Â�simplicity, the auditory and proprioceptive dimensions will not be Â�systematically Â�considered in this paper.

Figure 3.╇ Phonosemantic diagrammer. The articulatory features of the Italian phonological system are represented within the mouth of a speaker placed in the speech act space

The picture represents the phonological system of Italian. The place of �articulation of the phonemes is represented on the abscissa: lips are on the left, the tongue is in the middle and the soft palate is on the right. Nasality is represented as a post-velar place of articulation, according to the fact that it is produced by a movement of the velum. The degree of aperture of the vowels and the manner of articulation of �consonants are merged and represented on the ordinate: the plosive consonants are at the top, the approximants are in the middle, the open vowels are at the bottom. �Voicing is �represented by writing the voiceless phonemes upon the lines and the voiced �phonemes

 Luca Nobile

under the lines according to the fact that the former are more closed than the latter. For simplicity, we have omitted some secondary traits, such as the labiality of the back vowels /u, o, f/, of the nasal consonant /m/ and of the prepalatal affricates /v/ and /ʤ/, as well as the dentality of the nasal /n/ (sometimes labiovelarity or velarity) and the laterality of the approximant /l/.

2.â•… Descriptive aspects 2.1â•… Vowel monophonemes1 Each of the vowel phonemes of Italian, except for /u/, represents, when taken in Â�isolation, at least one grammatical lexeme. These are: i /i/ ‘the’ {article; defined ; Â�plural}, e /e/ ‘and’ {conjunction ; copulative}, è /ˈ7/ ‘is’ {P32 ; essere}, a /a/ ‘to, at’ {preposition ; locative}, ha /ˈa/ ‘has’ {P3 ; avere}, ho /ˈf/ ‘(I) have’ {P1; avere}, o /o/ ‘or’ Â�{conjunction; disjunctive}. The central phoneme /a/ gives rise to two lexemes, a /a/ and ha /ˈa/, which can be distinguished only by stress. The back phoneme /u/ does not mean Â�anything, so it can represent the non-linguistic sound of the human voice in the Â�typical Â�onomatopoeia of the wild man’s call: uh! {meta-onomatopoeia}. The grammatical word most similar to /u/ is un /un/ ‘a’ {article; indefinite; singular}; as it is the only Italian monosyllable starting with /u-/ it can be employed as an alternative to uh! for a morphological comparison among homogeneous features. These items are a good starting point to analyze the linguistic value as a product of a system of oppositions because they constitute a closed, small, simple and important set. It is closed because it includes all the Italian words formed by one phoneme. It is small because it includes only eight words, connected by only ten minimal oppositions, easy to analyze in full. It is simple because all the oppositions can be described by only two distinctive features (aperture and place of articulation), which can be easily represented by a single two-dimensional picture. Finally, it is important because it includes the on average most frequent group of words among all the groups formed from words of the same number of phonemes: they alone represent about 7.7% of all tokens in a normal utterance; you can hear one of them every 12.9 words. If one puts the vowel monophonemes on the phonosemantic diagrammer, one sees that the distribution of the logical-semantic values, represented by distinctive morphological features, tends towards isomorphism with the distribution of the

1.â•… This chapter summarizes Nobile 2003; see also Nobile 2008. 2.â•… P1, P2, P3, P4, P5, P6 mean first, second, third, fourth, fifth, sixth person.



Words in the mirror 

phono-articulatory values. The analysis can be conducted on two levels (see Figure 4). First, one can describe the purely formal isomorphism between the ‘opposition axes’ of phonemes and lexemes; for example, the fact that the {conjunctions} e and o face each other on the same degree of aperture (an iconic diagram). Secondly, one can describe a more substantial isomorphism between the ‘opposition directions’; for example, the fact that the {positive, copulative} conjunction has a [front, single] place of articulation while the {negative, disjunctive} conjunction has a [back, double] place of articulation, and not the contrary (an imagic diagram).

2.1.1â•… Opposition axes The isomorphism between the axes can be summarized as follows: 2.1.1.1â•… A  perture.â•… The [degree of aperture] of the phonemes tends to distinguish the {grammatical category} of the lexemes: [closed : middle : open]â•… ≈â•… {article : conjunction : verb}

Note that articles are the only representatives of the nominal group in this system: they entirely take up the [closed] side of it, while all verbs take up its [open] side; �conjunctions are in the middle.

2.1.1.2â•… P  lace.â•… The [place of articulation] of the phonemes tends to distinguish the {morphological variation} of the lexemes inside each category: {definite : indefinite} {plural : singular} [front : back]â•… ≈â•… {copulative : disjunctive} {P3 : P1} {essere : avere}

All these semantic pairs are ‘polar’: they are oppositions between contraries, exactly like the articulatory pair [front : back]. The pair {P3 : P1} can be viewed as polar if it is regarded as an {not-I (P2, P3) : I (P1)} opposition. This is possible because the Â�Italian morphological P3 can also give, semantically, an impersonal (as in English) or a Â�courtesy P2 (as in German): è caldo ‘he is hot/it is hot/you are hot’.

2.1.1.3â•… Accent.â•… The [accent] distinguishes the {verbs}: {P3 avere : locative preposition} [stressed : unstressed]â•… ≈â•… {P3 essere : copulative conjunction} {P1 avere : disjunctive conjunction}

Only for the central open homophones (ha : a) is the stress the only distinctive feature; otherwise (è : e; ho : o) it is in addition to the degree of aperture.

 Luca Nobile

2.1.1.4â•… Centrality.â•… The phonological [central] position of /a/ (with respect to [front] and [back] vowels) is reflected by the semantic central positions of ha (with respect to è and ho) and a (with respect to e and o): {P3 essere : P3 avere : P1 avere} [front : central : back]â•… ≈â•… {copulative : locative : disjunctive}

In fact, the semantic value of ha {P3, avere} is formed by half of the features of è {P3, essere} and half of the features of ho {P1, avere}, while the semantic value of {locative} is logically intermediate between {conjunctive} and {disjunctive}. This relatively constant use of analogous phonological features to distinguish analogous logical-semantic features falls within the case of diagrammatic iconicity. Below one can see how this diagrammaticity gives rise in turn to secondary imagic structures.

Figure 4.╇ Iconic and imagic diagram of the vowel monophonemes. The distribution of the morphological features tends towards isomorphism with the distribution of the phonological features. Moreover, the latter is concretely similar to the former: the {positive} values are [front] while the {negative} are [back], and the {relational} values are [open] while the {unrelated} are [closed]

2.1.2â•… Opposition directions The isomorphism between the directions can be summarized as follows: 2.1.2.1â•… Aperture.â•… The [degree of aperture] tends to distinguish the {grammatical categories} in accordance to the level of topological-relational complexity. The more [open (F1 acute)] degree characterizes the {more-relational} grammatical categories,



Words in the mirror 

such as verbs and preposition, while the more [closed (F1 grave)] degree connotes the {less-relational} elements, such as articles and onomatopoeia. [closed : middle: open]â•… ≈â•… {article : conjunction : verb} less-relational : middle: more-relational

From the point of view of both physiological experience (eating, talking…) and Â�popular metaphors (opening to China, closed mind…), the motor pair [open : closed] can be considered a transparent image of the semantic pair {more relational  : Â�less-relational}, especially related to spatiality. Furthermore, the homologous timbre opposition [acute : grave] can easily evoke a space opposition {open : close} (and then {more-relational : less-relational}), because, for reasons of physical acoustics, a sound emitted indoors assumes a more grave timbre than when emitted outside (think of a voice from a cave, or of the noise of a train entering a tunnel).

2.1.2.2â•… Place.â•… The [place of articulation] tends to distinguish the {morphological variation} in accordance with a ‘positive’ or ‘negative’ spatial polarity. The place [front (F2 acute)] connotes the spatially {positive} side of each pair (plurality of i, additivity of e, convexity and exteriority of è), while the place [back (F2 grave)] connotes the Â�spatially {negative} side (singularity of un, subtractivity of o, concavity and interiority of ho). As a factor of redundancy, lip protrusion gives to all back vowels a contradictory gait (tongue back, lips forward), which adds a {separative} connotation to all {negative} elements (singulativity of un, separativity of o, transitivity of ho), as Â�opposed to the {unitive} connotation of the {positive} elements (collectivity of i, unitivity of e, intransitivity of è).  {definite : indefinite} [front : back]â•…â•… {plural : singular} â•… ≈â•…  {copulative : disjunctive} [palatal : labio-velar]  {P3 : P1} {essere : avere} positive-unitive : negative-separative

From the point of view of both physiological experience (advancing, going Â�backwards…) and popular metaphors (the progress, the recession…), the motor pair [front  : back] can be considered as a transparent image of the semantic pair Â�{positive : negative}, while the motor pair [palatal : labio-velar] (i.e. [single : Â�double] place of articulation) can be considered as a transparent image of the semantic pair {unitive : separative}. Furthermore, the timbre opposition [acute : grave] can Â�easily evoke a semantic opposition {positive  : negative} (think of the grave tones that Â�usually accompany the villain’s stage entrance), perhaps because, for reasons of physical acoustics, it urges perceptually more [high  : low] regions of the body (and this is another image of {positive : negative}; e.g. ‘high quality’ vs. ‘low quality’; see also Lakoff & Johnson 2003:â•›147).

 Luca Nobile

One must also remark that at least in the case of the verbs (è: ho; but not only) the motor pair [front : back] works also like an index distinguishing the physical positions of the grammatical subjects (P3 : P1) within the speech act space. I will return to this point later on.

2.1.2.3â•… Accent.â•… The [accent] distinguishes two complementary aspects of the Â�aperture as power to contain. The [stressed] accent connotes ha /’a/ ‘has’ as an active Â�{containing}, while the [unstressed] accent connotes a /a/ ‘at, to’ as a passive {contained} (see Â�Figure 5). lui ha

casa

lui

a casa

Figure 5.╇ Accent on /a/ differentiates two aspects of the power to contain

For example, looking at the picture of a man beside a house, one can distinguish by the accent Lui ha casa /ˈluiˈakˈkasa/ ‘He has a home’ {containing} from Lui a casa /ˈluiakˈkasa/ ‘Him at home’ {contained}. Likewise, watching a bunch of kids dividing themselves into teams, one can distinguish by the accent Marco ha Fabio / ˈmarkoˈafˈfabjo/ ‘Marco has Fabio’ {containing} vs Marco a Fabio /ˈmarkoafˈfabjo/ ‘Marco to Fabio’ {contained}. [stressed : unstressed]â•… ≈â•… {P3 avere : locative preposition} active-containing : passive-contained

From the point of view of both physiological experience (breathing reflects activity) and popular metaphors (activity can be stressful), the motor pair [stressed : unstressed] can be considered a transparent image of {active : passive}. Furthermore, as it opposes the respiratory cavity to the vocal cavity contained in it, the pair can easily evoke the conceptual opposition {containing : contained}.

2.1.2.4â•… Centrality.â•… The [central open] position of the phoneme /a/ corresponds to the central position of the notion of a {structured spatiality} as the power to contain or be contained (ha : a ‘has : at’), which is the base of any set-element or structure-item system (traditional symbol: ʘ).

[front : central open : back]â•… ≈â•…

{P3 essere : P3 avere : P1 avere} {copulative : locative : disjunctive}

positive direction : structured spatiality : negative direction

It is not surprising that the [open centre] of the vowel system depicts the idea of {structured spatiality}. Not only because the concept of centre (centre of a figure, of a town, of a problem…) implies the idea of a space contained in another space. But also



Words in the mirror 

because the opening of the mouth connects the internal cavities allowing us to contain (lungs, stomach, intestine, bladder…) with the external cavities allowing us to be contained (holes, homes, houses, habitations, hotels, halls, hangars…). The mouth is the anatomical junction between inner space and outer space, the combination of which give rise to our physiological experience of spatiality as structured. The pure exhibition of the oral cavity containing the tongue (/a/) can thus operate as the original figure of that concept (ha : a ‘has : at’).

2.2â•… System of persons3 Among the two hundred grammatical monosyllables of standard Italian, the persons differentiation system constitutes the largest set of forms. It includes 89 words divided into two major subsystems: the verbal one and the pronominal one. The subsystem of verbal persons also presents in its turn two main subtypes (see Table 2): the subtype of the verb essere ‘to be’ (P1 son4 /ˈson/, P2 sei /ˈs7i/, P3 è /ˈ7/) and the subtype of the verbs avere ‘to have’, fare ‘to do’, andare ‘to go’, sapere ‘to know’, dare ‘to give’, and stare ‘to stay’ (P1 ho, fo, vo,5 so, do, sto /ˈ-f/; P2 hai, fai, vai, sai, dÀi, stai /ˈ-ai/; P3 ha, fa, va, sa, dÀ, sta /ˈ-a/). As they are usually disyllabic, the plural grammatical persons of the verbs are excluded from this analysis. Table 2.╇ Differentiation of persons in verbs

essere avere, fare, andare, sapere, dare, stare

P1

P2

P3

/ˈson/

/ˈs7i/

/ˈ7/

/-ˈf/

/-ˈai/

/-ˈa/

The subsystem of pronominal persons presents four main subtypes (see Table 3): singular subject (P1 io /ˈio/, P2 tu /ˈtu/, P3 lui, lei /ˈlui, ˈl7i/), plural subject (P4 noi /ˈnoi/, P5 voi /ˈvoi/, P6 lor6 /ˈlor/), singular object (P1 mi, me /mi, me, me/; P2 ti, te /ti, te, ˈte/; P3 lo, la, si, se, sé /lo, la, si, se, ˈse/) and plural object (P4 ci, ce /vi, ve/; P5 vi, ve /vi, ve/; P6 li, le, si, se sé /ˈli, le, si, se, ˈse/).

3.╅ This chapter summarizes Nobile 2011. .╅ Son is a literary and somewhere popular form with apocope of the more standard (but disyllabic) sono. . ╅ Fo and vo are literary and somehow popular short forms of the more standard (but disyllabic) faccio and vado. .╅ Lor is a literary form with apocope for the more standard (but disyllabic) loro.

 Luca Nobile

Table 3.╇ Differentiation of persons in pronouns P1

P2

P3

Singular subject

/i-/

/t-/

/l-/

Plural subject

/n-/

/v-/

/l-/

Singular object

/m-/

/t-/

/l-/, /s-/

Plural object

/v-/

/v-/

/l-/, /s-/

2.2.1â•… Subsystems differentiation Considering the two subsystems, one can observe that (with the partial exception of son, sei and io) the verbal persons differentiation depends on the vowel termination, while the pronominal persons differentiation depends upon the initial consonant: [-vowel] : [consonant-]â•… ≈â•… {verbal} : {pronominal}

The initial consonant of verbs (/f-/, /v-/, /s-/, /d-/, /st-/) does not distinguish Â�persons but, in principle, the semantics of the verb (e.g. /fa/ : /sa/ = ‘does’: ‘knows’). The vowel termination of pronouns does not distinguish persons but several other things Â�including number, gender, case and logical role. In particular, if we exclude some interferences from the differentiation of genders (/l7i/ and /la, lo/), the logical role of the subject is usually distinguished by the presence of a back, grave vowel (/io/, /tu/, /lui/, /noi/, /voi/, /lor/), while the logical role of the object is distinguished exclusively by front, acute vowels (/mi/, /me/, /ti/, /te/, /si/, /se/, /vi/, /ve/, /vi/, /ve/, /li/, /le/). The distinction of persons is thus based on vowels in verbs and on consonants in pronouns. This can be viewed as an imagic diagram because the phonological relationship between opening and closure ([vowel]: [consonant]) is an Â�appropriate image for the ontological relationship between relations and entities ({verbal}: {pronominal}): [-vowel] : [consonant-]â•… ≈â•… {verbal} : {pronominal} open : closed

This diagram is particularly clear among monophonemes, where all verbs are vowels (ho /ˈɔ/, ha /ˈa/, è /ˈε/) and all pronouns are consonants (m’ /m/, t’ /t/, s’ /s/, l’ /l/, c’ /v/, v’ /v/);7 so, in minimal phrases, the object is distinguished from the subject-in-action as a consonant is distinguished from a vowel (l'ha /ˈla/ ‘it + he has’, l'è /ˈlε/ ‘it + he is’↜, m’hai /ˈmai/ ‘me + you have’). If one considers distinctive features, this diagram characterises the whole system of grammatical persons in Italian (see Figure 6).

.╅ These are forms with elision of the pronouns mi, ti, si, lo, la, ci, vi before words starting with a vowel.



Words in the mirror 

Figure 6.╇ Imagic diagram of the grammatical persons (I). The entities {pronouns} are closings [consonants] while the relationships {verbs} are openings [vowels]

Besides, the diagram is an extension of the one we already observed in vowel monophonemes, where the motor pair [open  : closed] distinguished the different degree of spatial and relational complexity of {verb : article}. Now the articles are replaced by pronouns and the closed vowels by consonants, but [open : closed] is still an imagic diagram of {more-relational  : less-relational}. The exceptions (/io/, /son/ and /s7i/) reflect the exceptional nature both of the locutor subject among Â�pronouns and of copula among verbs: the person who is doing the action of Â�speaking is distinguished by a (closed) vowel that approximates it to a verb, while the verb that predicates qualities, not actions, is distinguished by a (open) consonant which approximates it to a pronoun.

2.2.2â•… Subtypes differentiation Considering the subtypes, one can observe that both in verbs and in pronouns, the phonological distinction of persons is governed by non-symmetrical patterns, and that these asymmetries reflect asymmetries of semantics. 2.2.2.1 Verbal persons asymmetry.â•… In the phonology of verbs like avere, the first Â�person {P1} /-ɔ/ is distinguished primarily from the third person {P3} /-a/, which

 Luca Nobile

is then distinguished, secondarily, from the second person {P2} /-ai/ by adding a �phoneme, in accordance to the schema:

P1 /-/

P3 /-a/ P2 /-ai/

This asymmetry in the phonological differentiation of verbal persons seems to reflect an asymmetry in their logical-semantic differentiation. This is not only because, universally, the locutor, for himself, remains always {P1}, while others, for him, from time to time may become {P2} or {P3}, but also because in Italian, as remarked before, the morphological {P3} may mean, besides an impersonal (ha tuonato ‘it thundered’) and a semantic {P3} (ha tuonato ‘he thundered’), even a courtesy semantic {P2} (ha tuonato ‘you thundered’), so it can be opposed to {P1} as a generic {not-I : I}.

2.2.2.2â•… Pronominal persons asymmetry.â•… In the phonology of pronouns, only the first person distinguishes, by the initial phoneme, the singular subject (/ˈio/), the plural Â�subject (/ˈnoi/), the singular object (/mi, me, ˈme/) and the plural object (/vi, ve/). The second person radically distinguishes only the singular (/ˈtu, ti, te, ˈte/) and the plural (/ ˈvoi, vi, ve/) while the third person does not distinguish by the first phoneme singular from plural nor subject from object (/ˈlui, ˈl7i/; /ˈlor/; /lo, la/; /li, le/).8 So pronominal persons are asymmetrically differentiated in accordance to Table 4. 

Table 4.╇ Asymmetrical differentiation of pronominal persons

P1 P2 P3

Singular subject

Singular object

Plural subject

Plural object

/ˈi-/

/m-/

/n-/

/v-/

/t-/

/v-/ /l-/

This asymmetry in the phonological differentiation of pronominal persons seems to reflect an asymmetry in their logical-semantic differentiation. Indeed, if P3 is the

.â•… Although, among the objects, P3 distinguishes the reflective (/si, se/) and the masculine dative (/ˆi/). For reasons of simplicity here we overlook these two cases.



Words in the mirror 

only one which expresses singular and plural by the same initial phoneme (/l-/), it is also because its plural is the only one which can be obtained by multiplying singular (a genuine plural: {P6} lor = {P3} lui + {P3} lui). Conversely, if the first and second persons distinguish singular and plural by different initial phonemes, it is also because their plurals are obtained by the addition of other persons (hybrid plurals: {P4} noi = {P1 io + P2 tu} or {P1 io + P3 lui} or {P1 io + P2 tu + P3 lui}; {P5} voi = {P2 tu + P2 tu} or {P2 tu + P3 lui}). Also, if the first person is the only one distinguishing by the initial phoneme subject from object, it is also because, from the point of view of the locutor, the grammatical subject of {P1} is the only one referring to an actual, ontological Subject (i.e. the subject of the speech act and of the perception of the world), and, as such, it is radically distinct from the grammatical object. Conversely, if the second and third persons do not distinguish the subject from the object by the first phoneme, it is also because, for the locutor, the grammatical subjects of {P2} and {P3} are just ontological Objects (i.e. objects of his speech act and perception), and as such are not radically distinguished from their grammatical object.

2.2.3â•… Persons differentiation Considering the single persons, one can observe that, in all subsystems and subtypes, their phonological differentiation is isomorphic to their logical-semantic differentiation (see Figure 7).

Figure 7.╇ Imagic diagram of the grammatical persons (II). The speaker {P1} is [back], the listener {P2} is [front], the rest of the world {P3} is [central, open]

 Luca Nobile

In fact, in each subsystem, the first person {P1} is distinguished by the most [back] and [grave] (usually also [labial]) articulations (verbs: /-o, -f/; pronouns: /i-, m-, n-, v-/), the second person {P2}, by the most [front] (usually also [closed] and [acute]) articulation (verbs: /-i/; pronouns: /t-, v-/), the third person {P3}, by the most Â�[central] (and usually [open]) articulation (verbs: /-a/; pronouns: /l-, s-/). The phoneme /i/ Â�distinguishes {P2} in the vowel subsystem of the verbs, with respect to which it is [front], while it distinguishes {P1} in the consonant subsystem of the pronouns, with respect to which it is [back]. The distribution of persons within the articulatory space, thus, for both verbs and pronouns, singulars and plurals, subjects and objects, may be regarded as an imagic diagram of their distribution within the speech act space. Indeed, the differential relationship among the places of articulation [front  : central-open  : back] is concretely similar to the differential relationship among the places of the speech act {P2 : P3 : P1} since, from the point of view of the locutor, he himself {P1} always occupies the {back} region of his visual field, the allocutor {P2} normally occupies the {front} region, and the rest-of-the-world {P3}, distinctively occupies the elsewhere, that is it can be found not before or behind, but in the {aperture}. Note that the hybrid plurals {P4} /n-, v-/ and {P5} /v-/ (see above 2.2.2.2), while respecting the general distribution, tend to distinguish themselves from their Â�singulars {P1} /i-, m-/ and {P2} /t-/ by a greater proximity to the hybridizing persons, of which they incorporate some phono-articulatory features: /n-/ includes the Â�alveodental Â�articulation place of {P2} /t-/ and {P3} /l-/; /v-/ almost entirely includes the Â�phonological features of {P2} /t-/ and {P3} /s-/; /v-/ include the sonority and aperture features that characterize {P3} /l-/.

2.3â•… System of adverbs The system of adverbs involves 23 forms, traditionally classed in 8 categories, which one can group in 3 types (see Table 5): (I) the spatial adverbs, including adverbs of general place (1 – lì /ˈli/ ‘there [punctual]’; lÀ /ˈla/ ‘there [areal]’; qui /ˈkwi/ ‘here [punctual]’; qua /ˈkwa/ ‘here [areal]’; via /ˈvia/ ‘away’), vertical place (2 – su /ˈsu/ ‘up’; giù /ˈʤu/ ‘down’) and pronominal or vectorial place (3 – ci, ce /vi, ve/ ‘to there [standard spoken]’; vi, ve /vi, ve/ ‘to there [formal written]’; ne /ne/ ‘from there’); (II) the non-spatial adverbs, including adverbs of time (4 – poi /ˈpfi/ ‘then’; giÀ / ˈ3a/ ‘already’; mai /mai/ ‘never’), quantity (5 – più /ˈpju/ ‘more’; men /men/ ‘less’; [un] po' /umˈpf/ ‘a bit’) and manner (6 – ben /ˈb7n/ ‘well’; mal /mal/ ‘badly’); (III) the holophrastic adverbs (7 – sì /ˈsi/ ‘yes’; no /nf/ ‘no’) and the negation (8 – non / non/ ‘not’). Every category except negation distinguishes at least a pair of contrary



Words in the mirror 

semantic values describable as a {positive-convex  : negative-concave} opposition; some constitutes a triad with an asymmetrical element.

2.3.1â•… Types differentiation The three types of adverbs can be distinguished phonologically by the way they distinguish their own pairs of semantic values. The adverbs of place tend to distinguish their semantic values by symmetrical pairs sharing final vowels and opposing initial oral consonants (lì : qui; lÀ : qua; su : giù; ci : vi; ce : ve); when a third asymmetrical value appears, it gives rise to partial exceptions (via shares its /-i-/ with qui, lì and its /-a/ with qua, lÀ; ne shares /-e/ only with ce, ve). The adverbs of time, quantity and manner tend to distinguish their semantic Â�values by asymmetrical pairs sharing the (labial) place of articulation and opposing the (post-velar) nasal switch of the initial consonants (poi : mai; più : men; ben : mal); when a third value appears (giÀ, un po'), it gives rise to partial exceptions (giÀ has a prepalatal initial consonant /3/ but this is realized with a labial redundant feature; un poˈ /umˈpf/ contains both oral and nasal labial consonants). The holophrastic adverbs distinguish their semantic values by both initial [oral : nasal] (but not labial) consonants (s- : n-) and final [front : back] vowels (-i : -o).

Figure 8.╇ Imagic diagram of adverbs. Places are tongue positions within the oral space; times, quantities and manners are gestures of the lips and velum on the boundaries of that space; {positive} values are [front], {negative} are [back] (exceptions are omitted)

 Luca Nobile

So the initial consonants play a crucial role (see Figure 8): the adverbs of place are mostly distinguished by tongue positions within the oral cavity (/l-/, /s-/, /3-/, /v-/, /k-/; except for /v-/ and /n-/) while the adverbs of time, quantity and Â�manner are mostly distinguished by labial or nasal articulations, beyond or before the oral Â�cavity (/p-, b-, m-/; except for /3-/). This can be viewed as an imagic diagram because the oral cavity is an appropriate image to represent the speech act space (see already above 2.2.3) in opposition with its own borders and nasal cavity which represent other, non-spatial dimensions. The exceptions are meaningful. Among the adverbs of place, non-lingual articulations characterize the labial via ‘away’ (beyond the oral cavity) and the nasal ne ‘from there’ (before the oral cavity), the meanings of which justify their phonological Â�eccentricity. Besides, the only non-spatial adverb articulated inside the oral cavity is the perfective giÀ ‘already’, the more present and less virtual value among the adverbs of time. These exceptions indirectly confirm thus that the oral cavity tends to be employed as an image of the speech act space.

2.3.2â•… Categories differentiation Among the adverbs of place, categories are distinguished by vowels. Adverbs of Â�general place exclusively employ central vowel /a/ (via, lÀ, qua) and stressed front vowel /ˈi/ (via, lì, qui), adverbs of vertical place exclusively employ stressed back vowel /ˈu/ (su, giù) and adverbs of vectorial place exclusively employ unstressed front vowels /e/ (ve, ce, ne) and /i/ (vi, ci). Among the adverbs of time, quantity and manner, categories are distinguished more weakly. Adverbs of time have the only pair of descending diphthongs (poi, mai) but they can also have a simple vowel (giÀ). Adverbs of manner exclusively employ voiced consonants and have closed syllables (ben, mal) but they are not the only ones (men). Adverbs of quantity create partially palindromic structures with labial (più) and nasal features (men). 2.3.3â•… Semantic values differentiation Every {positive-convex: negative-concave} logical-semantic opposition correspond to a [front acute : back grave] phono-articulatory opposition between initial consonants (see Figure 8 and Table 5). So the whole system of adverbs is structured in accordance with the sensorimotor polarities we already found in grammatical persons Â�(concerning type I) and vowel monophonemes (concerning types II and III). In several cases Â�(categories 1, 3, 4, 5) a third value is opposed consistently to the main pair in different, asymmetric ways. In one case (negation) the value is unique and it is opposed to its own absence.



Words in the mirror 

Table 5.╇ Phonosemantic differentiation of the adverbs semantic value {+ +} 1 general place

via

{– –}

qui, qua

/ˈli ˈla/

‘there’ su /ˈsu/ ‘up’

{–} /ˈkwi ˈkwa/

‘here’

vi, ve /vi, ve/ ‘(to) thr’ (formal)

giù /'ʤu/ ‘down’ ci, ce /vi, ve/ ‘(to) thr’ (informal)

ne /ne/ ‘from there’

4 time

poi /'pɔi/ ‘then’

giÀ /'ʤa/ ‘already’

mai /'mai/ ‘never’

5 quantity

più /ˈpju/ ‘more’

(un) po' /um'pɔ/ ‘a bit’

men /ˈmen/ ‘less’

6 manner

ben /'↜渀屮bεn/ ‘well’

2 vertical place 3 vectorial place 

ii

{0}

lì, lÀ

/ˈvia/

‘away’ i

{+}

mal /ˈmal/ ‘badly’ sì /ˈsi/ ‘yes’

7 holophrastic iii

no /'nɔ/ ‘no’ non /non/ ‘not’

8 negation

[external]

[front]

[center]

[back]

[internal]

phonological value (place of articulation of the initial consonant)

2.3.3.1â•… General place (qui, qua, lì, lÀ, via).â•… The velar plosive consonant Â�/k-/, Â�having a more [closed] and [back] articulation inside the locutor’s mouth, Â�distinguishes the place qui, qua {near to the locutor}; the alveolar lateral Â�approximant /l-/, Â�having a more [open] and [front] articulation, distinguishes the place lì, là {far from the Â�locutor}; the labiodental fricative /v-/ having an Â�[external] Â�articulation on the Â�locutor’s lips, distinguishes the place via {away}; this diagram is very Â�similar to the diagram of persons (see above 2.2.3 and Â�Figure 7). Moreover, the [front, closed] vowel /-i/, articulated at a small distance from the palate, Â�distinguishes the Â�precision degree qui, lì {punctual} while the Â�[central,

 Luca Nobile

open] vowel /a/, articulated at a great distance from the palate, Â�distinguishes the Â�precision degree qua, lÀ {areal}; this diagram is similar to the diagram of the verbs and pronouns (see 2.2.1 and Figure 6). 2.3.3.2â•… Vertical place (su, giù).â•… The phono-articulatory opposition between /s-/ [open, front, acute] and /3-/ [closed, back, grave] can be regarded as an imagic diagram of the logical-semantic opposition between {up} and {down}, because in the common sense the most important prototypes of {up} and {down}, are often distinguished in this way. For example, the air and sky {up} are typically more {open} than the ground and earth {down}, the hands and mouth {up} are more {open} and {front} than feet and anus {down}, the calls of animals living {up} are more {acute} than calls of animals living {down} (because their bodies and resonance cases are on average smaller). Moreover, the proprioception of more [acute] sounds (called indeed ‘high’) mostly involves regions of the body located {up}, like the head, while more grave sounds (indeed known as ‘low’), mostly involve regions located {down}, like the belly. 2.3.3.3â•… Vectorial place (ci, ce, vi, ve, ne).â•… The phono-articulatory opposition between /v/ (or /v-/) [front, acute; single place of articulation] and /n-/ [back, grave; double place of articulation] can be regarded as an imagic diagram of the logicalsemantic opposition between {positive, unitive} and {negative, separative}; this is similar to the diagram of monophonemes (see 2.1.2.2). 2.3.3.4â•… Time (poi, giÀ, mai).â•… The phono-articulatory opposition between /p-/ [front, acute] and /3-/ [back, grave], can be regarded as an imagic diagram of the logical-semantic opposition between {future} and {past}, because in the walking experience future is typically before, while the past is behind (see above 2.3.3.1 and 2.2.3; as well as Lakoff & Johnson 2003:â•›41–43). Moreover, the phono-articulatory opposition between /p-, 3-/ [oral front] and /m-/ [nasal back] can be regarded as an imagic diagram of the logical-semantic opposition between {positive: Â�negative}, in accordance with previous diagrams (see 2.1.2.2 and 2.3.3.3; as well as Lakoff & Johnson 2003:â•›147). 2.3.3.5â•… Quantity (più, men,9 un po').â•… The phono-articulatory opposition Â�between /p-/ [front acute] and /m-/ [back grave], can be interpreted as an Â�imagic Â�diagram of the logical-semantic opposition between {more} and {less}, in Â�accordance with Â�previous diagrams concerning {positive : negative} oppositions

.╅ Men, ben, mal are forms with apocope of the more standard (but disyllabic) meno, bene, male.



Words in the mirror 

(see 2.1.2.2). Â�Moreover, the intermediate phonological position of the consonant cluster /-mp-/ with respect to initial consonants /p-/ and /m-/ (generated by the fact that the adverb poˈ need an article) can be viewed as an image of the intermediate semantic position of un poˈ {a bit} with respect to più {more} and men {less} (see above the case of /a/, 2.1.1.4). 2.3.3.6â•… Manner (ben, mal).â•… The phono-articulatory opposition between /p-/ [front, acute; single place] and /m-/ [back, grave; double place] to signify the Â�logical-semantic opposition between {well} and {badly} is another special case of the general diagram [front acute : back grave] = {positive : negative} (see  2.1.2.2). 2.3.3.7â•… Holophrastic (sì, no).â•… The phono-articulatory opposition between both /s-, -i/ [front, acute; single place] and /n-, -o/ [back, grave; double place], to signify the logical-semantic opposition between {affirmative}  and {negative} is another special case of the general diagram [front acute : back grave] = {positive : negative} (see 2.1.2.2). 2.3.3.8â•… Negation (non).â•… The phono-articulatory value /non/ can be considered an imagic diagram of the semantic value {negation}, as opposed to its absence, because it involves some of the most [back grave] elements of the Italian phonological system. 3.â•… Conclusion The systematic analysis of the vowel monophonemes, of the grammatical persons and of the adverbs of the Italian language highlights a strong tendency to isomorphism between the system of phono-articulatory differences and the system of logicalsemantic differences, such that the first appears to be in many cases describable as an imagic diagram of the second. In particular, the [open : closed] phonological pair tends to constitute an imagic diagram of several {complex : simple} semantic pairs, such as {more-relational : less-relational}, {action : object} and {areal : punctual}; while the [front : back] phonological pair tends to constitute an imagic diagram of several {positive : negative} semantic pairs, such as {definite : indefinite}, {plural : singular}, {unitive : separative}, {convex : concave}, {P3 : P1}, {far : near}, {future : past}, {more : less}, {well : badly}. This result, which has been built in a neo-Saussurean perspective, seems to be consistent with the framework outlined by the mirror system hypothesis, in Â�particular regarding the centrality of the motor patterns and the gestural-mimetic origin of Â�language. On the one hand, indeed, the mirror system hypothesis can explain how the articulatory diagrams can be transmitted to the cognitive system of the listener, since

 Luca Nobile

it provides mirror neurons able to reproduce the motor pattern of other’s actions from noise that it produces (Kohler et al. 2002; Fadiga et al. 2002; Rizzolatti & Craighero 2007, citing Liberman & Mattingly 1985). On the other hand, all the evidence brought to light by the mirror system hypothesis about the neurophysiologic and Â�evolutionary Â�contiguity between gestural mimicry and language articulation may receive a more Â�economical explanation. Not only can one say that, in phylogenesis, a mimetic system of gestural communication would have preceded and originated the verbal Â�communication system in use today, but also that, in ontogenesis and psychogenesis, this would continue to operate, at least in part, as a gestural and imitative system for the unconscious differentiation of the percept (see already Fónagy 2001:â•›347 about this). The results of the research thus seem to prove Hurford’s prediction wrong (2004), in accordance to which the mirror system hypothesis should not explain the efficiency with which the child acquires the semantic distinctions by means of the phonetic ones. In the light of the foregoing, indeed, the child’s cognitive efficiency seems to be explicable as an effect of the isomorphism between the differential systems of signifiers and signifieds. In other terms, the acquisition of semantic distinctions would be made easier by the acquisition of the phonetic distinctions, because the latter provide a diagram of the former. That is all the more so if the logical-semantic distinctions between concepts are embodied in their turn in sensorimotor patterns, as Gallese and Lakoff claim (2005). In this case, indeed, the tendency to isomorphism that we found between phono-articulatory space and logical-semantic space, would result in a tendency to isomorphism between two types of sensorimotor patterns (e.g. paraphrasing Gentilucci 2008, between the pattern governing the opening of the oral cavity, connected to /a/, and the pattern governing the opening of the hollow of the hand, connected to ha ‘has’). It is to be observed, however, that the Gallese’s and Lakoff ’s ‘concepts’ cannot be identified with the linguistic ‘values’ so far treated. While the former, indeed, are Â�positive images of the signified things, the latter are negative images of their differences. One emphasizing the identity of things, to which languages refer by different means, and the other the diversity of means, by which languages refer to the same things. From the perspective of a neo-Saussurean linguistics, only the ‘values’ belong properly to langue, and therefore to linguistics: the ‘concepts’ are extra-linguistic realities, Â�pertaining to psychology. The evidence for this is that while the ‘values’, as differential systems of Â�signifier and signified, necessarily differ for each language, the ‘concepts’, as mental images of things or actions, do not necessarily differ. One thus belongs, the other does not belong to langue. And langue consists only of ‘values’. But how, then, are the linguistic ‘values’ able to refer to psychological ‘concepts’?. An example will clarify this. In the experiment by Hauk et al. (2004), the Â�reading of action verbs related to the hands, feet or face, activates in the brain the motor Â�neurons corresponding to the muscles of the hands, feet and face, somotopically, that is in accordance with the map of the body existing at the surface of the brain, called



Words in the mirror 

the homunculus. The ‘concept’ of each of these action verbs is therefore embodied, inter alia, in the sensorimotor pattern that allows them to be performed. Moreover, that ‘concept’ lies in a system of topological relationships with sensorimotor patterns of other actions, a system represented by the cortical image of the homunculus. Now, among the examples proposed by Hauk et al. a prominent position is occupied by a paronomasia, chosen in order to minimize the interference of the articulatory Â�movement with the activation of the motor neurons due to concepts: kick, pick, lick. The reading of these three verbs activates sensorimotor patterns involving respectively the feet, hands and tongue of the homunculus. This paronomasia can illustrate the way in which one can think that ‘values’ refer to ‘concepts’. In fact, if one looks at the phonological dimension, kick is distinguished by a [voiceless, velar, plosive] initial consonant, pick by a [voiceless, labial, plosive], lick by a [voiced, lateral, approximant]. So kick and pick are phonologically a pair of Â�opposites, distinguished only by the place of articulation, while lick is an asymmetrical Â�element, distinguished as such by the manner of articulation and voicing. But this can be regarded as a diagram of the fact that feet and hands are symmetric to each other, while the tongue is not. Moreover, if one goes into detail, one also sees that kick is opposed to pick as a [back] to a [front] articulation place. If it was accompanied by other analogous oppositions in English (e.g. go : bow) this opposition could be regarded as an image of the difference between an action performed with the feet, located at the bottom or back of the body, and an action made with the hands, located on the top and front. Between these two polarities, lick occupies an intermediate place, since the tongue can be considered neither as front as the hands are nor as back as the feet are. Moreover, it is distinguished by other significant traits of /l-/, which is the only [voiced] consonant in this group (the relationship with voice distinguishes the tongue from hands and feet) and the only phoneme mobilizing the sides of the tongue itself. So, while the ‘concepts’ of kick, pick and lick are embodied in sensorimotor patterns implying feet, hands and tongue, the ‘values’ of /k-/, /p-/ and /l-/ seem to refer to these concepts in this way: they distinguish their relative positions within the oral cavity, such as the relative positions of sensorimotor patterns identifying concepts are distinguished within the body. One could object that, while the ‘sensorimotor concepts’ are probably similar in different languages, the linguistic values that distinguish them are dissimilar enough to falsify the above. For example, some languages may distinguish hands from feet, not by the phonological opposition [front : back] which depicts their locations, but by the phonological opposition [back : front] which does not. Here one may respond that, even if the ‘sensorimotor concepts’ of the hands and feet do not change, the ‘semantic values’ that distinguish them from other concepts may change, and this may justify such a change in the phonological value. For example, hands can be distinguished from feet not only as {upper/front : lower/back} limbs, but also as {grasping/pulling : carrying/pushing} limbs. The first pair of semantic values is adequately depicted by

 Luca Nobile

the phonological diagram [front : back], the second pair, by the phonological diagram [back : front]. This answer is correct. But it gives rise to the counter-argument that [front: back] may therefore depict simultaneously a thing and its opposite, which is a contradiction in terms: thus it would not depict anything. One can answer to this counter-argument that the simultaneity and Â�contradiction exist only in the cognitive domain of the Observer (an epistemological Third Â�Person listening to others’ utterances), not in the cognitive domain of the Locutor (an Â�epistemological First Person producing its own utterances), and that the cognitive domain of the Observer is inappropriate to understand the object (i.e. the relationship between phonological and semantic values). On the one hand, indeed, only the Observer has a simultaneous cognition of several languages, from the comparison of which the Â�contradictions emerge: the Locutor has cognition only of the language he is Â�speaking, the use of which shows no contradictions. On the other hand, only the Â�Locutor has a simultaneous cognition of both the phonological and semantic systems of the Â�language he speaks, while the Observer has cognition only of the phonological systems of the languages he observes, since, in most cases, he tends to infer their semantic systems by analogy to the one of his mother tongue (see Proverbio et al. 2009). It Â�follows that the Observer’s cognitive domain includes several phonologies inconsistent with its Â�semantics: the Observer does not understand the relationship between sound and sense and feels it is arbitrary. Conversely, the Locutor’s cognitive domain usually includes one phonology consistent with its’ semantics: the Locutor understands the relationship between sound and sense and feels it is natural. It should be added, however, that the Observer is always also a Locutor, and can therefore understand the cognitive perspective of the latter, while the reverse is not always true. If one intends to study language as an activity of the human body, not only as the product of that activity, it is necessary to distinguish the cognitive domains of the Â�Locutor and Observer, and assume the perspective of the former. Although in the Observer’s cognitive domain languages seem to make an arbitrary use of the Â�phonological values, in the Locutor’s cognitive domain phonological values exhibit a tendency to specialize themselves in a few semantically coherent distinctions, so that, as a whole, the differential system of sounds of a language can be described as an image of the differential system of its meanings.

References Albano Leoni, F. 2009. Dei suoni e dei sensi. Bologna : Il Mulino. Arnauld, A. & Nicole, P. 16621. La logique ou l’art de penser. Paris: Savreux. 1683. La logique ou l’art de penser. Paris: Desprez. Auroux, S. 2007. Introduction: Le paradigme naturaliste. Histoire Epistémologie Langage 29(2): 5–15.



Words in the mirror 

Benveniste, E. 1966. Nature du signe linguistique. In Problèmes de linguistique générale I, 49–55. Paris : Gallimard. Bopp, F. 1833. Vergleichende Grammatik, I. Berlin: Dümmler. Bottineau, D. 2002. Les cognèmes de l’anglais: Principes théoriques. In Le système des parties du discours, Sémantique et syntaxe, R. Lowe (ed.), 423–437. Québec : Laval. Bouquet, S. 2008. Ontologie et épistémologie de la linguistique dans les textes originaux de Â�Ferdinand de Saussure. Texto ! 〈http://www.revue-texto.net/index.php?id=1850〉. de Brosses, C. 1765. Traité de la formation méchanique des langues et des principes physiques de l’étymologie, 2 vols. Paris: Saillant. de Condillac, E.B. 1775. Grammaire. In Cours d’étude pour l’instruction du Prince de Parme, I. Parme: Imprimerie Royale. de Cordemoy, G. 1668. Discours physique de la parole. Paris: Lambert. Diogenes Laertius, 1533. De vitis, decretis et responsis celebrium philosophorum Libri decem, nunc primum excusi. Basileae: Frobenius. Fadiga, L., Craighero, L., Buccino, G. & Rizzolatti, G. 2002. Speech listening specifically Â�modulates the excitability of tongue muscles: a TMS study. European Journal of Â�Neurosciences 15: 399–402. Fischer, O. 1999. On the role played by iconicity in grammaticalisation processes. In Form miming meaning. [Iconicity in Language and Literature 1], M. Nänny & O. Fischer (eds). Amsterdam: John Benjamins. Fischer, O. & Nänny, M. 2001. Introduction: Veni, vidi, vici. In The Motivated Sign. [Iconicity in Language and Literature 2]. Amsterdam: John Benjamins. Fónagy, I. 1983. La vive voix. Essai de psycho-phonétique. Paris: Payot. Fónagy, I. 2001. Languages Within Language: An Evolutive Approach [Foundations of Semiotics 13]. Amsterdam: John Benjamins. Fresina, C. 1991. La langue de l’Être: Essai sur l’étymologie ancienne. Münster: Nodus. Gallese, V. & Lakoff, G. 2005. The brain’s concepts: The role of the sensory-motor system in conceptual knowledge. Cognitive Neuropsychology 21(0): 1–25. Gassendi, P. 1649. Animadversiones in decimum librum Diogenis Laertii qui est De Vita, Moribus, Placitisque Epicuri. Lugduni: Barbier. Genette, G. 1976. Mimologiques. Voyage en Cratylie. [Poétique] Paris: Seuil. Gentilucci, M. & Corballis, M. 2006. From manual gesture to speech: A gradual transition. Â�Neuroscience and Biobehavioral Reviews 30: 949–960. Gentilucci, M., Dalla Volta, R. & Gianelli, C. 2008. When the hands speak. Journal of Physiology 102: 21–30. Givón, T. 1985. Iconicity, isomorphism, and non-arbitrary coding in syntax. In Iconicity in Â�Syntax [Typological Studies in Language 6], J. Haiman (ed.). Amsterdam: John Benjamins. Givón, T. 1995. Isomorphism in the grammatical code: Cognitive and biological considerations. In Iconicity in Language [Current Issues in Linguistic Theory 110], R. Simone (ed.). Amsterdam: John Benjamins. Hauk, O., Johnsrude, I. & Pulvermüller, F. 2004. Somatotopic representation of action words in human motor and premotor cortex. Neuron 41: 301–307. von Humboldt, W. 1836. Über die Kawi-Sprache auf der Insel Java. Berlin: Königlichen Â�Akademie der Wissenschaften.

 Luca Nobile Hurford, J.R. 2004. Language beyond our grasp: What mirror neurons can, and cannot, do for language evolution. In Evolution of Communication Systems: A Comparative Approach, K. Oller & U. Griebel (eds), 297–313. Cambridge MA: The MIT Press. Jakobson, R. 1949. L’aspect phonologique et l’aspect grammatical du langage dans leurs Â�interrelations. In Jakobson 1963. 161–175. Jakobson, R. 1963. Essais de linguistique générale. Paris: Editions de Minuit. Jakobson, R. 1965. A la recherche de l’essence du langage. Diogène 51: 22–38. Junzo, K. 1988. La voix: étude d’ethno-linguistique comparative. Paris: EHESS. Kilani-Schoch, M. & Dressler, W. 2005. Morphologie naturelle et flexion du verbe français. Â�Tübingen: Gunter Narr. Kohler, E., Keysers, C., Umiltà, A., Fogassi, L., Gallese, V. & Rizzolatti, G. 2002, Hearing sounds, understanding actions: Action representation in Mirror Neurons. Science 297: 846–848. Lakoff, G. & Johnson, M. 2003[1980]. Metaphors We Live By. Chicago IL: The University of Chicago Press. Leibniz, G.W. 1710. Brevis designatio meditationum de Originibus Gentium ductis potissimum ex indicio linguarum. In Miscellanea berolinensia ad incrementum scientiarum, I. Berolini: Papen. Leibniz, G.W. 1765. Nouveaux essais sur l’entendement humain. In Oeuvres philosophiques latines et françoises. Amsterdam: Schrender. Liberman, A., and Mattingly, I.G. 1985. The motor theory of speech perception revised. Â�Cognition 21: 1–36. Locke, J. 1690. An Essay Concerning Humane Understanding. London: Basset. De Mauro, T., Mancini, F., Vedovelli, M. & Voghera, M. 1993. Lessico di frequenza dell’italiano parlato. Milano: Etas. Magnus, M. 2001. What’s in a Word? Studies in Phonosemantics. Ph.D. dissertation, University of Trondheim. 〈http://www.trimegistos.com/dissertation〉. Monneret, P. 2003. Le sens du signifiant: implications linguistiques et cognitives de la motivation. Paris: Champion. Monneret, P. 2004. Essais de linguistique analogique. Dijon: ABELL. Nobile, L. 2003. L’origine fonosimbolica del valore linguistico nel vocalismo dell’italiano Â�standard. Rivista di filologia cognitiva. 〈http://w3.uniroma1.it/cogfil〉. Nobile, L. 2008. The grammatical monophonemes of standard Italian: A structural isomorphism between phonological and semantic oppositions? Cognitive Phylology 1 〈http://padis2.uniroma1.it:81/ojs/index.php/cogphil/index〉. Nobile, L. 2011. Sémantique et phonologie du système des personnes en italien. Un cas d’iconicité diagrammatique?. In Sémantique et lexicologie, L. Begioni et al. (eds). Rennes: PUR. Philps, D. 2006. From mouth to hand. In The Evolution of Language, A. Cangelosi, A. Smith & K. Smith (eds.), 247–254. Singapore: World Scientific Publishing. Proverbio, A.M., Adorni, R. & Zani, A. 2009. Inferring native language from early bio-electrical activity. Biological Psychology 80(1): 52–63. Rastier, F. 2005. Saussure au futur: Écrits retrouvés et nouvelles réceptions. Texto ! 〈http://www. revue-texto.net/index.php?id=1816〉. Rizzolatti, G. & Arbib, M. 1998. Language within our grasp. Trends in Neurosciences 21: 188–194. Rizzolatti, G. & Craighero, L. 2007. Language and mirror neurons. In Oxford Handbook of PÂ�sycholinguistics, G. Gaskell (ed.), 771–785. Oxford: OUP. de Saussure, F. 1922. Cours de linguistique générale. Paris: Payot.



Words in the mirror 

de Saussure, F. 2002. Écrits de linguistique générale. Paris: Gallimard. Thiébault, D. 1802. Grammaire philosophique, 2 Vols. Paris: Courcier. Toussaint, M. 1983. Contre l’arbitraire du signe. Paris : Didier. Vico, G. 1744. Principi di Scienza Nuova. Napoli: Muziana. Wallis, J. 1653. Grammatica Linguae Anglicanae. Oxoniae: Lichfield. Waugh, L. 1993. Les degrés d’iconicité diagrammatique dans le lexique. Faits de langue 1: 227–234.

part ii

General theoretical approaches

Un mélange genevois Tacit notions of iconicity in Ferdinand De Saussure’s Writings in General Linguistics Jui-Pi Chien

National Taiwan University This paper explores some notions of iconicity in the newly published Writings in General Linguistics by Saussure. It begins by revisiting the imputed opposition between symbol and sign, and then proposes some contexts which serve to discover iconic traits in Saussure’s theory. In order to relate Saussure’s ideas to the attempts at defining ‘the iconic sign’, the author draws on the controversy over ‘cognitive type’ – whether we should rely on concrete objects in reality or some established laws and principles in interpreting images. Moreover, some conceptual tools like langue, sème and sôme are discussed within the broadened schemes of signs which Saussure proposed in his manuscripts. Finally, the paper concludes with an analysis of ‘the symbolic logic’ and ‘analogical reasoning’, which not only dissolves the opposition between symbol and sign, but also recognizes rules or principles as keys to our perception of similarities between systems of signs.

1.â•…Symbols, signs, and iconicity The definitions of symbols and signs have been a matter of controversy in the Â�Saussurean corpus of general linguistics. The edited text of Saussure’s Course in Â�General Linguistics gives the impression that its author is straightforward in denying the Â�theoretical Â�relevance of symbols to his conceptualization of semiology – so it appears that he Â�criticized the symbol as an ‘awkward’ notion, which takes for granted the motivated connections between signifiers and signifieds (Saussure 1959; 1983:â•›68–69). The science which Â�Saussure was developing appears to have been based solely on the more Â�avant-garde notion of ‘sign’ which precludes motivation. However, it has been discovered that Saussure actually left this issue open in his notes and manuscripts, which implies the coexistence of symbols and signs in the actual functioning of his system (Benveniste 1971:â•›43–48; Â�Todorov 1982:â•›268–270; De Cuypere 2008:â•›35). Accusations against Â�Saussure for the missing notion of iconicity – defined in terms of resemblance, Â�contiguity or Â�transitivity between

 Jui-Pi Chien

sounds and objects – can be attributed to the misfortune of his edited texts: editors like Bally and Séchehaye were enthusiastic about presenting the most definitive distinguishing features of a new science and method, but they somehow left out the Â�discursive Â�contexts which may give rise to an appreciation of the dialectical – oppositional but fruitful – relationships between symbols and signs. What is at issue here are the changing conditions of intentionality, motivation, and meaning rather than showing preference to a certain type of sign (Benveniste 1971:â•›11; Eco 1979:â•›217). Apart from the symbol/sign issue, the Saussurean notion of arbitrariness is often seen as the major obstacle in the attempt to theorize the ‘isomorphic’ – factual and Â�natural – composition between signifiers and signifieds (Jakobson 1990; Sebeok 1976:â•›152; Nöth 1998:â•›337). Arbitrariness is perceived as ‘the absence of Â�connection’ between sound and meaning or merely ‘a chance correlate’ of the two made by Â�nonnative Â�language users (Jakobson 1990:â•›410–411; Benveniste 1971:â•›44). To prove the necessity of a link, which supposedly has been dismissed by Saussure, Jakobson and his followers place Â�emphasis on a great amount of iconic formations as revealed in actual Â�speech-act circumstances, offering a variety of examples on both linguistic and poetic levels. Â�However, there have been voices arguing for a ‘tacit notion’ of Â�iconicity Â�underlying Â�Saussure’s accounts of signs (Hutton 1989:â•›71; De Cuypere 2008). These authors refuse to take isomorphism as the only legitimate way of conceptualizing iconicity in languages – they rather draw our attention to the aspect of grammar in Saussure, i.e. analogy, which functions in Â�accordance with the similarity and imitation of ‘units’ and thus gives rise to ‘extra meaning’ (De Cuypere 2008:â•›41–42, 48, 81–82). According to this alternative Â�viewpoint, we do not take imagic iconicity (or the iconic ground) as an end in itself: the Â�quantitative aspect – the affluent number of similarities between sounds and objects – is actually very marginal in the notion of iconicity in Â�Saussure. Instead, they suggest that we explore the laws, principles or criteria which govern the process of learning, recognizing, and interpreting patterns in order to introduce linguistic iconicity into Saussurean discourse. Following this approach, we value meaning and interpretation as a key to making the shift from imagic to linguistic iconicity – it takes into consideration the variations of functions, contexts, and discourses.

2.â•… Discourses on iconicity 2.1â•… Iconicity divided between linguistic and iconic signs In parallel with the controversy over the notion of arbitrariness, semioticians and Â�philosophers have argued against the undifferentiated endowment of the functions and features of linguistic signs to ‘icons’. In order to break away from the constraints of Â�linguistic concepts, some philosophers refined notions about the degrees of similarity



Un mélange genevois 

and quantified the amount of information carried by images in different media Â�(Wallis 1973; Goodman 1970). However, the idea that images refer to, imitate or Â�resemble things in the world appeared absurd to those who have followed the Peircean and Greimassian sense of ‘iconicity’. According to Peirce, Greimas, and their followers, the making of images is a quality, capacity or possibility, which is not yet settled within its medium of representation – relating images to the world of concrete objects does not help in generating an enlightening theory of art Â�(Bierman 1962:â•›249). Â�Therefore, they dispelled the “referential illusion” by putting forward Â�several Â�features of Â�iconicity: (1) it is deprived of denotations and connotations, i.e. the procedure of making sense out of linguistic signs is not applicable; (2) it is Â�asymmetric, which draws our Â�attention to its formative potential; (3) it stands for itself and exhibits its own meaning, i.e. it is Â�self-referential (Bierman 1962; Eco 1976; Sebeok 1976; Â�Greimas & Courtés 1979; Sonesson 1998). Emphasizing the fact that logical Â�semantics – based on the distinction between denotations and connotations – fails to hold good in defining iconicity, they suggested that we shift to Â�socio-biological contexts: just like all other signs, icons are thus construed as ways of modeling and representing the world which are somehow constrained by the specific societies or cultures from which they emerge (Greimas & Courtés 1979; Bouissac 1986, 2004).

2.2â•… Iconicity construed as cognitive type by Groupe µ Iconic signs defined within these strictures turn out to have essentially the same features as Saussurean linguistic signs – both are deemed arbitrary or unmotivated. Debates developed in the 1960s and 1970s expose a paradox in the nature of iconic signs: they are torn between the two extremes of being a similitude and a convention. In the early 1990s, the Belgian Groupe µ made a drastic move to explore visual Â�perception as an independent domain from that of language cognition: they declared that we abolish an excessive use of linguistic models in the study of visual arts and seek to redefine the nature of linguistic phenomena based on a thorough study of human Â�perception (Groupe µ 1992:â•›147). Moreover, they accused Umberto Eco of having wrongly Â�simplified the relation between images and objects (129–131, 142). Instead of Â�dismissing such relation as being merely ‘referential’ (which they thought happens only in the designating function of natural languages), they justified it as a two-way ‘transformation’ (132–133; Figure 1). On the one hand, producers select (t2) some features in their models and incorporate (t1) them into their iconic signs. On the other hand, for viewers to make good sense, they have to interpret the signs in two ways: the one (t2) leading them back to the cognitive types of producers; the other (t1) to the conventional ways of transforming objects. In addition to such two-way postulations, Groupe µ proposed a triadic model in which ‘types’ govern the transformation of objects into ‘referents’ (or cultured objects) and of images into

 Jui-Pi Chien

‘signifiers’ (or collections of visual stimuli) (Figure 2). They emphasized the fact that there are more transparent and stable links between types and referents than those between types and signifiers: referents can expand the paradigms (repertoires) of types while signifiers are so ambiguous – much more coded – that it takes conjecture to Â�recognize types through them (140–141). By introducing types into the structure of iconic signs, they annulled the naïve idea that iconic signs are merely ‘mirrors’ of objects. In their terms, iconic signs have transformed the perception of objects in accordance with some innate organizing principles (types); meanwhile, they are dynamic and even Â�rhetorical, because artists negotiate between their experiences of objects and Â�paradigms in their own styles – the former open up possibilities for inventions while the latter may set limits upon them.

Model

t1

t2

Producer

Figure 1.╇ Production of iconic signs (Groupe µ 1992:â•›132) Type Recognition (conjectural)

Stabilization (conventional)

Conformity

Conformity Referent

Signifier Transformation

Figure 2.╇ The model of iconic signs (Groupe µ 1992:â•›136)

2.3â•… Iconicity redefined as nuclear content and pragmatic law by Eco Later in the 1990s, Eco responded to Groupe µ and the rising cognitive science in a Â�cynical tone: he claimed that it is unfounded to take ‘cognitive types’ as a self-evident starting point because their nature has been made a black box (Eco 2000:â•›131–34, 138).



Un mélange genevois 

He suggested that we follow his common-sense approach in order to discuss Â�cognitive types (CTs) as constructed end products in the same fashion as trying to describe some unknown species. His approach introduced another term, nuclear contents (NCs), into the process of discovering CTs. According to Eco, NCs Â�manifest Â�themselves as our multimedial presentations – through drawings, gestures, sounds, and words, etc. – of something. These perceptual experiences or memories are made public when a human subject manages to convey them to another. Therefore, Â�contrary to what Â�scientists have postulated – biologically specific devices functioning in us all the time – Â�according to Eco, CTs remain in absentia most of the time. It is not until subjects start to Â�communicate, to recognize, and to refer felicitously – in order to suit circumstances – that CTs come to be in presentia. Following this argument, Eco reversed the fortune of CTs from being a black box to a white one, with which Â�subjects make visible and tangible their treasured ideas and help each other imagine things and formulate ideas. The notion led to the assumption that Eco had taken a ‘realistic turn’ in the 1990s as regards iconic signs. However, a more critical look into his statements reveals that he actually preferred another aspect of CTs, which is not limited to perceptual experiences or semioses (NCs) but can be ‘transmitted’ across cultures (Eco 2000:â•›138, 166, 175–176). Thus, it seems that Eco has undermined his argument – the presence of CTs is made up by NCs – and thus revived the controversy that he and others had initiated in the 1960s and 1970s: iconic signs are once again detached from immediate physical Â�environments and the actual world. Eco’s efforts are consistent since he has always argued that referents are far from concrete objects – as opposed to Groupe µ which retained the binary distinction that types are conceptual and categorical while referents are physical. Even if referents have to be rigid designations in extreme cases, they simply constitute the starting point of a larger and more evolving action, which is about making contracts – establishing ‘on trust’ relationships – with those who have before used certain signs in one way or another (Eco 2000:â•›293). Moreover, instead of feeling uncertain about and even Â�arguing over the objects referred to, viewers and speakers can examine the laws of modeling and representing which have evolved between communities: referents are about Â�conceptual and legal domains where subjects – across time and space – negotiate the uses of certain signs. With such an enlarged notion of referentiality, Eco Â�recognizes two prospects of iconic signs: (1) they follow and bring together certain laws; (2) they are effective in forming discourses, arguments, persuasions, and even creating effects of meaning and truth (Eco 2000:â•›420). Iconic signs restructured in this fashion stop relying on objects referred to as the only legitimate source of meaning; it is rather meaning – at its multiple levels and in different contexts – which guides viewers and speakers into applying laws and principles.

 Jui-Pi Chien

2.4â•… Meaning as effects of tangible laws It is not unusual to take meaning as effects which have emerged from procedures, organizations, and reorganizations. Semiotics does not take meaning as something given or something that can be easily told from the surface – it rather encourages more explanations so as to go deeper into the layers of a text or to discover its latently conflicting discourses. Hermeneutics, on the other hand, has taken intuition, Â�empathy, pre-understanding, and philosophy as primary conditions in dealing with a text. It stood against following a procedure to help with explanation and to enhance or vary understanding (Greimas & Courtés 1982 [1979]: 187–88). Paul Ricœur in the late 1980s and early 1990s endeavored to bring semiotics and hermeneutics into one Â�harmonious whole, advocating the necessity of generating meaning and understanding in accordance with laws and procedures. As he declared, the benefit, though somewhat extreme, is to transform interpretation and understanding as a matter of ‘pedagogy’: a process or reconstruction made up by intermediary stages, which we can repeat for ourselves and teach to others (Ricœur 1990:â•›117). Likewise, Eco took a similar turn in the way that he has made CTs – laws and referents – so stable and tangible that they can be ‘systems of instructions’ with which interpreters learn how they can manage texts (Eco 2000:â•›168, 171). Basing our development on such a methodological convergence between hermeneutics and semiotics, we will explore tacit notions of iconicity in Saussure.

3.â•… Exploring Saussurean ideas 3.1â•… Transferring the notion of langue into visual images As regards distinctions between linguistic signs and visual images, Saussure has presented fragmentary but insightful discussions in his notes and course lectures. Before he started elaborating on the schema of sign in his third course lecture (1910–1911), he hypothesized the existence of an ‘associative center’ in our brain, which links the Â�passive part (from hearing to associative) and the active part (from associative Â�center to Â�hearing) of interhuman communications. Such a hypothetical model facilitates Â�defining his object of study, langue, as a passive and social code, which governs the active but individual performances of speech (Saussure 1993:â•›67). In this context, Â�Saussure Â�apparently emphasized the aspect of the visual because he used ‘verbal image’ (image verbale) and ‘acoustic image’ (image acoustique) – in the sense of changing Â�sensory impression – before introducing still another term, ‘signifying’ (signifiant), into the entity of sign (139). Moreover, he was considering whether our visual impressions are more lasting – much more harnessed as a social code – than our speech: he took for



Un mélange genevois 

example the extreme case that those who were deprived of speech can still write (70). To prove the reality of langue from the perspective of the visual, Saussure proposed to amend his conceptualization as follows: In [la langue] we have a fact 〈object〉 of a concrete nature. These signs are not abstractions, even though they are in the mind. The set of associations ratified socially which constitutes [la langue] is located in the brain; it is a set of realities like other mental realities. It must be added that [la langue] is tangible, that is to say, translatable into [est traductible en] fixed images such as visual images, which would not be possible for acts of speech, for example. The utterance of a word involves all sorts of movements in the air, the muscles, etc. which it would be extremely difficult to identify. But in [la langue] there is only the acoustic image, and that can only be translated into [cela peut se traduire en] a fixed image.  (Saussure 1993:â•›71a)

The nature and functioning of langue fascinated Saussure: it is both an object which he examined and a heuristic tool which he used to invent aspects of his system. In his improved rationalization of langue, he showed a preference for the visual because of its obvious and reliable links to the social code – he implied that we can describe or gather the actual functioning of langue by ‘interpreting’ visual images. Â�Nevertheless, he still affirmed that sound impressions – though they are in a less advantageous Â�position than the visual in capturing the reality of langue – are crucial and Â�indispensible, and that langue is actually reinforced by both the visual and sound impressions. Such an idea, which portrays langue as a perceptual structure or a certain socially and Â�culturally defined condition, serves to argue for the interdependence of linguistic and Â�non-linguistic systems in his concept of iconicity. Although he initiated the Â�generalization of langue according to linguistic phenomena, he also wrote that the ‘vocal Â�apparatus,’ ‘the original contract,’ and ‘the initial convention’ are ‘of lesser (or least) importance’ when it comes to the co-existence of linguistic and non-linguistic systems in semiology Â�(Saussure 2006:â•›201–202). The semiological system in his terms blurs boundaries and Â�distinctions: its Â�fundamental features which are derived from linguistic systems offer an index to the ways many other systems function. Furthermore, it is far from the truth that Saussure has imposed ‘linearity’ or ‘unidimensionality’ on our perceptions and interpretations of visual patterns (Groupe µ 1992:â•›59). A look into his writings and miscellaneous notes reveals that he not only theorized the succession of units but also demonstrated the value of ‘juxtaposition,’ ‘pluriform divisibility,’ and ‘multidimensionality’ in appreciating musical compositions, magic lantern slides, allegorical paintings, and ideographic writings – he was actually considering ‘the spatial’ rather than the temporal as a more fruitful way of forming and understanding signs in general (Saussure 2006:â•›16, 64, 74–79).

 Jui-Pi Chien

a.

b.

Figure 3.╇ “Pluriform divisibility […] in a ‘direct visual’ system [… (a)] was compared with whatever meaning with (b)” (Saussure 2006:â•›75)

In addition, he has devised an alternative term such as sème to address both the vocal and visual aspects of langue (70). He made the point that we should abandon the principle of succession in time and references to concrete entities (sômes) for the Â�benefit of transforming visual figures into patterns (sèmes) in our mental space. The sème as a coming together of several different units or figures is entirely guided by meaning but can go beyond its constraints (Figure 3). The general characteristics of sème as a Â�meaningful entity or a synthesized impression are more ‘iconic’ than Â�‘logical’, which present asymmetrical and irregular patterns rather than segmentalizing Â�identical cuts or units. Iconicity in this instance is not about reducing visual stimuli to geometrical patterns or setting limits to what we can see – Saussure’s idea of a semiological Â�system constituted by sèmes is actually imbued with the potentials for multidimensional Â�combinations and recombinations.

3.2â•… The actual functioning of arbitrariness It should be appreciated that arbitrariness is a feature or character which Saussure observed and considered at the beginning of his inquiry – it does not explain every aspect of his system. We should not assume that he simply dismissed the possibility of a link between signifier and signified. Quite the contrary, such a connection or relation is every bit his concern (Saussure 2006:â•›58–59). After discovering Peirce in the 1950s, Jakobson indicated that ‘the iconic’ is of the essence in the study of languages and that this is a topic for the future. Mainly drawing on Peirce in simplified terms, he illustrated patterns of iconic similarities in words and sentences based on hierarchical (presuming a higher intelligence, such as the interpretant in Peirce), sequential (forming sentences in accordance with the temporality of actual happenings), and even simply coincidental grounds. Although he was favorable to Peircean terms, Jakobson admitted that certain Saussurean ideas are conducive to conceptualizing iconicity: (1) the ‘optimum semiotic process,’ i.e. semioses, fulfilled by arbitrary signs; (2) ‘relative motivation,’ that is analogical reasoning, which brings order and regularity to signs – in contrast with the



Un mélange genevois 

mass growing from absolutely arbitrary and unmotivated signs (Jakobson 1990: 412, 415). Nevertheless, weighing Saussure against Peirce, he still concluded that Saussure’s theory fails on iconicity unless we revise his dogmas – the arbitrariness of sign and the linearity of signifier. Jakobson proposed to improve our knowledge of them respectively as: (1) a “patent and compulsory […] yet latent and virtual” system; (2) “the dissociation of phonemes into distinctive features” (1990:â•›419–420). Because of the presumed lack of a clear notion on iconicity in Saussure – in the Course onomatopoeic words are recognized as symbols rather than signs – and the aphoristic style of his Â�writings, some points he has already developed can have been ignored and demand further interpretation. We may wonder whether Jakobson’s revised notions are in any sharp contrast to Saussure’s original ideas. The idea of distinctive features within a system provides a larger context to approach the functioning of arbitrariness. In Saussure’s third course lecture, a major portion of which touches upon langue (the rule or law summarized from different languages), he called his students’ attention immediately to the matter of choice and a posteriori connections between signifier (acoustic image) and signified (concept). For him, they not only shape the learning and speaking of a language but also explain the fact that every language is different in its own right. By employing the notion of arbitrariness to define his idea of signs, he was able to defend his system against the philosophy that there is a shared and predetermined origin or biological function of languages. Furthermore, within a specific system, he saw that arbitrariness functions in a solitary and limited fashion: it gives rise to the making of terms which cannot be related to – associated with – others (“il fait appel À rien” [Saussure 1967:â•›86]). The outreaching and sociable force en vivant goes to relative arbitrariness, i.e. analogy, which is able to reproduce shared units in different terms. Arbitrary and relatively arbitrary states of mind make up for a system – the former serves as a force of generating terms, while the latter explains the process. Although it has been generally conceived that the latter sets limits to the former, they both are governed by langue. Langue in this context means the rule or law of generating a posteriori, artificial, and reworkable connections within one sign entity and across many others.

3.3â•… The network of differences, games, and increase of relations We can confirm the existence of such seemingly contradictory but actually compatible states – between unmotivated and motivated signs – by looking into Emile Constantin’s notes along with Saussure’s diagrams and writings. At the beginning of his lecture on the nature of the linguistic sign (dated May 2nd, 1911), Saussure declared that he would reject approaches to langue which do not start from the two-term structure, i.e. a form and a meaning (Saussure 1967:â•›150; 1993:â•›75). However, as he gradually elaborated on this notion along with his critical tools, he indicated at one point that such structure (Figure 4) is not the “starting point” (“n’est donc pas initial”) in the

 Jui-Pi Chien

study of langue at all. Instead, he put forward the idea that the schema should be made to play out and to address phenomena of all kinds – beyond the formation of vocabulary only (“ce schéma peut entrer en jeu […]; il s’agit de n’importe quoi”, [dated July 4th, 1911] 1967:â•›264; 1993:â•›140). Looking further back in time, as revealed in his orangery manuscripts – written soon after he took the professorship in 1891 but not discovered until 1996 and published in 2002 as Writings in Â�General Linguistics – we find that already at this early stage he criticized the schema as something that is “entirely rough” and shows a “profound misconception” of langue (2006:â•›22, 24). He then proposed “a range of possible forms and possible meanings” (2006:â•›24), which interact through their distinctions but in no way correspond to each other, as a better representation of signs (Figure 5).

Signifié Signifiant Figure 4.╇ The schema of sign in the third course lecture (Saussure 1993:â•›139) General difference of meanings (only exists in relation to forms) General difference of forms (only exists in relation to meanings) Figure 5.╇ Saussure’s proposed approach to the study of signs in his orangery manuscripts (Saussure 2006:â•›24)

For unknown reasons, Saussure propagated the self-criticized structure (and schema) (Figure 4) to his students, putting aside most of his ramifications on the Â�network of differences. The ‘true thoughts’ (pensée intime) (Figure 5), which he Â�cherished and used as the starting point of his manuscripts, appear only in Â�elliptical lines towards the end of his third course lecture (Saussure 2006:â•›59;â•›1993:â•›140–143). Â�Furthermore, in his manuscripts he already conceptualized the functioning of a Â�system in accordance with the idea of “a complicated game leading to a final balance” (“les différences qui résultent du jeu compliqué et de l’équilibre final” [Saussure 2006:â•›43]). Such a game generates networks among signs without limit, starting point or any fixed point. It not only enables signs to absorb and locate new ideas at any time but also serves to change the boundary of general meaning from time to time. Despite having introduced such a prospect into the actual functioning of his system, he still affirmed that the whole thing is the ‘result’ of the schema which he has defined and criticized



Un mélange genevois 

(2006:â•›43). Langue in this context therefore governs both the formation of solitary entities (arbitrary signs) and the increase of relations – via analogical reasoning and making distinctions – among them. The manuscripts reveal how he then sought to define langue in more abstract terms in an attempt to develop it into a working tool for theorizing differentiations, multiplications, fragmentations, and discontinuities in the continuous currents of languages – he was planning a drastic move away from contemporary received notions about language.

4.â•…The search for ‘symbolic logic’ We gather from Saussure that the precondition for the formation of signs and their system is heterogeneity among elements – as he wrote in his orangery manuscripts, a sign entity appears in the form of unmotivated mixture (mélange) of diverse Â�elements (Saussure 2006:â•›4). Symbols on the other hand are thought to be derived from homogeneous elements, which are isolated and do not necessarily form an entity or system – it takes our motivation or intention to put them into certain categories (Ducrot & Todorov 1979:â•›102). The division of signs and symbols into such extremes has aroused suspicion among cultural and social semioticians concerning the practicality of Saussure’s methodology on two aspects: (1) the derogation of symbols – symbols are thought to be less complicated and more transparent than signs; (2) the process of signification solely governed by linguistic terms – the specificity of non-linguistic meaning cannot be fully appreciated (1979:â•›91–92). However, our examination of Saussure’s notes and manuscripts reveals that: (1) he was enthusiastic about theorizing the visual aspect of signs; (2) he has put forward the idea that the formation of visual sèmes is not at all constrained by linearity and meaning. Our hypothesis is that Saussure may have incorporated features of symbols into signs in his revised Â�conceptualization of langue, and thus drawn on the idea to reflect upon aspects of culture. Moreover, considering the fact that the metaphor of ‘a complicated game’ is dominant in his manuscripts and course lectures, it is essential to discuss how the play of a Â�Saussurean game helps explore symbols in a more abstract sense, i.e. the logic which not only structures Â�linguistic systems as such but also sheds light on Â�cultural systems as diverse as intersystemic and intermedial phenomena. It has been asserted that “Saussure took games to be secondary to rules” (Pietarinen 2006:â•›80), which justifies his forte in establishing synchronic linguistics and his weakness in conceptualizing diachronic linguistics. However, based on our reading of his notes and writings, which reveal some quite radical lines of thinking, this does not appear to be the whole truth. Games in their wide and far-reaching forms Â�actually take precedence over rules or laws on both synchronic and diachronic axes. At one point in his notes, Saussure reflected upon the formation and comprehension

 Jui-Pi Chien

of a ‘proposition’ in accordance with several perspectives: logical semantics, grammar, and linguistics. Logic (in its philosophical sense) and grammar share the same tools with subject and predicate in describing and analyzing a statement or a proposition. However, Saussure made the point that neither of the two approaches stops us from applying the third one: we are actually free to use a noun, a pronoun, an adjective, or an infinitive as either subject or predicate. From the perspective of linguistics, he saw that the Â�making of a syntagm actually allows various ways of combining, recombining, and even Â�omitting certain parts of speech. Such liberty manifested in the form of a game played inside speaking subjects occurs frequently in speech-act circumstances, but it does not affect our comprehension of messages – the ‘overall logic’ of a linguistic system stays unchanged (“Cela ne changerait rien aux choses logiques” [Saussure 2006:â•›66–67; 2002:â•›101]). As regards the diachronic axis, he argued that psychologists have misunderstood the nature of laws – langue is far from being primary, fixed, or conventional; it actually is a result of social exchange, which constantly creates a whirlpool (an anarchy, disorder or mutation) of signs beyond the choices of individual speaking subjects (“le résultat incessant de l’action sociale, imposé hors de tout choix” [2006:â•›67; 2002:â•›102]). These notes shed light on the fact that Saussure does not entirely neglect logic and social reality in conceptualizing his system. In his efforts to revitalize Saussure, Benveniste urged to find the best means to bring Peircean and Saussurean approaches together through defining the operation of ‘symbolic logic’: it avoids “segmentalizing a culture of a statement into discrete elements” and discusses “possible combinations” among these elements (Benveniste 1971:â•›11–12). The metaphor of game which he used for such unification calls attention to a Â�certain hidden substructure, latent mechanism or shared logic between the science of Â�linguistics and that of culture (1971:â•›14). Reconceptualizing the notion of iconicity in the context of ‘motivated’ combinations and recombinations, we find that it is concerned with the carrying out of a certain logic which Saussure practiced in his science: analogical reasoning functions to bring together diverse mediums, systems, and even innovations as a series of homology (A: B = C: D, etc.). Preliminarily, these equations reveal a certain ‘viewpoint,’ ‘proportion,’ shared ‘code,’ or practical ‘purpose’ with regard to social convention or cultural habit (Eco 1979:â•›195–216). We then move on to refine the logical possibilities between these analogous systems in accordance with the ways these systems exist and our knowledge of them: (1) they exist in the same way and we know them equally well; (2) they still exist in the same way, but we knew one of them before the others and use it to discover the rest; (3) we already knew one of them and transmit our knowledge of it unto the others which seemed irrelevant; (4) none of them exists nor is known and we invent one of them so as to bridge the connections (Itkonen 2005:â•›15–19). Analogy as a pedagogy or a system of instructions should not be considered as a tool which simply reduces the specificities of diverse mediums to the linguistic model. In a larger sense, it is a way of reasoning which serves our interests to extend our knowledge



Un mélange genevois 

to new mediums or domains. The sense of homogeneity or isomorphism which was thought to be missing in Saussure lies here: langue – in the sense of a mental model – has the intelligence to discover, imitate, and invent diverse systems. Although langue has been changing over the course of time, it sustains the equilibrium of a system and perceives its relevance to others at a certain time.

References Benveniste, E. 1971. [1939]. Recent trends in general linguistics. The nature of the linguistic sign. In Problems in General Linguistics, M.E. Meek (transl.), 3–15, 43–48. Coral Gables FL: University of Miami Press. Bierman, A.K. 1962. That there are no iconic signs… Philosophy and Phenomenological Research 23(2): 243–249. Bouissac, P. 1986. Iconicity and Pertinence. In Iconicity: Essays on the Nature of Culture: Â�Festschrift for Thomas A. Sebeok on His 65th Birthday, P. Bouissac, M. Herzfeld & R. Posner (eds), 193–213. Tubingen: Stauffenburg. Bouissac, P. 2004. Saussure’s legacy in semiotics. In The Cambridge Companion to Saussure, C. Sanders (ed.), 240–260. Cambridge: CUP. De Cuypere, L. 2008. Limiting the Iconic: From the Metatheoretical Foundations to the Creative Possibilities of Iconicity in Language [Iconicity in Language and Literature 6]. Amsterdam: John Benjamins. Ducrot, O. & Todorov, T. 1979. [1972]. Critique. Sign. In Encyclopedic Dictionary of the Sciences of Language, C. Porter (transl.), 90–92, 99–105. Baltimore MD: Johns Hopkins University Press. Eco, U. 1979. [1976]. Theory of sign production. In A Theory of Semiotics, 151–313. Â�Bloomington IN: Indiana University Press. Eco. U. 2000. [1997]. Kant and the Platypus: Essays on Language and Cognition, A. McEwen (transl.). London: Vintage. Goodman, N. 1970. Seven strictures on similarity. In Experience and Theory, L. Foster & J.W. Swanson (eds), 19–29. Cambridge MA: University of Massachussets Press. Greimas, A.J. & Courtés, J. 1982. [1979]. Semiotics and Language: An Analytical Dictionary, L. Crist, et. al. (transl.). Bloomington IN: Indiana University Press. Groupe µ. 1992. Traité du signe visuel: Pour une rhétorique de l’image. Paris: Seuil. Hutton, C. 1989. The arbitrary nature of the sign. Semiotica 75(1–2): 63–78. Itkonen, E. 2005. Analogy as Structure and Process: Approaches in Linguistics, Cognitive Â�Psychology, and Philosophy of Science [Human Cognitive Processing 14]. Amsterdam: John Benjamins. Jakobson, R. 1990. [1966]. Quest for the essence of language. In On Language, L.R. Waugh & M. Monville-Burston (eds), 407–421. Cambridge MA: Harvard University Press. Nöth, W. 1998. Ecosemiotics. Sign Systems Studies 26: 332–343.

Pietarinen, A.-V. 2006. Peirce’s game-theoretic ideas in logic. In Signs of Logic: Peircean Themes on the Philosophy of Language, Games, and Communication, V.F. Hendricks & J. Symons (eds), 77–102. Dordrecht: Springer. Ricœur, P. 1990. Between hermeneutics and semiotics: In homage to Algirdas J. Greimas. Â�International Journal for the Semiotics of Law 3(8): 115–132.

 Jui-Pi Chien de Saussure, F. 1959. Course in General Linguistics, C. Bally, A. Séchehaye & A. Riedlinger (eds), W. Baskin (transl.). New York NY: McGraw-Hill. de Saussure, F. 1967. Cours de linguistique générale. Fascicules 1 & 2, R. Engler (ed.). Wiesbaden: Otto Harrassowitz. de Saussure, F. 1968. Cours de linguistique générale. Fascicule 3, R. Engler (ed.). Wiesbaden: Otto Harrassowitz. de Saussure, F. 1983. Course in General Linguistics, C. Bally, A. Séchehaye & A. Riedlinger (eds), R. Harris (transl. & annotated). London: Duckworth. de Saussure, F. 1993. Troisième cours de linguistique générale (1910–1911) d’après les cahiers d’Émile Constantin/Saussure’s Third Course of Lectures on General Linguistics (1910–1911) from the Notebooks of Émile Constantin, E. Komatsu (ed.), R. Harris (transl.). Oxford: Pergamon. de Saussure, F. 2002. Écrits de linguistique générale, S. Bouquet, R. Engler & A. Weil (eds). Paris: Gallimard. de Saussure, F. 2006. Writings in General Linguistics, S. Bouquet, R. Engler & A. Weil (eds), C. Sanders, M. Pires, & P. Figueroa (transl.). Oxford: OUP. Sebeok, T.A. 1976. Contributions to the Doctrine of Signs. Bloomington IN: Indiana University Press. Sonesson, G. 1998. [Entries on] icon, iconicity, index, indexicality. In Encyclopedia of Semiotics, P. Bouissac (ed.), 293–297, 306–311. Oxford: OUP. Todorov, T. 1982. [1977]. Language and its doubles. Saussure’s semiotics. In Theories of the Â�Symbol, C. Porter (transl.), 222–246, 255–270. Ithaca NY: Cornell University Press. Wallis, M. 1973. [1968]. On iconic signs. Recherches sur les systèmes signifiants: Symposium de Varsovie, 1968. [Approaches to Semiotics 18], J. Rey-Debove. K. Fenton & T.A. Sebeok (eds), 481–98. The Hague: Mouton.

How to put art and brain together Mark Changizi 2AI Labs

Books on the psychology of art tend to identify known principles of cognitive or brain science, and then show instances of art exemplifying the principles. This approach suffers from two deep problems, the first is that most of the supposed principles of psychology are suspect, and the second that art is so rich and varied that one can find examples of it to fit any principle. Here I suggest a new way forward for putting art and brain together, one within the general theoretical framework that I call “Nature Harnessing.” It is that the arts have been culturally selected over time to be a “good fit” for our brain, and our brain has been naturally selected over time to be a good fit to nature… so, perhaps the arts have come to be shaped like nature, exactly the shape our brain came to be highly efficient at processing. I provide examples from my research on the structure of writing, speech and music.

A generation ago it was only a brave eclectic minority of psychologists and neuroscientists who dared to address the arts. Things have changed considerably since then. ‘Art and brain’ is now a legitimate and respected target of study and is approached from a variety of viewpoints, from reductionistic neurophysiology to evolutionary approaches. Things have changed so quickly that late 20th–century conversations about how to create stronger art-science collaborations and connections are dated only a decade later – everyone is already doing it! And the new generation of students being trained is at home in both the arts and sciences in a way that was rare before. Although we are all now more culturally comfortable bathing in conversations about art and brain, are we making progress? Has looking into the brain helped us make sense of the arts? Here I will briefly explain why I believe we have made little progress. And then I will propose an alternative route to understanding art and its origins. Perhaps the most common modus operandi in the cognitive and brain sciences approach to art is (i) to point to some known principle of brain science, and then (ii) to provide examples of art showing conformance with that principle. As fun as it may be to read explanations of art of this kind, the approach suffers from two fundamental difficulties – one on the brain side, one on the arts side.

 Mark Changizi

To start with the ‘brain’ difficulty, it is simply that we do not understand the brain. Although the field is jam-packed with fantastically clever experiments giving us fascinating and often valid data, there is usually very little agreement (or ought to be little agreement) about how to distill the data into broad principles. And the broader and higher-level the supposed principle, the more controversial and difficult-to-defend it is. Consequently, most of the supposed principles in the brain sciences remotely rich enough to inform us about the arts are deeply questionable. If we are so ignorant of the brain, why is the modus operandi above sometimes seemingly able to explain art? Because there is a lot of art out there, and it comes in a wide variety. Consequently, given any supposed principle from neuroscience or psychology, one can nearly always cherry pick art pieces fitting it. What very few scientific studies do is attempt to quantitatively gauge whether the predicted feature is a general tendency across the arts. The fundamental difficulty on the ‘arts’ side is that we often do not have a good idea what facets of art are universal tendencies that need to be explained. These difficulties for the brain and arts make the common modus Â�operandi a poor way to make progress comprehending art and brain. What initially looks like neuroscientific principles being used to explain artistic phenomena is, more commonly, suspect brain principles being used to explain artistic phenomena that may not exist. (A second common approach to linking art and the brain sciences goes in the other direction: to begin with a piece of art, and then to cherry-pick principles from the brain sciences to explain it.). How, then, should we move forward in our quest to understand the arts? Here I will suggest a path, one that addresses the brain and art difficulties above. The ‘arts’ difficulty can be overcome by identifying regularities actually found in the arts, whether universals, near-universals, or statistical tendencies. One reason large-scale measurements across the arts are not commonly carried out may be that any discipline of the arts tends to be vast and tremendously diverse, and it may seem prima facie unlikely that one will find any interesting regularity. With a strong Â�stomach, however, it is often possible to collect enough data to capture a signal through the noise. The ‘arts’ difficulty, then, can be addressed by good-old-fashioned data collection, and distillation of empirical regularities. But even so, we are left with another big problem to overcome. ‘Good-old-fashioned data collection’ involves more than simply collecting data. Which data should one collect? And which kinds of regularities should be sought after? There are effectively infinitely many ways of collecting data, and effectively unlimited ways of analyzing any set of data. Without theory as a guide, one is not likely to identify empirical regularities at all, much less ones that are Â�interesting. Good-old-fashioned theory is required in good-old-fashioned data Â�collection. We need predictions about empirical regularities, and then need to gather data in a manner designed to test the prediction. But this brings us back to our first difficulty, the ‘brain’ one. If we are so ignorant of the principles of the brain, then how can we hope to use it to make predictions about regularities in art?



How to put art and brain together 

We are, indeed, woefully ignorant of the brain, but we can make progress in explaining art. Here is the fundamental insight I believe we need: the arts have been culturally selected over time to be a ‘good fit’ for our brain, and our brain has been naturally selected over time to be a good fit to nature… So, perhaps the arts have come to be shaped like nature, exactly the shape our brain came to be highly efficient at processing. For example, perhaps music has been culturally selected to be structured like some natural class of stimuli, a class of stimuli our auditory system evolved via natural selection to process (see Figure 1).

Nature

Natural selection

Brain

Cultural selection

Culture

Figure 1.╇ The brain was shaped by natural selection for nature, and culture was shaped by cultural selection for the brain. (The arrows here indicate the “explanatory” direction, not the causal direction. E.g. we look to nature to understand the brain, and thus the arrow points from brain to nature.)

