Language and History Integrationist Perspectives (Routledge Advances in Communication and Linguistic Theory)

Language and History

When linguistics was first established as an academic discipline in the nineteenth century, it was envisaged as an essentially historical study. Languages were to be treated as historical objects, evolving through gradual but constant processes of change over long periods of time. In recent years, however, there has been much discussion by historians of a ‘linguistic turn’ in their own discipline, and, in linguistics, integrationist theory has mounted a radical challenge to the traditional notion of ‘languages’ as possible objects of inquiry. Language and History develops the integrationist critique of orthodox linguistics, at the same time extending its implications to the field of history. By doing so, it throws light on what is now recognised by many historians to be a ‘crisis’ in their own discipline. Underlying the post-modernist scepticism about traditional forms of historiography, the integrationist approach reveals a more deep-seated problem concerning the interface between philosophy of history and philosophy of language. With chapters from a range of leading international contributors, Language and History represents a significant contribution to the developing work of the integrationists. Nigel Love is Associate Professor of Linguistics at the University of Cape Town, South Africa. Topics on which he has been published include Welsh phonology, the work of J.L. Austin, language and cognition, and the linguistic theory implied by procedures of statutory interpretation. He is editor of the journal Language Sciences.

Routledge Advances in Communication and Linguistic Theory Series Editor: Roy Harris University of Oxford

This series presents an integrationist approach to problems of language and communication. Integrationism has emerged in recent years as a radically innovative theoretical position. It challenges the most basic assumptions underlying orthodox twentieth-century linguistics, including those taken for granted by leading structuralists, post-structuralists and generativists. According to integrationists, human communication is an essentially creative enterprise: it relies very little on the ‘codes’, ‘systems’, ‘habits’ and ‘rules’ postulated by orthodox theorists. Instead, integrationists see the communicative life of each individual as part of a continuous attempt to integrate the present with the past and the future. The success of this attempt depends crucially on the ability to contextualise on-going events rather than on any mastery of established conventions. The books in this series are aimed at a multidisciplinary readership comprising those engaged in study, teaching and research in the humanities and social sciences, including anthropology, the arts, education, linguistics, literary studies, philosophy and psychology. 1. Words – An Integrational Approach Hayley G. Davis 2. The Language Myth in Western Culture Edited by Roy Harris 3. Rethinking Linguistics Edited by Hayley G. Davis and Talbot J. Taylor 4. Language and History Integrationist perspectives Edited by Nigel Love 5. The Written Language Bias in Linguistics Its nature, origins and transformations Per Linell

Language and History Integrationist perspectives

Edited by Nigel Love

First published 2006 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2006 Nigel Love, selection and editorial matter; the contributors, their own chapters This edition published in the Taylor & Francis e-Library, 2006.

“To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”

All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book has been requested ISBN10: 0–415–31762–2 (Print Edition) ISBN13: 978–0–415–31762–7

Contents

List of contributors Preface 1 Language, history and Language and History

vii ix 1

NIGEL LOVE

2 The end of linear narrative? Reflections on the historiography of English

19

ROGER LASS

3 History and Comparative Philology

41

ROY HARRIS

4 Word-stories: etymology as history

60

CHRISTOPHER HUTTON

5 Language: object or event? The integration of language and life

71

MICHAEL R. WALROD

6 Indeterminacy of meaning and semantic change

79

EDDA WEIGAND

7 On the cusp: Antoine Meillet as a sociologist of language

99

ROBERT McCOLL MILLAR

8 ‘The grammatical being called a nation’: history and the construction of political and linguistic nationalism JOHN E. JOSEPH

120

vi

Contents

9 How to make history with words

142

ROY HARRIS

10 Talking about what happened

156

TALBOT J. TAYLOR

11 Part of the meaning/history of euro: integrational corpus linguistics

172

MICHAEL TOOLAN

12 Language and prehistory

188

G E O F F R E Y G A LT H A R P H A M

13 Bridges to history: biomechanical constraints in language

200

STEPHEN J. COWLEY

References Index

223 239

Contributors

Stephen J. Cowley works at the universities of Hertfordshire in England and KwaZulu-Natal in South Africa. After completing his PhD on the interpersonal effects of Italian prosody, he turned to how external language influences human cognition. His current empirical work investigates how linguistic activity transforms infant agency. Geoffrey Galt Harpham is Director of the National Humanities Center in Research Triangle Park, North Carolina, and Research Professor at Duke University and the University of North Carolina. He is the author, most recently, of Shadows of Ethics: Criticism and the Just Society (Duke University Press, 1999) and Language Alone: The Critical Fetish of Modernity (Routledge, 2002). Roy Harris is Emeritus Professor of General Linguistics at the University of Oxford and Honorary Fellow of St Edmund Hall. He has also held university teaching posts in Hong Kong, Boston and Paris and visiting fellowships at universities in South Africa and Australia and at the Indian Institute of Advanced Study. Christopher Hutton is Associate Professor and Head of the Department of English at the University of Hong Kong. He is the author of Abstraction and Instance (Pergamon, 1990), Linguistics and the Third Reich (Routledge, 1999) and Race and the Third Reich (Polity, 2005). John E. Joseph is Professor of Applied Linguistics at the University of Edinburgh. His most recent book is Language and Politics (Edinburgh University Press, 2006). He currently holds a three-year Major Research Fellowship from the Leverhulme Trust. Roger Lass is Emeritus Professor of Linguistics at the University of Cape Town. He is the author of many works on historical linguistics, the history of English and phonological theory, and he edited the third volume (1999) of the Cambridge History of the English Language. Nigel Love teaches linguistics at the University of Cape Town. He is co-author, with John Joseph and Talbot Taylor, of Landmarks in Linguistic

viii Contributors Thought II: The Western Tradition in the Twentieth Century (2001) and editor of the journal Language Sciences. Robert McColl Millar lectures in linguistics at the University of Aberdeen. He has published widely on language change from a social and cultural perspective as well as on Scottish language policy. His second book, Language, Nation and Power was published by Palgrave Macmillan in 2005. Talbot J. Taylor is the author of several books on language theory, sociolinguistics, the rhetoric of language science and the history of ideas. He is Director of the Linguistics Program at the College of William and Mary, Williamsburg, and co-editor, with Roy Harris, of the interdisciplinary journal Language and Communication. Michael Toolan is Professor of Applied English Linguistics at the University of Birmingham. He is the author of Total Speech: An Integrational Linguistic Approach to Language (1996) and editor of the Journal of Literary Semantics. Michael R. Walrod is President of the Canada Institute of Linguistics at Trinity Western University, Langley, British Columbia. He lived and worked for many years in the Philippines, where he was involved in translation and community development. His research interests and writings are in the areas of philosophy of language and discourse analysis. Edda Weigand is Professor of Linguistics at the University of Münster. She has published widely on pragmatics, speech act theory, dialogue theory and lexical semantics.

Preface

This volume is in part a record of the second conference of the International Association for the Integrational Study of Language and Communication (IAISLC), held in New Orleans, 26–28 March 2002. It is therefore the second in the series inaugurated by The Language Myth in Western Culture, edited by Roy Harris and published by Curzon in 2002. But one or two of the chapters have been contributed by scholars who were not present at the conference, and who would not necessarily share all or any of the theoretical positions taken by integrationists. It is hoped that bringing these into conjunction with the chapters written by members of IAISLC will both illuminate the integrationist argument and make for a more interesting book. My thanks to all the contributors, who along with the public at large have awaited its much overdue publication with patient forbearance. It is also only a partial record in that not every item on the New Orleans agenda is represented here. The saddest omission is ‘The prehistory of grammar’ by George Wolf, who despite being seriously ill, planned, arranged and attended the conference, where he delivered what he had been able to write of his paper. He died shortly afterwards. It would not have been fitting to present his thoughts in the form of an unfinished draft and it has not been possible to reconstruct a publishable version. This volume could never have existed without George’s devoted and in the circumstances heroic organisational work. It is dedicated, with affection and gratitude, to his memory. NL

1

Language, history and Language and History Nigel Love

Defining a topic for discussion by conjoining two large, vague nouns with ‘and’ is not calculated to elicit narrowly focused debate. On the contrary, it is a procedure that apparently invites contributors to say anything that in some way relates one designated thing or phenomenon to the other; and when the things or phenomena are in themselves complex and multifaceted, and the connections and relationships between them numerous and varied, the likeliest outcome is no more than an enhanced and perhaps rueful appreciation of how many disparate issues can be made to fall under the heading in question. This shortcoming of the ‘X and Y’ formula may appear especially salient in the case of ‘language and history’. To take only the most obvious point, it seems to embrace questions pertaining to the language of history and questions pertaining to the history of language. Aren’t those two (at least two) different subjects? Is there anything to be gained by bringing them together in a single volume? Does this collection go beyond merely bringing them together and perform the altogether more useful and significant operation of yoking them together? That is for its readers to say. At any rate, without claiming that everything in this book is directed towards sustaining the thesis that the two kinds of linguo-historical question are related, an attempt can be made to spell out that thesis and to draw attention to those aspects of the following chapters that support or illustrate it. A commonsensical view of the relation between language and history might run as follows. In the primary sense of the word, history is what happened in the past. In a secondary sense, sometimes covered by the word ‘historiography’, history is what historians produce: in essence, collections of statements, factual or interpretative, as to (some of ) what happened. Language serves the historian, as it serves any other species of statementmaker, by providing the linguistic equipment for making statements. That equipment consists of words whose meanings convey the substance of the statements the historian wishes to make. A language is a device for encoding systematically recoverable semantic content. What could be simpler than that? A famous epistemology – Popper’s – has been founded on treating language and languages as thus unproblematically capable of enshrining ‘objective knowledge’ (see e.g. his 1972 book of that title).

2 Nigel Love That matters are not quite so simple may be initially insinuated with the observation that languages themselves (allegedly) have histories; and this, if true, must be expected to complicate their role as repositories of stable meanings. But once again common sense offers a bluff response. Languages do indeed change over time, and the reader of non-contemporary texts must be alert to the interpretative difficulties this may cause. But can we not rely on linguistic historians (philologists, historical linguists) to reconstruct the history of languages and thereby alleviate such difficulties? The temporal instability of languages may in certain ways be problematic, but the problems can in principle be overcome. Christopher Hutton’s chapter in this volume raises by implication the question of how far common sense of this order is the product of a linguistic culture whose writing happens to involve the deployment of characters that have long since come to be treated as visually arbitrary. This is not so with Chinese, for instance. Chinese characters offer a ‘fundamental semiotic challenge’, being read ‘in complex ways as having both a pictorial and a phonetic element’. That is to say, the meaning and history of the word taken to be represented by a character is bound up with questions about what the character depicts or portrays and how that portrayal has evolved. The whole enterprise of reconstructing linguistic history is thus from the outset liable to be conceived in ways that differ radically from those taken for granted in other cultures, with a range of consequences not just for how in detail ancient texts are interpreted but for what it is to interpret them. Even if we confine ourselves to the world of Western languages and linguistics, one would have more confidence in the deliverances of common sense on these questions if linguistic historiography were theoretically better founded than it is. The purpose of that enterprise is to investigate and describe the history of languages, i.e., how they change through time. But the fact is that there is no agreement among historical linguists as to either (a) what a language is, or (b) what it is that changes when a linguistic change occurs or (c) what kind of event, if any, a linguistic change might be. This is an unpromising state of affairs. It is a linguists’ platitude that what counts as a language in ordinary parlance is a matter of indefinitely variable sociocultural definition; that no linguistic (i.e. internal structural) justification can be found for distinguishing Swedish and Danish as separate languages if Mandarin and Cantonese are to be acknowledged as versions of the same language; that what happened to Dutch when taken to South Africa was on the face of it no more cataclysmic than what happened to English when the Normans invaded, and yet South African Dutch became Afrikaans, while English marched on as English. Historical linguists, for their part, have in general never been particularly concerned to impose some special definition of their own, uniquely required for their historians’ purposes,1 or to displace ordinary parlance by defining the term ‘a language’ so as to give unequivocal answers to questions about the boundary between one language and another (whether in time, in space, in