If the arts are as I describe just above – selected to harness our brains by Â�mimicking nature – then we can pursue the origins of art without having to crack open the brain. We can, instead, focus our attention on the regularities found in nature, the regularities which our brains evolved to competently process. I will suggest in a moment that we can do exactly this, and give examples where I have been successful at doing so. But let us first deal with a potential problem. Don’t brains have quirks? And if so, could the arts not tap into our quirks, and then no analysis of nature would help explain the arts? What do I mean by a quirk? Brains possess mechanisms selected to work well when the inputs to the mechanisms are natural stimuli, the kind of stimuli regularly encountered by our ancestors over evolutionary epochs. What happens when the inputs are not natural, i.e. what happens when the inputs are of a kind the mechanism was not selected to accommodate? (For example, when a screwdriver handle is used as a hammer.) The answer is, ‘Who knows?’ The mechanism never was selected to accommodate non-natural inputs, and so the mechanism may carry out some arbitrary, inane computation – a ‘quirk’ (e.g. the result of hammering a nail with a screwdriver handle is a quirk of the screwdriver, not part of the screwdriver’s design specifications). To grasp what the mechanism does on these non-natural inputs, we may have no choice but to crack open the hardware and figure out how it actually works. If the arts tended to be culturally selected to tap into the brain’s quirks, then nature wouldn’t help us, and we’d be bound to the brain’s enigmatic details in our grasp of the arts. There is, however, a good reason to suspect that cultural selection won’t try to harness the brain’s quirks, and the reason is this: quirks are stupid. When your brain

 Mark Changizi

mechanisms are running as nature ‘intended,’ they are exceedingly sophisticated machines. (Just as a screwdriver is good at screwing.) When they are run on inputs not in their design specs, however, the behavior of the brain’s mechanisms (now quirks) are typically not intelligent at all. (And, similarly, screwdrivers are ‘stupid’ hammers.) The quirks will usually be embarrassing in their lack of sophistication for any task because they were not designed for any task. And that is fundamentally why we expect the arts to have culturally been selected to tap into our ‘functional’ brain mechanisms, running roughly as nature intended. If we can set aside the quirks, then we can side step the brain in our attempt to grasp the origins of the arts. If I am correct about this, we can remove the most complicated object in the universe from the art equation! With the brain put on the shelf, the goal is, instead, to analyze nature, and use it to explain the structure of the arts. Is this really possible? And is not nature just as complicated as the brain, or, at any rate, sufficiently complicated that we are headed for despair? I think not. Nature is filled with simple regularities, many of them having Â�physical or mathematical foundations. And although it may not be trivial to discover them, our hopes should be far greater than our hopes for unraveling the brain’s mechanisms. Our presumption, then, is that our brains evolved to ‘know’ these regularities of nature, and if we, as scientists, can unravel the regularities, we have thereby unraveled the brain’s competencies. What regularities from nature am I referring to? Below I will present three brief examples from my research. Only one is explicitly about the arts, but all three concern the cultural evolution of human artifacts, and how they harness our brains via mimicking nature (see Figure 2). Nature

Brain

Culture

Figure 2.╇ By shaping culture to look like nature, culture will tend to end up shaped well for the brain. And, importantly, we scientists can hope to get a handle on this without having to understand the detailed brain mechanisms. The arrow cutting through the brain and going from culture to nature is meant to symbolize my nature-harnessing theoretical approach. It means that I pretend there is a single arrow like this, where culture has been selected to the shaped like nature. This is a simplification of the more detailed picture in Figure 1, and the greater simplicity is a boon to a scientist because the most complicated object in the universe – the brain –has been removed from the “equation”

The first concerns the origins of writing, and why letters are shaped as they are. Our visual systems evolved for more than a hundred million years to be highly competent at visually processing natural scenes. One of the most central features of these natural scenes was simply this: they are filled with opaque objects strewn about. And that is enough to lead to visual regularities in nature. For example, there are three



How to put art and brain together 

Â� junction types having two contours: L, T and X. Ls happen at many object corners, Ts when one edge goes behind an object, and these two are accordingly common in natural scenes. X, however, is rare in natural scenes. Matching nature, letter shapes with L and T topologies are also common across languages, but X topologies rare. More generally, the shapes found more commonly in natural scenes are those found more commonly in writing systems (see Changizi et al. 2006a; Changizi 2009). The second concerns the origins of speech, and why speech sounds as it does. Our auditory systems evolved for tens of millions of years to be highly efficient at processing natural sounds. Although nature consists of lots of sounds, one of the most Â�fundamental categories of sound is this: solid-object events. Events among solid objects, it turns out, have rich regularities that one can work out. To begin with, there are primarily three kinds of sound among solid objects: hits, slides and rings, the Â�latter occurring as periodic vibrations of objects that have been involved in a Â�physical Â�interaction (namely a hit or a slide). Just as hits, slides and rings are the fundamental atoms of solid-object physical events, speech is built out of hits, slides and rings – called plosives, fricatives and sonorants (but made via air-flow mechanisms mimicking solid-object event sounds). Similarly, just as solid-object events consist of a physical interaction (hit or slide) followed by the resultant ring, the most fundamental simple structure across language is the syllable, most commonly of the CV, or Â�consonant-sonorant form. More generally, and as I describe in my recent book, Harnessed: How Language and Music Mimicked Nature and Transformed Ape to Man (2011), spoken languages share a wide variety of solid-object event signatures. Written and spoken language look and sound like fundamental aspects of nature: opaque objects strewn about and solid-objects interacting with one another, Â�respectively. Writing thereby harnesses our visual object-recognition mechanisms, and speech harnesses our event-recognition mechanisms. Neither opaque objects nor solid objects are especially evocative sources in nature, and that’s why the look of most writing and the sound of most speech are not evocative (Changizi 2009). Music – the third cultural production I have addressed with a nature-harnessing approach – is astoundingly evocative. What kind of story could I give here? A natureharnessing theory would have to posit a class of natural auditory stimuli that music has culturally evolved to mimic, but have I not already dealt with nature’s sounds in my story for speech? In addition to general event recognition systems, we probably Â�possess auditory mechanisms specifically designed for the recognition of human behavior. Human gait, I have argued, has signature patterns found in the regularities of rhythm. Doppler shifts of movers have regularities that one can work out, and these regularities are found in music’s melodic contours. And loudness modulations due to Â�proximity predict how loudness is used in music. These results are described in Harnessed. For example, just as faster movers have a greater range of pitches from their directed-toward-you high pitch to their directed-away-from-you low pitch, faster tempo music tends to use a wider

 Mark Changizi

range of pitches for its melody (see Figure 3 for a summary of the nature-harnessing approach to explaining the origins of writing, speech and music). Nature

Culture

Opaque-object, 3D

Visual object recognition

Writing

Solid-object events

Auditory event recognition

Speech

Human movers

Auditory human movement recognition

Music

Figure 3.╇ The structure of my nature-harnessing arguments for writing, speech and music. For example, for the first row, writing shaped itself (via cultural selection) for our visual object Â�recognition mechanisms in the brain, and these mechanisms were, in turn, shaped (via natural selection) for recognizing three-dimensional scenes with opaque objects strewn about. Supposing that writing shaped itself mostly for the brain’s selected for functions and not the quirks, then writing is expected to principally shape itself to look like three-dimensional scenes with opaque objects. The arguments for the lower two rows have an analogous form

Many other aspects of the arts are potentially treatable in a similar fashion. For example, color vision, I have argued (Changizi et al. 2006b; Changizi 2009), is Â�optimized for detecting subtle spectral shifts in other people’s skin, indicating Â�modulations in their emotion, mood or state. That is, color vision is a sense designed for the emotions of other people, and it is possible to understand the meanings of colors on this basis, e.g. red is strong because oxygenated hemoglobin is required for skin to display it. The visual arts are expected to have harnessed our brain’s color mechanisms via using colors as found in nature, namely principally as found on skin. Again, the strategy is to understand art without having to unravel the brain’s mechanisms. One of the morals I want to convey is that you don’t have to be a neuroscientist to take a brain-based approach to art. The brain’s competencies can be ferreted out without going inside, by carving nature at its joints, just the joints the brain evolved to carve at. One can then search for signs of nature in the structure of the arts.

References Changizi, M.A., Zhang, Q., Ye, H. & Shimojo, S. 2006a. The structures of letters and symbols throughout human history are selected to match those found in objects in natural scenes. The American Naturalist 167: E117-E139. Changizi, M.A., Zhang, Q. & Shimojo, S. 2006b. Bare skin, blood, and the evolution of primate color vision. Biology Letters 2: 217–221.



How to put art and brain together 

Changizi, M.A. 2009. The Vision Revolution: How the Latest Research Overturns Everything We Thought We Knew About Human Vision. Dallas TX: Benbella. Changizi, M.A. 2011. Harnessed: How Language and Music Mimicked Nature and Transformed Ape to Man. Dallas TX: Benbella.

Image, diagram, and metaphor Unmined resource and unresolved questions Vincent Colapietro

The Pennsylvania State University It is far from implausible that some of C. S. Peirce’s other distinctions might eventually prove to be as heuristically fruitful as the distinction of icon, index, and symbol has proven itself to be. This distinction is based upon the relationship between a sign and its dynamical object, a relationship which constitutes the basis (or “ground”) of signification since it accounts for why anything is accorded the status of a sign. Anything functions iconically insofar as the basis of signification is an intrinsic relationship between the perceptible properties of a sign and its object. Anything functions indexically insofar as the basis of its signification is a causal connection, whereas anything functions symbolically insofar as this basis (or “ground”) is a disposition. These distinguishable functions are, more often than not, dynamically integrated in actual signs. They name not separate signs but distinct functions of irreducibly complex processes.

1.â•… Introduction “The most fundamental [division of signs] is”, Peirce claims, “into Icons, Indices, and Symbols” (Peirce 1931–1958, CP 2.227). This division has also proven to be his most influential classification. Its applicability to diverse fields of inquiry goes some distance toward rendering plausible the promise of articulating a truly general theory of signs. After defining an icon, Peirce distinguishes three kinds of iconic signs (images, diagrams, and metaphors).1 While the classification of icon, index, and symbol is based on the relationship between a sign and its object, that of image, diagram, and “metaphor” is also based on the relationship between an iconic sign2 and its dynamical

.â•… Strictly speaking, this is a classification of hypoicons, not icons. But, for our purpose, this distinction can be ignored for the time being. In the body of this paper, however, the meaning of this neologism will eventually be explained. .â•… “A possibility alone is”, Peirce asserts, “an Icon purely by its quality; and its object can only be a Firstness. But a sign may be iconic, that is, may represent its object mainly by its similarity,

 Vincent Colapietro

object, from that (in the case of images) in which the relationship is most immediate to that (in the case of metaphors) in which it is differentially mediated (mediated most markedly by difference; cf. Anderson 1984; Shapiro 1983). The text in which these distinctions are drawn is a notoriously difficult one: [Those icons (or hypoicons)] which partake of simple qualities, or First Firstnesses, are images; those which represent the relations, mainly dyadic, or so regarded, of the parts of one thing by analogous relations in their own parts, are diagrams; those which represent the representative character of a representamen (CP 2.277)3 by representing a parallelism in something else, are metaphors.

The passage in which Peirce presents this classification is, indeed, a tough nut to crack, but the fruit enclosed within a seeming impenetrable shell might be sustaining as well as delectable (if only we can reach it!). To alter the metaphor, a rich vein of ore might be contained in the depths and darkness of this text. My goal is accordingly to win access to a portion of this vein, so that we might begin to assess its richness or value. Though my aim on this occasion is to explore this less familiar trichotomy (image, diagram, and metaphor), I want to do so in reference to the more famous one (icon, index, and symbol). Hence, I am compelled to touch upon a classification of signs, no doubt, very familiar to most readers. But I do so in the interest of indicating that the less familiar classification is an unmined resource of potentially great riches,4 also in the interest of raising (but not resolving) two fundamental questions regarding central features of the Peircean account of iconic signs. As Max H. Fisch notes, Peirce came in time to realize that “signs do not fall into mutually exclusive kinds but that a given sign may have elements or aspects of all three kinds” (1986:â•›320).5 The iconic function of any complex sign (of anything we are likely

not matter what its mode of being” (CP 2.276). Moreover, its object itself can be something other than a firstness (e.g. the actual layout of a particular city). .â•… The division of firstness into a first firstness, a second firstness, and a third firstness might appear to pose a problem, since firstness precludes (or so Peirce seems to insist in other Â�contexts) such a recursive application of his categorial scheme. Since the division is not of pure firstness but the first of a third (an icon being a sign in its firstness, a sign being itself paradigmatically an instance of thirdness), there is here only an apparent difficulty. .â•… Unquestionably, diagrams, metaphors, and (to a less extent) images, as conceived by Peirce, have been topics on which there is a far from insignificant literature. But the Â�trichotomy of image, diagram, and metaphor has, with a few notable exceptions (e.g. Haley 1988, 1995 and Hausman 1989) not been explored in the depth and detail it deserves. Though the Â�explorations of these three types of icons (or hypoicons) in conjunction with one another and illuminating, more (arguable, much more) needs to be done. .â•… Elsewhere, Fisch writes, “just as the world does not consist of two mutually exclusive kinds of things, signs and non-signs, so there are not three mutually exclusive kinds of signs: icons,



Image, diagram, and metaphor 

to identify simply or unreflectively as a sign) is integrated with indexical and symbolic functions. Whether a photograph is an iconic or indexical sign (that is, whether it is a sign in which the iconic or the indexical function is predominant) depends partly �(perhaps wholly) on the context of interpretation and use. Insofar as the photograph is being taken as evidence of a crime, insofar as it is being used in a specific way in a �juridical context, its indexical function is of the utmost important.6 Its iconic �function is, even in this context, far from insignificant (the appearance of the individual in the photograph must, after all, markedly resemble that of the suspect). But, insofar as a �photograph is being experienced as a work of art (cf. Dewey 1989), the complex �interplay between its iconic and symbolic functions tends to eclipse its indexicality (cf. Ransdell). It does not seem, at least to me, especially important to say which of these functions �predominates; rather what is significant concerns, above all, the specific ways in which these distinguishable functions are integrated. For instance, the �shifting �position of a weathervane in response to the varying direction of the wind offers us an instance in which the iconic and indexical functions are intricately �connected. For �purposes of certain analyses or interpretations, however, we can focus on one of another of these functions more or less to the exclusion of other ones. But it is more often than not critical to remind ourselves that icon, index, and symbol in anything remotely approximating their categoreal purity are abstractable features of complex signs, not concrete or actual signs. To speak of an icon is, accordingly, a shorthand way of referring to the iconic function of a sign whose mode of signification always encompasses more than this specific function.7 What is most frequently worthy of attention and exploration is, then, the complex interplay among distinct functions.

indices, and symbols. These are rather elements or aspects of semeiosis that vary greatly in relative prominence or importance from semeiosis to semeiosis. We may therefore call a sign, for short, by the name of that element or aspect which is most prominent in it, or to which we wish to direct attention, without thereby implying that it has no other element or aspect of the other two kinds” (1986:â•›332–333). Note that the emphasis falls on process or activity. Again, Fisch succinctly makes this point when he insists: “The fundamental distinction is not between things that are signs and things that are not, but between triadic or sign-action and dyadic or dynamical action (5.473). So the fundamental conception of semeiotic is not that of sign but that of semeiosis […]” (330). .â•… “Photographs, especially instantaneous photographs, are very instructive, because we know that they are in certain respects exactly like the objects they represent. But this resemblance is due to the photographs having been produced under such circumstances that there were physically forced to correspond point by point to nature. In that aspect, then, they belong to the second class of signs [i.e. the class of indices], those by physical connection”, rather than intrinsic resemblance (CP 2.281). .â•… In manuscript 283, Peirce asks his readers to consider “the aggregate formed by a sign and all the signs which its occurrence carries with it. The aggregate will itself be a sign; and we may

 Vincent Colapietro

2.â•… Classification of hypoicons One of the classifications of icons or, more precisely, hypoicons seems to me and indeed other theorists (e.g. Shapiro 1983; Haley 1988, 1995; Hausman 1989) to be especially suggestive, even if Peirce failed to elaborate or even explore in much depth this trichotomy. A pure icon can only be a possibility: its mode of being is that of possibility.8 (Likewise, a pure index can only be an actuality: its mode of being is that of existence or actuality). But signs other than those whose mode of being is possibility can fulfill an iconic function. In order to distinguish a pure icon in the strict sense from the iconic function in its various (ontological) guises, Peirce devises the neologism hypoicon. In sum, then, a hypoicon need not be a possibility; it might be an actuality or a regularity. The implications of this for understanding the classification of image, diagram, and metaphor are far too complex and intricate to explore here, but they are worthy of investigation. In identifying the classification of image, diagram, and metaphor as a classification of hypoicons, we might take Peirce to be stressing that he is stepping away from a consideration of icons in their categorial purity. That is, he can be read as dealing here with embodied and embedded signs, indeed, ones whose embodiment and embeddedness are integral to their mode of signification.9

call it a perfect sign, in the sense that it involves the present existence of no other sign except such as are ingredients of itself. Now no perfect sign is in a static condition; you might as well suppose a portion of matter to remain at rest during a thousandth of a second, or any other long interval of time. The only signs which are tolerably fixed are non-existent abstractions. We cannot deny that such a sign is real; only its mode of reality is not that active kind we call existence. The existent acts, and whatever acts changes […]” (EP 2: 545 n25). .â•… Peirce’s classification of signs as iconic, indexical, and symbolic is one based on the Â�relationship between a sign and its object. In addition to this basis of classification, he devised one based on what a sign is in itself (i.e. in its firstness). A possibility might serve as a sign, but then so too might an actuality or ‘regularity’ or ‘law’. (The introduction of hypoicon needs thus to be considered most immediately in reference to this classification.) Finally, Peirce devised a classification of signs in terms of their interpretants (signs in their thirdness), yielding rheme, dicent, and argument. The intricate architecture of Peirce’s semiotic theory is perhaps nowhere more evident than in the complex ways in which these three tripartite classifications intersect. While Peirce’s various distinctions (say, that of icon, index, and symbol or that of posisign, sinsign, and legisign) might be detached from this architecture, it is time and again Â�illuminating to recall the context in which Peirce originally envisioned them. .â•… In a late manuscript, now published in EP 2 (Peirce 1998), Peirce draws a distinction between a medium and a vehicle of communication. “A medium of communication is something, A, which being acted upon by something else, N, in its turn acts upon something, I, in a manner involving its determination by N, so that I shall thereby, through A and only through A [this being the most paramount point, for our purpose], be acted upon by N” (EP 2: 391). He



Image, diagram, and metaphor 

The impulse toward formalization and taxonomization (so prominent in so much of Peirce’s investigation of signs) is thus countered by contextualization. The image as a sign is hypoiconic. An image is, as such, an image of something (though its object might be at any point in our engagement with this sign indistinct or indecipherable). It bears by virtue of its form a relationship to something other than itself, whether this relationship is perceived or not (for example, the form inscribed on the surface of a rock, in the instance of a fossil, resembles, however remotely, that of species of marine organisms present on the contemporary scene). In general, the relationship of signs to their objects is integral to the structure and functioning of signs themselves, not something primarily conferred upon them by an agency other than themselves. The implications of this are far-reaching and deep-cutting. None is more so than this one: intentionality in its most pervasive and primordial sense is not a feature of consciousness; rather it is an aspect of signs. Consciousness is intentional by virtue of being semiotic, rather than signs deriving their intentionality from consciousness. In any event, the image as a hypoicon (as something embodied in some medium and, as a consequence, embedded in some locale) is more than a pure possibility and thus, other than an ungrounded possibility. At the heart of signs, there is always to some extent a conflict between their specific modes of embodiment and the Â�unrealizable demand of being a purely diaphanous medium. What a sign is in itself, apart from all else (especially apart from its status and function as a sign), both facilitates and Â�undermines its defining function(s). What a sign is in its firstness transcends, Â�whatever mode of actuality it exemplifies and, moreover, whatever functions (semiotic or Â�otherwise) it fulfills. In other words, a sign is always more than a sign.10 What a sign is in the nexus of relationships in and through which it constitutes itself a sign (put more simply, what a sign precisely as a sign is) depends upon, first, an assemblage of Â�qualitatively Â�immediate

offers “a somewhat imperfect example” of what he means by this: “one animal, say a mosquito, is acted upon by the entity of a zymotic disease, and in its turn acts upon another animal, to which it Â�communicates the fever”. In this instance, “the active medium is in some measure of the nature of a vehicle, which differs from a medium of communication in acting upon the transported object and determining it to a changed location […]” (EP 2: 391). Peirce stresses, a sign “just so far as it fulfills the function of a sign, and none other, perfectly conforms to the Â�definition of a medium of communication […]”. But I am inclined to argue that, insofar as any medium is embodied and embedded in a context in which its particular mode of embodiment Â�implicates it in particular ways, it is inescapably a vehicle. In other words no sign is a perfectly Â�diaphanous medium in which something other than itself is disclosed, without distortion. .â•… Stated differently, a thing functioning as a sign is always more than a sign: its being is not exhausted in this function. But, then, such a thing is also never a mere thing, it being always more than it immediately is.

 Vincent Colapietro

features, second, a concatenation or purely existential Â�relationships, and, third, an Â�indeliminably expansive field of symbolic linkages.11 Returning to the image, there is in this instance of iconicity the perpetual threat of collapse (the image becomes so inherently absorbing as to draw us into its Â�qualitative immediacy) along with the periodic disruption of such absorption. There is, indeed, more than the threat of such collapse: there is, time and again, the Â�inevitable drift into monadic consciousness (or a dreaming state) in which the sense of otherness (or difference) is all but eradicated. There is typically then the Â�awakening from this absorption, the rude shock of the experiential world driving toward a Â�recognition, however tacit or unverbalized, of the distinction between image and imaged (the iconic sign and its dynamical object). The dynamism of images is partly due to the tension between such absorption and disruption. It is also partly due to the tensions and conflicts, struggles and threats, generated within the field of the image itself (think here of a nightmare). Images are hypoicons in which the qualitative features inhering in the images (but inhering in the images not so much as images of anything other than itself, as qualitative ensembles in their own right) tend to absorb our attention, rendering us oblivious of all else, whereas diagrams are hypoicons in which dyadic relationships predominate. Diagrams relate, principally by means of indexical signs, one set of relationships to another set. For instance, the relationship between the parallel lines on the hastily drawn map relate to just this relationship between two streets. In the diagram, then, we have on exhibit the relationship (in its simplest terms) of this to that, in its relationship to this other this and this other that. Such a sign directs our attention to what is other than itself, rather than (as does the image) entice our consciousness to forget all else but the image in its qualitative immediacy, in its sheer firstness. Of course, dyadic relationships within any given diagram can form intricate and complex patterns of such relationships (this runs parallel to that and, in turn, these parallel lines are perpendicular to those two other parallel lines, etc.). While qualitative immediacy is predominant in images and dyadic oppositions are paramount in diagrams, triadic relationships define metaphors, in Peirce’s sense (at least metaphors insofar as they are instances of hypoicons). Michael Haley argues, in my judgment convincingly, that Peirce’s characterization of metaphor in CP 2.277 is not intended as a global definition or adequate characterization of this species of sign (Haley 1988:â•›18–19). Rather Peirce is, according to Haley, considering metaphor only in a certain respect (metaphor insofar as it exemplifies a species of the genus hypoicon). The metaphor is as much a symbol as it is an icon or, stated more exactly, its symbolic function is not necessarily submerged in, or eclipsed by, its iconic work.

.╅ That is, any sign possesses its firstness, secondness, and thirdness.



Image, diagram, and metaphor 

That is, when Peirce characterizes metaphors as hypoicons “which represent the Â�representative character of a representamen [or sign] by representing a parallelism on something else” (CP. 2.277), he is not setting out to define this term, but only to characterize this instance of semiosis from a certain perspective. This is, in Peirce’s judgment (at least at one phase in the course of his development), what a metaphor is, but only insofar as it is such a species of icon. The classification of hypoicons into images, diagrams, and metaphors is a richly suggestive, though still largely unexplored,12 one. Its exploration might be undertaken in, at least, two quite different ways, one essentially archival and the other primarily imaginative. On the one hand, this exploration might ransack Peirce’s writings, Â�especially his unpublished manuscripts, for any clues these might provide for a systematic articulation of this seemingly quite fundamental trichotomy; then collate these scattered clues and integrate them into a systematic reconstruction of what might reasonably be considered, on basis of such textual evidence, to be Peirce’s basic position. On the other hand, we might use Peirce’s categorial scheme and scattered remarks, as they are more or less readily available; and, then, reconstruct imaginatively how Peirce would have developed these insights, had he himself had the time and interest to do so. Within the limits of this paper, I can do neither. I will rather Â�suggest, in an all too hasty manner, the most promising direction in which anyone committed to an imaginative reconstruction should travel; and, then, raise two questions, one pertaining to the very basis of iconicity, the other to the function of metaphor. What ties these three matters together is, above all else, the heuristic and exploratory character of each one. The imaginal, diagrammatic, and metaphoric (or, to use Haley’s expression, Â�metaicon) are aspects of iconic signs, in varying relationship to their dynamical object. The imaginal sign is the most fundamentally iconic, the metaphor the least so. With respect to immediate resemblance, the metaphor is at the greatest distance from its object (it being, indeed, an iconic sign in which different things are conjoined). The image resembles its object most straightforwardly and obviously, whereas the Â�metaphor (paradoxically) resembles its object through its difference from that object. Diagrams are, in this respect, intermediate, akin in some respects to images and in others to metaphors.

.╅ In asserting this, I do not intend to slight the work of many dedicated and careful �expositors (least of all, Joseph Ransdell and Michael Haley). But, I suspect, most (if not all) of them would agree with me: much exploration remains for us to do here. The group around the Iconicity in Language and Literature project have, however, explored in depth and detail some of the most basic distinctions, above all, the distinction between images and diagrams.

 Vincent Colapietro

For example, Virginia Woolf in To the Lighthouse (1927) describes the struggle of one of the characters to paint. In an early passage, Lily Briscoe asked herself: “But why different and how different?” – why and how was one of the other Â�characters (Mrs. Ramsey) different “from the too perfect shape one saw there” (76)? She raised these questions while “scraping her palette of all those mounds of blue and green which seemed to her like clods with no life in them now, yet she vowed, she would inspire them, force them to move, flow, do her bidding tomorrow” (76). The Â�reflection on the disparity between Mrs. Ramsey and “the all too perfect shape one saw there” is itself reflected in Lily’s judgment regarding the distance between “those mounds of blue and green” from the elusive image haunting her artistic imagination. In a much later passage, she returns to this canvas: Can’t paint, can’t write, she murmured monotonously, anxiously considering her what her plan of attack should be. For the mass loomed before her; it protruded; she felt is pressing on her eyeballs. Then […] she began precariously dipping among the blues and umbers, moving her brush hither and thither, but it was now heavier and went slower, as if it had fallen in with some rhythm which was dictated to her (she kept looking at the hedge, at the canvas) by what she saw, so that her hand quivered with life, this rhythm was strong enough to bear her along with it on its current. Certainly she was losing consciousness of outer things, and her name and her personality and her appearance, and […] her mind kept throwing up from its depths, scenes, and names, and sayings, and memories and ideas, like a fountain spurting over that glaring, hideously white space, while she modeled it with greens and blues.  (Woolf 1927:â•›237–238)

It is impossible, for me at least, not to imagine – thus to image – Woolf before the blank page as Lily Briscoe before the blank canvas, haunted by self-doubt (“[…] she heard some voice saying she couldn’t paint, saying she couldn’t write” [237]) but compelled by a drive stronger than any doubts. In this single passage from To the Lighthouse, we can discern nothing less than image, diagram, and metaphor in their interaction with one another. That is, we can discern the subtle interplay of the power of images to absorb us so completely that all consciousness of all else is lost, the power of Â�patterns of lines to forcefully direct our attention to kindred patterns elsewhere (so that Â�discerning the one pattern orients us in the midst of the other), finally that of differential Â�likeness (kinship marked by difference). Eventually I shall return to other examples from this novel, but for the moment the basis of images, diagrams, and Â�metaphors merits attention. At bottom, images, diagrams, and metaphors, as species of icons (as instances of iconicity), are based on resemblance or similarity. This much is well known. But Peirce offers what will likely strike most readers as a counterintuitive account of resemblance itself. This is not at all widely recognized. Given the dependence of icons on resemblance or similarity, the dependence of our recognition of resemblances or likeness on



Image, diagram, and metaphor 

something other than these perceptible features of phenomena is a matter of wide and deep significance.13 It is, accordingly, imperative to probe this topic.

3.â•… The Basis of Resemblance or Similarity Peirce is clear that resemblance or similarity is the basis for the relationship of iconic signs to their dynamical objects. Insofar as a sign functions as such by virtue of the similarity between its own inherent features and its dynamical object, that sign is an instance of iconicity. But what is the basis of resemblance or similarity itself?14 That is, what makes two objects (in particular, a sign and its object) resemble one another? This must sound, at least to some, like an odd question. Once we are at the level of immediately felt or discerned resemblance, there seems to be nothing more to say: either the resemblance is immediately perceived or it is not. We might even say that resemblance is a self-grounded or, perhaps, ungrounded relationship: whenever it obtains, nothing underlies or underwrites it, but it underlies and qualifies any number of other kinds of relationship. We might also say that the relationship of resemblance is a logically primitive one: while other relationships are derived from it, it is itself not derived from anything more basic or underlying than inherent resemblance. The problem with holding such a position is that Peirce stands at the portals of this position, explicitly and indeed emphatically barring entrance.15 In one of his Â�earliest yet strongest papers, he insists: “Every thought […] is, so far as it is Â�immediately Â�present, without similarity to any other, but incomparable with any other and absolutely sui generis” (CP 5.289). In being wholly incomparable, it is wholly inexplicable. But, in insisting upon the immediacy of thought or anything else being incomparable and thus inexplicable, Peirce does not intend to deny what everyone in fact experiences – the object thought (the object-thought or thought object) today strikingly resembles that of

.╅ The treatment of this topic can only be complex, since it cuts across a number of �disciplines. For example, it is in part a psychological question. In his account of resemblance, Peirce is both drawing upon, and reacting to, the associationalists (primarily to John Locke and David Hume). But it is also in part a strictly logical (or formally semeiotic) question, hence one removed from considerations of the particular form of animal mentality by which resemblance as such is realized or recognized. .╅ Resemblance is implicitly, but inherently, a triadic relationship, albeit such a �relationship in which the third in reference on which similarity is based is eclipsed by the qualitatively �arresting or manifest similarity of these terms. .╅ Of course, Peirce might be mistaken about this matter. But, in developing an account of his position, it is certainly requisite to take fully into consideration his explicit �pronouncements, especially about such critical topics.

 Vincent Colapietro

yesterday. He thus takes pains to stress: “Observe that I say in itself ” (CP 5.289 n2) – i.e. thought in its firstness. Then he makes a stunning suggestion: “I am not so wild as to deny that my sensation of red today is like my sensation of red yesterday. I only say that the similarity can consist only in the physiological force behind consciousness – which leads me to say, I recognized this feeling as the same as the former one, and so does not consist in a community of sensation” (CP 5.289 n2). It is telling that, in The Symbolic Species, Terrence Deacon argues along the same lines as Peirce. He notes: “Usually, people explain icons in terms of some respect or other in which two things are alike”. “But,” he immediately adds, “the resemblance doesn’t produce the iconicity. Only after we recognize an iconic relationship can we say exactly what we saw in common, and sometimes not even then” (1997:â•›74). Then Deacon makes an intriguing suggestion of his own: “The interpretive step that establishes an iconic relationship is essentially prior to this, and it is something negative, something that we don’t do” (ibid, p. 74, emphasis added). What we do not do is make any distinction between one thing and another. In this way, Deacon is shifting emphasis “from the relationship [of similarity or resemblance] to the process behind it” (pp. 75–76). What are we to make of these assertions? Everything resembles everything else in some respect. We perceive resemblances not simply or even primarily because they are present or perceptible,16 but because there are acts or processes of perceiving or instituting liknesses. Iconicity is at least proximately rooted in indexicality. But such indexicality is itself stabilized or solidified by nested relationships of (at least) a Â�triadic form. On the one hand, everything is similar to everything else, in some respect and to some degree. Hence, the basis for resemblance between (or among) various things is, in a sense, simply given with the character (the defining qualities, actual Â�relationships, and distinctive tendencies or propensities) of those things. On the other hand, the Â�discernment of likenesses, resemblances, and kinships is always a function of the Â�interests, objectives, and values of some agent. There is some basis in things, Â�however superficial or tenuous, for attributions of likeness. But (to take only one example of beings susceptible to the intimating force of iconic signs) we tend to discern the Â�likenesses we do because of our interests, concerns, and goals. We do not simply impose a grid upon an amorphous stuff. But, then, we do not simply read off patterns

.╅ I want, ultimately, to maintain that we perceive or discern resemblance because it is truly inherent in the relationship between one thing and another. But, out of the myriad possible resemblances between (or among) any complex phenomena, the ones that strike us as especially salient or pertinent are ones whose salience or pertinence calls for an explanation beyond the immediately felt sense of their inherent resemblance (cf. Bouissac 1986).



Image, diagram, and metaphor 

of resemblance from the manifestations of things, as though our presence and indeed constitution are irrelevant to the process of discerning likeness. Such a view cuts across realist and anti-realist accounts of icons, arguably doing justice to the complementary intuitions underlying these rival positions. The work of Paul Bouissac and others, drawing upon experimental studies, Â�provides us with a model for addressing how iconicity is yoked to indexicality, how Â�inherent resemblance can be linked to perspectival salience, as well as other questions. In Â�particular, Bouissac’s work on iconicity and pertinence points the way toward an explanation of the phenomenon of resemblance (see, e.g. 1986, 2009). Apart from his Â�intention, it also Â�provides invaluable resources for imaginatively reconstructing a Peircean account of one of the most basic relationships there is, the one in which Â�iconicity is grounded.

4.â•… A Topical Reinterpretation of Metaphorical Significance In discussions of metaphorical significance, much is ordinarily made of how this Â�figure of speech (more accurately, this mode of apprehension) involves exhibiting Â�previously hidden or unsuspected likenesses between apparently disparate Â�phenomena. What is rarely stressed, at least adequately, is the degree to which the Â�significance of Â�metaphors is topical and, with respect to the relevant topoi, heuristic and interrogative.17 The Â�likenesses presented by metaphors are, on such a topical reinterpretation, Â�primarily intimations of intelligibility yet to be disclosed, not one simply and fully grasped in our initial understanding of the semantic transference of an expression from one Â�discursive domain to another. There is, without question, an immediate (or Â�spontaneous) Â�understanding of, and indeed delight in, the more or less jarring Â�juxtaposition involved in any apt instance of metaphorical innovation.18 Several striking examples can be taken from Virginia Woolf ’s To the Lighthouse. At one point, we encounter this passage describing one of the characters (Wiliam Bankes, who is visiting his mentor, Mr. Ramsey): “He felt rigid and barren, like a pair of boots that have been soaked and gone dry so that you can hardly force your feet into them. Yet he had forced his feet into them” – that is, he must force himself into a set of Â�constraining and unwelcome expectations (“He must make himself talk. Unless

.â•… I am indebted to Donald Verene for this suggestion and for pointing out the relevance of Ernesto Grassi’s Philosophy as Rhetoric (1980) to this approach to metaphor. Though he made these suggestions in conjunction with another paper of mine, I have only now had the opportunity to take them up. .â•… Though this is not simply a version of Max Black’s (1962) interactionism, my view is in some respects akin to his and to other versions of interactionism.

 Vincent Colapietro

he were very careful, she would find out this treachery of his; that he did not care a straw for her [Mrs. Ramsey]…”) (135). As much as this metaphor of shrunken boots discloses a facet of this character’s personality, it directs us to attend to the specific ways in which this individual has already and has yet to force himself into the Â�constraint of such expectations. That is, the heuristic, topical function of this metaphor is (at least) as important as its directly disclosive and indicative force. Consider as an example also the reflections of another character (an example to which I have already referred), Lily Briscoe, who wants to establish the difference between Mrs. Ramsey and “the Â�perfect shape one saw there”: “How did she differ? What was the spirit in her, the Â�essential thing, by which, had you found a crumbled glove in the corner of a sofa, you would have known it, from its twisted fingers, hers indisputably” (76). The image of a pair of all too Â�constraining and hurtful boots into which one has to force one’s feet is, accordingly, to be Â�juxtaposed to that of a crumbled glove, at least momentarily misplaced, the twisted fingers of which are telltale signs. More than this, the fingers of such a glove bear a uniquely identifying signature (“ […] you would have know from it, from its twisted fingers, hers Â�indisputably” [76, emphasis added]). Now consider the thoughts of yet another Â�character, Mrs. Ramsey herself: “And, then, while the children rummaged among her things, she looked out of the window at a sight which always amused her – the rooks trying to decide which tree to settle on”. Such a Â�decision however proved to be difficult: “Every time, they seemed to change their minds and rose up into the air again, because she thought, the old rook, the father rook, old Joseph was her name for him, was [like her husband] a bird of a very trying and difficult Â�disposition” (121). While Lily Briscoe imagines herself to be “a bird for speed, an arrow for directness” (76) Mrs. Ramsey sees “old Joseph as “a disreputable old bird, with half his feathers missing. He was like some seedy old gentleman in a top hat that she had seen playing the horn in front of a public house” (121–23). Does seeing this bird in this light enable her (at least in part) to hide from herself (while revealing to readers) the fact of her husband having “a trying and difficult disposition” (121)? Finally, there is the example of Lily Briscoe, standing before a canvas, outdoors. “She looked blankly at the canvas, with its unpromising white stare; from the canvas to the garden. There was […] something she remembered in the relations of those lines cutting across, slicing down, and in the mass of the hedge with its green cave of blues and browns, which had stayed in her mind, which had tied a knot in her mind so that at odds and ends of time, involuntarily, as she walked along the Brompton Road, as she brushed her hair, she found herself painting that picture […]” (234). The author is in effect providing us with a diagram, a map of the relationships, spatial, mnemonic, and otherwise, in which this character, poised before the canvas, is enmeshed (cf. also Ljungberg 2010). From this description, we can more or less easily be situated in the particulars of her setting (she is framed by Woolf as a mass in relationship to other masses much as the painter is struggling to mass colors into vibrant and vital form). But Woolf



Image, diagram, and metaphor 

gives us more than a diagram of abstractable relationships; she provides us also with an array of absorbing images and heuristic metaphors (cf. Colapietro 2009:â•›110–111). Metaphors thrust upon our attention a perspective. They often seem to offer a compelling disclosure. I however want to argue that their characteristic power is not logical compulsion, but heuristic intimation. What makes such intimation heuristic is that metaphors, at least as I conceive them, open a field of inquiry,19 define a topos in one of the principal senses given to this term in Renaissance rhetoric (Grassi 1980). The foregrounding of surprising likenesses invites by its inherent power the exploration of additional likenesses – and differences. The Platonic dialectic of likeness and difference, envisioned as a logic of disclosure, was re-imagined in the Renaissance as a rhetoric of likeness and difference, ordained to the identification of topoi (these topoi being understood primarily as sites of inquiry). Logical conceptions are, from the Â�perspective of a regrettably neglected tradition, the offspring of rhetorical tropes (Grassi 1980). Heuristic fields in which the most rigorous methods of experimental inquiry are instituted, employed, and refined are themselves discoverable only by the metaphorical imagination. What could be more Peircean than a conception of Â�metaphor as an instrument for opening a field of inquiry, rather than decisively defining or adequately characterizing phenomena? The aptness of a metaphor is more than anything else an unrealized but (seemingly) trustworthy promise: the delight and insight it accords upon initial encounter points toward the open-ended work to be undertaken for the sake of even deeper understanding. The more apt the metaphor, the more trustworthy does the implicit promise feel, so much so that its issuance seems inseparable from its fulfillment. But, like all other signs, metaphors keep their word by what they generate, not what they are or seem to be at the moment of utterance. Consider, for example, R. W. Emerson’s characterization of symbols and language: Here is the difference between the poet and the mystic, that the last nails a symbol to one sense, which was a true sense for a moment, but soon becomes old and false. For all symbols are fluxional; all language is vehicular and transitive, and is good, as ferries and horses are, for conveyance, not as farms and houses are, for homestead. (Emerson 1982:â•›279)

Or consider another of his arresting images: Language is fossil poetry. As the limestone of the continent consists of infinite masses of animalcules, so language is made up of images or tropes, which now, in their secondary use, have long ceased to remind us of their poetic origin.  (Emerson 1982:â•›271)

.╅ As I will suggest at the conclusion of this paper, this metaphor is itself the one Peirce used to characterize his work in semeiotic (see CP 5.488; or EP 2: 413).

 Vincent Colapietro

The aptness of these metaphors is, at once, immediate and beckoning (or �promissory). That is, their aptness if immediately felt but ultimately promissory. It is not secured in the moment of inaugural utterance or initial reception, but unfolds over time.

5.â•… Conclusion While icons are signs by virtue of a resemblance between their own inherent qualities and the qualities of their dynamical object (in brief, while resemblance is the basis of iconicity), the question of the basis of resemblance itself is a pressing and, to date, unresolved one. My own inclination (tentative and undeveloped, as it is) is to take Â�seriously the hypothesis that the basis of resemblance involves nothing less than all three functions (the iconic, the indexical, and the symbolic). While metaphors are instances of semiosis in which hidden or unnoticed similarities or affinities are brought to light, a topical conception of metaphorical significance suggests not so much the compelling disclosure of such similarities or affinities, as Â�heuristic intimations. That is, such a conception of metaphors identifies, above, all topoi, places to be investigated, sites to be mined or field to be cultivated. The very metaphor used by Peirce to describe his work in semeiotic might be taken to imply just this. “I am,” he confessed, “a pioneer, or rather a backwoodsman, in the work of clearing and opening up what I call semiotic […] and I find the field too vast, the labor too great, for a first-comer” (CP 5.488; also in EP 2, 413). It may even be the case that parts of the field he opened have grown over again, requiring the renewal of his inaugural efforts. This may be especially true of the trichotomy of image, Â�diagram, and metaphor. An archival investigation promises to be rewarding. In the spirit of Peirce, however, an imaginative reconstruction of his categoreally nuanced position might be even more apposite. My hope is that I have offered here some Â�specific Â�suggestions about how this delicate yet promising task might be carried out.

References Anderson, D.R. 1984. Peirce on metaphor. Transactions of the Charles S. Peirce Society 20:  453–468. Black, M. 1962. Metaphors and Models. Ithaca NY: Cornell University Press. Bouissac, P. 1986. Iconicity and pertinence. In Iconicity: Essays on the Nature of Culture: Â�Festschrift for Thomas A. Sebeok on his 65th birthday, P. Bouissac, M. Herzfeld & R. Posner (eds), 193–213. Tübingen: Stauffenburg. Bouissac, P. 2009. Static images as agencies: Performative, algorithmic and normative functions. Open Semiotics Resource Center. 〈http://www.semioticon.com/people/articles/Static%20 Images.htm〉.



Image, diagram, and metaphor 

Colapietro, V. 2009. Pointing things out: Exploring the indexical dimensions of literary texts. In Redefining Literary Semiotics, H. Veivo, C. Ljungberg & J.D. Johansen (eds), 109–133. Newcastle upon Tyne: Cambridge Scholars Press. Deacon, T. 1997. The Symbolic Species: The Co-Evolution of Language and the Brain. New York NY: W.W. Norton. Dewey, J. 1989. Art as Experience. The Later Works of John Dewey, Vol. 10. Carbondale IL: SIU Press. (Cited as LW 10). Emerson, R.W. 1982. Selected Essays [Penguin American Library], L. Ziff (ed.). New York NY: Penguin Books. Fisch, M.H. 1986. Peirce, Semeiotic, and Pragmatism: Essays, K.L. Ketner & C.J.W. Kloesel (eds). Bloomington IN: Indiana University Press. Grassi, E. 1980. Philosophy as Rhetoric: The Humanist Tradition. University Park PA: Â�Pennsylvania State University Press. Haley, M. 1988. The Semeiosis of Poetic Metaphor. Bloomington IN: Indiana University Press. Haley, M. 1995. Iconic functions of the index in poetic metaphor. Journal of Pragmatics 24: 605–25. Hausman, C.R. 1989. Metaphor and Art: Interactionism and Reference in the Verbal and Â�Nonverbal Arts. Cambridge: CUP. Ljungberg, C. 2010. Dynamic instances of interaction: The performative function of iconicity in literary texts. Sign System Studies: The semiotics of resemblance. (Special issue). Peirce, C.S. 1931–1958. The Collected Papers of Charles Sanders Peirce, vols. 1–6, C. Hartshore and P. Weiss (eds); vols 7–8 Arthur W. Burks (ed.). Cambridge, MA: Belknap Press of Â�Harvard University Press. [Cited as CP]. Peirce, C.S. 1998. The Essential Peirce, Vol. 2, the Peirce Edition Project (ed.). Bloomington IN: Indiana University Press. (Cited as EP 2). Ransdell, J. 1986. On Peirce’s conception of the iconic sign. In Iconicity: Essays on the Nature of Culture: Festschrift for Thomas A. Sebeok on his 65th birthday. P. Bouissac, M. Herzfeld & R. Posner (eds), 53–74. Tübingen: Stauffenburg. Shapiro, M. 1983. The Sense of Grammar: Language as Semeiotic. Bloomington IN: Indiana Â�University Press. Woolf, V. 1927. To the Lighthouse. San Diego CA: Harcourt Brace Jovanovich.

part iii

Narrative grammatical structures

The farmers sowed seeds and hopes Element order in metaphorical phrases Yeshayahu Shen & Elad Kotzer Tel Aviv University*

Conceptual prominence plays an important role in determining word order in metaphorical sentences: conceptually prominent items tend to precede less prominent ones. In: ‘the farmers sowed seeds and hopes’ the order of the two noun in the conjunctive noun phrase (seeds and hopes) seems more natural than its inverse (hopes and seeds) since seeds (the more concrete noun) is conceptually more prominent than hopes. This linear precedence of prominent items iconically mirrors their ‘cognitive precedence’, namely, the fact that they are retrieved from memory before less prominent counterparts (Kelly et al. 1986, Osgood & Bock 1977). Three factors contributing to conceptual prominence affect word ordering: abstractness – concrete terms tend to precede more abstract ones; animacy – animate terms tend to precede non-animate ones; and salience – salient terms tend to precede less salient ones. We discuss the findings of a series of psychological experiments and corpus studies that lend support to this argument.

1.â•… Introduction Conceptual prominence plays an important role in determining word order in literal sentences, in that conceptually prominent items tend to precede less prominent ones. Consider, for example (1) and (2): 1. ‘the man bought an orange and a grapefruit’ 2. *‘the man bought a grapefruit and an orange’ The ordering of the nouns “an orange and a grapefruit” in (1) seems more natural than the corresponding ordering in (2) since an orange is a more prototypical, hence Â�conceptual more prominent, fruit than a grapefruit (Kelly et al. 1986). Arguably, there is an iconic correspondence between the linear ordering of the nouns in the *╇ This research was supported by the Israel Science Foundation administered by the Israel Academy of Sciences and Humanities, nº 969 – 07.

 Yeshayahu Shen & Elad Kotzer

Â� conjunctive noun phrase and their conceptual prominence. It has been previously pointed out (e.g. Rosch 1999; Kelly et al. 1986) that prominent items are retrieved from memory before less prominent ones (e.g. Rosch 1999); placing the former before the latter in the sentence linear structure, then, mirrors this cognitive retrieval precedence (Kelly et al. 1986; Osgood & Bock 1977). Note, that this preference of a certain ordering over its inverse is not restricted to literal sentences such as (1) and (2), but is exhibited by metaphorical sentences as well, as illustrated in 3 and 4: 3. She came to the party with a friend and a smile. 4. The farmers sowed seeds and hopes. The two nouns in (3) (‘friend’ and ‘smile’) belong to the same syntactic category (nouns) and have the same syntactic role in the sentence (direct object), yet the Â�natural order is the one in which ‘friend’ appears before ‘smile’ rather than vice versa. A similar situation exists in (4). Although ‘seeds and hopes’ are both nouns and have the same syntactic role in the sentence (direct object), people prefer the order ‘seeds and hopes’ to the opposite one, ‘shirts and sorrow’. These sentences are based on the figure of speech known as the ‘zeugma’, in which one word stands in the same relation to two other words, but with two different meanings. The most common form of zeugma is a sentence with a verb that has two direct objects, such as ‘I lost my virginity and the key to my apartment’ or ‘She forgot her bicycle and her love’. There are also less frequently used forms, as in the examples ‘Her hair was long and comfortless’, in which there is a different meaning relationship between the subject and each of its two adjectives. In many zeugmas, such as (3) and (4), people prefer one ordering of the nouns that serve as objects over the other. We call this the ‘canonical order’. In this paper, we discuss three factors contributing to the conceptual prominence of items in conjunctive noun phrases that affect this preference: salience, animacy, and abstractness. Past research on the order of words in literal expression (such as [1])1

.â•… Most of these studies are not directly relevant to the conjunctive structure illustrated in (1) and (2), which is the focus of the present paper. For example, a considerable number of the Â�linguistic studies in this area (e.g. Hawkins 1983) focused on rules applying to the order of various syntactic categories, such as the differences in the basic syntactic order in Â�different languages (e.g. SVO – Subject-Verb-Object, as in English and Hebrew, versus Â�SOV – Â�Subject-Object-Verb, as in Turkish and Japanese). Such cases involve the order of units with different syntactic roles (e.g. subject and object) or units that belong to different syntactic Â�categories (e.g. noun and verb). Other studies (Bock and Warren 1985) investigated the factors that affect the syntactic role played by words in the same syntactic category, such as what Â�determines which of two nouns is chosen to represent the grammatical subject and which the object (e.g. ‘The boy is walking with the dog’ versus ‘The dog is walking with the boy’).



The farmers sowed seeds and hopes 

revealed various factors that affect this order, including prosodic factors (e.g. words with fewer syllables tend to appear first; Cooper & Ross 1975), semantic factors (e.g. words that denote time tend to appear before words that denote space; Cinque 1999), more general semantic-pragmatic principles such as ‘me first’ (Cooper & Ross 1975), and even social-pragmatic principles such as the placement of men’s names before women’s names (e.g. ‘Romeo and Juliet’ rather than ‘Juliet and Romeo’; cf., Wright et al. 2005). Here we will focus on three conceptual factors – salience, animacy, and Â�abstractness – which many studies have shown to be important in determining the order of the Â�linguistic elements in literal expressions. These factors have almost never been Â�investigated in metaphorical expressions – in particular, in the zeugma. The importance of studying metaphorical language (and figurative language in general) is reflected in the fact that this sort of language has recently become a major research area in the Â�cognitive sciences, especially cognitive linguistics and cognitive Â�psychology, as a result of the Â�realization that it provides an important window onto the Â�understanding of Â�various sorts of cognitive processing, including language processing (e.g. Gibbs 1994). One of the most basic questions that arises in the study of metaphoric language is whether Â�various sorts of cognitive processing are common to the understanding of both literal and metaphorical language (e.g. Giora 2002). This is why it is important to investigate the factors that affect word order in metaphorical expressions and see if they are similar to those that operate in literal expressions. The major aim of this study is to present the results of a number of studies that investigated whether these conceptual factors have the same affect on word order in metaphorical expressions as in literal expressions. Section 2 discusses the three Â�factors and their effect on word order in literal expressions. Section 3 presents three pilot studies that were the first to investigate the connection between salience and element order in metaphorical expressions. One study explored various poetic corpora, while two additional investigated different aspects of the memory for and interpretation of metaphorical expressions. Section 4 presents several follow-up studies that investigated the other two factors as well as salience, in a deeper and more comprehensive exploration. The conclusion, Section 5, discusses various implications of these findings for other types of figurative expression, as well as the relationship between literal and metaphorical language.

2.â•… S  alience, animacy and abstractness, and their effect on word order in literal expressions 2.1â•… Salience Salience is the characteristic of information in the mental lexicon that allows us to retrieve the most available information most quickly (Giora 2003). Among the factors

 Yeshayahu Shen & Elad Kotzer

contributing to the salience of particular meanings of words are the frequency with which the word is used with that meaning, and the familiarity, conventionality and prototypicality of that meaning. Although these factors are distinct, the measures of salience obtained when they are studied are highly correlated with one another (for a detailed description, see Giora 2003). Salience itself, as a complex concept, has not yet been researched as a factor affecting word order, although some of its components have been. For example, a study by Kelly et al. (1986), which investigated the effect of the degree of prototypicality on word order, found that words for more prototypical objects tended to appear first, so that a sentence like ‘The man bought an orange and a lemon’ was perceived as more natural than ‘The man bought a lemon and an orange’ because oranges are more prototypical fruits than lemons. In addition, Kelly et al. found that subjects who read sentences in the non-standard word order, with the less Â�prototypical object appearing first, tended to reverse the order of the items in a recall test more often than those who had read the sentences in the standard order. The explanation that Kelly et al. offered for this finding is that the retrieval of information from memory is sensitive to the greater availability of the more prototypical items. Since the more prototypical members of a set are more readily available for retrieval, it is natural that they will tend to be placed first in the linear structure of a sentence. In another study (Fenk-Oczlon 1989), word frequency was found to affect word order. In this paper we consider the concept of saliency as including all these constituent factors.

2.2â•… Animacy The animacy hierarchy is an organization of knowledge about ontological entities, such as human beings, other animals, plants, and inanimate objects (e.g. Keil 1979). Studies of animacy have focused on the hierarchy that includes these four levels, in the order listed, from highest to lowest. Their hypothesis is that the location of a concept in the hierarchy is correlated with its conceptual salience: Concepts that are higher in the hierarchy are more basic to our conceptual system and are learned earlier in human conceptual development (Keil 1979).2

.╅ In linguistics this hierarchy is one of the versions of the animacy hierarchy, which plays a major role in various linguistic phenomena, such as pronominalization patterns, the �ascription of various syntactic roles, and effects on word order. See, e.g. Deane (1992, 1993), Chapter 5, which summarizes some of these phenomena, as well as Bock and Warren (1985), and many other studies in this area, which indicate the central role of the hierarchy in language �production.



The farmers sowed seeds and hopes 

Various studies (McDonald et al., 1993; Cooper & Ross 1975) have pointed to the concept of ‘animate leaders’: people’s preference for placing words for concepts with higher animacy first in sentences.3 See also Byrne and Davidson (1985).

2.3â•… Abstractness Various studies have revealed the sensitivity of the order of the nouns in a sentence to their degree of abstractness. For example, Mayerthaler (1988) and Benor and Levi (2006), who discussed the semantic factors that affect the order of nouns in Â�binomials, claim that there is a general principle (similar to Cooper and Ross’s ‘me first’ principle) that unmarked elements – those that represent the neutral value of a particular dimension – tend to appear before marked ones, because the first location is less marked than the second one. In the semantic realm the unmarked elements are those that are closest to the speaker or those that she can grasp. According to Mayerthaler’s analysis, which is also accepted by Benor and Levi, words for concrete objects are less likely to be marked than words for abstract objects because they are more easily grasped, and therefore they tend to appear first in phrases (see Benor & Levi 2006). Only a few studies empirically investigated the connection between the abstractness hierarchy and the order of the nouns in a sentence in Â�comprehension and Â�production, and here too the studies focused on literal sentences (e.g. Bock & Â�Warren 1985). Our goal in the present study was to discover whether the Â�abstractness Â�hierarchy affects the order of the nouns in metaphorical sentences as well. This issue is especially interesting because it is possible that this sort of effect depends on the type of cognitive processing involved. For example, at least one of the few Â�studies that explored the issue experimentally by analyzing the Â�understanding and Â�production of language (Bock & Warren 1985) found that the abstractness hierarchy does not Â�necessarily affect the order of nouns in memory tasks. In this study they presented the participants with sentences containing nouns with Â�different degrees of Â�abstractness and in different orders, but with the same syntactic role. These Â�sentences contained a pair of nouns, as in ‘The lost hiker fought time and Â�winter’. It turned out that the degree of abstractness of the nouns did not affect the order in which they were remembered. These findings indicate that the connection between abstractness and noun order is unclear, and that this effect may depend on the type of task and the type

.â•… This is a special case of the ‘me first’ principle formulated by Cooper and Ross, which claims that people prefer to place words for things that are close to their self-image earlier in phrases. For example, students from Illinois call the game they are watching the Â�‘Illinois-Michigan game’, whereas students from Michigan call the same event the ‘Michigan-Illinois game’. Â�According to Cooper and Ross, word order reflects the salience of objects in people’s worldview.

 Yeshayahu Shen & Elad Kotzer

of �cognitive processing involved. Thus, in addition to the investigation of this �connection for �metaphorical sentences, the present study should also contribute to our �understanding of the connection in general.

3.â•… Salience and word order in the zeugma: Pilot studies 3.1â•… Distribution in poetic language Our first question involves the frequency of canonical and non-canonical word order in poetry and prose corpora, and asks which one is more frequent. Several studies of poetic corpora in recent years suggest that the canonical order is more frequent. In one study (Shen 1998, 2008) a sample of 350 zeugmas taken from Hebrew poems written between 1900 and 1980 was assembled and analyzed for its word ordering. Several different periods in modern Hebrew poetry were represented in the poetry corpus, which were otherwise markedly different from one another (e.g. in their historical contexts, themes, and style). The hypothesis was that contextual factors, such as the particular poem, poet, or school of poetry in question, would not affect the structural pattern emerging from the study’s results. In addition, we analyzed a large number of zeugmas from the work of an Â�American poet who is known for his use of this figure – Allen Ginsburg. The main finding revealed by this analysis is unequivocal. The vast majority (83%) of the zeugmas that were Â�analyzed were in the canonical order in which the more prototypical element appeared first, while only 17% had the opposite order. This pattern occurred in all Â�different Â�corpora, with very little variance among them. To expand the number of corpora for analysis, an additional study analyzed poetic corpora from nineteenth- and twentieth-century Russian poetry (a field study performed with the help of Lisa Chodnovsky; see Shen 1998, 2008a). The corpus of Russian poetry included a number of major poets representing five different periods of Russian poetry (the Pushkin period, the second half of the nineteenth century, the early twentieth century, the period from the 1920s to the 1950s, and the 1960s). The poets from these different periods vary greatly in their poetic characteristics. For example, they variously display the influence of Romantic poetry (e.g. Lermontov), of neo-classical poetry (Batyushkov and Bertinsky, as well as Pushkin), symbolist poetry (e.g. Ansky, who is considered one of the pioneers of symbolism in Russian poetry, and Block), axiomistic poetry (Gomelb), and social poetry (Maikovsky). The total number of zeugmas collected from these poets was 287 (ranging from 2 to 22 for each poet). These zeugmas were analyzed in the same way as the samples from Hebrew and English poetry, so as to see which of the possible word orders was more frequent.



The farmers sowed seeds and hopes 

The results of this study clearly match the results of the other two studies. The analysis showed that there were more phrases in the canonical than in the Â�non-canonical order (172 vs. 115, respectively). Moreover, in each one of the periods the number of phrases that appeared in the canonical order was greater than those in the non-canonical order. A poet-by-poet analysis showed that 14 of the 17 poets Â�followed this pattern, while only 3 did not. These findings, taken together, show that, with a very small number of exceptions, the most frequent word order in all the corpora that were analyzed is the one in which the more salient element appears before the less salient one – regardless of the language, geographical area, historical period or poetic characteristics of the poems.

3.2â•… Salience: Recall and interpretation tasks Previous studies investigated some of the cognitive implications of the association between salience and word order. These studies revealed some clear findings that are consonant with the greater frequency of the canonical word order in the various poetic corpora. In general, these initial studies show that the canonical word order in zeugmas is more basic and cognitively simpler. There are three main findings that support this generalization. Preference. A previous study (Shen 1998; see also Shen 2008, 2008a) found that when participants are presented with the two possible word orders and asked to judge which one sounds more ‘natural’, they say that the canonical word order is more natural. Recall. Another study (Shen 1998; see also Shen 2008a) tested the hypothesis that, when participants are asked to recall non-canonical zeugmas, they have a greater Â�tendency to invert the order of the elements (so that they fit the canonical order) than they do to invert the order of the elements in canonical zeugmas. The rationale for this hypothesis was that the more salient noun is more accessible to memory (see Kelly et al. 1986), and therefore tends to appear in the first position in the sentence, before the noun with lower salience, and therefore lower accessibility. The linear structure of the sentences produced by the participants were found to be in accordance with the Â�accessibility of the nouns in question in most cases, thus confirming the hypothesis. In the experiment the participants were asked to read a set of zeugmas that appeared either in their canonical or non canonical version. Consider, for example: ‘The baby imbibed milk and kisses/kisses and milk’. Here, the noun ‘milk’ is more closely Â�associated with the verb ‘imbibe’ than the noun ‘kisses’ is. Thus, the structure which places ‘milk’ before ‘kisses’ is the canonical one, and the reverse structure is the Â�non-canonical one. Half of the participants received the canonical version (e.g. ‘The baby imbibed milk and kisses’), while the other half received the non-canonical Â�version (e.g. ‘The baby imbibed kisses and milk’). Half of the zeugmas read by each Â�participant were canonical ones,

 Yeshayahu Shen & Elad Kotzer

while the other half were non-canonical (for details see Shen 1998). After the reading phase the participants were asked to recall the zeugmas they read. In accordance with the hypothesis, they inverted the non-canonical sentence Â�significantly more often than the canonical sentences. Interpreting zeugmas. Another study (Shen and Shalev, submitted) investigated the way people interpret or paraphrase zeugmas. The purpose of the study was to find out whether the non-canonical sentences would be inverted in a task which did not involve the limitations of memory. Twenty-six native Hebrew speakers participated in the experiment. They were divided into two groups, where the first group was presented with five zeugmas in the canonical order and five in the non-canonical order, while the second group was presented with the same zeugmas in the opposite order. For example, if the first group read the canonical sentence ‘The soldier going into battle packed his shirt and his sorrow’, then the second group read the same sentence ending ‘packed his sorrow and his shirt’. Each of the participants was asked to write a one-line interpretation of each sentence presented to them. The sentences written by the participants were analyzed to find out if the order of the elements in the sentence was the same as the order in the zeugma that had been presented, or in the opposite order. For example, one of the participants who was presented with the non-canonical zeugma ‘The soldier packed his sorrow and his shirt’ offered the interpretation, ‘While packing his shirts he thought about his sorrow’. This is an example of inversion, as the order of the elements is opposite to that in the zeugma presented to the participant. Another example of inversion is the interpretation offered by another participant, ‘The soldier got ready to leave and tried to get over his sadness’. Our assumption was that an inverted interpretation (which does not preserve the word order of the original), indicates that the original sentence is less natural than an original sentence that generated a ‘preserving’ interpretation (an interpretation that preserve the word order of the original). When we analyzed the interpretations we compared the participants’ responses to each canonical zeugma and its inverted counterpart. We found that the number of cases in which the participants inverted the order of the elements was significantly greater for the non-canonical than for the canonical zeugmas (for an extended discussion, see Shen 2008, 2008a).

4.â•… Salience, abstractness and animacy The studies reviewed so far were pilot studies and thus had a number of limitations, such as having a fairly small number of experimental items (about 10) and participants (20–26). Other, more significant limitations are the following:



i.

The farmers sowed seeds and hopes 

These studies did not distinguish sufficiently between salience and abstractness. For example, in the case of the zeugma ‘The soldier packed his shirt and his sorrow’, it is not clear whether people’s preference for this order over the opposite one should be attributed to the fact that ‘shirt’ is more salient than ‘sorrow’ or the fact that it is more concrete.

ii. They did not systematically investigate the factors of animacy and abstractness (see the previous note), which are also important in determining word order. iii. Obviously, the relative weight of the three factors could not have been investigated. A more comprehensive set of studies (for an extended discussion see Shen and Kotzer, submitted) was therefore undertaken to address these limitations. These studies used the same three experimental tasks as the pilot studies – preference, interpretation, and recall. Our hypothesis was that each of the three semantic factors at issue affects word order in the various aspects of understanding zeugmas. The research included six experiments. The first three (1–3) studied salience and abstractness separately using all three tasks, and then compared the effects of the two factors. The other three experiments (4–6) used the same three tasks to study and compare the factors of salience and animacy. One set of experimental items was used for the first three experiments and another set for the last three. The principal measure we used in the recall and interpretation experiments was the ‘inversion measure’ discussed above. We checked in which sentences the participants inverted the word order more frequently – for example, read ‘You held your breath and the door for me’ but wrote down ‘You held the door and your breath for me’. Our assumption was that people have a tendency to ‘revise’ sentences so as to put them in the canonical order, which is cognitively stronger. This measure was used in other studies (Glucksberg et al. 1997; Chiappe et al. 2003) to test hypotheses about canonical word order, mainly in similes.

4.1â•… E  xperiments 1–3: The effect of salience and abstractness on word order: Recall, interpretation and preference Sixteen zeugmas were composed for the purposes of these three experiments, with the two elements in each sentence having a different level of salience and a different level of abstractness. In eight of the sixteen zeugmas the more salient element was also the more abstract one (e.g. in the sentence ‘David tried to control his anger and his gun’, ‘anger’ is both more salient and more abstract), while in the other eight the more salient element was less abstract (e.g. in the sentence ‘The farmers sowed seed and hope’, ‘seeds’ is more salient, while ‘hope’ is more abstract).

 Yeshayahu Shen & Elad Kotzer

The first eight sentences were used to create the following two types of sentence: Eight sentences in which the first element is both more salient and more abstract. This order is called ‘canonical salience – non-canonical abstractness’. For example: Danielle’s appearance stopped my breath and the traffic. Danielle’s appearance stopped the traffic and my breath. Eight sentences in which the second element is both more salient and more abstract. This order is called ‘non-canonical salience – canonical abstractness’. For example: When he awoke from his sleep he remembered that he had lost his keys and his virginity. When he awoke from his sleep he remembered that he had lost his virginity and his keys. The last eight sentences were used to create the following two types of sentence: Eight sentences in which the first element is more salient and the second element is more abstract. This order is called ‘canonical salience – canonical abstractness’. For example: In her free time Elaine wove a plot and a tablecloth. In her free time Elaine wove a tablecloth and a plot. Eight sentences in which the second element is more salient and the first element is more abstract. This order is called ‘non-canonical salience – non-canonical abstractness’: In the spring the farmer sowed seeds and hope. In the spring the farmer sowed hope and seeds. The thirty sentences that were constructed this way were used to construct four Â�different versions of the questionnaire, each with sixteen question-answer pairs, where each answer was a zeugma. For example: What did the farmer sow in the spring? In the spring the farmer sowed seeds and hope.

4.1.1â•… Experiment 1: Recall This experiment tested the effects of salience and abstractness, both separately and together, on people’s ability to recall the order of the words in zeugmas. We hypothesized that sentences would be preferentially recalled by participants if they were in the canonical order for either salience or abstractness. We further hypothesized that sentences in the canonical order for both factors would be recalled best, while those in the non-canonical order for both factors would be recalled least well. The question of whether sentences that were canonical for salience would be recalled better or less well than sentences canonical for abstractness was left open, without a hypothesis. The experiment had two stages. First the participants read the stimulus pairs, each of which included a question and an answer, as in the above example. Then, after a short



The farmers sowed seeds and hopes 

break, they were presented with the original questions (e.g. ‘what did the farmers sow in the spring?’) and asked to reconstruct the original answers from memory. In general, the results accorded with the hypotheses. There was a significant effect of salience in the measure of inverted element; that is, there were more inversions when the order of salience was non-canonical. There was a similar effect of abstractness, but here the effect was not quite significant. There was also a significant interaction, derived from the fact that, among the items where the salience order was canonical, there was no difference between the items with canonical or non-canonical abstractness order, while among the items where the salience order was non-canonical, there was a significant difference between the items with the two different orders of abstractness. The simple scores for remembering the two elements in the sentences did not show any effect of either salience or abstractness (for details, see Kotzer 2007).

4.1.2â•… Experiment 2: Interpretation This experiment was similar to Experiment 1, except that it used an interpretive task to test how the factors of salience and abstractness, separately and together, affect word order in zeugmas. The hypotheses were the same as those in Experiment 1. The participants in this experiment were asked to write one sentence of interpretation for each zeugma. The results were similar to those obtained in Experiment 1. The inversion measure revealed a significant effect of salience – there were more inversions when the salience order was non-canonical. There was also a significant effect of abstractness – there were more inversions when the abstractness order was non-canonical. A significant interaction was found as well – when the salience order was Â�non-canonical, the difference between the canonical and non-canonical salience orders was much greater than when the salience order was canonical. 4.1.3â•… Experiment 3: Preference The aim of this experiment was similar to that of the previous two, but it used the task of judging which sentence sounds more natural. We compared sentences that are canonical with those that are non-canonical for both factors of salience and abstractness, as well as sentences that are canonical for salience with those that are canonical for abstractness. We hypothesized that sentences that are canonical for both factors would be considered more natural than those that are non-canonical for both. The question of whether sentences canonical for only one factor would be considered more natural than those canonical only for the other factor was left open. The participants were presented with two versions of zeugma and asked to mark the one they considered more natural. The analysis of the responses revealed a significant effect of canonicity – that is, participants prefer sentences that are canonical for both factors over those that are non-canonical for both. In contrast, no preference was found for sentences that are canonical for either one factor or the other.

 Yeshayahu Shen & Elad Kotzer

4.2â•… E  xperiments 4–6: The effects of salience and animacy on word order in three tasks: Recall, interpretation and preference The task and procedure in these three experiments were exactly the same as in the first three. The only difference between the two groups was that the last three experiments investigated salience and animacy instead of salience and abstractness. The experimental items in Experiments 4–6 were based on sixteen zeugmas in which the two elements had different levels of salience and a different position in the animacy scale. In eight of the sentences the more salient element was also higher on the animacy scale (e.g. ‘The mouse and the cheese ran into the hole’, where ‘mouse’ is both more salient and higher on the animacy scale), while in the remaining eight sentences the more salient element was lower on the animacy scale (e.g. ‘The dog and the man barked at the mail carrier’, where ‘dog’ is more salient but lower on the animacy scale). The first eight sentences were used to create the following types: Eight sentences in which the first element was both more salient and higher on the Â�animacy scale. This order is called ‘canonical salience – canonical animacy’. For example: The robbers and the sacks escaped around the corner. The tiger and the land were awakened by the rain. The cat and the houseplant shrieked with thirst. Eight sentences in which the second element was both more salient and higher on the animacy scale. This order is called ‘non-canonical salience – non-canonical animacy’. For example: The sacks and the robbers escaped around the corner. The land and the tiger were awakened by the rain. The houseplant and the cat shrieked with thirst. The remaining eight sentences were used to create the following types: Eight sentences in which the first element was more salient but lower on the Â�animacy scale. This order is called ‘canonical salience – non-canonical animacy’: The cypress and the bird were uprooted from their homes. The hen and my aunt squawked incessantly. The sapling and the rabbit put down roots at our house. Eight sentences in which the second element was both more salient but lower on the animacy scale. This order is called ‘non-canonical salience – canonical animacy’. The bird and the cypress were uprooted from their homes. My aunt and the hen squawked incessantly. The rabbit and the sapling put down roots at our house.



The farmers sowed seeds and hopes 

The thirty-two sentences constructed in this way were used to formulate four �questionnaires, each with sixteen question-answer pairs, where the answer was always a zeugma. An example of one of these pairs is: What happened after I left? After I left, my mother and the houseplant withered. Since the task and the procedure were the same as those in the previous �experiments, we will present only the main findings of the last three experiments.

4.2.1â•… Experiment 4: Recall There was a significant inversion effect involving salience – that is, the participants tended to invert the elements of the zeugma in sentences that were non-canonical for salience. The same occurred for animacy. There was no interaction between these two effects. There was no effect of either salience or animacy for the memory scores, nor was there any interaction. 4.2.2â•… Experiment 5: Interpretation The results were similar to those of Experiment 4. There was a significant inversion effect involving both salience and animacy. There were four times as many inversions for canonical than for non-canonical zeugmas, with the two variables contributing about equally. There was no interaction between the two effects. 4.2.3â•… Experiment 6: Preference The participants rated the sentences that were canonical for both salience and animacy as significantly more natural than ones that were non-canonical for both. They also rated the sentences that were canonical for salience but not for animacy as more natural than the ones that were canonical for animacy but not for salience. Taken together, the two types of sentence that were canonical for salience were rated more natural than the two types of sentence that were canonical for animacy.

5.â•… Conclusion Three conceptual factors – salience, abstractness and animacy – have been shown to affect word order in literal expressions. The purpose of the research presented here was to investigate whether they also affect word order in metaphorical expressions. Our general hypothesis, for each of the three factors separately, was that participants would prefer the canonical sentences over the non-canonical ones (Hypothesis 1). We also hypothesized that when the factors were tested together, the participants would most prefer the sentences that were canonical for two factors, and least prefer the sentences

 Yeshayahu Shen & Elad Kotzer

that were non-canonical for two factors (Hypothesis 2).4 We left open the question of whether participants would prefer sentences that were canonical for only one factor over sentences canonical for the other factor. Hypothesis 1 thus predicted that participants would prefer zeugmas in which the word order for each of the factors was canonical. This hypothesis was �generally �supported in all the experiments (although the effect of abstractness was only �marginally significant in Experiment 1). In the recall and interpretation experiments the participants inverted the word order of the non-canonical sentences to accord with canonical word order, while in the judgment experiments they found the canonical sentences to be more natural than the non-canonical ones. This means that each of the three tested factors affects word order in zeugmas across all three tasks. Hypothesis 2 predicted a hierarchy of preferences among the types of sentence that pit one factor against another. Both of the predictions were supported. We also tested the question that was left open, as to whether there would be a hierarchy among the various factors, but we did not find any clear hierarchy. This finding can be explained in several possible ways. First, there may not be any such hierarchy, or there may be a hierarchy with such small differences between the levels that they are not significant. Second, the figure of speech used in the present research may not be suitable for revealing such a hierarchy. In sum, the present research was an initial attempt to study the effect of certain semantic factors on the structure of metaphorical sentences. More research is �necessary in this area, both to try to support canonical principles for other factors in figurative speech, and to investigate new ways for answering the question of whether there is a hierarchy among the various factors. This can be done by testing new factors that can effect word order but which were not investigated in the present research, making use of different experimental methodologies, and testing other figures of speech. These findings show that the major conceptual factors that affect word order in �literal sentences also play an important role in the understanding and production of metaphorical language. The findings accord with the general view that has been developed in other contexts (Giora 2007; Glucksberg & Keysar 1990), which claims the processes underlying understanding and production are shared by literal and �metaphorical language, and refute the traditional view (Grice 1975; Searle 1979) that different processes underlie these two forms of language. However, a large proportion

.â•… It might seem as if Hypothesis 2 could be derived from Hypothesis 1 – that is, if each factor separately would support the canonical order, then surely this would be the case for two factors taken together, but in practice the factors might cancel each other’s effect, or one factor might turn out to be much more dominant than the other. In the latter case there would be no difference between the participants’ preference for sentences that are canonical for both factors and those that are canonical for the dominant factor alone.



The farmers sowed seeds and hopes 

of the evidence for the claim that the processes are shared by the two forms of language  are based on stimuli consisting of conventional metaphors (e.g. Â�Keysar 1989; Gibbs 1994) or on an extended context that makes it easier to understand Â�metaphorical expressions of this kind (Ortony et al. 1978). In contrast, the research presented here shows clearly that even when the metaphorical stimulus items are novel rather than conventional, and even when the participants are not provided with any sort of Â�context, the principles that explain the understanding of Â�metaphorical expressions are common to them and to literal expressions, at least as far as word order is concerned.

References Benor, S. & Levy, R. 2006. The chicken or the egg? A probabilistic analysis of English binomials. Language 82: 233–278. Bock, K. & Warren, R. 1985. Conceptual accessibility and syntactic structure in sentence Â�formulation. Cognition 21: 47–67. Byrne, B. & Davidson, E. 1985. On putting the horse before the cart: exploring conceptual bases of word order via acquisition of a miniature artificial language. Journal of Memory and Language 24: 377–389. Chiappe, D. Kennedy J. & Smykowski, T. 2003. Reversibility, aptness, and the conventionality of metaphors and similes. Metaphor and Symbol 18: 85–105. Cinque, G. 1999. Adverbs and the Universal Hierarchy of Functional Projections. Oxford: OUP. Cooper, W.E. & Ross, J.R. 1975. Word order. In Papers from the Parasession on Functionalism, R.G. Grossman, L. San & T. Vance (eds.) 63–111. Chicago IL: Chicago Linguistic Society. Deane, P. 1992. Grammar in Mind and Brain: Explorations in Cognitive Syntax. Berlin: Mouton de Gruyter. Deane, P. 1993. On metaphoric inversion. Metaphor and Symbolic Activity 8: 111–126. Fenk-Oczlon, G. 1989.Word frequency and word order in freezes. Linguistics 27: 517–556. Gibbs, R.W. Jr. 1994. The Poetics of Mind: Figurative Thought, Language, and Understanding. Cambridge: CUP. Giora, R. 2002. Literal vs. figurative language: Different or equal. Journal of Pragmatics 34: 487–50. Giora, R. 2003. On Our Mind: Salience, Context, and Figurative Language. Oxford: OUP. Giora, R. 2007. Is metaphor unique? Neural correlates of non-literal language. Brain and Â�Language 100: 111–114. Glucksberg, S. & Keysar B. 1990. Understanding metaphorical comparisons: Beyond similarity. Psychological Review 97: 3–18. Glucksberg, S., McGlone, M. & Manfredi, D. 1997. Property attribution in metaphor Â�comprehension. Journal of Memory and Language 36: 50–67. Grice, H.P. 1975. Logic and conversation. In Syntax and semantics: Vol. 3. Speech acts, P. Cole & J.L. Morgan (eds), 41–58. New York NY: Academic Press. Hawkins, J.A. 1983. Word Order Universals. New York NY: Academic Press. Keil, F.C. 1979. Semantic and Conceptual Development: An Ontological Perspective. Cambridge MA: Harvard University Press.

 Yeshayahu Shen & Elad Kotzer Kelly, M.H., Bock, J.K. & Keil, F.C. 1986. Prototypicality in a linguistic context: Effects on Â�sentence structure. Journal of Memory and Language 25: 59–74. Keysar, B. 1989. On the functional equivalence of literal and metaphorical interpretations in discourse. Journal of Memory and Language 28: 375–385. Kotzer, E. 2007. Conceptual Factors Influencing Word Ordering in Metaphorical Expressions. MA thesis, Tel Aviv University. Mayerthaler, W. 1988. Morphological Naturalness. Ann Arbor MI: Karoma. McDonald, J.L., Bock K. & Kelly M.H. 1993. Word and world order: Semantic, phonological, and metrical determinants of serial position. Cognitive psychology 25: 188–230. Ortony, A., Schallert, D., Reynolds, R. & Antos, S. 1978. Interpreting metaphors and idioms: Some effects of context on comprehension. Journal of Verbal Learning and Verbal Behavior 17: 465–477. Osgood, C.E. & Bock, J.K. 1977. Salience and sentencing: Some production principles. In Â�Sentence Production: Developments in Research and Theory, S. Rosenberg (ed.), 89–140. Hillsdale NJ: Lawrence Erlbaum Associates. Rosch, E. 1999. Principles of categorization. In Concepts: Core Readings, E. Margolis & S. Laurence (eds), 189–206. Cambridge MA: The MIT Press. Searle, J. 1979. Expression and Meaning. Cambridge: CUP. Shen, Y. 1998. Zeugma: prototypes, categories, and metaphors. Metaphor and Symbol 13: 31–47. Shen, Y. 2008. How come the shirt is being packed before the sorrow? On one central aspect of the relation between cognition and poetic structure. Language and Literacy 1: 159–175. Shen, Y. 2008a. Cognition and poetic figures. In: Cambridge Handbook of Metaphor and Thought, R. Gibbs (ed.), 295–307. Cambridge: CUP. Shen Y. & Kotzer, E. Submitted. You held your breath and the door for me: Word order in Â�figurative expressions. Shen Y. & Shalev, H. Submitted. Why ‘A magnet is like a smile’ is turned into ‘A smile is like a magnet’: Inversions in metaphor comprehension. Wright, S., Hay, J. & Bent, T. 2005. Ladies first? Phonology, frequency, and the naming Â�conspiracy. Linguistics 44: 531–561.

Non-iconic chronology in English narrative texts Vyacheslav Yevseyev

Eurasian Institute for Humanities, Astana/Kazakhstan This article presents the results of a corpus investigation into the phenomenon of non-iconic chronology, which is understood here as the reversal of natural order of events at the textual micro-level, e.g. ‘I conquered after I came and saw’ as opposed to the well-known example of iconic order ‘I came, I saw, I conquered’. Although the general implications of chronological, or temporal, iconicity were widely discussed as early as the 1980s, little is known so far about how frequently non-iconic chronology occurs in real (rather than constructed) narrative texts, what sorts of text tend to be more temporally iconic or less temporally iconic, and what syntactic structures (coordinate or subordinate with particular temporal conjunctions) are most typically used to depict sequential events in a reversed textual order. An answer to these questions is attempted here on the basis of some 150 English literary texts.

1.â•… The problem Since Jakobson (1971 [1965]: 350) quoted Caesar’s words Veni, vidi, vici to illustrate the iconic principle of sequence, it has almost become a commonplace in the literature to make the observation that in narrative texts events are presented primarily in the order in which they occur, or are assumed to occur, in the real world. The Â�phenomenon of temporal iconicity has appeared in different scientific paradigms under quite Â�different labels, e.g. “order of mention principle” (Clark & Clark 1977:â•›358) or “natural clause ordering” (Opačić & Osgood 1984:â•›295) in psycholinguistics, ordo naturalis in Â�discourse analysis (e.g. Brown & Yule 1983:â•›125), “principle of temporal sequence” (Tai 1985:â•›50) or “chronological sequencing” (Haiman & Thompson 1988: passim) in the study of Â�syntactic iconicity, “iconic sequence” in narratology (Fleischman 1990:â•›131), “principle of chronological order” in tense-aspect theory (Klein 1994:â•›45), as well as “principle of unmarked temporal interpretation” (Declerck 2006:â•›427), “general iconic principle of linear successivity” (Johanson 1995:â•›327) etc.1 At the same time, most authorsÂ� .â•… Even if sometimes doubts are voiced as to whether iconicity can explain different Â�phenomena of language change and the typological distribution of linguistic categories, its

 Vyacheslav Yevseyev

underline that this principle is not absolute and deviations from the strict chronological order are possible if special markers are involved: temporal Â�conjunctions, e.g. after and before, pluperfect forms, e.g. had done, or temporal adverbs like already and first; these deviant cases are referred to as ordo artificialis (being the opposite of ordo naturalis, e.g. Dressler 1989:â•›15), hysteron proteron (e.g. Klein 1994:â•›45), “countersequentiality” (Givón 2001:â•›295), “reverse-order reports” (Bardovi-Harlig 1994: passim) or anachronie (Genette 1972:â•›79). However, we still remain in the dark about how such special cases, which will be called in the present article ‘instances of non-iconic chronology’ (or simply: ‘non-iconic cases’), are distributed in literary texts: what sorts of texts they are typical of, what literary epochs are ‘more iconic’ or ‘less iconic’ in this respect, and whether certain authors deliberately avoid non-iconic chronology or, on the contrary, employ it quite frequently. What is more, there do not seem to exist many detailed descriptions of the interaction of conjunctions, tense forms and temporal adverbs within such non-iconic cases in real texts. In some previous studies (Yevseyev 2003, 2008, 2009) attempts were undertaken to systematically describe different aspects of iconic and non-iconic chronology in German and Russian narrative texts using corpus data, but it is the present study that aims at developing this approach a step further primarily on English material.

2.â•… Corpus and method To describe non-iconic chronology in English in quantitative terms, a special corpus of narrative texts was compiled and manually indexed. A search of an electronic Â�corpus, e.g. of the British National Corpus (BNC),2 would be insufficient for our purpose since automatic queries are confined to certain lexical items (temporal conjunctions or Â�temporal adverbs first, then etc.). The problem is that roughly one fourth of structures expressing non-iconic chronology does not contain any clear formal markers (such structures are based on the pattern ‘[Past Indefinite] → [Past Perfect]’, see below), and therefore they cannot be retrieved effectively via electronic corpus search. For this reason a number of most significant or illustrative works of English prose (British and American) – preferably of the nineteenth and twentieth century – were selected for

subtype ‘iconicity of sequence’, to which also ‘iconicity of temporal sequence’ belongs, is left untouched (Haspelmath 2008:â•›3). .â•… Though corpus linguistics has made dramatic progress in the last decade (cf. Lüdeling & Kytö 2008, 2009) and electronic corpora are a good instrument for finding rare linguistic phenomena which can be presented as a chain of alphabetical symbols, there seem to remain areas where manual analysis is indispensable, e.g. the study of chronological iconicity.



Non-iconic chronology in English narrative texts 

investigation (153 texts in total). As it would have been impossible to carefully analyse these texts in full, representative text samples of about 15,000 to 40,000 characters were compiled following a unified method (see Table 1). Table 1.╇ The procedure for compiling text samples Length of text (in pages)

Pages selected for analysis

up to 25 25 to 50 51 to 100 101 to 250 251 to 500 501 to 1000 over 1000

all pages 4–5, 10–11, 14–15, 20–21, 24–25, 30–31, 34–35 etc. 10–11, 20–21, 30–31, 40–41, 50–51, 60–61, 70–71 etc. 20–21, 40–41, 60–61, 80–81, 100–101, 120–121 etc. 50–51, 100–101, 150–151, 200–201, 250–251 etc. 100–101, 200–201, 300–301, 400–401, 500–501 etc. 200–201, 400–401, 600–601, 800–801, 1000–1001 etc.

Obviously, text pages whose numbers end in certain symbols (e.g. 0 or 1) are not likely to be reserved by authors for special purposes such as an excessive use of Â�non-iconic chronology or, on the contrary, its deliberate avoidance; therefore the resulting text samples can be regarded as having been taken randomly, provided that the same method has been rigorously applied to extracting samples from all texts. The total length of all text samples in the English corpus is over four million characters (roughly 2,000 ordinary book pages). Text samples were searched for pairings of non-simultaneous events, fulfilling the following conditions: – they should belong to the past context (including historical present) since only past situations are quite definite in relation to such factors as factuality vs. Â�counter-factuality, semelfactivity vs. iterativity and punctuality vs. durativity; thus, only pairs of factual situations, having one phase, were selected for further consideration since they can be regarded as typical links in a narrative chain (cf. Myhill 2001:â•›168); punctual situations (i.e. all ‘achievements’ and many ‘accomplishments’, cf. Vendler 1967:â•›106) are prototypical links in a narrative chain, e.g. ‘He took a book from his pocket when he entered the room’, but also durative situations (i.e. ‘states’, ‘activities’ and some ‘accomplishments’ – ibid.) may be involved in Â�establishing non-simultaneous temporal relations when conjunctions of Â�anteriority or posteriority are used: ‘He sat in the armchair after he had a cup of tea’; – they should be expressed in neighbouring clauses since a close conceptual Â�relationship can be discerned only between items that are closely related Â�textually (reversing the ‘First Behaghel’s Law’ – Behaghel 1932:â•›4, alias the ‘Proximity Â�Principle’ – Givón 1985:â•›202); both clauses may be placed within a

 Vyacheslav Yevseyev

–

–

sentence, e.g. ‘He took a seat when he entered the room’ or ‘He took a seat having entered the room’ or within a smaller text fragment exceeding the boundaries of a sentence, e.g. ‘He took a seat. He had entered (stealthily)’; among complex sentences, only those with temporal conjunctions were considered since temporal clauses can both precede and follow the main clause, thus making possible either the iconic or the non-iconic variant with the same conjunction, whereas other types of subordinate clauses may have a more or less fixed position; consequently, including clause types other than temporal would seriously distort the picture;3 a simultaneous reading of the temporal relation should be excluded; in order to distinguish non-simultaneous pairs of events from simultaneous ones, the theory of aspect and Aktionsart has to be consulted (e.g. Comrie 1976; Sasse 1991): in most cases both situations constituting a temporal relation of successivity must be telic, i.e. having an intrinsic limit for their dynamics and exhausting themselves after reaching this limit; in many cases, however, only general world knowledge or the notion of ‘scripts’ (usual order of events during a meal in a restaurant, during an air flight etc. – Abelson 1981) helps to interpret the order of events correctly.