Language, history and Language and History

3

terms of social stratification . . .), or to specify how exactly the concept of a dialect relates to the concept of a language, or to worry over whether the sense of ‘a language’ in which Spanish, say, is a form of Latin, should or should not take priority over the sense in which Spanish and Latin are different languages. Such issues have traditionally either not been thought to matter or else have been set aside sine die, or at any rate for a rainy one. In his chapter in this collection Roger Lass considers, from the perspective of a historical linguist working (albeit critically) within the orthodox scholarly tradition that starts with nineteenth-century comparative philology, some of the consequences for linguistic historiography of supposing that a language name like ‘English’ uncontroversially identifies a domain susceptible to historical treatment in terms of a linear narrative. English is especially problematic in this respect because of the relative wealth of documentary evidence available from different times and places, and because of the depth and intensity of study to which surviving texts have been subjected. A paradox emerges: the more that is known of a language’s history the less reason there is to suppose that it has one. Published histories of English are replete with statements to the effect that A ‘becomes’ B in cases where there is no reason to suppose, even granting the standard assumption that in certain circumstances the formula A ⬎ B has a coherent interpretation,2 that A and B stand in any identifiable relationship of temporal succession. Forms attested in eighth-century Northumbrian are treated as directly ancestral to forms attested in tenth-century West Saxon. It is as if items from one period in Italian were to be linked by the A ⬎ B formula to items from a later period in Portuguese. The reason that would be unacceptable is that Italian and Portuguese are well known to be different ‘languages’, whereas from a standpoint that starts by identifying ‘English’ as the object whose history is to be mapped, Northumbrian and West Saxon are not. They are merely ‘dialects’. Thus the historiographical enterprise is distorted from the outset by the requirement that a folk language name correspond to a real-world entity whose history can be likened, in Lass’s trope, to a tube or cylinder whereby ‘the historiographical act consists of pouring in forms at one end . . . and looking to see what occurs at the various graduation marks (the points at which the surviving texts allow us to intersect the “stream” of history), and then finally seeing what comes out of the bottom’. Lass further observes that ‘historians tend to adopt a manner of speaking in which the collection of materials whose ancestry we are reconstructing, no matter how complex and variable, always at some remove appears to go back to a single ancestral object whose “story” we then find ourselves telling’. We know quite well from reflection on our own first-order experience of language that it is hard to find ‘single objects’ at any level of generalising description. Lass takes the example of ‘the’ Modern English vowel ‘short a’, as in bat, cab, sap, etc., which, looking at the English-speaking world as a whole, corresponds at the very least to all of [a], [ ], [a], [␧], [e]. But as we go back in time such variation dwindles away: ultimately all of these will be held to have a a

4

Nigel Love

single, unitary origin in the ‘comparativist fiction of the “dialect-free protolanguage” ’. There is no such phonetic object as ‘short a’ (nor for that matter any unitary phonological analysis in terms of which ‘short a’ would emerge as a single pan-English phoneme), but the very act of setting it up for the modern language establishes it as the reflex of some definite ancestral entity. Thus does a language’s history emerge as an artefact of the particular synchronic analysis taken as a starting point. Why are these conceptual problems that swirl round the concept of ‘a language’ largely ignored? Partly because they are intractable, and partly because the central proposition on which Western historical linguistics is founded is not that it is possible to identify languages across transformations in time but linguistic substance constitutive of linguistic units, i.e., the component microparts of languages. The issue of which language any units thus historically located count as belonging to can indeed be left to politics, or societal Selbstverstehung, or whatever; the historical linguist is in business on the footing that, given any two états de langue, however crudely identified, it is possible to say whether they do or do not contain a core of material, at the meaningful level of articulation, which counts as ‘the same’ in that one can plausibly postulate continuous direct or indirect historical transmission of linguistic units combining a form with a meaning. Non-meaningful sameness, i.e., the subject matter of phonetic or phonological history, has in practice tended to get the lion’s share of attention in historical linguistics, but in fact it is logically subordinate to meaningful sameness in that one has no basis for historically relating two ‘sounds’ unless there is some justification for relating meaningful units in which those ‘sounds’ appear. So what makes English Indo-European is hundreds of etymologist’s facts such as that the word thatch contains the same root as the Latin verb tego, that feather is ultimately the same, or contains some of the same linguistic substance, as Welsh adar ‘birds’, that cow and the stem of bovine are ultimately cognate, and so on and so forth. What makes English Germanic is hundreds of facts such as that although English father, German Vater, Latin pater are all ‘the same word’, as are English foot, German Fuss, Latin pes, the English forms stand closer to the German than they do to the Latin, in that there are forms, attested or reconstructible, taken to be ancestral to both the English and the German that are less historically remote than forms ancestral to the English, the German and the Latin. And what makes Modern English and Old English the same language historically speaking (even if they are different languages according to most other kinds of speaking) is that many Old English forms are themselves identifiable as the recent forebears of Modern English forms. But where do such ‘facts’ come from? Feather and adar, for instance, have neither a single phonetic segment in common (granted a non-rhotic English speaker) nor anything but the loosest semantic connection. Crucially required, it might seem, is some way of perceiving the ‘sameness’ allegedly lurking behind obvious – in many cases, gross – discrepancies. But this is the wrong way to look at it. It is not a matter of first noticing a (non-existent)