The results from the English corpus will be compared with those obtained from a Â�German and a Russian corpus, which were partially documented in Yevseyev (2003, 2005, 2008, 2009).

3.╅ General features of non-iconic chronology in English narrative texts 3.1╅ Main findings 1. In the course of analysis, 3,765 cases of conjoined pairs of non-simultaneous events were identified, among them 3,509 iconic (93.2 percent) and 256 noniconic (6.8 percent). These figures can be regarded as solid statistical evidence that the principle of temporal iconicity is dominant in narrative texts and that non-iconic chronology is the exception. In Yevseyev (2003: Ch. 7) very similar results were documented for German and Russian narrative texts; updated figures .╅ Existing studies demonstrate that conditional clauses are almost exceptionally placed in accordance with the iconicity principle, final and consecutive clauses show a clear tendency to follow the iconicity principle (even if less consistently than conditional clauses), whereas causal clauses usually produce non-iconic chronology as they tend to follow their respective main clauses (e.g. Dik 1997:╛399; Degand 2000:╛175).



2.

3.

4.

5.

Non-iconic chronology in English narrative texts 

are as follows: 4.7 percent of non-iconic cases for Russian and 6.2 percent for German. These comparatively minor differences may have been caused not only by diverging narrative styles typical for English, German and Russian, but also by the peculiarities in the use of available grammatical means which are characteristic of the three languages (cf. Section 5.2). This issue will be examined in more detail in forthcoming publications. About one fourth of all texts analysed does not contain any non-iconic cases, i.e. they seem to follow the principle of temporal iconicity consistently, at least at the textual microlevel. Obviously, this result can be only relative for texts of which only samples were analysed; however, if no instances of non-iconic chronology were detected in a randomly selected text sample of about 10 to 20 pages in length, their probability in the rest of the text must be very low. In most texts the ratio of non-iconic cases to all cohesively expressed events (iconic and non-iconic together) lies between 2 and 20 percent (provided the text contains any non-iconic cases); this fact should be interpreted as an additional argument in favour of temporal iconicity as the ruling principle in narrative texts. Only very few texts analysed have a ratio of non-iconic cases of about 30 Â�percent and more, but this is caused by the small number of iconic cases in those texts. The results from the Russian and the German corpus are very similar: about one fourth of all texts does not contain non-iconic cases, the bulk of texts with Â�non-zero values is distributed evenly between circa 1.5 percent and 20 percent; in German several outliers with even higher values occur (very short texts), whereas in Russian no outliers have been identified. Realist and modernist texts seem to be similar with respect to the ratios of Â�non-iconic cases. This result challenges my previous findings, according to which realist texts are characterized by higher rates of non-iconic cases than Â�modernist texts. These differences were rather significant in German and even more Â�significant in Russian. The approximately equal ratios of non-iconic cases in realist and modernist English texts may be caused by the similarities of Â�narrative styles specific to both epochs in English, whereas in Russian and German the epoch-specific styles are different. Some authors tend to avoid non-iconic chronology, since in their texts zero or very low rates of non-iconic cases have been discovered, e.g. in J. Steinbeck, R.L. Stevenson, J.D. Salinger, Th. Dreiser, G. Eliot. On the other hand, some authors seem to make use of non-iconic chronology relatively often, though the number of non-iconic cases is also in the texts of these authors by far smaller than the number of iconic cases: A. Conan Doyle, J.F. Cooper, C. Dickens, J. London. The existence of different author styles in relation to chronological iconicity has been already demonstrated on German and Russian material (see Yevseyev 2008:â•›345–346): such authors as S. Lenz, Ö. Horváth, M. Bulgakov,

 Vyacheslav Yevseyev

A. Tchekhov almost always follow the principle of temporal iconicity, whereas such authors as J.W. Goethe, H. Böll, F. Dostoyevsky, L. Tolstoy tend to deviate from this principle. 6. Interestingly, texts belonging to the genre ‘travelogue’ are as a rule highly iconic from the standpoint of microstructural temporality: the corresponding text Â�samples do not contain any non-iconic cases, e.g. Gulliver’s Travels by J. Swift or Travels with Charley by J. Steinbeck. This result is supported by evidence from Â�German: Â�absolutely iconic with regard to micro-structural temporality are The Harz Â�Journey by H. Heine, Wanderungen durch die Mark Brandenburg by Th. Â�Fontane, The Voices of Marrakesh by E. Canetti. Although there is one Â�conflicting piece of evidence from the Russian corpus, we can suggest the tentative Â�explanation that the author of a travelogue is not willing to deviate from the natural order of events even on the micro-level of narration since (s)he is used to narrating the scenes of the journey in the order as (s)he experiences them.

3.2â•… Discussion Recently, De Cuypere (2008:â•›166) scrutinized Givón’s predictions for the Â�numerical distribution of iconic and non-iconic cases among sentences containing after. In Givón’s terms, cases like “After she shot him, he died” should be much more frequent than cases “He died after she shot him”. Using the search results obtained from the BNC, De Cuypere found out that among 50 after-sentences 36 cases (72 percent) were non-iconic (ibid.). This somewhat unexpected result may be tentatively explained by different functions of after in iconic and non-iconic cases. In iconic cases this Â�conjunction either excludes possible simultaneous readings, as in ‘He had a meal and read a newspaper’ or helps to highlight the later event by downgrading the syntactic status of the clause conveying the earlier event (cf. Tomasello 2003:â•›271), as in ‘After he finished proofreading his manuscript he decided to re-write it completely’. If the two events cannot be thought of as simultaneous and no syntactic focussing of one of the two clauses is necessary, after is redundant and can be readily omitted. On the other hand, after cannot be dropped so easily in a non-iconic case; even if our world knowledge would suggest the correct interpretation of the temporal relation, the resulting compound sentence is by far more difficult to process: ‘He took a seat after he came in’. In the light of these considerations, non-iconic cases with after should be more numerous than iconic ones, and De Cuypere’s results would have been more plausible if the sample had contained more cases (e.g. 500). However, my corpus results demonstrate preference for iconicity, which ranges from slight to strong in particular conjunctions (see Table 2); among after-sentences, only 41.3 percent are non-iconic, and this finding is more in line with Givón’s opinion, though it contradicts my ad hoc theoretical construct.



Non-iconic chronology in English narrative texts 

Table 2.╇ Iconic and non-iconic cases among structures with particular temporal �conjunctions Conjunction

when as before till after as soon as until

Iconic cases

Non-iconic cases

Total

Tokens

Rate

Tokens

Rate

Tokens

Rate

237 63 56 53 27 30 31

76.0% 51.6% 75.7% 98.1% 58.7% 83.3% 96.9%

75 59 18 1 19 6 1

24.0% 48.4% 24.3% 1.9% 41.3% 16.7% 3.1%

312 122 74 54 46 36 32

100% 100% 100% 100% 100% 100% 100%

All in all, the distribution of iconic and non-iconic sentences with after, before and other temporal conjunctions does not seem to demonstrate conclusively whether or not the principle of temporal iconicity is effective in general: selecting the appropriate syntactic form of the utterance (‘main → main’ vs. ‘sub → main’ vs. ‘main → sub’) during text production seems to be just as important as ordering the clauses in correspondence with or against the principle of temporal iconicity. If a temporal marker is not necessary from the pragmatic point of view, it will be omitted in accordance with the principle of economy. Calculations based on my corpus results obtained from English, as well as Russian and German show that syntactic structures without temporal conjunctions represent, roughly speaking, one fourth of non-iconic cases and three fourths of iconic cases (see Table 3). Table 3.╇ Iconic and non-iconic cases with or without a temporal conjunction With a temporal conjunction

Without a temporal conjunction

iconic

≈ 25 percent

≈ 75 percent

non-iconic

≈ 75 percent

≈ 25 percent

In non-conjunctional iconic cases the correct interpretation of the temporal relation is supported by the linear succession of clauses (cf. “maxim of succession by default” – Simone 1995:â•›162), and non-conjunctional non-iconic cases are based on the pattern ‘[Past Indefinite] → [Past Perfect]’. The latter is not very easily accessible via electronic corpora, for the query ‘had’ – it is the only formal marker all cases of the mentioned type have in common – would bring too many irrelevant results, e.g. ‘He had a house’, ‘She had to get up early’ etc. However, replacing the temporal Â�conjunction with a pluperfect form in a non-iconic case does not make the two events less Â�correctly ordered temporally and the corresponding syntactic structure less important as a

 Vyacheslav Yevseyev

piece of evidence for or against the principle of temporal iconicity. For that reason, results obtained from manual texts analyses seem to be more reliable for an exhaustive description of the principle of temporal iconicity. The effectiveness of the temporal iconicity principle is also backed up by the fact that the reversal of the sequential order is confined to pairs of events; obviously, a whole story cannot be told backwards (cf. Haiman 1985:â•›4). In narratology, such notions as “anticipation”, “prospection”, “flash-forward”, on the one hand, and “retrospection”, “flashback”, on the other (cf. Prince 1987) are used to refer to shifting one event or a whole passage forward or backward in the linear dimension of the narrative text; nevertheless, within the shifted passage the events are arranged iconically. In my Russian corpus there is at least one example for three conversely ordered events (C-B-A): Rovno v devjat’ časov Razumixin javilsja v numera Bakaleeva. Obe damy ždali ego davnym-davno s isteričeskim neterpeniem. Podnjalis’ oni časov s semi ili daže ran’še. (F. Dostoevskij, Prestuplenie i nakazanie, 201) ‘At exactly nine o’clock Razumikhin presented himself at Bakaleyev’s Tenements. Both ladies had been long awaiting him in hysterical impatience. They had risen at seven, or even earlier’. (F. Dostoyevsky, Crime and Punishment, 255)

World knowledge helps restore the real succession of events: the ladies got up, then started to wait for their visitor, his coming was the final point of their waiting. It is not quite clear at the moment whether sequences with four and more events in reverse order, i.e. (E)-D-C-B-A, can be detected in real narrative texts.

4.â•… Incidence of non-iconic chronology 4.1â•… Method One of the essential characteristics of the subject under consideration is the Â�parameter called here ‘incidence of non-iconic chronology’, which describes the relation of Â�non-iconic cases to a certain text length. Advancing this notion, I depart from my Â�earlier concept defined in Yevseyev (2005), where the parameter ‘density of Â�minimal narrative units’, i.e. frequency of both iconic and non-iconic event pairs in a text passage of certain length, was analysed. Here the same approach is projected on non-iconic cases in the hope of discovering some significant correspondences. To calculate the incidence of non-iconic chronology in text samples, the following formula has been used:



Incidence of non-iconic chronology =

Number of non-iconic cases in a text sample Length of the text sample in 1,000 characters



Non-iconic chronology in English narrative texts 

The approximate length of each text sample was estimated as a product of the total number of lines in the sample and the average number of characters in a line. For example, the text sample for “The Sea-Wolf ” by J. London contains three non-iconic cases, its estimated length is 24.6 thousand characters; the density of non-iconic cases is therefore 3/24.6 = 0.12 non-iconic cases per 1,000 characters or, in other words, one non-iconic case per 4 book pages, provided one ordinary book page contains 2,000 characters.

4.2â•… Results Apart from those texts that do not contain non-iconic cases and accordingly have a zero value on this parameter, and some outliers with very high density of non-iconic cases, the substantial majority of the texts demonstrate values that are almost evenly Â�distributed between 0.03 and 0.25. In other words: reading a narrative text, one Â�encounters Â�micro-structural deviations from the principle of chronological order as often as once in every 16 ordinary book pages, at the very least, and up to once in every 2 ordinary book pages, at the very most. Only three texts of the corpus display values over 0.25; all three texts are short stories, e.g. Mark Twain’s “How I edited an Â�agricultural paper”, which has a value of 0.50 (one non-iconic case per page, on average). What is more, in this story the three non-iconic sequences are placed in the immediate Â�vicinity of each other, thus forming a kind of ‘non-iconic passage’ of the type B-A-D-C-F-E (compare the example of the type C-B-A from the text by Dostoyevsky above): The next morning I found a similar group at the foot of the stairs, and scattering couples and individuals standing here and there in the street and over the way, watching me with interest. The group separated and fell back as I approached, and I heard a man say, “Look at his eye!” I pretended not to observe the notice I was attracting, but secretly I was pleased with it and was purposing to write an account of it to my aunt. I went up the short flight of stairs, and heard cheery voices and a ringing laugh as I drew near the door, which I opened, and caught a glimpse of two young rural-looking men, whose faces blanched and lengthened when they saw me, and then they both plunged through the window with a great crash. I was surprised.

As this passage contains 742 characters, the parameter “incidence of non-iconic Â�chronology” would have a value of 4.0 non-iconic cases per 1,000 characters, or 8 Â�non-iconic cases per page, if the whole text were organized in a similar manner. My results suggest, however, that so many non-iconic cases are never grouped so closely together in real narrative texts.

4.3â•… Discussion Very similar results were obtained for the German and Russian corpora. In Yevseyev (2009:â•›290–291) values from 0.02 up to 0.25 were documented for the bulk of Â�German

 Vyacheslav Yevseyev

narrative texts, provisional results for Russian narrative texts are likewise from 0.03 up to 0.25. As the data drawn from English tally almost exactly with those from Â�German and Russian, we may assume some hidden regularities behind the Â�incidence of Â�non-iconic chronology. These rules may issue from the principle of temporal Â�iconicity, which prohibits a too high occurrence of deviations from this principle in real Â�narrative texts. Only in separate texts (short stories) or within special passages of longer texts non-iconic cases can be immensely ‘dense’, as in the above-mentioned text by Mark Twain or in Günther Grass’s short story “Der Slawist”, the latter showing the highest value for whole texts so far documented: 1.36 non-iconic cases per 1,000 characters, i.e. almost three non-iconic cases per page.

5.â•… Typical structures for non-iconic chronology 5.1â•… Results One of the linguistically relevant results of the present study is a clear picture of what syntactic structures are used to convey non-iconic chronology and how frequently they are used. All 256 non-iconic cases were indexed according to the parameters ‘verb forms’, ‘conjunction’, ‘punctuation mark’ (as in English no strict rules exist for using commas with temporal clauses, cases with and without a comma were subsumed under one type). The analysis has produced 39 types, 17 of which are represented by only one case (see Table 4). Table 4.╇ Typical structures for non-iconic chronology in English Rank

Type

Number of cases

1

[Past Ind](,) as [Past Ind]

2

[Past Ind](,) when [Past Ind]

48

18.8

3

[Past Ind]. [Past Perf]

39

15.2

4

[Past Ind](,) when [Past Perf]

12

4.7

58

Percentage 22.7

5-6

when [Past Ind](,) [Past Perf]

11

4.3

5-6

before [Past Ind](,) [Past Ind]

11

4.3

7

[Past Ind](,) after [Past Perf]

10

3.9

8

[Past Ind], [Part I]

6

2.3

9–10

[Past Ind]; [Past Perf]

5

1.9

9–10

[Past Ind](,) after [Past Ind]

5

1.9

11–14

before [Past Perf](,) [Past Ind]

4

1.6

11–14

before [Past Ind](,) [Past Perf]

4

1.6

(Continued)



Non-iconic chronology in English narrative texts 

Table 4.╇ (Continued) Number of cases

Percentage

[Past Ind]: [Past Perf]

4

1.6

11–14

[Past Ind](,) as soon as [Past Ind]

4

1.6

15–16

[Past Perf] after [Past Perf]

3

1.1

15–16

[Past Ind](,) [having Part II]

3

1.1

17–22

[Past Perf], [Part I]

2

0.8

17–22

[Past Perf] when [Past Perf]

2

0.8

17–22

[Past Ind], but [Past Perf]

2

0.8

17–22

[Past Ind], [Past Perf]

2

0.8

17–22

[Past Ind] – [Past Perf]

2

0.8

17–22

[Past Cont]. [Past Perf]

2

0.8

23–39

types represented by only one case

17

6.6

256

100

Rank

Type

11–14

Total

Rank 1. ‘[Past Ind](,) as [Past Ind]’, e.g. “A frown of scorn darkened quickly his forehead as he heard again the silly laugh of the questioner” (J. Joyce, A Portrait of the Artist as a Young Man, 361) or “‘We can’t argue about it here,’ Tom said impatiently, as a truck gave out a cursing whistle behind us” (F.S. Fitzgerald, The Great Gatsby, 120). Here the juxtaposing of the same tense forms and the meaning of the conjunction may suggest a simultaneous reading of the temporal relation, so that in the course of analysis very many candidate cases of this type had to be excluded from further consideration. In the 58 remaining cases there was a clear connotation of causation; as a result the situation conveyed in the immediately preceding clause is thought of as only possible due to the situation expressed in the immediately following clause. Even if the two situations overlap temporally, the central phases of them are not simultaneous and a non-iconic case is the result. Rank 2. ‘[Past Ind](,) when [Past Ind]’, e.g. “‘I thought I’d better catch up, Mrs.  Lawrence,’ he said rather ambiguously when he arrived” (F.S. Â�Fitzgerald, This Side of Â�Paradise, 201) or “‘I’ll tell him that,’ said Mr. Yorke, when his Â�foreman Â�mentioned the rumour […]” (Ch. Bronte, Shirley, 600). Also in this type Â�simultaneous readings are possible, so that an additional meaning of causation or general world knowledge is necessary for an interpretation of the temporal relation as sequential and non-iconic. Rank 3. ‘[Past Ind]. [Past Perf]’, e.g. “‘What’s that for?’ Flem Snopes said. He had approached” (W. Faulkner, The Hamlet, 300). This type – as well as the type ‘[Past Ind]; [Past Perf]’ (see below) – is very important as it is not always mentioned in comprehensive grammars and is very difficult to find via electronic search. Though

 Vyacheslav Yevseyev

the two clauses are not combined within a coherent syntactic structure, the use of Past Perfect and its reference to the previous event (temporal anaphora) produces a clear textual connection. Often special words such as of course, apparently, perhaps are used to show that the situation given in the immediately following clause is an explanation: “‘You must not get wet,’ she smiled, speaking Italian. Of course, the hotel-keeper had sent her (E. Hemingway, Cat in the Rain, 160). Rank 4. ‘[Past Ind](,) when [Past Perf]’, e.g. “‘I like to dine here,’ he said to his aunt when he had completed his meal […]” (W. Thackeray, Vanity Fair, 400) or “‘I regret, though,’ said Gino, when they had finished laughing, ‘that I toppled him onto the bed […]’” (E.M. Forster, Where Angels Fear to Tread, 40). Rank 5–6 (a). ‘before [Past Ind](,) [Past Ind]’, e.g. “Before he left the house he wrote a short note to the District Nurse […]” (A. Cronin, The Citadel, 150) or “But it came not, that awakening moment! It came not, even when, just before they left the house, Clifford stole to the parlor-door, and made a parting obeisance to the sole occupant of the room (N. Hawthorne, The House of the Seven Gables, 251). Rank 5–6 (b). ‘when [Past Ind](,) [Past Perf]’, e.g. “When we made that start for Bankok [sic] we had been already three months out of London” (J. Conrad, Youth, 10) or “When he returned to Mrs. Warren’s rooms, the gloom of a London winter evening had thickened into one grey curtain […]” (A. Conan Doyle, The Adventure of the Red Circle, 71). The only difference to Type 4 is that the conjunction when is used within the clause referring to the temporally posterior event; however, in Type 4 Past Perfect is replaceable with Past Indefinite, whereas in this type the use of Past Perfect (often in association with already) is obligatory to support the Â�correct Â�interpretation of the Â�temporal relation (otherwise the temporal relation could be interpreted as iconic). This fact provides additional evidence to the observation that, all other things being equal, the anterior event is associated with the Â�subordinate clause, and the posterior event with the main clause (Jou & Harris 1990:â•›21–22 or Jansen 2008:â•›83); to neutralize this rule, special lexical or grammatical markers are needed. Significantly, when is not likely to be replaceable with as in this type (at least according to my corpus data) since the connoted causal semantics of as would Â�contradict the temporal order Â�suggested by the juxtaposed tense forms of Past Â�Indefinite and Past Perfect. Rank 7. ‘[Past Ind](,) after [Past Perf]’, e.g. “He felt extremely bold after the dull girl in so vain a disguise of rose-colour had responded to the call of her mother […]” (H. James, The Portrait of a Lady, 401) or “‘And then there’s the Butterfly,’ Alice went on, after she had taken a good look at the insect with its head on fire […]” (L. Carroll, Through the Looking-Glass, 160). Rank 8. ‘[Past Ind], [Part I]’, e.g. “Richard’s trance was broken. ‘Mr. Blaize!’ he said, recognizing the farmer’s voice” (G. Meredith, The Ordeal of Richard Feverel, 201). As a rule, the primary function of Participle I is to express simultaneity with the Â�situation of the main clause, but the meaning of causation suggests a sequential interpretation of this relation. Sometimes the non-simultaneous reading is supported by special



Non-iconic chronology in English narrative texts 

lexical means: “Morris went away, taking another kiss first” (H. James, Washington Square, 250); among iconic cases, structures with a preposed ‘Part I’ clause, which Â�self-evidently conveys the anterior event, are with 61 tokens 10 times as frequent as the non-iconic structures of the described type, this finding being a piece of evidence for the effectiveness of the temporal iconicity principle. Rank 9–10 (a). ‘[Past Ind]; [Past Perf]’, “She said, ‘Yes, father.’ She seemed quite satisfied; she had brought him down by fifty centavos a head” (G. Greene, The Power and the Glory, 200). Rank 9–10 (b). ‘[Past Ind](,) after [Past Ind]’, e.g. “The Court discharged me from this case right after he sentenced you this afternoon” (W. Faulkner, The Mansion, 50). These most frequent 10 types represent 80 percent of all cases. Among the less frequent types, the following are most interesting: –

–

–

–

‘before [Past Perf](,) [Past Ind]’, e.g. “Before her eyes had adapted themselves to the obscurity, a hand grasped her own, with a firm, but gentle and warm pressure […]” (N. Hawthorne, The Scarlet Letter, 300); in this type there is a certain contradiction between the meaning of the tense form and the temporal conjunction: the highlighted last phase of a process is represented as posterior to a punctual event; ‘[Past Ind], [having Part II]’, e.g. “Vaguely he felt it was his mother’s fault, and he felt stronger, having blamed her […]” (A. Miller, I Don’t Need You Any More, 30); among iconic cases, structures with a (preposed) ‘having Part II’ clause are with 44 tokens more than 10 times as frequent as non-iconic structures of the described type (3 tokens – cf. Table 4), this fact strongly supporting the Â�assumption of Â�temporal iconicity in narrative; in some cases a colon signals that the previous event is intended as explanation: ‘[Past Ind]: [Past Perf]’, e.g. “This was known at Briarfield: the newspapers had reported it […]” (Ch. Bronte, Shirley, 600); in some cases dashes or brackets are used to show that the previous event is only conceived as supplementary information: ‘[Past Ind] – [Past Perf]’, e.g. “The Â�cabman put their bags down in the passage – they had left heavy luggage at the station – and strolled about till he came on the landlady’s room and woke her […]” (E.M. Forster, Where Angels Fear to Tread, 80) or ‘[Past Ind] ([Past Perf])’, e.g. “Montag walked from the subway with the money in his pocket (he had visited the bank which was open all night every night with robot tellers in attendance) and as he walked he was listening to the Seashell Radio in one ear”  (R. Bradbury, Fahrenheit 451, 120).

In many cases one of the situations is presented as durative (atelic), but the semantics of conjunctions or tense forms makes it possible to interpret the temporal relation �correctly (as non-simultaneous and non-iconic):

 Vyacheslav Yevseyev

–

–

–

‘until [Past Ind] [Past Ind]’, e.g. “Until the barrel-organ stopped playing Constantia stayed before the Buddha, wondering, but not as usual, not vaguely” (K. Mansfield, The Daughters of the Late Colonel, 118); the situation of the main clause is atelic (as it has no intrinsic limit) and is presented as limited extrinsically by the situation of the subordinate clause; ‘[Past Ind]. [Past Perf Cont]’, e.g. “‘I wish I could gather knowledge as carelessly,’ Martin broke out half an hour later. He had been closely analyzing Brissenden’s mental equipment” (J. London, Martin Eden, 300); ‘[Past Cont]. [Past Perf]’, e.g. “We were now returning from the long walk. We had reached the middle of a clean Faubourg, where the houses were small but looked pleasant” (Ch. Brontë, Villette, 700); the situation given in the main clause is presented as a process, whereas the meaning of the Past Perfect produces a Â�non-simultaneous reading.

5.2â•… Discussion The significance of the results obtained on English material is partially supported by the findings from the German corpus. In German the most frequent type is Â�‘[Präteritum], als [Präteritum]’, represented by 38.8 percent of cases (cf. Yevseyev 2009:â•›291). The two most frequent types in English constitute together approximately just as much (41.5 percent). This may be explained by the fact that in German the conjunction wenn Â�(etymologically equivalent to when) is used only for iterative situations in the past and is replaced by als (etymologically equivalent to as) for semelfactive, i.e. happening only once, actions. Thus, one could assume a natural rule according to which approximately 40 percent of all non-iconic cases are to be conveyed by structures with simple past forms and with a conjunction neither specifying simultaneity nor non-simultaneity; as in English two temporal conjunctions of this kind are available, they are both used in these structures; in German, the only conjunction of this type is responsible for nearly the same percentage of such cases. Strikingly, temporal conjunctions of posteriority and anteriority (before and after, as well as till/until and as soon as), which suggest a non-ambiguous reading of theÂ� temporal relation as sequential, are used only in a relatively small part of Â�non-iconic cases (46 cases of 256, i.e. 18.0 percent). Cases without a temporal conjunction, which are based exclusively on the contrast of the meanings of Past Indefinite and Past Â�Perfect, are even more numerous in comparison (61 cases of 256, i.e. 23.8 percent). In Â�German, structures with conjunctions of anteriority and posteriority (bevor/ehe ‘before’ and nachdem ‘after’ as well as bis ‘till/until’ and sobald/kaum dass ‘as soon as’) are Â�somewhat more frequent among non-iconic cases (25.6 percent), whereas in Russian corresponding structures (with conjunctions posle togo kak ‘after’, prežde čem ‘before’, pered tem kak ‘before’, kak tol’ko ‘as soon as’, poka (ne) ‘till/until’) make



Non-iconic chronology in English narrative texts 

up only 6.5 percent of all non-iconic cases; the reliability of these findings cannot be doubted, since the figures obtained for the three corpora are very similar in many other Â�parameters, such as the ratio of non-iconic cases and ‘incidence of non-iconic chronology’ (see above). It must be stressed here again that intensive manual studies, rather than searches of electronic corpora, are necessary to draw an adequate picture of structures conveying non-iconic chronology. Interestingly enough, in English narrative texts no formally unmarked cases were identified; these can be called ‘hysteron proteron’ with reference to classical examples such as “Moriamur et in media arma ruamus” (‘Let us die, and charge into the thick of the fight’, Vergil, Aeneid). The only candidate could be the following case, which, Â�however, contains two pluperfect forms and an adverbial phrase: “But by the time Connie had found it [the note put on the jug of hot water] the steam had made the writing too faint to read […]” (K. Mansfield, The Daughters of the Late Colonel, 117); here it is shown only that both situations were completed at a certain moment, but world knowledge helps us to conclude that the reaction of ink with the steam precedes the girl’s discovery of the note, which marks an end to this process. In my German corpus there are several cases of this type, e.g.: [ich] fuhr mit dem nächsten Zug weg. Ich bezahlte dem Hotelier sogar den angebrochenen Tag (H. Böll, Ansichten eines Clowns, 125) ‘[I] left by the next train. I even paid the hotel for the full day’.  (H. Böll, The Clown, 118)

In my Russian corpus, there is even a larger number of cases of this type (more than 10 percent of all non-iconic cases), e.g. in A. Bely’s Petersburg, in A. Grin’s She who runs on the waves, in some other modernist texts and, last but not least, in folk tales. It cannot be an accident that formally unmarked non-iconic cases are found Â�primarily in texts that either originate from oral narrative tradition or intentionally break with written narrative tradition, the latter being characterized by the elaborateness and Â�explicitness of linguistic expression (Roberts & Street 1997:â•›168). The reversal of Â�temporal order without explicit marking may therefore be a feature of oral narrative tradition Â�resulting from the impossibility to ‘improve’ the texts in the editing phase by means of adding necessary temporal markers in order to make the temporal relations in the text as transparent as possible. The assumed absence of formally unmarked non-iconic cases in English narrative texts may be explained also by the tendency to use an exact temporal orientation with every predicate. This is consistently followed in English in the form of the ‘sequence of tenses’ rule both in temporal clauses and in reported speech. In Russian, which practises the relative use of tenses in subordinate clauses, the inversion of the Â�temporal relation may remain formally implicit more often than in English, thus producing a fair number of ‘hysteron proteron’ cases. As far as the frequency of such formally unmarked

 Vyacheslav Yevseyev

non-iconic cases and the relative vs. absolute use of tense forms is Â�concerned, German may lie between Russian and English, since in German the ‘sequence of tenses’ rule is followed in temporal clauses, but not in reported speech, where it is usually replaced by special forms of the subjunctive mood, which are Â�temporally oriented to the main clause. The explanation given here is only provisional and the problem needs further elucidation.4 Jou and Harris’s (1990:â•›21–22) remark that the subordinate clause tends to describe the anterior event and the main clause the posterior event (see above), is strongly backed up by the statistical data from my corpus. Among iconic cases, the Â�linearisation Â�pattern ‘sub → main’ is represented by 459 cases, and the pattern ‘main → sub’ by only 158 cases (numerous cases based on the pattern ‘main → main’, e.g. ‘[Past Ind], [Past Ind]’, were not counted here). In contrast, non-iconic cases Â�demonstrate the inverse distribution: The pattern ‘sub → main’ occurs in 33 cases, the pattern ‘main → sub’ in 161 cases (the pattern ‘main → main’, e.g. ‘[Past Ind], [Past Perf]’, was not taken into account). Â�Consequently, cases in which the anterior event is conveyed by the Â�subordinate clause and the posterior event by the main clause are three times and more as frequent as cases in which the reverse is true. This concealed regularity also seems to explain Ford’s (1993:â•›146–147) observation that temporal clauses, which usually occur Â�sentence-finally if they express simultaneous situations, tend to occur sentence-initially if sequenced events are involved: Being associated with the previous event in a segment of the narrative chain, a temporal clause is inevitably placed before the respective main clause in accordance with the principle of temporal iconicity.

6.╅ Summary and outlook This study has discussed the results of a corpus investigation into the phenomenon of non-iconic chronology understood as micro-structural deviation(s) from the natural order of event presentation in narrative texts. Based on a manually indexed corpus of samples extracted from literary texts, an exhaustive description of all instances of non-iconic chronology, both formally marked and unmarked, has been given in order to scrutinize some implications of the iconicity principle. Manual evaluation of real narrative texts (primarily English, but also German and Russian) also made it �possible

.â•… As far as possible cases without explicit marking of the ‘reversed’ temporal relation are concerned, it must be mentioned here that Bardovi-Harlig (1994:â•›273) documents 3 cases of this type within a sample of 103 “reverse-order reports” (i.e. ‘non-iconic cases’); this sample represents, however, not literary texts but spontaneous utterances of learners of English as a second language. All in all, this observation demonstrates that the frequency of formally Â�unmarked cases of non-iconic chronology is relatively low in English.



Non-iconic chronology in English narrative texts 

to outline the numerical distribution of different structures conveying non-iconic chronology. One of the most important results is that the principle of temporal iconicity is rather effective in narrative texts, since non-iconic cases make up on average about 5 to 7 percent of all cohesively expressed pairs of events; many texts do not seem to contain non-iconic cases at all, and even in those very few texts that deviate very significantly on this parameter the ratio of non-iconic cases is relatively low (30 percent). Temporal conjunctions of anteriority and posteriority (before and after, as well as till/until and as soon as), which express the order of events unambiguously, are used in only a small part of all non-iconic cases (about one fifth), the rest of the structures Â�conveying non-iconic chronology is characterized by a complex interaction of the semantics of tense forms, conjunctions and adverbs. One fourth of all cases is based on the non-conjunctional structure ‘[Past Indefinite] → [Past Perfect]’. In this article, the notion “incidence of non-iconic chronology” was put forward. This term describes how often syntactic structures with the reverse order of events occur in relation to a certain amount of text. The results show that a reader can expect to encounter such cases as a rule only once per two ordinary book pages or even less frequently. This finding provides indirect evidence for the effectiveness of the Â�iconicity principle, which seems to prohibit a too high occurrence of deviant cases in real Â�narrative texts. As the three corpora have demonstrated nearly the same distribution of the corresponding values, there seems to exist a certain cross-linguistic regularity behind these data. The empirical results presented in this article may shed new light on the Â�temporal structure of literary texts. In future, data obtained from automatic searches of Â�electronic corpora could be used to complement the picture, which was drawn from manual analyses, by numerically characterizing particular aspects of non-iconic Â�chronology, e.g. the ratio of non-iconic cases among structures containing certain temporal conjunctions.

Narrative texts quoted Böll, H. Ansichten eines Clowns. München: dtv, 1985. Böll, H. The Clown. Translated by L. Vennewitz. New York NY: Penguin, 1994. Bradbury, R. Fahrenheit 451. New York NY: Simon and Schuster, 2003. Brontë, C. Shirley. Oxford: Clarendon, 1979. Brontë, C. Villette. Oxford: Clarendon, 1984. Carroll, L. Through the Looking-Glass. In The Complete Works, L. Carroll, 129–249. London: Nonesuch, 1939. Conan Doyle, A. The Adventure of the Red Circle. In His Last Bow: Some Reminiscences of Â�Sherlock Holmes, A. Conan Doyle, 62–82. London: Penguin, 1997. Conrad, J. Youth. In Youth and Two Other Stories, J. Conrad, 3–42. New York NY: Doubleday, Page and Co., 1924.

 Vyacheslav Yevseyev Cronin, A. The Citadel. Moscow: Foreign Languages, 1963. Dostoevskij, F.M. Prestuplenie i nakazanie. Moskva: Xudožestvennaja literatura, 1983. Dostoyevsky, F. Crime and Punishment. Translated by D. McDuff, London etc: Penguin, 2003. Faulkner, W. The Hamlet. New York NY: Random House, 1940. Faulkner, W. The Mansion. London: Chatto and Windus, 1961. Fitzgerald, F.S. The Great Gatsby. London: Penguin, 1990. Fitzgerald, F.S. This Side of Paradise. In This Side of Paradise, The Crack-Up and Other Â�Autobiographical Pieces, F.S. Fitzgerald, 13–270. London etc.: Bodley, 1971. Forster, E.M. Where Angels Fear to Tread. London: Arnold, 1975. Grass, G. Der Slawist. In Werkausgabe in zehn Bänden, Vol. 1, G. Grass, 318–319. Darmstadt und Neuwied: Luchterhand, 1987. Greene, G. The Power and the Glory. London: Heinemann, 1971. Hawthorne, N. The House of the Seven Gables. Columbus OH: Ohio State University Press, 1971. Hawthorne, N. The Scarlet Letter. Columbus OH: Ohio State Univ. Press, 1971. Hemingway, E. Cat in the Rain. In Selected Stories, E. Hemingway, 159–162. Moscow: Â�Progress, 1971. James, H. The Portrait of a Lady. London: Bodley, 1976. James, H. Washington Square. In The Europeans. Washington Square, H. James, 197–392. Â�London: Bodley, 1974. Joyce, J. A Portrait of the Artist as a Young Man. In Dubliners, A Portrait of the Artist as a Young Man, J. Joyce, 223–441. Moscow: Progress, 1982. London, J. Martin Eden. Moscow: Foreign Languages, 1953. London, J. The Sea-Wolf. In Novels and Stories, J. London, 481–771. New York NY: Literary Classics of the U.S., 1982. Mansfield, K. The Daughters of the Late Colonel. In The Garden Party and Other Stories, K. Mansfield, 88–119. Melbourne: Penguin, 1952. Twain, M. How I Edited an Agricultural Paper. In Nineteenth-Century American Short Stories, 308–313. Moscow: Progress, 1970. Meredith, G. The Ordeal of Richard Feverel: A History of a Father and Son. London: Constable, 1914. Miller, A. I Don’t Need You Any More. In I Don’t Need You Any More: Stories, A. Miller, 3–52. New York NY: Viking, 1967. Steinbeck, J. Travels with Charley: In Search of America. New York NY: Viking, 1962. Swift, J. Gulliver’s Travels. New York NY: MacMillan, 1924. Thackeray, W.M. Vanity Fair: A Novel Without a Hero. New York NY: Garland, 1989.

References Abelson, R.P. 1981. Psychological status of the script concept. American Psychologist 36: 715–729. Bardovi-Harlig, K. 1994. Reverse order reports and the acquisition of tense: Beyond the Â�principle of chronological order. Language Learning 44: 243–282. Behaghel, O. 1932. Deutsche Syntax, Vol. 4: Wortstellung, Periodenbau. Heidelberg: Winter. Brown, G. & Yule, G. 1983. Discourse Analysis. Cambridge: CUP. Clark, H.H. & Clark, E.V. 1977. Psychology and Language: An Introduction to Psycholinguistics. New York NY: Harcourt.



Non-iconic chronology in English narrative texts 

Comrie, B. 1976. Aspect: An Introduction to the Study of Verbal Aspect and Related ProbÂ�lems. Cambridge: CUP. De Cuypere, L. 2008. Limiting the Iconic: From the Metatheoretical Foundations to the Creative Possibilities of Iconicity in Language [Iconicity in Language and Literature 6]. Amsterdam: John Benjamins. Declerck, R. 2006. The Grammar of the English Verb Phrase, Vol. 1: The Grammar of the English Tense System. Berlin: de Gruyter. Degand L. 2000. Contextual constraints on causal sequencing in informational texts. Functions of Language 7: 173–201. Dik, S. 1997. The Theory of Functional Grammar, Vol. 1: The Structure of the Clause. Berlin: de Gruyter. Dressler, W.U. 1989. Semiotische Parameter einer textlinguistischen NatürlichkeitsÂ�theorie. Wien: Akademie der Wissenschaften. Fleischman, S. 1990. Tense and Narrativity: From Medieval Performance to Modern Fiction. Â�London: Routledge. Genette, G. 1972. Figures, Vol. 3 : Discours du récit. Paris: Seuil. Givón, T. 1985. Iconicity, isomorphism and non-arbitrary coding in syntax. In Iconicity in Â�Syntax, J. Haiman (ed.), 187–219. Amsterdam: John Benjamins. Givón, T. 2001. Syntax, Vol. 2. Amsterdam: John Benjamins. Haiman, J. 1985. Natural Syntax: Iconicity and Erosion. Cambridge: CUP. Haiman, J. & Thompson, S.A. (eds). 1988. Clause Combining in Grammar and Discourse Â�[Typological Studies in Language 18]. Amsterdam: John Benjamins. Haspelmath, M. 2008. Frequency vs. iconicity in explaining grammatical asymmetries. Â�Cognitive Linguistics 19: 1–33. Jakobson, R. 1971. [1965]. Quest for the essence of language. In Selected Writings, Vol. 2: Word and Language, R. Jakobson, 345–359. The Hague: Mouton. Jansen, F. 2008. Order of mention and the position of temporal clauses in Dutch comÂ�plex Â�sentences. In LineÂ�aÂ�risation and SegÂ�menÂ�tation in Discourse: Multidisciplinary Approaches to Discourse, W. Ramm & C. Fabricius-Hansen (eds), 77–85. Oslo: Oslo University Press. 〈http://www.hf.uio.no/ilos/forskning/konferanser/mad08/proceedings/MAD08-Â� Proceedings_Complete.pdf〉 (15, March, 2010). Johanson, L. 1995. On Turkic converb clauses. In Converbs in Cross-Linguistic Perspective: Â�Structure and Meaning of Adverbial Verb Forms, M. Haspelmath & E. König (eds), 313–347. Berlin: de Gruyter. Jou, J. & Harris, R.J. 1990. Event order versus syntactic structure in recall of adverbial complex sentences. Journal of Psycholinguistic Research 19: 21–42. Klein, W. 1994. Time in Language. London: Routledge. Lüdeling, A. & Kytö, M. (eds). 2008. Corpus Linguistics: An International Handbook, Vol. 1 [Handbooks of Linguistics and Communication Science 29]. Berlin: de Gruyter. Lüdeling, A. & Kytö, M. (eds). 2009. Corpus Linguistics: An International Handbook, Vol. 2 [Handbooks of Linguistics and Communication Science 29]. Berlin: de Gruyter. Myhill, J. 2001. Typology and discourse analysis. In The Handbook of Discourse Analysis, D. Schiffrin, D. Tannen & H. Hamilton (eds), 161–174. Oxford: Blackwell. Opačić, G. & Osgood, C.E. 1984. Natural order in cognizing and clause order in sentencing. Folia Linguistica 18: 295–343. Prince, G. 1987. A Dictionary of Narratology. Lincoln NB: University of Nebraska Press.

 Vyacheslav Yevseyev Roberts, C. & Street, B. 1997. Spoken and written language. In The Handbook of Sociolinguistics, F. Coulmas (ed.), 168–186. Oxford: Blackwell. Sasse, H.-J. 1991. Aspect and aktionsart: A reconciliation. Belgian Journal of Linguistics 6: 31–45. Simone, R. 1995. Iconic aspects of syntax: A pragmatic approach. In Iconicity in Language [Current Issues in Linguistic Theory 110] R. Simone (ed.), 153–169. Amsterdam: John  Benjamins. Tai, J. 1985. Temporal sequence and Chinese word order. In Iconicity in Syntax [Typological Studies in Lanuage 6], J. Haiman (ed.), 49–69. Amsterdam: John Benjamins. Tomasello, M. 2003. Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge MA: Harvard University Press. Vendler, Z. 1967. Linguistics in philosophy. Ithaca NY: Cornell University Press. Yevseyev, V. 2003. Ikonizität und Taxis: Ein Beitrag zur Natürlichkeitstheorie am Beispiel des Deutschen und Russischen. Frankfurt: Peter Lang. Yevseyev, V. 2005. Measuring narrativity in literary texts. In Narratology Beyond Literary Â�Criticism: Mediality, Disciplinarity, J.C. Meister (ed.), 109–124. Berlin: de Gruyter. Yevseyev, V. 2008. Kodierung von Taxisrelationen und narrative Eigenschaften russischer Texte des Realismus und des Modernismus. In Textkohärenz und Narration: Untersuchungen russischer Texte des Realismus und der Moderne, R. Hodel & V. Lehmann (eds), 331–354. Berlin: de Gruyter. Yevseyev, V. 2009. Neikoničeskaja xronologija v nemeckix povestvovatel’nyx tekstax Novogo vremeni. Russkaja Germanistika 5: 287–296.

A burning world of war How iconicity works in constructing the fictional world view in A Farewell to Arms Xinxin Zhao

Kunming University of Science and Technology How do literary works access the iconic potential inherent in language to construct their fictional world view? A case study of Hemingway’s A Farewell to Arms is presented to show that instead of its commonly accepted dual themes of love and war, the novel’s overriding theme is war – the universal and pervasive condition of human existence, itself cast as a war against the forces that conspire to make life shorter. The burning world of war is produced by the iconic force of literary language, from lexical choice to syntactic structure, from internal paragraph patterning to the patterning of paragraph sequences, from chapter level to the overarching narrative structure itself. Through quantitative and sequential iconic means, the abstract theme is made accessible, and the formless fictional world is given shape.

1.â•… Introduction Iconicity is defined as a semiotic notion [that] refers to a natural resemblance or analogy between the form of a sign (‘the signifier’, be it a letter or sound, a word, a structure of words, or even the absence of a sign) and the object or concept (‘the signified’) it refers to in the world or rather in our perception of the world.  (Fischer & Ljungberg 2010)

In language, the arrangement of linguistic elements is often found to resemble the concept they stand for, from relatively direct mirroring (e.g. the sequence of verbal signifiers usually imitates the sequential order of occurrences in reality) to much more suggestive imitation (e.g. a cluster of abrupt sentences may iconize speed). Contrary to Saussure’s (2006:â•›788) axiom that “the linguistic sign is arbitrary”, considerable Â�semiotic and linguistic research in the 20th century has shown that iconicity operates at every level of language (phonology, morphology, syntax, etc.) As for the poetic use of language, the notion of icon “is ultimately related to the Platonic process of mimesis”: Aristotle broadened the notion “from a chiefly visual

 Xinxin Zhao

Â� representation to embrace all cognitive and epistemological experience” (Sebeok 2001:â•›50). A literary text, especially a self-conscious one, constitutes an ideal field for iconicity study. As Müller indicates: In the conflict between the arbitrarist model and the non-arbitrarist or ‘iconicist’ model the literary critic will rather find himself on the side of those who accept the concept of iconicity, because iconic forms of expression are so frequent, almost omnipresent, in aesthetic or poetic language that their denial would be tantamount to a denial of the nature of the verbal artifact. (1999:â•›393)

Thus, it is the role of literature to exploit the potential of iconicity, since the Â�interaction and interdependence of form and meaning is one of the fundamental Â�characteristics of the aesthetic use of language. According to Nöth (2001:â•›26), when iconicity is adopted in literary analysis “as a prerequisite of mutual understanding”, it could achieve a parallelism between the writer’s message and the reader’s interpretation of it. During the writing Â�process, the writer employs the iconic force in language to enact the fictional reality through the form of the text. During the reading process, the reader does not only perceive the revelation of the fictional world, but enters into it iconically through recognition of the similarities between linguistic structure and the underlying Â�concept. The basis for this “mutual understanding” is the fact that iconicity is often founded on what is considered as consistent universal correspondences between form and meaning (e.g. symmetrical structure may iconize correlated concepts, as in antithesis, parallelism, and chiasmus). This is the main reason why this paper adopts theories of iconicity as its theoretical framework. Compared with other literary theories which tend to be culture-oriented, iconicity is more text-oriented, and the formal evidence from the text is both substantial and universally recognizable. In this study, Hemingway’s A Farewell to Arms (AFTA) is chosen as a case study. The burning-log metaphor, the major icon of the fictional world in this novel, will be taken as a starting point. Next, on the basis of some iconic examples, the whole world-view/theme of the novel will be illustrated by means of textual analyses. As Halliday and Hasan (2001:â•›332) state, the analysis of a text is “a means to an end” and “not an end in itself ”. The aim of textual analysis is to analyze the panorama of the fictional world. This paper does not undertake to give a full account of all the iconic usages in the novel, but wishes to demonstrate how the abstract world-view can be made visible through the form of language. If it is reasonable for Plato to compare a poet with a painter in his Republic, one can suppose that a literary text is like a painting: instead of it being depicted with colors and lines, it is portrayed by means of words and sentences. Iconicity is a tool to exhibit this painting, of presenting it to the reader. However, in order to decode the literary icons, it also appeals to readers’ interpretative ability.



A burning world of war 

2.â•… The burning-log metaphor – the major icon of the novel Usually, iconicity exists in the ‘natural’ similarities between the signifier and the Â�signified; or, as often as not, the similarities are inherent in both. The metaphor, however, establishes an analogous relationship between two dissimilar things by the Â�similarities conceived. It is this conceived similarity that makes the source and the Â�target domains adopt the same form. For example, from ‘he broke the chair’ to ‘he broke the rules’, from a concrete source domain to an abstract target, the same signifier ‘break’ is adopted because of the conceived common ground – something is damaged. Hence, the iconic aspect of metaphor lies in the fact that the sameness of form stands for the sameness of meaning. And from ‘breaking’ a material chair to breaking Â�immaterial rules, Â�according to Givón (1979:â•›316), “the process of metaphoric extension quite often yields Â�nonconcrete senses” from “the original concrete sense”. From the concrete to the abstract, from a formal signifier to a formless signified, the illustrative function of iconicity is highlighted. In this way, the abstract can be explained and constituted by the concrete through the similarities conceived. In a literary work, can a single metaphor be used to explain the theme and Â�constitute the whole fictional world? This is a possibility, according to Haiman, because as a Â�diagrammatic icon, metaphors “need only represent, and not necessarily Â�reproduce, reality, and it is possible because in order to do this, they need to reproduce only the essential attributives of the objects they denote” (1985:â•›11). For example, in the Â�metaphor ‘Life is a journey’, the event category or the source model is ‘a journey’ and the abstract category is ‘life’. ‘A journey’ serves as the basic level supplying the salient Â�attributes for the superordinate categories, such as: a journey is the distance between one place and another (life is the passage from birth to death), a journey may have obstacles Â�blocking the way (life tends to be beset by difficulties), and a journey has a destination to arrive at (in a life it is usual to achieve some objective). The salient Â�concrete Â�features of a Â�journey (distance, obstacles, destination) are thus mapped onto the abstract model of life Â�(passage, difficulties, objective). In this way, the salient Â�attributes of the source model are used to structure the abstract target model; and the isomorphism between the Â�structures of source and target is established. In other words, a metaphor Â�iconizes reality based on a point-to-point correspondence with the world experienced by Â�people. In the concluding chapter of AFTA, it is the burning-log metaphor that is used to Â�represent the fictional world with all the “essential attributives” of that world.

2.1â•… Twin-poles, no escape In the last chapter of the novel, the narrator, Frederic, finds out that the baby he had with Catherine is dead and also that “Catherine would die” (Hemingway 1957:â•›327). The moment when Frederic is losing everything that matters to him, a revelation comes

 Xinxin Zhao

to him in an epiphany: what the mechanism of this world is to be – “they killed you in the end” (327). And the burning-log metaphor, which is presented in this epiphanic moment, plays a summarizing role for the whole book: Once in camp I put a log on top of the fire and it was full of ants. As it commenced to burn, the ants swarmed out and went first toward the centre where the fire was; then turned back and ran toward the end. When there were enough on the end they fell off into the fire. Some got out, their bodies burnt and flattened, and went off not knowing where they were going. But most of them went toward the fire and then back toward the end and swarmed on the cool end and finally fell off into the fire. I remember thinking at the time that it was the end of the world and a splendid chance to be messiah and lift the log off the fire and throw it out where the ants could get off onto the ground. But I did not do anything but throw a tin cup of water on the log, so that I would have the cup empty to put whiskey in before I added water to it. I think the cup of water on the burning log only steamed the ants. (1957:â•›327–28)

The reason why this metaphor could be considered the major icon of this novel is that the overall structure of the novel and the fate of the main characters have an analogous relationship with this metaphor. It condenses the whole fictional world into a burning log. As Haiman states, iconic signs “have to simplify and abbreviate what they stand for” (1985:â•›11). Despite reducing the fictional world to a burning-log, the metaphor still includes all the essential elements of the novel. Table 1 shows the general correspondence between the source (the textual signifier) and the target (the world-view). The specific correspondence between the metaphor and each major character will be analyzed in more details in the next section. Table 1.╇ General correspondence between the metaphor and the world-view Source

Target

The burning-log

The general living condition – war

The burning end

The painful experience caused by war

The temporary cool end

The occasional relief from pain

Falling off into fire finally

The world breaks and kills everyone

The burning log is a miniature fictional world with twin-poles – the burning end and the temporary cool end, but without any escape. People living in their personal pain caused by war are like the ants driven and tortured by the omnipresent fire. Occasionally, they can turn to something for help, hoping to be released from pain for a while. However, no matter what kind of anesthesia they use, the pain can only be eased temporarily; they are constantly forced back to the burning end. Â�Consequently, the escape from this tragedy is a cruel illusion; there is no way off



A burning world of war 

the burning log, there is only the occasional cool spot. The metaphor is therefore Â�significant. For most people, an entire lifetime is spent going back and forth between the twin-poles. Finally the world breaks everyone impartially – everyone will fall into the fire sooner or later. In the metaphor, the ants “went first toward […] the fire”; then turned back “toward the (cool) end”. This process repeats itself (“most of them went toward the fire and then back toward the end and swarmed on the cool end”), and “finally (the ants) fell off into the fire”. The overall arrangement of the five books of this novel perfectly coincides with this sequence. As Nänny and Fischer (1999: xxii) observe, one of the main forms of diagrammatic iconicity is sequential form, i.e. the order of verbal signifiers and their local relations imitate the sequential order of occurrences in the fictional world. In this novel, this metaphor strictly adheres to the iconic sequential principle: the course taken by the ants on the burning-log imitates the sequence of occurrences in the fictional world, as Table 2 shows: Table 2.╇ Correspondence between the metaphor and the narrative structure Burning-log metaphor

Overall structure of the novel

Burning end

Book I: Wounded at the Italian front

Temporary Cool end

Book II: Short happy time in Milan

Burning end

Book III: Back to the front

Temporary Cool end

Book IV: Isolated peace in Switzerland

Falling off into fire

Book V: Death of Catherine

In the left column, the ants go back and forth between the two poles and then finally fall into the fire; in the right column, the occurrence of the major events in each book also follow this sequence. In Book I, Frederic, who is hit by a trench mortar (Chapter IX), is literally burned by the fire of war. Because of his wound, in Book II, he enjoys a short happy time with Catherine in the Milan hospital. In Book III, war separates the couple; Frederic is sent back to the burning end and almost killed by Â�Italian battle police for “treachery” (Chapter XXX). Frederic flees, not out of Â�cowardice but out of unwillingness to make a sacrifice for a cause that seems meaningless to him. It appears that he escapes from war by achieving an isolated peace with Catherine in Switzerland. However, because of the transient nature of the cool end, in Book V, the death of Catherine makes him lose everything that matters to him. The world kills, breaks, and leaves everyone dying or living alone. Frederic, like the ants, moves between the burning war end and the temporary cool end (staying with Catherine), but finally no one can escape destruction. Thereby, a point-to-point isomorphism is established. Thus, we could say that the burning-log metaphor iconizes the narrative structure of this novel.

 Xinxin Zhao

Narrative is a tool for knowing as well as telling. As Abbott notes, “we do not have any mental record of who we are until narrative is presented as a kind of armature, giving shape to that record” (2007:â•›3). Thus, in a novel, narrative is a way for the author to organize his understanding of the fictional world. In this novel, the metaphor Â�mirrors the way the narrator organizes his narrative, it puts the narrative in a nutshell and condenses the fictional world into a burning-log with twin-poles but no escape.

2.2â•… The burning-log metaphor and other major characters Since Frederic is the I-narrator and the main character of this novel, those major events in Table 2 are all narrated from Frederic’s point of view. Table 2 also could be used to show how the burning-log metaphor iconizes Frederic’s life span within the story. Similarly, as far as Catherine is concerned, a point-by-point correspondence can also be observed as shown by the following table: Table 3.╇ Correspondence between the metaphor and Catherine Burning-log metaphor

Major events in Catherine’s life

Burning end

The pain of losing her fiancé to the war

Temporary cool end

Playing ‘the rotten game’ with Frederic

Burning end

Frederic leaving her for the front

Temporary cool end

Short idyllic time in Switzerland

Falling off into fire

Death from childbirth

At the very beginning, Catherine has been wounded by war because she lost her fiancé in the Battle of the Somme. During their first meetings, Catherine and Â�Frederic begin flirting with each other in order to forget their personal troubles. Â�Flirting, which Frederic compares to bridge (“This was a game, like bridge, in which you said things instead of playing cards” [1957:â•›30]) and Catherine refers to as “the rotten game” (31), diverts Catherine’s thoughts from the death of her fiancé. Ironically, this Â�diverting game later becomes her source of suffering. When Catherine and Frederic begin Â�seriously loving each other and accomplish the oneness (“‘There isn’t any me. I’m you. Don’t make up a separate me’” [115]), war tears them apart – Frederic leaves her alone and pregnant. The transformation of love from distraction into pain signals the Â�transitory nature of happiness (or cool end). In the last book, Catherine fulfils her fantasy of lÂ�iving in a place where no one knows them. Soon, she loses the battle of Â�life-Â�and-death and dies in childbirth. Again, the attributives of the burning-log metaphor are in accordance with the occurrences in Catherine’s life. Another important character is Rinaldi, a surgeon in the army and Frederic’s “best friend and war brother” (171), whose source of pain or burning end is the enormous



A burning world of war 

pressure from war. Just as Rinaldi himself says, “‘[t]his war is killing me, […], I am very depressed by it’” (167), because, “‘All summer and all fall I’ve operated. I work all the time. I do everybody’s work. All the hard ones they leave to me’” (167). Â�Tragically enough, Rinaldi relieves the heavy burden of work by making himself overloaded with more work. Hard work ironically reduces the sensitivity to strain caused by work (“I don’t think; I operate.” [167]). Work is one of his cool ends – “I am only happy when I am working” (170). Once the work is over, he’s going back to the painfully burning end (“But now, baby, it’s all over. I don’t operate now and I feel like hell. This is a Â�terrible war, baby. You believe me when I say it.” [167]). Besides working, Rinaldi seeks solace in two things – “I only like two other things; one is bad for my work and the other is over in half an hour or fifteen minutes” (170) – namely, alcohol and sex. But neither of them provides him with a long-term remedy. As for the alcohol, it’s nothing but “self-destruction day by day”, and, “It ruins the stomach and makes the hand shake. Just the thing for a surgeon.” (172). As for sex, it only gives Rinaldi syphilis, which was still a terminal disease at the time. Though different people may use different means to alleviate their personal pain, the transient nature of the cool end does not vary with each individual. The final destruction definitely comes, just as Frederic says, “Or they killed you gratuitously like Aymo. Or gave you the syphilis like Rinaldi. But they killed you in the end” (327). In that way, Rinaldi is another ant on the burning log, who can not escape the fate of being destroyed by war. It seems that the twin-poles-no-escape is the life pattern in this fictional world. When living in a world in which war is endless and war fire is everywhere, to find an even evanescent comfort is an effort to survive. Frederic and Catherine look to love for escapism, to shut the war away whereas Rinaldi indulges in work, sex and alcohol to stop thinking about the war. No matter what kind of cool end they take, the temporary relief can never cure the pain for good. Destruction is the inevitable outcome of war; and war is the inevitable outcome of a cruel, senseless world. To sum up, the fictional world iconized by the metaphor is a place where people are permanently at the risk of war; one may die quick and early (like Catherine); or late and scarred (like Frederic and Rinaldi); and there is little one can do about it. The best thing that can happen in this world is a temporary relief, which can only postpone the pain, not abrogate it. The metaphor is iconic on two levels: on a macro level, the metaphor visualizes the narrator’s comprehension of the world; and on a micro level, the sequence of the wording, the twin-poles-no-escape process coincides with the overall structure of the novel and with the sequence of events happening to each major character. The diagrammatic icon (metaphor) here is used as an explanatory device – to explain something big and abstract into something specific and vivid. By means of metaphor, as Ivan Fónagy says, “the poet pretends to discover the external or internal object to which he assigns a new name at that very moment” (1999:â•›22). In our case, the burning-log is the new name given to the fictional world – the icon of the

 Xinxin Zhao

world, and the metaphor is the way by which the narrator Frederic uses his own words to figure out the world filtered through his eyes. By replacing the abstract world-view in the narrator’s head with this visible burning scene, the reader may literally see what the world is like.

3.â•… A burning world of war Contrary to most critics’ view of the two themes (love and war), A Farewell to Arms is only a war story; it is about joining in the war, fighting in the war, being wounded in the war, loving in the war, escaping from the war, and dying in the war. The war is no World War I or any specific war mentioned in the novel, but the burning permanent living condition of mankind. Human existence is a kind of war against the forces that aspire to destroy the individual. Just as Miss Ferguson says to Catherine, “Fight or Die. That’s what people do. They don’t marry” (108). In the context of the novel, the Â�burning log is an endless war with no escape, and the ants are people who are burned by the omnipresent war fire.

3.1â•… Universal war The opening of a novel usually provides the time and place in which the work is set. Besides, the main theme is always fixed, as Nobel laureate Gabriel García Márquez notes, In the first paragraph, you solve most of the problems with your book. The theme is defined, the style, the tone. At least in my case, the first paragraph is a kind of sample of what the rest of the book is going to be.  (cited in Kirszner & Mandell 2006:â•›381)

This is also the case in AFTA. However, the setting is exceptional in that no �specific time or place is suggested; instead, a fully detailed description is given to an �unspecific time and place. In this novel, it is the first three paragraphs that �summarize the theme: In the late summer of that year we lived in a house in a village that looked across the river and the plain to the mountains. In the bed of the river there were pebbles and boulders, dry and white in the sun, and the water was clear and swiftly moving and blue in the channels. Troops went by the house and down the road and the dust they raised powdered the leaves of the trees. The trunks of the trees too were dusty and the leaves fell early that year and we saw the troops marching along the road and the dust rising and leaves, stirred by the breeze, falling and the soldiers marching and afterward the road bare and white except for the leaves.



A burning world of war 

The plain was rich with crops; there were many orchards of fruit trees and beyond the plain the mountains were brown and bare. There was fighting in the mountains and at night we could see the flashes from the artillery. In the dark it was like summer lightning, but the nights were cool and there was not the feeling of a storm coming. Sometimes in the dark we heard the troops marching under the window and guns going past pulled by motor-tractors. There was much traffic at night and many mules on the roads with boxes of ammunition on each side of their pack-saddles and gray motor trucks that carried men, and other trucks with loads covered with canvas that moved slower in the traffic. There were big guns too that passed in the day drawn by tractors, the long barrels of the guns covered with green branches and green leafy branches and vines laid over the tractors. To the north we could look across a valley and see a forest of chestnut trees and behind it another mountain on this side of the river. There was fighting for that mountain too, but it was not successful, and in the fall when the rains came the leaves all fell from the chestnut trees and the branches were bare and the trunks black with rain. The vineyards were thin and bare-branched too and all the country wet and brown and dead with the autumn. There were mists over the river and clouds on the mountain and the trucks splashed mud on the road and the troops were muddy and wet in their capes; their rifles were wet and under their capes the two leather cartridge-boxes on the front of the belts, gray leather boxes heavy with the packs of clips of thin, long 6.5 mm. cartridges, bulged forward under the capes so that the men, passing on the road, marched as though they were six months gone with child. (Hemingway 1957:â•›3–4)

3.1.1â•… Being instead of fighting The dominant feature of this opening description is its highly repetitive syntactic Â�pattern. The following list presents the verbs included in the predicates of all the clauses: be: was/were (12) (33%). There be: There was/were (8) (22%). Intransitive verb [Vi]: lived, looked, went, fell, moved, look, came, fell (8) (22%). Transitive verb [Vt]: raised, powdered, saw, see, heard, see, carried, splashed (8). (22%). The dominant pattern is as follows: Subject + be (+ rest of predicate) + Â�prepositional phrase (e.g. “The plain was rich with crops” [3] my emphasis); sometimes the Â�prepositional phrases precede be (e.g. “beyond the plain the mountains were brown and bare” [3] my emphasis); or, they come both before and after (e.g. “In the bed of the river there were pebbles and boulders, dry and white in the sun” [3] my Â�emphasis ). Â�Including the be-clauses, Vi-clauses far outweigh Vt-clauses. Of the intransitive and transitive verbs, 44% of them are static verbs (namely, “lived”, “looked”, “saw”, “see”, “look”, “see”, “heard”).

 Xinxin Zhao

All of the above features are in accordance with a nominal style, which is defined by Lanham as “a style based on nouns, on stasis” (2004:â•›11). Lanham also notes that one of the noun strategies for suppressing action is that “the sentence’s forward motion sinks into the da da dum prepositional-phrase-string pattern” (13). The monotony of the prepositional-phrase strings makes the action disappear into the nouns. (For example, compare the original sentence “The plain was rich with crops” with a much more active version with similar meaning “Crops grew richly and covered the plain”). In this way, the nominal style presents the setting like a static world without much action. However, in the novel, a tension is produced between the form and the content. According to the content, fighting is pervasive, which is supposed to be full of action, but the syntactic pattern is marked by a lack of dynamic verbs. The conflict between form and content sends out a message: in this world, war is not fought, but it is. It seems that the conflict between form and meaning violates the form-mimingmeaning iconicity – one may expect a war scene to be iconized not by a nominal style, but by a style full of action. If, however, we take the narrator’s standpoint into consideration, the iconic effect becomes clear. The narrator, Frederic, is an American who has joined the Italian army. In addition, strictly speaking, he is not a soldier at all, but an ambulance driver. The war between Italy and Austria has little to do with him, and he hardly fights in any real combat. As a result, he is a bystander or outsider. The war in Italy is more like a scene before his eyes than an action he is involved in. In this way, the nominal style has a very strong iconic impact here – it represents not only the event but the narrator’s sense of the event. As Müller suggests, “it is just the deviation from the iconic norm which manifests iconicity most conspicuously” (2001:â•›319). This is iconicity, on a level different from miming external reality, it mimes internal reality. Throughout the novel, there is little face-to-face combat, but the fictional world is filled with the effects of war. And the story is narrated by Frederic, who seldom joins in any actual fighting and sees the war from a limited perspective. However, Frederic still has to find ways to survive the war. As the opening setting shows, this story is not about fighting in war, but living in the war – “we lived in” a world where war is a universal living condition.

3.1.2â•… Iconicity in landscape description Besides the syntactic structure, the setting of this novel is iconic in that the landscape description imitates the way in which the narrator ‘sees’ this world. As Wolf (2001:335) says, “landscape does not reside in their geographical exactness but mainly in that their verbal representations follow […] the structures of subjective visual perception”. In the setting of this novel, the textual focus and word order mimic the narrator’s Â�subjective visual perception. From the three paragraphs cited above, we can see that



A burning world of war 

visual sense is the major kind of perception the narrator adopts (“we saw the troops marching”, “we could see the flashes”, “we could look across a valley and see a forest” [3–4] my emphasis), and consequently, the narrator is our focalizer. Using the words of Abbott’s (2007:â•›66), “we […] see the action through (his) eyes”. Although the landscape description here takes on the external focalization, with the narrator/focalizer just giving an account of external happenings without any insight into thoughts or feelings, the recording of the scenery is still a subjective Â�experience. Because focalization (focus of narration) is indeed the selection of the information conveyed, whether by external or internal means, the focalizer records selectively what interests or impresses him most. According to Wolf (2001:â•›335), “[a] first element in the imitation of perception is the specificity and subjectivity of the gaze the reader is permitted to follow”. Based on the focalizer’s subjective selection, in our case, only two images are given a detailed description and permitted as the representation of the overview of the fictional world – nature image and war image. Interestingly, these two images appear alternatively throughout the whole setting. We see this in the first paragraph cited above: groups of words denoting nature and war show themselves in the following sequence: summer, river (×2), plain, mountains, pebbles, boulders, sun, water, channels ↓ troops ↓ dust, leaves (×2), trees (×2), trunk, year ↓ troops marching ↓ dust, leaves, breeze ↓ soldiers marching ↓ leaves The sequence reproduced with its original word order embodies one basic iconic Â� principle: contiguity in discourse mirrors contiguity in perception. War and nature coexist in this fictional world – troops marching are as natural as leaves falling. Besides, the alternation of the two groups of words is structured in a way that follows the eye movements of the focalizer, who shifts back and forth between nature imagery and war imagery: both war and nature are equal constitutive parts of this fictional world – war and nature intertwine through the narrator’s shifting focus. The first chapter of AFTA is short, but it could hardly be more significant. It Â�introduces the general setting of this novel: the narrator, unidentified so far, describes

 Xinxin Zhao

a war scene in an unnamed country during an unspecific period of time. The setting indeed sets the tone for the entire novel to follow. Through lexical choice (the verb ‘be’,), the universality of war is highlighted. The subjective focalization mirrors the position of the narrator: he is living in the war rather than actively fighting in it. The endless marching troops move on and on from summer to winter while the narrator remains stationary expressed by the nominal style. He is on the periphery, physically and emotionally. This story is not about fighting in a war as a soldier but living in a war as a human being.

3.2â•… War between ‘Us’ and ‘Them’ Apart from the setting in the first paragraphs, war is also pervasive in the rest of the book: there is war between Frederic and Catherine, between Frederic’s opposing selves (one self as a soldier and one as a civilian who is through with war), and finally between the couple and the world. When the war stops being fought between the couple or within Frederic himself, Frederic becomes one with Catherine, and they stand side by side fighting together against their ultimate enemy – ‘the world’. Besides being referred to as ‘the world’, this ultimate enemy also shows itself in the more ambiguous form of the third-person Â�singular (‘it’) or plural (‘they’). As Catherine says, “there’s only us two and in the world there’s all the rest of them. If anything comes between us we’re gone and then they have us” (139; my emphasis). In this war between ‘us’ and ‘them’, who represents ‘the world’/ ‘it’/ ‘they’?. In fact, throughout the novel, although we are in the world of war, we never get a clear impression of the enemy. In World War I, as described in the book, Germans and Austrians are the literal enemies. However, there is hardly any direct Â�description of them. Anything Frederic knows about them is just hearsay (e.g. “We heard that Germans and Austrians had broken through in the north and were Â�coming down the mountain valleys toward Cividale and Udine.” [188] my Â�emphasis). On the only chance of confronting them, he only sees their helmets (e.g. “Along the top of the stone bridge we could see German helmets moving.” [211]). The literal enemy is so vague, let alone the ultimate enemy. Furthermore, Germans and Austrians are not the only enemies in the world of AFTA. The characters have also to cope with the Italian battle police (They “execute officers of the rank of major and above who were separated from their troops” [224], and Frederic is almost killed by them), the Â�doctors (who send Frederic back to the front when his recovery is not Â�completed yet, which Rinaldi points out, saying that “It’s a crime to send you back. They ought to get complete articulation” [166]), biology (Catherine’s pregnancy makes the Â�couple feel “trapped biologically” [139] and “the end of the



A burning world of war 

trap” [320] is Â�Catherine’s death of childbirth), and the doomed fate. ‘The world’ Â�incorporates all of them. In a word, the ultimate enemy, or ‘the world’, is everything that serves to make human life shorter and less pleasant. Frederic flees from the army to avoid being killed and goes back to Catherine for a ‘separate peace’. All these endeavors are indeed against ‘the world’, which insists on breaking and killing everyone; it is as if ‘the world’ cannot bear to let anyone remain happy and safe and tries every possible means to intrude into their temporary happiness. In the following passage loose and tight cohesion between sentences is used to reflect iconically how ‘the world’ breaks the couple. It is taken from Chapter XXXIV, where Frederic and Catherine reunite after Frederic’s desertion. [1] That night at the hotel, in our room with the long empty hall outside and our shoes outside the door, a thick carpet on the floor of the room, outside the windows the rain falling and in the room light and pleasant and cheerful, then the light out and it exciting with smooth sheets and the bed comfortable, feeling that we had come home, feeling no longer alone, waking in the night to find the other one there, and not gone away; all other things were unreal. [2] We slept when we were tired and if we woke the other one woke too so one was not alone. [3] Often a man wishes to be alone and a girl wishes to be alone too and if they love each other they are jealous of that in each other, but I can truly say we never felt that. [4] We could feel alone when we were together, alone against the others. [5] It has only happened to me like that once. [6] I have been alone while I was with many girls and that is the way that you can be most lonely. [7] But we were never lonely and never afraid when we were together. [8] I know that the night is not the same as the day: that all things are different, that the things of the night cannot be explained in the day, because they do not then exist, and the night can be a dreadful time for lonely people once their loneliness has started. [9] But with Catherine there was almost no difference in the night except that it was an even better time. [10] If people bring so much courage to this world the world has to kill them to break them, so of course it kills them. [11] The world breaks every one and afterward many are strong at the broken places. [12] But those that will not break it kills. [13] It kills the very good and the very gentle and the very brave impartially. [14] If you are none of these you can be sure it will kill you too but there will be no special hurry. (Hemingway 1957:â•›249; my emphasis)

Cohesion, as defined by Hoey (2000:â•›266), “is a property of text whereby certain Â�grammatical or lexical features of the sentences of the text connect them to other Â�sentences in the text”. Halliday and Hasan (2001:â•›333–38) list almost all the types of cohesion. There are five main types, namely, reference (R), substitution (S), ellipsis (E), Â�conjunction (C), and lexical (L). The specific subtypes will not be listed here, but will be given in the following table right after the main types.

 Xinxin Zhao

Table 4.╇ Cohesive ties (S.1–S.9) Sentence No. No. of ties Cohesive ties

1 2

0 7

3

3

4

4

5

2

6

4

7

4

8

6

9

4

we (×3) slept woke the other one alone alone (×2) we we (×2) alone (×2) It that I (×2) alone lonely we (×2) lonely together I night (×3) lonely (loneliness) kill It kill with Catherine night it better

Type of cohesion

Presupposed item

R (Pronominal) L (collocation) L (same item) S (Nominal) L (same item) L (same item) R (Pronominal) R (Pronominal) L (same item) R (Pronominal) R (Demonstrative) R (Pronominal) L (same item) L (synonym) R (Pronominal) L (same item) L (same item) R (pronominal) L (same item) L (same item) L (same item) R (pronominal) L (same item) R (pronominal)

We sheets and bed waking the other one alone alone We We alone S. 4

L (same item) R (pronominal) R (comparative)

Night Night Day

I/me alone alone We Lonely together I night lonely Kill The world Kill We

Table 5.╇ Cohesive ties (S.10–S.14) Sentence No. No. of ties Cohesive ties Type of cohesion Presupposed item

10 11

0 4

12

4

the world break (broken) every one But Break It

L (same item) L (same item)

The world Break

R (pronominal) C (adversative) L (same item) R (pronominal)

people S. 11 Break The world

(Continued)



A burning world of war 

Table 5.╇ (Continued) Sentence No. No. of ties Cohesive ties Type of cohesion Presupposed item

13 14

2 3

kill

L (same item)

kill

It

R (pronominal)

The world

kill

L (same item)

kill

These

R (pronominal)

The very good, the very gentle, the very brave

It kill

R (pronominal) L (same item)

The world kill

From the tables, we can see that sentence [10] has no cohesive link with the Â� preceding sentences. Besides, starting from S.10, the cohesive ties in S.10–14 have nothing to do with the ones in S.1–9: before S.10, the cohesive ties are ‘we’, ‘alone’, ‘I’, etc. and after S.10, ‘the world’, ‘break’, and ‘kill’. According to these features, sentence 10, as a boundary line, divides this paragraph into two groups – Group A (S.1–S.9) and B (S.10–S. 14). Among each group, the sentences are tightly cohesive, whereas between the two groups, there is hardly any connection. In principle, according to Halliday and Hasan (2001:â•›297–298), “we shall expect to find a greater degree of cohesion within a paragraph than between paragraphs”. Â�However, in our case, when there is no cohesion, two distinct groups of sentences are put into the same paragraph. The question then is: why is there such a break in Â�cohesion within the paragraph? First of all, based on the cohesive ties, let us summarize the main idea for each group: Group A – only when we are together, we are never alone, and Group B – the world intends to break and kill everyone sooner or later. Judging from both the form and the content, the comment in Group B seems to come out of nowhere. In Group A, the couple enjoys their happy togetherness; suddenly, without any notice, ‘the world’ intrudes and tries to break and kill. As Halliday and Hasan (2001:â•›295) note, “if a sentence shows no cohesion with what has gone before, […] we might choose to regard such instances as discontinuities”. Iconically, with this break of cohesion, ‘the world’ breaks into the couple’s happiness without any reason or logic, which is not the first time in this novel. Throughout the book, the couple never have a chance to be settled for long: when Frederic starts getting emotionally involved with Catherine in Chapter VII, he has to leave her for the war in the next chapter and gets wounded; when he is seriously in love with her during his recovery in hospital, he has to go back to the front; when they finally achieve their ‘separate peace’ in Switzerland, Catherine dies in childbirth. ‘The world’ strives to break them up and kill ‘the good, gentle and brave one’ – Catherine. As reflected in the texture of this paragraph with less cohesion, whenever the ‘we’ are together ‘the world’ breaks into them, irrationally.

 Xinxin Zhao

Analyzing a text for cohesion, Halliday and Hasan (2001:â•›331) observe, “it is useful to note not only the type of tie […] but also the distance separating the presupposing from the presupposed”. Based on this distance, they divide the cohesive instances (or ties) into the ‘immediate’ – “presupposing an item in a contiguous sentence” and the ‘remote’ – “having one or more intervening sentences not involved in the presupposition” (339). When we look closely at the actual words that enter into cohesive ties, we find that two opposing sides, ‘we/I’ and ‘the world’, work respectively as the main ties for each group. However, the distance between the ties is different, as shown in the following two tables: Table 6.╇ Cohesive ties in Group A Sentence No. Cohesive tie (‘we’) Cohesive tie (‘I’) 1

we

0

2 3 4 5 6 7 8 9

we (×3) we we (×2) 0 0 we (×2) 0 with Catherine

0 I 0 Me I (×2) 0 I 0

Table 7.╇ Cohesive ties in Group B Sentence No. Cohesive tie (‘the world’) 10 11 12 13 14

The world/it The world It It It

Table 6 shows that in Group A the cohesive tie – ‘we’ could be coded as ‘remote’ in S.7 and S.9, since there are one or two sentences occurring in between; similarly in the case of ‘I’, which appears also with one or two intervening sentences. When we contrast these with Group B in Table 7, ‘the world’ is definitely an immediate tie, linking together without any interruption in every neighboring sentence. Besides, it is interesting to note that in Group A ‘we’ and ‘I’ emerge alternatively – whenever the string of ‘we’ is broken (in S.5, S.6 and S. 8), ‘I’ takes its place. Or to put it in another way, the link between ‘we’ is broken by ‘I’ over and over again. Furthermore, whenever



A burning world of war 

‘we’ intermits, the time line is broken too. In Group A, the ‘we’/‘I’ pronoun shift is Â�accompanied by past/present tense shift, as illustrated in Table 8; whereas in Group B, all the sentences are of the uniform present tense. Table 8.╇ Pronoun shift along with tense shift Sentence No. Cohesive tie (Pronoun) Tense 1 2 3 4 5 6 7 8 9

we we I we me I we I with Catherine

Past Past Present Past Present Present Past Present Past

Pronoun shift combined with tense shift sends out a message: the happy time when we were together and not alone is only the past, at present I am left on my own and feel lonely. Fónagy (1999:â•›18) observes that “[m]ovement in time – the irregular use of tenses always […] suggests evidential distance”. With the remote ties and the back-and-forth pronoun and tense shift, Group A is much less integrated than Group B – thus a Â�broken image is created here. Halliday (2002:â•›223–224) suggests, “[c]ohesion is a necessary but not a sufficient condition of coherence […], the mere presence of cohesive ties is not by itself a guarantee of a coherent texture”. A passage full of cohesive ties may still give a fragmented impression – the text could be cohesive but not necessarily coherent. Even with all their efforts to stick to each other, Frederic and Catherine can still not escape from having to part from each other. Meanwhile, in Group B, with the immediate ties and the unified tense, ‘the world’ is built up as a solid and indestructible enemy. Thus, in the war between the couple and ‘the world’, how could the breakable couple win and survive? This paragraph indeed foreshadows the coming tragedy. Furthermore, this paragraph is another instance of structural analogy with the burning-log metaphor. Through the switch between ‘I’ and ‘we’, Frederic moves between the burning end – the ‘I′-state (the lonely state without Catherine) and the cool end – the ‘we′-state. The temporariness of the cool end (we-state) is presented by the past tense, whereas the timelessness or permanence of the burning end (I-state) is reinforced by the present tense. At last, both Frederic and Catherine fall into the fire – the world breaks Frederic and kills Catherine. The present tense in this part shows the inevitability and factualness of the tragic ending awaiting everyone. The grim worldview of twin-poles-no-escape is once again embodied in an iconic paragraph.

 Xinxin Zhao

During their second meeting, Frederic suggests, “Let’s drop the war”, and Catherine answers, “It’s very hard. There’s no place to drop it” (1957:â•›26). Her insightful statement sets the underlying tone for this novel. In nearly every chapter, the war machine rolls on in the front line or in the background. As Passini says before his death, “It doesn’t finish. There is no finish to a war” (50). The best thing that can happen on the burning log is the transitory cool end. It cannot salvage the ants from the endless burning fire; however, sometimes it still makes life worth living, such as love, even if the war fire will lick up everything in the end, leaving everyone to die or to live alone.

4.â•… Conclusion Instead of a love story in war, A Farewell to Arms is a war story. Unlike many war stories, however, this novel does not oppose war nor glorify the experience of combat. It simply depicts the hostility and violence of a universe in which war is a permanent condition of existence and human existence is itself a war. The war between ‘us’ and ‘the world’ never ends and we never win (as Frederic says: “Perhaps wars weren’t won any more. Maybe they went on forever.” [118]). If there is anything to make living in such a world less painful, it is something that could create an illusion of “dropping the war” (26) even just for a little while. In their striving for transient happiness, Frederic and Catherine aspire to escape from war; an escape that is prohibited by ‘the world’, whose Â�punishment is death, just like the ants on the burning log. With the death of Catherine, ‘the world’ deprives each of them of the person loved. Contrary to the Â�idealization of love, as Â�Stolrzfus comments, “human love dies in the flesh” (2007:â•›123). Thus, the war, which is endless and could never be won, is the major theme of the novel. Any attempt to drop out of it is made in vain – everyone will be broken or killed in the end. Â�Hemingway himself wrote in his introduction to the 1948 edition of A Farewell to Arms: “The fact that the book was a tragic one did not make me unhappy since I believe that life was a tragedy and could only have one end” (qtd. in Spanier 2007:â•›92). In this paper I have illustrated how the study of iconicity can shed new light on even a much-studied text like Hemingway’s and how it can help us understand the ways in which a literary work, distant as it may be from us, affects us as readers. As Max Nänny (1997 and elsewhere) has convincingly shown, a writer like Hemingway was extremely conscious of his texts’ architecture. By investigating the iconic aspect of a literary text, i.e. the similarities between the linguistic or textural forms used and the meaning intended to express, the theme of a novel becomes visible, and the imaginary fictional world becomes tangible. When an author uses linguistic signs and narrative structures to mirror the idea in his head, the formless idea is concretized through formal iconic means. Through it the reader may recognize the formal features, each of the apparent clues left on the pages leading to a more general reflection, and thus



A burning world of war 

Â� contributing to the final effect and impression. From formless thought in the author’s head to the formal text, then to formless thought in the reader’s head, the distance between author and reader has never been so close. When a literary work is read by a non-native speaker, iconicity works especially effectively to reduce the distance between the reader and the work/author. Â�Without a cultural or historical common ground with the author, a non-native reader Â�inevitably has to face the gap lying between him/her and the work. Even native readers still have this problem when they read a work which is written long ago. If the reader can read a work only from a perspective limited by age and background, how can (s)he Â�interpret the work and get to author’s original idea as closely as possible? Iconicity could be the answer, because the similarity between the iconic signifiers and the signified is universally recognized. Concerning the iconic quantitative principle, the more form the more meaning can readily be identified. As for the iconic sequential principle, the linearity of language never changes in different times or countries. With this universal correspondence between form and meaning, the reading and interpreting of the text can bridge the temporal and cultural gap between the reader and the work. As Umberto Eco argues, “some texts are ‘open’ (Joyce’s Finnegans Wake, atonal music) and invite the reader’s collaboration in the production of meaning, while Â�others are ‘closed’ (comics, detective fiction) and predetermine the reader’s response” (qtd. in Selden 2004:â•›50). As far as I am concerned, no matter how ‘open’ or ‘closed’ the texts are, the author, through the form-miming-meaning devices, exercises Â�control over the way readers perceive the text. The difference between the ‘open’ and ‘closed’ texts lies in how much interpretive freedom the author entitles the reader to. In other words, it is the iconic nature of literary texts that allows a spectrum of possible Â�readings. On the one hand, the self-conscious form of a literary work should and could arouse the reader’s attention and make a certain impression on the reader; on the other hand, the reader’s creative and interpretative capacity is also necessary and desirable. The creation of meaning is a cooperative enterprise towards which, reader and author both contribute. If literary critics justify their interpretation by referring to the text and using the text-grounded iconic evidence, they will not readily be accused of Â�subjectivity in a profession that prizes objective methods.

References Abbott, H.P. 2007. The Cambridge Introduction to Narrative. Cambridge: CUP. Donaldson, S. (ed.). 2007. New Essays on A Farewell to Arms. Cambridge: CUP. Fischer O. & Ljungberg, C. 2010. Iconicity in Language and Literature. 〈http://www.iconicity.ch〉. Fischer O. & Nänny M. (eds). 2001. The Motivated Sign [Iconicity in Language and Literature 2]. Amsterdam: John Benjamins.

 Xinxin Zhao Fónagy, I. 1999. Why Iconicity? In Nänny & Fischer (eds), 3–36. Givón, T. 1979. On Understanding Grammar. New York NY: Academic Press. Haiman, J. 1985. Natural Syntax: Iconicity and Erosion. Cambridge: CUP. Halliday, M.A.K. 2002. On Grammar. Jonathan J. Webster (ed.). London: Continuum. Halliday, M.A.K. & Hasan, R. 2001 [1976]. Cohesion in English [English Grammar Series 9]. London: Longman. Hemingway, E. 1957. A Farewell to Arms. New York NY: Charles Scribner’s Sons. Hoey, M. 2000 [1991]. Patterns of Lexis in Text. Oxford: OUP. Kirszner, L.G. & Mandell S.R. 2006. Literature: Reading, Reacting, Writing, 5th edn. Beijing: Peking University Press. Lanham, R.A. 2004. Analyzing Prose. Beijing: Peking University Press. Müller, W.G. 1999. The Iconic Use of Syntax in British and American Fiction. In M. Nänny & O. Fischer (eds), 393–408. Müller, W.G. 2001. Iconicity and rhetoric: A note on the iconic force of rhetorical figures in Shakespeare. In Fischer & Nänny (eds), 306–22. Nänny, M. 1997. Hemingway’s architecture of prose. Chiastic patterns and their narrative Â�functions. North Dakota Quarterly 64(3): 157–176. Nänny M. & Fischer O. (eds). 1999. Form Miming Meaning. [Iconicity in Language and Â�Literature 1]. Amsterdam: John Benjamins. Nöth, W. 2001. Semiotic foundations of iconicity in language and literature. In Fischer & M. Nänny (eds), 17–28. de Saussure, F. 2006. Course in General Linguistics. H. Adams & L. Searle (eds). Talahassee: University Press of Florida. Sebeok, T.A. 2001. Signs: An Introduction to Semiotics, 2nd edn. Toronto: University of Toronto Press. Selden, R., Widdowson P. & Brooker, P. (eds). 2004. A Reader’s Guide to Contemporary Literary Theory, 5th edn. Harlow: Pearson Longman. Spanier, S.W. 2007. Hemingway’s unknown soldier: Catherine Barkley, the critics, and the great war. In Donaldson (ed.), 75–108. Stoltzfus, B. 2007. A sliding discourse: The language of A Farewell to Arms. In Donaldson (ed.), 109–36. Wolf, W. 2001. The emergence of experiential iconicity and spatial perspective in landscape description in English fiction. In Fischer & Nänny (eds), 323–48.

part iv

Cognitive poetics

Aesthetic qualities as structural resemblance Divergence and perceptual forces in poetry Reuven Tsur

Tel Aviv University When we say “The music is sad”, we report that we have detected some resemblance between the structure of the music and the structure of an emotion. In this sense, “sad” refers to an aesthetic quality of the music. In poetry, “sad” may refer either to the mere contents of the poem, or to an aesthetic quality arising from an interplay of divergent structure, low energy level, slow motion, sad contents. The paper explores such questions as “How do systems of music-sounds and verbal signs assume perceptual qualities endemic to other systems, such as human emotions or animal calls?” “What may a critic mean when asserting that a certain metric configuration is ‘more dignified’ than some other; that is, what may ‘dignified’ mean in a context of metric configurations?” The paper is focused on two structural phenomena found in both poems and emotions: “divergence” and “perceptual forces”.

1.â•… Introduction: Emotional qualities and onomatopoeia In this paper, I will explore two sets of related features of poetic structure which have run all through my work, from my first English publication to my latest articles: convergence-and-divergence, and perceptual forces.1 I will mention convergence for the sake of comparison only, and concentrate on divergence and perceptual forces. Traditional poetics has largely described the structures underlying these features, but had to resort to impressionistic means to point out their contribution to emotional qualities. Cognitive poetics, by contrast, is tailor-made to deal with that aspect of poetry in a principled manner. As to aesthetic qualities, the following example may illuminate their nature. The adjective “sad” has different meanings in the sentences “My sister is sad”, and “The music is sad”. In the former, it refers to the mental processes of a flesh-and-blood person. In the latter it does not refer to the mental processes of the sound sequence. Nor does it refer to the mental processes the music arouses in the

.╅ The sound files for this paper are available online: http://www.tau.ac.il/~tsurxx/Structural_ Resemblance/within_thine_own.html

 Reuven Tsur

listener. One can be perfectly consistent when saying “That sad piece of music inspired me with happiness”. It reports, rather, that the listener has detected some structural resemblance between the music and an emotion, such as a low energy level, slow movement, and a withdrawn, unassertive attitude suggested by the minor key.2 In this sense, “sad” refers to an aesthetic quality of the music. Cognitive poetics provides a conceptual system that allows to explore similar aesthetic qualities in specific instances of poetry. We will isolate two structural aspects of emotions: relative disorganization and deviation from a normal energy level. How do systems of music-sounds and verbal signs assume perceptual qualities endemic to other systems, such as human emotions or animal calls? At the present stage of my argument, I only want to point out that the resources available in the Â�target systems impose severe strictures on the process. Usually only very few Â�features or Â�configurations thereof are available in the target systems that may be shared with the source phenomena. So, the best one can do is to choose the nearest options Â�available in the target system. Minute differences may suffice to transform the Â�perceived Â�character of a complex whole. As Krueger (1968:â•›100–101) observed, the overall Â�perceived Â�qualities of “total complexes” is determined by minute differences: “It has been observed over and over that the smallest changes in experience are felt emotionally long before the change can be exactly described”.

i

a

u

ku

ku

Figure 1.╇ Wave plot, and the first and second formants of the cardinal vowels i-a-u, and of the European cuckoo’s call. (Formants are concentrations of overtones that determine vowels and sound colour). Note that the formants of the bird’s call are most similar to, but not identical with, the vowel [u] (produced on SoundScope)

.â•… For a cognitive explanation (in an evolutionary perspective) of the “withdrawn, unassertive” affective character of the minor key see, e.g. Cook & Fujisawa 2006:â•›9–16; Cook & Hayashi 2008:â•›318–319. As to the affective character of the major and minor modes, Cook and Hayashi (2008:â•›311) quote Jean-Philippe Rameau, the French composer and author of an influential book on harmony, who wrote in 1722: “‘The major mode is suitable for songs of mirth and rejoicing,’ sometimes ‘tempests and furies,’ […] as well as ‘grandeur and magnificence.’ The minor mode, on the other hand, is suitable for ‘sweetness or tenderness, plaints, and mournful songs.”



Aesthetic qualities as structural resemblance 

In onomatopoeia, the phonological system of a language cannot reproduce the actual sounds of, e.g. the cuckoo’s call: neither the minor-third interval, nor the sound quality, nor the abrupt onset. The bird says neither [k] nor [u]. The only thing one can do is to choose the speech sounds with the nearest formant structure (see Figure 1). A symphony orchestra, by contrast, can reproduce the minor-third interval, but not the formant structure of the call.

ku

ku

i

a

u

Figure 2.╇ Wave plot and pitch abstract of the European cuckoo’s call and of the cardinal vowels read by a professional reader

ku

ku

Figure 3.╇ Sound waves and pitch extract of the imitation of the cuckoo’s call in L. Mozart’s Toy Symphony (produced on Praat)

The nearest option to codify in human speech the abrupt onset of the call is the abrupt consonant [k] – all the other features of [k] are irrelevant. In the orchestra,

 Reuven Tsur

the abrupt onset is indicated more directly (see Figure 3). Thus, the voiceless plosive [k] is a bundle of perceptual features, a subset of which is frequently exploited by the context to suggest some abrupt metallic noises as “ticktack” or “click”; but in the case of “cuckoo” only the perceptual feature [+abrupt] is utilized. Thus, the same elements or configurations in a target system may serve as the “nearest option” for a wide range of source phenomena. In a recent paper, I discuss Milton’s verse line

(1) And sát as Princes, whom the supréme King w x w x w x w x w x and quote Milton’s 1809 editor, Henry J. Todd’s comment: “I conceive that Â�Milton also intended the last foot of [this] verse to be a spondee, as more dignified and Â�impressive than the accentuation […] of súpreme on the first syllable” (Todd 1970:â•›199). As a side issue, I raise the question of what “dignified” may mean with reference to metric structures. It suggests, I claim, that the listener may detect some structural resemblance between the consecutive heavy stresses and the outward manifestations of dignity in humans, such as weightiness, reserve of manner, and clearly-articulated gestures. “Weightiness” in a context of two consecutive stressed syllables encumbered by an unstressed syllable of a disyllable in a strong position suggests “massive”, “hard to deal with” or “demanding great effort”. In a context of dignified human behavior it suggests “of much importance or consequence”. As to “reserve”, a stressed syllable in a weak position (followed by another stressed Â�syllable) “holds back” the rhythmic movement of the line, whereas a dignified person “holds back”, “controls” his responses, the expression of his emotions or thoughts. As to “clearly articulated gestures” in poetry, I argue in that paper that the rhythmical performance of such constructs as a disyllable with its second, stressed, syllable in a weak position requires exceptionally clear articulation. According to the Â�foregoing analysis, then, within the limited resources of metric structures, Â�consecutive stresses with some additional difficulty are the nearest perceptual options for suggesting a general quality of muchness, slowness and Â�articulateness that can be individuated through the meaning of words as expressing dignity. But notice that there is no iconic relationship here between metric structure and Â�contents. Rather, the Â�contents individuates as dignified the generalized qualities suggested by the metric structure. Pope’s.

(2) And ten low words oft creep in one dull line

is not perceived as “dignified”; rather, the two verse lines exploit different potentials of slowness suggested by successive stressed syllables.3

.â•… As to the desirable potential of slowness see, e.g. Shakespeare’s Sonnet 94:



Aesthetic qualities as structural resemblance 

2.â•… Convergent and divergent styles The terms “convergent” and “divergent” are taken from optics, where they are applied to rays of light which meet or tend to meet in a focus, and to rays which continually depart from one another. Guilford (1970) adopted those terms in the phrases “convergent-thinking” and “divergent-thinking abilities”, referring by them to logical and creative thinking, respectively. One may add that emotional processes are typically more divergent than non-emotional mental processes. This suggests a spectrum: Â�logical thinking creative thinking emotional processes. Emotional and non-emotional processes do not constitute a rigid dichotomy, but a continuum. Â�Elizabeth Duffy states that. [t]here is no point on this continuum where a ‘non-emotional’ degree of disorganization of response changes suddenly to an ‘emotional’ degree of disorganization; and there is no point at which a ‘non-emotional’ conscious state changes suddenly to an ‘emotional’ one. These characteristics of experience and behavior show continuous variation rather than separation into hard and fast categories. (1968:â•›138)

The same holds true of the energy-level continuum. I borrowed Guilford’s terms (1970) to describe the structural resemblance between certain poetic structures on the one hand, and convergent and divergent mental processes on the other. Now notice that “disorganization” in divergent poetry is, still, severely constrained by regular meter. I propose to introduce the distinction “convergence vs. divergence” by comparing two passages in which other things are really equal, in fact, literally identical, where the only difference is the line division: (3) But wherefore thou alone? Wherefore with thee Came not all Hell broke loose? Is pain to them Less pain, less to be fled, or thou than they Less hardy to endure? Courageous Chief, The first in flight from pain, had’st thou alleg’d To thy deserted host this cause of flight, Thou surely had’st not come sole fugitive.  (Paradise Lost IV. 917–923) They that have power to hurt and will do none, That do not do the thing they most do show, Who, moving others, are themselves as stone, Unmoved, cold, and to temptation slow, They rightly do inherit heaven’s graces…

 Reuven Tsur

(4)

But wherefore thou alone? Wherefore with thee came not all Hell broke loose? Is pain to them less pain, less to be fled, Or thou than they less hardy to endure? Courageous Chief, the first in flight from pain, Had’st thou alleg’d to thy deserted host This cause of flight, thou surely had’st not come Sole fugitive.

Excerpt 3 consists of a series of “straddled lines”. These sentences run on from one line to another which, themselves, when isolated, form an iambic pentameter line. The run-on lines of Excerpt 3 are rearranged (by James Whaler [1956]) into endstopped lines in Excerpt 4. This rearrangement affects the perceived quality of the passage. Excerpt 3 is perceived as fluid, whereas Excerpt 4 as more stable. When the syntactic unit and the verse line coincide, they reinforce each other’s shape, yielding “strong gestalts”. When the syntactic unit is run on from one line to another, they blur each other, yielding “weak gestalts”. I have asked students “Is irony equally subtle in the two passages?” Some students could discern no significant difference. But the rest were in agreement that irony seems to be ‘somehow subtler’ in Excerpt 3. How can we explain this? Semantically and syntactically the two passages are identical. Gestalt psychologists have produced evidence that strong gestalts are typically perceived as rational, non-emotional, whereas weak gestalts typically display an emotional quality. A similar correlation emerges from findings of the Rorschach inkblot test. In Excerpt 3, the sentences run on from line to line, and the line boundaries intruding upon the sentences blur each other, weakening each other’s shape. Leonard B. Meyer, who applies gestalt theory to music, accounts for the association of weak and strong gestalts with emotional and intellectual qualities as follows. “Because good shape is intelligible in this sense, it creates a psychological atmosphere of certainty, security, and patent purpose, in which the listener feels a sense of Â�control and power as well as a sense of specific tendency and definite direction” (Meyer 1956:â•›160). Poor shapes generate an opposite atmosphere. We have noted, however, that the divergent structure in Excerpt 3 seems to affect not only emotional qualities, but irony too, rendering it subtler. Meyer’s formulation may account for this effect too, precisely because it refers to a general psychological atmosphere, rather than a specific attitude. The ironic attitude typically involves some kind of pretended ignorance. The “psychological atmosphere of patent purpose” inspired by the stronger gestalts in Excerpt 4 subverts, therefore, the tone of elusive ignorance in irony. Weak gestalts, divergent structures, may enhance, then, quite diverse attitudes. Rather than indicating an iconic



Aesthetic qualities as structural resemblance 

relationship between form and content, divergent structures generate a “a psychological atmosphere of uncertainty, lack of patent purpose and definite direction”, concreted by various kinds of contents in a variety of more specific attitudes.

3.â•… Perceptual forces (large scale) Our other term is “perceptual forces”. At the beginning of his book Art and Visual Â�Perception, Arnheim demonstrates “the hidden structure of a square” by placing a black cardboard disk in various positions on a white square.

â•…â•…

Figure 4.╇

Figure 5.╇

Thus he “maps out” regions of tension and balance. In Figure 4, the disk lies slightly off the centre. “In looking at the disk”, he says, “we may find that it does not merely occupy a certain place but exhibits restlessness. This restlessness may be experienced as a tendency of the disk to get away from where it is placed or, more specifically, as a pull in a particular direction – for example, toward the center” Â�(Arnheim 1967:â•›2). “Psychologically”, says Arnheim, “the pulls in the disk exist in the experience of any person who looks at it” (ibid.: 6). He continues: “There is no point in calling these forces ‘illusions’. They are no more illusory than colors, which are attributed to the objects themselves, although they are actually nothing but the reactions of the nervous system to light of particular wave lengths” (ibid.: 8). And he concludes: “The disk is most stably settled when its center coincides with the center of the square. In Figure 5 it may be seen as drawn toward the contour to the right. With changing distance this effect will weaken or even turn into its opposite” (ibid.: 3). Do perceptual forces exist in verbal structures as well? Fodor and his colleagues used this principle to test the psychological reality of constituent or phrase structure of sentences. The technique is based on the Gestalt assumption that a perceptual unit tends “to preserve its integrity by resisting interruptions” (Fodor & Bever 1965:â•›415). In the experiment, subjects listened to a sentence during which a click occurred, and immediately afterward were required to write down the sentence and indicate where

 Reuven Tsur

the click had occurred. If a phrase is a perceptual unit, subjects should tend to hear a click which occurred during a phrase as having occurred between the phrases. One of their sentences was “That he was happy was evident from the way he smiled”. This sentence has a major break between “happy” and “was”. A click was placed at various positions in this sentence. […] Each subject heard the sentence with only one click on it. Fodor and Bever found that subjects were most accurate in locating the click which occurred between the two major phrases of the sentence – i.e. between “happy” and “was” in the above example. Clicks occurring before this break tended to be displaced towards the right (i.e. into the break), and those occurring after the break towards the left (i.e. again into the break). A later experiment indicates that even where there can be no acoustic cues for a break, mere syntactic knowledge may evoke such perceptual forces (see Appendix). Fodor et al., however, overlooked one crucial point. If the intrusion occurs in the middle of a perceptual unit, it induces balance and stability; it is only when it occurs between the middle and the boundary that it induces perceptual forces. For our present purpose, these results have two important implications: first, that perceptual forces do exist in a linguistic environment; second, that perceptual forces in a linguistic environment are crucially influenced by the placement of the Â�intruding event relative to the boundary of the perceptual unit. In poetic prosody there is a Â�further complication. One cannot elicit perceptual forces with the help of some Â�extra-linguistic click. However, syntactic boundaries may intrude upon verse lines perceived as ‘wholes’; and line boundaries may intrude upon syntactic units. Here the exponents of both the intruding and the disrupted events are conveyed by the same noises, by the same words. In a verse line, a syntactic break at the caesura reinforces balance and Â�symmetry; the nearer to the end of the line, the more it presses toward the end for Â�completion. Consequently, our relief will be greater when the missing part is supplied. This may generate, in certain circumstances, a sharp, witty effect, turning the last string of Â�syllables into a “punch-phrase”, so to speak. In Excerpt 5, from Pope’s An Essay on Â�Criticism, there is little that can account for the wit of the second line, except the requiredness of the last word: (5) Some foreign writers, some our own despise, The ancients only, or the moderns, prize. This is a characteristic feature of Pope’s wit. In divergent poetry, the effect is much more sophisticated. A syntactic break near the line boundary exerts pressure toward the boundary; but the verse line is not end-stopped: the sentence runs on to the next line. In such cases a sense of sweeping movement may be generated. Consider, for instance, the following excerpt from Milton’s “On his blindness”, and note the Â�placement of the two tokens of “best”. (6) “God doth not need Either man’s work, nor his own gifts. Who best





Aesthetic qualities as structural resemblance 

Bear his mild yoke, they serve him best. His state Is kingly: thousands at his bidding speed, And post o’er land and ocean without rest; They also serve who only stand and wait.”

The sentence “Who best bear his mild yoke, they serve him best” could constitute an iambic pentameter line, and the repeated “best” would enhance the symmetry and stability of its segments. Had Milton divided this run-on line into 6+4 syllables, (7) Who best bear his mild yoke, They serve him best he would have generated a relatively mild divergent movement. As it stands, Â�straddled over two lines, beginning in the ninth position and ending in the eighth position of the next line, the repeated pair of words introduces asymmetry and great instability into the sequence. The nearer an intruding break to the middle of a perceptual unit, the more it enhances symmetry; the nearer to its boundary, the more it enhances asymmetry and instability. Here the straddled line begins near the line boundary and ends just before the next line boundary. The first token of “best” occurs at the line boundary which, in turn, intrudes upon the complex sentence near its beginning; the second token of “best” occurs at the sentence boundary which, in turn, intrudes upon the line near its end. Now compare this to the following excerpt from Coleridge’s “The Rime of the Ancient Mariner”: (8)

Farewell, farewell! but this I tell To thee, thou Wedding-Guest! He prayeth well, who loveth well Both man and bird and beast.



He prayeth best, who loveth best All things both great and small; For the dear God who loveth us, He made and loveth all.

While in Milton’s sonnet the two tokens of “best” disturb the balance and induce Â�fluidity, Coleridge’s repeated “well” and “best” generate stability: their two tokens occur at the precise middle and end of the line, generating a sharp, epigrammatic quality. The two syntactic units (ending with “well” or “best”) converge with the two half-lines; in Milton, they diverge. Furthermore, while in Coleridge the relative clause follows the main clause: “He prayeth best, who loveth best”, Milton inverts this order, so as to increase the predictive load of syntax, generating suspense: “Who best bear his mild yoke, they serve him best”. Milton’s poem conveys a theologically-laden inner struggle. The theological ideas assume an emotional character, the sententious tone of the dictum becomes affectionate, owing to the highly divergent structure of the

 Reuven Tsur

text, suffused with impetuous perceptual forces, generating a perceptual quality that bears a structural resemblance to powerful emotions. The adjective “mild” softens the Â�utterance both by its meaning, and by blurring the iambic metre. We have contrasted Milton’s divergent sentence-and-versification structure to a similar but convergent structure in “The Ancient Mariner”. We may, however, contrast it more immediately to the structure of the last line in the same poem: “They also serve who only stand and wait”. Here the elements of language and versification act in consonance to generate an atmosphere of stability: as in Coleridge, the relative clause comes last; this is the only case in this divergent poem in which a whole sentence entirely converges with the line; and one of the exceptional cases in Milton in which stressed syllables occur only in strong positions, and in all strong positions. Thus, the juxtaposition of diametrically different configurations of language and versification points up the contrast between them, generating a powerful sense of fluidity leading up to an intense sense of stable closure. We have discussed perceptual forces at some length, in visual perception, the psycholinguistic laboratory, and enjambment in poetry. In music Cooper and Meyer (1960) pointed out similar perceptual forces: a steeply rising pitch sequence or intensity sequence (crescendo) has a marked forward grouping effect (it leads, so to speak, Â�forward). The same phenomenon we find in speech, in rising intonation contours.

4.â•… Perceptual forces (minute scale) But the perceptual forces can be demonstrated at the sub-phonemic level too, in the alignment of intonation and syllable crest. At this level, the intruding event, the peak of the pitch contour, normally occurs in the middle of the syllabic crest, generating stability; in some instances, however, it occurs late in the vowel, or even on a sonorant after it; and sometimes it occurs earlier than the middle. I have found in my corpus of poetry-readings that late peaking generates an impetuous forward drive; in fact, the later the peaking, the more impetuous is the forward drive. An early peak effects backward grouping and stability. Let us observe peak delay in three recordings of line 7 from Shakespeare’s second Sonnet: (9) To say within thine ówn déep-súnken éyes Let us listen to the Marlowe Society’s reading of this line (1959). We will focus here only on certain aspects of the words “say”, “within”, and “own”. The second syllable of “within” (being part of a function word) is perceived as unduly prominent, cued by stress, Â�rising intonation and late peaking. Apparently, metric regularity in the first six Â�syllables “To say within thine own” is straightforward enough, and there would appear to be no Â�reason for such extravagant devices. But the reciter has a real problem



Aesthetic qualities as structural resemblance 

here: the Â�caesura occurs in the middle of a prepositional phrase. One must indicate an intruding event after “within”, without disrupting the phrase. The over-articulation of “within”, coupled with an undue stress, without a pause, seems to have here one purpose: to indicate a caesura, without stopping. Â�Listening again to the line confirms that after both words (“within” and “say”) there is an impetuous drive across a discontinuity. In the case of “say” it is the contents that compels the reciter to separate the reporting phrase from the reported speech; but if he wants to preserve the line’s perceptual integrity, he must preserve the first four syllables, up to the caesura, as one unit. There is no pause after “say” or after “within”. In both instances, discontinuity is indicated by an exceptionally long Â�word-final sonorant, /j/ and /n/ respectively, and a drastic change of the direction of the pitch contour. In “say” there is a late peak on the second sound of the diphthong. In the second syllable of “within” there is a double peak, one occurring at the end of /i/, the other on /n/. This late peak bestows extreme prominence on “within”, emphasizing the strong position, but also displays an impetuous forward push. On “own” in Â�position 6 there is an additional late peak: it serves the need to group the word forward with the next two words, so as to begin the string of stressed syllables in a strong position.

to

s

a

y

with

i

n

th I ne

o w

n

Figure 6.╇ Wave plot and pitch extract of “to say within thine own” in the Marlowe Society’s reading. The markers indicate diphthong and vowel boundaries. Notice the late peaks on say, -thin and own

Admittedly, the pitch movement the Marlowe Society assigns to the sequence “to say within” is exceptional: these pitch movements cannot be predicted from spoken English prosody, or from any possible metric deviation; it is, indeed justified solely

 Reuven Tsur

by the evasive problem of caesura and line integrity. Notwithstanding this, Callow (1999) has recourse to similar intonation contours. He seems to identify exactly the same Â�problems in the first six syllables of the line; and offers exactly the same kind of Â�solution. Here too there is a conspicuous late peak on the second syllable of “within”, and the word-final /n/ is exceptionally long. Figure 8 reveals a rather high and Â�exceptionally late peak on “say”.

to s

ay

wi th i

n

thine

ow

n

Figure 7.╇ Wave plot and F0 extract of “to say within thine own” in Callow’s reading. The pairs of markers indicate vowel boundaries. Notice the late peak on -thin

s

a

y

Figure 8.╇ Wave plot and pitch extract of “say” in Callow’s reading (produced on Praat)

Listening to Gielgud’s performance (2006) indicates a conception that is rather Â�similar to that of the other two reciters, but with considerably different emphases. Pitch resets very high on “say”, with a late peak on /j/ of the diphthong, imposing a



Aesthetic qualities as structural resemblance 

� forward �impetus to the whole line. This drives across an enhanced break indicated both by a pitch �discontinuity and a straightforward 100-msec pause.

to s

ay

with i

n thine ow n d ee p sunk en eyes

Figure 9.╇ Wave plot and F0 extract of “to say within thine own deep-sunken eyes” in Â�Gielgud’s reading. The pairs of markers indicate vowel boundaries. Notice the late peaks on -thin and deep

In Gielgud’s performance (2006), the peak on the stressed /i/ of “within” is rather Â� moderate; but it occurs very late in the vowel, and there is an additional peak on /n/. The two syllables /(∂)in/ and /(∂)ajn/ belong to two consecutive function words. They happen to be very similar, but of different duration: the sequence /–(∂)in/ is 239 msec long, of which /n/ takes 148 msec; whereas /(∂)ajn/ is only 155 msec long, of which /n/ takes 67 msec. This relatively long duration contributes to the perception of great stress on “within”, also signaling conspicuous discontinuity. This combination of cues indicates a caesura after “within” and, at the same time, an impetuous drive across it. A similar story can be told of “own” in the sixth position. Late peaking is a rare, relatively little-understood phenomenon. Gerry Knowles (1992) notes that in ordinary speech it usually occurs in the middle of tone groups after a pause. As to its function, Robert Ladd says: “peak delay is said to signal that the utterance is in some way very significant or non-routine” (Ladd 1996:â•›99). The more remarkable it is that in this line three leading British actors have recourse to it several times at the same places in the line, most notably on a preposition, at places where no pauses precede them, but some forward thrust is called for, for rhythmic reasons. These actors utilize, then, a kind of vocal manipulation available in language for semantic emphasis – to cope with rhythmic complexities.

 Reuven Tsur

5.â•… Materials and structures I have proposed an approach to emotional qualities in poetry that closely Â�resembles iconicity. It does not, however, pursue an iconic relationship between form and Â�content. The form-and-content approach allows the critic to handle only those instances in which the similarity between form and content exists, or else compels him to read the similarity into them. The present approach replaces this Â�dichotomy by the Â�materials-and-structures dichotomy proposed by Wellek and Warren (1956). It regards both the contents and the formal elements of versification as aesthetically neutral materials that can be combined into aesthetic structures. According to the Wellek and Warren model a wide range of elements (which are independent Â�variables) may occur in any combination, and thus the tools offered here may serve to describe any unforeseen combination of elements in a poem. Unforeseen Â�combinations may Â�display unforeseen gestalt qualities, and cognitive poetics may systematically account for them. When we say “The music is sad”, we refer to an aesthetic quality of the music. When we say “The poem is sad”, we may refer either to the mere contents of the poem, or to an aesthetic quality arising from a configuration of divergent structure, low energy level, slow motion, sad contents. Contents, “projected world” word meanings, phonetic structure, metaphor, meter, rhyme, alliteration, are all materials. Structures are their various combinations. Poetic effects arise from the subtle interaction of a great variety of materials. The sequence of stressed and unstressed syllables may converge with or diverge from the sequence of strong and weak metric positions; syntactic units may coincide with verse lines, or may run on from one line to another; alliteration may work in conjunction with, or against, meter; and so forth. Briefly, they may act in convergence reinforcing each other, yielding exceptionally strong gestalts, sometimes with a pervasive witty Â�quality, sometimes suggesting simplified mastery of reality, as in nursery rhymes. Or they may act in divergence blurring each other, so as to yield an exceptionally weak gestalt with a pervasive emotional or subtle ironic quality. Such divergence may be reinforced by abstract nouns in a landscape defined here and now. “Hypnotic” poetry typically involves exceptionally regular meter–stress mappings, end-stopped lines but unpredictable groupings of lines, alliterations that work both in conjunction with and against meter, frequent repetition of key phrases, high energy level, the irruption of the irrational in the world stratum; and so forth. Finally, the various configurations need not necessarily comprise homogeneous elements. Consider again “Who best bear his mild yoke, they serve him best”. The strained enjambment, the relative clause preceding the main clause, the Â�consecutive stressed syllables “mild yoke” and “best bear” with the alliteration in adjacent stressed syllables, induce a sense of fluidity and uncertainty. At the same time, the superlative “best”, its symmetrical repetition at the two extremes, and the



Aesthetic qualities as structural resemblance 

epigrammatic Â�formulation, all inspire an atmosphere of certainty and stability. Such opposing effects need not generate a conflict between fluidity and stability or Â�mitigate each other. On the contrary, where the powerful drift is established as dominant, the robust Â�stabilizing elements may be perceived as vigorous intrusions at the “wrong” places, enhancing fluidity rather than stability. This divergent, fluid structure has, in Coleridge’s terms, the unpredictability of life and nature, of a “feeling profound and vehement” which, at the same time, is brought “under the irremissive, though gentle and unnoticed, control of will and understanding” – effected by the closure, after the event. But notice this: we do not perceive an analogy between the verbal structure and emotional processes; rather, we perceive the verbal structure as displaying an emotional quality. There are no rules to infer the aesthetic qualities emerging from configurations of aesthetically neutral elements. According to Frank Sibley (1962), we decide that a piece of music is sad by listening or that a piece of poetry is dignified by reading, just as we decide that the book is red by looking or that the tea is sweet by tasting. According to the present conception, disagreement whether a piece of convergent poetry is hypnotic, witty, playful, monotonous, cheerful or suggests simplified mastery of reality may be due to different mental organizations of the same aesthetically neutral elements.

Appendix Observations on Larsen’s criticism of the click experiment Larsen (1971) criticised Fodor and Bever’s and related experiments, pointing out that due to insufficient precision of the postulate, results obtained by the click Â�technique seem uninterpretable at present. He pointed out two flaws in their Â�procedure. First, two crucial categories of responses were excluded from their data analyses: responses to clicks objectively located in the major syntactic break itself; and Â�correct subjective locations of clicks objectively located outside the major break. Second, the notion of “constituent” is rather fuzzy. Every word boundary is at the same time a constituent boundary – though at different levels in the Â�hierarchy of constituents. The simplest prediction, therefore, on the assumption that Â�perceptual units resist interruption would be: Noise heard during speech should tend to be located Â�perceptually between words rather than within words. Larsen’s experiment refutes this hypothesis. I agree that the exclusions mentioned by Larsen are hard to understand. I also agree that the notion of “constituent” is rather fuzzy (I have written on this, back in 1973). But it is clear from Fodor et al.’s examples that they are not dealing with just any syntactic constituent, but with constituents delimited by major syntactic Â�boundaries. Thus, for instance, Garret, Bever and Fodor (1966) recorded pairs of sentences such as:

 Reuven Tsur

1. As a result of their invention’s influence the company was given an award. * was given an award. 2. The chairman whose methods still influence *the company * * When subjects were asked where they hear the longest pause in the sentences, they report – as one might expect – that they hear a pause in (1) between “influence” and “the”, and in (2) between “company” and “was”. The perceived pause thus corresponds to the major constituent boundaries in the two sentences. Then, the two italicized segments were interchanged. “Subjects’ perception of pause location, however, was unchanged. The same was true of click displacement. As indicated by asterisks in the two sentences above, a click occurred either during “company” or “was”. The perception of click location, however, was significantly different for the two sentences. The click in sentence (1) tended to be heard between “influence” and “the”, and in (2) between “company” and “was”. But remember the sentences were acoustically identical” (Slobin 1971:â•›25–26; italics in original). At any rate, click displacement occurs not even toward the boundary between NP and VP, but toward “major syntactic boundaries” – even when the same acoustically identical words are used. According to the predictions of gestalt theory, the stronger a boundary, the greater its effect on perceptual phenomena. Thus, we should not expect word boundaries and clause boundaries have the same effect on an intruding click, even though both “words” and “clauses” can be labeled “constituents”. Anyway, Larsen’s criticism does not affect my argument for two reasons. First, both Fodor et al. and Larsen overlook a crucial point. Perceptual forces are expected to be caused by intrusion not anywhere inside a unit, but between the middle and the boundary of a unit. Intrusion in the middle is expected to reinforce rather than upset stability. Thus, Larsen’s finding that clicks tend to be perceived in midword rather than between words does not necessarily jeopardize my position. Second, my main concern here is not the psychological reality of linguistic units, but two more Â�immediate phenomena: whether perceptual forces do exist in a linguistic environment; and whether they tend to push toward the nearest boundary when dislocated from the middle. In this respect, Larsen’s experiment replicated Fodor et al.’s findings. Larsen concludes that in view of his findings two alternatives and less arbitrary decisions are possible: (1) to reject the hypothesis that linguistic constituents are perceptual units; or (2) to reject the assumption that perceptual units resist interruptions. In view of the importance attributed here to the point in the middle, a third possibility is well worth experimenting too: that an intruding event reinforces perceptual units in their middle, but generate a “push” when they occur between their middle and their boundaries. Thus, for instance, one may expect that a major syntactic boundary reinforces the stability of a verse line when it falls at the caesura, but generates fluidity when it occurs between the caesura and the line boundary. This appears to apply to the alignment of intonation peak and



Aesthetic qualities as structural resemblance 

syllabic crest as well. On one point in this respect all researchers appear to agree: that an intonation peak hitting the middle of a syllabic crest is the unmarked option, whereas late and early peaking are marked. Moreover, I have predicted and empirically found that if you add two syllables at the end of an iambic pentameter line with a caesura after the fourth position, there is a strong tendency to move the caesura to the right, after the sixth position—if the prevailing conditions permit (and sometimes even if they don’t). Consider: (10) Invoke thy aid to my advent’rous song.

(Paradise Lost I. 13)

In this line, the caesura occurs after aid in position 4. Suppose however, that we add two more syllables to the verse line, turning it into an iambic hexameter, thus: (11) Invoke thy aid to my advent’rous song of praise. If one continues to observe a caesura after aid, Excerpt 11 is liable to fall apart. Here the caesura, in harmony with the perceptual needs of the iambic hexameter, is Â�automatically shifted to after my in position 6, even though this happens in Â�mid-phrase. Here a feeling is generated of a caesura as well as a “sense of impulsion across the [Â�non-existent] break”. At any rate, Larsen gives no information where exactly in mid-word do the Â�dislocated clicks occur in his experiment; and neither Larsen, nor Fodor et al. Â�indicate whether correct subjective locations of clicks do or do not occur at the middle of a larger syntactic constituent (e.g. clause). This problem is aggravated by another, Â�methodological difficulty: while it is relatively easy to determine where is the Â�boundary of a clause or the middle of a word or a verse line, we have no criteria to determine where is the perceptual middle of a clause. Thus, neither Fodor et al.’s nor Larsen’s click experiments may explain why some clicks are correctly located in mid-clause and some not, nor may, perhaps, prove that syntactic constituents have Â�psychological Â�reality; but they strongly suggest that when clicks are subjectively displaced in Â�mid-clause, they move in the expected direction, and seem to corroborate my assumptions. I said “seems to”, because to get more compelling results one should re-devise the whole click experiment in view of the importance attributed to the middle point.

References Arnheim, R. 1967. Art and Visual Perception. London: Faber & Faber. Coleridge, S.T. 1951. Selected Poetry and Prose, D.A. Stauffer (ed.), Â�New York NY: The Modern Library. Cook, N.D. & Fujisawa, T.X. 2006. The psychophysics of harmony perception: Harmony is a three-tone phenomenon. Empirical Musicology Review 1(2): 1–21.

 Reuven Tsur Cook, N.D. & Hayashi, T. 2008. The psychoacoustics of harmony perception. American Scientist 96: 311–319. Cooper, C.W. & Meyer, L.B. 1960. The Rhythmic Structure of Music. Chicago IL: Chicago Â�University Press. Duffy, E. 1968[1941]. An explanation of emotional phenomena without the use of the concept emotion. In The Nature of Emotion, M.B. Arnold (ed.), 129–140. Harmondsworth: Penguin. Fodor, J.A. & Bever, T. 1965. The psychological reality of linguistic segments. Journal of Verbal Learning and Verbal Behavior 4: 414–420. Garrett, M., Bever, T. & Fodor, J.A. 1966. The active use of grammar in speech perception. Â�Perception and Psychophysics 1: 30–32. Guilford, J.P. 1970[1959]. Traits of creativity. In Creativity, P.E. Vernon (ed.), 167–188. Â�Harmondsworth: Penguin. Knowles, G. 1992. Pitch contours and tones in the Lancaster/IBM spoken English Corpus. In New Directions in English Language Corpora – Methodology, Results, Software Â�Developments, G. Leitner (ed.), 289–299. Berlin: Mouton de Gruyter. Krueger, F. 1968. [1928]. The essence of feeling. In The Nature of Emotion, M.B. Arnold (ed.), 97–108. Harmondsworth: Penguin. Ladd, R.D. 1996. Intonational Phonology. Cambridge: CUP. Larsen, S.F. 1971. The psychological reality of linguistic segments reconsidered. Scandinavian Journal of Psychology 12(1): 113–118. Meyer, L.B. 1956. Emotion and Meaning in Music. Chicago IL: Chicago University Press. Sibley, F. 1962. Aesthetic qualities. In Philosophy Looks at the Arts: Contemporary Readings in Aesthetics, J. Margolis (ed.), 63–88. New York NY: Scribner. Slobin, D.I. 1971. Psycholinguistics. Glenview IL: Scott, Foreman & Co. Todd, H.J. 1809. Dr. Johnson’s remarks on Milton’s versification, with remarks by the editor. In The poetical Works of John Milton, with Notes of Various Authors. H.J. Todd (ed.), t.2, 154–209. London: Law & Gilbert. Wellek, R. & Warren, A. 1956. Theory of Literature. New York NY: Harcourt, Brace & Co. Whaler, J. 1956. Counterpoint and Symbol: An Inquiry into the Rhythm of Milton’s Epic Style [Anglistica 6]. Copenhagen: Rosenkilde. Recorded Readings Callow, Simon (1999). reading Shakespeare’s Sonnets. Hodder Headline AudioBooks HH 185. Gielgud, Sir John (2006). reading Sonnets of William Shakespeare. Caedmon SRS 241 C–D. The Marlowe Society and Professional Players (1959). reading Shakespeare: The Sonnets. Argo ZPR 254.

Audio processors SoundScope 16/3.0 (ppd) Praat 5.0.43

Mental space mapping in classical Chinese poetry A cognitive approach Han-liang Chang

Fudan University and National Taiwan University Primary verbal composite modelling, as manifested in cognitive poetics, raises serious theoretical questions over the nature and function of the linguistic sign. This chapter attempts to assess Chinese poetics popularized by North American sinologists in the 1960s to 1980s, re-read the so-called ‘Old Style’ Chinese poetry produced before the sixth century, and, finally, investigate why classical poetry in general lacks figurative and imagistic intricacy, characteristic of highly Â�conceit-laden Western poetry, e.g. in the metaphysical and modernistic traditions. Specifically, the essay analyzes the ways in which mental spaces in classical Chinese poetry are mapped and examines how vital relations, scales, force-dynamics, and image-schemata, are integrated or ‘blended’ in creating mediated poetic ‘space’. Through close readings of sample poems in terms of current Language and Space studies, the author argues that the commonly assumed iconicity in classical Chinese poetry should be more properly called poetic indexicality rather than iconicity.

1.â•… Introduction Primary verbal composite modelling, as discussed by cognitive poetics, raises Â�serious theoretical questions about the nature and function of the linguistic sign Â�(Stockwell 2002). If language as a conventional symbolic system qualifies as the Peircean Â�Thirdness, it cannot possibly be at the same time a Firstness. Rather than engaging in Â�endless debate on the naturalness or arbitrariness of the linguistic sign, and rather than compromising with a ‘both/and’ solution, I have chosen to delve into the empirical reading of classical Chinese poetry, mainly that produced before and during the Tang Dynasty of the seventh and eighth centuries. The special use of imagistic language in Chinese poetry, quite differently from its various English counterparts, has evoked heated debate on poetic iconicity over the past four decades. The debate has been complicated by the supposedly ‘ideographic’ (and popularly but wrongly held ‘iconographic’) nature of the Chinese writing system,

 Han-liang Chang

which has remained virtually unchanged since the second century when the script was codified. A classic statement on poetic diction from the 1970s reads as follows: “Chinese nouns are close approximations of universals” (Kao & Mei 1971:â•›104). The underlying assumption is a kind of simplistic iconicity existing between substantives – “unadorned archetypal nouns” – and natural phenomena (Kao & Mei 1971:â•›81). The quoted statement was made, ironically, during the heyday of structural Â�linguistics and poetics, when they were belatedly introduced and applied to the study of Chinese poetry (Jakobson 1966; Kao & Mei 1971, 1978; Guillén 1971–72; Cheng 1982 [French original 1977]). At that time, such claims about lexical and syntactical iconicity, when introduced into the Chinese speaking world, were warmly received by traditional interpreters, who, harbouring a similar vision of mimesis, found the idea congenial to their favourite shi hua (i.e. ‘poetic talks’ or critical fragments), as Â�evidenced by such expressions as qing jing jiao rong (“emotion and scenery Â�convergence”), jing jie (“poetic boundary”), etc. The irony lies in the fact that, in the 1960s and 1970s, structural poetics based on the Saussurian linguistic model had not caught on and had never taken root, probably due to traditional literary scholars’ general lack of Â�training and interest in linguistic analysis and of their suspicion of linguistics-informed Â�poetics, especially when it is imported from the West. The only exceptions are probably Kao and Mei (1978), but their application remains largely eclectic, marred by the burdens of traditional poetic theories. Curiously, the next paradigm, cognitive linguistics, has rarely been appropriated to deal with classical poetry either, partly because it takes to commonalities that Â�operate across all kinds of discourses, and partly due to researchers’ predominant interest in the more immediate contextuality of language cognition and its concern with Â�“common operations in everyday life” (Fauconnier & Turner 2002:â•›17). However, perhaps lacking the refined sophistication of textual analyses prevalent in a former generation, the cognitive approach, with its focus on human conceptualization and language cognition, however, may have something to say on the reading of Chinese poetry. For instance, the renewed interest in categories and prototypes, and their novel articulation, may shed light on both the semantic and pragmatic aspects of poetry, and may therefore provide theoretical input on the traditional concept of iconicity. And the kind of poetic ‘space’ (e.g. “scenery” and “boundary” included), whose ‘iconicity’ has appealed to professional exegetes and common readers, when examined in the light of contemporary thinking on spatial cognition and its language representation, may turn out to be a misnomer and will have to be re-conceptualized and re-articulated (Landau & Lakusta 2006; Levinson 2003; Levinson & Wilkins 2006). One could re-read classical Chinese poetry, especially that which deals with the prototypes of time and space, in terms of cognitive ‘commonalities’, and rethink the general lack of figurative and imagistic intricacy of classical Chinese poetry – especially when compared with the highly conceit-laden English poetry of the Â�metaphysical and



Mental space mapping in classical Chinese poetry 

modernist traditions. Specifically, it would be interesting to analyze the ways in which mental spaces in Chinese poetry are mapped, for example, how vital relations, scales, force-dynamics, and image-schemata are integrated or ‘blended’ in creating Â�mediated poetic ‘space’ (Fauconnier 1997). Through close reading of sample poems which are noted for their representations of ‘space’, in terms of current Language and Space Â�studies (Jarvella & Klein 1982; Svorou 1994; Bloom, Peterson, Nadel & Garrett 1996; Pütz & Dirven 1996; van der Zee & Slack 2003; Levinson 2003; Levinson & Wilkins 2006; Hickmann & Robert 2006), I hope to show that the commonly assumed iconicity in classical Chinese poetry, to return to the Peircean terms used above, should be more properly called poetic indexicality.

2.â•… Classical Chinese poetry and its American mediators Accordingly, this chapter will be divided into two parts. This section provides some background information about Chinese poetry studies over the past four decades, and by so doing exposes some of their shortcomings, whereas the second part offers an application of current Language and Space studies to the so-called gu ti shi (Ancient Style poetry). For readers who are not familiar with them, a short history of Chinese poetic traditions seems necessary, but for technical reasons, including space limits, it cannot be given here. Readers are referred to writings by Liu (1962), Yip (1969), Kao and Mei (1971, 1978), Owen (1985), and Yu (1987) in the References section. I will however provide a brief account of the scholarship by North-American sinologists in the 1960s and 70s, with particular reference to their conceptualizations of Chinese poetic semantics, mostly under the misnomer of concrete-universal. I have chosen these scholars because they are generally regarded as academic celebrities, who served as mediators between Chinese poetry and its English readership, at a time when Roman Jakobson was experimenting with his analytical method, especially on poetic parallelism (Jakobson 1966). The earliest example in this line of ‘modernist’ thinking is arguably the late James J. Y. Liu of Stanford University. In his Art of Chinese Poetry (1962), Liu takes note of the lack of inflection of classical Chinese: This is at once a source of strength and of weakness, for on the one hand it enables the writer to concentrate on essentials and be as concise as possible, while on the other hand it leads easily to ambiguity. In other words, where Chinese gains in conciseness, it loses in preciseness. As far as poetry is concerned, the gain is on the whole greater than the loss, for, as Aristotle observed, the poet is concerned with the universal rather than the particular, and the Chinese poet especially is often concerned with presenting the essence of a mood or a scene rather than with accidental details. (Liu 1962:â•›40; our emphasis)

 Han-liang Chang

The example Liu gives is from the eighth-century Wang Wei (701–761). As the Â�modern interpreter sees it, in the lines: (1)

yue chu jing shan niao Moon rise surprise mountain bird shi ming chun jian zhong Occasionally cry in spring valley 

(quoted in Liu 1962:â•›40)

it is of no consequence whether “mountain”, “bird”, and “valley’” are singular or plural. According to Liu, one can translate these lines as: “The moonrise surprises the mountain bird/That cries now and again in the spring valley”, or “The moonrise surprises the mountain birds/That cry now and again in the spring valley (or valleys)” without changing the meaning (Liu 1962:â•›40). Liu asserts that this sense of timelessness and universality created by lack of inflection is further enhanced by the frequent omission of the subject in Chinese poetry (Liu 1962:â•›40). How these grammatical features can be said to contribute to Chinese poetry’s “impersonal and universal quality” remains open to debate (Liu 1962:â•›41). However, Liu’s curious argument, also lacking linguistic grounding, has turned out to be quite popular among Chinese scholars mediating their own source language and a Euro-American academic community using an alien target language, be it English or French. A little later than Liu, another promulgator of a distinctively Chinese poetics based on non-inflective language is Wai-lim Yip of the University of California at San Diego, who had published his Princeton doctoral thesis on Ezra Pound’s rewriting of Cathay in 1969. Quite similar to Liu, albeit without being aware of the fact, Yip (1969) believes that Chinese poetry has “a special mode of representation of reality constituted or made possible by the peculiarity of the Chinese language itself.” (Yip 1969:â•›12). Â�Commenting on various English renditions of Li Po’s (701–762) “Taking leave of a friend”, Yip observes that “[i]n the original, or in the translations that observe the original structure, we see things in nature” whilst in some Westernized versions, “[w]e see the process of analysis at work rather than the things acting themselves out before us” (Yip 1969:â•›16). To Yip, the grammaticalization of Chinese, modelled on a Western language, such as English, is a “syntactical commitment” (Yip 1969:â•›19), it shows “analysis at work” and “the logic of succession” – all of which destroy “the drama of things” (ibid.: 19), “unanalytical presentation” (ibid.: 20), “simultaneous presence of two objects” (ibid.: 22), and “objects in their purest form uncontaminated by intellect or subjectivity” (ibid.: 25). Yip concludes that classical Chinese verses, as represented by (2)

ji sheng mao dian yue cock crow/thatch inn/moon ren ji ban qiao shang man trace/wood bridge/frost

(quoted in Yip 1969:â•›25)



Mental space mapping in classical Chinese poetry 

present a special poetic vision. “[L]ike the shots in the movies and the montage Â�technique, [they] have touched upon the realms of painting and sculpture, although, unlike the movies, the objects are projected only on the screen of imagination, not Â�literally before our eyes” (Yip 1969:â•›26). This is what Yip famously describes as the unmediated pure experience. Throughout his writings, Yip has untiringly pushed the same argument that de-syntactization represents a primordial mode of Â�cognition unknown to Western poeticians. This unique lexical philosophy against the grain of language and human cognition, e.g. temporality, has provoked a Â�torrent of reactions from various fronts. In this way, Yip can be said to have ironically Â�forestalled even Â�cognitive linguistics well ahead of the linguistic paradigm-shift in the 1980s. At about the time when Yip was working on his idiosyncratic theory, Kao Yu-kung of Princeton University and Mei Tzu-ling of Cornell University were collaborating on the analyses of so-called Recent Style poetry of the Tang Dynasty (Kao & Mei 1971, 1978). In the earlier essay, they invoke the American New Critic William K. Â�Wimsatt’s “verbal icon” to support their argument that words in Tang poetry represent “the universals” whilst being “concrete” (Kao & Mei 1971:â•›69–79). The co-authors further Wimsatt’s rehash of the paradoxical concrete-universal (Wimsatt 1954:â•›73–83) in the Aristotelian and Hegelian tradition. They elaborate on this paradox by moving upward and downward in the Aristotelian-Porphyrian hierarchy of biological classification, but they can go no farther than genus and species, in the same way as Kant and Hegel have limited their discussions to ‘species’ and ‘individuals’ (Wimsatt 1954:â•›72). The problem is that ‘individual’ is not a biological class in opposition to species, because, as a vague quantity concept rather than a class concept, it can also be accommodated by domain, kingdom, phylum, class, order and family. Such terminological fault-finding may not be fair to Kao and Mei because it would miss their assertion on Chinese poets’ general penchant for universality. Kao and Mei argue that “Chinese nouns are close approximations of universals” (Kao & Mei 1971:â•›104); and the poet uses “unadorned archetypal nouns,” or “archetypal or primitive terms that stand at the head of each genus” (Kao & Mei 1971:â•›81), such as “man”, “bird”, “flower”. Wang Wei’s four-line poem, already cited earlier by Liu (Liu 1962), is a typical example. (3)

ren xian gui hua luo Man at leisure, laurel flower fall. ye jing chun shan kong Night silent, spring mountain empty. yue chu jing shan xiao Moon rise, startle mountain bird, shi ming chun jian zhong Time to time sing amidst spring brook.

(quoted in Liu 1962:â•›40)

 Han-liang Chang

Here “man”, “night”, “bird”, and even the qualified “mountain” and “flower” are “unadorned archetypal nouns” that constitute the World – with a capital! This kind of “undifferentiated” “imagistic language”, as Kao and Mei see it (1971:â•›128), testifies to the principle of equivalence. And – here follows a non-sequitur – “making things equivalent is the attempt to restore the primordial oneness after the fall.” (Kao & Mei 1971:â•›129) It is inconceivable that such a bizarre reading of Jakobson based on free association should have been published in the prestigious Harvard Journal of Asiatic Studies. Rather, the circumstance shows how Asian Studies have often been ghettoized in certain journals, without becoming integrated in mainstream poetics. With historical hindsight, or perhaps in an attempt to redeem their earlier whims, the co-authors attempt to apply Jakobson’s principle of equivalence in the 1978 sequel. Here they observe: “Since Chinese is a language weak in syntax […] the result is that the metaphoric relation dominates over its complement, the analytic relation.” (Kao & Mei 1978:â•›287). Kao and Mei apply this principle of equivalence to both the lexical and syntactical levels of poems whose “themes” and “motifs” may attract the cognitive linguist interested in the domain of space. Certain common themes in T’ang poetry also call for the use of contrast [as a manifestation of equivalence]; the very nature of themes such as bidding farewell, looking into the distance, and mediating on history invites the poet to make comparisons – between the past and the present, the far-away and the near-athand, or the imagined and the real. (Kao & Mei 1978:â•›287)

However, to deal with such themes, Kao and Mei retrieve – albeit a bit belatedly – the Jakobsonian model of poetic principle, i.e. the projection of selection onto Â�combination, as a way of fine-tuning the Ransomian dialectic of local texture and Â�logical structure, advanced in their 1971 essay (Kao & Mei 1978:â•›286). Kao and Mei’s application covers two major figures, metaphor and allusion, which are “special instances of the principle of equivalence in action” (Kao & Mei 1978:â•›293). Thus the famous lines by Li Po. (4)

fu yun you zi yi Floating cloud, wanderer’s mind; luo ri gu ren qing Setting sun, old friend’s feeling.

(Kao & Mei 1978:â•›293)

are explicated to be containing a pair of metaphors where human sentiments are likened to natural phenomena, or the other way around (Kao & Mei 1978:â•›289). By the same logic, Wang Changling’s (circa 698–756) couplet, (5) dan shi long cheng fei jiang zai If Winged General of Dragon City were present bu jiao hu ma du yin shan He would not let the Hunnish cavalry cross Mount Yin.   (Kao & Mei 1978:â•›289)



Mental space mapping in classical Chinese poetry 

which involves a historical allusion to the Han Dynasty General Li Guang, links the present to the past, presumably through the operation of equivalence. For all the Â�richness of their materials – just imagine the huge corpus of Tang poetry! – Kao and Mei never go beyond the structural model of Jakobson, whether or not their reading is acceptable being another matter. This is reflected in their commentary on the principle of equivalence: Equivalence, consisting of similarity and contrast, is one of the two basic modes of arrangement in ordinary language. In poetry it assumes an even more important role. For example, rhyme and alliteration, prosody and parallelism, are all constituted at least in part by the principle of equivalence. In the general area of meaning, we noted several promising avenues of analysis. When two terms are related by similarity and contrast, new meaning is generated.  (Kao & Mei 1978:â•›293)

3.â•… Equivalence: A structural feature or a cognitive category? This last sentence from the above quotation – with its implications of conceptual blending or coupling, can lead us out of the anachronistically introduced structural paradigm to the cognitive paradigm which was on the rise at the time when Kao and Mei were writing (cf. Rosch 1975; Rosch & Lloyd 1978). But before addressing the cognitive approach, we should examine Kao and Mei’s approach in greater detail to see what is missing. One may have several reservations about Kao and Mei’s ‘formalistic’ approach, but I shall highlight just two theoretical issues: equivalence and Â�universals. Firstly, their subject matter is the Recent Style poetry or what is generally called lü-shi Â�(Regulated poetry) because of its dominant feature of equivalence, which is Â�manifested in Â�diction, couplet verse line, and balanced stanzaic form, as Â�demonstrated by examples (4) and (5) above. It would be quite easy to identify parallelism, on any linguistic level, in this kind of poetry. This dominant feature is so cast in foreground that one tends to lose sight of the poetic language’s Â�pragmatic Â�functions, such as the cognitive process performed by the speaker/actor but entrusted (or indeed Â�initiated) by the reader. For instance, the metaphors in (4) do not generate Â�themselves automatically by the two pairs of terms: “floating clouds” and Â�“wanderer’s mind”; “setting sun” and “old friend’s feeling”. The two domains have to be blended – through speech act perhaps, and crossed to make metaphors possible. Similarly, in example (5), the couplet suggests two spatial-temporal domains, two ‘mental spaces’, that of the enunciation and enunciator (i.e. here and now), and that of the enunciated (i.e. there and then), which have to be negotiated pragmatically in the first place. And the ‘space builder’ that mediates and blends

 Han-liang Chang

the two domains is none other than the textually suppressed enunciator ‘I’, whose presence is however suggested by the conditional “If ” (dan shi). The much-abused equivalence constitutes one of the two poles of language Â�structure conceptualized by Saussure, and elaborated by Jakobson and others. Â�Saussure calls it “rapports associatifs” in opposition to “rapports syntagmatiques” (Saussure 1959); these are later codified by Roman Jakobson and André Martinet as paradigma versus syntagma or metaphor versus metonymy. However, its foundation is Â�opposition, including both the positive and negative dimensions – identification and differentiation – that begin on the supposedly ‘lowest’ phonemic level, and move up to other phonological, semantic, and syntactic levels, and as such can be said to be ubiquitous in language. To show equivalence in poetry is in fact to state the obvious, and one sees equivalence everywhere to confirm only viciously this equivalence model. Moreover, this linguistic phenomenon, albeit based on basic binary logic, cannot be freely appropriated to blend with other metaphysical entities, such as the “primordial oneness” before the Fall, or the Aristotelian concrete-universal. Finally, there are as many ways of conceptualizing the world as there are many natural languages. Therefore, cognitive universality cannot be derived from the particularity of a natural language, such as classical Chinese. This is what I have termed elsewhere in my critique of Leibniz, Hegel and Derrida “hallucinations” (Chang 1988). The reference to the concrete-universal leads us to the problem of category, a Â�common concern of all the Chinese critics I have discussed above – Liu, Yip, Kao and Mei. Their interest in linguistic categorization is, however, unnecessarily Â�complicated by the fuzzy concept of concrete-universal charged with value, and the assumption that Chinese poetry (or one type of it) represents primordial images only. It may not be strange that the Chinese, as other homo loquens, have been concerned with Â�categorization. Kao and Mei’s 1971 discussion of genus and species was echoed by James Liu’s student Pauline Yu (Yu 1987) from an indigenous Chinese perspective. Yu comments on the Confucianist notion of lei, or “categorical correspondences” (Yu 1987:â•›41) in the following statement: In other words, meaning is not attached externally and arbitrarily to an image but follows logically from the fact that objects and, situations were believed traditionally to belong to one or more non-mutually exclusive, a priori, and natural classes. (Yu 1987:â•›42)

One could detect herein the same line of thinking that subscribes to natural morphology, iconicity included. Category is probably the first and foremost, but also the most challenging issue for cognitive studies. The Kantian a priori categories of time and space have received renewed attention from cognitive linguists. Ronald Langacker calls these



Mental space mapping in classical Chinese poetry 

“basic domains” (1987). As such they are capable of forming more complex Â�concepts through blending. Where terminology is concerned, John R. Taylor’s (1989) Â�“prototype” may not be a more beautiful word than Kao and Mei’s “archetype” and “universal”, but the currency gained by ‘prototypical categorisation’ in the late 1980s clearly points to a paradigm-shift. Interestingly, universal categories discussed by Kao and Mei in 1971, and metaphor and allusion, discussed in 1978, continue to be Â�considered in Â�different linguistic and literary contexts, but by cognitive linguists and critics from other Â�perspectives. Researchers have suspended purely theoretical discussions, but tend to adopt an eclectic approach which reconciles the a priori and a posteriori positions. For instance, deictics or indexicality can be regarded as a language universal, but it is at the same time culture-specific and under sociohistorical, and, surely, Â�linguistic Â�constraints (Fillmore 1997). The common themes of space and time, treated in Â�passing by Kao and Mei, are prime categories, as has long been established by Â�Aristotle, Â�Leibniz and Kant, but one no longer cares about bird, flower, mountain, and river, and Â�primordial images, with researchers rather turning to more specific space Â�configurations, Â�addressing such topics as vector grammar, representation of direction, blending of mental spaces, etc. The second part of my paper will be a preliminary exploration into this area, and my data will be taken from the Ancient Style poetry before the Tang Dynasty.

4.â•… Sample analyses of space in Ancient Style poetry Since the earliest recorded times, from the twelfth to fifth centuries BCE, Chinese poets have been obsessed with space and have established a network of spatio-temporal configurations as poetic convention. The oldest anthology Shi Jing (The canon of poetry) abounds with examples showing directions, orientation, and movement. Example (6) is the opening of an ancient rhyme about an event which took place presumably in the pre-Confucian Zhou Dynasty, and it was recorded no later than the first century A.D. in the Grand Historian’s Annals. (6)

den bi xi shan xi [I] am climbing up yonder West Mountain cai qi wei yi To gather its wild beans! (Ancient Rhymes, Anonymous. Shen 1975:â•›16)

There are at least two instances of spatial language, i.e. linguistic expressions describing spatio-temporal configurations: (i) the spatial variable “West Mountain” that indicates “object location”, and (ii) the present participial phrase “climbing up” that indicates “object movement”. These two variables are common motifs in Ancient Style poetry. A most popular example is probably.

 Han-liang Chang

(7) cai ju dong li xia [I] plucked chrysanthemums under the East hedge you ran jian nan shan And by chance glanced South Mountain  (Tao Qian [circa 365–427] Shen 1975:â•›213.) While both (6) and (7) involve topology, they evoke, respectively, two states of human existence, kinesis and stasis, with (6) suggesting frames of reference and Â�vector and motion, and (7) only frames of reference established by the speaker/actor’s Â�perspective. Now a whole spectrum of spatial concepts that constitute the larger Â�spatial domain can be evoked: frame (Lee 2001), profile (Langacker 1987, 1999), orientation, movement, frames of reference, construal, perspective-taking (Levelt 1996), ground and figure (Levinson & Wilkins 2006), landmark and Â�trajector Â�(Langacker 1987), trigger and target (Fauconnier 1997), base space and projected space (Stockwell 2002). In Example (6) West Mountain is the landmark, and the speaker/actor, or his persona, will be the trajector. One could also label “West Â�Mountain” as target 1, the “wild beans” target 2, both of which serve to trigger the speaker/actor’s kinesis, i.e. bodily movement from the ‘here and now’, the moment of enunciation to the ‘there’ of ‘yonder’ West Mountain. The case of Example (7) is more subtle. First of all, one would have to pose the questions: Whose orientation do the lines suggest? Does the perspective belong to the landmark or the trajector? In either case, who or what are they? A chain of relationships can be arrived at. In line 1, “I” is the trajector (T1), and “chrysanthemum” the landmark (L1) but once at the East hedge, “chrysanthemum” becomes the trajector (T2) and “East hedge” the landmark (L2). Beyond the semantic universe, the whole of line 1 becomes a trajector whose landmark is line 2. Thus in the reading process, the action of “plucking chrysanthemums by the East hedge” serves as a trajector (T3) for the new landmark (L3) “glancing in leisure South Mountain”. This is where semantics and pragmatics merge. How are domains blended? Jakobson would suggest his projection of selection onto combination. Fauconnier is never very clear about this. When we compare (6) and (7), (6) looks dynamic because of the space-builders “climbing up” and “plucking” and because there is no indication of initial bodily location, whilst at first glance (7) looks static (Levinson & Wilkins 2006:â•›3), because there is no bodily movement that changes the speaker/actor’s location, save the relatively static “plucking” and perspective-taking “glancing”. However, if we did a detailed analysis, we would find (7) more dynamic than (6) because of the chain-reactions of landmark/trajector or target/trigger. The series of landmark and trajector belong to different types of mental spaces: “time spaces”, as in (5), “space spaces”, as in both (6) and (7), “domain spaces”, as indicated by “climbing” in (6) and “plucking” in (6) and (7), “glancing” in (7), and “hypothetical spaces”, as the subjunctive in (5) (Stockwell 2002:â•›96). A diagram to Â�compare the conceptual domain in (6) and (7) can be borrowed from Levinson & Wilkins (2006:â•›3).



Mental space mapping in classical Chinese poetry  Conceptual subdivision of the domain

Stasis

Non-angular

Kinesis

Angular

Topology

Frames of reference

Intrinsic

Relative

Motion

Absolute

First of all, in both (6) and (7), the spatial domain is angular rather than topological. There are relative frames of reference between landmark and trajector, East hedge, South Mountain, and West Mountain, and there is motion in the trajectors of both. If classical Chinese poetry used universals only, then its spatial domain would be �topological, as in the sub-domains in (8). (8)

da mo gu yan zhi Large desert lone smoke straight [vertical] chang he luo ri yuan Long river setting sun round [circular] (Wang Wei. Gao 1970:â•›425.)

Although da (big, large, broad, etc.), chang (long), zhi (straight-line), and yuan (round, circle, spheric) are primordial, once set in relation to other components, i.e. once Â�syntactic relations are established, the whole spatial domain becomes angular and kinetic. This would be a counter example to Yip’s and Kao and Mei’s argument for Wang Wei’s “non-analytical” landscape. In (6) and (7), the chain reactions between landmarks and trajectors, or between mental spaces that are to be blended, can only be explained, if explained away, to be culture-specific, the formalistic constraint, such as genre, being in itself a Â�second-order signification on top of the primary model of Chinese (which Â�Chinese?). One such Â�constraint is – as all students of Chinese poetry would immediately Â�recognize – the Â�timeless motif (or “domain space”) of flower “plucking”, indeed plucking Â�chrysanthemum, in particular, has become almost a literary institution (a domain larger than allusion) since Tao Qian. In other words, and with due respect to Â�Fauconnier et al., there is a time space of the timeless space of domain space which remains a

 Han-liang Chang

Â� hypothetical space because it belongs to another reality space which is poetry. Now all of these need blending. What may serve as a better blender than our old but outlawed friend syntax or the axis of combination? Another culture-specific constraint is surely directions, a key domain in vector grammar. In both (6) and (7), there are the “space spaces” of directions, which John O’Keefe defines, not without a touch of irony, as “parallel, infinitely long vectors” (O’Keefe 2003:â•›70–71). For thousands of years, Chinese houses, especially in Â�northern China, have been built according to one almost exclusive formula: they are seated North, facing South, with gardens on the East, and ‘towers’ on the West. Without this knowledge one would not be able to understand (7). This spatial cognition is infinitely reflected in poetry, that is, once it has entered language or, more precisely, having been already created by language at the first instance. This would lead us again into the a priori versus a posteriori circular argument. Let us stop that and accept the diagram by van de Zee & Slack (2003:â•›2) (cf. Jackendoff, 1996).

Space-time

Formal/cognitive representation of space-time

Language-space set of linguistic expressions describing spatio-temporal configurations

We concede that the Chinese, in the fourth to fifth-century Jin Dynasty when Tao Qian flourished, had an agreed construal of spatial configurations expressed in the poetic language of songs (yue fu) and Ancient Style poetry, whose later appellations they could not have anticipated. These expressions and the cognition involved are mutually implicated, as indicated by the double arrow. The two-way traffic can be suggested by Tao’s use of the ancient formulaic “plucking” to blend his daily life domain and the creative domain. In fact, this recurrent motif was so popular that it ceases to be realistic, but serves only as a space builder, a xing or poetic rising, which blends the mental space that is the larger and longer lyrical tradition and the mental space which is the innermost recesses of poetic imagination. Other such domain spaces include, no doubt, the poet’s sense of direction. In the following section, I shall read a most famous Ancient Style poem “Jiangnan” (South of the Yangtze River), supposedly a folk song collected by literati during the Six Â�Dynasties (third to sixth centuries). The lyric was obviously quite popular during the time, because scholars agree that it was tuned and sung in public performances. Although it was admitted into most anthologies, few people take the verse seriously as an example of good poetry. It is time to right a wrong.



(9)

Mental space mapping in classical Chinese poetry 

jiang nan ke cai lian One could pluck lotus seeds in Jiangnan (River South) lian ye he tian tian How rich the lotus leaves grow [How like ploughed fields] yu xi lian ye jian Fish frolic amidst lotus leaves yu xi lian ye dong Fish frolic to the east of lotus leaves yu xi lian ye xi Fish frolic to the west of lotus leaves yu xi lian ye nan Fish frolic to the south of lotus leaves yu xi lian ye bei Fish frolic to the north of lotus leaves. (Shen 1975:â•›77)

The poem’s authorship is unknown, and the only plausible conjecture is that it was a popular song in the xiang he (‘echoing’) style; and the refrain formed by the last four lines suggests that it was sung by the chorus or by several parts of it in alternation. Because of its seeming simplicity and lack of artistic sophistication, the poem has been regarded as belonging to the folk tradition. But let us turn to the study of the poem’s space domain. As with (6) and (7), the first thing one notices is the directions. All four directions are evoked in the refrain in a rather mechanical order. For Chinese language users throughout the centuries, however, the four directions follow an old colloquial usage. Even today, we still give directions in that order: starting with the east, proceeding to its opposite west, next moving counter-clockwise to the south, and finally ending in the latter’s opposite north. Another more recent order is the clockwise east/south/west/ north. In both cases, one starts with the east and ends in the north. This space configuration, traceable probably to ancient fertility myth and the cyclical concept of time, is lent to the fish which frolic in the same order. Now the space-blender is obviously the lexical string in colloquialism, linked not by syntax, but dictated by inherited cultural convention. It is an unconscious meme, as it were, that the north is least preferred. This fore-knowledge leads us to the title, “Jiangnan”, literally, the region to the south of the Yangtze River, i.e. the warmer part of China. In Chinese literature, Â�Jiangnan is not only represented as a land of plenty, but it has also become an Â�equivalent to the symbol of the cornucopia, partly as a result of the politicized north-south divide. For this reason, one can safely surmise that there is a reality space belonging to the Â�hidden impersonal speaker, and that this reality space is the north, Jiangbei, or the region to the north of the Yangtze River. From the northerner’s point of view, Jiangnan is an imagined space, a space of alterity. The blender of these two mental spaces is one modal word: the concessional and conditional adverbial ke (may, could), showing a possibility rather than actuality. The relation of the two spaces is shown in the diagram.

 Han-liang Chang

Reality space (sp 1) Jiangbei here/now ipseity

Imagined space (sp 2) Jiangnan there/future alterity

In Fauconnier’s terms, Space 1 serves as a trigger and Space 2 as its target. Once the two spaces are built and blended by the conditional ke, and by two modalities, the alethic and the axiological, respectively showing probability and value, the poem is ready to move to a third space, lotus ponds, where people can gather (cai) lotus seeds and where fish frolic (xi), or where people watch fish frolic, another culture-specific space domain that goes back to the Daoist philosopher Zhuangzi (Chang 2005).

Reality space (sp 2) Jiangnan (space space) ipseity (territorialized)

Imagined space (sp 3) lotus ponds (space space) harvesting (domain space) fish playing (domain space?) alterity (fishs Õ space)

Between line 1 and line 3, where two spaces are blended, i.e. the human pleasure of gathering lotus seeds and that of fish frolicking are identified, there is a strange intrusion by line 2. The Chinese reads: /lian/ /ye/ /he/ /tian/ /tian/, which can give rise to two different renditions due to the polysemy of /tian/: “How rich the lotus leaves grow!” or “How the lotus leaves look like tian (i.e. ploughed fields in the north or water rice pads in the south)!” One reading says the repeated tian is an adjective, meaning ‘rich’ or Â�‘vigorous’ [in growth], to qualify lotus leaves. But originally the written word of tian ⊞ was an icon, a pictograph of a divided ploughed field, with two footpaths crossing in the middle, thus dividing the plot of land into four pieces, east, west, south and north, provided that one moves the square sign slightly clockwise until it becomes a diamond or lozenge, with an angle facing north, and another facing south. That is why a secondcentury lexicographer explained the written sign as being composed by four reduplicated square mouth, /kou/ ⎕, divided by a vertical footpath and a horizontal footpath.



Mental space mapping in classical Chinese poetry 

The two intersecting lines that make right angles and a cross + that divides the spacespace (as the signified concept, le signifié) are in fact Â�morphographemes (as the signifying visual form, le signifiant, with or without the sound signifier) of another graphic sign which is pleremic (meaning-ful) in its own right. The sign + means Â�number ten, and the idea is derived from its components: the horizontal line , referring to the West and the East, and the vertical line of |, referring to the North and the South, with their intersection pointing to the centre. This explains beautifully the four, nay, five directions in which the fish play (Guo 2006:â•›42. Xu Shen, Shuowen jiezi). Now what kind of mental space is this? We are back to the realm of signs, not the Saussurian linguistic sign constituted by the signifying sound and the signified Â�concept, but a more intricate phonographic sign, consisting of the triply articulated phonic and graphic signifiants and their shared signifié, or, more precisely, two Â�coexisting signifiants, pointing to two signifiés. Our spatial cognition is therefore put to the test. There are now three other pairs of mental spaces, developed on the logical semantic level, and the purely formalistic, semiographemic level (Chang 1996).

Reality space (sp 2) Jiangnan ipseity

Semiographemic space (sp 5) iconic written sign 田

imagined space = reality space (sp 4) lotus pool = ploughed field 田

Reality space (sp 6) icon of lotus leaves

For a poet who has not been to Jiangnan or whose experience is mediated by previous texts, there is, however, another route to satiate his/her vicarious experience. Perhaps

 Han-liang Chang

by accident, or through his unconscious (partially mental) space, the poet manages to get to the imagined space of Jiangnan through writing, indeed through the magical function of a written sign. The fish frolic in both fields, the lotus leaves that resemble tian, a quartered watery rice field, which is iconic to the pictograph sign. Thus, line 2 lian ye he tian tian can be translated – rather than “How rich the lotus leaves grow!” or “How the lotus leaves look like ploughed fields!” – as, Â�miraculously, “How lotus leaves are like the tian word?” This kind of special cognition where Â�literacy and Nature merge is quite common in Chinese. Even today, we often say, “Lo, the geese fly like a human word [ren zi]!”, “She is wearing slippers with a human word [ren zi] Â�(flip-flop)”, or “He sleeps like a big word [da zi].” Does writing map Nature, or the other way round? The semiographemic mediation is also seen in the two Â�space-building words, cai (pluck) and xi (frolic), which link the human world and the piscine world by pointing to the identical biological domain of Umwelt. Such semiosis bears witness to the mental space mapping of Chinese poetry in general.

References Bloom, P., Peterson, M.A., Nadel, L. & Garrett, M.F. (eds). 1996. Language and Space. Â�Cambridge MA: The MIT Press. Chang, H.-l. 1988. Hallucinating the other: Derridean fantasies of Chinese script. Working Paper No. 4. Milwaukee WI: The University of Wisconsin-Milwaukee, Center for Twentieth Â�Century Studies. Chang, H.-l. 1996. Semiographemics: A Peircian trichotomy of classical Chinese script. Â�Semiotica 108 (1–2): 31–43. Chang, H.-l. 2005. Intersubjectivity in Controversy: A story from the Taoist philosopher Zhuangzi. In Controversies and Subjectivity [Controversies 1], P. Barrotta & M. Dascal (eds), 173–184. Amsterdam: John Benjamins. Cheng, F. 1982. [1977]. Chinese Poetic Writing. Trans. D.A Riggs & J.P. Seaton. [Studies in Â�Chinese Literature and Society]. Bloomington: Indiana UP. Fauconnier, G. 1997. Mappings in Thought and Language. Cambridge: CUP. Fauconnier, G. & Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Â�Hidden Complexities. New York NY: Basic Books. Fillmore, C.J. 1997. Lectures on Deixis. Stanford CA: CSLI. Gao, B.-y. (ed.) 1970 [1890]. Tang Song Shi Juyao (Essential Poems from the Tang and Song Dynasties). Reprint. Taipei: Yiwen Yinshuguan (Publishing House). Guillén, C. 1971–72. Some observations on parallel forms. Tamkang Review 2 (2) – 3(1): 395–415. Guo, J. 2006. Xie Zhi de tou ying: zhongguo de fa wenhua [Projections of the fabulous animal Xiezhi: Legal culture in China]. Shanghai: San lian Bookstore. Hickmann, M. & Robert, S. (eds) 2006. Space in Languages: Linguistic Systems and Cognitive Categories [Typological Studies in Languages 66]. Amsterdam: John Benjamins.



Mental space mapping in classical Chinese poetry 

Jackendoff, R. 1996. The architecture of the linguistic-spatial interface. In Language and Space, P. Bloom, M.A. Peterson, L. Nadel & M.F. Garrett (eds), 1–30. Cambridge MA: The MIT Press. Jakobson, R. 1987 [1966]. Grammatical parallelism and its Russian facet. In Language in Â�Literature. K. Pomorska & S. Rudy (eds), 145–179. Cambridge MA: Belknap Press. Jarvella, R.J. & Klein, W. (eds). 1982. Speech, Place, and Action. London: Wiley. Kao, Y.-k. & Mei, T.-l. 1971. Syntax, diction, and imagery in T’ang poetry. Harvard Journal of Asiatic Studies 31: 49–136. Kao, Y.-k. & Mei, T.-l. 1978. Meaning, metaphor, and allusion in T’ang poetry. Harvard Journal of Asiatic Studies 38: 281–356. Landau, B. & Lakusta, L. 2006. Spatial language and spatial representation: Autonomy and Â�interaction. In Space in Languages: Linguistic Systems and Cognitive Categories, M. Â�Hickmann & S. Robert (eds), 309–333. Amsterdam: John Benjamins. Langacker, R.W. 1987. Foundations of Cognitive Grammar, Vol. 1: Theoretical prerequisites. Â�Stanford CA: Stanford University Press. Langacker, R.W. 1999. Grammar and Conceptualization [Cognitive Linguistic Research 14]. Berlin: Mouton de Gruyter. Lee, D. 2001. Cognitive Linguistics: An Introduction. Melbourne: OUP. Levelt, W.J.M. 1996. Perspective taking and ellipsis in spatial descriptions. In Language and Space, P. Bloom, M.A. Peterson, L. Nadel & M.F. Garrett (eds), 77–107. Cambridge MA: The MIT Press. Levinson, S.C. 2003. Space in Language and Cognition: Explorations in Cognitive Diversity. Â�Cambridge: CUP. Levinson, S.C. & Wilkins, D.P. 2006. The background to the study of the language of space. In Grammars of Space: Explorations in Cognitive Diversity, S.C. Levinson & D.P. Wilkins (eds), 1–23. Cambridge: CUP. Liu, J.J.Y. 1962. The Art of Chinese Poetry. Chicago IL: University of Chicago Press. O’Keefe, J. 2003. Vector grammar, places, and the functional role of the spatial prepositions in English. In Representing Direction in Language and Space, E. Van der Zee & J. Slack (eds), 69–85. Oxford: OUP. Owen, S.1985. Traditional Chinese Poetry and Poetics: Omen of the World. Madison WI: Â�University of Wisconsin Press. Pütz, M. & Dirven, R. (eds). 1996. The Construal of Space in Language and Thought [Cognitive Linguistics Research 8]. Berlin: Mouton de Gruyter. Rosch, E. 1975. Cognitive representations of semantic categories. Journal of Experimental Â�Psychology 104: 192–233. Rosch, E. & Lloyd, B.B. (eds). 1978. Cognition and Categorization. Hillsdale NJ: Lawrence Â�Erlbaum Associates. de Saussure, F. 1959. [1916]. Course in General Linguistics, C. Bally, A. Sechehaye & A. Riedlinger (eds), W. Baskin (transl.). New York NY: MacGraw-Hill. Shen, D.-q. (ed.). 1975. (c.1739). Gushi Yuan (Sources of Ancient Poetry). Reprint. Taipei: Â�Huazheng Shuju (Bookstore). Stockwell, P. 2002. Cognitive Poetics: An Introduction. London: Routledge. Svorou, S. 1994. The Grammar of Space [Typological Studies in Language 25]. Amsterdam: John Benjamins.

 Han-liang Chang Taylor, J.R. 1989. Linguistic Categorization: Prototypes in Linguistic Theory. Oxford: Â�Clarendon Press. Van der Zee, E. & Slack, J. (eds). 2003. Representing Direction in Language and Space. Oxford: OUP. Wimsatt, W.K. 1954. The Verbal Icon: Studies in the Meaning of Poetry. Lexington KY: University of Kentucky Press. Xu, Shen [c. 58–c.147] 100AD. Shuowen jiezi zhu (Annotated ancient lexicography). Glossed by Duan Yucai (1735–1815). 〈http://shuowen.chinese99.com〉 Yip, W.-l. 1969. Ezra Pound’s Cathay. Princeton NJ: Princeton University Press. Yu, P. 1987. The Reading of Imagery in the Chinese Poetic Tradition. Princeton NJ: Princeton University Press.

Iconicity in conceptual blending Material anchors in William Morris’s News from Nowhere David Glyn Wilson

Universities of Neuchâtel and Geneva In 1890, William Morris, pioneering writer, designer, and revolutionary socialist, published News from Nowhere. In this utopian romance, he expresses his aesthetic and social convictions most convincingly not through propositional argumentation, but instead through iconic conceptual/material anchors set up to support conceptual blending networks through which the reader is guided to conjure up human-scale scenes and contexts. These conceptual integration networks achieve compression (and decompression) of vital relations (such as identity, change, or time) in blended mental spaces, heightening the aesthetic, emotional, and hence persuasive potential of the text, often at key points in its structure. This paper explores several passages from News from Nowhere, arguing that motivated non-metaphorical conceptual/material anchors can be just as iconic as metaphorical ones.

1.â•… Introduction: The aims of this study Within a framework of Cognitive Poetics (for example, Freeman 2007; Brône & Â�Vandaele 2009), this study aims to examine the iconicity of conceptual blends with Â�material anchors in William Morris’s utopian romance News from Nowhere (NN). Using blending theory (for example, Fauconnier & Turner 2002), I will consider the contribution of these blends to structuring NN and fulfilling Morris’s aim of moving the reader (to action) by giving the “semblance of felt life” (Freeman 2009). In so doing, I hope to throw light on how NN’s material anchors make a direct appeal to readers’ emotional cognition, prompting for a compelling embodied vision of ‘Nowhere’ and its hoped-for “epoch of rest”. In examining the iconicity of non-metaphorical conceptual blends, the Â�question arises of how blending relates to a general theory of iconicity. Within iconicity studies, although metaphor and metaphorical blends have often been considered (for Â�example, Herlofsky 2003; Hiraga 2005; Freeman 2009), so far the relationship

 David Glyn Wilson

between iconicity and blending does not appear to have been fully theorised, since Â�‘non-metaphorical blends’ seem to remain unaccounted for in iconicity theory – for example, in the work cited above and in the diagrammatic typology first drawn up in Nänny and Fischer (1999:â•›xxii) and redrawn in Freeman (2009:â•›172). This seems problematic, as the Â�present study reflects the fact that iconic non-metaphorical blends are Â�ubiquitous and can be as important in the creation of ‘poetic iconicity’ as the metaphorical blends discussed by Freeman (2009). If non-metaphorical blends can be accounted for in the current typology of iconicity, it needs to be made explicit how this can be done; and, of course, if it cannot be done, some other solution will  be needed. This question is beyond the scope of the present study. Here, I will Â�content myself with preparing the way for this discussion by looking at the iconicity of Â�non-metaphorical material anchors in News from Nowhere.

2.â•… William Morris (1834–96) and News from Nowhere (1890) Although primarily thought of as a poet in his own time, Morris was “remarkably at the forefront of five historic movements in western culture” (Latham 2007:â•›x). This range and versatility provided Morris with the raw material to help create striking conceptual blends in his later writing. Before discussing blending theory and applying it to NN, it will be useful to give an overview of the many and varied influences that led Morris to write this utopian romance, as well as to sketch a brief summary of its structure and themes.

2.1â•… Time, place, and politics: The Thames, Kelmscott, Iceland, Nowhere Obliged by professional commitments to his decoration and furnishing firm to spend most of his time in London, Morris was sickened by the injustices of Â�Victorian England that he saw all around him, with its industrial exploitation of the poor and degradation of the natural environment. Initially, his writing was characterised by a withdrawal from this horror – witness the often quoted lines from his long poem The Earthly Paradise: “Dreamer of dreams, born out of my due time,/Why should I  strive to set the crooked straight?” (Morris 2003). Increasingly, though, he became Â�interested in Â�precisely this question of setting the crooked straight. Morris’s Â�substantial Â�commitment to social change came about with his joining the Â�Democratic Â�Federation in 1883. Between 1888 and 1890, on top of onerous political campaigning, other Â�writing and of course the running of Morris & Co., he wrote his two socialist Â�dream-visions, both initially Â�published in The Commonweal, the newspaper of the Socialist League. The second of these dream-visions was News from Nowhere, in which Morris would use ‘story’ to blend past and present to yield a vision of the future.



Iconicity in conceptual blending 

Already published in 1888, Edward Bellamy’s widely read and influential socialist Â�utopia Looking Backward had portrayed the future in terms of an efficient but mechanistic, coercive and somewhat joyless form of state socialism. Morris, however, had been appalled at Bellamy’s meagre vision. He set about offering an alternative vision embodying his own hopes for a socialist future in Britain that would follow a necessary revolution. News from Nowhere was no knee-jerk reaction to Looking Backward, though. The ideas and the emotions developed in NN had been blending in Â�Morris’s mind for decades. Morris the romantic medievalist dreamed of an idealised past. While at Oxford he had read widely in medieval history. But already aged nine he could be seen dressed in a miniature suit of armour riding a pony in Epping Forest, and had apparently read all of Walter Scott’s novels by the age of seven. So it is no surprise that A Dream of John Ball (1888), the first of his two socialist dream-visions, should be the tale of a nineteenth-century narrator transported back to the fourteenth century to meet the revolutionary priest John Ball, soon to be one of the leaders of the Â�unsuccessful 1381 Peasants’ Revolt. NN, the second of his socialist propaganda tales, also has prominent elements of the medieval dream-vision, (recalling, for Â�example, Piers Plowman and Winner and Waster) in which “the narrator, wrestling with a Â�particular question, goes to sleep and has a dream in which he encounters a guide and authoritative figures who lead him through a sequence of events in a fantastic landscape” (Cowan 2007:â•›138–139). In this respect, NN has been seen as “a conscious anti-novel, hostile to Â�virtually every aspect of the great tradition of Victorian fiction” (Brantlinger 1975, cited in Parrinder 1976:â•›1). It has also been criticised for a lack of in-depth characterisation (Wilmer 1993:â•›xxix). But this deliberate choice arises from Morris’s chosen form and more generally from the fact that – rather than seeking the minute dissection of Â�character and motive in the manner of a Henry James, for example – Morris “looks to literature for the timeless and symbolic patterns of human experience which we find in myth, in folk literature and in dreams” (ibid.: xxx). Indeed, this conventionality in the dream-vision lends itself to a kind of compression involving people and place and time and thus a representative iconicity that is more difficult to achieve with the individual and the particular. This will be examined in more detail in the analysis in Section 4. As Wilmer points out, although in time Morris was to recognise more fully the negative aspects of medieval life, he retained the belief that the middle ages were “rarely guilty of the specifically social and economic evils that deform the industrial world” (ibid.: xxvi). Indeed, after his conversion to a socialism which looked to the future, for Morris “the medieval world is no longer a refuge from the present; it helps him, rather, to understand the present and construct, in imagination, an alternative future” (ibid.: xxvii). A Dream of John Ball (1888), as well as the fantasy romances The House of the Â�Wolfings (1888) and The Roots of the Mountains (1889), all look back in time

 David Glyn Wilson

for Â�inspiration and hope on the problem of achieving a noble, just and sensually Â�pleasurable human society. Equally, these works all reflect the influence of places that were Â�significant in Morris’s developing thought. First and foremost, Kelmscott – Â�probably the most iconic of all the places associated with Morris – linked by the River Thames to Hammersmith, where Morris had his London home. But, oddly enough, there was also Iceland, which he visited twice. It might seem a far cry from the gentle Cotswold countryside of Kelmscott, but what Morris saw as this simple, heroic and more egalitarian society was to provide elements of social structure that he would project to his picture of a utopian fellowship of the future with its roots in the past. NN integrates all these strands in Morris’s thought, blending place and Â�elements from the distant past with the nineteenth-century present, and from this blend emerges a dream-vision of the future with a haunting feeling of utter change yet strong Â�continuity with the past.

2.2â•… News from Nowhere: Structure and themes This utopian romance, set in the late nineteenth century and the year 2102, tells of a post-revolutionary society, reborn out of the squalor and inequality of capitalist Â�Victorian England to “days of peace and rest, and cleanness and smiling goodwill” (NN: (4)1, where “mastery has changed into fellowship” (NN: 228). As the narrative begins, William Guest has just left a factious meeting of the Socialist League in late nineteenth-century London. He returns discontentedly to his Â�Hammersmith home and falls asleep, awaking to find himself still in London, but in the twenty-second century. A revolution had taken place in 1952 and the new egalitarian society is unrecognisable in almost all respects from the Victorian capitalist regime that Guest has left behind. In ‘Nowhere’, work is synonymous with art, and “all work is now pleasurable […] because there is conscious sensuous pleasure in the work itself; it is done, that is, by artists” (NN: 122–123) – a stark contrast to Bellamy’s Looking Â�Backward (1888), where people work out of duty or coercion. In medieval dream-vision style, Guest is led on two journeys, the first from Â�Hammersmith via Trafalgar Square to Bloomsbury. All has changed radically, but not quite beyond recognition. The second journey is a long navigation of the Thames by rowing-boat into the idyllic setting of the Cotswolds, to Kelmscott Manor, and Â�ultimately Kelmscott church, in which a quasi-medieval hay-time festival and feast (the haysel) is about to begin.

.â•… Quotations from News from Nowhere (from the Wilmer 1994 edition) are throughout cited thus: NN followed by the page number. The longer extracts in Section 4 also include the chapter number.



Iconicity in conceptual blending 

Ultimately, the dreamer Guest must fade from the consciousness of his hosts in Nowhere, tantalisingly without participating in the feast, returning to awaken once more in the nineteenth century, and this is not without some anguish. Nonetheless, Guest at last finds that he is “not so despairing” (NN: 228): Or indeed was it a dream? If so, why was I so conscious all along that I was really seeing all that new life from the outside, still wrapped up in the prejudices, the anxieties, the distrust of this time of doubt and struggle? […] if others can see it as I have seen it, then it may be called a vision rather than a dream. (ibid.)

Wilmer claims that NN “achieves its effect not by sermonizing but by creating a Â�compelling dream-world to stand in opposition to the world we know” (1993:â•›xli). While an “opposition” is certainly partially explanatory, this analysis of the effect of NN can be refined by applying the cognitive theory of conceptual integration.

3.â•… C  onceptual integration: Story, projection, mental spaces, blending, and material anchors In this section, the theory of ‘conceptual integration’, also referred to as ‘blending’, will be briefly introduced. More details will emerge in Section 4 when the framework is applied to News from Nowhere.

3.1â•… Story – projection – parable Mark Turner has claimed that “narrative imagining – story – is the fundamental Â�instrument of thought. Rational capacities depend upon it. It is our chief means of Â�looking into the future, of predicting, of planning, and of explaining” (Turner 1996:â•›4–5). He goes on to demonstrate how we use ‘story’ projected as ‘parable’: “projection of one story onto another may seem exotic and literary, and it is – but it is also, like story, a fundamental instrument of the mind. Rational capacities depend upon it. It is a Â�literary capacity indispensable to human cognition generally” (ibid.: 5). As Turner acknowledges, C.S. Lewis had observed in 1936 that parable – the result of the projection of story – belongs “not merely to expression and not exclusively to literature, but rather, […] to mind in general” (ibid.: 7). I will cite just two further examples of the recognition of the cognitive power of ‘story’. The first is from Donald Schön, who realised that the question of how social problems are set should precede the question of how they are solved. He noted that “Over the last thirty or forty years, people have told some very different Â�stories about urban housing, and there have been some very dramatic shifts in ideas in good Â�currency about the problem” (1979:â•›262). Interestingly, looked at afresh from today’s perspective, Figures 6 and 7 in Schön’s paper (ibid.: 276) strongly prefigure

 David Glyn Wilson

the Â�heuristic blending diagrams used to illustrate Fauconnier and Turner’s theory of conceptual integration (see Figure 1). Generic space

Input I1

Input I2

Blend Figure 1.╇ Basic blending diagram (from Turner 2007a:╛379)

A second example, from medical science, will serve to demonstrate the widespread (but still insufficiently acknowledged) acceptance of the importance of ‘story’ to human thought processes. In the BMJ (British Medical Journal) of November 2009, the Â�“Editor’s choice” focused on “the power of stories” in public health decision making: “The brains of human beings seem built to process stories better than other forms of input […] they seem to offer a solution – a way to extract some meaning and redemption from tragedy by preventing its reoccurrence.” So said Thomas B Newman, professor of epidemiology and biostatistics at the University of San Francisco, in the Christmas BMJ six years ago, in a plea for a broad perspective when making public health decisions (BMJ 2003; 327:â•›1424–7).  (Groves 2009:â•›b4887, my emphasis)2

So how does story come to define the way we think? How can one story affect another? To tackle this question, we can turn to the theory of ‘conceptual integration’. .â•… I am grateful to Dr Williamina Wilson for this reference, as well as for extensive Â�discussion of the ideas in this paper.