Language, history and Language and History

5

sameness and then working back to ancestral forms whose divergent development in different branches of Indo-European has caused the differences. It is a matter of demonstrating that, provided it is not on semantic grounds wholly implausible to associate feather with adar, a story can be told whereby given, among many other things, a certain reconstructed Indo-European root or form, certain established sound changes distinguishing Germanic from Celtic . . . , continuous transmission through many generations might be expected to yield feather or something like it as the contemporary end point of one line of development and adar or something like it as the contemporary end point of the other. The alleged ‘sameness’ is an artefact of the notion of ‘continuous transmission’. What is claimed to have been continuously transmitted? The only possible answer is: a metalinguistic reification located in the mind of the historical linguist. In trying to rationalise this onto-epistemological curiosity, two kinds of explanation suggest themselves. In ‘History and comparative philology’ Roy Harris discusses issues in the historiography of historical linguistics itself. He suggests that the main achievement of early scholars in the tradition of comparative philology, and their main claim to have founded a science of language, was that, going beyond merely observing many formal-semantic correspondences across Indo-European languages, they devised ways of systematising those observations and reducing them to simple correlational formulae. But this immediately invites the question what the correlations mean and why they should hold. ‘Interlinguistic correspondence formulae (of the kind typified by Grimm’s Law) are no more than that: i.e., formulae. In themselves they contain no historical information and capture no “facts” other than the observations and assumptions (whether accurate or inaccurate) that went into their formulation.’ So an interpretation is required. The preferred interpretation, much influenced by a contemporary revolution in biological science, was to suppose they represented the different development, along separate evolutionary lines, of linguistic entities that could be traced back to a common ancestor. Thus did comparative philology find itself committed to belief in the survival across millennia of sounds and forms which although in many cases transformed beyond recognition nonetheless mystically retained their identity. In the biological case items that did indeed retain their identity across such time spans were eventually identified and called genes. In the linguistic case, needless to say, nothing analogous to genes has ever turned up. Another kind of explanation would point out that the ‘sameness’ of linguistic forms across long time spans is no more than a logical extension of a commonplace feature of our understanding of how languages work ‘synchronically’. If you and shortly afterwards I each produce an utterance that might be written down as ‘feather’, then irrespective of any phonetic differences between them we have both said ‘the same thing’. That in itself, properly understood, may be uncontentious enough. What is highly contentious is that ‘saying the same thing’ in such a case is to be interpreted as instantiating an abstract

6

Nigel Love

invariant, viz. the ‘thing’ (in this case the word feather) that has been said twice.3 Once one makes that reifying move there is nowhere to stop. For if the abstraction in question remains the same across minor phonetic differences (as between your pronunciation and mine) and tiny time gaps (as between your utterance and mine), at what specifiable degree of enlargement do the differences and the gaps begin to matter, and why? A viable concept of the limits of ‘a language’ would help here. But it is not clear that there is any to be had. What, then, is a linguistic change, granted a historical linguistics crucially founded on identifying linguistic continuities through time? Is there a theoretical perspective in which it makes sense to conceive of linguistic objects as changing and yet mysteriously staying the same? Let us approach an answer by sketching the kind of account a historical linguist might give of a small lexical change in English that seems to be taking place right now, if indeed it has not already been accomplished. English contains a word font, which belongs primarily to the technical vocabulary of printers, and which according to the second edition (1989) of the Oxford English Dictionary4 refers to ‘a complete set or assortment of type of a particular face and size’. It has a number of homonyms, and a variant fount. As against that, the word-processing program with which I am now writing offers me (i) a choice of ‘fonts’, and (ii) a choice of ‘font sizes’, each choice independent of the other. This suggests that, at any rate for the designers of Microsoft Word, the word font does not mean what the OED says it means, in at least two respects. For the fonts in question are not ‘assortments of type’, nor are they of any particular size. What does the linguist make of these discrepancies? He is likely to detect a paradigmatic case of semantic change, recent or ‘in progess’, induced by developments in the technology of printing. The use of movable type requires discrete collections of physical objects (‘types’) each consisting of an adequate supply of the various characters needed for printing in a given size and ‘face’. Such a collection is, as the OED tells us, called a ‘font’. Any desired change in either the size of the type, or the typeface, or both, calls for a different font. So in this usage 10 pt Garamond is as much a different font from 12 pt Garamond as either is from any size of Baskerville. But computer simulation of the effects brought about by changing fonts does not depend on substituting one set of physical objects for another. Moreover, the software works in such a way that size is most conveniently treated as an independent variable.5 Hence the world of electronic printing has no use for the word font in its old face-and-size sense, and the way is open for it to change its meaning. Font has come, or is well on its way to coming, to mean ‘typeface’. It remains to be seen whether that meaning is destined to be merely additional to the old one, or whether it will eventually supplant it. A story along these lines is supported by the fact, readily ascertainable by inquiry among, e.g. contemporary university undergraduates, that there are many members of the generation familiar from childhood with personal

Language, history and Language and History

7

computers who are unaware that font ever meant anything else. Moreover, one of the semantic developments here can be slotted into a familiar category: old linguistic expressions are commonly retained, with consequently different meanings, across changes in the way things are done, as when we carry on ‘dialling’ telephone numbers even though what we actually do is press buttons, or when ships continue to have ‘sailing times’ despite having no sails. Let us briefly consider just some of the assumptions on which such an account rests. The first assumption is baldly announced in the opening observational statement: ‘English contains a word font’. This is a statement about a certain public codification6 of the English lexicon. That is to say, in Michael Walrod’s image, font is one of the objects currently on display in the lexicographer’s museum. But one may ask why the relevant part of the lexicon is codified in the way it is, and whether alternative ways would not be at least equally appropriate. Simplifying what the OED has to say, it appears that there are (at least) two separate nouns with the form font. One of them – the one with which we are concerned – has fount (exactly the same meaning as font in the printing sense) as a variant. The other (‘baptismal water container’) has connections with, but is not apparently to be seen synchronically as a mere variant of, the word fount as a poetic or metaphorical version of fountain. So two forms, font and fount, are analysed as three words: font, fount and font/fount. What is the rationale for this analysis? It seems to be largely etymological, but etymology is not taken as definitively settling the matter: if it did, presumably font in the printing sense would count as the same word as fondue7 – which of course is precisely what the macrohistorical linguist would want to say. We seem to be working here with some synchronic notion of wordhood and word-identity which, while partially or intermittently deferential to the macrohistorian’s notion, nonetheless diverges from it in accordance with criteria that are far from obvious. Such considerations have a bearing on the putative linguistic change we are considering, in as much as they are involved in settling one kind of question as to what exactly it is that is held to be changing. Fount as a variant of font in the printing sense seems never to have had much currency in the United States. And since it may be surmised that the change whereby font comes to mean ‘typeface’ originates in America, we may ask whether it extends to the variant fount at all. Certainly those for whom font means ‘typeface’ do not seem to use fount in this way. If that is so, it would be good reason for ceasing to treat fount as a mere variant of font. In other words, the change in the meaning of font has arguably brought about a lexical differentiation of font and fount. All this, of course, against the background of the macrohistorian’s insistence that font and fount in the printing context just are and will continue to be the same word. A second assumption is that when a word-processing program offers its users separate drop-down menus of fonts and sizes and thereby implies that font means ‘typeface’, it speaks with the authority necessary for having its usage acknowledged in a public codification of English words and their

8 Nigel Love meanings. This is by no means the only possible reaction. It could be treated as a simple mistake – a solecism not to be countenanced by a reputable dictionary. In which case there has been no semantic change, and a whole generation of English speakers has been woefully misled. This conclusion would of course be powerfully assisted, not to say made definitively true, if when the next version of Microsoft Word comes along we find that its authors, perhaps having encountered some such discussion as this, have solemnly taken a decision systematically to replace font with typeface. Meanwhile, and absent such decisive metalinguistic intervention, we are left wondering just when and on what grounds to declare that a change has ‘occurred’. A third assumption is that font did indeed mean what the OED says it meant, up at least until when Mr Gates and his confrères started to mess around with it. One question that arises here, put in terms of a much-discussed dichotomy, is whether or how far a lexicographer’s statement of the meaning of a word is supposed to be descriptive or prescriptive of usage.8 It is not impossible that there have always been reputable English-speakers for whom, and irrespective of whether one thinks there are rights and wrongs in the matter, font just did or does mean ‘typeface’.9 If this is so, then the amendment required of a descriptive dictionary is the acknowledgement not of a new sense of the word but of failure to record an old one. But stating the third assumption in this way itself implies a fourth, which is that words ‘have’ one or more definitely statable meanings. The fact is, though, that alongside those for whom, erroneously or not, font may have meant ‘typeface’ long before the invention of word-processors and the associated printing technology, there will be those, especially outside the world of old-style printing, for whom font, albeit in some sense or degree known to them as an English word, never had any clear or precise meaning at all.10 (These, doubtless, are disreputable speakers, but that is a kind of judgement unknown to modern linguistic science.) Most people’s vocabularies have more or less extensive penumbral areas in which there dwell words about whose meaning and usage they are to varying degrees vague.11 Our acquaintance with what the dictionary codifies as the words of our language is not all or nothing: it runs the whole gamut from knowledge of the precise wording of the dictionary definition to sheer oblivious ignorance. Positions in between these end points of the scale are especially interesting in the present context, since they are likely to be a potent source of lexical change. Perhaps the originators of the change in question knew that font meant something to do with the letterforms of printed text, and found it handy for pressing into service as, in effect, a synonym of typeface. Or perhaps they knew perfectly well that it meant ‘a complete set or assortment of type of a particular face and size’, but preferred to use font anyway. It is, after all, a shorter word than typeface. The issue is what account we can give of the role of (a) the language-users whose usage ushered in the change in question; (b) the languageusers whose widespread adoption of the new usage is what licenses, if anything does, recognising a ‘change’ at all.