Iconicity in conceptual blending 

3.2â•… Mental spaces and blending Blending operates on ‘mental spaces’ – “very partial assemblies” of meaning “constructed as we think and talk […] Mental spaces are connected to long-term Â�schematic knowledge, such as the frame for walking along a path, and to long-term specific knowledge, such as a memory of the time you climbed Mount Rainier in 2001” Â�(Fauconnier 2007:â•›351). So in Fauconnier’s example, this particular mental space would include “you, Mount Rainier, the year 2001, and your climbing the mountain” (ibid.). “Running multiple mental spaces, or, more generally, multiple constellated Â�networks of mental spaces […] and blending them […] is at the root of what makes us human” (Turner 2007a:â•›378). So what does it mean to blend mental spaces? Â�Figure 1 shows the usual basic heuristic diagram for blending (from Turner ibid.: 379). In this basic diagram, only four spaces are represented. But often, as Turner points out, many spaces are involved, creating complex networks of meaning – see Fauconnier and Turner (2002) for a comprehensive account of the theory they jointly founded in 1993 (Turner 2007a:â•›377). “While this static way of diagramming aspects of the process is convenient, such a diagram is always a snapshot of an imaginative and complicated development that can involve deactivating previous connections, reframing previous spaces, and other actions” (ibid.: 379). One of the most accessible and comprehensive summaries of blending, highlighting the role of the blend as a ‘conceptual anchor’ for the whole complex integration network, runs as follows: Conceptual integration, also called ‘blending’, is a basic mental operation that works on conceptual arrays to produce conceptual integration networks. Certain conceptual arrays provide inputs to the network. Selective projection from the input conceptual arrays and from the relations between them carries elements and relations to a blended conceptual array that often has emergent structure of its own. This blended conceptual array is often referred to as ‘the blend’. The blend typically does not obliterate the inputs. It provides a human-scale, integrated scenario that serves as a conceptual anchor for the conceptual integration network. (Turner 2008:â•›13)

3.3â•… Compression in the blended space The overarching goal of conceptual integration is to “achieve human scale” Â�(Fauconnier & Turner 2002:â•›312ff.). This is achieved largely through compression. We do not establish mental spaces, connections between them, and blended spaces for no reason. We do this because it gives us global insight, human-scale understanding, and new meaning. It makes us both efficient and creative. One of the most important aspects of our efficiency, insight, and creativity is the compression achieved through blending. (ibid.: 92)

 David Glyn Wilson

Compressed in the blended space are a number of very basic conceptual relations that we perceive and impose constantly on things and processes in the world. Such Â�relations include cause–effect, change, time, and identity among others (ibid.: xiii). Â�Fauconnier and Turner (2002) provide numerous examples of how compression enables us to give human-scale shape and meaning to otherwise diffuse phenomena and vast Â�time-spans. For example, Fauconnier and Turner (ibid.: 95) cite a visual Â�presentation of dinosaur–bird evolution showing a single creature, beginning as a dinosaur and moving through positions on a path while it becomes more bird-like at each stage, until finally it is a Â�fully-fledged bird: in the picture, it is literally ‘turning into’ a bird. As they point out, none of us believe that any single dinosaur mutated into a bird, but the picture is Â�useful for us in compressing otherwise barely manageable relations of time, change, and Â�identity. Likewise, Christmas Day 1982 and Christmas Day 2009 are linked in our minds by the vital relation of ‘time’, allowing us to bring these two separate mental spaces together under compression in the identity of ‘Christmas Day’. Time, change, identity, space, and cause–effect are some of the most important of the vital relations that, under Â�compression, help create the human-scale blended space of William Â�Morris’s Nowhere, as we shall see in Section 4.

3.4â•… Conceptual anchors and material anchors Culturally established blends with both conceptual and material anchors include the cyclic day, money, the grim reaper, and many others (see Fauconnier & Turner 2002). Material anchors are typically physical objects which stand for the conceptual blend – in the case of money, for example, coins and notes: these relatively worthless objects stand for and facilitate a highly-developed system of relative values and complex rules for exchange, and so on. Blends such as money are so long-established in culture that of course we do not run the blend in our minds when we use money. Such blends involve actual physical objects, such as coins (or watches, in the case of the cyclic day): these material anchors are iconic for the conceptual anchors which are themselves iconic conceptual blends, standing for complex integration networks. So in what way can a blend arising in a creative work of literature (or other art) be said to constitute a conceptual or material anchor? This is the question that will be tackled in the next section.

4.â•… News from Nowhere: A blending analysis 4.1â•… The hyper-blend of Nowhere News from Nowhere has an overall blending network comprising many smallerscale ‘local’ blends. Fauconnier and Turner (2002) refer to such an overall network as a ‘mega-blend’, while in Turner (2007b) it is referred to as a ‘hyper-blend’: I will



Iconicity in conceptual blending 

use the latter term. Utopias and dream-visions are hyper-blends of varying degrees of Â�complexity, and coming out of Morris’s amazingly varied background, NN is not Â�surprisingly a complex case. As Yuri Cowan points out, [t]he dream-vision, like allegory, demands serious consideration on the part of its reader, and rewards thought with an insight into its particular concern (whether that be the nature of mourning, as in [the medieval] Pearl [poem] and Chaucer’s Book of the Duchess, or the distribution of wealth and work, as in Winner and Waster and News from Nowhere) that is thus individualized. (Cowan 2007: 139)

Cowan emphasises the tension in the dream-vision form (seen in the very title of the genre) between the absolutist vision itself and the narrator’s relativist point of view: “the landscape of the dream-vision partakes of the ideal, remaining rooted all the while in the real” (2007:â•›140, my emphasis). The ‘insight’ Cowan refers to is reminiscent of Fauconnier and Turner’s ‘emergent structure’ in the blend. Indeed, as already seen, they have also used the word ‘insight’ to refer to this emergent structure: “We do not establish mental spaces, connections between them, and blended spaces for no reason. We do this because it gives us global insight, human-scale understanding, and new meaning” (Fauconnier & Turner 2002:â•›92, my emphasis). Likewise, Cowan’s phrase “remaining rooted all the while in the real” recalls Â�Turner’s observation that the blended space does not work in isolation from the Â�network: blending creates “an imaginative blended space that, however odd or even impossible, is nonetheless connected to its inputs and can illuminate those inputs. A blend can Â�produce knowledge”. It provides “full access to the input structures Â�without requiring continual recourse to them” (Turner 1996:â•›83). More recently, Turner has phrased the same idea slightly differently: “The blend typically does not obliterate the inputs. It Â�provides a human-scale, integrated scenario that serves as a conceptual anchor for the conceptual integration network.” (Turner 2008:â•›13, my emphasis). In Morris’s romance, the imaginary blended place ‘Nowhere’ is a conceptual anchor supporting a highly intricate blend network, contributing to which are many simpler, local blends. The very disparate inputs and their relations compressed in these blends include elements from the medieval past, nineteenth-century England, and Icelandic social organisation. In the hyper-blended space of Nowhere, we vicariously experience England (not an imaginary place) through Guest, but in a future strongly linked with the past, with new values and practices and a healthy and joyful fellowship of inhabitants.

4.1.1â•… Emergent structure in the hyper-blend In the hyper-blend network for NN, the overarching element of emergent structure appears to be a feeling: a conclusion – more an emotional response than a rational

 David Glyn Wilson

deduction – that the nineteenth-century capitalist version of human society was not inevitable, or at the very least was un-desirable (to reawaken the word). By contrast, life in such a society as that of Nowhere is made to feel extremely desirable. “Blended spaces are sites for central cognitive work: reasoning […], drawing inferences […], and developing emotions” (Fauconnier & Turner 1998, my emphasis). The twentieth-century socialist theoretician G.D.H. Cole testified to just this experiential effect of NN as a factor in his becoming a socialist: I became a Socialist more than fifty years ago when I read News from Nowhere as a schoolboy and realised quite suddenly that William Morris had shown one the vision of a society in which it would be a fine and fortunate experience to live.  (cited in MacCarthy 1995:â•›xvi)

The emergent structure may also include something much more revolutionary. “For the Morris of News from Nowhere in 1890, to dream of such a London [and England] had revolutionary implications” (Wilmer 1993:â•›xv). The “epoch of rest” in Nowhere had come about through revolution not reform, and revolutionary implications would, for some readers at least, be part of the emergent structure of the blend.

4.1.2â•… Variability in blending My cautious use of modality above in ‘may’ and ‘would’ takes us back to Turner’s reminder that active blending is a dynamic process. This implies that understanding and experience of blends must often vary to some extent between individuals who are constructing them (in production or reception). For practical purposes, we may not often notice such variability, but outside culturally established blends, and even within some of them, there must be degrees of variability where more original blends are concerned. Michael Reddy’s (1979) study of the ‘conduit metaphor’ of language would seem to require that blending is not uniform or entirely predictable. So in the case of a blending analysis of a literary work, such as NN, we can look at the linguistic prompts and interpret them, taking into consideration contexts of production and reception. But, when it comes to such creative blends, analysts can only suggest a range of potential blending processes that may be triggered and run for any given set of linguistic or other prompts. 4.2â•… T  extual analysis: Conceptual and material anchors for aspects of Nowhere In this section, I will examine extracts from NN which prompt for blends crucial to its structure and experiential effect. The first three extracts come from the opening, the fourth from the central section, and the last two from near the end, the climactic visit to Kelmscott in the Cotswolds. In each case, material anchors (places, objects, and people) described in the text support and stand for complex integration networks.



Iconicity in conceptual blending 

I have used underlining to highlight some of the main elements potentially prompting for blending in each extract. Of course, the effect of prose fiction often depends on the building-up of context, and in order that the ‘feeling’ generated in the blended spaces can be experienced, I have included a little textual context.

4.2.1â•… The River Thames Parrinder (1976) has discussed purposeful Dickensian echoes in the river setting of NN, referring especially to Guest’s initial boat trip on the renewed Thames as setting the tone for Morris’s tale: This boat-trip is a negative counterpart to the opening chapter of Our Mutual Friend, in which Gaffer Hexam, a predatory Thames waterman, and his daughter Lizzie are disclosed rowing on the river at dusk on an autumn evening. Southwark and London Bridges, made of iron and stone respectively, tower above them. The water is slimy and oozy, the boat is caked with mud and the two people are looking for the floating corpses of suicides which provide a regular, indeed a nightly, source of livelihood. Dickens created no more horrifying image of city life. His scavengers inaugurate a tale of murderousness, conspiracy and bitter class-jealousy. Morris’s utopian waterman, by contrast, guides his Guest through a classless world in which creativity and a calm Epicureanism flourish.  (Parrinder 1976:â•›4)

Even before Guest’s initial boat trip, the Thames is already a lightly blended space at its first mention in NN. At this point, Guest is still meant to be in Victorian London, but already the river appears to have lost the foulness (physical and moral) of Dickens’s Thames. (1) If I could but see a day of it [the new society after the revolution],” he said to himself; “if I could but see it!” As he formed the words, the train stopped at his station, five minutes’ walk from his own house, which stood on the banks of the Thames, a little way above an ugly suspension bridge. He went out of the station, still discontented and unhappy, muttering “If I could but see it! if I could but see it!” but had not gone many steps towards the river before (says our friend who tells the story) all that discontent and trouble seemed to slip off him. It was a beautiful night of early winter, the air just sharp enough to be refreshing after the hot room and the stinking railway carriage. The wind, which had lately turned a point or two north of west, had blown the sky clear of all cloud save a light fleck or two which went swiftly down the heavens. There was a young moon halfway up the sky, and as the home-farer caught sight of it, tangled in the branches of a tall old elm, he could scarce bring to his mind the shabby London suburb where he was, and he felt as if he were in a pleasant country Â�place – pleasanter, indeed, than the deep country was as he had known it.

 David Glyn Wilson

He came right down to the river-side, and lingered a little, looking over the low wall to note the moonlit river, near upon high water, go swirling and glittering up to Chiswick Eyot: as for the ugly bridge below, he did not notice it or think of it, except when for a moment (says our friend) it struck him that he missed the row of lights downstream. Then he turned to his house door and let himself in; and even as he shut the door to, disappeared all remembrance of that brilliant logic and foresight which had so illuminated the recent discussion; and of the discussion itself there remained no trace, save a vague hope, that was now become a pleasure, for days of peace and rest, and cleanness and smiling (NN, Ch. I: 44) goodwill.

It is notable here that, although Guest is still in nineteenth-century London, some aspects of the river’s renewal that he is to encounter in detail in the next chapter when he awakes in Nowhere, are already suggested. This blend seems to facilitate the slip from real world to blended vision, involving the vital relations of place, change, and identity (the place is ostensibly Victorian London, but is no longer perceived in the old “discontented” way). There is a similar portentous blend at the end of NN: amid the beauty of Nowhere, a grimy, enfeebled labourer hobbles past Guest touching his hat “with much servility” (NN: 227). Both the physical condition and the deferential attitude of this incongruous figure presage Guest’s imminent return to the inequalities of Victorian England. In each case, the incongruity in the blend prompts for an emotional response, which is enacted by Guest: in the first case, it is one of “pleasure” (NN: 44); in the second, he is “inexpressibly shocked” (NN: 228). When he wakes in Nowhere, it is no longer winter, but a bright June morning. Guest emerges from what used to be his Hammersmith home (a version of Â�Morris’s own London home named ‘Kelmscott House’): it is now called the ‘Guesthouse’ and home to a community, rather than a nuclear Victorian family – a blend clearly Â�involving vital relations of place, identity, and change, highlighting the completely new social organisation. Down by the river, he meets the boatman Dick: (2) I felt that I must make some conversation; so I pointed to the Surrey bank, where I noticed some light plank stages running down the foreshore, with windlasses at the landward end of them, and said, “What are they doing with those things here? If we were on the Tay, I should have said that they were for drawing the salmon nets; but here – ” “Well,” said he, smiling, “of course that is what they are for. Where there are salmon, there are likely to be salmon-nets, Tay or Thames; but of course they are not always in use; we don’t want salmon every day of the season.” I was going to say, “But is this the Thames?” but held my peace in my (NN, Ch. II: 47) wonder…



Iconicity in conceptual blending 

It is this revived Thames that will later lead Guest in a literal and metaphorical journey through the transformed countryside of Nowhere from London to Kelmscott. Projection of story, in this case a journey narrative with scenes of rural life along the way, enables compression to human scale of essentially complex and diffuse ecological and political arguments. Of course, argumentation and narrative are routinely conceptualised using a ‘path’ image schema (for example, Turner 1996). As blended story, ideas become not only graspable in terms of understanding, but can also generate emotions (Fauconnier & Turner 1998).

4.2.2â•… N  ineteenth-century coins: Decompressing the culturally established money blend In extract (3), Dick peers “with some curiosity” (NN: 49) at the crudely-formed and strangely oxidised nineteenth-century coins Guest offers in payment for his Â�ferrying services. The culturally-established blend of money, key to capitalist culture, is being decompressed in this passage. Instead of it being taken for granted, Dick tries to understand money as “something which I am not to give to a neighbour […] a troublesome and roundabout custom; and we don’t know how to manage it” (NN: 49–50). This old custom is portrayed as anti-social and even pointless. In this passage, the material anchor is a negative one, to be decompressed and exposed as representing a set of Â�self-centred practices. Here, two stories are projected into the blend – one in which money is taken for granted as the basis of social relations, the other in which it no longer even exists (except in museums): the emergent structure is initially a lack of recognition, an unease, and finally a feeling that the whole history of money has led to the current undesirable system of social relations. The repudiation of the system of money, with its material anchors of coins, is here iconic of a radically new social order: (3) I think I know what you mean. You think that I have done you a service; so you feel yourself bound to give me something which I am not to give to a neighbour, unless he has done something special for me. I have heard of this kind of thing; but pardon me for saying, that it seems to us a troublesome and roundabout custom; and we don’t know how to manage it. And you see this ferrying and giving people casts about the water is my business, which I would do for anybody; so to take gifts in connection with it would look very queer. Besides, if one person gave me something, then another might, and another, and so on; and I hope you won’t think me rude if I say that I shouldn’t know where to stow (NN, Ch. II: 50) away so many mementos of friendship.

The irony in the final line of the extract arises from faulty decompression of the blend. Dick, representing in his innocence the new vision of work and the production and sharing of goods, is unable to reconstruct the conceptual integration network that led

 David Glyn Wilson

to the acceptance of money in the first place: he assumes that the only use for money would be as tokens of friendship. How could it possibly determine human �relationships and human fates? The thought does not occur to him.

4.2.3â•… Trafalgar Square Extract (4) again appeals to feeling, centring on Guest’s “strange sensation” of Â�observing a placid orchard and suddenly finding himself projecting upon that scene the Â�prelude to what became “known as ‘Bloody Sunday’, [when] Trafalgar Square was the scene of the most ruthless displays of establishment power that London has ever seen” Â�(MacCarthy 1995:â•›567). Trafalgar Square had been banned as a place for public meetings, and in extract (4) Guest apparently throws his mind back to the events of 13 November 1887 – events which Morris had experienced at first hand, as he had been marching with the Socialist League and thousands of other political protesters towards the Square when the police, who had cordoned it off, brutally attacked along with Â�cavalry and foot Â�soldiers, leaving hundreds injured. (4) We came presently into a large open space, sloping somewhat toward the south, the sunny site of which had been taken advantage of for planting an orchard, mainly, as I could see, of apricot-trees, in the midst of which was a pretty gay little structure of wood, painted and gilded, that looked like a refreshmentstall. From the southern side of the said orchard ran a long road, chequered over with the shadow of tall old pear trees, at the end of which showed the high tower of the Parliament House, or Dung Market. A strange sensation came over me; I shut my eyes to keep out the sight of the sun glittering on this fair abode of gardens, and for a moment there passed before them a phantasmagoria of another day. A great space surrounded by tall ugly houses, with an ugly church at the corner and a nondescript ugly cupolaed building at my back; the roadway thronged with a sweltering and excited crowd, dominated by omnibuses crowded with spectators. In the midst a paved Â�Â�befountained square, populated only by a few men dressed in blue, and a good many singularly ugly bronze images (one on the top of a tall column). The said square guarded up to the edge of the roadway by a four-fold line of big men clad in blue, and across the southern roadway the helmets of a band of horse-soldiers, dead white in the greyness of the chilly November afternoon – I opened my eyes to the sunlight again and looked round me, and cried out among the whispering trees and odorous blossoms, “Trafalgar Square!” “Yes,” said Dick, who had drawn rein again, “so it is. I don’t wonder at your finding the name ridiculous: but after all, it was nobody’s business to alter it, since the name of a dead folly doesn’t bite. Yet sometimes I think we might have given it a name which would have commemorated the great battle which was fought on the spot itself in 1952, – that was important enough, if the historians don’t lie.”  (NN, Ch. VII: 77–78)



Iconicity in conceptual blending 

For contemporary readers familiar with the recent events of November 1887 – and particularly for those who had been personally involved – the date 1952 must have come as something of a shock, since Morris’s preceding description fits the prelude to the events of 1887 to a tee. Let us consider what Morris, as one personally implicated in the 1887 demonstration, may be attempting here. The vital relations compressed in this blend predominantly concern time, place, identity, and change. Not only is the orchard the same physical space where the attack took place in 1887, it even still bears the same name in 2102. However, the times of the orchard and the attack are different and great change has occurred, so that in Guest’s fleeting moment of recognition of the identity of the old and new Trafalgar Square there are powerful compressions of vital relations. The blend may well, however, be more complex still. Alluding to Dick’s remarks in extract (4) above, MacCarthy comments that “the events of ‘Bloody Sunday’ have been brought forward to 1952 [referring to the fictive revolution that brought about the new order]” (MacCarthy 1995:â•›569). As the nineteenth-century Guest could not have seen 1952, this seems an odd comment at first sight, and it would seem more straightforward to interpret this passage as Guest simply recalling the Bloody Sunday of 1887 (and indeed the 1887 freedom of speech march is referred to in the following paragraph). It is not entirely clear quite what MacCarthy means by “brought forward”, as she does not elaborate. But the notion of some kind of relation between 1887 and 1952 is worth investigating. It seems plausible that Morris is prompting in the blend for an even tighter compression of the relation of identity into one of uniqueness – one vital relation can often be compressed into another (Fauconnier & Turner 2002:â•›314). Thus, the prelude to the routed Bloody Sunday demonstration is compressed into a relation of uniqueness with the prelude to the fictive revolution of 1952. Note that Guest imagines a very static and unresolved preliminary scene; he does not call up an image of the 1887 protesters being savagely beaten and defeated by the establishment forces – and if the aim were merely to contrast the ugliness of 1887 with the beauty of 2102, it might be supposed that a vivid description of the historically attested violence would be a forceful way of achieving such a contrast. If compression to uniqueness is a plausible analysis here, then yet another vital relation, cause–effect, is compressed in the blend, strengthening the causal connection between the grim “phantasmagoria of another day” (i.e. the continuous and ultimately successful socialist struggle) and the resultant sunny orchard. In this context, it is instructive to recall that Morris had written A Dream of John Ball in 1888 just prior to NN (1890). As mentioned earlier, A Dream of John Ball links the failed Peasants’ Revolt of 1381 to the ongoing nineteenth-century Â�socialist Â�struggle. The narrator in A Dream of John Ball travels back in time in his dream and looks Â�forward from the fourteenth century in his exchanges with the priest John

 David Glyn Wilson

Ball, one of the leaders of the unsuccessful revolt. Thus MacCarthy’s interpretation of Guest’s Trafalgar Square vision in NN as incorporating allusion to an imagined 1952 victory, as well as to the 1887 rout, seems in keeping with Morris’s acute sense of the interconnectedness of history, of the relation of the past to the present and to the future. Indeed, the relation here is quite complex, 1952 being ‘past’ in the story of Nowhere but ‘future’ for its readers. Note, furthermore, that as Guest comes to after his vision, Dick speaks explicitly of the revolution of 1952. Guest’s vision, however, without excluding consideration of the 1952 revolution, appears to evoke equally the demonstration of 1887. Indeed, Guest puts no exact time on his vision, Â�referring Â�simply to a “phantasmagoria of another day”, without specifying or even hinting which day. In fact, at no point is this vision explicitly identified with 1887. It is possible that Morris intends Guest, in his limited perspective of ‘dreamer in need of enlightenment’, to be aware only of 1887 in his vision of Trafalgar Square. Â�Alternatively, and more likely, the strangeness of this unresolved vision floating vaguely in time may indicate that Guest (or at least the reader) is supposed to be Â�subliminally aware of Â�experiencing more than a simple memory of the events of 1887. The Â�simultaneous evocation – indeed the blending and compression – of 1887 and 1952 is made entirely possible through the indeterminate description of the eerily static and only potentially revolutionary scene. It seems to be suspended in time. In this way, allowing for Â�compression of the two events from the same physical space – one Â�historically real and one hoped for – into a relation of uniqueness is a powerful way for Â�Morris to evoke an emotional response to what he believed, at least in his more optimistic moments, to be the inevitability of socialist change through a Â�continuous historical process. Just imagine how you would feel if you had been involved in, or Â�concerned by, the discouraging events of 1887 and you now read that they are Â�assimilable to a Â�historical process leading to eventual victory. I think this is what Â�Morris was aiming at – to encourage, through arousing positive, hopeful emotions towards the Â�socialist struggle. (It also needs to be remembered that NN was first serialised with overt Â�propagandist intent in the socialist newspaper The Commonweal.)3 Fauconnier and Turner (2002:â•›312) list “come up with a story” as one of the Â�sub-goals of compression in the blend. In extract (4), we see the power of story to make sense of a long and hugely complicated and tragic series of events and their possible resolution. It is worth, in this context, recalling the words of Thomas Newman quoted earlier, cited by Groves (2009) in the BMJ, but this time focusing on the second part of the quotation: “The brains of human beings seem built to process stories better than

.â•… Thanks are due to Professor Olga Fischer for a query concerning the extent of the Â�compression in the Trafalgar Square episode. This led to the inclusion of the discussion of A Dream of John Ball and Morris’s sense of the interconnectedness of history.



Iconicity in conceptual blending 

other forms of input […] they seem to offer a solution – a way to extract some Â�meaning and redemption from tragedy by preventing its reoccurrence” (Groves 2009:â•›b4887, my emphasis). Because the inputs to this tightly compressed blend remain accessible, we are able to use the blend to reassess these inputs: in this particular case, to Â�reassess the political and social organisation of (Victorian) England and, in Morris’s ideal case, hope to prevent the reoccurrence of events such as Bloody Sunday by instead Â�mounting a successful revolution.

4.2.4â•… An old house amongst new folk:4 Kelmscott Manor and Ellen It was mentioned earlier that blending performs “central cognitive work: reasoning […], drawing inferences […], and developing emotions” (Fauconnier & Turner 1998, my emphasis). This is again clearly evident in NN in the final pages. Naomi Jacobs has drawn attention to Morris’s peculiar view of beauty and the human body: A radical reconceptualization of beauty and of the body was central to [Morris’s] maturation from aesthetic dreaming to active work on behalf of social change. Ultimately, Morris’s recognition of the importance of embodiment – what he sometimes calls ‘animal life’ – necessitates a conception of beauty as not merely a quality to be perceived or appreciated in the material world, but as a quality to be (Jacobs 1997:â•›26) passionately experienced in the physical self.

The following two passages from NN embody these views in story form; indeed, they are literally embodied in the character of Ellen. Kelmscott is the goal of the journey – both the river journey and the spiritual journey. To see Kelmscott Manor was Ellen’s reason for joining the expedition up-river. (5) Once again Ellen echoed my thoughts as she said: “Yes, friend, this is what I came out for to see; this many-gabled old house built by the simple country-folk of the long-past times, regardless of all the turmoil that was going on in cities and courts, is lovely still amidst all the beauty which these latter days have created; and I do not wonder at our friends tending it carefully and making much of it. It seems to me as if it had waited for these happy days, and held in it the gathered crumbs of happiness of the confused and turbulent past”. (NN, Ch. XXXI: 220)

In extract (5), compression achieved in the blend sets up this ancient Tudor house as a material anchor for a blend involving relations of time, place, and identity, Â�telescoping widely separated ages and making their relations comprehensible on the human scale of a house that has not only survived but been integrated into a future happiness. The house then represents the idea that in the midst of highly unpromising

.╅ The title of NN Chapter XXXI.

 David Glyn Wilson

Â� conditions (“the turmoil”), it is not unreasonable to hope when looking to the future. More importantly, Kelmscott Manor is a material anchor for the emotional experience of this idea. (6) She led me up close to the house, and laid her shapely sun-browned hand and arm on the lichened wall as if to embrace it, and cried out, “O me! O me! How I love the earth, and the seasons, and weather, and all things that deal with it, and all that grows out of it, – as this has done!” I could not answer her, or say a word. Her exultation and pleasure were so keen and exquisite, and her beauty, so delicate, yet so interfused with energy, expressed it so fully, that any added word would have been commonplace and futile. I dreaded lest the others should come in suddenly and break the spell she had cast about me; but we stood there a while by the corner of the big gable of the house, and no one came. I heard the merry voices some way off presently, and knew that they were going along the river to the great meadow on the other side of the house and garden. We drew back a little, and looked up at the house: the door and the windows were open to the fragrant sun-cured air; from the upper window-sills hung festoons of flowers in honour of the festival, as if the others shared in the love for (NN, Ch. XXXI: 220–221) the old house.

In extracts (5) and (6), postponement of action, and indeed a highly charged Â�inactivity, is iconically reflected in the slightly archaic language, which takes its time and is longer than functionally needed for contrasting the activity of “the others” with the intense inactivity of Guest and Ellen. There is an energy and a stillness in the blend of Ellen, and of course in Morris’s energetic “epoch of rest” in general, where “all work is now pleasurable … because there is conscious sensuous pleasure in the work itself ” (NN: 122–123). Guest has, of course, fallen in love with Ellen. “When Bruce Glasier told him he had fallen in love with Ellen, Morris replied ‘that he had fallen in love with her himself!’” (MacCarthy 1995:â•›588). Morris’s marriage with Jane Burden, one of the Â�Pre-Raphaelites’ favourite models, was not a success, though they remained together, Morris stoically accepting her affair with Rossetti. Jane was said to be “silent, Â�languorous, frequently unwell or supposedly unwell” (Thompson 1977:â•›75). In the light of this, and his feelings on beauty and the body, his invention of the blend of Ellen – “feminine and boyish, sexual and yet almost asexual, intelligent and animal” (MacCarthy 1995:â•›588) – makes her a fitting counterpart to the old Tudor house: two linked blends at the climax of the tale standing as material anchors for Morris’s Â�personal and political hopes and dreams, but also crucially for what he wished to be thought of not as a dream, but as a vision.



Iconicity in conceptual blending 

5.â•… Conclusion Conceptual integration involves the projection of multiple stories into a blended Â�mental space. In this space, a new blended story integrates elements of each of the Â�projected inputs, yielding insights, conclusions, and emotions. Through compression of vital relations, the blend stands as a graspable conceptual anchor for an Â�integration Â�network. This compression puts the blend in an iconic relationship to the highly Â�complex, more diffuse integration network. In practical terms, iconic blends allow us to function efficiently in the world; they give us the means to tell blended stories to yield new meaning and emotional experience. In the hyper-blend of News from Nowhere, many conceptual blends with material anchors provide iconic threads and moments in the narrative, of which the Thames, coins, Trafalgar Square, Ellen and Kelmscott Manor have been considered in this study. However, it seems at present unclear exactly how non-metaphorical blends should be represented in a general theory of iconicity. Examples such as those discussed in this study show the existence of iconic phenomena that could arguably belong in the “diagrammatic – semantic” category in Nänny and Fischer’s (1999:â•›xxii) “types of Â�iconicity” figure. However, under the “semantic” branch of that category there is only “metaphor” (cognitive and grammatical), whereas the blends discussed in this study are iconic but not metaphorical. Refining our understanding of the relations between blending and iconicity would seem to be a worthwhile next step in exploring these two important areas of cognition.

References Brône, G. & Valdaele, J. (eds). 2009. Cognitive Poetics: Goals, Gains, Gaps. Berlin: Mouton de Gruyter. Cowan, Y. 2007. ‘Paradyse Erthly’: John Ball and the medieval dream-vision. In Writing on the Image: Reading William Morris, D. Latham (ed.), 137–153. Toronto: University of Toronto Press. Fauconnier, G. 2007. Mental spaces. In Geeraerts & Cuyckens (eds), 351–376. Fauconnier, G. & Turner, M. 1998. Blending as a Central Process of Grammar. 〈http://Â�markturner. org/centralprocess.WWW/centralprocess.html〉 (2 August, 2008). Fauconnier, G. & Turner, M. 2002. The Way We Think: Conceptual Blending and the Mind’s Â�Hidden Complexities. New York NY: Basic Books. Freeman, M.H. 2007. Cognitive linguistic approaches to literary studies: State of the art in Â�cognitive poetics. In Geeraerts & Cuyckens (eds), 1174–1202. Freeman, M.H. 2009. Minding: feeling, form, and meaning in the creation of poetic iconicity. In Brône & Valdaele (eds), 169–196.

 David Glyn Wilson Geeraerts, D. & Cuyckens, H. (eds). 2007. The Oxford Handbook of Cognitive Linguistics. Oxford: OUP. Groves, T. 2009. Editor’s choice. The power of stories. BMJ 339: b4887. Herlofsky, W.J. 2003. What you see is what you get: Iconicity and metaphor in the visual Â�language of written and signed poetry: A cognitive poetic approach. In From Sign to Signing Â�[Iconicity in Language and Literature 3], W.G. Müller & O. Fischer (eds), 41–61. Amsterdam: John Benjamins. Hiraga, M.K. 2005. Metaphor and Iconicity. Houndmills: Palgrave Macmillan. Jacobs, N. 1997. Beauty and the body in News from Nowhere. Journal of the William Morris Society 12(2): 26–30. MacCarthy, F. 1995. William Morris: A Life for Our Times. New York NY: Alfred A. Knopf. Morris, W. 1994. [1890]. News from Nowhere and Other Writings. C. Wilmer (ed.). Hammondsworth: Penguin Classics. Morris, W. 2003 [1868]. The Earthly Paradise: Apology. In Representative Poetry Online, I.  Lancashire (ed.). 〈https://tspace.library.utoronto.ca/html/1807/4350/poem1502.html〉 (Accessed May 13, 2010). Nänny, M. & Fischer, O. (eds). 1999. Form Miming Meaning [Iconicity in Language and Â�Literature 1]. Amsterdam: John Benjamins. Ortony, A. (ed.). 1979. Metaphor and Thought. Cambridge: CUP. Parrinder, P. 1976. News from Nowhere, The Time Machine and the break-up of classical realism. Science Fiction Studies 3(3). 〈http://www.depauw.edu/SFs/backissues/10/parrinder10art. htm〉 (August 4, 2009). Reddy, M. 1979. The conduit metaphor: A case of frame conflict in our language about language. In Ortony (ed.), 284–324. Schön, D. 1979. Generative metaphor: A perspective on problem-setting in social policy. In Ortony (ed.), 254–283. Thompson, E.P. 1977. William Morris: Romantic to Revolutionary. London: Merlin Press. Turner, M. 1996. The Literary Mind: The Origins of Thought and Language. Oxford: OUP. Turner, M. 2007a. Conceptual integration. In Geeraerts & Cuyckens (eds), 377–393. Turner, M. 2007b. The way we imagine. In Imaginative Minds, I. Roth (ed.), 213–236. London: OUP & The British Academy. Turner, M. 2008. Frame blending. In Frames, Corpora, and Knowledge Representation, R. Rossini Favretti (ed.), 13–32. Bologna: Bologna University Press. Wilmer, C. (ed.). 1993. William Morris: News from Nowhere and Other Writings. London: Â�Penguin Books. (Edited with an introduction and notes by C. Wilmer.)

part v

Acoustic and visual iconicity

Thematized iconicity and iconic devices in the modern novel Some modes of interaction John J. White

King’s College London The following paper offers case studies of forms of thematized iconicity in the modern realist novel. Taking its cue from Eco and Sebeok’s The Sign of Three (1983), it explores some of the ingenious, yet plausible, ways in which acoustic and visual iconicity have played a substantial role in modern fiction. Examples of acoustic iconicity and indexicality are explored in David Lodge’s Deaf Sentence (2008), whereas visual iconicity is shown to be of overarching importance in Umberto Eco’s The Mysterious Flame of Queen Loana (English translation: 2005). Emphasis is placed on the importance of a self-deprecating satirical perspective in Deaf Sentence and on the way in which copious intertextual material and visual images are used to defamiliarize the historical context of Eco’s novel. Readerresponse is considered in the case of narrated iconicity in both works. The paper is predicated on the working hypothesis that thematized iconicity deserves serious consideration as a literary topos in its own right.

This paper responds to the fact that thematized iconicity has been neglected in recent Iconicity Studies despite its importance as a plot-ingredient in modern fiction from Vladimir Nabokov’s Signs and Symbols (1948) to David Lodge’s Deaf Sentence (2008). While Umberto Eco’s The Name of the Rose (1983) still remains literature’s best-known “semiotic novel” by virtue of its protagonist’s ingenious interpretation of indexical clues, iconicity is also thematized at various junctures in the work. I argue that while this also generally holds true for Eco’s subsequent fiction, his “illustrated” novel, The Mysterious Flame of Queen Loana (2005), being an exception on account of its prolific combinations of thematized iconicity and supporting visual material. The main sections of the present paper explore selected aspects of the interplay between thematized iconicity and iconic devices in Deaf Sentence and Queen Loana. The structural and narratological implications of such interaction are considered with the reference to C. S. Peirce’s conception of iconicity and indexicality as hybrid sign-aspects.

 John J. White

1.â•… Introduction: A lacuna in recent studies of iconicity in fiction There is a distinct imbalance in the field of literary semiotics between the treatment of iconic and indexical material. Those studies concerned with iconicity have tended to concentrate on iconic patterns and devices, while generally neglecting Â�iconicity’s Â�thematic role in fiction. At the same time, a number of our colleagues, including Umberto Eco, Thomas A. Sebeok, Jean Umiker-Sebeok, David H. Richter (1986) and Rocco Capozzi (1997) , have over the years produced impressive major studies of fiction’s Â�thematization of Â�indexicality. At the founding Iconicity in Language and Literature symposium in Â�Zurich, I gave a paper (White 1999) on iconic aspects of prose and poetry, together with some related evidence of iconicity’s occasional importance as a plot-element. However, I failed on that occasion to distinguish adequately between iconicity as method and its constitutive role as theme. What follows is a belated attempt at a corrective. Talk of thematized iconicity might at first seem like a deliberate challenge to the Iconicity Research Project’s two-pronged definition of iconicity as either “form Â�miming meaning” or “form miming form”. My intention with this paper, I should therefore stress, is simply to draw attention to a recurrent phenomenon in modern fiction which work on literary iconicity has tended to underestimate. The present paper explores examples of thematized iconicity alongside iconic devices per se, and at times looks even at instances of thematized indexicality. Nevertheless, I would want to argue that thematized iconicity also deserves serious consideration as a literary factor in its own right. Umberto Eco would seem to agree: “After having let semiotics speak abundantly about texts”, he once wrote, “it is [only right] to let a text speak by itself about its semiotic strategy” (Eco 1979:â•›40).1 If this originally sounded like a trailer for his novel The Name of the Rose (published the year after the above comment was made in The Role of the Reader), I should like to point out that Eco’s position on this question remains relevant to other, more recent works of fiction by him, including The Mysterious Flame of Queen Loana (Eco 2005), material from which will be considered later on. If one were to ask any group of people with a serious interest in contemporary literary semiotics which sign in Charles Sanders Peirce’s Second Trichotomy most Â�frequently occurs as either a theme or a significant plot-element in modern fiction, most of them would probably point to the index. Not because Peirce, during one of his rare excursions into literary territory, chose to demonstrate some of the defining Â�features of indexicality by analysing the various interpretive possibilities of Man

.â•… The reference to Eco’s further “fabulation of semiotic […] strategies” in the fiction Â�following The Name of the Rose (Capozzi 1997:â•›230) also allows for the kinds of thematized iconicity and indexicality discussed below in the case of The Mysterious Flame of Queen Loana. Cf. also Bondanella 1997:â•›101.



Thematized iconicity and iconic devices in the modern novel 

Â� Friday’s footprint in the sand, as discovered on his desert island by Robinson Â�Crusoe.2 More likely, it would be due to the generous coverage indexicality had received in The Sign of Three (Eco & Sebeok 1983), an important compendium of essays on the importance for criminal detection of indexical evidence and associated processes of what Peirce called ‘abduction’. The works of fiction discussed in The Sign of Three, from Defoe’s Robinson Crusoe and Â�Voltaire’s Zadig to Poe’s Murders in the Rue Morgue and Conan Doyle’s Sherlock Holmes stories, all contain ingenious illustrations of indexicality’s pivotal role at plot-level in detective fiction, a feature usually commented on (or, to put it another way, vicariously thematized) by the narrator or in the latter instances in exchanges between the detective-protagonist and an admirer.3 During the second half of the twentieth century, the detective novel had thus for obvious reasons become the preferred hunting-ground for those interested in thematized semiosis in general and the role played by narrated indexicality in particular. There are other putative Â�reasons for thematized indexicality’s paramount semiotic status in recent literary research. Indexical material of the kind Â�discussed in The Sign of Three and in some chapters of Eco’s The Role of the Reader readily lends itself to the construction of intricate, yet as a rule largely plausible, plots, whereas iconicity is notoriously elusive, protean and often difficult to thematize Â�convincingly. To compound matters, understandably few fictional characters in modern literature show much interest in iconicity – either in the language domain or in other contexts. The academic protagonist of the first novel I shall look at is, fortunately, a rare exception. Although discourse analysis and corpus linguistics are two of his specialist areas of academic interest, he also alludes knowledgeably, if not always enthusiastically, to other branches of modern, post-Chomskyan linguistics. While it would be rash to claim representative status for my present, rather restricted illustrative material, the two works discussed in the main sections of this paper do arguably offer insights into some of the more ingenious guises thematized iconicity has assumed in modern fiction. Both Deaf Sentence and The Mysterious Flame of Queen Loana are in their different ways comic novels, thereby suggesting that literary semiosis can be more fun than the work of certain heavyweight semioticians would lead us to believe. The extent to which the amusement factor is a product of thematization will concern us later.

.â•… “That footprint that Robinson Crusoe found in the sand […] was an Index to him that some creature was on his island, and at the same time as a Symbol, called up the idea of a man” (CP 4:â•›531). Surprisingly, the element of iconicity also involved in Crusoe’s response is nowhere invoked by Peirce, although it is central to Eco 1983a. .â•… Eco laid the foundations for some parts of The Sign of Three in “Rhetoric and Ideology in Sue’s Les Mystères de Paris” and “Narrative Structure in Fleming” (Eco 1979:â•›125–43 and 144–72, respectively).

 John J. White

2.â•… Dysfunctional acoustic semiosis: David Lodge’s Deaf Sentence The following sequence of illustrations from David Lodge’s widely acclaimed Deaf Â�Sentence focuses on the often underestimated sub-category of acoustic Â�iconicity. Although the well-known English author and academic David Lodge (since Â�retirement Honorary Professor of Modern English Literature at the University of Birmingham) himself suffers, like the novel’s protagonist, from a substantial hearing deficit, a Â�postscript makes it clear that Deaf Sentence is not to be read in any Â�narrowly Â�biographical spirit.4 Initially the work looks rather like just one more variation on the once fashionable ‘campus novel’ genre with which the twinned reputations of David Lodge and the late Malcolm Bradbury are invariably associated. However, in this case there is an added twist: the work’s hero Desmond Bates, a former Professor of Applied Linguistics at an unnamed university in an unspecified northern English city, has already taken what nowadays tends to be euphemistically referred to as ‘voluntary’ early retirement from his institution. The term is crassly misleading. He was in fact forced to do so both by his hearing handicap and the re-assignment of his Department of Linguistics to the English Department during an administrative re-shuffle justified, as usual these days, ‘on managerial grounds’. Desmond Bates unfortunately comes to regret having opted for retirement. He soon begins to miss the familiar comforts of an academic Â�routine, the institutional companionship that he had hitherto taken for granted, as well as the Â�reassuring sense of being intellectually valued in his specialist field. Despite Â�initial appearances to the contrary, the unsettling predicament of isolation that retirement imposes on him is presented as having much in common with suffering from a Â�progressively deteriorating inability to hear what people try to say to him. Desmond Bates, we learn at the outset, suffers from high-frequency deafness: he cannot hear consonants, only vowels.5 Lodge’s novel offers a vivid, often moving, account of what it feels like to be afflicted by extreme deafness in the latter years of one’s life. In the protagonist’s particular case, it is a predicament compounded by repeated reminders that he is becoming progressively senile, even in the eyes of his nearest and dearest. The novel’s theme of mortality is to some extent displaced onto Bates’ father (always sympathetically referred to as “Dad”), a tragi-comic alter-ego figure living mostly in the past, afflicted by paranoia

.â•… In his concluding Acknowledgements, the author stresses that “the narrator’s deafness and his Dad have their sources in [his] own experience, but the other characters in the novel are fictional creations, as is the nameless northern city” (Lodge 2008:â•›293). .â•… I am grateful to William J. Herlofsky (Nagoya Gakuin University, Japan) for confirming in discussion that the various technical details in Lodge’s novel add up to a scientifically accurate portrayal of deafness.



Thematized iconicity and iconic devices in the modern novel 

and the early symptoms of senile dementia. “It’s as if he’s presenting to me a Â�mocking effigy of myself […] in dire straits twenty-odd years from now”, Bates Â�complains after one of his painful, dutifully filial visits to his father (Lodge 2008:â•›47). The Â�literary Â�landscape of the past two decades is, of course, littered with geriatric protagonists Â�pondering their senility. Witness the recent works of fiction treating infirmity and old age by Julian Barnes, Saul Bellow, Nick Hornby, Howard Jacobson, Marina Lewicka, Philip Roth and John Updike. Yet deafness plays a less prominent role in such works than one might expect. Even when it is a factor, the debility is seldom the “kind of Â�pre-death” evoked in Deaf Sentence (Lodge 2008:â•›19). Three parallel narrative strands, reinforced by a series of extended metaphors, structure the unfolding narrative in Deaf Sentence: 1. the insinuation that deafness is ‘a kind of pre-death’ 2. the bleak picture of retirement as virtual academic brain-death and 3. the presentation of Bates senior’s isolation, paranoia and incipient senile dementia as a parallel caricature-version of the loneliness that in his son’s case is essentially a consequence of hearing difficulties. The prominent role played by analogy and various extended metaphors in Lodge’s Deaf Sentence is not immaterial to my present reading, especially in the light of Peirce’s (opaquely presented) conception of metaphor as a sign “which represents the representative character of a representamen [i.e. another sign] by representing a parallelism in something else” (CP 2: 277).6 What Peirce posits as the common ground between any metaphor’s two contributory signs effectively lends itself to thematization at various junctures in the present novel, particularly given that each metaphor-strand displays an iconic “parallelism” (to use Peirce’s term) to the other two. The iconic element in this case is therefore endophoric (i.e. ‘sign miming sign’), as it generally tends to be in Peirce’s thinking on metaphor.7 What complicates the picture is the embarrassing way in which the Â�protagonist’s disability creates situations that usually result in a grotesque mismatch between Â�people’s actual words and what he wrongly assumes they said. (Neither retirement nor old age necessarily causes extreme problems of the kind that plague the Â�present novel’s protagonist.) If this acoustically iconic feature is to receive the attention it deserves, many sequences in Lodge’s comic novel are best appreciated when read aloud.

.â•… For a less rebarbative explanation of this definition and its semiotic implications, see Â�Johansen and Larsen (2002:â•›40–43). Cf. also “The semantics of metaphor” (Eco 1979:â•›67–89). .â•… For a linguistic account of exophoric iconicity in language, see Nöth 2001:â•›22–23.

 John J. White

Desmond Bates’ areas of academic specialization (applied linguistics, discourse analysis and corpus linguistics) have a marked influence on the way he views his increasing deafness, as well as on many of the defensive strategies he adopts, when occasionally obliged to appear in public. He sometimes talks endlessly at people, thus giving them little opportunity to reply, while generally trying to avoid situations where the loud buzz of other people’s conversation makes comprehension in closed spaces virtually impossible. He goes to the Quiet Carriage on trains, yet still switches off his hearing aid in order to insulate himself from other people’s intrusive mobile-phone conversations. He has a whole arsenal of other ingenious self-camouflaging devices which he deploys as part of an overall ‘belt and braces’ strategy. When trapped by his interlocutor into having to reply to a remark, he tends to resort to token signals of agreement (“precisely”, “I agree”, or just a nod of the head) to imply that he has understood and agreed with a remark, although obviously he has not. When suspicions are eventually aroused, he has a habit of feigning interest in some irrelevant side-attraction in order to deflect any further attempt at invading his silent territory. Ultimately, when deafness becomes intolerable, he agrees to follow his wife’s advice and starts to attend lip-reading classes. One such class is recalled in the concluding paragraph of Deaf Sentence: We had a session on homophones which could cause misunderstanding, for example, married and buried, wet suit and wedding suit, big kiss and biscuits. […] Members volunteered their own stories of misunderstandings. Marjorie was asked at the supermarket checkout if she would like a “free gateau” and eagerly accepted the offer, which turned out to be a “free catalogue”. Violet was baffled when her friend enthused about “laxative porridge”, which turned out to be “wax-free polish”. I told my story of the “long-stick saucepan”. We had a short talk on towers. […] The Eiffel Tower was built as a temporary structure for the 1889 Paris Exhibition, and was much criticized at the time. It was supposed to be demolished afterwards, but the populace became fond of it and it was saved when a wireless transmitter was put on the top. I always learn something new at the lip-reading class. (Lodge 2008:â•›291)

Homophones that could cause a misunderstanding, the topic chosen on this Â�particular occasion to encourage group-discussion, is something in which Bates immediately shows a professional interest, even though the subject revives painful memories.8 The phrase “long-stick saucepan” harks back to an embarrassing occasion (Lodge 2008:â•›78)

.â•… The types of misunderstandings cited in this passage are the very opposite of the Â�conditions identified in the section on “Iconicity as a prerequisite of mutual understanding (Nöth 2001:â•›26–27).



Thematized iconicity and iconic devices in the modern novel 

when he had misunderstood his wife’s request to pass her a “non-stick frying pan”, and had as a result become an object of family ridicule from then on. A similar misunderstanding occurs elsewhere in the novel, but at a time when – perhaps because it is the festive season – Desmond Bates is able to cope with his Â�difficulty in a more relaxed frame of mind: “So what did Father Christmas bring you, Daniel?” I asked him, crouching down to bring myself to his level. “Father Christmas bringed Daniel an icicle”, he said. “An icicle? That doesn’t sound like much of a present”, I said. “A tricycle, Desmond”, Marcia said, and everybody around us laughed. One thing we deafies can do at a party is give people a few laughs with our mistakes, and I did not begrudge them this one. Daniel, however, didn’t laugh, but turned his wide eyes on the grinning grown-up faces with a puzzled and faintly disapproving air. “And it’s ‘brought’, not ‘bringed’, Daniel”, his mother added. “Father Christmas brought you a tricycle”. Being a teacher, Marcia thinks it her duty to correct her children at every opportunity. Of course Daniel’s mistake was perfectly logical and shows that he has already mastered the way to form the past tense of regular English verbs. You’ve got to grasp the rules before you learn the exceptions.  (Lodge 2008:â•›177, abridged)

Although some of the lip-reading group’s anecdotal recollections of homophones likely to cause confusion are, technically speaking, textbook examples of ‘minimal pairs’,9 the term is studiously avoided. It would certainly not be part of the group’s vocabulary; Â�perhaps the noun ‘homophone’ was only used by the person Â�organizing the discussion in order to impress those present. In all probability, the few other Â�technical terms that occur in the lip-reading class described at the end of the novel are the Â�protagonist-narrator’s, not the discussion leader’s. In his narrating capacity, Desmond Bates uses comparable linguistic jargon when admitting to having “played on the phonetic near-equivalence of the homophones ‘death’ and ‘deaf ’” throughout his narrative (Lodge 2008:â•›289) and during his occasional forays into corpus Â�linguistics in order to sidestep a current Â�embarrassing ethical predicament brought about by the repeated attentions of one Â�particularly determined female research Â�student (Lodge 2008:â•›232–33.). Double iconicity operates in the case of the repeated word-play on ‘deaf ’ and ‘death’, for just as the two nouns are similar, so too for Bates are the two conditions to which they refer. There is one important rider, however.

.â•… Minimal pairs are two words that are completely similar in sound with the exception of one phoneme, e.g. ‘rook’ and ‘look’. The term is used in order to distinguish phonemes from allophones. Only phonemes make a meaningful distinction between words, allophones do not. For example a word like ‘cool’ pronounced with a palatal [k], as found in ‘keep’, will not make a difference in English, it is still the same word as ‘cool’ pronounced with a regular velar [k]. In other languages palatal and velar [k] can serve to make a distinction in meaning.

 John J. White

According to his way of thinking, arrived at on the basis of personal experience, “deafness is Â�axiomatically comic”, i.e. to the non-deaf observers of a deaf person’s behaviour (Lodge 2008:â•›13). Yet in the majority of his anecdotal accounts of his recent difficulties as a “deafie” (one of his favourite terms of self-deprecation), it is usually both mortifying and at the same time uncomfortably comic, the latter thanks to the particularly amusing way in which incidents are recalled. Being able to present his deafness from a more disinterested linguistic Â�perspective usually affords Desmond Bates the distance he needs to be able to rise above his Â�embarrassing predicament. For a similar reason, the individual chapters of Deaf Â�Sentence are told alternately in either the first or the third person, with the protagonist thus controlling the picture of his disability by presenting from the outside as well as subjectively experienced. One memorable feature of the more distanced, Â�protective Â�perspective are the various parodistic literary games Bates repeatedly plays with the minimal pair death/deaf. He comes up with such spoof literary titles and mock Â�quotations as Deaf in the Afternoon (pace Hemingway), “Deaf, where is thy sting?” (St Paul), “After the first deaf, there is no other” (Dylan Thomas), “To every man upon this earth, deaf cometh soon or late” (Macaulay) etc. (Lodge 2008:â•›19). He also seems to be self-mockingly aware that he shares the name Bates with a Â�similarly Â�burdened fictional character in Jane Austen’s Emma, though this time the practical joke is the author’s rather than part of a narratorial game. By narrating – or Â�thematizing – Â�deafness and death in this way, Bates deploys his sense of humour as a protective shield against advancing debility and old age itself. As for the Â�function of the novel’s Â�anecdotally ambiguous homophones and minimal pairs, these can be read within a semiotic Â�framework as amusing examples of endophoric iconicity. Seen from the Â�protagonist’s viewpoint, however, these are more than just illustrations of ‘form Â�miming form’. Because of the resemblances Desmond Bates sees between their various iconic Â�signifiers and the underlying experiences to which they allude, these Â�language games also entail complex patterns of what Fischer and Nänny once Â�tentatively referred to as “form miming form miming meaning”.10 At this juncture, I want to move on from single-phrase examples of Â�thematized iconicity and various wittily juxtaposed homophones to a more self-contained Â�illustration, albeit one still based on acoustic similarities, though this time between

.â•… The case for such a supposition is set out in the editors’ Introduction to The Motivated Sign: “The miming of one form by another (repetition) is only noted when it is in some sense meaningful. […] mere repetition should not be considered an example of endophoric iconicity. It is clear that the repetition only becomes meaningfully iconic when it rises above conventional expectations. For this reason, the difference between endophoric and exophoric iconicity may be less clear in practice than it is in theory […], because even endophoric iconicity consists of a structure (form miming form), which, combined into a new sign, then mimes a third element (a signified) in the world as we perceive it” (Fischer & Nänny 2001:â•›2, my emphasis).



Thematized iconicity and iconic devices in the modern novel 

whole sequences of dialogue and how they are (mis)interpreted. In the case of the bravura passage which follows, we might perhaps initially assume that the Â�misunderstandings in this mindless exchange (a further example of dysfunctional semiosis) are primarily indicative of the listener’s degree of deafness. But this would only be the case, if the dialogue were as straightforward as the examples recalled at Bates’ lip-reading class (although even here there is arguably some rather Â�improbable material). In the case of the abortive Senior Common Room exchange recorded in the first third of the novel, the narrator’s dismissive references to Dada and to Noam Â�Chomsky’s “impossible sentences” would seem to suggest something very different. For reasons Â�considered later on, although Deaf Sentence is generally a quasi-realist novel, the passage I now want to consider marks a radical break with the general pattern. The following extract involves a highly contrived scenario, one more likely to have been created by the novel’s implied author than a recollection supplied by Desmond Bates in his role as participant, for he is a generally reliable narrator. The reception in the Senior Common Room […] was the usual ordeal by Lombard Reflex.11 There were several fellow sufferers among the elderly guests whom these occasions tend to attract, and I had some exchanges along the familiar lines of ‘Terribly noisy in here’ – ‘What?’ – ‘I said it’s terribly noisy in here’ – ‘Sorry, can’t hear you, it’s so damned noisy in here…’ Then Sylvia Cooper, wife of the former Head of History, engaged me in one of those conversations in which your interlocutor says something that sounds like a quotation from a Dadaist poem, or one of Chomsky’s impossible sentences, and you say ‘What?’ or ‘I beg your pardon?’ and they repeat their words, which make a banal sense the second time round. ‘The pastime of the dance went to pot’, Sylvia Cooper seemed to say, ‘so we spent most of the time in our shit, the cows’ in-laws finding they stuttered.’ ‘What?’ I said. ‘I said, the last time we went to France it was so hot we spent most of the time in our gîte, cowering indoors behind the shutters.’

.â•… As the protagonist-narrator explains at the beginning of Deaf Sentence, “the Lombard Reflex [is] named after Etienne Lombard, who established early in the twentieth century that speakers increase their vocal effort in the presence of noise in the environment in order to resist degradation of the intelligibility of their messages. When many speakers display this reflex simultaneously, they become, of course, their own environmental noise-source, adding incrementally to its intensity” (Lodge 2008:â•›3). Taking his cue from Lombard, Lodge’s Â�protagonist postulates a spoof Bates Reflex, “named after Desmond Bates, who established early in the twenty-first century, that users develop an unconscious hostility towards their hearing aids which causes them to punish these devices by carelessly allowing the batteries to run down” (Lodge 2008:â•›9).

 John J. White

‘Oh, hot, was it?’ I said. ‘That must have been the summer of 2003.’ ‘Yes, we seared our arses on bits of plate, but soiled my cubism, I’m afraid.’ ‘I’m sorry?’ ‘We were near Carcassonne. A pretty place, but spoiled by tourism, I’m afraid.’ ‘Ah, yes, it’s the same everywhere these days,’ I said sagely. ‘But I do mend sherry. Crap and Sargasso pained there, you know. There’s a lovely little mum of modern tart.’ ‘Sherry?’ I said hesitantly. ‘Céret, it’s a little town in the foothills of the Pyrenees’, said Mrs Cooper with a certain impatience. ‘Braque and Picasso painted there. I recommend it.’ ‘Oh yes, I’ve been there’, I said hastily. It has a rather nice art gallery.’ ‘The mum of modern tart.’ ‘Quite so’, I said. I looked at my glass. ‘I seem to need a refill. Can I get you one?’  (Lodge 2008:â•›108, underlinings are my emphasis, the italics are the author’s)

The humour in the above disastrous exchange depends largely on our first being told what Bates thought his interlocutor said, only to learn subsequently of her relatively innocent, ham-fisted attempts at making polite conversation. An added complication occurs near the end of the extract, when the protagonist comes close to second-Â�guessing what his interlocutor did actually say. That notwithstanding, the dominant narrative order ensures that readers first experience deafness vicariously: from a variation on what Henry James would call the “obtuse” narrator’s standpoint. Thereafter, a kind of distortingly perverse acoustic decoding ploy appears to be at work in Desmond Bates’ warped misinterpretations of what was said to him. There are a number of occasions when the relationship between what Â�Sylvia Â�Cooper is thought to have said and her actual words could still be described as ‘loosely motivated’, much as that between onomatopoeic words and the sounds they purport to represent is (something that does not usually deter us from regarding onomatopoeia as an acceptable form of acoustic iconicity).12 While we do not automatically laugh at the mismatch between onomatopoeic words and their acoustic objects, much of the present passage’s humour certainly depends on a series of ludicrous discrepancies between the addressee’s crass misconstructions and the source-remarks. The kinds of Â�meanings Desmond Bates comes up with are above all

.â•… One reason usually suggested for this discrepancy is that onomatopoeic forms “are made to conform to the phonological systems of particular languages, rather than being directly imitative of what they […] stand for” (Lyons 1977:â•›101).



Thematized iconicity and iconic devices in the modern novel 

indexical in another, more revealing sense than that of being prima facie evidence of his hearing disability: they reflect his gut Â�hostility to the particular speaker and the social event. For once, the protagonist is Â�himself no longer the figure of fun he once was at the family Christmas gathering. This time the principal target would appear to be the chattering classes, especially the sort of thick-skinned people who insist on trying to engage in polite social small-talk, even with someone as notoriously hard of hearing as Desmond Bates is. Since most of us Â�chatting and exchanging pleasantries with colleagues at academic receptions of the kind described in Deaf Sentence are unlikely to be reminded of Dada poetry or Chomsky’s “green ideas Â�[sleeping] furiously”, rather than seeing the Â�protagonist’s Â�misinterpretations as an attempt at single-handedly keeping a polite conversation going under difficult Â�circumstances, it makes more sense to interpret the above passage as an implausible (but still iconically configured) satirical sequence. To sum up: in this episode from Deaf Sentence we find a specific sub-type of Â�endophoric iconicity being exploited during the course of the work’s various Â�fictive Â�scenarios, one based on a grotesquely distorting, yet still loosely motivated Â�relationship between spoken utterance and ‘assumed’ (i.e. usually misheard) counter-version. Once again, the material is perspectivized, partly through the prism of the Â�linguistic Â�technical terminology that the protagonist brings to bear on various Â�incidents, but also because his account does more than just replicate the conversations he Â�unsuccessfully engages in. To give a simple example from the passage above, the act of thematization – in this case, subjectivized narration – makes good use of Â�various Â�self-effacing Â�reporting Â�gestures. Subtly positioned verba dicendi (“I said”, “I said sagely”, “I said Â�hesitantly”, “‘Quite so’, I said and I looked at my [empty] glass”) play a role in Â�making this Â�improbable exchange more than just a one-sided game conducted at the Â�protagonist’s expense. This is not the only place in Deaf Sentence where thematization is a Â�product of subjective perspective. Quite early in the work, Desmond Bates admits to Â�regularly word-processing what he modestly refers to as “some kind of journal, or notes for an autobiography, or perhaps just occupational therapy” (Lodge 2008:â•›38). He Â�usually works on this project late into the night, cathartically recording the events and Â�impressions of the previous day in what seems to be a sort of retiree’s blog. Many of the novel’s individual chapters and anecdotes are specifically dated. The author’s choice of genre (in this case, a serial blog-cum-diary narrative) has important Â�repercussions for the status of the entire novel: the modern computerized Â�equivalent of an intimate diary, projected out into the blue with no particular addressee in mind, potentially risks becoming as solipsistic and subjectivist as the various Â�asymmetrical exchanges Desmond Bates engages in, whenever he is professionally obliged to make an appearance in public. On the protagonist’s own admission, “as aural-oral Â�communication becomes more and more difficult, the total control one has over written discourse

 John J. White

becomes more and more appealing, especially when the subject is deafness” (Lodge 2008:â•›12). This is arguably the key to the novel’s Â�presentational strategy. In the case of Deaf Sentence, ‘narrated iconicity’ might seem to be a more Â�appropriate term than ‘thematized iconicity’. Although acoustic iconicity becomes an integral part of the novel’s plot, it is the narrator’s academic interest in the Â�internal organization of sound patterns, misleading homophones and minimal pairs that informs a large proportion of the work’s presentational strategy. While his humorous smokescreen of technical jargon at times helps him come to terms a little with his deteriorating condition, he manages on most occasions to do so by exercising his sense of humour. ‘Thematization’, I would argue, is by no means a neutral concept. It can cover a whole spectrum of discourse modes and narratorial behaviour patterns.

3.â•… T  ext thematizing image and image illuminating text: Two-way traffic in Umberto Eco’s The Mysterious Flame of Queen Loana Umberto Eco’s La Misteriosa Fiamma della Regina Loana (English translation 2005), is a 460-page work of fiction containing over 150 illustrations. Despite the fact that the publisher’s blurb presents the work as a “graphic novel”, the actual sub-title, An Â�Illustrated Novel, is more accurate. That is to say, Eco’s novel has little in common with the “graphic novel”, as the term is understood in the twentieth and twenty-first centuries, or with the results of the recent trend in text-plus-image fiction. For a start, the ratio of image to text in Eco’s work is exceptionally generous by comparison with that in other visual intermedial works of fiction. This needs to be stressed because the proliferation of visual material in The Mysterious Flame is clearly a key part of the work’s principal iconic strategy. Whereas many recent novels include indexical Â�photographic material, The Mysterious Flame most frequently deploys reproductions of comic images, book-illustrations and poster images rather than actual photos. Over 90% of the novel’s visual material is reproduced in bold colour, with monochrome and sepia only being used to reproduce newsprint or a handful of old family photographs. The resultant bold impact of the novel’s illustrative component makes an encounter with The Mysterious Flame substantially richer than one with any run-of-the-mill graphic novel or even much text-plus-image fiction. Because The Mysterious Flame forms part of a larger project-in-progress on Â�thematized iconicity, I shall restrict myself in what follows to a handful of Â�illustrative scenarios. These will be discussed in some detail, after which the paper will conclude with some general observations on the ways in which Eco thematizes his Â�(iconic-Â�cum-indexical) visual material and on that material’s narrative function. First, some introductory remarks on the images’ narrative context are called for. Eco’s novel is the story of a sixty year-old Milanese antiquarian bookseller diagnosed



Thematized iconicity and iconic devices in the modern novel 

as suffering from selective memory loss after emerging from a coma-like trance.13 In his doctor’s words, “your memory is good. Except you don’t know who you are” (Eco 2005:â•›8). While possessing total recall of every book he has read (the novel’s rich array of literary quotations, bookish allusions and intertextual associations serve to prove the point), the protagonist cannot even remember his own name, what he Â�previously did for a living, the fact that he has a wife and family, or what kind of a biography he had ammassed prior to the novel’s starting-point in the Â�present. Â�Sensing that by methodically reconstructing his past he will eventually find the key to his present and future identity – “you can only anticipate the future if you can call the past to mind”, he is advised (Eco 2005:â•›36) – the protagonist embarks on a Â�systematic Â�journey of Â�self-rediscovery that is likened to Tom Sawyer’s descent into the cavern (Eco 2005:â•›38, 119). “Perhaps I could learn in [the cavern of] that attic about the formation of my sexual consciousness” is one of the thoughts that Â�motivates his quest, as he rummages about in an attic in search of evidence of his early years (Eco 2005:â•›127).14 In his search for hidden clues to his identity, Eco’s protagonist (known as Yambo to his family and close friends) sees himself as Â�having become a sort of Â�do-it-yourself Sherlock Â�Holmes, for once searching for clues to his own Â�identity rather than that of some elusive Â�criminal outsider. The following Â�analogy sums up his introspective train of thought on discovering a run of The Strand Magazine Â�containing all of Sherlock Holmes’ serialized adventures: Those images said to me: de te fabula narratur. At that very moment Sherlock Holmes was me, intent on retracing and reconstructing remote events of which he had no prior knowledge, while remaining at home, shut away, perhaps even in an attic. He too, like me, motionless and isolated from the world, deciphering pure signs. He always succeeded in making the repressed resurface. Would I be able to? At least I had a model.  (Eco 2005:â•›152)

In order to escape the contemporary world in which he does not yet feel at home, Yambo has by this time retreated to the family’s ancestral home of Solara in the hills to the north of Milan, there to immerse himself in material from a past that has for so many decades eluded his grasp. At this point, the detective novel starts to metamorphose into something of a modern Gothic novel, with trap-doors, hidden rooms full of papers, mementos, books, ghosts from the past, and even a house with a supersti-

.â•… A detailed medical diagnosis of the protagonist’s mental condition is offered in the novel’s opening chapter “The Cruellest Month” (Eco 2005:â•›3–27). His self-diagnosis largely follows in Chapter 2 “The Murmur of Mulberry Leaves” (Eco 2005:â•›28–44). .â•… In the following discussion, the phrase “Yambo senior” is used as shorthand for the Â�protagonist’s sixty year-old self. “Yambo junior” is used for his younger self only where the context is ambiguous.

 John J. White

tious old retainer. The trail of clues Yambo follows during his Â�assiduous Â�detective work up at Solara comprises a whole treasure-trove of forgotten evidence from his past. This material by and large assumes the form of visual images Â�(illustrations from the comics and adventure books that captured his imagination when he was an inquisitive young boy), then, with passing decades, first the Italian nationalist, then the fascist Â�propaganda publications that formed part of his early teens, as well as Â�documentary material Â�relating to Italy’s role during the Second World War and the local partisans’ Â�resistance activities, and, from a post-war period, a seductive array of fashion Â�magazines and advertisements that opened his eyes to a modern adult world of Â�fantasies. Some of this material arouses strong feelings of nostalgia in Yambo senior (the self who is Â�rediscovering his young self ’s past); other items merely Â�confirm the voracious young reader’s budding love of literature and of virtually all things Â�antiquarian. In contrast, some (especially those dating from the fascist and immediate post-1945 eras) raise ethical questions about how an early obsession with all things fantastic and romantic could be politically exploited during the teenager’s subsequent formative years. Stirring images of boyish heroism, Italian patriotism, adventure and sexual prowess are the common denominator of both the visual component and the passages that thematize the Â�material’s iconic and indexical significance. The novel’s juxtaposition of a narrative text evoking feelings triggered off by these nostalgic images from Yambo’s early life with the reproduced images themselves is calculated to encourage a certain voyeurism. For as the protagonist lovingly dwells on the various iconic echoes of his past – his younger self ’s “epiphanies of the exotic”, as they are called (Eco 2005:â•›95) – the reader is also invited to join him in walking down Memory Lane. We look at the same visual images that fascinate young Yambo. Although some of the work’s text-illustration configurations are specific to Italian history (especially, from the 1930s to the early 1950s), much of the other visual material is at the same time part of the twentieth century’s collective past, or at other times typical of the kind of matter that all bright young boys must have consumed with enthusiasm. The work’s material is not without its equivalents in many of the readers’ childhood experiences. Stamps, comics, adventure tales, patriotic images, remembered film stills and glossy magazines with erotic overtones all retain their power to trigger off powerful memories in both reader and protagonist. Even music and popular songs of the time are recalled visually via record- and sheet music-covers. This visually orchestrated nostalgia accounts for the sheer quantity of images on display in The Mysterious Flame alongside the narrator’s descriptions of their impact. The visual images and their thematizing counterparts are equally important. They complement one another and in doing so encourage us to join Yambo senior in feasting our eyes on many visual echoes of the past, even if some of them do not have the same sentimental value for us as they do for the protagonist. As we are repeatedly reminded (Eco 2005:â•›25, 53, 72, 87), the novel’s convalescing protagonist sees himself as consciously searching for some sort of epiphany Â�comparable



Thematized iconicity and iconic devices in the modern novel 

to Marcel’s “madeleine”-experience in Marcel Proust’s A la recherche du temps perdu.15 “I don’t remember images or smells, or flavours” he confesses to his doctor at one early stage, “I only remember words” (Eco 2005:â•›26). However, he is assured that this could all soon change. And it does.

4.â•… Yambo’s treasure trove The discovery up at Solara of what was probably Yambo junior’s first stamp album is one of the early occasions when he becomes extremely receptive to powerful Â�stimuli from the past. The encounter with the long forgotten album and Yambo senior’s response to one particular page is the subject of the following extract: Poking around haphazardly, I picked out a clothbound album, in landscape format. As soon as I opened it I saw that it was a collection of stamps. Clearly mine, as it bore my name at the beginning, with the date, 1943, when I presumably began collecting. […] I probably invested all my scant capital with some stationer in the city, one who specialized in selling to budding collectors, and so those stamps that to me seemed straight from fairytales represented currency. […]. For me, that album must have been, more than a material object, a receptacle of oneiric images. I was seized with a burning fervour at the sight of each figure. Forget the old atlases. Looking through that album I imagined the clear blue seas, framed in purple, of Deutsch-Ostafrika; I saw the houses of Baghdad, framed by a tangle of Arabian-carpet lines, against a night-green background; I admired a profile of George V, sovereign of the Bermudas, in a pink frame against blue fields; the face of the bearded pasha or sultan or rajah of the Bizawar State […] subdued me with its shades of terracotta; certainly the little pea-green rectangle from the Labuan Colony was rich in Salgarian echoes; perhaps I was reading about the war that was started over Gdansk as I was handling the wine-coloured stamp marked Danzig; on the stamp of the state of Indire, I read FIVE RUPEES in English; I daydreamed over the strange native pirogues outlined against the purple backdrop of some part of the British Solomon Islands. I invented tales involving a Guatemalan landscape, the Liberian rhinoceros, another primitive boat, which filled a large stamp from Papua […], and I wondered where Saargebiet was, and Swaziland.  (Eco 2005:â•›253–55)

.â•… The role of the madeleine epiphany in bringing back Marcel’s Combray past is recounted in the “Overture” section of Swann’s Way (Proust 1939:â•›58–59). As Beja (1971:â•›61) points out, “it is not on the meaning of the past moment that [Marcel] centres his attention, but on the significance of his recapture of that moment – not so much on lost time as on the search for it”. The same holds true for Yambo’s quest, with the rider that in his case past time often remains a distinctly political concept.

 John J. White

This unforeseen encounter with a once much-treasured possession triggers Â�emotional memories of a long-lost world as well as the young boy’s yearning for surrogate Â�locations, exotic ones that he was unlikely ever to visit, given the current state of the world in the mid-twentieth century. “I was seized with a burning fervour at the sight of each figure” (Eco 2005:â•›254), Yambo senior confesses of the philatelic images on the stamps depicted. Having chanced upon such a once prized Â�possession, the man almost becomes a boy again. While we may not all share his former enthusiasm for the rich colours of early British Colonial stamps or be intrigued by the Â�historical Â�significance of other significant items on the page in question, a young collector’s Â�emotions are Â�palpably reflected in this passage. That said, much of the vocabulary used in the above passage’s second paragraph and in many of the ensuing trains of Â�associative thoughts are essentially ones that go through Yambo senior’s mind as he contemplates images from his, and the mid-twentieth-century’s, past. “Perhaps I  was reading about the war that was started over Gdansk as I was handling the Â�wine-coloured stamp marked Danzig”, Yambo senior imagines (ibid.). His thoughts then turn to another historically important stamp in the collection: one from German East Africa Â�(German Colonial territory confiscated by the Versailles Treaty at the end of World War 1). Then, thinking associatively about Central Europe’s more recent Â�territorial disputes, he wonders where the Saargebiet might be on a map. Of all the 12 stamps reproduced alongside the narrative account of the stamp-album episode, only the Danzig stamp carries weighty political significance. The following, more general Â�paragraph, recording Yambo junior’s joy at the freedom to travel in his imagination, nevertheless gives his hobby a more personal, essentially political complexion: During the years when we seemed to be penned in by insuperable barriers, pinched between two clashing armies, I travelled the wide world through the medium of stamps. […]. When even the train connections were interrupted […], there I was, soaring from the Vatican to Puerto Rico, from China to Andorra.  (Eco 2005:â•›256)

The fact that his album is a veritable magic carpet, allowing the boy to escape from a threatening outside world, is probably something only Yambo senior would have been able to appreciate in retrospect. What is more, only he has the language to express this idea adequately. The picture of Yambo junior, trapped in the vice of history, yet finding an outlet in a seemingly politics-free hobby, is echoed later on: A new flutter of tachycardia seized me at the sight of two stamps from Fiji […]. They were no prettier or uglier than the others. One showed a native, the other bore a map of Fiji. […] Perhaps I was struck by the map’s precision, like a chart of treasure islands, perhaps I had encountered the unheard-of-names of those



Thematized iconicity and iconic devices in the modern novel 

territories for the first time upon those little rectangles. […] I kept staring at these two stamps.  (Eco 2005:â•›253–56)

Though hardly obvious material for a “Proustian moment”, the two Fijian stamps that Yambo junior once jealously coveted have now been acquired and are about to be assigned their rightful place in his album. As in the first passage devoted to Yambo senior’s philatelic discovery, historical events of the time give even such an unpretentious domestic sequence personal significance. The passage quoted above evokes a dark time when much of Europe was hemmed in by invading armies and a prey to internal threats. That reproduced below now clears the air, being set at a later time when the Allies, with partisan support, are set to take Italy, and Mussolini’s Â�fascism is on the verge of collapse: I told [Don Cognasso, one of Yambo’s wise mentors] about my stamps, and he said that once upon a time he had collected them, too, but after the war he lost interest and threw them out. He had twenty or so left and would happily give them to me. I went to his house and was amazed by my windfall: it included the two Fiji stamps. […] I took them home to put them on a new page of my album. […] I was sitting at the table tending my collection. […] The radio was on. We had just heard the end of a propaganda programme from the Republic of Salò that featured the members of a single family discussing politics and concluding, of course, that the Allies were our enemies, that the Partisans were bandits resisting the draft out of sloth, and that the Fascists in the north were defending Italy’s honour alongside their German comrades. [Then] the Voice of London special messages began, as mysterious and evocative as a Montserrat stamp, for the Partisan brigades: Messages for Franchi, Happy is not happy, The rain is past, My beard is blond, Giacomo kisses Mohammad, The eagle flies, The sun also rises… I see myself still adoring the Fiji stamps when suddenly between ten and eleven the sky starts buzzing, and we turn out the lights and run to the window […]. Some said it was an English reconnaissance plane, some, an American plane that came to drop packages, food and arms for the Partisans in the mountains, perhaps not far from us, on the slopes of the Langhe.  (Eco 2005:â•›336–39)

In the first sequence, stamp-collecting is the personal consolation of a lone boy Â�mentally escaping into an exotic faraway world. In the second passage, stamps and album take their place in the family circle alongside other mundane consolations. This Â�re-connecting of various images from the 1930s and mid-1940s to create an Â�associative sub-narrative is not untypical of the way in which both iconic and thematized strands of the novel are interwoven throughout The Mysterious Flame. Images comment on each  other, the narrative comments on them, and our relating these strands to one another – both implicitly and via commentary – also opens up further interpretive Â�possibilities, as we shall shortly see. The one passage anticipates what is to come, the other serves as more of reminder of darker times. Patterns of pre- and

 John J. White

Â� post-Â� figuration come together here, as the stamp album becomes a leitmotif.16 Rather than Â�commenting Â�further on the interrelated passages cited so far, I now want to Â�consider some Â�intriguing features of the album-page reproduced in Figure 1, since these take us off on a further, hitherto unforeseen interpretive trail. This is not the familiar terrain of text-and-image fiction. To start with some obvious points: one might wonder whether anyone, even a juvenile collector, would put such a ragbag assortment of stamps together on the same page. Indeed, the images beg a further question as to which young collector would be able to mount his acquisitions so professionally and using such quality backing at an apparently tender age. More important, even a novice would surely notice that the 1932 Cook Island stamp at the top right of the set displays a mouth-watering error: the frame contains the inverted centre-image of a canoe (the kind of ‘invert’ most collectors can only dream of!). Yambo senior may be a rare books expert, but when outside his field of expertise, he appears unaware that the Danzig stamp he is Â�looking at is extremely rare, as are the £25 British Central Africa stamp and most other items on this page. Eco is playing a game with the reader here, and that game has much to do with the interplay between iconic illustrations and their thematization at Â�various Â�junctures in the entire narrative. Put bluntly, what is the point of pretending to Â�illustrate material from what was allegedly a first stamp album, housing the kind of modest assembly of beginner’s stamps one would expect of someone of his age? We hardly need Sherlock Holmes’ help to speculate on why the images embedded within the passage devoted to a single page in young Yambo’s collection are so important. Instead, we need to look to what some of the novel’s comic-book characters, Mickey Mouse, Clarabelle Cow and the crew of the good ship Hispaniola tell us earlier on in the novel. Let me explain briefly what is being cryptically alluded to here, as this will lead us to one further form of interweaving between visual iconicity and thematized iconicity in The Mysterious Flame: the patterns of pre- and post-figuration which such visual-textual cross-referencing that such effects often entail. A comic-book story entitled Clarabelle’s Treasure was, we learn, one of Yambo junior’s favourite reads. It involves an adventurous search for hidden treasure every bit as involved as the one at the end of Robert Louis Stevenson’s Treasure Island. The details are given in an episode devoted to the recollection of comic-book stories that appears right at the beginning of the novel:

.â•… Given that The Magic Mountain is invoked in The Mysterious Flame, Thomas Mann’s Â�description of leitmotif as “die vor- und zurückdeutende magische Formel” (the magic formula that points both forwards and backwards [Mann 1946: xi]) may not be irrelevant to Eco’s strategy of interlacing prefigurative and postfigurative iconic visual material with Â�passages thematizing iconicity.



Thematized iconicity and iconic devices in the modern novel 

I told the story from memory. ‘You see, Mickey Mouse and Horace Horsecollar, taking an old map, went in search of a treasure that had been buried by Clarabelle Cow’s grandfather or great uncle, and they’re racing against the slimy Eli Squinch and the treacherous Peg-Leg Pete. They come to the place and consult the map. They’re supposed to start from a big tree, make a line to a smaller tree, then triangulate. They dig and dig but find nothing. Until Mickey has a flash of inspiration: the map is from 1863, more than seventy years have passed, this little tree couldn’t possibly have existed at that time, so the tree that now appears big, must have been the little one then, and the big one must have fallen down, but may have left traces. And indeed they look and look, find a piece of old trunk, redo the triangulation, start digging again, and there it is, exactly in that spot, the treasure.’ ‘But how do you know all that?’ ‘Doesn’t everyone?’ ‘No, everyone does not’, said Paola [Yambo’s wife], excitedly. ‘That isn’t semantic memory, that’s autobiographical memory. You’re remembering something that made an impression on you as a child! And this cover Â�[below-left] sparked it’.  (Eco 2005:â•›21)

Various tales of adventure and the reward of treasure trove are contained in The MÂ�ysterious Flame, many of them seeming to prefigure Yambo’s experiences up at Solara. The chapter in which Yambo leaves Milan in order to go up to Solara to try to confront and solve his identity issues is entitled “Clarabelle’s Treasure”. (It is immediately after this that Yambo and the reader are bombarded with visual images that from then on proliferate in every chapter and grow larger in format until they fill entire pages.) If the visual material itself is not the metaphorical treasure the protagonist has been Â�desperately searching for, the interpretation he places on them could be. Yambo’s personal treasure-hunt is in fact figurative, rather than resembling the pragmatic one Clarabelle and her cartoon assistants engage in or that for Captain Flint’s treasure in Stevenson’s Treasure Island. We have been invited to follow a Â�misleading “paper-trail” (a term used repeatedly in the novel). Eco lays a further trap for the reader later on by allowing his protagonist to chance upon a copy of the first folio edition of Shakespeare’s plays (dating from 1623). The dream of chancing upon material rich beyond one’s wildest dreams is central to Yambo’s antiquarian activities, but it has also been prefigured in the discovery of a collection of valuable stamps without the lucky finder having any sense of their real worth. The real treasure at the end of the figurative quest in the case of The Mysterious Flame comes, I would argue, in the final 26 pages of Yambo’s search, when, encountering a phantasmagorical cavalcade of real and comic-book figures from his past, the protagonist is reconciled with his earlier lives and experiences. As for the lost loves of his earlier years that so often haunt him, he is finally at one with the past: “I will

 John J. White

be reunited. I will be at peace” is the final thought with which he consoles himself (Eco 2005:â•›448).

5.â•… Concluding thoughts The lack of scholarly attention to thematized iconicity is surprising, given the role played by iconic signs in our everyday social lives.17 However, as some of the Â�examples in the present paper have demonstrated, thematized iconicity in literature usually involves a far richer spectrum of semiotic devices than one encounters in modern life. For example, prefiguring, and at times postfiguring18 events that occur later on in The Mysterious Flame is a vast, complex task performed both by the visual material (combining iconic and indexical functions) and the narrator’s commentaries on them. Thematizing iconicity in such cases clearly does not take place in a vacuum, it is part of a complicated interweaving of various strands: the free-standing individual iconic image, its narrative exegesis, the endophoric interplay between various image-patterns evolving across large tracts of Eco’s novel, and the changes in genre and text-type that mirror the protagonist’s progression towards self-understanding. Few modern novels demonstrate more convincingly than The Mysterious Flame that thematized iconicity performs an important role in modern fiction.

References Beja, M. 1971. Epiphany in the Modern Novel. Seattle WA: Washington University Press. Bondanella, P. 1997. Umberto Eco and the Open Text: Semiotics, Fiction, Popular Culture. Â�Cambridge: CUP. Capozzi, R. 1997. Interpretation and overinterpretation: The rights of texts, readers and implied authors. In Reading Eco: An Anthology. R. Capozzi (ed.), 216–34. Bloomington and Â�Indianapolis IN: Indiana University Press. Eco, U. 1979. The Role of the Reader: Explorations in the Semiotics of Texts. Bloomington IN: Indiana University Press. Eco, U. 1983. The Name of the Rose. London: Secker & Warburg. (Italian: Eco, U. 1980. Il nome della rosa. Milan: Bompiani). Eco, U. 1983a. Horns, hooves, insteps: Some hypotheses on three types of abduction. In The Sign of Three: Dupin, Holmes, Peirce, U. Eco & T. Sebeok, (eds), 198–220. Bloomington IN: Indiana University Press.

.â•… For examples, see Hodges & Kress 1988 and Dines Johansen & Larsen 2002. .â•… On strategies of prefiguration and postfiguration in the modern novel, see White 1971:â•›11–14.



Thematized iconicity and iconic devices in the modern novel 

Eco, U. 2005. The Mysterious Flame of Queen Loana. London: Secker & Warburg. (Italian: Eco, U. 2004. La misteriosa fiamma della regina Loana. Milan: Bompiani). Eco, U. & Sebeok, T. (eds). 1983. The Sign of Three: Dupin, Holmes, Peirce. Bloomington IN: Indiana University Press. Hodge, R. & Kress, G. 1988. Social Semiotics. Oxford: Polity. Johansen, J.D. & Larsen, S. 2002. Signs in Use: An Introduction to Semiotics. London: Routledge. Lodge, D. 2008. Deaf Sentence. London: Harvill Secker. Lyons, J. 1977. Semantics, Vol. 1. Cambridge: CUP. Mann, T. 1946. Einführung in den Zauberberg. In Der Zauberberg, Mann, T, ix–xxvi. Stockholm: Bermann-Fischer. Nänny, M. & Fischer, O. (eds). 2001. The Motivated Sign. [Iconicity in language and literature 2]. Amsterdam: John Benjamins. Nänny, M. & Fischer, O. 2001. Introduction. In Nänny & Fischer (eds), 2–14. Nöth, W. 2001. Semiotic foundations of iconicity in language and literature. In Nänny & Fischer (eds), 17–28. Peirce, C.S. 1931–1999. The Collected Papers of Charles S. Peirce. Vols. 1–6, C. Hartshorne & P. Weiss (eds); Vols. 7–8, A.W. Burks (ed.). Cambridge MA: Harvard University Press. (Quoted as CP. Reference is made to vols and paragraphs.) Proust, M. 1931. Swann’s Way, Vol. 1: Remembrance of Things Past. C.K.S. Moncrieff (transl.). London: Chatto & Windus. Richter, D.H. 1986. Eco’s echoes: Semiotic theory and detective practice in The Name of the Rose. Studies in Twentieth-Century Literature 10: 213–236. White, J. 1971. Mythology in the Modern Novel: A Study of Prefigurative Techniques. Princeton NJ: Princeton University Press. White, J. 1999. On semiotic interplay: Forms of interaction between iconicity and Â�indexicality in twentieth-century literature. In Form miming meaning. [Iconicity in Language and Â�Literature 1], M. Nänny & O. Fischer (eds), 83–108. Amsterdam: John Benjamins.

Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy Gabriele Rippl

University of Bern Charles Simic’s collection of prose poems Dime-Store Alchemy. The Art of Joseph Cornell (1992) pays homage to one of America’s greatest surrealist visual artists, whose work Simic reactivates in American cultural memory. Simic’s poems, which are presented as boxes to approximate Cornell’s visual techniques, have been investigated by literary scholars for their ekphrastic verbal means but the strategies of iconicity and intermediality have been neglected. Against the backdrop of recent debates about intermediality and iconicity in postmodern culture, the article sheds light on Simic’s highly self-reflexive intersemiotic translations and transformations and his intricate art of visual iconicity which foregrounds the iconic character of verbal signs. While trying to evoke visual and pictorial qualities via language, Simic’s poems, just like other ekphrastic texts which combine iconic and indexical functions of language, investigate the specificities of and tensions between verbal and visual media.