Language, history and Language and History

9

Here we come upon the question that has for many years been acknowledged as central to the theoretical debate in historical linguistics: how do we bring the behaviour of language-users into a discourse about linguistic history in terms of which languages and their component parts are treated essentially as if they were agents of their own temporal evolution?12 Current descriptions of English would have it that key, to vary the example, is one of a large number of nouns that can be used as premodifiers. And we are familiar with expressions such as a key idea or the key issue. Latterly – very recently, perhaps – key seems to have made a significant move towards becoming a fully fledged adjective, in being used predicatively (this problem is key), with intensifiers (that notion is really key), and as the head of an adjectival phrase (key to a solution here is . . .). It is likely that there are many speakers on whose ears such expressions fall as would this station is railway or the vicar’s collar is unusually dog. On the other hand, a pundit might be inclined here and now to announce a syntactic change. But note two salient points about any such pronouncement. First, its inherently normative nature; second, the fact that it will typically be based on no information whatever about the linguistic consciousness of the speakers in respect of whose usage it is proffered. The fact that we have no reason to believe that speakers of a language share the same linguistic consciousness is the most important obstacle to pinpointing the origin of linguistic changes in communicational interaction. A linguistic change is the result of metalinguistic judgements based on linguistic reflection. The ‘actuation problem’ that so troubles sociohistorical linguists13 is insoluble because a linguistic change requires not just an event, or even many events, but an interpretation. If culturally powerful enough the interpreter may get his interpretation written into the record. If not, not. In any case, it is not clear in what circumstances any particular speaker is obliged to pay attention to what the record says. A well known piège à cons for teachers of historical linguistics is the English verb weave. One will not infrequently find the form weaved, as in the car weaved through the traffic, cited as illustrating the occasional failure of morphological irregularities to carry across to figuratively extended senses of words (contrast he wove the fabric, where the verb in its literal sense retains the strong past tense form). But according to the ‘official’ history of English this analysis is based on a sort of professional folk etymology: there were historically two different verbs, one of which always had a weak past tense (as in he ducked and weaved ), and that is the one found in the traffic context. But is it? Suppose someone simply announces that when he says ‘the car weaved through the traffic’ he is, as a matter of ‘fact’, using an analogically levelled past tense form of weave/wove. What might in principle show him to be wrong about that? Here again we see the impossibility of reconciling a would-be descriptive but in reality essentially normative statement of the ‘facts’ with the unbounded fluidity of the relationships different speakers may contract with their language. The point of adverting to just a few of the plethora of queries that arise in connection with the hypothetical linguist’s story about font is not, in a sense,

10 Nigel Love to challenge the story itself. As far as it goes, and on its own terms, it is unexceptional. It is just the kind of story historical linguists are wont to tell about simple lexical developments of this kind, granted that they mostly discuss changes sufficiently far in the past to be treated as definitely accomplished. Taking a change either very recent or still in progress has the advantage of more readily revealing the nature of such stories. For, although unexceptional, it is far from unexceptionable. On the contrary, precisely because it draws attention to a currently live if trivial issue in English semantics, some may be expected to take exception to it. Or at least to treat it as open for discussion and debate. Some may even resolve to see to it that the alleged change is reversed or at any rate halted in its tracks. The starting point for the story is the observation of a mismatch between what an extant codification has to say about the meaning of a word and how the word seems actually to be used. The as yet uncodified new usage is so prevalent as to warrant, the linguist urges, public acknowledgement of a change. Far from being a scientist’s dispassionate statement of a fact or set of facts of some kind, such a story constitutes a move in a metalinguistic discourse which, if successful, will have a variety of social implications. Exactly the same once applied to comparable stories about established changes, albeit that their sociopolitical significance will in many cases have expired or been exhausted by the passage of time. On an account of this kind, a linguistic change just is a change in a codification. What has changed, or what the hypothetical linguist is proposing ought to be changed, is a dictionary entry. This is especially clear when the change in question is the complete loss of some item. Keller (1994) offers an interesting suggestion as to how, in general, to think about linguistic changes. Keller challenges the assumption – as old as Western thought itself, he suggests – that phenomena can be categorised in terms of a dichotomy of the natural (sunsets, earthquakes, flowers . . .) on the one hand, vs human artefacts (coins, computers, candlesticks . . .) on the other. For there is a category of things that straddle the boundary between the two: phenomena that arise ‘naturally’, but as an unintended consequence of concerted action, directed to other ends, on the part of human beings. Keller is not of course the first to identify such a category, which he dubs ‘phenomena of the third kind’, but, he claims, its relevance to linguistics has not previously been explored. In Keller’s view it is the category into which linguistic changes fall. As a simple non-linguistic example of a phenomenon of the third kind, Keller describes what road-traffic analysts apparently call the traffic jam ‘out of nowhere’. Imagine a steady stream of cars proceeding along a highway at roughly the same speed with ample separation between them. For some reason somewhere in the stream driver A slows down. In response driver B immediately behind applies his brakes too. To be sure of avoiding collision B slows down slightly more than would be required to maintain his distance from A. Behind B, driver C reacts in the same way . . . Each successive car ends

Language, history and Language and History

11

up travelling somewhat slower than the one immediately ahead, with the result that drivers L, M, N . . . come to a complete halt. The ensuing traffic jam is an artefact, but it has not been consciously or intentionally created: it is the unlooked-for outcome of many individual decisions to proceed with due prudence and caution in the light of the behaviour of the driver in front. The jam has been brought about by what Keller follows Adam Smith and (more recently) Robert Nozick in calling an ‘invisible hand’. Keller’s most detailed linguistic example concerns the fate of the German adjective englisch ‘angelic’ (⬍ Engel ‘angel’). Around the middle of the nineteenth century this word started to become infrequent as part of the vocabulary of the German-speaking community. This happened because it was no longer learned by German-acquiring children. It was no longer learned because those who knew the word tended to avoid it and to use alternatives such as engelhaft instead. They avoided englisch because they did not want to risk being misunderstood. That risk existed because englisch ‘angelic’ (call it englisch1) was homophonous with englisch2 ‘English’, and because the meanings of englisch1 and englisch2 were such that in almost any utterance containing englisch1 this word could be understood as englisch2 without making the utterance nonsensical. The opposite was not true: in most contexts englisch2 could not be interpreted as englisch1 without rendering the proposition senseless (bizarre, improbable, etc.). This being so, the use of englisch2 was much less likely to be misinterpreted. Because englisch1 was more of a hazard to successful communication than englisch2, and because there was a readily available alternative to englisch1 (engelhaft), speakers came to use this alternative instead. The clash of homonyms became critical when it did, says Keller, because cultural circumstances prevailing at the time led to an increased need for an use of words meaning ‘English’ and ‘angelic’. The important point about this explanation, from Keller’s point of view, is that there is no reason to suppose that any one individual deliberately intended to bring about the demise of englisch1. Speakers were simply trying to make themselves understood and to avoid being misunderstood: in the given situation these desiderata led them to avoid using englisch1, and this avoidance eventually had the unsought consequence that englisch1 disappeared, as if removed by an invisible hand. It is clear from this explanation that the workings of the invisible hand, as Keller presents them, are relative to a particular codification of the German vocabulary. Nobody seems to say ‘englisch’ any more to mean ‘angelic’, and there is no reason to dispute the broad outline of Keller’s account of why not. But what changed, exactly? Apparently, the willingness of German-speakers to tolerate the risk of saying ‘englisch’ and leaving their hearers doubtful whether they meant ‘angelic’ or ‘English’. This took a plunge in the midnineteenth century. But a change of tolerance level is not a linguistic change. The linguistic change only comes about when someone interprets the scarcity of new instances of ‘englisch’ ⫽ ‘angelic’ as showing that the word englisch1 has ceased to exist. What does that mean? It just means that the lexicographer

12 Nigel Love has been persuaded either to remove a dictionary entry or to adorn it with an obelisk. R.I.P. Moreover, stories of this kind are incompatible with a macrohistorical perspective in terms of which englisch1 exemplifies a fully productive word-formation process of which this particular instance happens to have fallen out of current use, and font, for all its formal and semantic vagaries, remains the ‘same’ linguistic item over whatever span of time etymologists are able to give a plausible account of it. There is a gross conceptual discontinuity here. The key to the enigma of linguistic objects that change while remaining the same is that the objects that are held to change are not the objects that are held to remain the same. Microhistorical changes are established in relation to synchronic codifications irrelevant to the macrohistory. The macrohistory deals in entities invisible to the microhistory. The rationale for the synchronic codifications is at best unclear, while the macrohistorical entities are artefacts of a dubious metaphysics. In short, historical linguistics as currently understood and practised is a conceptual and theoretical disaster area. Not only has it no viable foundations, it cannot even muster a modicum of internal coherence. But one point emerges clearly enough: the ‘internal’ history of a language is a metafiction constructed from diachronic projections of the fiction that lies at the heart of synchronic descriptive linguistics, namely that utterances instantiate enduring linguistic abstractions (‘short a’, feather, font, weave/wove vs weave/ weaved . . .). The difference between synchronic and any kind of historical linguistics is that the former is in practice concerned, as Saussure put it (or, given his theoretical context, was forced to concede), with ‘un espace de temps plus ou moins long pendant lequel la somme des modifications survenues est minime’14 – i.e., during which the temporal dimension can be ignored. The difference between macro- and microhistorical linguistics is essentially a matter of how long particular abstractions are held to endure. And since there can be no objective or indeed rational answer to this question, the result is an indefinite multiplicity of answers. Font and fondue are different words, perhaps several different words. On the other hand, they are one and the same word. Take your pick. Alternatively, overthrow the whole conceptual structure that gives rise to such quandaries. Meanwhile, let us observe that languages clearly inhabit a domain where there is no viable distinction to be drawn between history and historiography. A language’s history just is constituted by what its historiographers say. This is perhaps especially obvious where the historiography is patently in thrall to an ‘external’ sociopolitical agenda. Robert McColl Millar starts his discussion of the work of Antoine Meillet with the observation that in his Aperçu d’une histoire de la langue grecque (1913), which has been described as ‘arguably the first true “history” of the Greek language’, Meillet was in fact ‘developing by implication a theoretical position on the relationship between language, history and society’. That theoretical position becomes clearer when one turns to his later work Les langues dans l’Europe nouvelle, first

Language, history and Language and History

13

published in 1918 and then in a considerably revised edition in 1928. It amounts to proposing that the history of any given ‘language’ should be written in such a way as to subserve the requirements of a taxonomy of however many socially stratified linguistic varieties need to be recognised in the light of the sociolinguistic situation of the speech community or nation in question. Such a taxonomy is entirely the invention of the historiographer. At the bottom of Meillet’s pile are what he calls parlers, which owe their lowly status to the fact that they lack any connection with ‘civilisation’. Somewhere in between, among other distinguishable linguistic kinds, are dialectes, which may be ‘capable of literature in their own written form, but will not be used either as the language of administration or the language of intellectual debate’. In the case of some privileged communities their highest variety will earn the accolade of being dubbed a grande langue de civilisation, of which for Meillet the most illustrious example is French, as codified largely in the seventeenth century by the agencies of an authoritarian state. Meillet’s interest for Millar lies in the fact that unlike many predecessors he did not simply equate languages with grandes langues de civilisation, but acknowledged that where language is concerned there is a ‘dynamic relationship between historical processes and social patterns’, such that the connections and interactions between stratified varieties should be taken into account. Meillet may therefore be seen as a transitional figure ‘between the previous authoritarian and prescriptive discourse of the French linguistic tradition and a newer sociology of language which embraced diversity and at times actively encouraged, and engaged in, language development’. Meillet’s work overtly (and prescientifically, some contemporary linguists would scoffingly say) attempts to fuse the ‘internal’ history of languages with ‘external’ considerations of their cultural status. The fact that such a project should even be envisaged is an interesting comment on the epistemology of linguistic historiography. In his chapter in this volume John Joseph identifies the most salient driving force behind such projects: nationalism. His discussion focuses particularly on the debate among theorists of the modern concept of national identity between those who are ‘essentialist’ and those who are ‘constructivist’ about language. Both sides are agreed on the centrality of the idea of a national language: the question is whether, as essentialists would have it, the national language ‘can be taken for granted as what members of a nation share and what the rest of nationalism can be constructed on’, or whether ‘a national language is itself constructed as part of the nationalist project’. Within sociopolitical discourse constructivism about national languages, i.e., authoritatively codified ‘standard’ languages disseminated through formal education, is clearly a more sophisticated position than essentialism in so far as it recognises that the provenance of such entities cannot be ignored if one is to understand what they are and how they function. And constructivism about national languages is the orthodoxy within modern linguistics, which has long made a fetish of disparaging the ‘prescriptivism’ involved in their construction15 and in fact sometimes affects to ignore them