1.â•… Introduction For scholars working in the fields of iconicity, intermediality and ekphrasis, the Â�Belgrade-born American poet Charles Simic (*1938) and his collection of short prose poems published in 1992 under the title Dime-Store Alchemy. The Art of Joseph Cornell, present an interesting case (cf. Rippl 2005).1 Simic pays homage to one of America’s greatest surrealist visual artists, Joseph Cornell (1903–1972), whose work he Â�reactivates in the American cultural memory. But other writers, not only Charles Simic, have recently paid tribute to Joseph Cornell’s famous boxes – amongst them are Octavio Paz (cf. his poem “Objetos y apariciones” [1991]) and Gabriel Josipovici (cf. his short stories “The principle of order” [1987] and “That which is hidden is that

.╅ I would like to thank my colleagues Christina Ljungberg, Christine Schwanecke and John White who indefatigably commented on different versions of this article. My thanks also go to Alexander Sigrist and Hilary Sharp without whose logistical support this article would not have been submitted for publication.

 Gabriele Rippl

which is known” [1980]). In 2001, a whole collection of literary texts (including some by Robert Pinsky, Siri Hustvedt and Joyce Carol Oates) was published by Jonathan Safran Foer, celebrating Joseph Cornell’s boxes; however, Simic’s response to Cornell’s art can be regarded as the most comprehensive one. Although Cornell’s montages, assemblages and ‘magical’ boxes have clearly inspired Simic’s Dime-Store Alchemy, his rich iconic strategies have not yet been investigated properly. The aim of this article is to discuss these iconic strategies against the backdrop of recent debates about intermediality (Rippl 2010) and (post-)modern culture.

2.â•… T  ypes of iconicity: Charles Simic’s verbal responses to Joseph Cornell’s boxes Between 1921 and 1931 Joseph Cornell, the son of Dutch immigrants to New York, had a job with a textile manufacturer with offices in Madison Square. He killed time between appointments by foraging in used bookstores and junk shops for books, records, photographs, prints of old movies, and theatrical memorabilia (Simic 1992: xvi) and rearranged the treasures collected on his rambles in his famous enigmatic boxes and assemblages. Poet Charles Simic obviously sees himself as a curator of Â�Cornell’s art and thus in charge of stocktaking. Many of his poems are beautiful Â�verbal still lifes and function as inventories, which impart a fresh dignity and ‘magic’ to Â�Cornell’s art. The serial prose poems of Simic’s collection, in turn inspired by the serial quality of Cornell’s enchanting boxes, display iconic aspects at the same time as they are ekphrastic, which helps Simic to produce in his temporal verbal medium the visual effects of Cornell’s static spatial art objects. Not surprisingly, Simic’s Dime-Store Alchemy is replete with different kinds of intermedial relationships between texts and pictures. The collection actually includes several illustrations of Cornell’s boxes since many of Simic’s prose poems are ekphrases that refer to static visual works of art via purely verbal descriptions,2 and sometimes text and image even form a unit. As the majority of his poems in Dime-Store Alchemy are technopaignia or pattern poems whose lines are arranged in such a way that they seem to imitate the shape of Cornell’s boxes, they are staple examples of iconicity,

.â•… In his seminal article Heffernan distinguishes ekphrasis, pictorialism, and iconicity. While ekphrasis is “the verbal representation of visual representation” (1991:â•›2), i.e. the Â�representation of works of art, and pictorialism the representation of natural objects and artifacts which “generates in language effects similar to those created by pictures” (ibid.: 300), iconicity is “more complicated than pictorialism because it embraces sounds and sets relations as well as visual properties” (ibid.: 300).



Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy 

Â� displaying a visual resemblance between the arrangement of words or letters on a page and what they signify. This indulgence in iconic dimensions of language transforms Simic’s Dime-Store Alchemy into a prime example of contemporary poetry, precisely because it invites its readers not only to read, but also to view the spatial structure of the poems in order to grasp their full meaning. Literary scholar Christina Ljungberg has drawn attention to the fact that many contemporary writers, amongst them Â�Margaret Atwood and Michael Ondaatje, “are playing with the conventions of language in order to exploit the possibilities of linguistic signs and systems creatively. In particular, it is the diagrammatic aspect of iconicity that seems to attract contemporary writers, as they use form to add meaning” (2001:â•›351; 2003; 2005). Even though Ferdinand de Saussure has argued for the arbitrariness of the Â�linguistic sign, more and more scholars have maintained over the last decades that the arbitrary sign often carries traces of motivation and is not exclusively conventional. In their view, Charles Sanders Peirce’s concept of the sign offers a more viable way of Â�looking at the function of literary language in use. Peirce differentiated between the icon, the index and the symbol and introduced the concept of iconicity. As a Â�ubiquitous feature of language, “iconic signs are, to a certain degree, omnipresent in language and verbal communication” (Nöth 2001:â•›18). As we will see this is very pertinent to Simic’s poems, whose arrangement of lines imitates the shape of boxes, and display iconic features with self-referential functions. In this context, Werner Wolf ’s (2003) concept of “intermedial iconicity” is a Â�phenomenon especially relevant in connection with Simic’s poems. Wolfgang G.  Â�Müller and Olga Fischer called Wolf ’s concept of intermedial iconicity an “entirely new Â�departure in iconicity research” (Müller & Fischer 2003:â•›15) and defined it as a Â�phenomenon dealing with the interdependence and interaction between different art media on the level of form […]. This type of iconicity emerges (1) when a work of art is transferred from one medium to another one and in this process retains formal features inherent in the source medium, and (2) when a work of art adopts or imitates formal features characteristic of another medium.  (Müller & Fischer 2003:â•›15)

Hence, while Müller and Fischer consider intermedial iconicity as a subcategory to the category of diagrammatic iconicity and while Nöth would probably see it as a case of Â�iconicity based on the exophoric principle, which implies a reference to Â�something beyond language (this is the case with Simic’s ekphrases which are external references to a visual medium), Wolf ’s category of intermedial iconicity seems to go beyond both of these conceptualizations. According to Wolf, the category of intermedial Â�iconicity Â�covers such texts or text passages which mime “other arts and media” (Wolf 2003:â•›339). Wolf

 Gabriele Rippl

is very much interested in the modality and Â�mediality of forms, the media Â�underlying the forms.3 For him, forms can be Â�“specified Â�principally by the nature of their underlying semiotic systems (involving verbal Â�language, Â�pictorial signs, music, etc., or, in cases of ‘composite media’ such as film, a combination of Â�several semiotic systems)” (Wolf 2005:â•›253). With regard to Simic’s Dime-Store Alchemy it is important to note that the poems’ strong intermedial nature, which is Â�predominantly based on the Â�foregrounded functions of the iconic Â�(sometimes in Â�combination with the Â�indexical), leads to Â�self-reflexivity and invites readers to Â�ruminate on media Â�specificities and the Â�functions of the different media in our Â�postmodern world. Â�Moreover, Â�intermediality also Â�pertains the sensory modality and the “semiotic Â�register of sign functions” Â�(Ljungberg 2010:â•›82). Following Peirce’s Â�definition of the sign as a mediator or medium, Christina Ljungberg claims that intermediality Â�“concerns the Â�transgression of the border between […] media, e.g. between Â�different sign Â�systems and/or the iconic enactment of one medium within another. It also involves the modality of a Â�specific medium” (ibid). According to Ljungberg, Â�intersemiotic Â�translations are Â�“radically Â�performative”, “strongly self-reflexive” and “a highly Â�effective Â�communication Â�strategy” (ibid.: 83). As we will see in the sections below, by transgressing one sign system, Simic is indeed able to “create something new and unprecedented” (ibid.: 84) while at the same time reflecting the visual medium in the verbal one and, in doing so, displaying a high degree of self-referentiality which goes beyond the ordinary iconic self-reference Â�typical of every aesthetic sign.

3.â•… Readings In the subsequent readings of Charles Simic’s Dime-Store Alchemy, different iconic strategies will be investigated and intermedial instances discussed. The latter can either be based on imagic, on diagrammatic or on metaphoric iconicity; most often, however, intermedial effects in Simic’s poems are triggered by diagrammatic iconic aspects. In all cases, however, it calls upon a recipient’s mind in which the imitations and implicit references to other media “coalesce into an imaginary presence” (Wolf 2003:â•›353). As Max Nänny puts it, “the awareness of iconicity is dependent on the Â�recognition of similarity” (1986:â•›200), i.e. on the reader’s alertness. The first example from Simic’s poetry collection, “Untitled (Soap Bubble Set), 1936”, refers to Cornell’s 1936 construction Untitled (Soap Bubble Set, Figure 1) which was included in the Â�historic “Fantastic Art, Dada, Surrealism” exhibition of that year at the Museum of

.â•… Linguist Ludwig Jäger speaks of a displacement or repression of the problem of mediality, that is of the sensuous side of a sign (Jäger 1999:â•›13).



Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy 

Modern Art at New York. This work belongs to a whole series of boxes with soap bubbles and assembles objects that seem to represent the whole cosmos through their globe-like shape: “there are four weights hanging at the top, a narrow compartment to the left containing a wineglass holding an egg, and another to the right with a column bearing a child’s sculptured head. In the center is a shelf holding a clay pipe and, on the back wall, a geological map” (Ashton 1974:â•›91).

Figure 1.╇ Glass case with inserts, 15 3/4╛x╛14 1/4╛x╛5 3/8 in., Wadsworth Atheneum, Hartford, Connecticut

Like the objects in some of Giorgio de Chirico’s paintings, for instance, in his Love Song of 1914, Cornell’s trivial items from everyday life inhabit a solitary, calm world and are thus “elevated to the realm of the extraordinary” (Waldman 2002:â•›24). The soap bubble is, of course, a bubble, but it is also a canonical vanitas symbol in many still lifes. The ‘magic’ of Cornell’s boxes is brought about not only by the Â�associative quality of the everyday life objects they assemble, but also by the Â�structural device of the frame. The box in which Cornell has arranged the collected objects Â�functions as a demarcation of space within which the ‘magic’ can take place. Like framed Â�windows which open into another world, Cornell’s framed boxes allow us to dive into Â�‘magical’ worlds of childhood dreams, as we charge the boxes with our own imaginative narratives. Simic’s ekphrastic answer to Cornell’s Untitled (Soap Bubble Set) is a short prose poem with the same title: Untitled (Soap Bubble Set), 1936 A soap bubble went to meet infinity.

 Gabriele Rippl

Here’s a cabinet containing the implements you need. There’s a classroom map of the moon, a clay pipe, a blue egg in a wineglass, and a child’s head on a block of wood. There are many versions of this dream machine, but the map, the pipe, and the glass are almost always present. The heavenly bodies are soap bubbles. They float in the empyrean, cradling the dreamer. The ephemeral bubble ascends into the wintry cold and silence of the sky. It’s the soul of the world ascending. Cosmogenies are soap bubbles. The father of our solitude is a child. A soap bubble has no content. After it has burst, there’s nothing left of it. The leap of a ballet dancer is a soap bubble, too. “Our hearts leap,” we say. This is the “long bright silence” of Nietzsche’s before which we stand in awe. The far and beyond and the near momentarily touch. The world is beautiful but not sayable. That’s why we need art. (Simic 1992:â•›54)

Simic’s prose poem links together the individual objects of Cornell’s box and offers an explanation of infinity and the beautiful, the two concepts evoked by Cornell’s box for Simic. Since these two concepts, infinity and the beautiful, cannot be expressed by verbal language, Cornell’s art is, according to Simic’s anthropological aesthetics, Â�indispensable. This ekphrastic poem delivers not simply a description of the content of this particular box, but also an interpretation of Cornell’s artistic method in general and the way it affects the onlookers and operates on their imagination. The poem suggests the shape of a box and is, therefore, a technopaignium, that is an example of imagic iconicity. I would like to argue that the indentations, however, break the allusion to the shape of the box and attest to Simic’s play with the paradigm of imagic iconicity. While the indentations reduce the impression of a clear-cut visual form of a box, they at the same time confirm the literariness of the poem and as such are an example of the paragon. My second example from Dime-Store Alchemy is “Matchbox with a fly in it” which is, arguably, Simic’s most perfect homage to Cornell’s boxes and their ‘magic’. Matchbox with a Fly in It Shadow box Music box Pill box A box which contains a puzzle A box with tiny drawers, Navigation box Jewelry box Sailor’s box Butterfly box Box stuffed with souvenirs of a sea voyage Magic prison An empty box.

(Simic 1992:â•›33)

“Matchbox with a fly in it” does not describe a particular box of Cornell’s. The reader is therefore free to develop his or her own story about the boxes mentioned in the poem.



Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy 

The best trigger for the onlookers’ own dreams and fancies is of course the “empty box” mentioned in the last line, which symbolizes the total freedom of imagination. Due to the almost complete absence of verbs, from a formal point of view this poem is Simic’s most radical one. It consists of a list of names, an inventory of different types of boxes, which consciously subverts the discursive and narrative functions of language. The poem is iconic in several respects. First of all, I would like to suggest interpreting the arrangement of the lines in this poem as imagic iconic representations of a fly’s movement in a closed matchbox; form here mimes meaning and reinforces it. Secondly, as pointed out by Max Nänny, the size and overall structure of a literary text “may also function as a global icon of its theme” (1986:â•›200). The brevity of this poem can thus be interpreted as a diagrammatic icon of what it is about, namely a matchbox, precisely because the small number of words used creates iconic isomorphism. While the arrangement of the lines, which actually differ in length, does not produce a Â�technopaignium, as in George Herbert’s (1957) “Easter Wings”, the lineation Â�nevertheless invites the readers to ruminate on the fact that poetry is not just an oral (i.e. linear) medium, but also very much a visual (i.e. spatial) one. Finally, the poem is also an example of intermedial iconicity since the almost complete absence of verbs can be read as Simic’s verbal attempt to imitate the ‘pure’ spatiality of visual art. His poem is thus an example of form miming other arts and media. My third example is “Cléo de Mérode”, a poem on Cornell’s box L’Égypte de Mlle. Cléo de Mérode. Cours élémentaire d’histoire naturelle (Figure 2).

Figure 2.╇ L’Égypte de Mlle. Cléo de Mérode. Cours élémentaire d’histoire naturelle (1940). Â�Construction, 4 11/16â•›xâ•›10 11/16â•›xâ•›7 1/4 in., Collection of Richard L. Feigen, New York

 Gabriele Rippl

Cornell’s box is a perfect example of the raging egyptomania of the first half of the twentieth century. On the inside of the open lid of this solid wooden box there is a graphic representation of the Egyptian goddess Hathor, the goddess of the sky, of love, happiness, dancing, and music (Waldman 2002:â•›42), who sits on a throne-like pedestal. In the box itself there are three divisions – a narrow one on the left hand side, a large compartment in the middle, and another narrow one on the right side. The narrow divisions display a few small objects such as the red petals of a plastic rose, a couple of tiny spoons, etc. These objects frame twelve small corked glass bottles, which contain a collection of objects and substances such as shells and pigments whose proximity is rather unsuspected. Many of Cornell’s boxes and collages remind one of Lautréamont’s famous phrase, “Beautiful as the chance meeting of a sewing machine and an umbrella on a dissecting table”, a description the Surrealists loved (Waldman 2002:â•›23). This is Simic’s verbal approach to Cornell’s ‘Egyptian box’: Cléo de Mérode Joseph Cornell could not draw, paint, or sculpt, and yet he was a great American artist. He roamed the streets of New York from the late 1920s till his death in 1972, foraging in used bookstores and junk shops. “My work was a natural outcome of my love for the city”, he said. One day in 1931 he saw some compasses in one shop window and some boxes in the next, and it occurred to him to put them together. Here are some of the things he found and placed in a box called “L’Égypte de Mlle Cléo de Mérode cours élémentaire d’histoire naturelle”, which he constructed in 1940:  oll’s forearm, loose red sand, wood ball, German coin, D several glass and mirror fragments, 12 corkstopped bottles, cutout sphinx head, yellow filaments, 2 interÂ� twined paper spirals, cut-out of Cléo de Mérode’s head, cutout of camels and men, loose yellow sand, 6 pearl beads, glass tube with residue of dried green liquid, crumpled tulle, rhinestones, pearl beads, sequins, metal chain, metal and glass fragments, threaded needle, red wood disc, bone and frosted glass fragments, blue celluloid, clear glass crystals, rock specimen, 7 balls, plastic rose petals, three miniature tin spoons for a doll house. Cléo de Mérode, by the way, was a famous ballerina and femme fatale of the 1890s. (Simic 1992:â•›16)

Like the majority of Simic’s poems in Dime-Store Alchemy, this piece also comes in the shape of a box, however with indentations: the narrative top part represents the open lid, the lower part the box itself. Also the shape symbolizes the form of Hathor sitting on her throne (see Figure 3). The poem therefore displays aspects of imagic iconicity. Simic’s text starts off with a narrative frame that gives biographical details



Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy 

about Cornell as well as an explanation of the artist’s method of collecting and reassembling objects that have lost their value and usefulness for their owners. The actual ekphrastic description of Cornell’s Cléo box is a list and, as such, also testifies to the interaction between iconicity and indexicality, which is what allows the evocation of pictures in the first place. Moreover, the fact that the actual ekphrastic description is framed by a narrative, stages the issue of framing. Lineation is used to divide the poem into two parts: the narrative framing discourse and a static, picture-like framed inner Â�description of a visual work of art. The lineation furthermore reinforces the poem’s emphasis on the process of seeing as culturally conditioned activity. In addition to traits of imagic iconicity, we can also detect strategies of diagrammatic iconicity, because Cornell’s twelve glass bottles are represented by the twelve lines of the inner ekphrastic text. Moreover, Simic creates his ‘inner’ text, i.e. the Â�ekphrasis within the larger one, by again reducing language to nouns and Â�adjectives. This has the effect that narrative time and movement come to a standstill. This Â�abstinence from verbs is responsible for the curious reduction of Simic’s poetic Â�language to its selective/associative axis. Simic tries to overcome the mediality of Â�language itself, its Â�‘sequentiality’ – to use a term introduced by Lessing (1853) in his well known essay Laocoön in 1766 – and thus to approximate poetic language to the graphic/plastic arts.4 This of course means that we also deal with intermedial iconicity. My final example is The Hotel Eden (Figure 3). Like the majority of Cornell’s boxes, The Hotel Eden assembles once again diverse objects such as a hotel advertisement and a bird (symbolizing heaven and freedom) in a box which seems to represent a room and which actually belongs to a whole series of hotel boxes.

Figure 3.╇ The Hotel Eden (1945). Construction, 15 1/8╛x╛15 3/4╛x╛4 3/4 in., National Gallery of Canada, Ottawa

.â•… For a discussion of Lessing’s ideas and the mediality of language, see Rippl (2005).

 Gabriele Rippl

Now, with The Hotel Eden things become rather complicated since none of Simic’s poems from the third section of his collection, named “Imaginary Hotels”, actually describes Cornell’s The Hotel Eden. However, many of the poems appear to be inspired by that work; this is the poem Simic placed next to the representation of the box: The Magic Study of Happiness In the smallest theater in the world the bread crumbs speak. It’s a mystery play on the subject of a lost paradise. Once there was a kitchen with a table on which a few crumbs were left. Through the window you could see your young mother by the fence talking to a neighbor. She was cold and kept hugging her thin dress tighter and tighter. The clouds in the sky sailed on as she threw her head back to laugh. Where the words can’t go any further – there’s the hard table. The crumbs are watching you as you in turn watch them. The unknown in you and the unknown in them attract each other. The two unknowns are like illicit lovers when they’re exceedingly and unaccountably happy. (Simic 1992:â•›47)

“The Magic Study of Happiness” is a perfect example not only of imagic iconicity, but also of diagrammatic iconicity. In spite of the spatial proximity of the picture of Â�Cornell’s box and Simic’s poem, it is not easy to connect the two, which is why we need analogy, i.e. metaphor. Again, we deal with a poem shaped as a box with one indented line: we are confronted with an interplay of imagic and diagrammatic iconicity. In the first paragraph of the poem the reader is presented with the description of a scene in which s/he can detect a diagonal line running from the table with the crumbs, through the window and up into the sky. Since this can be interpreted as a representation of a bird’s flight line, it connects the poem with the bird in Cornell’s box in a diagrammatic and metaphorical way. Birds eat bread crumbs and a bird is in the centre of Cornell’s box, therefore, one connecting element between the box and the poem is the term “crumb”. Another allusion to Cornell’s work of art is the term “happiness” in the title of Simic’s poem “The Magic Study of Happiness”, which refers to a central word in Cornell’s work, namely “Eden”, which is a place where happiness reigns. There are also other associations and iconic features: Simic’s crumbs and Cornell’s bird may also be connected by the image of the love bird, to which the poem implicitly refers with its title (‘happiness’), as well as the whole second paragraph about illicit lovers and Â�unaccountable happiness. The bird can also be seen as a metaphor of poetry. Since a bird is a medium of song, the bread crumbs in Simic’s poem can be read as another of the poet’s intermedial references, this time to music.5 .â•… In this context Robert Pinsky’s poem “Song” (Foer 2007:â•›185) – another homage to Cornell – is of interest: Air an instrument of the tongue, The tongue an instrument



Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy 

4.â•… Conclusion: Functions of ekphrasis and iconicity In his poetry collection Dime-Store Alchemy, Charles Simic implicitly discusses problems of representation and the relationship between life and art by Â�engaging in and exploiting imagic, diagrammatic and metaphoric iconicity in order to Â�produce Â�intermedial effects. To evoke Cornell’s boxes and to create verbal equivalents for the visual artist’s method, Simic has developed a rather special poetics. His prose poems often “adopt features of the imitated medi[um], features that stand out as relatively alien to the literary medium” (Wolf 2003:â•›354). As we have shown, Simic’s poems are verbal attempts to approach the condition of the other, the visual medium of Cornell’s boxes and assemblages. Many of Simic’s intermedial poems are highly self-reflexive intersemiotic translations and transformations which Â�foreground the iconic Â�character of verbal signs, exactly because they involve an imitation of the other medium’s Â�structures and indirectly discuss the advantages as well as Â�restrictions of their own medium, written language. While trying to evoke visual and pictorial qualities via Â�language, Simic’s poems, just like other ekphrastic texts which combine iconic and indexical functions of language, investigate the Â�specificities of and tensions between verbal and visual media. In our Â�postmodern times, when the visual seems to have become the leading cultural paradigm, Simic’s intermedial iconic poems invite us to think about the collaborations and Â�competitions of different semiotic systems in contemporary Western culture. As such, Simic’s poetry functions as a meta-medium, which has the potential to reflect critically on the role of other media. This shows that ekphrasis and iconicity are not simply highly specialized literary modes, but rather meta-discourses, the investigation into which helps us to understand representation and communication in our postmodern world.

Of the body. The body An instrument of spirit, The spirit a being of the air. The bird a medium of song. Song a microcosm, a containment Like the fresh hotel room, ready For each new visitor to inherit A little world of time there. In the Cornell box, among Ephemera as its element, The preserved bird – a study In spontaneous elegy, the parrot Art, mortal in its cornered sphere.

 Gabriele Rippl

In order to explain the recent success of ekphrases, one needs to take their Â�functions into account. First, literary evocations of visual phenomena invite us to extend the reception time we usually dedicate to pictures, thus they reduce our Â�attention Â�deficit induced by technical and electronic media which produce deluges of digital Â�pictures. Within a media historical and anthropological framework of Â�interdisciplinary research these descriptions help us to investigate how people react and adjust to the new forms of verbal-visual communication and how they preserve their capacity of imagining and mentally visualizing things in the face of the deluges of preconceived and prescribed images we are now exposed to. Second, descriptions of images not only analyse differences between the media, but they also investigate our modern and postmodern cultural archives and are thus important with respect to cultural memory. Third, ekphrases, although they add a static element to the text, very often trigger a narrative and induce new aesthetic qualities. Finally, there is one additional important function of ekphrases, namely that they help us to reconstruct the concepts of representation and mimesis. In nineteenth-century realistic novels, ekphrases and descriptions in general were handmaids, supporting the illusion of a fictive world, semantic condensation and heightened memorability. Modernist and postmodernist literary ekphrases, however, often subvert this function and investigate the nature of representation itself, which accounts for the highly self-reflexive quality of contemporary ekphrastic texts. Their investigation of medial differences between texts and images, on the one hand, demonstrates the inadequateness of verbal representations of visual data and Â�acknowledges that each verbal representation is necessarily also a misrepresentation. On the other hand, the sheer number of ekphrastic descriptions in postmodernist texts seems to insist on the fact that the verbal medium cannot be forgone, that “all media are mixed media”, as W.J.T. Mitchell (1994:â•›94–95) has reminded us repeatedly over the last years. The imaginary power of literature manifests itself in the abundant intermedial transgressions of contemporary literature whose performative character, as Christina Ljungberg (2010) has recently pointed out, allow readers to access new levels of meaning, thus generating something novel and unique.

References Amsterdam–Zürich Iconicity Research Project. University of Amsterdam. University of Zurich. 2009. 〈http://www.iconicity.ch〉 (15 December, 2009). Ashton, D. (ed.). 1974. A Joseph Cornell Album. New York NY: Viking. Foer, J.S. (ed.). 2007. A Convergence of Birds. Original Fiction and Poetry Inspired by the Work of Joseph Cornell. London: Penguin. Heffernan, J.A.W. 1991. Ekphrasis and representation. New Literary History 22: 297–316.



Iconicity and intermediality in Charles Simic’s Dime-Store Alchemy 

Herbert, G. 1957. Easter Wings. In The Metaphysical Poets, H. Gardner (ed.), 121. Â�Harmondsworth: Penguin. Jäger, Ludwig. 1999. Die Sprachvergessenheit der Medientheorie: Ein Plädoyer für das Medium Sprache. In Sprache und neue Medien [Jahrbuch 1999 des Instituts für Deutsche Sprache], W. Kallmeyer (ed.), 9–30. Berlin: Walter de Gruyter. Josipovici, G. 1980. That which is hidden is that which is shown; that which is shown is that which is hidden. The London Review of Books 2(22) (20 November–3 December): 17. Josipovici, G. 1987.The principle of order. In P.N. Review 14(2): 44–45. Lessing, G.M. 1853. Laocoon: An Essay on the Limits of Painting and Poetry, E.C. Beasley (transl.). London: Longman. Ljungberg, C. 2001. Iconic dimensions in Margaret Atwood’s poetry and prose. In The Â�Motivated Sign [Iconicity in Language and Literature 2], O. Fischer & M. Nänny (eds), 351–366. Amsterdam: John Benjamins. Ljungberg, C. 2003. Diagrams in narrative: Visual strategies in contemporary fiction. In From Sign to Signing [Iconicity in Language and Literature 3], W. Müller & O. Fischer (eds), 183–199. Amsterdam: John Benjamins. Ljungberg, C. 2005. Diagrams and diagrammatization in literary texts. Recherches Sémiotiques/ Semiotic Inquiry 24: 99–115. Ljungberg, C. 2010. Intermedial strategies in multimedia art. In Media Borders, Multimodality and Intermediality, L. Elleström (ed.), 81–98. Basingstoke: Palgrave Macmillan. Mitchell, W.J.T. 1994. Picture Theory. Chicago IL: Chicago University Press. Müller, W. & Fischer, O. 2003. Introduction. In From Sign to Signing. [Iconicity in Language and Literature 3], W. Müller & O. Fischer (eds), 1–20. Amsterdam: John Benjamins. Nänny, M. 1986. Iconicity in literature. Word & Image 2(3): 199–208. Nöth, W. 2001. Semiotic foundations of iconicity in language and literature. In The Â�Motivated Sign [Iconicity in Language and Literature 2], O. Fischer & M. Nänny (eds), 17–28. Â�Amsterdam: John Benjamins. Paz, O. 1991. Objetos y apariciones. In The Collected Poems of Octavio Paz, 1957–1987, E. Â�Weinberger (ed. & transl.), 404–405. New York NY: New Directions. Rippl, G. 2005. Beschreibungs-Kunst. Zur intermedialen Poetik angloamerikanischer Ikontexte (1880–2000). München: Fink. Rippl, G. 2010. English literature and its other: Toward a poetics of intermediality. In Â�ImageScapes, C. Emden & G. Rippl (eds), 1–27. Frankfurt: Peter Lang. Simic, C. 1992. Dime-Store Alchemy. The Art of Joseph Cornell. Hopewell NJ: Ecco Press. Waldman, D. 2002. Joseph Cornell. Master of Dreams. New York NY: Abrams. Wolf, W. 2003. Intermedial iconicity in fiction. In From Sign to Signing, [Iconicity in Language and Literature 3], W. Müller & O. Fischer (eds), 339–360. Amsterdam: John Benjamins Wolf, W. 2005. Intermediality. In Routledge Encyclopedia of Narrative Theory, D. Herman, M. Jahn & M.-L. Ryan (eds), 252–256. London: Routledge.

Words, like shells, are signs as well as things Heilna du Plooy

North-West University, South Africa South African (Afrikaans) novelist Ingrid Winterbach is known for her individual style, the originality of voice in her novels, as well as for an unusual and ironic perspective on reality. This article focuses on Winterbach’s ability to integrate word choices and narrative technique so that passages in her fiction can be described as highly iconic. Winterbach exploits language from different domains, discourses, and semantic fields, as well as archaisms, resulting in incongruous phrasing, ambiguity, and irony. Word play and lists of words activate semantic values by relying on the materiality of words. In Winterbach’s novels, words do not only create unexpected and defamiliarized meanings on account of illogical systems of creative thought, but they also bear witness to the previous ideologically suspect meanings in each word’s history, thereby reactivating the kind of archaeology of thought and language described by Foucault.

1.â•… Introduction Iconicity is such a prominent feature of poetic texts that some theorists regard Â�iconicity as central to poeticity as such. The question then arises whether iconicity also figures in prose texts, and what forms iconicity would assume in a prose text. In this article I would like to explore specific structures and techniques in narrative prose texts which could possibly be regarded as manifestations of iconicity. I will analyse and interpret passages from two novels by the South African (Afrikaans) novelist Ingrid Winterbach and will argue that these passages can be considered as iconic on account of the way in which “form enacts meaning”, which implies that these passages might be seen as examples of “form miming meaning” (Fischer & Nänny 1999). Ingrid Winterbach is known for her individual style, the originality of voice in her novels as well as an unusual and ironic view of reality. Language and narrative technique are integrated to such an extent that specific passages in these novels can be described as iconic. Apart from exploiting more subtle aspects of language as such to achieve effects which can be described as iconic in different ways, Winterbach’s novels are also imbued with an acute awareness of what lies beneath the surface of things, Â�thematically as well as linguistically, and with the history of words and Â�history as such.

 Heilna du Plooy

In this article, I will therefore firstly analyse passages from the novels as Â�examples of specific forms of iconicity and then I would like to explore the way in which Â�Winterbach employs the archaeology of thought and discourse and the Â�archaeology of linguistic signs in these iconic passages. It appears that the Â�thematic development in the iconic passages in the novels also reflects the main themes which are developed in the novels. I do not intend to name or tag every manifestation of iconicity in the novels because the variety of linguistic and literary phenomena employed by poets and novelists is vast and variable (Du Plooy 2008). I take the diagram and explanation of the different kinds of iconicity in the introduction to Form Miming Meaning (Fischer & Nänny 1999) as my point of departure (see Figure 1),1 also keeping in mind, however, the following remark made by Burke (2001:â•›33): […] despite the fundamental differences between the two categories of iconicity [imagic and diagrammatic], it is important that they are not viewed as definite, isolated entities; rather, it is far more productive to see them as two concepts occupying the opposite poles of a single iconic continuum.

It seems that this might also be true for other forms of iconicity which lie between types or blend effortlessly or which could fit into different categories at the same time (Leech & Short 1981; Ungerer & Schmidt 1996). There are a few other theoretical concepts which will determine my argument. Nänny emphasises that: .â•… Figure 1.╇ Types of iconicity (Fischer & Nänny 1999: xxii) Imagic

– Oral/aural – Tactile – Visual

Iconicity – Isomorphism

Structural

– Centrality vs. peripherality – Distance/proximity – Sequential ordering – Markedness – Repetition

– Iconicity of motivation Diagrammatic Semantic

– Metaphor (cognitive) – Metaphor (grammatical)

– Analogy – Grammaticalisation



Words, like shells, are signs as well as things 

[…] all iconicity is semantically motivated. It is the meaning of a text or the semantic context that alone detemines whether a linguistic sign or literary element has iconic force. […] Hence, iconicity, like beauty, is in the eye of the beholder: the perception of imitative form depends on the reader’s capacity to see connections, to perceive similarities between the meaning of a sign or text and the formal means used for its expression. (1999:â•›174)

In my analysis, I will therefore keep in mind that creative writers are primarily Â�concerned with meaning, but meaning can be generated in ways other than using the semantic value of words. Also, any text becomes meaningful through a reader and therefore different readings of texts provide different interpretations. In literary studies, it is an accepted notion that it is the representational Â�function of language which is responsible for the literary quality in a text, so that referentiality plays a secondary though still indispensable role (Bronzwaer 1993:â•›25, 27; Wybenga & Cloete 1992:â•›178–181; Lotman 1972:â•›7–49). For instance, the narrative force that moves the story forward in a novel is kept in balance by introducing Â�elements and aspects (motives and indices in a multitude of forms) and techniques which Â�complicate and expand the meanings of the narrative elements and redirect the attention of the reader to the representation in the text. Iconic literary strategies or techniques, their value and their functions, have been described in different terms by theorists Â�working in Â�different traditions (see De Cuypere 2008), but Â�basically literary techniques are intended to have a defamiliarising effect on the reader, to redirect the attention from the referential content or the signified to the text itself, to the signifier and eventually to the represented meaning. In many instances, this redirection is a form of iconicity. Also, the attempts to describe and analyse literary techniques and functions will inevitably be theoretically and terminologically opaque to a certain degree, because literary texts are not and do not want to be exact and the playfulness with which language and textuality are used and abused constantly create new Â�possibilities for generating meaning. I will therefore feel safer to work within broader categories of iconicity than with closely defined concepts. I will attempt to describe the iconic features but more importantly my aim will be to interpret specific passages in relation to the text as a whole. Poetic texts create their own poetic vocabulary and conceptual space, which cannot always be forced into some preconceived theoretical paradigm. Using theoretical Â�concepts to explain and discuss aspects of Winterbach’s novels is therefore done with great care, so as not to erect rigid walls and pillars in a beautiful and fragile Â�novelistic landscape of moving light and shadow.

2.â•… The iconic force of language in juxtaposition A characteristic feature of Ingrid Winterbach’s oeuvre is the way in which she Â�defamiliarises language. This is done on the level of the choice of words and the Â�coining

 Heilna du Plooy

of phrases as well as on the levels of dialogue and the narrative itself. Â�Winterbach acknowledges her preference for unusual and estranged language when she says in an interview: “Language is important. The juxtaposition, in the dialogue used by different characters, of for instance the language of religion, of science, of the arts. Language that offers resistance” (Van Vuuren 2004, my translation). Winterbach uses words and phrases from different domains, discourses and semantic fields as well as archaic language, and often coins phrases in which the incongruous relation between the words is ambiguous and ironic. Archaic or at least old-fashioned or underminingly mock pedantic words and words which are playfully rowdy or even coarse, are used with the same artistic abandon. Between words which are syntactically adjacent but which originate from different semantic domains or registers, a distinct tension develops. In The Elusive Moth (2005, Â�original title Karolina Ferreira [Viljoen 1993]) the main character is described as a “begenadigde Â�danser” (‘a blessed dancer’) where “begenadigde” means blessed, but in Afrikaans the word “begenadigde” has the distinct religious connotation of being exceptionally and intensively blessed with an added connotation of holiness. She therefore refers to the character’s dancing ability by qualifying it in religious terms, thus creating an ironic result. The ambivalence between the words correlates with the ambivalences in the character’s personality. She is depicted as a person in a process of Â�‘becoming’, still neither here not there, like the moth of the title of the book, in a stage of transition. In this example I use the phrase from the original Afrikaans novel but in the Â�following discussion and analyses of the novels I will make use of the translated Â�versions of the novels, indicating where necessary salient aspects of the texts in the original language. As the iconic phenomena in these novels have to do with structure and style, the semantics of the language is not the only concern in an analysis. The way in which the novels are translated, retaining the structure and the pattern but using different words, emphasizes the iconicity even more. Other phrases in which a semantic tension between adjacent words can be detected are, for instance, “the woman is heart-rendingly beautiful” or “poignantly beautiful” (Winterbach 2005:â•›9). The author uses the phrases “veiled funeral Â�exuberance” to describe an emotion (Winterbach 2008:â•›284) and “unabashedly Â�sorrowful” to refer to a character’s disposition (Winterbach 2008:â•›287). In phrases like these, which are used frequently, the tension between or among adjacent terms draws the attention of the reader to the formal aspects of the language construction because the semantic incongruence which is manifested on the linguistic level mimes the thematic Â�incongruence which underlies the specific passage and even the text as such. The redirection of focus from meaning to text points at a similarity “between the meaning of a sign or text and the formal means used for its expression” (Nänny 1999:â•›174.). In these cases the



Words, like shells, are signs as well as things 

Â� narrator’s ambivalent view of the world and of life in general, which figures strongly in all these novels, is mimed by the ambivalent Â�language Â�constructions (see also Du Plooy 2000, 2006). In all her novels, Winterbach intersperses the ongoing discourse with scientific terminology or academic language (from different fields such as geology, zoology, linguistics, psychology, theology, or the theory of art), but also with words from South African history, political terms, military terminology, the language of women or specifically male language as well as words or phrases which can be regarded as taboo in contemporary language use (John 2004; Du Plooy 2006; Botha & Van Vuuren 2002, 2006). These techniques can be analysed as forms of diagrammatic iconicity, because the blending of terms from unrelated domains forces the reader to construct new meanings. That implies that the new meanings are called forth by the Â�defamiliarising use of language on the formal grammatical, syntactical and/or syntagmatic Â�levels so that the form indeed mimes meaning. More specifically, one could perhaps speak of iconic proximity (Ungerer & Schmidt 1996) or iconic juxtaposition which is described as “words that are physically adjacent, evoking a sense of temporal, Â�psychological or locative affinity” (Burke 2001:â•›31; see also Leech & Short 1981:â•›236). In these instances, the adjacent affinity is undermined and becomes an adjacent opposition or juxtaposition, creating irony through extra tension and stronger defamiliarisation. In a similar way, language in dialogue and narration is used in a defamiliarising way in Winterbach’s novels. Geological terminology is used to describe appreciation of the landscape in To Hell with Cronjé (2007:â•›10) and the fascination with shells is expressed in the appropriate scientific vocabulary in The Book of Â�Happenstance (2008:â•›301–302). The words, phrases, and terminology used by characters come from different and even unrelated discourses, so that the narration becomes Â�textured and functions on different levels. In ways which are similar to the Â�coining of phrases, Winterbach succeeds in allowing language to enact its own meaning. The word enactment seems to me to be the key to her poetics: she does not want to simply narrate: she wants the language as such to generate meaning in her narration as well. Burke (2001) uses the term enactment in his explanation of the Â�relation between iconicity and the emotions the literary text can evoke. In her novels Winterbach creates a Â�pervasive form of enactment because language is constantly foregrounded to stimulate in the reader an awareness of language as such and of the more subtle aspects of language. The author implements the metalinguistic abilities of language continuously so as to generate extra meanings. I want to illustrate this by referring to examples from two novels, To Hell with Cronje (2007) and The Book of Â�Happenstance (2008 [2006]).

 Heilna du Plooy

3.â•… To Hell with Cronjé To Hell with Cronjé (2007) (original Afrikaans title Niggie 2002) is set against the Â�backdrop of the last months of the Anglo-Boer War (1899–1902).2 Despite the historical context, the novel focuses on the specific circumstances and personal dilemmas of individual characters and their struggle to survive physically and psychologically. The novel is concerned with how people succeed in finding a way to move on and to continue with the very ordinary business of living, not about how people survive or can survive, but about the fact that they simply do survive. The Anglo-Boer War (1899–1902) is an important event in South-African Â�history and had an enormous influence on the social and political developments in the Â�country in the course of the twentieth century. The characters in the novel Niggie are victims of the war, but they are also survivors of the war. Shortly before the end of the war, three men, Willem Boshoff, Reitz Steyn, and Ben Maritz, are instructed to escort a severely traumatised young man, Abraham Fouché, to his family in the eastern Free State. They join a group of burgers in a mountain hide-out and wait for instructions from the general. Most of the men in the camp suffer from war wounds or trauma and cannot take part in active fighting. All of them, including Reitz and Ben, realise that they have probably lost everything dear to them on a personal as well as a political level. It is evident that they will lose the war against the mighty British Empire, and they do not know what fate has befallen their families. They are aware of the fact that most farms have been razed and burnt to the ground, and that women and children have been taken to concentration camps, which are notorious for the alarming death rate of the inmates. The problems worsen when Reitz and Ben are wounded in an ambush. It is not clear by whom the ambush is set up, and it might possibly be by their own men. They are found and taken to a farm by women, who care for them while they are recovering. Consequently, the two men spend the last months of the war outside the main stream of historical events and have to work through their personal traumas while the country as a whole is facing a bleak future. The dialogue used in the novel can be described as iconic in more than one way. The discussions among the men on their journey and in the mountain camp not only reflect the different reactions to the war and the imminent outcome of the war, but also the ways in which people thought and expressed themselves in 1902. These Â�conversations can be read as indicative of aspects of the psychological Â�profile of the Afrikaner community of the day. The naïve assumptions, the cultural and

.╅ In 2004 Winterbach received the Hertzog Prize, the most prestigious Afrikaans literary award for the original Afrikaans novel Niggie (directly translated: Cousin).



Words, like shells, are signs as well as things 

Â� religious beliefs, even the misconceptions, are expressed openly, and it is fascinating to Â�recognise strains of thought which still determine discourses in contemporary South Africa, often in a disguised or even unconscious or underlying form. Some of the language styles used by different characters are based on diaries Â�written during the war (Botha and Van Vuuren 2008). Japie Stilgemoed’s formal and stylised, almost pedantic way of talking, his “merciless solemnity” (Winterbach 2007:â•›72), Kosie Rijpma’s ineffectual piety and the devoted seriousness of Willem’s contributions to conversations are contrasted by Gert Smal’s coarse vocabulary, his unabating verbal abuse of the Boer leaders and the British enemy alike (Winterbach 2007:â•›63–67, 72–76). There is also the black man, Ezekiel, whose speech and actions indicate that he can be seen as the perfect example of the “colonised mind”, Â�reciting history and religious passages better than any of the Afrikaner men in the camp Â�(Winterbach 2007:â•›39). The young man, Abraham, lost his speech when his brother was killed next to him in a battle. The seemingly nonsensical words, which are the only form of language that he is capable of, provide an example of the unexpected Â�meanings created by language. The sounds he utters do not mean anything as such, but do resemble words, and parts of words and phrases actually used in the conversations around him, echoing the gist of their words back to the speakers. In all these cases, the characters are not described by a narrator, but by the Â�language that they use, so that the language in its formal aspects and register mime the Â�meaning of the passage. The concern is not only with the meaning in the context of what the characters are saying, but also with meaning in the sense that the language as such represents specific types of people and specific underlying discourses and Â�philosophies prevalent in the South African community.

4.â•… Survival strategies The novel suggests that some people can and do use resources of psychological and spiritual strength to activate the regenerative powers of the psyche in times of stress. Survival is partly dependent on a person’s ability, consciously or unconsciously, to find meaning or to create meaning despite difficult circumstances. The suggestion is that it is necessary to look at circumstances differently in order to find a different way of thinking. One of the ways in which new meanings can be generated is the Â�exploitation and reinvention of language, by using words in new contexts and by Â�playing with words. Creativity is seen as the key to the ability to approach life from another Â�perspective, to distance oneself momentarily from pressing circumstances and transcend the Â�immediate pressure and stress. When the four men ride over the endless plains of the Karroo on their way to the eastern Free State, they play word games. They start with one word and then Â�improvise

 Heilna du Plooy

by providing another which is morphologically or semantically related to the first word. The words begin with the consecutive letters of the alphabet so that the Â�emphasis is primarily on alphabetical order and meaning is seemingly random. Though it appears as if meaning is indeed less important than structure and sound, the words in the sequence are also linked semantically. Initially, the culturally Â�determined meanings are secondary to the underlying pattern of the alphabetical order and the loose semantic connections, but a new pattern of association develops, and in the end the sequence of playful words almost becomes a poem, creating surprising and Â�unexpected Â�meanings. An analysis of the excerpt shows that it starts off with Â�sombre, ominous, and even threatening words, but as the sequence develops, there is a Â�loosening up, and Â�eventually the words seem to indicate new spaces, as if there were an opening up instead of a closing down of meaning. On the fifth morning after they have left the farm, there are vultures in the sky. There must be a cliff nearby, Ben remarks, watching the birds. “Buzzard,” says Reitz, “predatory bird.” “Bird of prey, that hunts animals for food,” says Ben. “Bone,” says Reitz, “the remains after death.” “Botfly,” says Ben, “dipterous fly with stout body.” “Carrion,” says Reitz, “dead, putrefying flesh.” “Carrion crow,” says Ben, “bird feeding mainly on carrion.” “Devonian,” says Reitz, “geological period.” “Devil,” says Willem, “lord of the kingdom of evil.” “Devil’s coach horse,” says Ben, “large rove beetle.” “Eland piss,” says Reitz, “the piss of an eland.” “Everlasting,” says Ben. “plant used as remedy for a cold.” “Goldfield,” says Reitz, “district where gold is found.” “Gold” says Willem. “streets of jasper and gold.” “Good heavens, Willem!” Ben exclaims. “Goldcrest, with its heavenly warbling.” “Heavenly body,” says Reitz, “celestial object.” “Hay,” says Willem, “to feed the horses.” “Helpful,” says Ben, “more helpful than Peternella one could not hope for.” Even Willem smiles. They carry on in that vein, but keep their eyes on the distant low mountain range, steadily acquiring more substance. (Winterbach 2007:â•›32–33)

Although the alphabetical order is the primary pattern, semantic associations also determine the sequence. The first part of the game is dominated by words associated with death and decomposition, also with the way in which nature dispenses with �animal carcasses. The underlying suggestion is that if these men were to die on the veld the same would happen to their bodies. Then there is mention of the devil followed by a reference to the urine of an eland, the largest African antelope. Eland is followed by



Words, like shells, are signs as well as things 

everlasting and then an unexpected transition to another register. Everlasting is a plant that can cure a cold, cold becomes gold, and gold calls forth gold and jasper. Religious connotations are activated by gold and jasper, because in 1902 religious people would immediately associate gold and jasper with the biblical description of the walls of the New Jerusalem as described in the Book of Revelations 21:18. Ironically these heavenly associations are echoed by the meaningless exclamation “Good Heavens”, which is the implicit author’s way to ironise, even undermine and question the characters’ restricted religious frame of reference. The noteworthy aspect is that the game does indeed make them feel better. Even Willem allows himself to laugh. What is important to my argument is that the characters do not plan the transition in the sequence of words. It happens on account of the alphabetical pattern and the likeness in sound between consecutive words, so that the words acquire a power of their own and seem to act out the change in tone. What is also important is that the original Afrikaans version of the novel is not translated semantically in this passage. In the translation other words are used but the patterns, the alphabetical order and the movement from desolation and death to relief and Â�playfulness remain the same. That would be the ultimate proof that the words are not chosen for their semantic value alone, but for their iconic functions, that is to perform a process, to enact meaning so that the form mimes the meaning. The meaning must remain the same in the translation and therefore the pattern has to be kept intact and must be recreated irrespective of the words. The implicit author wants to illustrate how language can open up new passages of thought, that words can stimulate Â�creative thinking, that words can make an appeal on underlying psychological resources, which facilitate psychic regeneration. Another striking example of a word game with yet another result, is when the men on their journey on horseback play with the word “kaffir”, a word which has come to be associated with racial prejudice and is still regarded as an abusive term in South Africa. In 1902 the word was not used pejoratively, though some of its connotations might have been to refer to objects of lesser quality. It primarily indicated that something was indigenous and therefore was often used as part of the names of indigenous objects. In contemporary South Africa, the South Africa in which the novel was published in 2002, this word is an absolute taboo; you could get a lawsuit filed against you for using it. And yet Winterbach uses the word repeatedly and excessively in this passage. Once they have saddled the horses and resumed their journey, Willem declares: “Kaffir-melon preserve – mouth-watering to say the least.” “Kaffir thorn,” Ben says, “a kind of tree.” “Kaffir cow,” Willem says, “cow belonging to a Kaffir.” “Kaffir sheeting,” says Reitz, “a thick, soft cotton.” “Kaffir cherry,” says Ben, “a raisin bush.” “Kaffir beer,” says Reitz, “beverage drunk by Kaffirs.”

 Heilna du Plooy

“Kaffir work,” says Willem, “work not fit for white people.” “Kaffir copper,” says Ben,“a large russet butterfly.” “Kaffir hangman,” says Reitz, “an executor or oppressor of Kaffirs.” “Kaffir chief,” says Reitz, “bird with an extremely long tail.” “Kaffir pebble,” says Reitz, “pebble found in gravel to indicate the presence ╇ of diamonds.” “Kaffirboom leaf miner,” says Ben, “insect found on the kaffirboom.” “Kaffir grave,” says Willem, “dwelling place of Kaffirs.” “Kaffir swallow,” says Ben, “a kind of swift.” “Kaffir pound,” says Reitz, “nickname for a penny.” “Kaffir war,” says Willem, “war between white people and Kaffirs.” “Kaffer-corn midge,” says Ben, “small gallfly with bright wings.” “Kaffir corn,” says Reitz, “fine, diamond-bearing gravel.” “Kaffir missionary,” says Willem, “missionary that works among Kaffirs.” “Kaffir crane,” says Ben, “large bird with long legs and neck.” “Kaffir half-crown,” says Reitz, “another name for a penny.” “Kaffir nation,” says Willem, “nation consisting of Kaffirs”.(Winterbach 2007:â•›34)

The repetition itself forces the reader to consider different possibilities carefully, because it is exceptional to use a word with such an abusive reputation and connotation in such a conspicuous way. An analysis would, in the first place, have to point out that the unselfconscious way in which the characters use and play with the word, illustrates that the word was once (definitely in 1902) innocent of the later ideological baggage. Used in so many senses, which are simply denotative, it is as if the word is cleared of its narrow racial connotation and is revived to its original ability to refer to a wide range of indigenous objects. Clearly, in 1902 it was an ordinary word to indicate well-known objects and it was used to refer to things that people liked. It is, however, exactly the former innocence of the word, that irrevocably calls forth the ideological marking which developed during the past century. The passage calls forth and reminds the reader of all the abusive and pejorative ways in which the word has been used and in which the word has been perceived in the course of a century marked by increasing political and ideological injustice. Winterbach’s novel is imbued with a subtle ironic tone which is an indication of the implicit questioning of the ideological thinking behind the linguistic phrases used ‘naturally’ by the characters and also by contemporary speakers. This links up with the whole issue of language being the archive of ideological thinking, with what is known as the archaeology of the sign. The use of old and well-worn phrases in a new and ironic context questions the conceptual thinking behind the words. The unconscious of the language itself is examined in the novel, and the suppressed content is brought to the surface, recycled and used in a new way. The initial uncontaminated Â�meanings of words and phrases are thus foregrounded, but the reader is also constantly reminded of the ideological and political baggage accumulated by the same words and phrases



Words, like shells, are signs as well as things 

in the course of a century. The words thus also enact the history of the community in which they have been used; they cannot shed their baggage and therefore language is never innocent. This insight is not narrated or explained but enacted by the words, by language itself in the text. The passage is not primarily concerned with referential meaning and the characters themselves are not aware at all of what their game would mean or reveal to people in the twenty-first century. The narrator provides no comment on the Â�passage, but the reader cannot miss the implications which are acted out by the Â�repetition of a word which has become ideologically unacceptable and unusable with time. This is to me a form of diagrammatic iconicity, or structural iconicity on account of repetition. In both examples there are, however, aspects of the emotional Â�involvement in the enactment that is executed in the text. In the first case, the Â�characters as well as the readers will be affected by the relieving movement in the word game, and in the second example the reader cannot but feel ideologically confronted by the implied history of the word. In his book The Order of Things, Foucault (1973: xi) concerns himself with the underlying codes of knowledge, the epistemological space specific to a Â�particular period, which implies a positive unconscious of knowledge on a level which he describes as archaeological. In his view, the historical analysis of Â�scientific Â�discourse, and that goes for discourse in general as well as literary discourse, should be Â�concerned with “a theory of discursive practice” (Foucault 1973:â•›xiv). Applied to the literary use of language as well as to the analysis of literature, one could Â�therefore say that the Â�discourses of a specific historical period are marked by the underlying Â�epistemological structure of thought. Consequently, one can assume that Â�language, discourse, discursive practices and style are the symptoms and data of this  Â�underlying epistemology, that is of the archaeology of thought (Foucault 1973: xiv; see also Deely 1986). Another idea of Foucault’s which is important to my argument is how the accepted structure of thought can limit the possibilities of human thinking and creativity. He illustrates this by referring to a story by Jorge Luis Borges (1993) about a “certain Chinese encyclopaedia” which presents an alternative classification of animals. According to this encyclopaedia “animals are divided into: (a) belonging to the emperor, (b) embalmed, (c) tame, (d) sucking pigs, (e) sirens, (f) fabulous, (g) stray dogs, (h) included in the present classification, (i) frenzied, (j) innumerable, (k) drawn with a very fine camelhair brush, (l) et cetera, (m) having just broken the water pitcher, (n) that from a long way off look like flies” (Foucault 1973: xv). Foucault’s comment on this classification is: “In the wonderment of this taxonomy, the thing we apprehend in one great leap, the thing that, by means of the fable, is demonstrated as the exotic charm of another system of thought, is the limitation of our own, the stark impossibility of thinking that” (Foucault 1973: xv).

 Heilna du Plooy

The language games in To Hell with Cronjé illustrate how another system or Â�pattern influences the character’s thinking and mindset to become more creative and open to alternative reactions and solutions. In the second instance, in the case of the repetition of the word “kaffir”, one can also see how the reader is confronted by the underlying ideological structure of his thought: first when he is shocked out of his wits by the use of an abusive word, then realising that the word was absolutely normal in 1902, and eventually by the realisation that the word carries the history of the century in itself, that the word becomes the ideological archive of the wrongdoing of a century.

5.â•… The Book of Happenstance The last group of examples I want to discuss comes from The Book of Happenstance (2008). In this novel, the main character and narrator, Helena, is an assistant in a Â�project in which words that have fallen into complete disuse and words seldom used in Afrikaans are listed. Helena has a collection of shells for which she cares very much, which is actually exceptionally valuable to her. Her shells are stolen and, as a final insult, her flat is befouled by the thief. The novel also tells how Helena discovers her boss’s dead body in their office. While trying to come to terms with his death, she Â�considers all the traumas in her past, including the recent loss of her shells. She grieves for the shells in an excessive way which her friends and even her lover cannot understand. To her, the shells are symbols of all those seemingly useless things which people nevertheless value and which make their lives bearable and worthwhile. The same goes for the archaic words with which she spends her days. The novel contains numerous lists of these archaic words. In many cases, these words bear witness to cultural practices which have also become extinct and Â�symbolise the loss of cultural values. This is especially true of all the nouns which are compounds of the word ‘death’ (Winterbach 2008:â•›35–37).3 Helena and Theo Verwey are indeed Â�working on the letter d and the words about death when her shells are stolen. The Â�number and nature of words about death that are no longer used and have become archaic, makes her realise to what an extent death has become depersonalised in Â�modern times.

.â•… These lists are very difficult to translate: the Afrikaans words have been retained with Â�explanations in parenthesis which to my mind interferes with the rhythm of the reading process. Nevertheless, translating the words would serve no purpose because these specific compounds probably do not exist in another language nor would they necessarily contain “death” as one part of the compound. In Afrikaans, similar to Dutch, words denoting new concepts are often formed by using compounded words and these words are spelled as one word, because they are regarded as one concept. In English for instance such words are written separately and therefore the concepts of compounds would definitely get “lost in translation“.



Words, like shells, are signs as well as things 

The words referring to death rituals, the clothes, the music, the Â�procedures Â�surrounding death, have fallen into disuse because the actions are no longer practised and the objects no longer used. Eco (1983:â•›525) writes about the relation between signifiers and signifieds and meaning when he explains the title of his novel The Name of the Rose. The word ‘rose’ has been used so often and in so many different ways and it has been associated with so many meanings that the word as a sign has become empty. But the word still exists as a sign and can be filled with new meaning. Eco refers to the mediaeval nominalist view of language when he explains that in time all things disappear or cease to exist but that words are signs and the signs survive. The same argument can be used to explain the value of Helena’s shells. The snail which used to live in the shell has died, has had to die for the shell to become available to be kept as an object of beauty. The shell is therefore a sign of the snail, of the life which was lived within it (Winterbach 2008:â•›301–302). The snail has died but the shell survives. However, the shells are also valuable of their own account. They are beautiful signifiers, emptied from the original reason for their existence, but now they can have many different meanings. To Helena they symbolise everything that is beautiful and valuable, not in a pragmatic sense, because nobody else seems to think that this a great loss, but in a personal way they symbolise the right of an individual to a personal value system, to the right to cherish and care for things which are valuable according to a personal choice, according to a personal sense of beauty. Helena, who is by no means a religious person, explains to the policeman why she is so upset about her shells: “Why do you like them,?” he asks. “Because they are beautiful,” I reply, “and because God made them”.  (Winterbach 2008:â•›14)

In an indirect way the novel is concerned with the relation between signifiers and signifieds, between signs and their meanings. It can happen that when things cease to exist, the words denoting them can still survive for some time, but as with the rituals of death, the words will eventually also disappear. Because the words are no longer used, they will be forgotten. But this process can work both ways – in some cases the things, objects or actions, are lost first and then the words which denote them disappear, and in other cases the words are not used any more and eventually the meanings will also disappear from the cultural memory. If the shells are gone (and the replacements do not have the same value to Helena), what they mean to her, is also lost. What I find fascinating is that Winterbach uses the representational function and ability of words to mourn the loss of words in their referential capacity. It is as if there is an underlying awareness that the basis of the generation of meaning in any sign Â�system is dependent on the referential function of language. Without things there Â�cannot be signs, without signifieds there is no need of signs:

 Heilna du Plooy

The signifying contract is both the origin and the model of any kind of exchange. It gives foundations to signs in general and is actualized through each particular semiotic act since it represents the transcendental model of the functioning of sign systems. (Cinquin 1986:â•›181)

But in a subtle way the author of The Book of Happenstance is also concerned with the issue of cultural memory. Cultural memory is represented in words because words Â�preserve archaeological and cultural memories. Words can even be seen as the link to the cultural past of a community. The lists of words which are no longer used Â�frequently and are becoming extinct are as such a representation of the loss of cultural memory. In a country where the Afrikaans language is indeed under severe official pressure, this is a most sophisticated way of interacting with the problem. The main theme of the novel is loss, in many different guises and forms and the lists of unused and archaic words are iconic representations of a specific form of loss. This not only adds to the depiction of the theme of loss, but actually sets the tone in the novel as a whole. Loss of cultural memory is a sad thing but also an inevitable process in the sense of what some theorists in cultural studies describe as “original plenitude and subsequent loss” (Rigney 2005:â•›12), but there is also loss on a personal level. Ultimately, the novel is about the loss of underlying systems of thought: not only do words reflect the epistemic systems of thinking as Foucault described them, but the whole cultural fabric of a society reflect an underlying episteme or Zeitgeist. If language is lost, first the words and then the language itself, the archaeology of the culture of a society will also eventually be lost because the words are the containers of an essential part of history. Helena also realises and accepts that loss is an inevitable part of life – her mourning of the shells is in a sense a pretext, a compensatory Â�process. She is mourning all her losses through her mourning of the shells. But, as Rigney (2005:â•›12–14) explains, cultural and personal loss can be compensated for by means of constructing new meanings, by new language in this case. Eventually, Helena imagines hearing her daughter telling her that loss must be accepted, it can never be averted. “There will always be losses. Life continues” (Winterbach 2008:â•›316). And indeed the novel ends with an open ending, in the sense that Helena realises that there will always be new possibilities and new words with which to build a future.

6.â•… Conclusion In her novels, Winterbach does not merely narrate, she uses language in such a way that it illustrates and enacts what she really wants to say. The creative and representational function and ability of words are foregrounded, in some instances to indicate the archaeological baggage of the community and in other instances to refer to the loss of words in their referential capacity. Winterbach constantly redirects the attention away



Words, like shells, are signs as well as things 

from the meager referential capacity of language to sensitize the reader to the richness of the representational and iconic abilities of words and texts. One of the important themes of her whole oeuvre is the idea that there is a link between language and creativity and also a link between creativity and the regenerative abilities of the human psyche. She forces the reader into an awareness of new ways of thinking by the extent of the disconnectedness between the meaning of words and the way in which the words are used in the novels. Seemingly empty forms and structures (i.e. the alphabetic order of words or the lists of forgotten words) become meaningful and imbued with new meaning in ingenious structures and patterns in the novels. The reader does not only read the story or even the text, the reader becomes aware of the novel, of the narrative text as such, as an example of creative, experienced language. Winterbach uses the archetype of the trickster figure to unsettle her characters, but her novels are themselves tricksters who taunt, delight, scare, unsettle, and �console the reader in multiple ways. The apparent playfulness is however closely linked to the thematic development in the novels as I have tried to indicate in my discussion of the examples. The stylistic enactment thus represents a variety of forms of iconicity, because in specific passages and in the novels at large, form indeed mimes meaning.

References Borges, J.L. 1993. The analytical language of John Wilkins. L.G. Vázquez (transl.) (El idioma analítico de John Wilkins). 〈http://www.alamut.com/subj/artiface/ language/johnWilkins. html 〉 (23 March, 2009). Botha, M. & H. Van Vuuren. 2002. Die eksperimentele gebruik van taal in Ingrid Winterbach se Niggie (The experimental use of language in Ingrid Winterbach’s Cousin). Stilet 18(2): 35–56. Botha, M. & Van Vuuren, H. 2006. Teks op teks op teks: Intertekstualiteit in Ingrid Winterbach se Niggie (Text on text on text: Intertextuality in Ingrid Winterbach’s Cousin). Literator 27 (3): 1–25. Botha, M. & Van Vuuren, H. 2008. ‘Onze stryd, onze heldedood…’: Kreatiewe verwerking van patriotiese stof uit Celliers se oorlogsdagboek in Winterbach se Niggie (‘Our struggle, our death as heroes…’: The creative recycling of patriotic material in Winterbach’s Cousin). Tydskrif vir Geesteswetenskappe 48(1): 111–128. Bronzwaer, W.J.M. 1993. Lessen in Lyriek: Nieuwe Nederlandse Poëtica. Nijmegen: SUN. Burke, M. 2001. Iconicity and literary emotion. European Journal of English Studies 5: 31–46. Cinquin, C. 1986. The archaeology of the sign. In Semiotics 1986, J.N. Deely & J.D. Evans (eds), 179–191. London: University Press of America. De Cuypere, L. 2008. Limiting the Iconic: From the Metatheoretical Foundations to the Creative Possibilities of Iconicity in Language [Iconicity in Language and Literature 6]. Amsterdam: John Benjamins. Deely, J. 1986. Idolum. Archaeology and ontology of the iconic sign. In Iconicity: Essays on the Nature of Culture: Festschrift for Thomas A. Sebeok on his 65th Birthday, P. Bouissac, M. Herzfeld & R. Posner (eds), 29–49. Tübingen: Stauffenberg.

 Heilna du Plooy Du Plooy, H. 2000. Oopte en afbakening: Ruimtes en rame in die oeuvre van Lettie Viljoen. Tydskrif vir literatuurwetenskap 16(2): 86–106. Du Plooy, H. 2006. Afstand en belewenis: Liminale ruimtes en oorlewing in Niggie deur Ingrid Winterbach (Distance and endurance; liminal space and survival in Ingrid Winterbach’s Cousin). Literator 27(1): 1–22. Du Plooy, H. 2008. Woorde wat teken en beteken – ikonisiteit in die poësie (Words as signs and signatures – Iconicity in poetic texts). Literator 29(2): 65–88. Eco, U. 1983. Naschrift. In De Naam van de Roos, Umberto Eco, J. Tuin, P. de Voogd & T. Van Velthoven (transl.). Amsterdam: Bert Bakker. Fischer, O. & Nänny, M. 1999. Introduction: Iconicity as a creative force in language use. In Nänny & Fischer (eds), xv–xxxvi. Foucault, M. 1973. The Order of Things: An Archeology of the Human Sciences. New York NY: Random House. John, P. 2004. Meer dydelikhyt oor die punch en die vis: ‘n vergelyking van Niggie, Daar’s vis in die punch en Eilande (More clarity about slang and literary language; a comparison between Cousin, There’s fish in the punch and Islands), Literator 25(1): 23–46. Leech, G.N. & Short, M.H. 1981. Style in Fiction: A Linguistic Introduction to English Fictional Prose. London: Longman. Lotman, J.M. 1972. Die Struktur literarischer Texte. München: Wilhelm Fink. Nänny, M. 1999. Alphabetic letters as icons in literary texts. In Nänny & O. Fischer (eds), 179–198. Nänny, M. & Fischer, O. (eds). 1999. Form Miming Meaning. [Iconicity in Language and Â�Literature 1]. Amsterdam: John Benjamins. Rigney, A. 2005. Plenitude, scarcity and the circulation of cultural memory. Journal of European Studies 35: 11–28. Ungerer, F. & Schmidt, H.-J. 1996. An Introduction to Cognitive Linguistics. London: Longman. Van Vuuren, H. 2004. Interview: Ingrid Winterbach, wie se roman Niggie met die Â�Hertzogprys bekroon is, word deur Ons Paneel onder hande geneem. 〈http://www.oulitnet.co.za/Â� paneelklopper/ingrid.asp 〉 (1 April, 2009). Viljoen, L. 1993. Karolina Ferreira. Kaapstad: Human and Rousseau. Winterbach, I. 2002. Niggie. Kaapstad: Human and Rousseau. Winterbach, I. 2005. The Elusive Moth, I. Gouws & I. Winterbach (transl.). Kaapstad: Human and Rousseau. Winterbach, I. 2006. Die Boek van Toeval en Toeverlaat. Kaapstad: Human and Rousseau. Winterbach, I. 2007. To Hell with Cronjé, E. Silke & I. Winterbach (transl.). Kaapstad: Human and Rousseau. Winterbach, I. 2008. The Book of Happenstance, D. Winterbach & I. Winterbach (transl.). Â�Kaapstad: Human and Rousseau. Wybenga, G. & Cloete, T.T. 1992. Ikoon en ikonisiteit. In Literêre terme en teorieë, T.T. Cloete (ed.),178–182. Pretoria: HAUM-Literêr.

Unveiling creative subplots through the non-traditional application of diagrammatic iconicity An analysis of Kingsley Amis’s The Green Man Andrew James

Chikushi Women’s University In literary terms, diagrammatic iconicity has generally been understood to refer to a patterned series of linguistic or syntactic textual connections. But taking a broader view of diagrammatic iconicity reveals previously invisible creative subtexts. In Kingsley Amis’s novel The Green Man (1969), critics have failed to see that Amis once again offers a fictional portrait of a former publisher. In his first novel, Lucky Jim (1954) L.S Caton was based on R.A. Caton, the operator of the small press that published Amis’s own Bright November (1947). Caton would return for cameos in each of the next six novels, until he was killed off in The AntiDeath League (1966). Perhaps critics simply forgot about Amis’s habit of venting lingering irritation with previous publishers, but the depiction of Victor Gollancz in The Green Man, his ninth novel, has gone unnoticed. Gollancz’s presence in the text clearly conveys the importance of artistic freedom, which only becomes apparent when iconic principles are applied.

1.â•… Introduction In literary terms, diagrammatic iconicity has generally been understood to refer to a series of linguistic or syntactic connections within a text that suggest a pattern. But by going beyond pieces of language to pieces of life, and taking a broader, thematic view of diagrammatic iconicity, we can allow for pre-established patterns of Â�authorial behaviour and previously invisible creative subtexts to become visible. This is the case with Kingsley Amis’s The Green Man (1969), a novel that has been interpreted through a variety of theoretical frameworks, when it is not seen as muddled autobiography. Â�Critics have failed to see the repetition of a familiar pattern in the novel, as Amis once again offers a fictional portrait of a former publisher. He did so in his first novel, Lucky Jim (1954), by basing L.S. Caton on R.A. Caton, the operator of the small press that published Bright November (1947), Amis’s inaugural poetry collection. Caton would

 Andrew James

return for cameos in each of the next six novels, including The Egyptologists (1965), which was co-authored with Robert Conquest, until he was killed off in The Anti-Death League (1966). Perhaps critics simply forgot about Amis’s habit of venting lingering irritations with previous publishers, but no one has recognized the depiction of Victor Gollancz in The Green Man, his ninth novel, and his first after switching to Jonathan Cape. Gollancz’s presence in the text clearly conveys the importance of artistic freedom, though this does not become apparent unless one applies certain iconic principles. There are no direct references to Gollancz in the novel, but numerous clues suggest that he is represented through the hero’s cat, named Victor Hugo, and the analysis of this cat as a diagrammatic iconic sign shows that Amis was both arguing for the importance of creative freedom and celebrating his release from Gollancz’s publishing house.

2.â•… Establishing a pattern: Satirizing former publishers Critics have tended to see the nefarious L.S. Caton as a joke figure and Amis himself encouraged this view when he stated that “The fact that L.S. Caton finally gets bumped off [in The Anti-Death League] is a signal that it’s to be taken seriously” (Amis 1974:â•›25). However, serious commentaries on the importance of artistic freedom are imbedded in all of Amis’s novels along with frequent diatribes against hindrances to freedom and Caton’s true function should be related to those comments. He reminds us that barriers to creative independence are intolerable. In The Green Man, Victor Hugo fulfills the same function. According to C.S. Peirce, by observing an icon, other unnoticed truths about its nature can be discovered, thus, for example, one could draw a map based on the information in two photos (Duranti 1997:â•›205). Similarly, we can begin to map Amis’s Â�feelings about Gollancz by reconstructing a picture of the man and collating it with the character of Victor Hugo. First, characteristics such as fickleness, sensitivity, and Â�selfishness, which are generally associated with cats, are also found in Gollancz, as shall be shown. The fact that Amis was inviting the reader to connect Gollancz with the cat is suggested by its very name, for Victor Hugo is never mentioned in Amis’s Â�voluminous correspondence and Maurice Allington, the protagonist, expresses little interest in English literature and even less in its French equivalent. Although this tempts the reader to dismiss the name as a red herring, Amis enjoyed playing with names, as critics Â�recognized in the lexical alteration of Jim Dixon, the hero of Lucky Jim, to Jake Richardson in the 1976 novel Jake’s Thing (Bradford 2001:â•›305). He also deliberately chose character names that suggested their real-life models. While Amis was writing his final book, The Â�Biographer’s Moustache (1995), about an inept literary biographer, journalist Eric Jacobs was writing his life. Jacobs read the novel in draft and when he discovered that the fictional biographer was called



Unveiling creative subplots 

Cedric – a name that “embraced my own” (Jacobs 2000:â•›5) – he threatened to sue. This prompted Amis to rename the character Gordon. Similarly, the last two letters of ‘Hugo’ are the first two in ‘Gollancz, ’ and the word ‘cat’ is contained in ‘Caton.’ Furthermore, Amis tended to do things twice, particularly in his late fiction, thus Gollancz receives another Â�treatment in the novel Difficulties with Girls (1988). If one views Victor Hugo as the first metaphoric Â�representation of Gollancz, then it is important to remember that Peirce’s definition of metaphor is not based on resemblance but works as a “double sign” so that the relation is often symmetrical. Lance R. Factor (1996:â•›230) offers the following as an example: “If the daughter resembles the mother, then the mother resembles the daughter”. Thus, when considering the iconic portrayal of Amis’s second publisher, we must compare the two Victors in order to find ways in which Gollancz is catlike and Victor Hugo resembles a publisher.

3.â•… Amis in theory Before examining the set of similarities between the two Victors according to a Â�non-traditional application of diagrammatic iconicity, it would be helpful to explore the reasons other theoretical readings of The Green Man have failed to reveal complexities in the creative subtext. After earning a reputation as a writer of romantic comedies with satiric bite, Amis wrote Take a Girl Like You (1960), a novel characterized by its dark and uncertain ending. His desire to unsettle the reader doubtless contributed to the creation of devious, conniving characters and a penchant for playing tricks on the reader. Jake’s Thing and The Biographer’s Moustache feature examples of aporia, or textual knots that resist untying (Barry 2002:â•›79). In these novels, crucial speeches are blocked out by car and airplane engines. Once the roar of the engine has subsided, the monologue is finished, never to be repeated, and both the reader and the protagonist can only wonder about what was said. In spite of irresolution and trickery in Amis and his antipathy for modernism, one hesitates to call him postmodern because he prided himself on being an entertaining storyteller, once claiming that he wrote with one eye on the face of his ideal reader, “watch[ing] for any signs of boredom or impatience to flit across [it]” (1975:â•›71). His style did become dense and fractured in the 1970s and 1980s, with convoluted sentences leading some critics to again cry postmodern, and to dismiss him as obtuse and self-indulgent, though this was probably not a conscious stylistic decision. He was simply becoming sloppy and his ability to self-edit weakened as he grew older. The Green Man was written in 1969, when Amis was at his literary peak, but the novel’s status as a genre experiment – it is a ghost story – has combined with the author’s well-documented alcoholism and the aforementioned critical tendency to search for evidence of postmodernism to result in little consideration being paid to the

 Andrew James

question of artistic intent. The main components in the novel’s intricate plot are as follows. Maurice Allington, the alcoholic innkeeper of the Green Man and the Â�novel’s first-person narrator, is subject to visions. According to legend, the inn is haunted by the ghost of Dr. Thomas Underhill, a practitioner of dark magic. Even though he is drunk most of the time and sees things, Allington would have us believe that he also sees real ghosts, converses with God in his living room, and battles the legendary green man. When the story begins, he is trying to seduce his wife’s friend. Though he is disturbed by the death of his elderly father, a resident at the inn, and begins drinking more heavily thereafter, he succeeds in his seduction. He is then Â�visited by Underhill’s ghost, the green man, and God, who enters Allington’s living room in the form of a young man in a dark suit. The hero is told to dispose of Underhill so that the divine need not intervene and reveal the secret of His existence. Allington is again successful, defeating both the ghost and the green man and strengthening a shaky relationship with his teenage daughter, but in the process he loses his wife and lover. At the end of the book he decides to sell the inn and move to another town. One recent reading of the novel claimed that it follows the postmodern pattern of “deliberately undermining the confidence” of the reader, as Emma Hawkins (2003:â•›47) has pointed out. She sees “post-modern touches” (52) in Amis’s version of the green man legend, for the hero “de-centers readers all too well, depriving them of the opportunity to formulate and sustain a single certainty. Through Maurice, who is an unreliable narrator, Amis has unbalanced readers” (59). Zachary Leader drew on biographical evidence to show that by using episodes from his own life Amis was having “several types of fun” (2006b: 617) with the reader. At the time of composition, Amis, like Allington, was drinking a bottle of Scotch whisky each day and, in 1969 and 1970, frequently had trouble differentiating between real and imagined memories. Leader refers to an extended argument between Amis and his wife, Elizabeth Jane Howard, over the guest list for a dinner party held the evening before, when the party had neither taken place nor been planned (618). Leader calls The Green Man a “remarkable” account of the author’s own “wounding, egotistical, drink dependent” behaviour (2006b: 620). Robin Sims borrows from Freud and Lacan in offering another compelling interpretation. He argues that Allington desires what he most fears (2007:â•›88) and attempts to prove that his “vigorous appetite for sex is also an appetite for death, while also an attempt to escape both it and the unacknowledged desire for it. The green man […] appears to embody these qualities, in such a way as to entangle them in one another” (2007:â•›93).

4.â•… Explaining Amis’s art through diagrammatic iconicity The respective interpretations of Hawkins, Leader, and Sims are not wrong or Â�misguided, but they discourage us from looking at Amis’s artistry. The argument



Unveiling creative subplots 

seems to be that if the novel was written by an alcoholic who hallucinated, then the author could not have been in control of his craft. But by analyzing the psychology of Maurice Allington we can explain Kingsley Amis to himself. Postmodern, psychoanalytic and biographical readings do not give sufficient attention to the pattern that runs through all of Amis’s previous novels whereby he stated his artistic intentions through the subtext. Victor Hugo, the cat, is not precisely a symbol for Victor Gollancz, who died two years before the novel’s publication, but similarities in Amis’s relationship to Gollancz and Allington’s to his cat become apparent through the application of some of the principles of iconicity. Icons are based on “resemblance” between sign and object and on a putative sharing of “properties” (Innis 1985:â•›2), but these properties are not always immediately recognizable. They only become so if the interpreter is conscious of similarity when approaching the text (Givón 1989:â•›98). Thus, a pattern of diagrammatic iconicity will not reveal itself unless we have reason to suspect its existence and seek it out. And while the figure of L.S. Caton alerts us to the artistic potential of Victor Hugo, “an iconic feature in a literary text is not self-explanatory”, as Max Nänny (2005:â•›230) states. “It is always the reader who perceives whether a textual feature is iconic or not”. One will not perceive the iconic significance of Victor Hugo without prior knowledge of Amis’s feelings towards Gollancz and his habit of fictionalizing publishers. Simply reading the text with Gollancz in mind will not provide enough clues to make this connection. The fundamental iconic difference between the two fictional representations of Amis’s publishers helps to explain why no one has recognized Gollancz in Victor Hugo. As an icon, L.S. Caton became transparent and “immediately recognizable”, to borrow Nänny’s words, to any critic writing on Amis. Though the character was initially translucent, with only a few select friends like Philip Larkin and Robert Conquest recognizing the connection, Caton’s recurrence in subsequent novels and Amis’s reiteration of the negative feelings he had for R.A. Caton in the mid-1960s turned L.S. Caton into a transparent icon. However, Victor Hugo has remained translucent because the cat does not appear in any later novels and Gollancz’s next fictional depiction is completely different. For these reasons, Victor Hugo “needs a more careful scrutiny,” and the cat’s iconic status “must be discovered by an imaginative effort on the reader’s part” (Nänny 2005:â•›230–1).

5.╅ Retracing the pattern to L.S. Caton Amis satirized R.A. Caton in his first and perhaps most successful novel, Lucky Jim, out of frustration at him for both hindering his literary career and hurting him �financially. The hero of the novel, Jim Dixon, a reluctant history lecturer at an English �red-brick university, is on probation and hopeful of obtaining a more secure position. Security is contingent on the publication of an academic paper and Dixon has �submitted one on

 Andrew James

the economic influence of shipbuilding in the 1400s. The journal’s editor, L.S. Caton, has tentatively accepted it without giving Dixon a definite date of Â�publication. Caton inherits many of the unattractive traits of the editor of the Â�Fortune Press, R.A. Caton, whose evasion tactics Amis became familiar with during his Â�struggle to publish in 1947. Caton took advantage of Amis’s desperation to appear in print by forcing him to purchase fifty copies of his own book and sell them privately, as Amis reported to Philip Larkin in a 25 July 1947 letter (Amis: Bodleian). Caton did not place such Â�crippling contractual conditions on young authors out of necessity, for he owned ninety-one houses in Brighton and made a considerable profit by printing soft-porn novels (Jacobs 1995:â•›129). But he was a shrewd and elusive businessman, as Amis Â�discovered during a visit to his offices in the fall of 1947. When the phone rang, Caton urged him to answer and “‘tell ‘em Mr. Caton’s gone out for a few minutes’” (Amis 2001:â•›140). Â�Zachary Leader has summarized Caton’s character as “dilatory, Â�inefficient, mean, secretive and double-dealing” (2006b:â•›189), and the standard interpretation of his fictional role is the one offered by Christopher Hitchens (2002:â•›107): “Amis took revenge against an editor named Caton by using his name for hateful or shifty parts […] and then killing him off in The Anti-Death League”. In truth, it is only his role in Lucky Jim that is of interest, for in each subsequent novel he has only the briefest of roles, bothering the heroes from afar by writing or sending a message and asking for a favour. The initials L.S. stand for “lazy sod,” which is the way that Amis came to refer to his publisher in correspondence with Larkin, after first calling him “THAT SOD” once their Â�relationship soured (Amis 2001:â•›104–5). In fictionalizing R.A. Caton, Amis added one other component. In 1948 he had been commissioned to write a monograph on Graham Greene by an academic at the University of Tucuman in Argentina. He finished the monograph then learned that he would not be paid for it, and the work was never published. The incident Â�understandably upset Amis and he would interweave it with the character of L.S. Caton. In each of his first seven novels, Caton appears, though never physically. He contacts the Â�protagonists by telephone or letter, and behaves in an evasive, vague manner. He always requests something of the hero, such as research assistance or the arrangement of a lecture on his behalf. In Lucky Jim, Caton absconds to Argentina with Jim Dixon’s piece on shipbuilding, which he then translates into Italian and publishes in a journal under his own name (Amis 1954:â•›229). In each subsequent book he sends a communiqué from South America asking for his favour, which the hero denies, until he finally returns to England in The Anti-Death League to give a talk on his travels, but is killed en route by a stray bullet (Amis 1965:â•›289). Thus, L.S. Caton shares the same essential properties as R.A. Caton and to make sure that readers could trace the fictional character back to its model, he included the name ‘Caton.’ He continued the trend by using the name ‘Victor’ in The Green Man.



Unveiling creative subplots 

6.â•… Looking for Victor Gollancz One supporter of Gollancz, his biographer and former employee Sheila Hodges, Â�portrayed him as a wonderful, though domineering and prickly man. She admitted that some authors disliked him for his intense political, social, and religious views and left his publishing house “because they felt that he was giving too much attention to public affairs and too little to publishing” (1978:â•›227). However, she claims that “far more felt gratitude and admiration that he remained first and foremost a human being of great moral courage who never hesitated to raise his standard in defence of the good and the important things of life”. Hodges’ biography is an unabashed Â�apology for Gollancz’s character that contains few unqualified criticisms. At the end of the book she allows that “He was often autocratic in his relations with authors”, “too Â�paternalistic” (227), and “working with him sometimes nearly drove [her] mad”, but insists that he “enriched [her] life in very many ways” (226). Not coincidentally, the book is a Gollancz imprint. A far less flattering portrayal is offered by Paul Johnson in Intellectuals (1988). In the chapter devoted to Gollancz, he describes him as positively Dickensian. The secret of his professional success, says Johnson, was to pay low wages: “When real need was pleaded, he would make an ex gratia payment or offer a loan, rather than raise a salary or an advance” (1985:â•›273). Johnson found it incredible that Gollancz “believed himself a man of great, instinctive benevolence, a true friend of humanity [when he] was in fact incorrigibly selfish and self-centred” (274). He calls Gollancz’s “self-centred vanity insupportable”, citing it as the barrier preventing him from forming lasting friendships: “In fact he was fortunate that his business acumen enabled him to create a private autocracy where no one could challenge him and his inability to make other men like him did not matter so much” (286–7). Amis’s first biographer, Eric Jacobs, gleaned most of his information on Gollancz from his Â�subject and, not surprisingly, he sides more with Johnson than Hodges. He did concede that “Victor’s egotism had its good side. It made him capable of the big gesture which could mean more than cash advances” (Jacobs 1995:â•›274). But these gestures were not enough, and within weeks of Gollancz’s death in 1967, Amis exercised a clause in his contract to move to Jonathan Cape, Elizabeth Jane Howard’s publisher (Jacobs 1995:â•›275). The extent to which Amis felt liberated by his death is evident in a blunt letter written to Livia Gollancz in February 1967 cancelling both a lunch appointment and all affiliations with the publishing house: “Now, sadly, my loyalty is at an end, and I am off ” (Amis 2001:â•›677). In The Green Man, Victor Hugo is neutered and in perpetual “vague semi-flight” (1969:â•›22), which invites comparison with the often evasive, emasculated Gollancz. After Amis completed his third novel, I Like It Here (1958), he was angered by Â�printing delays. “I finished that damned thing against time, forgoing the revision I should

 Andrew James

have liked to give it,” he wrote to Gollancz, calling him “a perfect louse to have put off Â�publication till 13th January [1958]” (Amis 2001:â•›521). Such exchanges were not infrequent. Gollancz also used stalling tactics to discourage Amis from conducting genre experiments, which took time away from straight fiction, for which there was an established audience. Thus, he would delay the release of Amis’s science fiction lectures, New Maps of Hell, which appeared in America in 1960, a year earlier than in England. Gollancz admitted to being reluctant to publish Â�something “so eccentric and ticklish” (quoted in McDermott 1989, 40). These experiences helped to convince Amis that in order to freely experiment with different literary genres and forms he needed to change publishers. In the novel Difficulties with Girls, Amis presented another version of Gollancz in Simon Giles, a publisher without a sex drive. The hero, Patrick Standish, contemplates sleeping with the publisher’s wife, weighing the damage it could do to his career, should his boss Giles find out, against Â�possible Â�pleasure. Standish is surprised, then, to be directly propositioned by Giles, who offers up his wife, pleading exhaustion and the need to keep her satisfied (1988:â•›261). A  further similarity between the two Victors is embodied in the pressures they exert on Allington and Amis. While loyalty to Gollancz prevented Amis from leaving, Victor the cat places more literal pressure on Allington. When he is first introduced he falls on his master’s feet “like an elephant pierced by a bullet in some vital spot” (1969:â•›22). Gollancz was certainly aloof from his stable of writers, not unlike the Victor Hugo who is “the picture of a cat to whom nothing out of the way, almost nothing at all, had ever happened” (ibid.: 45). Amis changes this by forcing the cat to witness Allington’s extra-sensory perceptions. When Allington hears the “rustling noise” of the green man passing outside the house, the cat races through the room, “impelled perhaps by the fragments of some memory” (ibid.: 83). Later, Allington and his daughter seem to be the only ones to hear phantasmal screams until the cat again “galloped up, threw himself at [Allington’s] feet and fell on to his side” (ibid.: 163). The demise of Victor Hugo liberates Allington just as Gollancz’s death left Â�Amis free to pursue other publishing options. At the end of the book, the cat stands between the advancing green man and Allington’s daughter “in a posture of defiance, his back arched and tail swollen. A kick from a wooden foot smashed into him, with a snapping of twigs or bones, and he went skidding, a lifeless bundle, across the road and into the ditch” (1969:â•›157–8). Victor Hugo’s death gives Allington’s daughter Amy the opportunity to escape and her father sufficient time to remember a silver talisman in his pocket, which he hurls over a cemetery wall, preempting the green man’s attack. After the cat’s death, Allington enters a new phase in his life–one in which he is no longer haunted by ghosts, but will be without his second wife in a different location.



Unveiling creative subplots 

7.â•… Conclusion The goal of this paper has not been to show the evasive or feline properties of Â�publishers, but to prove that Maurice Allington’s relationship to his cat was a Â�deliberate attempt on Kingsley Amis’s part to discuss his own creative situation. By taking a Â�picture of Gollancz and placing it alongside the picture of Victor Hugo, the two may be seen to blend together into a composite of Amis’s desire for artistic independence. Just as a single kick from the green man eliminates Victor Hugo and leads to Allington’s Â�liberation, Amis must have felt as though Gollancz’s death removed a major restriction on his artistic freedom by enabling him to change publishers. Amis’s desire to enter an uncertain but also uninhibited creative future in which he would write an alternate world novel, a crime novel, and a political thriller in addition to straight Â�fiction is reflected in Allington’s own inability to deal with present time. He is waiting for something, but does not know what. Near the end of the novel, with four hours until his midnight meeting with Underhill, Allington attempts various diverting tasks and each shows his failure to properly engage with time in the past, present, or future. He watches part of a movie, reads a novel for forty minutes, and then has a serious car wreck as his journey home is aborted by an ill-advised attempt to overtake a car on a curve (ibid.: 150–1). As his wife Joyce notes, Allington’s biggest problem is his inability to deal with the here and now. She decides to leave him for her lesbian lover because she wants to be with a person “Who hasn’t always got somewhere more important to be in the next two minutes” (ibid.: 170). On an artistic level, the novel shows that Amis understood this feeling completely. Although Lucia Santaella Braga (1996:â•›212) has said of diagrammatic iconicity that we cannot make a report or offer an analysis of similarity for “the only criterion for deciding whether one quality is like another is someone’s sincere assertion that it really is”, the burden of inter- and extra-textual proof leads to the conclusion that Amis was examining his relationship with Victor Gollancz in this novel. He too believed that he had somewhere more important to be and hoped that by changing publishers he would get the necessary support to write more freely.

References Amis, K. 1946–1980. Letters to Philip Larkin. Oxford: Bodleian Library. Amis, K 1947. Bright November. London: The Fortune Press. Amis, K. 1954[2000]. Lucky Jim. London: Penguin. Amis, K. 1960. Take a Girl Like You. London: Victor Gollancz. Amis, K. 1967. The Anti-Death League. New York NY: Ballantine. Amis, K. 1969[2004]. The Green Man. London: Vintage.

 Andrew James Amis, K. 1974. Kingsley Amis: A Profile [interview], with Clive James. The New Review (July  4): 21–8. Amis, K. 1975. Interview with Kingsley Amis, with Michael Barber. The Paris Review 59(64) (Winter): 39–72. Amis, K. 1988 [1989]. Difficulties with Girls. London: Penguin. Amis, K. 1995. The Biographer’s Moustache. London: Flamingo. Amis, K. 2001. The Letters of Kingsley Amis, Z. Leader (ed.). London: Harper Collins. Amis, K. & Conquest, R. 1965. The Egyptologists. London: Jonathan Cape. Barry, P. 2002. Beginning Theory: An Introduction to Literary and Cultural Theory, 2nd edn. Â�Manchester: Manchester University Press. Bradford, R. 2001. Lucky Him! The Biography of Kingsley Amis. London: Peter Owen Limited. Duranti, A. 1997. Linguistic Anthropology. Cambridge: Cambridge University Press. Factor, L.R. 1996. Peirce’s definition of metaphor and its consequences. In Peirce’s Doctrine of Signs: Theory, Applications, and Connections, V.M. Colapietro & T. M. Olshewsky (eds.), 229–235. Berlin: Mouton de Gruyter. Givón, T. 1989. Mind, Code, and Context: Essays in Pragmatics. Hillsdale NJ: Lawrence Erlbaum Associates. Hawkins, E. 2003. Kingsley Amis: Where did that green man come from? Lamar Journal of the Humanities 28(1): 47–63. Hitchens, C. 2002. The man of feeling. The Atlantic Monthly 289: 103–108. Hodges, S. 1978. Gollancz: The Story of a Publishing House 1928–1978. London: Victor Gollancz. Innis, R. 1985. Semiotics: An Introductory Anthology. Bloomington: Indiana University Press. Jacobs, E. 1995. Kingsley Amis. London: Hodder and Stoughton. Jacobs, E. 2000. Dear Martin, yours Eric. The Times (May 11): 3–5. Johnson, P. 1988. Intellectuals. London: Weidenfelt & Nicolson. Leader, Z. 2006. The Life of Kingsley Amis. London: Jonathan Cape. McDermott, J. 1989. Kingsley Amis, an English Moralist. London: Palgrave Macmillan. Nänny, M. 2005. Diagrammatic iconicity in poetry. In Theory into Poetry. New Approaches to the Lyric, E. Muller-Zettelmann & M. Rubik (eds.), 229–249. Amsterdam: Rodopi. Santaella Braga, L. 1996. From pure icon to metaphor: Six degrees of iconicity. In Peirce’s Â�Doctrine of Signs: Theory, Applications, and Connections, V.M. Colapietro & T.M. Olshewsky (eds.), 205–228. Berlin: Mouton de Gruyter. Sims, R. 2007. Facing the thing: The Green Man, psychoanalysis and Kingsley Amis. Critical Survey 19(2): 82–100.

part vi

Intermedial iconicity

The iconic indexicality of photography Piotr Sadowski

Dublin Business School The unique visual appeal of photography results from combining the basically iconic code (resemblance between image and referent) with indexicality. Photography is indexical insofar as the represented object is “imprinted” by light and the chemical (or electronic) process on the image, creating a visual likeness that possesses a degree of accuracy and “truthfulness” unattainable in purely iconic signs such as painting, drawing, or sculpture. The indexical origin of the photographic image explains why discussions of the photographic media (including film and television) often employ categories normally reserved for the emotive and irrational effects produced in traditional societies by sympathetic magic, with its objectively wrong but psychologically compelling sense of direct causal link between objects once physically connected but later separated. The essay discusses the iconic indexicality in the context of its historic antecedents such as imprints of hands, death masks, wax effigies, shadow portraits, and experiments with camera obscura.

1.â•… Introduction: The semiotic status of photography Just as linguistic iconicity adds emotive expressiveness to the informational content communicated by language’s symbolic code, so in visual communication combining different types of signs can enhance the expressive power of images. A classic case in point is photography, whose unique visual appeal results from combining the Â�basically iconic code, involving close perceptual resemblance between image and its Â�referent, with indexicality. Photography is indexical insofar as the represented object is ‘imprinted’ by light and the chemical (or electronic) process on the resulting image, creating a visual likeness with a degree of accuracy and ‘truthfulness’ unattainable in purely iconic signs such as painting, drawing, or sculpture (Friday 2002:â•›49). The Â�indexical origin of the photographic image explains why discussions of the Â�photographic media (including film and television) often employ categories normally reserved for the emotive and irrational effects produced in traditional societies by sympathetic magic, with its objectively wrong but psychologically compelling sense of direct causal link between objects once physically connected but later separated (here: the object and its displaced photographic ‘imprint’). This paper will

 Piotr Sadowski

accordingly discuss the semiotic status of photography both in terms of its postulated iconic indexicality, and in the context of photography’s historic antecedents such as intentional imprints of hands, death masks, shadow portraits, and experiments with the camera obscura. Like all signs, photographic images operate first of all within the direct context of ‘contiguity’, in which the sign itself is physically co-present with the receiver, as when a person looks at a photograph. Contiguity is the most common and ubiquitous form of communication, and it is a prerequisite for indirect communication involving Â�displaced reference (Sadowski 2009:â•›32–34). The simplest form of displaced reference is a physical trace, or index, produced by one of the communicators in the immediate environment (Peirce 1998:â•›143; Sebeok 1994:â•›72; Sadowski 2009:â•›34, 94–97). Indexes can be either fully contiguous with the sender, as in the case of a cast shadow, mirror reflection, personal odour or voice, or they can be spatio-temporally displaced from the sender, as in the case of a footprint, a finger print, a DNA trace, a death mask, a bullet hole, a bomb crater, and so on. In another possible type of indirect communication, the sender produces a change in the environment (a sign) which is perceptually analogous, in other words similar, to some other object implied by the sign. A sign whose form is Â�perceptually analogous to its referent is called an icon (Fischer & Nänny 2001:â•›1–9; Sadowski 2009:â•›36–38). Familiar examples of iconic communication in language include Â�onomatopoeia, synaesthetic sound symbolism, or morphological and syntactic Â�iconicity (Anderson 1998). Imitative gestures too belong to iconic communication, as in pointing upwards or downwards. Visual icons are also a part of conventional sign languages (Corballis 2002:â•›163). However, the most spectacular examples of iconic communication among humans are permanent imitative visual representations, as in realistic painting, Â�sculpture, and photography. As I am going to argue, photography owes its exceptional perceptual and Â�emotive appeal to the fact that it combines – uniquely in the history of human visual Â�representations of the world – two important types of visual communication: indexicality and iconicity, in addition to being related in an important way to human contiguous Â�experience. Painting, for example, is by comparison limited to Â�iconicity, which explains why even the best executed painted portrait is Â�usually considered less ‘authentic’ and ‘truthful’ in relation to the represented person than a Â�poor-Â�quality Â�photograph – and that difference in perception has everything to do with the indexicality of photography and the iconicity of painting (Carroll 1995:â•›70; Walton 1997:â•›68). In addition to being physically if inadvertently caused by the Â�represented object through the medium of light, a photographic image also resembles its object to an extent Â�normally unattained in painting. From a semiotic point of view, Â�therefore, photography can be defined as an ‘iconic index’, as can Â�photography’s Â�animated Â�extensions such as film and Â�television. It is indexicality that



The iconic indexicality of photography 

makes these forms of visual representation so much more Â�efficacious in reflecting the outside world, and so much more powerful in their emotive effect on viewers than purely iconic Â�representations such as paintings, drawings, or Â�sculptures. The iconically Â�indexical character of Â�photography means that its images not only closely resemble the photographed objects, but that the images are also Â�physically related to the objects they represent in a way never achieved in painting. Even in painting from life, the painted scene reflects only the painter’s subjective beliefs of what is there, whereas a scene in front of the camera is not affected by the Â�photographer’s beliefs. In other words, Â�photographs depict Â�realities that already exist (although Â�obviously only the Â�photographer’s choice can disclose them), whereas Â�paintings Â�create Â�physically Â�non-existent realities (Bazin 1980; Marien 2006). The indexical, quasi direct nature of a photograph also creates a subjective impression of a Â�surrogate possession of a Â�photographed object – be it a picture of a lover, of one’s child, of an admired actor or a pop star (Sontag 1978:â•›155). This illusory if psychologically compelling sense of  Â�possession is based on an implicit assumption of identity, or at least of Â�inseparable Â�sympathy between a photograph and what it Â�represents, between object and its index.

2.â•… Brief history of iconic indexicality Considering this compelling effect of photographs, it is curious that it should have taken humanity so long to produce permanent intentional displaced iconic indexes, which is what photographs are in semiotic terms. The reasons for this delay were mainly practical – photography is a highly technical medium – and the right Â�conditions for it were only created as a consequence of a number of discoveries and inventions in Â�physics, optics, and chemistry made during the Scientific Revolution in Europe and North America in the last few centuries. Interestingly, the perceptual and cognitive prerequisites that enable people to Â�create and appreciate displaced iconic indexes have been in place ever since the homo Â�sapiens with their reflective consciousness, extended working memory, and the Â�resulting Â�symbolic culture emerged from their pre-symbolic primate ancestors around 170,000 years ago. The earliest purposefully executed displaced iconic indexes (as Â�distinct from accidental displaced indexes such as footprints) on Â�archaeological record are imprints of human hands with outstretched fingers, found among the Â�paintings of animals on the walls of the Chauvet cave from over thirty thousand years ago (Chauvet & Deschamps 1996:â•›79). In Chauvet, hands were placed flat on the cave’s wall, and paint was then applied around it and between the fingers. Such negatives of human hands have also been found in other Upper Palaeolithic caves in southern France and northern Spain, as have the positives: impressions of hands coated with pigment and pressed against the

 Piotr Sadowski

wall. In one Spanish cave, the Gargas, there are nearly a hundred and fifty red and black hands, some of them small, probably belonging to women or children Â�(Leroi-Gourhan 1968:â•›148). These ‘signatures’ are the first recorded Â�permanent Â�intentional iconic indexes: iconic in the sense that the impressions bear a close resemblance in actual size to the real hands, and indexical in the sense that the impressions were actually physically caused by living people, who evidently wanted to preserve a bit of themselves for posterity, not unlike those modern tourists who cannot resist inscribing or carving their names, with dates, on the walls of famous buildings they visit. The dream to be able to fix an iconic index of a person, in other words to preserve the appearance of a person after their death, in defiance of time, is also reflected in the widespread ancient myth of the Otherworld, where the souls of the dead continue their existence after the demise of the body. In the absence of Â�family photo-albums or home videos containing temporally displaced iconic indexes of dead relatives, the ancient Greeks for example invented Hades populated by living shadows of dead Â�people. A shadow is a fully contiguous and inseparable index of a living person, and it is easy to imagine the immaterial shadow surviving its master’s death, destined to dwell for ever in the “land of shadows”, which is how Homer describes Hades, visited by Odysseus anxious to embrace the shadowy and elusive indexical ‘holograph’ of his departed mother (Homer 1980:â•›115). The combined effects of iconicity and indexicality are also responsible for the notoriety surrounding the Shroud of Turin, a piece of cloth allegedly covered with an impression of the body of Christ as it was – according to Christian tradition – taken down from the cross to be buried. There also exists a similar legend of St Veronica, a woman from Jerusalem who wiped the face of Christ with a veil, which henceforth retained the imprint of Christ’s face. It is interesting to note that while there exist Â�thousands, probably millions of paintings, drawings, sculptures, together with their copies and prints showing the image of Christ, for Catholic Christians, none of these iconic representations possesses the magical appeal of the unique artefact of the Turin Shroud, and that again has everything to do with the alleged indexical character of this relic. One can speculate that an even greater furore among believers would be caused by something like a bone from Christ’s body or a lock of Christ’s hair – which in Â�semiotic terms would represent spatio-temporally displaced ‘metonymic contiguous signs.’ The Shroud of Turin may be a fourteenth-century hoax, and the legend of St Veronica may be just that: a legend; however, for Christian believers the alleged Â�indexical impression of Christ’s body on the Turin Shroud is the next best thing to actually witnessing the Crucifixion directly, while the countless number of merely iconic paintings and prints has a correspondingly much weaker emotional appeal. In early modern times, a direct precursor of displaced iconic indexes of Â�photography were the fully contiguous iconic indexes produced by the camera obscura, an optical device and a drawing aid used by realistic painters in the Â�seventeenth and



The iconic indexicality of photography 

eighteenth century. The camera obscura, literally, a ‘dark chamber’, was a portable box Â�fitted with a convex lens and an internal mirror, which righted the Â�upside-down image created by the lens, so that it could be traced on a piece of paper placed on translucent glass plate installed at the top of the device. Drawings thus made could help artists to trace the outlines of shapes to be transferred onto Â�canvas to achieve highly realistic and Â�accurate, truly “photographic” paintings, mostly of cityscapes, such as those by Canaletto (1697–1768) (Links 2005:â•›118). Before the invention of film and television, the effects produced by the camera obscura must have been truly astounding: on the two-dimensional screen the viewer could see a three-dimensional scene, in its natural colours, fully animated, reduced in size and neatly framed. The camera obscura could capture a moving image (that is, it could create a contiguous iconic index of a scene, like live television), but it could not fix it to make it Â�temporally displaced – this part was undertaken by the painter who Â�created a permanent icon out of a fleeting iconic index, losing the indexical Â�character of the image in the process. This is why Canaletto’s paintings of Venice, Rome, and Â�London, realistic as they are, are not as accurate, Â�faithful, and ‘real’ as would have been achieved even by a primitive, black-and-white grainy photograph, to say Â�nothing of today’s colour, Â�high-definition digital cameras. Interestingly, the practice of painting pictures based on optical, that is, indexical images continued after photography as we know it was invented. For example, in the late nineteenth century, the Italian artist Ettore Roesler Franz produced a series of highly realistic watercolour pictures of Rome, which were based on black-and-white photographs taken by the artist. The indexical origin of Franz’s watercolours provided them with a degree of truly ‘photographic’ realism that probably would not have been possible if the artist had simply painted from life. However effective the camera obscura was in tracing the outlines of buildings and perspectives of streets and city squares, it was of no use in trying to transfer onto canvas the image of a human face, which consists of tones and shades rather than of lines. This is why the device could help artists to produce city paintings of admirable accuracy, but it was of no help in painting portraits – a much more popular genre at the time. An early attempt to capture a human figure, especially the head if not actually the face, and to fix it in a permanent iconic index were the silhouette cut-outs, popularized in the late eighteenth century by the Swiss scientist Johan Caspar Lavater. The shadow of a person’s profile cast by sunlight or a candle was traced onto paper, and then the image was filled in by hand, or used by the artist as a template to cut a silhouette from black paper. These shadow portraits were part entertainment and part artistic venture, and their popularity drew on the widespread conception that human character could be read through the study of facial features (Marien 2006:â•›1, 5). Silhouettes were of course not as detailed and colourful as painted portraits, but were considered to be more faithful and truthful of a person’s appearance and personality than the latter, precisely because of their indexical rather than merely iconic origins.

 Piotr Sadowski

The missing indexicality in painted images may also be responsible for the Â�subject of the ‘invention of painting’, popular in early modern art. Accordingly, the origin of the art of portraiture was attributed to the observation and tracing of a Â�person’s shadow. The idea goes back to an ancient Roman poetic love tale recorded by Pliny the Elder (23–79 CE), in which a potter discovered with the help of his Â�daughter how to model portraits in clay. She was in love with a youth, and when he was Â�leaving the country she traced the outline of the shadow which his face cast on the wall by Â�lamplight. The father filled in the outline with clay and made a model (Pliny 1991:â•›187; Â�Gombrich 1995:â•›31). The fifteenth-century architect Leon Â�Battista Alberti in his Â�treatise De Pictura (1435) also expressed a belief that the earliest Â�painters used to draw around shadows made by the sun, and the art of portraiture Â�eventually grew by a Â�process of additions (Alberti 1972:â•›61). The popularity of this story in the Â�Renaissance is also testified by Giorgio Vasari, who in his Lives (1568), recounts another version of the legend of the origin of painting. Accordingly, painting was first brought to Egypt by one Gyges from Lydia, who once saw his own shadow cast by the light of a fire and instantly drew his own outline on the wall with a piece of charcoal (Vasari 1987:â•›27). The popular cut-out silhouettes of the Â�eighteenth century were based on the same quasi-photographic principle Â�combining the indexicality of a person’s shadow with the iconicity of traditional painted Â�portraits, to achieve the accuracy and fidelity of representation unattained by conventional painting or Â�sculpture. Pliny the Elder also records that in ancient Rome Â�ceremonies for the dead were organized, in which mimes who physically resembled the departed would impersonate them by putting on their death masks (Beyer 2003:â•›17). The ritual thus combined contiguity in the form of live performance, iconicity in the mimes’ Â�resemblance to the dead, and indexicality in the form of death masks. In this way the spectacle came as close as was physically possible to ‘reviving’ the dead, even closer than today’s home videos, which offer displaced iconically indexical images, but without the Â�illusion of contiguous, live contact with the dead. Photography in the true sense of the word came upon the scene in the early Â�nineteenth century, in 1826, when the Frenchman Joseph Nicéphore Niépce for the first time achieved a permanently fixed iconic index using a camera obscura and the photographic process of silver nitrate. The subject of the first ever taken photograph was nature, a courtyard seen from Niépce’s window, rather than a human figure, mainly because the image required eight hours of exposure to be recorded on the photosensitive plate. It was Niépce’s younger partner Louis Daguerre, who in 1829 developed the first precise photographic images which he called, after himself, daguerreotypes. In the decades that followed, photography became all the rage, chiefly among rich enthusiasts, who carried their cumbersome and expensive cameras and tripods to take Â�pictures of city squares, landscapes, historic monuments, and other large and Â�immovable objects. The exposure time was gradually reduced to fifteen minutes in



The iconic indexicality of photography 

full sunlight in the late 1830s, and then to 20–40 seconds by the early 1840s (Briggs & Burke 2005:â•›133; Freund 1980:â•›22; Benjamin 1980:â•›199–216).

3.â•… Photography versus painting The fascination with the new, initially expensive but very easy and quick way of Â�making accurate and realistic pictures went hand in hand with the recognition of the fundamental difference between photographic images and traditional paintings. Unlike the latter, photography was first perceived not as an art but as a mechanical process of Â�impersonally copying nature, while the photographer was thought to be merely a non-interfering observer – a scribe, not a poet. Indeed, while it obviously took human inventiveness to produce the camera in the first place, the very creation of the image was taken care of by the optical and photochemical processes largely beyond the Â�photographer’s control: light reflected from objects was focused through the lens and exposed to chemically coated glass or metal plate, later replaced by flexible celluloid, before undergoing a complex chemical process resulting in a photograph – a record (that is, an index) of reflected light at an instant in time. At the outset, photography was thus considered more of a natural phenomenon than an artifice, much like Â�magnetism or electricity – phenomena understood to be discovered rather than invented, and Â�subsequently applied to human use (Marien 2006:â•›23). Of course, the photographic process was not totally accidental, because it involved the photographer as the conscious agent who chose the object to be Â�photographed, placed the camera in front of it, and allowed the processes of nature to create a desired image. But it is these natural, physical processes, applying largely beyond the Â�photographer’s control, that determine the inherently indexical Â�character of Â�photography, so different from the iconicity of painting or drawing, which are created entirely by the artist. It is because of this fundamental difference between intentional and personal (that is, iconic) depiction, and natural and impersonal (that is, indexical) record of reality that we say that a painting is ‘made’, whereas a Â�photograph is ‘taken’ (Friday 2002:â•›39). While painting is never more than the stating of an interpretation, a photograph, in the words of Susan Sontag, is “a trace, something directly stencilled off the real, like a footprint or a death masks,” which turns a photograph into “a material vestige of its subject in a way that no painting can be” (1978:â•›154). Unlike paintings therefore, photographs are causally dependent on the objects they represent, whereas paintings are causally dependent only on the beliefs and skills of the painter. This also means that paintings, iconic as they are in relation to the objects depicted on them, are also indexical in relation to their authors, which explains why in popular perception, paintings are valued not so much for their subject matter,

 Piotr Sadowski

as because of who painted them. Art critics, dealers, and the general public seem to care less about what a particular painting by Da Vinci, Cézanne, or Picasso Â�represents, than about the fact that it is a painting by Da Vinci, Cézanne, or Picasso. By the same token, the indexicality of photography explains why we are more interested in what photographs depict than in who took them, so that the photographers’ names, even as important as those of Alfred Stieglitz, Robert Capa, or Henri Cartier-Bresson, are not as well known in popular culture as the names of some twentieth-century Â�painters. This is also why naïve, or commercial, or purely utilitarian photography is no Â�different in kind from photography as practised by the most accomplished professionals: there are pictures taken by anonymous amateurs which are just as interesting, just as Â�formally (if inadvertently) sophisticated as professional artistic photography. The stylistic, subjective characteristics of artistic photography, such as controlled studio lighting, skill of composition, clarity of subject, precision of focus, quality of print, elitism of gallery exhibition – all features borrowed from painting – are extraneous to the essence of photography. Its indexical nature, which guarantees a closer relation with the represented object than can be achieved in merely iconic painting, means that amateur, spontaneous, crudely lit, grainy and unfocused, asymmetrically composed photographs can be just as interesting and compelling, if not more so, than the most accomplished painting. The inherent ability of photography to transcribe external reality from its Â�inception gave the medium its documentary validity as both accurate and seemingly unbiased. The writer Edgar Allan Poe, whose own daguerreotype portrait was made in 1849, wrote in an article for a popular magazine that “the closest scrutiny of the photogenic drawing discloses only a more absolute truth, a more perfect identity of aspect with the thing represented” (Quoted in Trachtenberg 1980:â•›38). This is why taking photographs and manipulating them gives one a sense, illusory though it may be, of ordering and somehow controlling reality. A chronological order of snapshots in family albums, or the meticulous filing of photographs used in science, art Â�history, or police investigations turn indexical images of selected fragments of the world into systems of classification, possession, and control. At the same time, the indexical nature of photography gives the information thus obtained the quality of objective evidence rather than of just subjective impression, which is why photographs, film or video footage (from CCTV cameras for example) are accepted as indisputable legal evidence, together with fingerprints and DNA samples. The earliest recorded instance of the use of photography in police investigations was the roundup of the Â�Communards in Paris in June 1871. During the Franco-Prussian War of 1870 and the short-lived Paris Commune of 1871, hundreds of pictures were taken of the Communards, who willingly allowed themselves to be photographed on the barricades. When the Commune fell, the police used these photographs to identify the Communards, who were subsequently arrested and executed. Some were later



The iconic indexicality of photography 

photographed in their makeshift coffins. Since then photographic, film, and video evidence has been used as an effective tool of modern states in the surveillance and control of their increasingly mobile populations (Freund 1980:â•›108). The value of the photographic image both as a truthful record of physical Â�reality and as evidence in crime investigations has been both explored and challenged in Michelangelo Antonioni’s classic film Blow Up from 1966. In the film, a Â�professional photographer accidentally captures on his camera what he thinks is a record of a Â�murder committed in a park. Details such as a hand with a gun sticking out of a bush, and what looks like a dead body lying on the ground, are beginning to show once the photographer enlarges the pictures in his darkroom. However, the larger the Â�incriminating images are the less revealing they become, as the blown-up grain blurs the details, rendering them unintelligible and therefore unreliable as possible evidence of the alleged crime. Paradoxically, the closer the photographer wants to get to the ‘truth’ of the crime using the supposedly objective photographic process, the farther he moves from actually producing any convincing evidence. The largest blow-up of the crucial element of the crime scene, the dead body, ultimately looks on the print like a piece of abstract art – blurred and indistinct beyond recognition, just a chaotic Â�pattern of black and white spots rather than a realistic image that would persuade Â�anyone about the supposed murder. Antonioni seems to remind us of the limitations of Â�photography as a way of representing reality, by exposing the artificial and Â�ultimately elusive nature of the photographic process, whose indexical, displaced images are distinct from and should not be confused with contiguous reality.

4.â•… Photography and the human face As soon as the exposure time in early photography was reduced to 20–40 seconds in the early 1840s, the most popular photographic object became – predictably – the human face. I say “predictably” because our attention to face and facial expressions as signs of personal identity, intentions, and emotions appears to be governed by powerful innate cognitive modules, which make us instinctively alert and sensitive to the facial features of people known to us: relatives, friends, socially important people (what we today call celebrities), and of course ourselves (Ekman 2003). Family histories and Â�genealogies are an inseparable part of recorded history, and must have preoccupied humans ever since the beginning of consciousness and the first awareness of the Â�transience of Â�individual life. Portrait images, however, come late in the history of visual representation, to culminate in Roman sculpted marble busts – three-dimensional icons of remarkable realism, which included also the earliest portrait ‘photographs’: indexical wax masks of dead members of Roman families. With their no-nonsense approach to reality, the Romans insisted on an exactitude that included every wart, pimple, Â�wrinkle, and

 Piotr Sadowski

blemish in their busts, which were certainly done from life, like the much later painted portraits and photographs (Woodford 1982:â•›84). The development of the photographic portrait in the nineteenth century Â�corresponded with the rise of the middle class in Western Europe: by having one’s Â�portrait done, an individual of the ascending bourgeoisie could visually affirm his or her new social status. In this, the bourgeoisie followed the model of the earlier Â�aristocratic pÂ�ortraiture executed on canvas, and of portable, miniature portraits on powder boxes and pendants, with tiny portraits of friends, lovers, and relatives. Â�However, by 1850 the miniature painted portrait had all but disappeared under the influence of the new medium of photography. For one-tenth the price of a painted portrait, the Â�photographer could furnish a likeness which satisfied both the taste and the pocket of the widening spectrum of middle-class population (Freund 1980:â•›9–11). The long-established human preoccupation with family history thus received a truly miraculous boost in the form of a new medium which – for the first time in Â�history – allowed people to know exactly what they looked like when they were younger, and what their parents and grandparents looked like as children. Our Â�appearance Â�naturally changes throughout our lives, from day to day, from year to year. The famous series of Rembrandt’s self-portraits from youth to old age shows the artist studying this relentless process, but it was only with the advent of photography that we have become fully aware of this effect of time. We look at the snapshots of Â�ourselves and of our relatives and friends taken a few years ago, and we recognise with a shock that we all have changed much more than we tended to notice in the Â�day-to-day business of living. This is because the feeling of constancy completely Â�predominates over the changing appearance (Gombrich 1982:â•›107). Before the Â�invention of the camera, not even a tiny minority of the rich who commissioned paintings of their children could possess that knowledge, as even the best painted portraits, by virtue of their iconicity, were less informative than a casual family snapshot. At first, to have a photograph taken of themselves, people had to go to the photographer’s studio, with its choice of painted backgrounds and plaster pillars. The popularity of family photographs increased after the 1880s, when the camera became more mobile and could enter more freely the private and domestic space. The introduction of small, easy-to-operate cameras, notably by Kodak, towards the end of the nineteenth century eventually led to the decline of professional studio portrait photography. Hundreds of thousands of people were now learning to take their own pictures that recorded the main events of their lives, which in semiotic terms meant converting direct, contiguous events into indirect, displaced iconic indexes of these events. Private cameras became inseparable from tourism, both as a way of collecting visual trophies as evidence of trips taken and places visited, and as a way of relieving the stress of travelling and adapting, if only for a brief time, to new environments. The very activity of taking pictures while on holidays is soothing and reassuring, in assuaging general feelings of disorientation



The iconic indexicality of photography 

and anxiety likely to be exacerbated by travel. This is probably why given the choice between buying a good-quality postcard and taking an often inferior-quality picture of a visited place with one’s own camera, most people still prefer to do the latter. The popularity of postcards was the greatest in the early decades of the twentieth century, when relatively few people could afford their own cameras (Freund 1980:â•›99). Today, in an age of cheap storage of digital pictures on personal computers, the traditional postcard and even printed photographs are gradually disappearing, both as records of one’s travels and as a way of keeping family and friends informed about one’s holidays. (The latter Â�function is now being fulfilled by electronic mail, mobile telephoning, and text messaging.) While it can be argued that taking pictures all the time while on holidays gets in the way of contiguous, direct experience, photographing the visited places seems to establish a sense of instant appropriation and an illusion of control over the new and otherwise unfamiliar surroundings. In the long term, photographs will also exist as spatio-Â�temporally displaced indexical records and memory aids of events that have ended, of people who have gone away or died, conferring a kind of immortality and importance that the photographed objects would never otherwise have attained. The quasi-magical sense of appropriation and control over photographed objects is also responsible for the often uncomfortable feeling people have when being Â�photographed. The anxiety is due to the combined effects of photographs as iconic indexes and of our instinctive, if illusory and incorrect recognition, Â�underlying Â�universal magical thinking, that both indexes and icons possess a direct though Â�hidden sympathetic link with the objects they depict or refer to, in this case with Â�ourselves. In this sense a photograph, or a related medium such as film or the old Â�camera obscura, can exert an even more powerful ‘magical’ effect than painting, Â�precisely because of the combination of iconicity and indexicality in the photographic media. The ‘magic’ of photography is amusingly illustrated by an account given by the early-Â�nineteenthcentury English explorer Edward Dodwell during his visit to Â�Athens, then under Turkish occupation. While in Greece Dodwell remained under the supervision of a Turkish official, who constantly put difficulties in the explorer’s way, which could only be overcome by bribes, until one day Dodwell got rid of the mercenary Turk in a most extraordinary way. While examining the ancient Â�monuments Dodwell was using a camera obscura, and no sooner did the Turk see the Temple of Acropolis instantaneously reflected on the paper in all its lines and colour, than he imagined that I had produced the effect by some magical process. He was visibly astonished and alarmed. When again he looked into the camera obscura, some of his soldiers just happened to pass before the reflecting glass, and they were beheld by the astonished Turk walking upon the paper. He now became outrageous, and after calling me pig, devil, and Buonaparte he told me that if I chose I might take away the temple, but that he would never permit me to conjure his soldiers into my box. When

 Piotr Sadowski

I found that it was in vain to reason with his ignorance, I changed my tone, and told him that if he did not leave me unmolested, I would put him into my box, and that he should find it a very difficult matter to get out again.  (Gernsheim & Gernsheim 1969:â•›29)

Visibly alarmed, the Turkish official immediately retired discomforted, and henceforth carefully avoided Dodwell and his dangerous box. A person from the iconoclastic Islamic world could react with superstitious fear to the allegedly appropriating power of photographic images, but even in the West, with its long history of rich visual culture, people still tend to feel apprehensive and selfconscious at having their pictures taken. We know rationally that cameras do not ‘steal our souls’ and take us into the photographer’s possession, as tribesmen were reported to feel when first confronted with a white man’s camera, and therefore we do not panic when we are being photographed. Still, the residual magical fear remains, making us feel awkward and uncomfortable, as if our privacy has been invaded, and some trespass and disrespect have taken place. Few people in today’s Western societies share the primitive dread of cameras, but some trace of irrationalism is still there, as seen in our reluctance to tear up or throw away a photograph of someone once loved, or dead, or living far away. But no matter how anxious and uncomfortable we felt at the time of being photographed, and how much we disliked our portrait when we first saw it, we tend to like photographs of ourselves with time – as we grow older we always prefer images of ourselves when we were younger. In this way, portrait photography exerts a double psychological effect: on the one hand, it is a way of arresting the damages caused by time and of achieving a sort of perennial youth and, on the other hand, photographic portraits are a sobering reminder of how illusory and self-deceptive our desire to freeze the past is, as the difference between the face on the photograph and the one in the mirror grows wider and wider with time. One is reminded of the pathetic obsession with which an ageing silent film queen, played by Gloria Swanson in Billy Wilder’s Sunset Boulevard (1950) watches over and over again the old films in which she used to star, living totally in the illusion of the past and unable to accept the passage of time. The two opposite effects of photography, one giving us a pleasing illusion of timelessness and the other painfully reminding us of the ravages done by time, thus cancel each other out, leaving us not better off perhaps than were the people before the age of photography, who never had to experience the frustration of comparing their Â�current appearance with their younger image. Without photographic portraits, people are both denied the pleasure of seeing their young selves ‘immortalized’ and spared the anguish and nostalgic sorrow of seeing their youth gone. The popularisation of photography in mid-nineteenth century, together with the invention of moving pictures towards the end of the century, marked the beginning of iconic indexical culture that was to dominate Western societies, gradually to Â�dominate



The iconic indexicality of photography 

over the early modern culture based on iconic, painted images and the symbolic, printed word. Photography not only offered a democratic opportunity for anyone to possess and contemplate their own and their families’ realistic portraits – it also affected every aspect of social life, providing the technological basis for Â�twentieth-century visual mass media such as photojournalism, advertising, cinema, television, video, DVD, and the Internet. Still photography, documentary film and television, due to their iconic indexicality and wide dissemination, offered an opportunity of surrogate participation in events not directly experienced, thus enlarging the sphere of individual interaction with the increasingly global social environment.

References Alberti, L.B. 1972[1435]. On Painting and on Sculpture. The Latin Texts of De Pictura and De Statua, C. Grayson (transl.). London: Phaidon. Anderson, E.R. 1998. A Grammar of Iconism. Madison NJ: Fairleigh Dickinson University Press/ London: Associated University Presses. Bazin, A. 1980. The ontology of the photographic image. In Classic Essays on Photography, A. Trachtenberg (ed.), 237–244. New Haven CT: Leete’s Island Books. Benjamin, W. 1980. A short history of photography. In Classic Essays on Photography, A. Trachtenberg (ed.), 199–216. New Haven CT: Leete’s Island Books. Beyer, A. 2003. Portraits: A History. New York NY: Abrams. Briggs, A. & Burke, P. 2005. A Social History of the Media: From Gutenberg to the Internet, 2nd edn. Cambridge: Polity Press. Carroll, N. 1995. Towards and ontology of the moving image. In Philosophy and Film, C.A. Â�Freeland & T.E. Wartenberg (eds), 68–88. London: Routledge. Chauvet, J.-M., Deschamps, E.B. & Hillaire, C. 1996. Chauvet Cave: The Discovery of the World’s Oldest Paintings, P.G. Bahn (trans.). London: Thames and Hudson/New York: Abrams. Corballis, M.C. 2002. Did language evolve from manual gestures? In The Transition to Language, A. Wray (ed.), 161–179. Oxford: OUP. Ekman, P. 2003. Emotions Revealed: Understanding Faces and Feeling. London: Weidenfeld & Nicolson. Fischer, O. & Nänny, M. (2001). Introduction: Veni, vidi, vici. In The Motivated Sign. Â�[Iconicity in Language and Literature 2], O. Fischer & M. Nänny (eds), 1–14. Amsterdam: John Benjamins. Freund, G. 1980. Photography and Society. London: The Gordon Fraser Gallery. Friday, J. 2002. Aesthetics and Photography. Hants: Ashgate Publishing. Gernsheim, H. & Gernsheim, A. 1969[1955]. The History of Photography from the Camera Obscura to the Beginning of the Modern Era. London: Thames and Hudson. Gombrich, Ernst H. 1982. The Image and the Eye: Further Studies in the Psychology of Pictorial Representation. Oxford: Phaidon. Gombrich, Ernst H. 1995. Shadows: The Depiction of Cast Shadows in Western Art. A Companion Volume to an Exhibition at the National Gallery. London: National Gallery Publications/ New Haven CT: Yale University Press. Homer. 1980. The Odyssey, T.W. Allen (transl.). Oxford: OUP.

 Piotr Sadowski Leroi-Gourhan, A. 1968. The Art of Prehistoric Man in Western Europe. London: Thames and Hudson. Links, J.G. 2005 [1982]. Canaletto. London: Phaidon. Marien, M.W. 2006 [2002]. Photography: A Cultural History. London: Laurence King. Peirce, C.S. 1998[1931–1958]. Collected Papers of Charles Sanders Peirce, C. Hartshorne & P. Weiss (eds). Bristol: Thoemmes Press. Pliny the Elder 1991. Natural History: A Selection, J.F. Healy (transl.). London: Penguin Books. Sadowski, P. 2009. From Interaction to Symbol: A Systems View of the Evolution of Signs and Â�Communication [Iconicity in Language and Literature 8]. Amsterdam: John Benjamins. Sebeok, T.A. 1994. Contributions to the Doctrine of Signs. Bloomington IN: Indiana University/ Lisse: The Peter de Ridder Press. Sontag, S. 1978[1973]. On Photography. London: Allen Lane. Trachtenberg, A. (ed.). 1980. Classic Essays on Photography. New Haven CT: Leete’s Island Books. Vasari, G. 1987 [1568]. Lives of the Artists. London: Penguin Books. Walton, K.L. 1997. On pictures and photographs: Objections answered. In Film Theory and Â�Philosophy, R. Allen & M. Smith (eds), 60–75. Oxford: Clarendon Press. Woodford, S. 1982. The Art of Greece and Rome. Cambridge: CUP.

Unbinding the text Intermedial iconicity in Peter Greenaway’s Prospero’s Books Christina Ljungberg University of Zurich

Intermediality concerns either the transgression of the boundaries between conventionally distinct media or the iconic enactment of one medium within another. How does this function in such a complex multimedia work as Prospero’s Books, Peter Greenaway’s film adaptation of Shakespeare’s The Tempest, which not only comments on other media but also addresses all other adaptations of Shakespeare’s play? How is the figure of dislocation – shipwrecks, loss of home and culture, the disorientation generated by Prospero’s masques – translated into Greenaway’s postmodern adaptation? To what degree does adaptation itself involve dislocation as Shakespeare’s figuration of dislocation resonates throughout Greenaway’s multimedia reworking of this text into contemporary sensibilities? Not only does the film self-reflexively perform the very process of adaptation but by ‘destructuring’ – or dislocating – the text into images, it also creates a visual vocabulary articulating a new order of reading and suggesting a new visual literacy.

1.â•… Introduction Prospero’s Books, Peter Greenaway’s 1991 adaptation of Shakespeare’s The Tempest, is a remarkable production in many respects. Challenging its audience with a dazzling Â�neo-Baroque multimedia experience, the tension generated by the constant Â�border crossing among media such as writing, film, painting, and their digital varieties explores both the history and potential of each medium, in particular their Â�media-specific Â�materiality. In so doing, Greenaway not only gives us a commentary on how we should ‘read’ the film but he also openly interrogates the process of its adaptation while Â�displaying what is involved in such a project for a postmodern filmmaker adapting an almost four-hundred-year-old canonical play to contemporary sensibilities. What makes Greenaway’s Prospero’s Books so intriguing for an analysis of the intersemiotic processes involved in adaptation is that the relationship between Â�Prospero, the magician, and his books shows a conspicuous similarity to the relationship between

 Christina Ljungberg

Â� Greenaway, the director, and Shakespeare’s text. The film is, as Peggy Phelan (1992:â•›43) points out, about “a playwright leaving the theatre, performed by an actor [John Â�Gielgud] who may be thinking about retiring, and is directed by a man who wants to redefine the properties of the filmic frame”. What does Greenaway do with Â�Shakespeare? What strategies does he employ to adapt Shakespeare’s play, which already Â�self-reflexively interrogates the media of books, language, visual design – and, in particular, the role of the director – to evoke imagination and the “fancies of the brain” (Greenblatt 1997:â•›3048), which are the core elements of both magic and Â�theater – and film? The key issue here, I would argue, lies in the performative Â�process of Â�adaptation, in which both the play as well as its screen transmogrification are Â�performances, but in which the constituent elements of the pre-text are taken apart, analyzed, and Â�re-ordered. Thus, new configurations are generated from already existent parts, creating something entirely new and thereby often disclosing hidden or unsuspected relationships, which in turn leads to fresh insights. Such processes are characteristic of what Charles Sanders Peirce defines as diagrams, which together with images and metaphors form the three types of iconic signs. Peirce (EP 2: 212–13) describes the creation of a diagram as a process which involves construction of “hypothetical state[s] of things”, the forming “a plan of Â�investigation”, the selection of “the features which it will be pertinent to pay attention to”, and finally the repeated return to some of these features. Peirce’s description of these cognitive processes is easily applicable to those of intermedial adaptation. Â�Furthermore, the intersemiotic transformations involved are both self-reflexive and entail the (re-) mapping of structures; therefore, they have a conspicuously iconic Â�character. What kind of mapping takes place in such a highly complex adaptation as Prospero’s Books? This work would seem to pose a particular problem since its Â�intricate Â�structures interact on several planes, including the level of the media, which involves both film and digital media as well as those engaged in productions of Â�theater (and Â�literature). What new meaning does such deconstructing and Â�subsequent Â�reassembling of the Â�pretext’s constituent elements generate, and how do they affect our Â�understanding and interpretation of the film? How is the figure of dislocation – shipwrecks, loss of home and culture, as well as the disorientation generated by Prospero’s conjured masques – translated into Greenaway’s postmodern adaptation? And more specifically, to what degree does adaptation in itself involve dislocation, or the “disruption of an Â�established order” as Merriam Webster defines the word? I will argue that the figure of Â�dislocation governs Shakespeare’s play as well as it resonates throughout Â�Greenaway’s performative multimedia reworking of this canonical text into contemporary Â�sensibility. Not only does the text self-reflexively perform the very process of adaptation, but it also becomes ‘destructured’ into images, images with which Greenaway plays as with 1words, showing the extent to which verbal language is visual and calling for a new visual literacy in order to ‘read’ our visually dominated multimedia culture.



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

2.â•… Adapting the pre-text Although film adaptations of famous literary works have accompanied film since the medium’s invention, one would tend to agree with what Robert Stam (2007:â•›4) states in his seminal introduction to Literature and Film (Stam & Raengo 2007): “Too often, adaptation discourse subtly reinscribes the axiomatic superiority of literature to film”. Such discourse also plays on fears that the text will be appropriated by Â�visuality and other sensory modalities, enacting what W.J.T. Mitchell (1980:â•›43) once called “the struggle for dominance between pictorial and linguistic signs”. This Â�manifests itself in notions that, for instance, the source or pre-text will be overwhelmed by the Â�audio-visual-verbal spectacle on the screen, drowning it in lavish costumes and period décor; that the adaptation might either be too literal a transformation of the pre-text or too radical an interrogation of the text, unsettling and alienating the Â�audience; or that the translation of the pretext’s cultural memory and context into another medium is not possible. Often considered as mere ‘borrowings’, adaptations therefore have been judged as to their literary quality. Moreover, Lawrence Venuti (2007:â•›24) argues that a film adaptation of a novel is often considered an “unfaithful or distorted Â�communication of the author’s expressive intention” (ibid.: 26). But, as he points out, such a view more often than not entails the unconscious use of a third term, “a Â�dominant or authoritative interpretation of the text, which the critic applies as a standard on the assumption that the film should somehow inscribe that and only that interpretation” (ibid.: 26). Such critical practice is of course especially common when it comes to the Â�adaptation of a canonical literary text like Shakespeare’s The Tempest, which has already accumulated a body of commentary and ‘authorized’ interpretations. The very mixed reception of Greenaway’s film illustrates this clearly. The criticism of the film ranged from seeing it as “a kind of illustrated essay on Shakespeare and his hero” which Â�“references the masterpieces of the past in a manner that antagonizes our pleasure in the arts rather than enhancing it” (Washington Post’s Hal Hinson 1991), “a Â�cluttered spectacle [which] yields no overriding design but simply disconnected MTV-like Â�conceits or mini-ideas every three seconds” (Chicago Reader’s Jonathan Rosenbaum 2008) but also a “ravishing but incomprehensible adaptation of The Â�Tempest” (Joe Brown, also Washington Post 1991) and “a collection of tumultuous visions, suggesting a great library of the gods as imagined by Renaissance painters” (New York Times’ Vincent Canby 1991). This criticism testifies to the large variety of preconceived ideas, understanding and expectations of adaptations. Film adaptations must necessarily perform substantially diverse operations on the source text’s formal and thematic features in order to translate these into film language. This means that they, on the one hand, could be regarded as “parasites” or “hybrids” as Stam (2007:â•›3) suggests with reference to Spike Jonze’s and Charlie Kaufman’s take on orchids in their film Adaptation

 Christina Ljungberg

(2002). This film about writers and writing performs the procedure and problematics of adapting a book for the cinema both thoughtfully and hilariously. The orchid metaphor – the book in question is Susan Orlean’s The Orchid Thief (1998) – not only suggests the pervasive hybridity of adaptations but also that they can be seen as “mutations”. Such a mutation could ensure its survival by adapting to “changing environments, changing tastes, as well as to a new medium, with its distinct Â�industrial demands, commercial pressures, censorship taboos, and aesthetic norms” (Stam 2007:â•›3). Hence, what must be explored are the “unstated doxa which subtly construct the subaltern status of adaptation (and the filmic image)” (2007:â•›4).1 Stam’s use of the postcolonial term ‘subaltern’ in this context makes a telling Â�analogy between film adaptation and postcolonialism as well as their relationships with dominant discourses. The postcolonial ‘subaltern’ (Guha & Spivak 1988) always deals with a disadvantage towards hegemonic discourses and must therefore develop his on her own. Similarly, Stam’s (2005:â•›5) suggestive formulation of the “subaltern Â�status of adaptation (and the filmic image)” above establishes the (imagined) authority, control and appropriation discursively exercised over adaptation by the source or pre-text. But more recently, adaptation has been looked at as a kind of performative criticism, as a critical interrogation and evaluation of the pre-text which produces a new work of art. The concept of the performative as developed by J. L. Austin, Judith Butler, Jacques Derrida and others is that dimension of discourse which generates new Â�‘realities’. Â�Performative statements are neither true nor false since the reality to which they refer is only created by their being uttered. For example, a judge who Â�condemns a murderer to a penalty of prison utters neither a true nor a false statement; he or she creates the Â�reality of the consequences of this condemnation by uttering the Â�judgment under Â�certain necessary and appropriate circumstances. Likewise, authors of works of Â�literature, a film or a film adaptation of a literary work may be said to create Â�performatively. Whatever they present (i.e. perform) as real is a reality created by their very presentation. As Stam points out, Just as the literary utterance creates the state of affairs to which it refers – rather than merely imitating some pre-existing state of affairs – so the filmic adaptation might be said to create a new audiovisual-verbal state of affairs, rather than merely imitating the old state of affairs as represented by the source novel. (Stam 2007:â•›11)

1.â•… Stam (2007:â•›4–7) mentions eight sources of hostility to adaptations, namely literature’s “putative superiority”, “dichotomous thinking”, “iconophobia”, “logophilia”, “anti-corporeality”, “the myth of facility”, and “class prejudice” and “parasitism”, i.e. that films feed on their literary source text.



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

Rather than considering adaptations as ‘borrowings’ attempting to affirm the Â�continuity of traditions and genres, they could be analyzed as instances of Â�discontinuity, of Â�cultural conflict and mutation (to some extent even mutilation): in short, as Â�polyvocal dislocations that celebrate the hybrid and performative quality of adaptation. Such forms of hybridity and dislocation reveal adaptations to be highly dialogic speech acts in which the artistic utterance always mingles its own expressions with someone else’s (cf. Bakhtin 1984). Adaptation thus becomes an orchestration of discourses, talents, and tracks, but also a space in which all these various elements are located – and dislocated. This dialogic quality could also be the reason why – despite criticism to the Â�contrary – an adaptation is allowed relative independence from its pre-text because the two are both similar and different but never identical. A film adaptation is always an interpretation produced at a certain historical and cultural moment. Moreover, it must be produced according to certain medium-specific rules (screen play, film Â�format); it will depend on the intertextual or intersemiotic relations between film and other Â�cultural practices; and it will necessarily be influenced by the contemporary context and conditions of film production. The adaptation of a work therefore increases its intermedial complexity. It also highlights its performative dimension, as it can be said to create a new state of affairs rather than imitate its ‘pre-text’. Adaptation would therefore seem to Â�necessarily involve what Fauconnier and Turner (2002) call “double-scope blends”, blends that recquire polyvalent iconic representations or otherwise ambiguous Â�multiple Â�interpretations that must first be deconstructed and then reconstructed into new Â�syntheses. That is why adaptations entail intermedial iconicity, since they either Â�concern the Â�transgression of boundaries between conventionally distinct media of communication (while retaining formal features of the pre-text) or the iconic Â�enactment of one medium within another.

3.â•… Adaptation and translation An intriguing feature of Greenaway’s adaptation is that it is an almost verbatim Â�translation of Shakespeare’s text about Prospero, formerly the Duke of Milan, who was usurped by his brother and exiled, put in a boat that eventually, together with his daughter Miranda, lands him on an island. A sympathetic friend, Gonzalo, Â�provides him with books from his library, which enables Prospero to become a magus, a Â�powerful wizard who is able to find his way across the oceans, to combat the witch Sycorax, to colonize the island, and take her son Caliban as his slave, to educate Miranda and then, using his magic powers, to revenge himself on his brother with the help of Ariel, a spirit of the air.

 Christina Ljungberg

But Greenaway’s film, as James Tweedie (2000:â•›111) points out, “is not the Â�standard ‘literate’ adaptation, which attempts to emulate the canonical status of a literary object, but a ‘literal’ film, one that deconstructs its subject into infinitesimal elements and translates them, word by word, image by image”. Instead, he contends, the film, “at times literally word for word, both demonstrates how those words were themselves ‘put together’ and performs their disintegration into constituent elements”. By deconstructing the boundaries between verbal and visual culture as well as those of dance and music, Greenaway’s adaptation actually translates the text into such a cascade of images that it blurs the distinction between the narrative and the visual (as well as between dance and music). This opens up the question about the relationship between translation and adaptation, as well as their relation to their sources or pre-texts. Generally, translation is defined as the recreation or transposition of those features – semantic, rhetorical, stylistic and structural – that the translator deems important from a pre-text into a target text of media-specific equivalence, which produces a text in a new medium/language. In contrast, adaptations have to adjust the pre-text to a new medium but instead of being a transposition of the original it maintains core elements of the original in such a way that they are fully recognizable. However, as Lawrence Venuti (2007:â•›25) points out, “[t]he relation between such second-order creations and their source materials is not communicative but hermeneutic, depending on the translator’s or filmmaker’s application of an interpretant”. He suggests translation theory as a way of thinking about film adaptation as it offers “a more rigorous critical Â�methodology”, and more specifically, a more hermeneutic concept of language. In his view, we should think of translation and adaptation as an interpretation that fixes a form and meaning in the source text in accordance with values, beliefs and representations in the target language and culture (Venuti 2007:â•›28). For this operation, the interpretant of the sign, that is, the effects of the signs on the spectators (Peirce EP 2:â•›429), needs to be taken into account since. [i]nterpretants enable the film to inscribe an interpretation by mediating between its prior materials, on the one hand, and the medium and its condition of production, on the other – by providing, in other words, a method of selecting those materials and transforming them into the adaptation through the multimedial choices made by the filmmakers. (Venuti 2007:â•›33)

The interpretant enables a translation, which is a further development of the sign. the interpretant is not the interpreter who is the personal agent but much more that “in which a sign as such results” (Colapietro 1993:â•›122). An interpretant can be either formal or thematic: as Venuti (2007:â•›33) explains, a formal interpretant involves a “relation of equivalence, such as a structural correspondence maintained between the adapted materials and the film”; it may also include “a particular style, such as a distinctive set of formal features that characterize the work of a director or studio”,



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

or comprise a “concept of genre that necessitates a manipulation or revision of the adapted Â�materials”. A thematic interpretant involves “codes, values, ideologies” such as a previous interpretation of the adapted material, the moral or cultural stance adopted by the filmmakers or a political position: “a number of interpretants will be applied in any adaptation, even if in the long run they might be grouped into more general Â�categories as the analysis proceeds and the field is articulated in an overall Â�interpretation” (Venuti 2007:â•›34). The hermeneutic relation is therefore not only an interpretation of the pre-text but engages with it much more forcefully in that it interrogates it, and in so doing, exposes under what cultural and social conditions it has been produced, as well as what translation or adaptation that has processed them. The metaphor of projection seems to me to be fruitful in this context, not only because the interpretant involves projection in itself. Like maps, adaptations often entail modes of projection for their interpretation since they are mapped from the medium context of one medium to the one of another. Moreover, similar to a map projection, which must translate a three-dimensional reality into two-dimensional space,2 the projection at work in adaptation will necessarily always involve some kind of distortion. Although this is just as true of translations, the change in medium Â�necessary in adaptation poses additional hurdles. Features of the projected domain are dramatically altered through additions, deletions, or substitutions, which force us to negotiate – or renegotiate – continuously the relationships between the various media at hand, constructing a mental diagram by drawing a new map of interconnections. Projections are pervasive throughout Prospero’s Books because Greenaway Â�provides Prospero with a Book of Mirrors as well as using mirrors and reflections to conjure up and control his plots. Moreover, as Peggy Phelan points out, Combining the Renaissance convention that theater’s job is to hold a mirror up to nature and the intricate contemporary critical theory of the cinema as both a mirror and a screen for the spectator’s ‘identifications’, Greenaway’s film simultaneously projects a deeply accurate Renaissance worldview and an exhilarating illustration of the most technically innovative possibilities of contemporary cinema. (Phelan 1992:â•›44)

Consider, for instance, how the projection of dislocation is translated from the play to the film. Shakespeare’s text is a story of loss and gain; it is also a story of Â�geographical and cultural dislocation (and partial relocation), which gives it its character of Â�disorientation and Unheimlichkeit in the sense of the “un-housedness” or

.╅ A map projection will distort the relationships between five geographical features, namely area, angle, gross shape, distance and direction (see Ljungberg 2005; Monmonnier 1996).

 Christina Ljungberg

“not-at-home-ness” that is often used in postcolonialism (cf. Ashcroft, Griffiths & Tiffin 2000:â•›73). Prospero’s Â�predicament is the consequence of his being too absorbed in his books, losing touch with reality and therefore becoming ‘dislocated’ from his duties as the ruler of Milan. To Miranda, the dislocation from Milan makes her subject to her father’s complete control and magic powers, including the fact that he involves her in his revenge on his shipwrecked enemies. Caliban, who as the son of Sycorax is the island’s Â�legitimate ruler, is turned into a colonially dislocated subject who is forced not only into Â�subservience by his new master’s refusal of acknowledgement but also to adapt his master’s language and culture – and being ridiculed for his inability to do so (cf. Hulme 2003). In the play, the figure of dislocation thus rules the diegesis from its very first scene when the shipwreck brought about by Prospero’s violent tempest brings his enemies to his island. In Greenaway’s adaptation, dislocation pervades not only on the diegetic level but also on the structural and thematic levels. Deconstructing cinematic conventions, he undoes in particular those codes pertaining to the classical Hollywood cinema such as spatial perspective and temporal logic. Temporal continuity is suspended in relation to the dislocations of the historical references to art and art history, e.g. by including objects that were produced after Shakespeare’s time (like the Piranesian bath house, Prospero’s ‘prophetic borrowing’). In Greenaway’s cinema, time can be Â�dislocated, reversed and differently conjoined – or undone, as in a palindrome (cf. Ljungberg 2007:â•›249). To Greenaway, Phelan (1992:â•›44) argues, “time is a way of counting, an alphabet that can be recited backwards or forwards”. Instead of starting events on board the ship and have the plot develop sequentially, Greenaway hence opens his film with a dazzling montage in which time and space are taken apart and again juxtaposed to disjunctively voice the film’s concerns. What we see is a recurring frame of a Â�back-lit drop of water splashing in slow-motion into a black pool, followed by a quote from Shakespeare’s play which we simultaneously watch being written into a book and hear being read out by Gielgud/Prospero in voice-over: Knowing I loved my books, he furnished me From mine own library with volumes that I prized beyond my dukedom. 

(Act I, Sc 2, 167–169)

Then frames of dripping water intercut with those of a quill writing on paper until we arrive at A Book of Water, flipping its pages. The book, set within a rectangular frame superimposed on the background, exhibits da Vinci-like drawings of whirlpools, hydraulics, and diagrams of the Archimedes Screw; waterfalls and cataracts are intercut with animated frames displaying clouds and storms which then �alternate with the dripping water and close-ups on the blue ink-well. The flood of images then stops on a hand in the frame on which the water drops. Superimposed on a �handwritten �manuscript permeated by clouds, it is followed by a traveling wide shot



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

to the �Piranesian bathhouse and pans over to the pool in which Prospero is immersed, his hand stretched out to catch the drops of water, still within the rectangular frame (Figure 1).

Figure 1.╇ Sir John Gielgud as Prospero stretching out his hand

We see a hand writing ‘Boatswain’ on paper, superimposed on Prospero, who is still in the frame, then an intercut between Ariel and Prospero’s hand, followed by a full shot of the book and, in it, an animated film strip with a ship fighting the waves in a tempest; a subsequent cut to Prospero shows him pronouncing the lines as he is Â�writing them, creating the impression of a command echoed on board a ship. A close-up of the inkwell follows, then a cut out of the rectangular frame to Prospero in his magic robe, sitting at his pulpit surrounded by model tall ships in his study and, finally, a ship in storm, its captain issuing the order “Bestir! Bestir!”, repeated orally and visually; the book in the frame starts to undo itself, its pages loosening and swirling through the air. Venuti’s emphasis on the importance of the interpretant as the translating instance would indeed seem apt for an analysis of this sequence. This rapid montage has no direct counterpart in Shakespeare’s play. What is even more interesting here is Â�Greenaway’s take on dislocation as the disruption necessary for the creative act and for the film’s Ur-Szene: the very first image, the drop of water, the first book, The Book of Water – which is countered by the drowning of the twenty-two books (except for the two saved by Caliban) at the end, making water the beginning and (almost) the end of everything and thus functioning as a frame in itself. With this rapid sequence of images – water, inkwell, quill, paper, text and book – Greenaway confronts us from the very start with the process of writing, i.e. the production of texts, and books as

 Christina Ljungberg

the main objects of desire – which are also the site where conflicts are acted out and relationships contested. But the pen’s scratching noise on the page reminds us, too, that writing is not only a precondition for the writer’s imaginary, fictive world in which he can move; it is also the space in which he can place himself as the protagonist whose “fancies of the brain” (Greenblatt 1997:â•›48) can change the world. As Greenaway writes in his screenplay, The storm winds have been let into the library by the person of Prospero Â�himself – who now strides into the space of swirling papers […] we now have two Prosperos […] the Prospero – dressed in blue – sitting at the desk who has dreamt up the scenario […] and the Prospero as actor in his own drama – […] dressed in Â�red – walking past. (1991:â•›20)

Figure 2.╇ Sir John Gielgud as Prospero in his study as the books are unbinding themselves

By having Prospero appear in both roles amidst the papers eddying from the unbinding books (Figure 2), Greenaway enhances the dual function of Prospero as both the magician/writer and the protagonist/character. Prospero’s role as the creating instance is also reinforced by his taking on all the lines up to the play’s crucial turning point in Act 5, when Ariel convinces him to show mercy, which then gives his characters voice – and humanity. Moreover, Greenaway’s film problematizes Prospero’s complex relationship to his books. Functioning as the condition for both his survival on and his control of the island they are also the reason for his present predicament. If his books once led to his ousting from Milan, his new writing in his cell-like study both sets him apart and allows him to take part. The books therefore play a dual role; they Â�alienate him from the external world at the same time as they permit him to interact with it.



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

There is thus an ongoing resistance here between the book as object and the texts it attempts to control but which elide representation. As James Tweedie (2000:â•›107) points out, “Greenaway’s film presents an extended meditation on the relationship between the mystical authority of the book and the hybridizing act of reading and writing”, which is manifested in the film’s Baroque explosion of sumptuous visual details, a “visual and conceptual ‘cacography’” that matches the “cacophony of Prospero’s ‘isle full of noises’” (Romney quoted in Tweedie 2000:â•›108). This “cacography” contains disorder and dislocation – not only do the books’ contents refuse to stay contained but the books also themselves escape or disintegrate into pages swirling in the air, unbinding themselves into pages come alive. Rather than the scientifically ordered inquiry into knowledge that the books purport to be, they present themselves as the source of unbounded knowledge, which is unfolded in front of us in an endless unordered disarray defying location. Similar to the dislocation of spatial perspective and Â�temporal continuity, the ‘logic’ seemingly guiding the twenty-four books (their number being suggestively equal to that of film frames per second) is more reminiscent of Jorge Luis Borges’ “Chinese encyclopaedia” with its highly heterogeneous categorization of Â�animals into groups of those “(a) belonging to the Emperor (b) embalmed, (c) tame, (d) suckling pigs, (e) sirens, (f) fabulous, (g) stray dogs, (h) included in the present classification…” that fascinated Michel Foucault in The Order of Things (1970: xv). The idiosyncratic selection of books thus has Â�Prospero’s library come across as “a museum of image-making and intermediality founded by Â�alternatively organized principles; the book is a space where architecture coexists with related manifestations of ‘music’, and where text and image manifest the same lineage” (Tweedie 2000:â•›111). Such different logic also prevails in the seemingly clear division of the film into past, present and future. As Greenaway (1991:â•›13) argues, this overall Â�distribution means that “the Past corresponds to Prospero’s long explanation of his history, the Present deals with Prospero’s various real-time plottings and the Future Â�concerns those plans Prospero makes to guarantee the success of his dynastic ambitions for his daughter”.3 However, this organization – into three unequal parts – also Â�performs the “deliberate confusion between fact, memory, and fantasy”, as an old man’s Â�recollection of events played out in a space that is “basically dark and richly coloured, influenced by late seventeenth Italian and Dutch painting of interiors” (Greenaway 1991:â•›37). The references to, e.g. Piranesi (who was not born until 1720) is explained by Â�Prospero’s magic arts which make him perfectly capable of ‘prophetic borrowing’ (Greenaway 1991:â•›42).

.â•… The film’s more detailed structure consists of ninety-one sections, each designating a Â�different visual location and broken down into numbered shots.

 Christina Ljungberg

Greenaway’s dislocations on the film’s structural and thematic level hence both reflect and productively interpret those on the pre-text’s diegetic level. Apart from the obvious dislocations of Prospero, Miranda and the shipwrecks, the most Â�prominent sign of dislocation on the story level is of course the cultural dislocation of Caliban, whose language and gestures mark him as a slave under Prospero’s power. Long a debate between Early Modern and Postcolonial scholars, the problem of how to Â�interpret Caliban has generated sharp interchanges (Sharpe 1993; Loomba 1998; Hulme 2003). Greenaway’s move to have the dancer Michael Clark interpret Caliban’s Otherness and oscillation between hatred, lust and despair transforms the play’s Â�inherent colonial violence exerted by a dominant power towards its Others. It also beautifully reveals the constructedness of social and cultural hierarchies and systems. That Caliban is the Â�dislocated – and constructed – Other of Prospero’s own subjectivity is already Â�suggested in Shakespeare’s pre-text when Prospero recognizes that “This thing of Â�darkness I/Acknowledge mine” (5.3.278) as both irreparably authorial and human.4

4.â•… Self-reflexivity and performativity In Prospero’s Books, it is, however, most insistently Ariel’s disrupted flight that Â�dislocated him from his aerial characteristic. Prospero’s power over Ariel, his medium, stems precisely from his ability to project the spirit’s incarceration but also his liberation. First imprisoned by Sycorax to suffer twelve years in a “cloven pine”, then liberated by Prospero but constantly reminded of what he owes his ‘saviour’ (and threatened to be locked up again, should he not behave), Greenaway has Ariel appear in four different ages played by four different actors, as a child, a boy, a teenager and a young man. In so doing, Greenaway interlinks not only the history of avant-garde film in which often various actors play the same part but also Renaissance theater in which one actor would play several parts. But, as Peggy Phelan (1992:â•›48) points out, Ariel also functions as “Prospero’s surveying camera, his time keeper, his means of Â�projection, his ‘representational apparatus’”. The relationship between Prospero and Ariel is one fraught with conflicts and challenges but also one which is dynamic and transformative: it is Ariel in Act 5 who has Prospero realize that the revenge he wants to exact is too harsh and has Prospero show mercy by appealing to his emotions, which turns around the tragic play he is writing; finally, it is Ariel, not the books, who ends the film with his flight into the spheres.

.╅ For a discussion on reading and misreading Shakespeare, see Hulme (2003; Loomba 2003).



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

By shifting the focus from the relationship between Prospero and Caliban to that between Ariel and Prospero, Greenaway employs Ariel as his “means of Â�projection” and his “representational apparatus” (Phelan 1992:â•›48). This shift is significant as it puts the film’s emphasis on the relationship between the artist and her or his medium. It therefore indicates the specific interpretants created by Greenaway for his Â�adaptation which emerges as a film in which the pre-text’s expressions are first dislocated and then translated into the images in the books let loose. These unbinding archaic books are strongly suggestive of postmodernism’s ‘free-floating images’ at the same time as they also symptomatically portend Walter Benjamin’s age of mechanical reproduction as well as our own multimedia technologies. A crucial agent in this is Greenaway’s Graphic Â�Paintbox, whose rectangular cutout in the middle of the viewing screen Â�presents yet another defamiliarizing and thus dislocating device (Figure 3). Functioning as a Â�structuring tool – a hands-on mise-en-abyme –, it differentiates the various narrative levels, e.g. by presenting the various books and their content. It has us follow how, inside its frame, certain crucial events are acted out, analeptically and proleptically seen through Â�Prospero’s eyes. These are the events that stem from Prospero’s Â�writing, that is, they mirror his imagination and his virtual (re)collection and thus have no Â�connection with reality. As Greenaway (1991:â•›12) points out, the island is “full of superimposed images, of shifting mirrors and mirror-images – true mir-ages – where pictures conjured by the text can be as tantalizingly substantial as objects and facts and events, constantly framed and reframed”. The Graphic Paintbox therefore diagrammatically structures the film adaptation, informing us on what level an event

Figure 3.╇ The Graphic Paintbox with The Book of Water

 Christina Ljungberg

takes place. At the same time it reflects Prospero’s fantasies and his knowledge stored in the books which it inserts into the main narrative, granting us an insight into how Â�Prospero’s play develops from the books he has read. As a digital instrument, the Paintbox itself functions iconically since it Â�represents numerically translated information of pre-recorded images and animated films that can be manipulated in every possible way, be it by altering, combining, merging or superimposing them.5 As Herbert Klein (1996) states, “there is hardly an image or a shot in the film that is not generated from some previous image”. Â�Therefore, with the Graphic Paintbox, “film is no longer restricted to the mere reproduction of Â�once-recorded images; with this technique, new images are created and the Â�filmmaker becomes a painter” (which Greenaway already is, see Greenaway, Gras & Gras 2000:â•›132). The Paintbox allows free manipulation of previously recorded images that can thus be changed, combined and fused together. This gives the Â�filmmaker unprecedented access to libraries of images and texts that can be called up and merged with images that have been freshly shot or other categories of text or numbers (Figure 4)6. But the Graphic Paintbox also functions metaphorically, as an allegorical reading of the Renaissance convention that theater is a mirror as much as the contemporary critical stance that cinema is as a mirror and a screen for the spectator identification as was mentioned by Phelan (1992:â•›44) earlier. Furthermore, as she (1992:â•›49) points out, “as an allegory of the relationship between the artist and his medium, Greenaway’s film is also then an autobiography of his own role as Magus. In this sense, Prospero’s relation to Ariel mirrors Greenaway’s with his paintbox”. But Prospero’s mind and “archive” also mirror Greenaway’s. Thus, to look for authenticity or “originality” here would, as Klein (1996) argues, be out of place. Instead, “the film shows the process of artistic creation and is at the same time an example of this process. Just as Prospero scribens brings forth the drama through the manipulation of his ideas, the film arises through the manipulation of images”.

.â•… Although digital imaging has become much more sophisticated in the twenty years since Prospero’s Books was produced, Greenaway’s artful manipulation of time and space in the Graphic Paintbox images for the film was revolutionary at the time of the film’s production and is still impressive. For a discussion of his relationship with the Paintbox, see Greenaway, Gras and Gras (2000). .â•… Digital representations are always iconic since there is no longer a physical Â�connection between the subject and the image. The image is created by numerical translation. This is why the reliability of photography’s connection to the real has been so debated with the onset of digital photography. The same goes for digital manipulations – as soon as the Â�indexical anchoring is lost, so is the relationship with reality, which makes the image iconic.



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

Figure 4.╇ The Graphic Paintbox with the Book of Water in the center against the bathhouse background, framed by an exaggeratedly artificial rain

That such diagrammatic manipulation implies dislocation on several levels has already been demonstrated. However, it also means that choices must be made from the dislocated material which, taken out of context, is nothing but manifestations of the same basic stuff. The minute components dislocated in Greenaway’s ‘archive’ can therefore be seen as a visual vocabulary articulating a new order of reading. Â�Arising from what is already at hand, it attempts to open up space for Â�acquiring visual Â�literacy by deconstructing verbal literacy. This ‘destructuring’ not only plays with the Â�concept of images as words; it also shows the extent to which verbal Â�language is visual. What does visual literacy mean to us who belong to highly Â�textually Â�literate cultures? How visual is text and how textual are images? There is even increasing Â�evidence of the iconic character of letters: as Mark Changizi et al. (2006) have shown, natural Â�writing systems are partly visual7. At the same time, the visual language

.â•… Oliver Sacks (2010:â•›24) points out that alexia, word blindness or visual aphasia, is cased by damage to a specific visual word form area in the brain “dedicated to the visual image of words” which is activated in reading. This sudden inability to read by often highly skilled readers demonstrates the extent to which reading, although we might think of it as a Â�“seamless and indivisible act, and as we read we attend to meaning […] is in fact, dependent on a whole hierarchy or cascade of processes which can break down at any point” (Sacks 2010:â•›27). We are born with some sort of system which allows us to make sense of the world; Â�nevertheless, defining and recognizing objects, which takes place in the perceptual act, demands ”a whole hierarchy of functions”. As Sacks points out, ”[w]e do not see objects as such, we see shapes, surfaces, contours and boundaries presenting themselves in different illuminations or

 Christina Ljungberg

of cinema is text based, which Greenaway (2010) has repeatedly deplored, calling for an undoing of “the primacy of the text and the unprimacy of the notion of the image”. Hence, what Greenaway seems to be doing is to thwart our routine cinematic ‘reading’ to show that there are Â�alternatives ways of seeing. By translating the Â�verbal pre-text into wildly associative images, Greenaway’s adaptation is suddenly less an adaptation than a Â�creative translation of Shakespeare’s poetic and evocative text which he then forces us, his viewers/readers, to perform by syntactically combining the disparate elements to literally create new insights8. The experience of being at a loss is critical here since what Greenaway is doing is to expose his “adaptation” to make us aware of what we are unknowingly doing when we are watching movies, not only this one but in particular more conventionally structured ones. This function is connected to a distinctive kind of aesthetic engagement or even experience, which is prior to questions of cognition: we are thrust into a situation in which narrativity is replaced by visuality with the Â�sudden lack of linear narration pushing us to our limits of perception. There is, however, also another aspect of this interaction between adaptation and translation which could explain the curious fact that Prospero’s Books is almost Â�unintelligible to readers who are not very familiar with Shakespeare’s play − at the same time as it forces those well acquainted with it to return to the pre-text over and over again. This shows not only the indexical character of the relationship between Â�pre-text and adaptation, which is what enables orientation and therefore also accounts for the diagrammatic aspect of adaptation. Such mapping processes Â�necessarily involve diagrammatic reasoning, as Peirce (CP 2.272) points out: all thought Â�processes require the formation of an imaginary “sort of skeleton diagram, or outline sketch” of our choices from which we can then make our decisions and achieve new insights. However, this can only be done by projecting the various

Â� contexts, changing perspective with their movement or ours” and then create invariants from this ”complex, shifting visual chaos”. A similar visual recognition is most likely deployed in reading, Sacks says, and ties this to the discovery made by Mark Changizi (2006) and his colleagues who have shown that all natural writing systems share certain features that our neurovisual system can translate. Investigating more than a hundred ancient and new natural language writing systems, they found that all (including alphabetic and Chinese systems) share certain basic topological features with the environment. As they suggest, these ”have been selected to resemble the conglomerations of contours found in natural scenes, thereby tapping into our already-existing object recognition mechanisms” (2006:â•›E117). .â•… As Greenaway pointed out in his talk in Zurich in January 2010, after eight-thousand years of textual literacy he is firmly committed to ”have the textmaker stand in the shadow and the imagemaker proudly come forward”.



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

� outcomes of our choices which then in themselves produce new and different signs by new combinations and juxtapositions.

5.â•… Conclusion Such new assemblages create both dislocation and disorientation, which underlines the extent to which cinematic orientation in Prospero’s Books is only a provisional and Â�temporary achievement in the ongoing rush of intermedial art forms. These Â�projections are also what make Prospero’s Books different from more conventional Â�narrative Â�cinema and why it confronts us with a cognitive challenge. The projections are of course Â�decisive for the reception of a film adaptation, as they may be either Â�reinforcing the pre-text, or “disjunctive”, resulting in contradictory or even Â�opposing Â�interpretations that will be received differently: “The viewer’s interpretant thus becomes a central Â�factor in accessing the significance of an adaptation, raising the question of whether an academic critical discourse can or should take precedence over other, more popular forms of reception” (Venuti 2007:â•›34). Hence, Greenaway’s film involves the projection and performance of dislocation on at least three levels: 1. at the diegetic level, as it involves shipwrecks and people being lost in both time and space, including cultural (Caliban) and spiritual (Ariel) dislocation. 2. at the level of medium, since, in an adaptation such as Prospero’s Books, the Â�multimedia necessarily must adapt to postmodern sensibilities by generating the sense of dislocation so prominent in Shakespeare’s play. Furthermore, Greenaway clearly wants to create a new way of seeing, a new visual literacy that reorganizes our visual attitudes and extends beyond traditional Hollywood narrative cinema. In so doing, he avails himself of what Winfried Nöth (2009:â•›114) calls Â�“performative metareference” and defines as “the reference of the act of sign Â�production itself ”. 3. at the level of reception, since the Baroque flood of images seems to have resulted in both utterly disoriented spectators and admirers. However, as Klein (1996) Â�suggests, Greenaway perhaps “makes excessive use of the possibilities offered by the Paint Box, demanding visual skills that can only be acquired through extensive training in the appropriate codes”, and “even if the film … possibly asks too much of the viewer, it nevertheless does not manipulate him”. We may not be manipulated but we are certainly dislocated since Greenaway’s Â�unbinding of Prospero’s books undoes temporal continuity and, therefore, the fluency with which we Â�usually ‘read’ the events on the screen. The film may then be considered a series of dislocations reflecting not only the extent to which the

 Christina Ljungberg

source text insists on being enacted and re-enacted but also how the film performatively casts us in the roles of Caliban and Ariel, making us Prospero’s – and Greenaway’s – prisoners. These dislocations may be said to lie at the heart of the adaptation – and in turn, carry forward Shakespeare’s own adaptations and translations of stock narratives and other sources, making this a highly self-referential work. Greenaway does not so much attempt to manipulate viewers as engage them in new ways of experiencing film beyond traditional linear narrative by forcing them to performatively interact with it. His Â�self-reflexive manipulation of time and space creates a liminal space in between and thus the end of the film could be read as unwriting or rewriting Prospero’s Â�dialogue: Shakespeare’s Prospero can stop being a magus, a status he has obtained through his knowledge and the magic he obtained from books. He can give up his magus status and quietly return to being the Duke of Milan, whereas we are let free but in an exilic Â�position, barred from a return home. More generally, Â�thinking of Â�Greenaway’s Â�adaptation in terms of projection and space would suggest the extent to which Â�multimedia are liminal configurations. One medium merges into another and is inevitably consumed in the process: by fire, water or mildew or rot – and reconstructed? If adaptations, as Robert Stam (2007:â•›45) suggests, “engage the discursive energies of their time, they become a barometer of the ideological trends circulating during the moment of production”, then Greenaway’s adaptation points to the necessity to adopt a new visual vocabulary in order to be able to ‘read’ an ever increasingly Â�visually dominated multimedia culture. That is the “mutation” potential that adaptations offer and which implies that a hybrid art form, although it may ‘rob’ a canonical text of some of its meaning, can, in so doing, produce something new and thus contribute to its survival, regardless of medium. That is after all what is self-reflexively suggested by the film’s last frames capturing Ariel’s flight, as he is set free. Running across a full shot of the clapping court which freezes and first becomes a background still, then a world map, he gains speed to leap out of the frame to other spheres. The creative process will always find new ways of articulating itself; the artist’s spirit in whatever form will always survive. In Peter Greenaway’s adaptation of Shakespeare’s text, intermedial iconicity therefore performatively, diagrammatically and metaphorically reflects the narrative structure, pretexts and contexts of this self-interrogating work. By having Gielgud’s Prospero so ostentatiously colonize and control everyone else with his voice and his magic, Greenaway creates a parallel to his own casting of himself as an old-fashioned ‘auteur-style director’. Hence, he makes a caricature of himself in the all-powerful character of Prospero, and thus self-reflexively and ironically distances himself not



Intermedial iconicity in Peter Greenaway’s Prospero’s Books 

only from Hollywood conventions but also from the more typically European ‘auteur’ cinema he obviously wants to escape. But Greenaway’s adaptation achieves something more as he taps most radically into our ways of unthinkingly watching film adaptations. By jamming our systems of interpretation to the effect that even the most readily intelligible signs require slow and painstaking analysis, he makes us aware of the degree to which the visual Â�language of cinema is imbricated in verbal language. His deconstruction of cinematic Â�language thus demonstrates the complexity as well as the interdependence of both reading and Â�seeing processes and how much this is grounded in iconicity. In so doing, he not only lays bare the extent to which our textual literacy is acculturated at the expense of other kinds of reading but also points to new fields of investigation of how such iconic Â�processes work.

Acknowledgements I would like to thank Jan Baetens, Claus Clüver, Vincent Colapietro, and Winfried Nöth for Â�discussing the paper with me and for their many valuable comments and suggestions.

References Ashcroft, B., Griffiths, G. and H. Tiffin. 2000. Postcolonial Studies: The Key Concepts. London: Routledge. Bakhtin, M. 1984. The Dialogic Imagination. Transl. C. Emerson & M. Holquist. Austin TX: University of Texas. Borges, J.L. 1964. Other Inquisitions 1937–1952. Transl. R.L.C. Simms. Austin TX: University of Texas. Brown, J. 1991. Prospero’s Books. (Review). The Washington Post, 29 November. 〈http://www. washingtonpost.com/ 〉 (May 6, 2010). Canby, V. 1991. Movies with their own language. Review of Peter Greenaway’s Prospero’s Books. The New York Times, 15 November. 〈http://movies.nytimes.com〉 (6 May, 2010). Changizi, M., Zhang, Q., Ye, H. & Shimojo, S. 2006. The structure of letters and symbols throughout human history are selected to match those found in objects in natural scenes. The American Naturalist 167(5): E117-139. Colapietro, V. 1993. Glossary of Semiotics. New York NY: Paragon. Fauconnier, G. & Turner, M. 2002. The Way We Think. New York NY: Basic Books. Foucault, M. 1970. The Order of Things: An Archaeology of the Human Sciences. New York NY: Vintage. Greenaway, P. 1991. Prospero’s Books. New York NY: Four Walls Eight Windows. Greenaway, P. 2010. New possibilities. Guest lecture at the Zurich School of Art and Design (HGKZ). 7 January.

 Christina Ljungberg Greenaway, P., Gras, V. & Gras M. 2000. Peter Greenaway: Interviews. Jackson MS: University Press of Mississippi. Greenblatt, S. 1997. Introduction. In The Norton Shakespeare, S. Greenblatt, W. Cohen, J.E. Howard & K. Eisaman Maus (eds), 3047–3054. New York NY: W.W. Norton. Guha, R. & Spivak, G.C. (eds). 1988. Selected Subaltern Studies (With a foreword by Edward W. Said). Oxford: OUP. Hinson, H. 1991. Prospero’s Books. Review of Peter Greenaway’s Prospero’s Books. The Â�Washington Post, 27 November. 〈http://www.washingtonpost.com/ 〉 (6 May, 2010). Hulme, P. 2003. Stormy weather: Misreading the postcolonial Tempest. Early Modern Culture 3: 1–22. 〈http://emc.eserver.org/1–3/hulme.html〉 Jonze, S. & Kauffman, C. 2002. Adaptation. Columbia Pictures. Klein, H. 1996. ‘The far side of the mirror’: Peter Greenaway’s Prospero’s Books. Ehrfurt Â�Electronic Studies in English (EESE) 12. 〈http://webdoc.gwdg.de/edoc/ia/eese/artic96/klein/12_96. html〉 (6 May, 2010). Ljungberg, C. 2005. Cartographic strategies in contemporary fiction. Orientations: Space/Time/ Image/Word, C. Clüver, L. Hoek, V. Plesch & P. de Voogd (eds), 155–172. Amsterdam: Rodopi. Ljungberg, C. 2007. ‘Damn mad’. Palindromic figurations in literary narratives. In Insistent Images [Iconicity in Language and Literature 5], E. Tabakowska, C. Ljungberg & O. Fischer (eds), 247–265. Amsterdam: John Benjamins. Loomba, A. 1998. Colonialism/Postcolonialism. The New Critical Idiom. London: Routledge. Loomba, A. 2003. The postcolonial Tempest: Response to Peter Hulme’s ‘Stormy Weather’. Early Modern Culture 3: 1–22. 〈http://emc.eserver.org/1–3/loomba.html〉 (10 June, 2010). Mitchell, W.J.T. 1980. Iconology: Image, Text, Ideology. Chicago IL: University of Chicago. Monmonnier, M. 1996. How to Lie with Maps. Chicago IL: University of Chicago. Nöth, W. 2009. Metareference from a semiotic perspective. In Metareference across Media, W. Wolf (ed.), 89–120. Amsterdam: Rodopi. Orlean, S. 1998. The Orchid Thief. New York NY Random House. Peirce, C.S. 1931–1958. The Collected Papers of Charles Sanders Peirce, Vols. 1–6, C. Hartshorne & P. Weiss (eds); Vols 7–8, A.W. Burks (ed.). Cambridge MA: Belknap Press. (Cited as CP). Peirce, C.S. 1998. The Essential Peirce, Vol. 2, Peirce Edition Project (ed.). Bloomington IN: Â�Indiana University Press. (Cited as EP 2). Phelan, P. 1992. Numbering Prospero’s Books. Performing Arts Journal 14(2) (May): 43–50. Romney, J. 1991. Prospero’s Books. Review Sight and Sound 1(5) (September): 44–45. Rosenbaum, J. 2008. Prospero’s Books. Review Chicago Reader Movie. 〈http://www.chicagoreader. com〉 (6 May, 2010). Sacks, O. 2010. A neurologist’s notebook: A man of letters. New Yorker, June 28: 22–28. Shakespeare, W. 1623/1997. The Tempest. In The Norton Shakespeare, S. Greenblatt, W. Cohen, J.E. Howard & K. Eisaman Maus (eds), 3055–3107. New York NY: W.W. Norton. Sharpe, J. 1993. Allegories of Empire. Minneapolis MN: University of Minnesota. Stam, R. 2007. Introduction: The theory and practice of adaptation. In Literature and Film: A Guide to the Theory and Practice of Film Adaptation, R. Stam & A. Raengo (eds), 1–52. London: Blackwell. Tweedie, J. 2000. Caliban’s books: The hybrid text in Peter Greenaway’s Prospero’s Books. Cinema Journal 40(1) (Fall): 104–126. Venuti, L. 2007. Adaptation, translation, critique. Journal of Visual Culture 6(1): 25–43.

Argumentative, iconic, and indexical structures in Schubert’s Die schöne Müllerin Costantino Maeder

Catholic University of Louvain Schubert’s The Fair Miller-Maid (1823), based on a collection of lyrics by Wilhelm Müller, is characterized by piano accompaniments that consist mainly in the iconic imitation of real world sounds and of the main character’s mood, musical strategies which any listener accustomed to classical music may recognize without having studied harmony or counterpoint without having studied musicology. Many claim that Schubert’s music simply illustrates and heightens Müller’s original text but his interpretation of Müller’s poems cannot be reduced to a superficial illustration. The partition shows that Schubert created a simple world of iconically determined sound patterns (flowing brook, hunter’s horn, patterns signifying death), which are quoted unusually often and varied in the very first four Lieder. These iconic, musical words, created dynamically, can contradict the song texts in the following Lieder. Another iconic device uses the opposition of minor and major tonalities, tied to illusion and reality.

1.â•… Prelude In Die schöne Müllerin,1 his song cycle composed in 1823, Franz Schubert does not Â�simply illustrate a series of once extremely popular poems by Wilhelm Müller (Müller & Schubert 2001), an early romanticist (1794–1827). Many scholars assume that Schubert offers a mere background, a superficial accompaniment, and wonder why he chose a superficial, literary jest2 for one of his finest works (Reed 1978; Youens 1997). The plot of the song cycle is indeed deceptively simple; many scholars were quite expectedly baffled about Schubert’s choice: a young miller leaves his masters and wanders happily

.â•… All examples are taken from Schubert (1980). .â•… Müller’s cycle ensues from a series of encounters with friends at a literary salon. Â�Important figures like Achim von Arnim, Wilhelm Hensel, his sister Luise, and many others, decided to create a Singspiel on a very common topic at that epoch: a story about a miller and a miller’s maid. Hensel, for example, was the hunter clad in green, Müller the young miller, and so forth. Later, Müller developed his song cycle on this first material.

 Costantino Maeder

through the countryside, with his lute in his hand. He hurries down to a brook and, captivated by its sound, follows it until he finds a new mill. He falls in love with the miller’s daughter. He tries to impress her, but her response seems tentative. The young apprentice decides to abandon his lute, now that he has found his love. However, the maid prefers a hunter clad in green. The young miller experiences an obsession with the color green, which is linked to death in the text through various metaphors and cultural references. An extravagant death fantasy follows in which the miller sees himself in his own tomb, fantasizing that his beloved will pick the flowers growing on his tomb and then remember him as a faithful soul. The young man, discouraged, drowns himself in the brook. The brook itself sings a lullaby for the drowned miller. Müller, however, conceived his work with a metapoetic, self-critical dimension and a blatant ironic distance. When reading Schubert’s choice of texts and listening for the first time to his songs, many will wonder how Schubert could take the poems so seriously, without noticing the humorous touch. I would argue, however, that this is a first sign that Schubert did not illustrate Müller: by making use of several iconic means and by contemplating what an audience could perceive without being musicologists, the Viennese composer truly interprets and rewrites Müller’s cycle of poems. Hereby, he creates a new, autonomous masterpiece, albeit closely associated to the literary source text.3 The new signification developed by Schubert is not the consequence of several minor textual changes to the original collection of poems, but the result of intermedial interventions on the musical, textual, and dramatic structure which make consistent use of iconical and indexical means.4 The implicit author of the cycle stages a complex work of art, which is based on a collection of poems and combines text and music. It uses another work of art and musical tradition as a type of LEGO bricks. In this sense, Schubert is not a mere composer, but an intermedial artist who is not subjugated to a text, but takes possession of it: he forges a play on the human condition of an artist within the modern world, as will be elucidated below. It will be shown how music, through iconical and indexical means, can Â�create word-like patterns and, next, how a composer can tie musical patterns to Â�content through argumentative strategies (Lo Cascio 1991) and cognitive mechanisms

.â•… Schubert omits Das Mühlenleben (after nr. 6), Erster Schmerz, letzter Schmerz (after. 15) and Blümlein Vergißmein (after nr. 17). The most important change, however, is the omission of the Prologue and the Epilogue. .â•… We define intermediality in a broader sense. Schubert, by choosing and appropriating Müller’s text, adopts a role that goes far beyond musical composition. The conception of Die schöne Müllerin as a song cycle integrates three different “media”: literature, music, and Â�performance. Schubert assumes responsibility of all these three dimensions.



Argumentative, iconic, and indexical structures 

Â� explicable through Grice’s theories on meaning and his maxims of conversation (Grice 1975). The creation of intermedial meaning, resulting from the interaction of text, music, and drama, ensues from a handful of very simple strategies such as Â�repetition and uniqueness, norm and surprise.

2.â•… A  bout signifying iconic and indexical strategies in music, opera and Lied Even though the complete cycle is not so much musical as it is intermedial, it is Â�nonetheless important to understand some typical mechanisms and strategies of Â�western classical music as well as some characteristics and constituents of dramatic texts. Whether music can signify and therefore communicate or whether it Â�stimulates and triggers only emotional reactions and mental images is a much debated Â�question. Music can obviously imitate sounds and therefore refer to the natural world: in our cycle, we notice that when the miller relates what the master says, he chooses a low Â�register; however, when he relates the girl’s words, he prefers a high one (Am Â�Feierabend). It can be that simple and blatant. But usually, it is not as plain and Â�self-explanatory as we might imagine. Classical, autonomous, instrumental music has evolved from opera and oratorio, which were heavily influenced by madrigals. The influence is, roughly delineated, in two ways. On the one hand, music has begun to mimic the cadence and the prosody of spoken language, or better: the cadence of staged speech by actors (cf. for instance Leopold 2003). Many musical figurations replicate, model and formalize the cadence of a question, of bewilderment, of an assertion, and so on. Long intervals within an assertion – a simple ascending and descending curve progressing in small intervals usually renders an assertion – signal distress or any other deep stirring of emotions. Schubert uses prosodic melodic lines when singing characters acknowledge the harsh reality of existence or accept death. In such cases, a character is not passion-driven but appears well-balanced. On the other hand, music has developed a certain range of symbols (in the Peircean sense, or conventional signs according to Saussure) which are Â�culturally determined and based originally on iconic means (mostly onomatopoeia: the Â�imitation of natural sounds, e.g. a storm, singing birds, a galloping horse, or the Â�imitation of human sound registers) and indexical means. Certain patterns Â�(signifiers) stand for precise Â�signifieds, of which many were codified at the end of the sixteenth Â�century, for example in madrigals: the galloping horse rhythm signifies that the Â�singing Â�character is passion driven and refuses reason. These symbols are still productive in today’s Â�Hollywood movies. In Schubert, the flowing water pattern, as we will see later, belongs to this category.

 Costantino Maeder

Both strategies can develop in new forms of iconic devices: for example, the Â� imitation of psychic behavior. The repetition of words and phrases, together with the incapacity of concluding a sentence or a reasoning, as well as the insisted repetition of motifs, melodies, or the usage of strophic forms, reveal that a character has lost his reason (Maeder 2006). The absence of these forms of repetition hints that the singing character is a well-balanced person who sincerely strives to reason and communicate. Often, an apparent representation of natural sounds reveals a diagrammatic Â�character: a descending scale represents trespass or a descent into hell, into misery, whereas an ascending scale is related to redemption, ascension to God and Heaven, and so on. These ‘words’ were contextually triggered, for example through genres: oratorio, cantata, motet or through the setting of an opera. By and by, they became common, musical words. A third important strategy is formed by opposition, that is by the passage from one distinct musical section to another. Differences in rhythm, accompaniment, or any other musical aspect, create expectations and indicate audibly that something is happening. This works usually as an indexical sign.5 In this context, the intermedial

.â•… Let me illustrate this with an example. In 1851, the first Scena e aria from Rigoletto by Giuseppe Verdi to a libretto by Francesco Maria Piave, confronted the audience with a recitative, followed by an aria. This very procedure is significant and determines the reception: at that time, the audience automatically associated recitative to dialogue, communication, and imitation of a real world discourse. In a recitative, characters try to be reasonable, to dialogue, to get information. In this section, the accompaniment does not have an important role of its own but follows cultural expectations: it supports the voices, avoiding covering them. No symmetrical regularity determines the score, which makes information exchange crucial. The aria, on the other hand, was associated with a certain loss of rationality and of reasoning skills: when communication fails or is no longer productive, an aria follows the recitative: this can be an outburst of joy, of hatred, or of any other more complex emotional stirring. Repetition of lexemes, phrases and sentences, melodic lines that cannot be associated with regular speech, the symmetrical construction of melodies as well as the strophic texture, distinguish an aria. The very repetitions of lexemes and musical patterns illustrate iconically that the singer has lost his reasoning skills, to put it bluntly. No wonder that an aria conveys only a minor amount of new information. But in this very sequence, the character that sings the aria does not simply interrupt the dialogue: his aria is characterized by a repetitive, very simple, and monotonous accompaniment. It is a musical calamity, as is his other aria, La donna è mobile. The accompaniment mimics a galloping horse from the beginning to the end. This pattern has been used since the times of Monteverdi and the madrigalists, and it is therefore well known by the audience (and still used in Hollywood productions). It has been associated with characters unable to refrain their passions, resembling a galloping horse that no longer obeys the reins. This rhythm signifies ‘intemperance’, ‘uncontrollability’, ‘passion’ and reveals that the character is ‘instinctdriven’. This aria typifies the most notable character of the play. The preceding recitative



Argumentative, iconic, and indexical structures 

implicit author surfaces: it segments and dramatizes a song and orients the audience.6 In Schubert, we notice that several songs present an audible opposition of two Â�distinct musical parts, for example in Am Feierabend. It is fundamental to study how the implicit author distributes content on musical parts. With this information in mind, we can now proceed with Schubert’s song cycle.

3.â•… Gricean maxims, drama and communication Die schöne Müllerin or The Fair Maid of the Mill is a song cycle based on a narrative construction. Every single song is autonomous, but, when performed within the whole cycle, it is part of a drama with an opening, a plot, and a conclusion. Since it is a drama, and therefore the implicit author dialogues with the audience, Grice’s approach to meaning and communication allows us to understand how Schubert creates new meaning that is not directly encoded in Müller’s Singspiel. Because of the intended competent, but not specialist public,7 and the fact that this cycle is a drama and therefore a communicative act, we can assume that the ‘cooperative principle’ is operative

deals with love affairs, betrayal and fornication: a courtesan tries to dissuade the Duke from continuing an affair with a nobleman’s wife. The Duke interrupts the recitative, and therefore the dialogue, by singing a very dull, simple aria about the liberty of having as many affairs as one wishes. The music (simple, monotonous, repetitive) discloses to the audience that the Duke is dull, simple-minded, unbridled, volatile, excitable, uncontrollable, and, sure enough, not a responsible monarch. The indexical sign, the passage from the recitative to the aria and the structure of the aria itself confirms it:â•›in 1854, writing a recitative followed by a simple aria is very old-fashioned, and, again, connotes the duke negatively and indicates that its function is outdated and no longer desirable. Rigoletto has been for decennia mutilated by political and religious censorship. This shows that the audience was well able to understand the parabolic structure of the libretto that allowed the audience to link the story situated in the past to the criticism of monarchy in the nineteenth century. .â•… Mostly onomatopoeia has allowed the development of musical signs that work as ‘words’ or ‘propositions’. Arias and madrigals often relied on similitudes. A person’s character or emotion was compared to a certain natural phenomenon, easily translatable into musical Â�patterns. In the course of time, the simple musical sound alone transported the very sense of the Â�comparison, just as linguistic signs do. However, music still has severe semantic Â�limitations; it can never express the content of an idea or an announcement in a simple and plain-spoken way, as natural language can, as in ‘The down train leaves Paddington at 10:10am and arrives in Cheltenham at 12:04pm’, but it could now express ‘danger’, ‘soothing calmness’, ‘sadness’, ‘sighs’, and even more complex things. Combined with text, in a play, music can interact with it in various ways, contradicting or highlighting other layers of significance. .â•… A competent public knows, albeit unconsciously or superficially, the necessary artistic Â�conventions and rules, without being necessarily able to verbalize them or having studied

 Costantino Maeder

and governs the intermedial output of the implicit author and the reception of the audience. Grice defines this principle as follows: We might then formulate a rough general principle which participants will be expected (ceteris paribus) to observe, namely: Make your conversational contribution such as is required at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged. (1989:â•›26)

From this principle stem different maxims and submaxims, as for instance ‘be brief ’, ‘be orderly’, ‘avoid unnecessary prolixity’, ‘make your contribution as informative as the current purposes of the exchange requires’, or ‘be relevant’. However, in a communicative act, an enunciator is not always able to follow these maxims, because of social rules or of insufficient language competence. An addressee, however, will nonetheless try to understand the intentions of the enunciator: when an enunciator flouts the maxims ‘avoid prolixity’, ‘avoid ambiguity’, or ‘be relevant’, an addressee, because he assumes that his partner sticks to the cooperative principle, will nevertheless accept the linguistic output as ‘relevant’ and start a series of implicatures8 in order to explain the apparent inconsistencies of the enunciator’s output. Apparent errors and mistakes that flout conversational maxims can be interpreted, for instance, as tentatives of articulating meaning that, for one reason or another, cannot be conveyed according to these very maxims. In a song cycle, we can assume that a composer should try to offer a varied musical setting, and that music follows the linguistic content. The absence of variety, for instance, corresponds to flouting the expectations: the lyrics offer a wide range of possible topics, but musically the same patterns are repeated: this flouts for instance ‘make your contribution as informative as is required (for the current purposes of the exchange)’ or ‘avoid unnecessary prolixity’. The audience, however, according to the cooperative principle will start to develop implicatures and try to understand, why the composer has chosen certain musical strategies and how he links them (> musical strategies) to the lyrics.

4.â•… Another principle: Keep it simple Additionally, I hypothesize a second principle that governs song cycles or operas in the eighteenth and most of the nineteenth century: ‘keep it simple’. In complex objects – as operas, songs, or song cycles obviously are – sense-creating Â�strategies

them. A specialized public, for example, musicologists, can describe explicitly the functioning of music, for instance. .â•… “If the overt, surface meaning of a sentence does not seem to be consistent with the Gricean maxims, and yet the circumstances lead us to think that the speaker is nonetheless obeying the cooperative principle, we tend to look for other meanings that could be implicated by the sentence.” Cf. the entry Gricean Maxims, http://en.wikipedia.org/wiki/Gricean_maxims



Argumentative, iconic, and indexical structures 

have to rely on simple mechanisms, such as for example an audible shift from one type of Â�accompaniment and melody to another. This second principle derives from the fact that a song cycle combines several, apparently incompatible, Â�complex Â�semiotic systems: text, piano-accompaniment, voice handling, and the singer’s Â�performance. Simplicity is, therefore, crucial in order to allow immediate Â�comprehension. Â�Maxims like ‘be brief ’ or ‘make your contribution as informative as the current purposes of the exchange requires’ and its reverse ‘do not make your contribution more informative than required’, for instance, are simple and fundamental means which trigger comprehension and shifts in meaning. The fact that a performance is Â�teleological, since it is linear and tends to move towards an end point, is closely tied to these Â�maxims: the audience cannot determine the duration of the performance or even rewind or reread it at their own pace. For that reason, all the means used for Â�signifying have to be immediately understandable for most of the audience, and not for the happy few, as is possible in the case of a string quartet aimed by definition at an eminently competent listenership.

5.â•… Schubert’s exposition and the flouting of the maxims The exposition of the song cycle corresponds to the first four Lieder. The young miller leaves his former master and wanders along a brook until he finds another mill, and decides to stop there. All four Lieder are characterized by a monotonous Â�accompaniment, without inner tension or opposition. A strophic structure Â�dominates, which again evokes monotony. They arouse the idea of continuity, of flowing, and of eternal movement. The audience notices immediately that this is against the expected rules of composition: no variation, no dynamic Lied, and no diversity in the Â�accompaniment mark this section where repetition dominates. In other words, Schubert is not brief, he is prolix, and he gives more information than required: the lyrics would have allowed a very diversified musical illustration. The audience will make guesses about the Â�reason for this: either Schubert has no project in mind, or the repetition of the same, Â�monotonous and repetitive elements is relevant.

5.1â•… The flowing water pattern: Horizontality and eternity As we have pointed out elsewhere (Maeder et al. 2005), Schubert relies on the Â�rendering of flowing, running water as a recurrent discursive pattern in his Schöne Müllerin. In Wohin, we notice an ostinato in the piano accompaniment that underlies the whole song (Figure 1). This undulating line is a typical ‘word’. It represents running, flowing water. On this accompaniment the voice floats calmly, tending to avoid large intervals since it has comparatively fewer and longer notes, whereas the undulating accompaniment is formed by many short notes.

 Costantino Maeder

Figure 1.╇ Ostinato in the piano accompaniment of Wohin

We can see this ‘word’ in most other songs in this cycle. A competent audience knows that this pattern can be found throughout music history from the madrigals at the end of the Cinquecento (sixteenth century), through the Baroque era until today. Smetana’s Vltava (also known by its German name The Moldau), a symphonic poem first performed in 1875 (Figure 2) is a well-known example: in this piece Â�dedicated to a river, we notice that the accompaniment is formed by many waves and that the theme is formed by a long wave which floats calmly above the shorter and quicker ones.

Figure 2.╇ Smetana’s Vltava (The Moldau)

Two variants of this ‘word’ can be identified (Figure 1 and Figure 3) in Schubert. The first one is the more obvious: an undulating wave is clearly visible, whereas the second displays very short waves. As is the case with Figure 1, there are many earlier examples, i.e. in Gluck (Vertigo’s Aria, from Der Pilger von Mekka). In both



Argumentative, iconic, and indexical structures 

variants, we notice that the accompaniment fills space and occupies a whole tonal section, creating a wall of sound. Very often, musical pieces that refer to water and rivers are characterized by monotony, namely the lack of divergence and differentiation: the flow is continuous, from the beginning to the end of a piece, uninterrupted by Â�contrastive segments. Additionally, we notice strophic, repetitive structures. Another aspect is horizontality, implied in the notion of flowing. A real brook, however, does not produce this type of sound. We hear clusters, shapeless noises. Instead of imitating the sound of water – as modern music is able to do and actually does – the musical word opts for a two-dimensional graph which consists of a cross-section of waves: from real world sound to graphics.

Figure 3.╇ Undulating line evocative of water

But since a song cycle fits an audience who does not necessarily reflect on music, and maybe does not even know about the existence of musical ‘words’, Schubert provides more information than required. For example, in the second Lied, after a piano introduction based on the flowing water pattern, the first line confirms immediately the guesses of the audience (“Ich hört ein Bächlein rauschen”, I heard a brooklet rushing). Otherwise, the insisted repetition of these musical Â�patterns, together with the insisted horizontality and monotony of this piece, and the fact that the text refers to flowing water, encourage and force the audience to associate the pattern with flowing water. The first four songs have an argumentative character as they manipulate the reader by creating a series of parallelisms through repetition. The main Â�statement of the first Lied is confirmed by many arguments, provided by the following three Lieder. We could circumscribe the opinion that the reasoning tries to confirm the following set of arguments: ‘life is a continuous flow’, ‘it is impossible to stop the moment, to freeze time’, ‘only in nature can a human being be himself ’, ‘desire is the driving force of existence’, ‘fulfillment of desire corresponds to dying’, and so on. The first song (Das Wandern) presents four couplets, all with the same undulating pattern and flowing rhythm. The strophic, repetitive structure enhances the idea of a Â�continuous flow. The first four songs present all variations of the Â�flowing

 Costantino Maeder

water �pattern. All songs are marked by one accompaniment, they are not varied and they do not �present �different, contrasting sections, as should be the case. They are �horizontal. The texts link this accompaniment to water, and then to eternal time, expectation, movement �(navigare necesse est) and euphoria. The wanderer has to move, as water does, as even the millstones do. Wandering, as presented in the first four Lieder, implies �discovery, potentiality. Finally, all four songs are written in Major tonalities, also conventionally associated with euphoria. All those features are easily �recognizable and instantly audible.

5.2â•… Norm and transgression: Vertical disruption and death Once this has been stated and proven time and again over four and a half Lieder, the implicit author introduces something new, this time by indexical means: Am Feierabend (After work), the fifth song in the cycle, starts with vertical chords and pauses (Figure 4), in A minor. From horizontality, continuity, and Major tonalities, the music shifts to verticality, discontinuity (pauses), and minor tonalities, a very incisive and audible opposition. However, at first, the music switches back to what has been asserted in the preceding four Lieder. The young miller is extremely happy. If he had thousands of arms he could have moved thousands of millstones, only because he longs to be noticed by the miller’s maid. This is still rendered by the Â�flowing water pattern and in Major. Unfortunately, it is evening. All the Â�journeymen, the master, and the maid are resting. The music changes immediately: vertical chords, interrupted by pauses, dominate now, as exposed at the beginning of the Lied. In this section, the young miller acknowledges that he is weak, that everybody is as strong as he is, implying that he cannot impress his beloved miller’s maid. He then talks about what the master said about their work, and how the maid wishes good night to everybody. The young miller returns to his first, high-spirited mood. Of course, the music illustrates it as well by its reprise of the flowing water pattern. Most Â�important to our considerations is the passage from the first section, which relies on the Â�flowing water pattern, to the second.

Figure 4.╇ Vertical chords and pauses in Am Feierabend



Argumentative, iconic, and indexical structures 

This sequence of vertical chords, together with the pauses, incisively interrupts the horizontal and teleological progression of the preceding accompaniments, even if the passage is marked as ziemlich geschwind (quite fast): the discrepancy with the constant flowing of the preceding section is clearly perceptible. Pauses and vertical chords, in music, as well as a narrow pitch range (for the melody), are often tied to interruption, death, and other figurations: the reaper, in Schubert’s Death and the Maiden, uses a narrow pitch range and vertical chords (Figure 5). The passage from the first to the second section corresponds with the change of tonality: from Major Schubert turns to minor, a tonality that is Â�commonly linked to Â�dysphoria. This very song presents for the first time in the cycle two clearly Â�contrastive Â�sections (A – B – A). The expectations of the audience, built up by the preceding four Â�Lieder and the beginning of the fifth, are upset. The audience will again begin to tie what happens musically to content, this time in a quite dynamic way, and through a ‘backward reading’ (“lettura retrospettiva”), as Eco (1979) would call it.9 The first songs were essentially strophic; they had a rather simple harmonic structure. The accompaniment was a continuous flow. Movement was thematized in many ways in the Â�lyrics: water, flowing brook, wandering, rolling mill stones, and wheels. The narrator is a subject in motion, he moves in an almost natural Â�totality: nature, earth, sky, and woods. The music creates the illusion of eternal, repetitive movement, and it is euphoric, rendered with major tonalities. The fifth song Â�presents an abrupt Â�inversion. It is in minor: the vertical structure is tied to reality and the consciousness of it, as well as to death. From the moment he saw the mill and the maid, the miller had tried to settle down, to be noticed by her. He forgets his first goal, just to wander, to ‘flow’. But coming to a halt implies stasis: forming a family, working, real projects within society. The young miller, symbol of the poet, contemplates to renounce his lute. In Pause [Interlude], the narrator tells us that he has hung it upon the wall, indicating that he can sing no more, that his heart is too full, and that he does not know how to transform his feelings into verse lines, creation, and Â�discovery, closely tied to movement. In the minor parts, the poet is aware of his unfulfilled longing. Reality is negative and cannot be steered, adapted to one’s own wishes. The singer has to acknowledge it, he cannot bend the outer world to his will: his suicide is the consequence.

.â•… Texts often mislead a reader’s expectation at first. When a reader detects an Â�inconsistency between the text and his expectation, he will ‘reread’ the text, in order to reconcile his Â�interpretation with the reality of the text: this act is called ‘retrolecture’.

 Costantino Maeder

Figure 5.╇ Narrow pitch range and vertical chords in Der Tod und das Mädchen

In Trockne Blume (Figure 6), for example, the young miller talks to flowers. He asks them to be put on his grave. This section, in minor, shows how the narrator perceives reality: death figurations mark the accompaniment, but a last sparkle of expectation arises: the music turns from minor to major, the flowing water motive surfaces again and marks the left hand, as the narrator fantasizes that his beloved maid will perhaps pass by and see the flowers and the grave, and will understand



Argumentative, iconic, and indexical structures 

that he was devoted to her. But ‘negativity’ has not been cancelled: the right hand continues the death motive, and the last song will show that the miller has drowned himself.

Figure 6.╇ Trockne Blume

 Costantino Maeder

The following table summarizes these findings. Dichotomies in Die schöne Müllerin Major Flowing water motive Horizontal Motion Teleological Euphoric Interpreting, fantasizing subject Nature Imagination, expectation Discovery, curiosity, potentiality The subject accepts nature as guide and sender Ecstasy, rapture Unity, monotony (of a Lied). In this case dominate constituents of this column dominate. Stable (music)

Minor chords, pauses Vertical Halt Static Dysphoric Interpreting, conscious subject Human society Reality Settlement The subject’s actions have no effect on others Disenchantment A Lied has distinct, contrastive sections: a part where the constituents of the first column dominate, another where the constituents of the second column prevail. Unstable (music)

6.â•… When music challenges words Once the audience has understood the links between both signifiers and their signifieds, Schubert, as intermedial author, can play with their meanings. For instance, he can use the death and reality patterns while the young miller, in the lyrics, seems to be happy, imagines a future, and misinterprets the world. The lyrics alone are monodimensional; together with music, the comprehensive meaning can be far more complex, and ambiguous. The eleventh Lied, Mein! (Mine!) is an example: the miller asks nature (the brook and the birds) to cease their sounds. He is happy; he is convinced that the beloved is his. But, nature does not understand him, as he states at the end of the lyrics. The accompaniment is explicit: the flowing water pattern and its signifieds dominates. Therefore, the miller’s attempt to settle himself is doomed. The situation is more complex in number thirteen. The young miller tells us that his green lute-ribbon ‘fades’ on the wall. The audience knows from the preceding song that the narrator has abandoned lute playing because of his love. The narrator is happy; the song is in B flat major. However, the accompaniment makes extensive use of the death pattern (verticality and pauses mark the accompaniment): the music shows that the wish of the narrator will not be fulfilled, that he is deceiving himself: while in the lyrics he states that the maid loves him, the music clearly tells the opposite.



Argumentative, iconic, and indexical structures 

7.â•… Conclusions The running water motive represents change and the eternal course of time. At the beginning of the cycle, the wanderer is in accordance with it. He is in harmony with nature and feels positively inclined, which the tonalities reflect. He is observing the countryside, he is fantasizing, he is a poet, and he is the one who is acting by interpreting nature. As the whole cycle is seen from his perspective, we must trust him, which is why the dichotomy of tonalities is very important. This cycle ties major tonalities to illusion, and illusions generate positive feelings. The subject sees and interprets, and he does it in very subjectively, in his own way. Novalis, a contemporary, states: […] so erfährt auch der Dichter die Begebenheiten der äußren und innern Welt auf eine sehr verschiedne Weise vom gewöhnlichen Menschen. […] Alle Töne, die die Natur hervorbringt sind rauh – und geistlos – nur der musikalischen Seele dünkt oft das Rauschen des Waldes, das Pfeifen des Windes, der Gesang der Nachtigall, das Plätschern des Bachs melodisch und bedeutsam.  (Novalis 2001:â•›393)10

As long as the narrator wanders through nature, he is happy. He correctly interprets what nature tells him; he understands the signs. But once the wanderer, symbol of the poet,11 sees the mill, a human construct, he halts and tries to stop time, to grasp Â�reality, determined by minor tonalities.12 He is no longer able to understand the signs. He has to accept his human frailty and incompetence when it comes to understanding human society. The last song, however, after the subjects’ narrative program has failed, is sung by the brook, which soothes the suicide. The accompaniment is formed again by an ostinato, typical of running water, and it is also written in major. Â�Apparently, there is still hope. Maybe there is still a place for poetry, for art, for illusion, and for the imagination. In Winterreise ‘A Winter’s Journey’, Schubert offers an analogous Â�construction. But now the wanderer is more and more confronted with reality, in a bitter winter landscape. And the end is in minor, vertical chords dominate, long pauses in the

.â•… “ […] so too the poet experiences the events of the outer and the inner world very Â�differently from the ordinary person […] All sounds produced by nature are rough – and empty of spirit – only the musical soul often finds the rustling of the forest – the whistling of the wind, the song of the nightingale, the babbling of the brook melodious and meaningful” Novalis (1997:â•›71). .â•… The audience knew the context: the poet’s name was Müller, to wit Miller (cf. Youens, 1997). This fact is another trigger that obliges the reader to interpret the cycle Â�metapoetically. .â•… “It would be nearer the mark to call it [Die schöne Müllerin] a statement about the human condition, in particular about the age-old conflict between the emotions of the natural man and the conventions of polite society” (Reed 1978:â•›413).

 Costantino Maeder

accompaniment figuring death, as well as a very monotonous, recitativic Â�treatment of the voice. Schubert’s last cycle no longer fosters illusions: a poet or an artist has no effect on his fellow humans, he has lost its function in society. Schubert presents other strategies with additional iconic features inviting Â�investigation such as, for example, the importance of the pitch or diastematic patterns (this indicates the pitch of notes by their vertical placing on the page). However, our aim was to show, how Schubert, by very simple iconic and indexical devices, which are perfectly audible, was able to create his own, autonomous song cycle.

References Eco, U. 1979. Lector in fabula: La cooperazione interpretativa nei testi narrativi. Milano: Bompiani. Grice, H.P. 1975. Logic and conversation. In Speech Acts, J.L.C. Morgan (ed.), 41–58. New York NY: Academic Press. Grice, H.P. 1989. Studies in the way of words. Cambridge MA: Harvard University Press. Lo Cascio, V. 1991. Grammatica dell’argomentare: Strategie e strutture. Scandicci (Firenze): Nuova Italia. Leopold, S. 2003. Die Anfänge von Oper und die Probleme der Gattung. Journal of SeventeenthCentury Music 9(1). 〈http://www.sscm-jscm.org/jscm/v9/no1/leopold.html〉 Maeder, C. 2006. Metastasio and the melodramma. In A History of Italian Theatre, J. Farrell & P. Puppa (eds), 186–93. Cambridge: CUP. Maeder, C., Fischer, O. & Herlofsky, W.J 2005. Introduction. In Outside-In – Inside-Out. Â�[Iconicity in Language and Literature 4], C. Maeder, O. Fischer & W.J. Herlofsky (eds.), 1–12. Amsterdam: John Benjamins. Müller, W. & Schubert, F. 2001. Die schöne Müllerin/Die Winterreise. Reclam, Ditzingen. Novalis. 2001. Werke. München: C.H. Beck. Reed, J. 1978. ‘Die schöne Mullerin [sic]’ reconsidered. Music and Letters 59(4): 411–419. Schubert, F. 1980. Die schöne Müllerin: Opus 25, D 79: Hohe Stimme (Original): High voice (Original). Mainz: Schott. Youens, S. 1997. Schubert, Müller, and ‘Die schöne Müllerin’. Cambridge: CUP.

John Irving’s A Widow for One Year and Tod Williams’ The Door in the Floor as ‘(mult-)i-conic’ works of art Christine Schwanecke University of Heidelberg

This article explores the various layers of iconicity in John Irving’s novel A Widow for One Year, which abounds in iconic images, diagrams and metaphors, and Todd Williams’ movie adaptation, The Door in the Floor. Intra-medial and intermedial forms of iconicity are brought into focus, as they weave the complex web of cross-, self- and meta-references out of which the novel’s semantic density emerges, building up the illusion of an enhanced ‘visuality’. The media transfer of the novel to film is also examined: to what extent has the novel’s predominant feature, its ‘(mult-)i-conicity’, been transferred to Williams’ movie? Beyond aesthetic questions, medial idiosyncrasies have required structural and material changes and Williams had to abandon some iconic signs, modify others, or establish new forms of iconicity, i.e. audio-visual ones. Despite the differences, the movie’s web of semantic and structural correspondences has become as dense as the one in Irving’s text.

1.â•… Introduction This article deals with the work of a novelist who has been critically acclaimed for the “stunningly visual” prose (Mcalpin 2005) with which he brings fiction to life (see Argendeli 1998) – or iconically refers to reality.1 Now, in a context where manifold forms and functions of iconicity are explored, it might be particularly fruitful to examine the work of such a novelist – John Irving.

.â•… As Winfried Nöth briefly outlines, for centuries, the iconic quality of literature has been a topic in aesthetic and philosophic discourse (2001:╛╛22–24). John Irving’s novels appear to be exemplary artifacts in this context, since the author’s literary style, his detailed realism, has not only caused especially plausible analogies to the (or a possible) real world – as the critics’ Â�opinions attest to – but has also iconically referred to the sphere of literature itself, for example, to aesthetic notions or traditions and to contemporary processes of literary Â�production and marketing.

 Christine Schwanecke

Irving’s ninth, highly appraised novel A Widow for One Year (1998) seems to be especially suitable: it features the main categories of iconicity in various forms and on all levels of the text. Iconic images, diagrams and metaphors occur very frequently, not only on the story-level, but also on the discourse-level, for example in paratextual elements, such as chapter headings, or in Irving’s ‘visual’ style. They establish both Â�intra-medial and intermedial references, which, on the one hand, help to enhance the story’s immediacy and, on the other, account for the novel’s semantic and formal density. Therefore, this article aims at exploring the multi-layered dimensions of iconicity in A Widow for One Year – its ‘(mult-)i-conicity’.2 By way of introduction, the novel’s most important plot lines and features will be detailed, followed by an outline of the theoretical framework underlying this discussion. Then some of the most important iconic forms that are displayed in the text will be paradigmatically examined and their functions for the narrative discussed. In so doing, I will distinguish between the effects iconicity triggers with regard to both intra-mediality and intermediality. In the chapter that follows, the film adaptation of the novel, Tod Williams’ The Door in the Floor, will be taken into account. By comparing the iconic forms embedded in the verbal narrative to the way they materialize in the movie of 2004, questions will be raised which concern the qualities of iconic signs in different media: for example, to what extent does the ‘intermedial transfer’ influence the characteristics of the iconic sign and its functions? To conclude, the most important findings will be briefly summarized.

2.â•… Short synopsis of A Widow for One Year The novel’s multiple plot lines, stories within stories, and extensive excursions all Â�centre on a character named Ruth Cole, who is the widow of the title. A Widow for One Year is divided into three sections, or sub-books, presenting Ruth at the ages of four, thirty-six and forty-one. In each part, pivotal episodes of the main character’s life are described. The first part of the book deals with the events of the summer of 1958, after which Ruth’s mother, Marion, abandons her daughter. She and her husband Ted Cole, a Â�successful writer of children’s books and an excessive womanizer, have grown apart, ever since their teenage sons, Thomas and Timothy, died in a car crash. Although

.â•… I would like to thank Gabriele Rippl and Christina Ljungberg, who were so kind as to invite me to their post-conference discussion of iconicity and to comment on different Â�versions of this article. I would also like to thank my supervisor Vera Nünning for her advice on my conference paper and Christoph Schöneich for his help Â�concerning the article version.



John Irving and Tod Williams 

they tried to mend their relationship by having Ruth as a substitution for their dead boys, their lives have never returned to normal: neither has grief-stricken Marion ever Â�managed to relate to her husband again nor has she established a loving, maternal relationship to Ruth. In the summer of 1958, Ted sets up an affair between his beautiful wife and his assistant, 16-year-old prep-school boy Eddie O’Hare. This has far-reaching Â�consequences which Ted never would have dreamed of: by the end of the summer, Marion has found the strength and will to leave her family. With her she takes the photographs of her sons, which, up until then, have covered every inch of wallpaper in their house. It is only in the last sub-book, 37 years later, that Marion returns to her daughter and Eddie. To overcome the loss of a mother and a lover, both of them have become novelists, as the second sub-book reveals. The same holds true for Marion, who, likewise, deals in her fiction with the loss of her sons. When Ruth finally meets her mother again, she has gone through a lot. She has suffered both physical abuse and emotional distress caused by a series of bad boyfriends; she has had to witness a murder, and to deal with the suicide of her father as well as the loss of her husband, Allen, leaving her to take care of their child on her own. At the end of A Widow for One Year – upon the return of Ruth’s mother on Thanksgiving in 1995 – the novel’s lose ends are tied together: having been a widow for one year, 41-year-old Ruth is happily married again, this time, to the love of her life, Harry Hoekstra, a retired police officer and ardent reader. At last, Ruth is able to overcome the trauma of the disruption of her family she has suffered early in her life. She now enjoys the bliss of family life, her husband, son, best friends, among them Eddie O’Hare, and her mother gathering around her.

3.â•… Theoretical preliminaries It is quite striking that all of the main characters are novelists or are at least very keen on literature. This hints already at the fact that there is a high degree of self-reflexivity in the novel. Yet, A Widow for One Year is not only focussed on intra-medial Â�questions. It is as much about images as it is about stories – which points to another of its most essential features, namely, to the intermediality of its composition: Irving’s novel, being a verbal medium, refers not only to a variety of issues concerning its own medial Â�tradition; it also contacts various visual media and represents both of these spheres with its own, innate strategies. Iconicity serves as one of those strategies; it is actually the most important and predominant one. The definition of iconicity that will be adhered to in this context is the Â�pragmatist semiotic one, according to which the iconic relationship between a sign and the object it represents is founded on the concepts of ‘similarity’ and ‘analogy’ (see Linke,

 Christine Schwanecke

Â� Nussbaumer & Portmann 2001:â•›19, 21–22; Nöth 2000:â•›193).3 Accordingly, this Â�discussion’s findings will be based on Peirce’s subdivision of the iconic sign. Firstly, there is the class of ‘iconic images’, which entail aural, oral, visual and Â�tactile icons, and feature an analogy to their object of reference which is perceptible at once. The second category is called ‘iconic diagrams’ and their object relation is either Â�structural or relational. Thirdly, there are ‘iconic metaphors’, which relate to a form of mediated iconicity, that is, the idea “conveyed by a sign and the idea of its object are mediated by a third idea, the tertium comparationis between the tenor and the vehicle of the metaphor” (Nöth 2001:â•›21).4 As the various forms of iconicity in A Widow for One Year are intertwined in a complex manner, it is necessary to distinguish at least two spheres in which iconic signs occur in the narrative. On the one hand, there are icons which self-referentially and meta-referentially centre on the literary medium itself.5 On the other, there are instances in which iconic signs further the crossing of medial boundaries between the literary medium and other, predominately visual media. Using Irina Rajewsky’s terminology, the former sphere will be termed “intra-medial”, the latter “intermedial” (2002:â•›11–15). Finally, it must be noted that iconicity is a transdiciplinary and transmedial Â�phenomenon, that is, a phenomenon which is neither bound to specific media nor to certain academic or scientific disciplines. Therefore, it seems especially well suited for an application to contexts in which two or more media are involved: it may provide an excellent tool for examining both A Widow for One Year and its film adaptation.

.â•… I will refrain from more detailed categorisations, such as the Peircean differentiation between ‘general icon’ and ‘hypoicon’ (Nöth 2000:â•›193–194; Nöth 2001:â•›19), as I will always refer to signs which are materially actualized in the novel or film. These ‘hypoicons’ are Â�basically called ‘icons’ or ‘iconic signs’. .â•… Compare also Olga Fischer’s and Max Nänny’s conceptual framework, which focuses on the distinction between iconic images and iconic diagrams (1999: xxi–xxii). They understand metaphors as instances of diagrams (semantic ones) and, therefore, subsume them under the second category. The question of whether metaphors are on one level with diagrams and icons or whether they are a subcategory of diagrams cannot and does not need to be dealt with here. .â•… The term ‘self-reference’ refers to incidents in which iconic relations are solely established within the novel, whereas the term ‘metareference’ applies to a special kind of self-reference; a kind of self-reference which establishes a secondary reference to texts and literature, viewing them from an outside or meta-perspective. ‘Metareference’ might relate to issues that concern to a greater or lesser extent the mediality of novels and related issues. For a detailed account of the difference between self- and meta-reference see Wolf (2009:â•›17–25).



John Irving and Tod Williams 

4.â•… ‘(Mult-)I-conicity’ in John Irving’s A Widow for One Year In what follows, the novel’s complex web of iconic references – its ‘multiconic’ layers – shall be analysed. To start with, the area of intra-medial iconicity will be illuminated. Afterwards, the sphere of intermedial analogies will be explored.

4.1â•… I conicity as a means to establish a multilayered web of intra-medial references As mentioned before, the intra-medial forms of iconicity in A Widow for One Year are employed self-referentially in order to expand the narrative’s semantic web of cross references and meta-referentially in order to raise questions that relate to literature in general and to the process of fiction making in particular. Irving, for example, Â�frequently makes use of the stylistic device of mise en abyme. This device is commonly defined as a form of literary recurrence or similarity and is, therefore, self-referential on a level subordinate to its diegetic frame – in short: on a ‘hypodiegetic’ level. The mise en abyme thereby ‘mirrors’ at least one element on the superordinate, that is, diegetic level (Wolf 2004:â•›461). In A Widow for One Year, these self-referential forms materialize as stories within the story of Ruth and enrich the novel with semantic possibilities and a net of Â�analogies. The stories on the hypodiegetic level hereby serve as iconic diagrams that proleptically or analeptically parallel important facts on the diegetic level, the level on which Ruth’s story is presented. The contents of fictional books, which are either paraphrased or sometimes even rendered in their entirety, help to reflect past events, to echo present circumstances and to foreshadow events that are yet to come. Ted’s children’s book “The Door in the Floor”,6 for example, is about a foetus, a boy who does not know whether he wants to be born and about his mother, who does not know if she wants her baby to be born either. This is due to the fact that they live in a world in which terrible things can happen: there is a door in the floor, behind which the little boy’s brothers have once disappeared and behind which the most awful things are said to lurk. The story within a story, which is rendered in its entirety as mise en abyme (Irving [1998] 2001:╛╛47–49), mirrors the fact that Ruth was half-heartedly conceived as a Â�compensation for her dead brothers, reflects the loss of the selfsame boys, and, at the same time, reinforces the poor relationship between Marion and her daughter. The iconic link between the story on the hypodiegetic level and the situation on the diegetic level is even made explicit by the authorial voice. It hints at the fact that Eddie, reading .â•… Every time I talk about the fictional children’s book I have used inverted commas (“The Door…”) so as to make clear that I am not talking about the actual film with the same title.

 Christine Schwanecke

the story – unlike the reader of A Widow for One Year –, does not realize “that, in the real story [i.e. the story on the diegetic level], the little boy was a little girl. Her name was Ruth […]. There was another kind of door in the floor […]” (ibid.: 49; emphasis in original). With this mise en abyme, Irving employs an iconic form which, as a fictional book, parallels the factual first book or section of A Widow for One Year by echoing its topics and, in so doing, emphasizes the dread of the life the Cole family leads. Yet, an even more important, more complex example of diagrammatic self-Â� reference in A Widow for One Year must be mentioned here. It concerns the very title of the novel, which, as a motif, consistently recurs throughout the novel in Â�various forms and on different textual levels. It is only in the second part of the novel that the paratextual “Widow” from the title emerges in the narrative. Ruth’s short story and, at the same time, first chapter of one of her books, which is entitled “The Red and Blue Air Mattress”, begins with the words: “She’d been a widow for one year […]” (ibid.: 240). This mise en abyme foreshadows not only events to come but also characters to be introduced on the diegetic level: one of Ruth’s readers, a widow, feels insulted by the way Ruth treats widowhood in “The Red and Blue Air Mattress”. This is why she writes an anonymous letter to Ruth trying to intimidate her by telling her that she will pray every day for Ruth to become as unhappy as her. She hopes that Ruth will find a husband, that she will learn to love him, and that she will eventually lose him; in short, that she will be a widow for the rest of her life. This sequence of events, which has been thematically introduced by the paratext, that is, the title of the novel, and then diagrammatically reintroduced as the topic of a mise en abyme, maps all the events that are soon to come: Ruth will marry her editor Allan, after initial difficulties she will find herself truly loving and finally losing him. She will become a widow – if not “for the rest of [her] life” (ibid.: 343), at least, as the title indicates, for one year. These observations show that a complex web of self-referential analogies, like the one displayed in A Widow for One Year, certainly intensifies the literary experience in the process of reception: as the recipient is constantly faced with ever new variations on the motif of ‘widowhood’ or the ‘broken family’, the semantic density of the novel is enhanced. What seems to be a challenge at first turns out to be a highly stimulating reading experience at last. Admittedly, readers do have to constantly re-evaluate the knowledge they have gained so far, especially in the light of topics being presented from incessantly changing perspectives or in different contexts. Yet, in so doing, they gain the possibility of becoming intellectually involved, of speculating, and of Â�establishing possible connections – in short, of actively exploring the dense web of semantic layers that is caused by means of iconic self-reference.7

.╅ This has also been pointed out by Ljungberg (2001). For a complementary approach, compare Ljungberg (2010). Whereas the complexity of iconic semantic strategies in �narratives



John Irving and Tod Williams 

These processes, however, become even more complex as Irving does not only employ purely self- but also metareferential iconicity, making literature and related topics issues in their own right: in discussions between Ruth and other characters, for example, the question arises over and over again to what extent novelists draw on their own biographies in order to create fiction (e.g. Ruth’s best friend’s remark: “You and what’s ‘real’ and what’s ‘invented,’ […] [your] favorite subject…”; ibid.: 274). Although these discussions carry a high metareferential potential (because they draw the reader’s attention from the diegetic plot to a meta-level on which he or she can diagrammatically map these questions, which are related to fiction per se, and consider them for him- or herself), it remains questionable to what extent they are iconic; they may be only weakly so, as it depends on the recipient whether a (and what kind of) diagrammatic mapping is actually carried out. There are, however, cases of metareferential iconicity that are more evident: one of those, for example, relates to matters of ‘story-telling’. In A Widow for One Year, the characters’ style of writing, which is ‘shown’ on the hypodiegetic level, is often paralleled by explicit meta-narrative remarks on the diegetic level. Above this, Irving’s actual style of writing seems to echo the idiosyncratic manner of presenting stories. This complex of metareferential iconicity serves to enhance both the literary illusion and the seemingly ‘realistic’ quality of the plot. In Irving’s ninth novel, the reader is confronted with uncountable ‘meta-narrative’ remarks, that is with explicit comments on story-telling that are interspersed in the novel itself (Nünning 2001:â•›126). At one instance, the focalizer Ruth is, in the mode of ‘showing’, making up a story that would appear realistic, or true to a potential Â�recipient, whereupon the authorial voice, in brackets, comments on her efforts in the mode of ‘telling’. At first, the reader bears witness to Ruth’s thoughts: The story should be that she was struck by her opponent’s elbow. In order for that to happen, Ruth would have had to have been standing too close to her opponent – crowding him from behind. In such a circumstance, Ruth’s imaginary opponent would have to have been a left-hander – in order to hit her in her right eye. (ibid.: 334; emphasis in original)

Right at the end of the story, which is being made up by Ruth, the recipient is Â�confronted with a remark by the authorial voice, which points to her way of Â�rendering tales: “(To tell a believable story, the novelist knew, you just have to get the details right.)” (ibid.). Paralleling Ruth’s style by referring to its essential quality, namely its Â�focussing on details, the authorial voice draws the reader’s attention not only to “the process is dealt with here, Christina Ljungberg’s article relates to the effectiveness of ‘intermedial’ Â�narrative strategies.

 Christine Schwanecke

of fiction-making”, as Campbell states (see 1998:â•›162), but also to Irving’s actual style of writing, which is full of comprehensive diagrams. These establish a Â�‘possible world’ in the recipient’s mind, which, in turn, adds to the ‘visual’ character of his prose. Obviously, Irving – just like his fictional novelists Ruth and Â�Ted – stands in the Â�tradition of what has been called Beschreibungsliteratur: a kind of aesthetics that avoids literary modes such as dramatic action or epic narration and favours Â�description instead. From antiquity onwards, this style has been regarded as “a mode of speaking that [is] able to bring absent objects before the [recipient’s] Â�mental eye” (Rippl 2000:â•›522–523). Irving has his characters obviously imitate this style, which is also his own. In the doubling of the descriptive literary mode, which abounds with iconically mapped details, the author manages to enhance both the plot’s plausibility and its Â�credibility. Furthermore, he enables the reader to truly ‘see’, that is to diagrammatically map Â�cognitive images of the events narrated,8 thereby enhancing the literary illusion and sustaining it throughout the narrative.

4.2â•… I conicity as a means to establish a multilayered web of intermedial references and as a means to affect their function However, the interesting iconic web of A Widow for One Year, which causes the novel’s ‘visuality’ and its demanding as well as stimulating nature, does not solely depend on intra-mediality. Rather, intermedial references are just as important: snapshots, Â�portraits, drawings, postcards, Polaroids, picture books – John Irving incorporates a large variety of references to visual media into his novel, some of which are iconic. Intermedial references are, according to Rajewsky, references between a medial product and another medium. The medium which refers to or represents the other medium does so with its own, innate means (Rajewsky 2002:â•›16–18). That is to say, literary language establishes links to visual media – to their concrete products, and once again, also to processes of their production as well as their reception. The relation between the two media can be called iconic, when, as in our case, literature represents another medium mimetically. This way of representation may result in “a seeming

.â•… In this context, it might be interesting to consider another text passage in which the mapping of the ‘full picture’ depends on a detail as small as a shoe. Ted tells Eddie about the fatal accident of his boys, in which Timothy’s leg got severed. His leg and its attached shoe remained in the car after the corpse had been removed. Since Ted does not provide Eddie with a detailed description of the shoe, the 16-year-old, to his benefit, fails to map a mental picture of this scene; he fails to “see” (Irving [1998] 2001:â•›177). The full dread of the scene, however, finally closes in on Eddie (and the reader) when Ted, in a completely different context, Â�unexpectedly provides even more details than the boy can bear: “‘About the shoe… […] It was a basketball shoe[;] […] Timmy called them his high-tops’” (ibid.: 187).



John Irving and Tod Williams 

Â� ‘imitation’ or ‘dramatization’ of the non-dominant medium [here, any visual media], its quality, structure or typical effects, in the mode of implicit ‘showing’” (Wolf 1999:â•›45). In Irving’s ninth novel, this intermedial mode of implicit showing appears in all shades of iconicity: imagic, diagrammatic, and metaphorical. In A Widow for One Year, intermedial iconic diagrams with imagic qualities are constructed by typographic means. When, in the second sub-book, Ruth travels through Europe in order to promote her newest novels, the picture postcards she sends to her family and friends at home play an important role as they, once again, add to the novel’s persuasion strategy. The fictional author’s postcards, rendered as iconic diagrams, are visually set apart from the general typeface, as they are enclosed by ‘white space’. Being indented and framed by blank lines above and below, they appear as almost independent formal and semantic units. In the following, a text passage is given to illustrate the relative independence of the literary postcards from their surrounding text:

Hannah may have been born without a biological clock, but she certainly responds to all the other things her body thinks it wants – if not a baby. [In a postcard to Hannah, which was a display of sausages in Munich’s Viktualienmarkt.] I FORGIVE YOU, BUT YOU FORGIVE YOURSELF A LITTLE TOO EASILY. YOU ALWAYS HAVE.  LOVE,  RUTH The drive from Munich to Stuttgart; the pronunciation of Schwäbische Alb; the farmland with red and blue and green cabbages. In Stuttgart, the hotel is on the Schillerstrasse – a modern hotel with lots of glass.  (Irving [1998] 2001:â•›346–347; typographical layout as in original)

Logically and typographically obviously disconnected from the surrounding text, the iconic sign diagrammatically maps a picture postcard from Munich with literary means. Its typeset refers to the two sides the medium typically consists of: the iconic image’s upper half refers to the postcard’s pictorial side, as the square brackets signify the frame of a postcard motive; the text inside verbally conveys its content, namely “sausages in Munich’s Viktualienmarkt”. The lower section of the icon refers to the side on which the sender’s message is written; the capital letters probably indicate a handwritten message or one written in capitals. The form of a visually independent textual unit conveys the impression of an actual manifestation of a postcard, which

 Christine Schwanecke

is existent on the diegetic level, that is, in the fictional world. With regard to this, the iconic diagram reveals its imagic quality. As signs like these also refer to a medium that characteristically appeals to human vision, they add substantially to the vividness of Irving’s narrative. However, it is not only the single units which imitate picture postcards that can be seen as intermedial iconic diagrams but also their entirety (one postcard succeeding the other): they diagrammatically map Ruth’s itinerary. This, once again, adds a certain air of ‘authenticity’ to her story: Munich, Hamburg, Cologne, Frankfurt – these are places the fictional novelist must have ‘really’ visited, one after the other; after all, the verbally conveyed ‘picture postcards’ appear to bear witness to this. Besides imitating ‘products’ of visual media, Irving’s narrative also strives to Â�imitate the ‘process of receiving’ those. Thus, the visual process of examining Â�pictures, which may be displayed side by side on a wall, is imitated by verbal means. In some paragraphs, the author accumulates ekphrases, or verbal depictions of visual Â�representations (Wagner 2004:â•›137), and ties them together by means of Â�punctuation. Thus, the process of focalizer Eddie looking at one photograph of Thomas and Â�Timothy after the other is diagrammatically imitated. Eddie perceives Ruth’s brothers “[w]restling in a pile of boys in a dormitory butt room […], clowning on crutches, posing with snow shovels, or playing cards” (Irving [1998] 2001:â•›42). And he Â�studies the pictures of the Cole family before the fatal car accident, when they are on “a holiday in Tortola, and one in Anguilla; [at] a summer picnic at the pond in New Hampshire; and Thomas and Timothy, when they were both younger than Ruth, in a bathtub together […]” (ibid.: 72). Each ekphrasis consists of nothing but a few words, catching the most important feature of the fictional visual objects it refers to. These minimalistic bits of information, these sentence fragments might parallel the pieces of information a recipient of real pictures would be able to absorb, on being faced with a whole wall of pictures, that is, with too many details to take in at once. Separated by commas and colons only, the individual images Eddie looks at are listed in sequential order. On reading this list, the reader has to follow Eddie’s gaze, moving naturally and smoothly from one picture to the other. As the reader and the focalizer’s gazes appear to have merged, the perception process of the diegetic pictures is left to be seemingly ‘unmediated’ and, therefore, highly immediate. What is more, the fleetingly presented, yet highlighted iconic diagrams, which are linked so closely to each other, give a vivid impression of what Barthes called the noema of photography, the ‘this-is-the-way-it-was’ (cf. Barthes [1980] 1989:â•›87). As suddenly as Eddie, the reader realizes what the pictures of the Coles reveal: “They’d been happy!” (Irving [1998] 2001:â•›42; emphasis in original) – an astonishing fact, given the desolateness of the family the reader has witnessed so far. Beside these examples, in which iconicity serves as a means to establish intermediality, there are also instances of intermediality which are, indeed, not caused by, but,



John Irving and Tod Williams 

nevertheless, related to iconicity. The iconic images and diagrams examined above are attempts of the literacy medium to establish analogies to another, a visual medium by imitating characteristics traditionally associated with it. Whereas these endeavours relate to the literary mode of ‘showing’, the phenomenon that is considered now belongs to the realm of ‘telling’, in which iconicity, in principle, does play a less essential role. As indicated before, A Widow for One Year features manifold examples of intermedial references to pictures, drawings, photographs. By way of ekphrasis these fictional medial products are often narratively rendered. Images are an issue – props on the diegetic level; their medial characteristics, however, are not imitated. Yet, these ekphrases may serve as iconic metaphors by establishing analogies between themselves and the objects, mainly fictional characters they represent. Thus, an ekphrasis may be used as a means of implicit characterization by iconically paralleling and, in so doing, intensifying what has been said about a character so far. For example, the ekphrases of the pornographic nudes that Ted likes to draw of his girlfriends serve as iconic metaphors echoing and reinforcing the descriptions of the women’s physical and mental state given at other instances. When Eddie happens to see one of Ted’s lovers, Mrs Vaughn, towards the end of their affair, she appears to be broken and degraded. The sixteen-year-old describes her clumsy movements, her physical appearance and its implications: “And when Mrs Vaughn left Ted’s workroom, she staggered to her car with a carelessness that implied her pride was irretrievable; it was as if the particular pose of the day had defeated her” (ibid.: 79). The impression of Mrs. Vaughn which the reader gains, by being presented with numerous descriptions like this, is paralleled and even reinforced by ekphrastical renderings of Ted’s ink drawings: […] the later nudes were as unconcealed as the frankest photographs of a cadaver. Her arms hung loosely at her sides, as if her shoulders had been savagely dislocated in a violent fall. Her exposed and unsupported breasts drooped; the nipple of one breast seemed larger and darker and more down-pointed […]. Her vagina was gaping and slack. (ibid.: 100; emphasis in original)

This ekphrasis serves as an iconic metaphor echoing Mrs Vaughn’s state of Â�degradation which was described earlier on. In the drawing, Mrs Vaughn is reduced to her sexuality, appearing limp, lifeless and even in a state of decay. The general Â�vocabulary used is one of violence (“savagely”, “violence”) and distortion (“dislocated”, “[one] Â�nipple […] seemed larger and darker”). By emphasizing the vertical lines in the fictional drawing, which are pointing downwards (“hung loosely”, “drooped”, Â�“down-pointed”, “slack”), the ekphrasis’ vocabulary verbalizes Mrs. Vaughn’s metaphorical defeat and fall caused by this relationship with a man who has exploited her body. Hence, in A Widow for One Year, an intermedial device such as ‘ekphrasis’ – albeit not being caused by iconicity – still may fulfil iconic, metaphorical functions. This

 Christine Schwanecke

might allow the author to draw even more powerful images of his characters, which, in turn, provide the reader with a large range of perspectives, illuminating a fictional character’s personality both in detail and comprehensively.

5.â•… I conicity in medial transfer: From A Widow for One Year to The Door in the Floor Finally, the medial transfer of A Widow for One Year shall be taken into account. Since iconicity is a transmedial phenomenon and, as such, not specifically bound to one medial source (see Rajewsky 2002:â•›12–13), it appears especially interesting to look at it from different medial perspectives – in particular, when a source medium is involved that abounds in iconicity. It might be rewarding to examine to what extent an iconic sign is changed in being adapted for an audio-visual medium from a novel. In the case of A Widow for One Year and its screen version The Door in the Floor,9 the actualized icons (with their respectively imagic, diagrammmatic and metaphorical characteristics), which are so important to both the novel’s form and content, seem to undergo a change in form and/or quality when transferred from their medial source to the target medium. By comparing some of the verbally conveyed iconic forms to the way they materialize in their movie adaptation, this assumption shall be tested.

5.1â•… Medial transfer as ‘iconic process’ However, another important issue concerning the basic nature of any medial Â�transfer has to be addressed first. ‘Medial transfer’ seems to be a process in which iconicity plays an important role. It results, after all, in the emergence of an iconic product. The target medium generally imitates at least one component of the source Â�artefact – be it an idiosyncratic topic, parts of its plot, or another characteristic feature. Hence, an iconic diagram is produced that to a greater or lesser extent parallels its original. This fact, however, must not be mixed with the ‘quality of adaptation debate’ or the Â�‘fidelity’ Â�discussion, which, in the discourse of Â�adaptation, have been strongly Â�contested, Â�anyway.10 Quite on the contrary: the fact mentioned

.â•… The screenplay of The Door in the Floor was written by Tod Williams, who also directed the movie. Both the screenplay and the film adaptation were established, however, in close collaboration with John Irving. .â•… For example, McFarlane discusses this topic, arguing that the “insistence on fidelity has led to a suppression of potentially more rewarding approaches to the phenomenon of Â�adaptation” (1996:â•›10).



John Irving and Tod Williams 

above serves to raise questions which are more fruitful to analysis, such as ‘In what respect has the film Â�diagrammatically mapped the novel?’, ‘Where has the film departed from its original?’, and ‘Why is that? – because of medial necessities or for aesthetic reasons?’11 In the case of The Door in the Floor, the author and the screenwriter decided – for aesthetical reasons – to refrain from mapping the whole of the novel’s story as presented in all three sub-books; instead, they tried to adhere to the novel’s most important topics and its general air (cf. Irving [2004] 2005; Williams [2004] 2005). Therefore, certain adjustments had to be made. Not wanting to drift off to the genre of crime12 but to stay true to the novel’s most important topics of trauma, loss, loneliness and self-discovery, the screenwriter, Tod Williams, has focused on the first part of the novel – which, in Irving’s opinion, best captures the whole novel’s essence (Irving [2004] 2005). Moreover, he changed certain scenes and even added new ones (ibid.), thus establishing selected analogies to the original artefact to his aesthetical Â�liking and, in so doing, imitating the novel’s essential mood. Williams has thereby Â�created an iconic representation that, while focusing on the first sub-book, captures the novel’s basic topics and quality.

5.2â•… Adapting the novel for the screen Having briefly touched upon ‘medial transfer’ as a process involving iconicity, I will now focus on this process’ outcome. As regards iconicity, first and foremost, three general trends can be noted: firstly, the translation of certain verbally evoked iconic signs into the audio-visual medium is sometimes not possible. Secondly, if it is, the respective iconic signs change in their materiality and/or in their quality. Thirdly, in order to establish an artefact that equals the novel in its atmospheric density, which is generated through a complex web of semantic layers, the audio-visual adaptation has to draw on other forms of iconicity – forms that can be created with a movie’s inherent means and, in their effect, equal those of the novel. Concerning the first general trend, we will look at the following example: the long passage of time which A Widow for One Year covers is iconically rendered by its sheer

.â•… The last question roughly parallels McFarlane’s distinction between (a) “those Â�elements of the original novel which are transferrable because [they are] not tied to one or other semiotic system” and (b) “those which involve intricate processes of adaptation because their effects are closely tied to the semiotic system in which they are manifested” (ibid.: 20). .â•… For example, a focus on the second and third book of A Widow for One Year, which had been proposed, would have resulted, as Irving speculates, in the movie becoming a thriller ([2004] 2005).

 Christine Schwanecke

size (cf. Nänny 200).13 Three books encompass thirty-seven years of Ruth’s life. This mimesis of duration, according to Irving, cannot be done in a movie convincingly or without any artistic losses: Some things […] work in novels splendidly well – the passage of time is one of them, for example. Most of my novels entail a considerable passage of time, as does A Widow for One Year. Passage of time isn’t great in movies. Movies don’t do passage of time exceedingly well. Novels do. (Irving [2004] 2005; Irving 2000:â•›131)

Irving hints at the fact that one would have to reduce Ruth’s childhood to a series of very short flashbacks in order to render a time span as great as nearly forty years (cf. Irving [2004] 2005). Thereby important characters, such as Marion, and topics, such as the relationship between Eddie and Marion and respectively, Ted and Marion as well as their shared history of trauma, would have to be neglected. Furthermore, the ageing of the characters would have posed a problem. Either an actor would have to be made old or one would have to find two actors, a younger and an older one, to embody one and the same character at different times of his or her life. Irving dismisses all of these strategies as dissatisfying (cf. ibid.).14 Thus, in adapting A Widow for one Year, a compromise was made: the diagrammatic mapping of duration was given up. In consequence only one third of the novel, its first part, was transferred to the audio-visual medium. Although this caused a shift of focus, this strategy was accepted because – and this was the author’s and director’s keenest wish – the novel’s tone could be “kept” (Williams [2004] 2005). Yet, there are even more ‘losses’: iconic diagrams, such as the one verbally Â�imitating the ‘process of receiving’ pictures (see 2.2), are, again for aesthetic reasons, in most cases annihilated. The visual process of a person looking at pictures is no longer paralleled by verbal means; instead, the camera presents them at one instance, one by one as close-ups. Thus, the audience gets to directly look at the photographs and the need for its gaze to be diagrammatically mapped has ceased to exist. This instance of iconic loss caused by an aesthetic decision rather than by restrictions the audio-visual medium presupposes can be seen as a typical case of a signifier coming to equal its signified and, consequently, eliminating itself, as Christina Ljungberg has shown in another context (Ljungberg 2005:â•›143–144).

.â•… Certainly, there are novels of six hundred pages dealing with the events of a single day, such as Joyce’s Ulysses (1992). In this case, the relationship between the time span dealt with and the relative number of pages is just not an iconic one. .â•… It has to be emphasized, though, that the iconic mapping of duration in this case was given up for aesthetical reasons and not because of medial necessities: a movie could very well iconically render the duration of a time span as long as forty years (although financial aspects or convention might forbid a film maker to produce a movie of, for example, five hours length).



John Irving and Tod Williams 

The account of ‘lost’ icons may give the impression that the movie lags behind the novel with regard to its density. This is not true, however: the medial transfer of an iconic sign does not inevitably result in its total loss; certain icons persist. Yet, as the means of iconic presentation differ from one medium to another, some signs change in their materiality and, consequently, in their character.15 So does the mise en abyme of Ted’s picture book “The Door in the Floor”. Being presented as an intra-medial iconic diagram in A Widow for One Year, it appears as an intermedial iconic diagram in Tod Williams’ movie. In the novel, the ‘book within the book’ is read by focalizer Eddie (see Irving [1998] 2001:â•›47), through whose eyes the reader is presented with the story, s/he reading along with him. Conversely, in the movie adaptation of 2004, Ted’s fictional book is, of course, not read by the audience itself: Jeff Bridges in the role of Ted Cole reads it. With its presentation in an audio-visual medium, the self-reflexive, medial analogy between the book and the ‘book within the book’ ceases to exist. The iconic analogy which is established now relates, other than in the novel, merely to the movie’s content. The doubling of the novel’s quality as a predominately visual medium has ceased to exist because the visual form is suddenly presented in an audio-visual one. Once again, the adaptation of the mise en abyme for the screen from the Â�literary medium is not only a story of losses. Compared to the novel, the movie has not Â�forfeited its story’s intense layers of iconic cross-references. A Door in the Floor proves to be just as immediate to the recipient as the novel, since a lot of iconic features have been added. To begin with, the scene’s staging establishes an eerie atmosphere, which parallels the uncanny tone of Ted’s story. It is set in a large, dark room in which Ted gives a reading to an audience. Only the light of a small reading lamp illuminates the scene, that is, the book Ted reads from and half of his face. He is rendering his story in a solemn voice, which, again, increases the gloom of the scene and Ted’s reading. The fictional audience sits in a semi-circle around Ted. Yet, when the author starts rendering his story, the camera focuses solely on him so that the real audience and the fictional audience seem to merge: the recipient is literally being drawn into the scene – a clever trick of cinematography which heightens the effect of immediacy. Behind Ted’s armchair there is a screen that is also illuminated, showing Ted’s ink drawings. With these iconic images, the movie establishes another analogy to the children’s story and, consequently, to the story on the diegetic level, which the novel does not. In A Widow for One Year, Ted’s illustrations to “The Door in the Floor” are not described in detail. They are mostly parenthetically mentioned, even put into brackets: “(There was an indescribable illustration of […] horrible things as well)”

.â•… These instances, unlike the ‘losses’ before, belong to a greater extent to McFarlane’s second category (see Footnote 8).

 Christine Schwanecke

(Irving [1998] 2001:â•›48). On the other hand, the ink drawings presented in the movie adaptation are actually present and even attention-grabbing. They Â�parallel the scenes Ted reads and appeal to the viewer’s eye, thus involving the audience more wholly: first, they are required to both bring their aural and visual perception together. Second, they are forced to establish links between the images, the story on the hypodiegetic, and on the diegetic level – and finally, between the movie’s content and its title. As the analysis of this scene indicates, the audio-visual adaptation of Irving’s novel does not lose the density the novel displays. The Door in the Floor makes use of its own medial possibilities in order to create iconic signs which enhance its effect. Cinematography, music, even costume – all of these means are employed in order to parallel the movie’s content. With regard to cinematography, wide, long shots have been opted for in which the characters, in contrast to their vast surroundings, appear very small and forlorn. Thus, their mental states of unhappiness and loneliness, which also show in the actors’ words and expressions, are made visually perceptible. Above that, even colours are employed as iconic signs echoing the mental states of the Â�characters. All of them are very soft; nothing is highly contrasted or saturated. These faint, often grey, blue, and greenish colours appear to parallel the characters’ resignation and sadness, feelings that have been caused by the loss of loved ones and the disruption of a marriage. Costumes and music also play an important role in the movie’s signification Â�complex, as both communicate the character’s personality and feelings in a certain situation. Ted’s lover, Mrs. Vaughn, for example, wears mainly black dresses. The colour, stereotypically used for sad and/or evil characters, establishes visual Â�analogies to her mental state and the outside action. When Ted, for example, breaks up with the angry, distressed woman, whom he has sexually exploited, she first goes after him with her kitchen knife. As he, fearing for his life, runs away, she pursues him with her car, obviously trying to run him over. This chaotic scene, in which the furious black-clad woman (in her black car!) tries to take revenge, is highly Â�threatening; but it is also somehow comical. This tension between threat and humour is Â�paralleled and even heightened by the underlying score. Composer Marcelo Zarvos has striven to create music in which both emotions could be captured at the same time and presented with equal importance, just as they were in the actual movie scene (cf. Williams [2004] 2005). He decided to draw on the genre of dance, which has the power to oscillate between a lightness of rhythm and the emotionality of melody; the former, thus, echoes the scene’s humour, the latter its severity – the woman is trying to kill the man, after all. To sum up, all of these iconic features put together have created a highly enjoyable scene and, ultimately, movie. Overall, Irving’s and Â�Williams’ adaptation of the novel for the screen has proven to be a successful method of achieving their goals: they largely stick to the novel’s most important topics and render its basic tone.



John Irving and Tod Williams 

6.â•… Conclusion This analysis has explored the dense web of iconic features in both John Irving’s novel, A Widow for One Year, and in Tod Williams’ movie adaptation, The Door in the Floor. The novel is characterized by its high degree of visuality, which is, in turn, often caused by various types of iconicity. This paper has focussed on the novel’s predominant types, namely, its intra-medial and intermedial ones. In practice, these types Â�cannot be separated as clearly from each other as has been implied here. Iconic images, diagrams and metaphors are combined in order to establish a complex web of self-, meta- and cross-references, which accounts for, among other things, the plot’s effect of immediacy and credibility. Furthermore, it makes the novel a highly Â�challenging but also rewarding read. As iconicity is a transmedial phenomenon, it seemed likely that A Widow for One Year’s most important characteristic, its Â�abundance of iconic features, would also have been transferred to its movie adaptation. This is why, in the previous section of this analysis, I have examined the way in which their transfer was achieved. The investigation of the medial transfer has shown that the translation of Â�certain iconic forms and functions depend on the one hand on aesthetical considerations and on the other hand on the medial conventions and the material with which the icons are constructed. In adapting the novel for the film, director Williams had to lose some iconic signs, to modify others, or to establish new forms of iconicity, namely, Â�audio-visual ones. As a result, the movie’s web of semantic and structural correspondences has, while different from that of the novel, become as dense as the one woven in A Widow for One Year. Both medial products show such a wide range of iconic forms that it is certainly not disproportionate to close with the remark that both novel and film are truly ‘multiconic’.

References Argendeli, J. 1998. A Widow for One Year by John Irving. Rev. of A Widow for One Year, by J. Irving. CNN.com. June 2. 〈http://www.cnn.com/books/reviews/9806/02/widow.for.year. cnn/index.html〉 (28 April, 2009). Barthes, R. 1989[1980]. Die helle Kammer: Bemerkungen zur Photographie, D. Leube (transl.). Frankfurt: Suhrkamp. Campbell, J.P. 1998. John Irving. A Critical Companion. London: Greenwood Press. Fischer, O. & Nänny, M. 1999. Introduction: Iconicity as a creative force in language use. In Form Miming Meaning [Iconicity in Language and Literature 1], M. Nänny & O. Fischer (eds), xv–xxxvi. Amsterdam: John Benjamins. Irving, J. [1998] 2001. A Widow for One Year. New York NY: Ballantine. Irving, J. 2000. My Movie Business: Mein Leben, meine Romane, meine Filme, I. Rumler (transl.). Zürich: Diogenes.

 Christine Schwanecke Irving, J. [2004]2005. Vom Buch zum Film (Adapting the Novel for the Screen). In The Door in the Floor. Dir. by T. Williams. Perf. by J. Bridges & K. Basinger. DVD. Tobis. Joyce, J. 1922. Ulysses. Paris: Shakespeare and Company. Linke, A., Nussbaumer, M. & Portmann, P.R. (eds). 2001. Studienbuch Linguistik. Tübingen: Niemeyer. Ljungberg, C. 2001. Iconic dimensions in Margaret Atwood’s poetry and prose. In The Â�Motivated Sign [Iconicity in Language and Literature 2], O. Fischer & M. Nänny (eds), 351–66. Amsterdam: John Benjamins. Ljungberg, C. 2005. Photographs in narrative. In Outside-In – Inside-Out [Iconicity in Language and Literature 4], C. Maeder, O. Fischer & W.J. Herlofsky (eds), 133–149. Amsterdam: John Benjamins. Ljungberg, C. 2010. Intermedial strategies in multimedial art. In Media Borders, Multimodality and Intermediality, L. Elleström (ed.), 81–95. Houndmills: Palgrave Macmillan. Mcalpin, H. 2005. The damaged child inside the man: Until I Find You, by J. Irving [Book review]. Los Angeles Times Book Review. Los Angeles Times, July 10, 2005. 〈http://articles. latimes.com/2005/jul/10/books/bk-mcalpin10〉 (30 April, 2009). McFarlane, B. 1996. Novel to Film: An Introduction to the Theory of Adaptation. Oxford: Clarendon Press. Nöth, W. 2000. Handbuch der Semiotik. Stuttgart: Metzler. Nöth, W. 2001. Semiotic foundations of iconicity in language and literature. In The Â�Motivated Sign [Iconicity in Language and Literature 2], O. Fischer & M. Nänny (eds), 17–28. Amsterdam: John Benjamins. Nünning, A. 2001. Metanarration als Lakune der Erzähltheorie: Definition, Typologie und Grundriss einer Funktionsgeschichte metanarrativer Erzähläußerungen. Arbeiten aus Anglistik und Amerikanistik 26(2): 125–164. Rajewsky, I.O. 2002. Intermedialität. Tübingen: Francke. Rippl, G. 2000. Visuality and ekphrasis in A.S. Byatt’s Still Life and ‘Art Work’”. In Anglistentag 1999 Mainz: Proceedings, B. Reitz & S. Rieuwerts (eds), 519–534. Trier: WVT. The Door in the Floor. [2004] 2005. Dir. by T. Williams. Perf. by J. Bridges & K. Basinger. DVD. Tobis. Wagner, H. 2004. Ekphrasis. In Metzler Lexikon Literatur- und Kulturtheorie, A. Nünning (ed.), 112–113. Stuttgart: Metzler. Williams, T. 2005[2004]. Hinter den Kulissen [Behind the Scenes]. In The Door in the Floor. Dir. by T. Williams. Perf. by J. Bridges & K. Basinger. DVD. Tobis. Wolf, W. 1999. The Musicalization of Fiction: A Study in the Theory and History of Intermediality [Internationale Forschungen zur Allgemeinen und Vergleichenden Literaturwissenschaft Series 35]. Amsterdam: Rodopi. Wolf, W. 2004. Mise en abyme. In Metzler Lexikon Literatur- und Kulturtheorie, A. Nünning (ed.), 461–462. Stuttgart: Metzler. Wolf, W. 2009. Metareference across media: The concept, its transmedial potentials and problems, main forms and functions. In Metareference across Media: Theory and Case Studies [Studies in Intermediality 4], W. Wolf (ed.) in collaboration with K. Bantleon & J. Thoss, 1–88. Amsterdam: Rodopi.

Author index

A Abulafia, Abraham╇ 104 Amis, Kingsley╇ 343–351 Anderson, Douglas╇ 160 Aristotle╇ 104, 211, 253, 255, 258, 259 Ashcroft, Bill╇ 376 Augustine (saint)╇ 104 Auroux, Sylvain╇ 103 Austin, J. L.╇ 372 B Bakhtin, Mikhail╇ 373 Benjamin, Walter╇ 381 Benveniste, Émile╇ 105, 135, 146 Bopp, Franz╇ 104 Borges, Jorge Luis╇ 379 Bouissac, Paul╇ 137, 167 Brosses, Charles de╇ 104 Brown, Joe╇ 371 Bühler, Karl╇ 41, 44, 45, 47 Butler, Judith╇ 372 C Canby, Vincent╇ 371 Capozzi, Rocco╇ 292 Caton, R.A.╇ 343 Changizi, Mark╇ 383–384 Chirico, Giorgio de╇ 317 Colapietro, Vincent╇ 374 Coleridge, Samuel Taylor╇ 241–242, 247 Condillac, Étienne Bonnot de╇ 104 Cornell, Joseph╇ 313–324 Cratylus╇ 103–105 D Dante Alighieri╇ 104 Deacon, Terrence╇ 166 De Cuypere, Ludovic╇ 136, 196, 329

Derrida, Jacques╇ 258, 372 Dewey, John╇ 159 E Eco, Umberto╇ 137, 229, 291–292, 302–310 Emerson, Ralph Waldo╇ 169 Epicurus╇ 104 F Fauconnier, Gilles╇ 260, 264, 274–275, 277, 373 Fisch, Max╇ 158 Fischer, Olga╇ 106, 107, 215, 270, 287, 298, 315, 327, 408 Fónagy, Ivan╇ 106 Foucault, Michel╇ 337, 341, 379 Freeman, Margaret╇ 270 G Genette, Gérard╇ 103 Givón, Talmy╇ 106, 192, 196, 213, 347 Gollancz, Victor╇ 343 Gras, Marguerite R.╇ 382 Gras, Vernon╇ 382 Grassi, Ernesto,╇ 167, 168 Greenaway╇ 369, 370, 373–374, 376–387 Greenblatt, Stephen╇ 370, 378 Greimas, Algirdas Julien╇ 137 Griffiths, Gareth╇ 376 Groupe µâ•‡ 137, 138 Guha, Ranajit╇ 372 H Haley, Michael Cabot╇ 158, 160, 162, 163 Hausman, Carl R.╇ 158 Hemingway, Ernest╇ 211, 228 Hermogenes╇ 103–104

Hinson, Hal╇ 371 Hulme, Peter╇ 380 Humboldt, Wilhelm von╇ 104 Hurford, James╇ 102–103, 126 J Johnson, Paul╇ 349 Jakobson, Roman╇ 106, 142 Jonze, Spike╇ 371 K Kaufmann, Charlie,╇ 371 Klein, Herbert╇ 382, 385 L Lakoff, George╇ 84, 126 Lavater, Johan Caspar╇ 359 Leader, Zachary╇ 346 Lévy-Bruhl, Lucien╇ 39, 41, 52 Liberman, Anatoly╇ 19 Ljungberg, Christina╇ 168, 211, 313, 315, 316, 375, 376, 406, 410, 418 Locke, John╇ 104 Loomba, Anita╇ 380 M Meyer, Leonard B.╇ 238 Milton, John╇ 236, 241–242 Mitchell, W. J. T.╇ 324, 371 Monmonnier, Mark╇ 375 Morris, William╇ 269–287 N Nänny, Max╇ 107, 215, 229, 270, 286, 298, 316, 319, 327–328, 330, 347, 356, 408, 418 Niépce, Joseph Nicéphore╇ 360 Nigidius Figulus╇ 104 Nöth, Winfried╇ 212, 296, 315, 385, 405, 408

 Author index P Peirce, Charles Sanders╇ 4, 44–45, 106, 137, 142–143, 157, 158, 160, 161, 162, 163, 164–165, 169–170, 251, 291, 293, 295, 315, 344–345, 370, 374, 384, 391, 408 Phelan, Peggy╇ 370, 375, 376, 380, 383, 383 Piranesi╇ 376, 377, 380 Plato╇ 103, 169, 211, 212 Poe, Edgar Allan╇ 293, 362 Pope, Alexander╇ 236, 240 Pound, Ezra╇ 39–52 R Ransdell, Joseph╇ 159, 163 Ricœur, Paul╇ 140

Rizzolatti, Giacomo╇ 101–102 Romney, Jonathan╇ 379 Rosenbaum, Jonathan╇ 371 D Sacks, Oliver╇ 383 Saussure, Ferdinand de╇ 46, 83, 90, 91, 97, 102, 105, 106, 135–147, 211, 258, 315, 391 Schubert, Franz╇ 389–404 Sebeok, Thomas A.╇ 291, 292 Shakespeare, William╇ 237, 242, 309, 369, 370, 376, 380, 386 Shapiro, Michael╇ 160 Sharpe, Jenny╇ 380 Simic, Charles╇ 313–324 Spivak, Gayatri Chakravorty╇ 372 Stam, Robert╇ 371, 372, 386

T Thomas Aquinas (saint)╇ 104 Tiffin, Helen╇ 376 Todorov, Tzvetan╇ 135 Turner, Mark╇ 273, 275–277, 284, 373 Tweedie, James╇ 374, 379 V Vasari, Giorgio╇ 360 Venuti, Lawrence╇ 371, 374–375, 377, 385 Verene, Donald╇ 167 W Wallis, John╇ 104 Waugh, Linda╇ 106 Wolf, Werner╇ 315 Woolf, Virginia╇ 164, 167, 168

Subject index

A abduction╇ 293 abstractness ╇ 179–180, 182–183, 187 adaptation╇ 369, 370, 371, 373, 374, 375, 384, 387, 405, 415–416, 419–421 film adaptation╇ 369, 370, 371, 372, 373, 374, 385, 386 intermedial adaptation╇ 370 postmodern adaptation╇ 370 adverbs 120–125, 207 place╇ 121, 122 time╇ 121–124, 192 Afrikaans╇ 62, 71, 327, 331, 339 alexia (word blindness, visual aphasia)╇ 383 allocutor╇ 120 American Sign Language╇ 20–21 analogy╇ 49, 65, 84, 128, 136, 147, 211, 247, 295, 323, 356, 419 animacy╇ 178–179, 187 arbitrariness╇ 46, 48, 65, 70, 73, 77, 83–84, 89–90, 97, 101–103, 104–106, 107, 142–143, 211–212, 251, 315 archetype╇ 252, 341 association╇ 40, 67, 89, 104, 105, 106, 258, 306, 322, 334, 384, associative center╇ 140 audio-visual-verbal (see also verbal and visual)╇ 371, 372 avant-garde film╇ 380 B blending see conceptual blending borrowing╇ 23, 25, 30, 36 brain╇ 151–152

C category╇ 158, 159, 160, 163, 170 Chinese language╇ 44, 83–97, 254, 256, 258, 263 chronological order╇ 191 cinema╇ 379, 385 cinematic convention╇ 376 cinematic language╇ 387 cinematic reading╇ 384 cohesion╇ 195, 207, 223–227 competence╇ 151, 154, 393, 395, 396 complexity╇ 35 concept╇ 55, 83, 127, 260, 265, 375 conceptual╇ 379 conceptual blending╇ 257, 269, 274, 275–276, 287 conceptual integration see conceptual blending conceptual prominence╇ 175–176 condensation╇ 40, 214, 324 cognition╇ 125, 384, 385 cognitive aspects╇ 257, 370, 385, 391 cognitive poetics╇ 232, 234, 246, 251, 269 cognitive precedence╇ 175, 176 cognitive processing╇ 179, 257, 285, 371 cognitive type╇ 135, 138, 357 connotation╇ 43, 55, 63, 69, 113, 137, 201, 331, 335 consonant╇ 6, 15, 24, 116 D defamiliarisation╇ 291, 327, 329, 331, 381 denotation╇ 55, 137, 213, 221, 336, 338 diagram╇ 106–107, 157–167, 370, 376, 384 diagrammatic╇ 107, 383, 386 diagrammatic aspect╇ 384 diagrammatic reasoning╇ 384

iconic╇ 111 imagic╇ 107, 111, 117, 120, 121 diminutives╇ 7, 59, 62–63, 71 dislocation╇ 369, 370, 376, 377, 379, 383, 386 double-scope blends╇ 373 dyadic opposition╇ 162 relationships╇ 158, 159, 162 dynamical see object E Early Modern╇ 380 ekphrasis╇ 314–316, 318, 321, 323, 341, 414–415 emotion╇ 39–40, 79, 154, 233–234, 237, 238, 246, 278, 285, 287, 331, 358, 391, 421 exophoria╇ 315 expressives╇ 4 F firstness╇ 157, 158, 159, 160, 161, 162, 166 foregrounding╇ 169, 313, 316, 323, 331, 336 G gestalt╇ 47–49, 51, 59, 238–239, 246 grammatical lexeme╇ 108, 110 H Hollywood cinema╇ 376 narrative cinema╇ 385 homophones╇ 111, 296–297, 298, 302 homology╇ 146 hybrid╇ 371 hybridity╇ 373 hybridizing╇ 379 hypoicon╇ 157–163, 408

 Subject index I icon╇ 88, 106, 157, 158–160, 211–212, 214, 319, 347, 356, 359, 413 iconic character╇ 370, 383 enactment╇ 369 function╇ 159, 160, 381 processes 387 representation╇ 373 (see also polyvalent iconic representation), 383 sign╇ 135, 370, 373 thinking╇ 19, 36 words╇ 3–16 iconicity╇ 106–107, 313, 321, 323, 369, 386 acoustic╇ 47, 103, 107, 140–141, 291, 294, 298, 300, 302 chronological╇ 192, 195 diagrammatic╇ 36, 47, 49, 64–67, 107, 113, 163, 213, 215, 217, 315, 321, 331, 337, 343, 346–347, 411, 414 endophoric╇ 295, 298, 301, 310 imagic╇ 4, 36, 46–47, 64, 106, 318, 320, intermedial╇ 315, 369–388 linguistic╇ 136 metareferential╇ 411 phonaesthetic╇ 91–94 pictographic╇ 94–95 poetic╇ 251, 271 temporal╇ 83, 191, 194–197, 198–200, 203, 208 ideograms╇ 39, 41–42, 45 ideographic writing╇ 52, 141, 251 ideophones╇ 3, 39–52, 71, 77–78 image╇ 16, 106, 135, 157–159, 369, 370, 372, 374, 377, 378, 379, 381, 382, 383, 384, 385 imitation╇ 20, 26, 30, 45, 46, 51, 70, 126, 136, 211, 235, 300, 329, 323, 356, 392, 413 index╇ 157, 158–159 indexical function╇ 159, 357, 384, 391–393, 404 indexicality╇ 157, 159, 166, 253, 259, 291–293, 321, 355–361, 365, 367, 382, 384 intentionality╇ 136, 161, 346 intermedial╇ 302, 370, 373, 390–391, 408, 421 art forms╇ 323, 375

intermediality╇ 313, 369, 379, 412–414 interpretant╇ 374, 375, 377, 381, 385 formal interpretant╇ 374 thematic interpretant╇ 375 isomorphism╇ 64–65, 72, 77, 110–111, 136–137, 147, 212 J Japan Sign Language╇ 19–36 K Kabbalah╇ 104 L langue╇ 105–106, 126, 135, 140–142 letter shapes╇ 153 lexicalization╇ 56, 71–72, 74, 79 likeness see similarity linearity╇ 141 locutor╇ 119, 129 M mapping╇ 43–47, 213, 246, 251–253, 370, 378, 384, 411, 412, 417–418 markedness╇ 4–5, 16, 83 media-specific materiality╇ 316, 369 mediality╇ 316, 321 memory╇ 177, 179, 181, 183, 304, 309, 357, 365, 379 cultural╇ 313, 324, 340, 371 metalanguage╇ 41, 323, 331, 411 metaphor╇ 27, 56, 65–66, 78, 157–170, 213, 256, 269, 270, 323, 370, 375, 416 metaphorical blending╇ 269–270, 287 metonymy╇ 56, 65, 67, 69, 258, 358 mimesis╇ 41, 211, 253, 324, 419 mimetics╇ 3, 8, 11, 44, 101, 103, 125, 412 mimicry╇ 39, 41, 45–47, 59, 101, 127, 151, 153, 220, 391–392 mirror neurons╇ 101 modality╇ 316 monophonemes╇ 110, 116, 117, 122 motivation╇ 7, 21, 30, 36, 48, 55, 71, 73, 75, 77, 90, 97, 135–136, 143, 146, 300–301, 315, 329

multimedia╇ 139, 369, 370, 374, 381, 385, 386 multidimensionality╇ 141 music╇ 153, 233, 238, 304, 339, 374, 389–403, 420 N narrative╇ 191, 200, 203, 205, 207, 211, 216, 228, 273, 281, 295, 297, 300, 302, 304, 306, 308, 319, 321, 327, 329, 374, 385, 406 nature harnessing╇ 149, 153 naturalism╇ 107 neurovisual system╇ 384 O object dynamical╇ 157–161, 162, 163, 165, 170 onomatopoeia╇ 4, 46, 58, 65, 97, 108, 110, 143, 233–235, 300, 356, 391, 393 opacity╇ 55–56, 63, 68, 70–71, 153, 339 ordo artificialis╇ 192, 198, 206, 207 P painting╇ 51, 141, 212, 255, 355–357, 369 parallelism╇ 158, 160, 163 pattern╇ 35 pattern poem╇ 314 perception╇ 47–48, 101, 220–221, 239, 242–245, 329, 356, 357, 384, 412, 420 performance╇ 369, 370, 372, 385 performative╇ 370, 373, 373, 386 criticism╇ 372 metareference╇ 385 statement╇ 372 performativity╇ 380, 386 phonological system╇ 107, 109 photography╇ 159, 302, 355–367, 382, 414 picto-phonetic characters╇ 86, 90–96 pictographic writing╇ 86–87, 89, 97, 266, 313, 316 pictorial╇ 371 poetry╇ 40, 180, 233–247, 251–266, 315, 319, 323, 403 Chinese╇ 251–266

polyvalent iconic representation╇ 373 postcolonialism╇ 372 postmodern╇ 313, 314, 380 projection╇ 375, 376, 386 R reduplication╇ 4, 43, 48, 55–79 referentiality╇ 3, 5, 15–16, 137–138, 315, 337, 339, 341, 386, 408–409 Renaissance╇ 360 convention╇ 382 painting╇ 371 rhetoric╇ 169 theater╇ 380 resemblance╇ 159, 163, 164, 165, 166 immediate resemblance╇ 165 inherent resemblance╇ 165, 167 basis of resemblance╇ 165–170 reverse order see ordo artificialis S salience╇ 177–178, 181–182, 187 secondness╇ 162 self-reflexive╇ 370, 386 self-reflexivity╇ 316, 369, 380 semi-iconic words╇ 8 semiosis╇ 163, 170, 266, 293, 294–302 sensory imagery╇ 39–41 modalities╇ 45, 316, 371 perception╇ 43–44

Subject index  sequentiality╇ 23, 29, 142, 198, 211, 215, 229, 321, 414 signification╇ 45, 145, 157, 159–160, 261, 390 signified╇ 83, 101, 103, 105, 106, 126, 135–136, 229, 265, 298, 329, 391, 403, 419 signifier╇ 97, 101, 106, 135, 138, 211, 215, 265, 298, 329 signifier-signified relationship╇ 107, 142, 143, 213, 339, 391 similarity╇ 95, 157, 164, 166, 168–169, 335, 355, 364, 386 sound symbolism╇ 2, 6, 11, 44–45, 49 sounds╇ 153 spatiality╇ 114–115, 379 speech act╇ 109, 120, 122 still life╇ 317 subaltern╇ 372 symbol╇ 160–161, 135 symbolic╇ 161 symbolically╇ 157 symbolic logic╇ 145 system╇ 145–146 T technopaignium╇ 314, 318 temporal continuity╇ 379 tension╇ 162, 220, 239, 313, 323, 330, 369, 395, 420 thirdness╇ 158, 160, 162 transfer╇ 21, 24–30 form-meaning╇ 26 transformation (intersemiotic)╇ 370

translation╇ 369, 371, 374–375, 382, 384, 385, 386 rhetorical, semantic, structural, stylistic features of translation╇ 374 transparency╇ 19, 21, 36, 56, 59–78, 113, 145, 180, 205, 347 trichotomy╇ 106, 107, 157, 158, 163, 170, 292, 408 V value theory of╇ 106 vehicle╇ 160-161, 408 verbal╇ 374 visual╇ 370, 374, 379, 386 visual aphasia╇ 383 visual attitudes╇ 385 visual image╇ 96, 140–141, 291, 304, 309, 383, visual language╇ 370, 383, 387 visual literacy╇ 369, 370, 383, 385 visual recognition╇ 384 visual skills╇ 385 visual system╇ 142, 152, 384 visual vocabulary╇ 369, 383, 386 visual word form╇ 383 visual writing systems visuality╇ 371, 384 vowel╇ 5–7, 12–13, 24, 110, 116 Z zeugma╇ 176, 180