14

Nigel Love

altogether on the ground that, being constructed, they cannot be objects of natural-scientific study. Whether or when the next step will be taken – of grasping that this applies to all languages merely in virtue of the linguist’s having identified them as such – remains to be seen. If a language’s history is determined by its historiography, is there any object of historical inquiry of which this is not true? In ‘How to make history with words’, Roy Harris conflates history and historiography in respect of history itself. He argues that the activity of historiographers is itself a way of making history. This is in one sense a lame truism – something one might be especially inclined to say in a case where the historiography is famous and widely read. That is not what Harris intends. Nor is it that he rejects the usual distinction between history as what happens in the past and history (historiography) as what gets written by historians. Rather, he asks whether that distinction is properly construed as a difference between two meanings of the word history. His point is that the historian does not merely report or interpret objectively given facts or events. Historiography is a languagebound enterprise, and facts and events, for the historian, are from the outset constituted by the language in which they are formulated. What is more, any such formulation undergoes an incessant process of contextualisation and recontextualisation as it is recycled through the minds of readers and the works of subsequent writers. There is no anchorage, says Harris, for the notion that we can ‘identify the established facts of history’ and somehow hold them constant through the vicissitudes of interpretation. We have nothing to hold them constant in except linguistic formulations, and linguistic formulations offer no guarantee of context-neutral stability. There is no such extralinguistic thing as ‘what really happened’, not just because to ask ‘what really happened?’ is to call for a linguistic answer, but also because there are at least as many answers as there are contextually determined understandings of the question. Or, to put it another way, the fact that some codification of English may imply that two instances of ‘what really happened?’ are instances of the same expression – the sentence what really happened? – by no means guarantees that the same question has been asked. Thus do problems in the conceptualisation of language link up and mesh with problems in the philosophy of history. While sceptical of the general proposition that no historical event can be identified independently of some verbal formulation, Talbot Taylor nonetheless acknowledges, in ‘Talking about what happened’, that verbal formulations may in a number of ways be crucially implicated in history. In the first place, ‘historical events very often involve language’. One might go further and say that historical events may sometimes be largely or wholly constituted by language. A recent notorious example concerned the death by suicide of a British civil servant in circumstances that threatened to cast an ugly light on the decision-making process leading to British involvement in the invasion of Iraq in March 2003. The ensuing judicial inquiry16 made it clear that from start to finish the whole matter was about language: about the rights and

Language, history and Language and History

15

wrongs of saying or refraining from saying what, when and to whom.17 Taylor’s focus, however, is on historical events whose occurrence is typically reported in utterances containing verba dicendi – verbs such as promise, threaten, mention, warn and their numerous like, which not only indicate that something has been said but also categorise the act performed in saying it. Such verbs, says Taylor, are responsible for constructing, at least in part, the cultural significance of the historical events reported by using them; and this is a circumstance historiographers ought to take account of, especially if, as Taylor argues, the import of a particular verbum dicendi, even where it happens to have established translation equivalents in different languages, cannot be taken as a cultural universal. Taylor does not discuss the interesting question how verbs of saying report speech acts. It seems that in very many languages the most general verb of saying (in English, say) may indifferently take as object direct speech, indirect speech, or a mere paraphrase of what was said, as in he said “the vicar’s collar is unusually dog” vs he said that the vicar’s collar is unusually dog vs he said that parson sports a preternaturally ecclesiastical neckpiece. This consideration may have a bearing, one way or another, on Taylorian scepticism about the total language-dependence of history, in as much as any culture whose language has a verb of saying that works in this way is clearly committed by its most fundamental machinery for making metalinguistic statements to recognising the contextual equivalence of different verbal formulations. But this is not the place to pursue this matter further. Michael Toolan, for his part, is sceptical of the proposition that the history of a verbal formulation can be clearly distinguished from the history of what it is used to talk about. In ‘Part of the meaning/history of euro: integrational corpus linguistics’, Toolan discusses this and in so doing raises, overtly or by implication, a number of relevant issues. In the first place, euro poses some striking problems of wordhood and wordidentity (cf. font, fount and font/fount), of which the most salient is that its use in connection with a new pan-European currency seems quite overtly intended to challenge, in a way that goes well beyond the commonplace conundrum presented by a case like Anglo-French fondue, the widely held notion that a given word belongs to a given language. (Is fondue a French word sometimes used as such by English-speakers? Or is it a French word that has been ‘borrowed’ into English, and thus cloned itself?) As regards euro, the idea seems to have been to come up with a form that would transcend petty nationalisms and make itself equally at home in the languages of all or most participating nations.18 But the descriptive linguist is nonetheless left asking whether there is one word euro, adopted by many different languages, or as many different words of that written form as there are languages that have adopted it. The modern possibility of the instant pan-global lexical prescriptivism that has ushered euro into existence casts a glaring light on the conceptual deficiencies of a metaphysics of language that obliges this to be treated as a serious question. Notice that the act of lexical prescription itself does not answer it: the prescribers can be presumed to be neutral as to what exactly linguists are to

16

Nigel Love

make of the situation thus brought about. And, precisely because euro has been imposed by a sort of lexicographical fiat on everyone at once, it raises in a rather obvious form, at the level of the individual language as conventionally understood, a question that ultimately arises for the individual speaker – namely, even if for the sake of argument one entertains the possible identification of a single international linguistic entity called ‘the word euro’, how far is it in fact ‘the same’ in the various different tongues that have adopted it?19 Won’t that depend on the lexical or lexicogrammatical company it finds itself keeping in different cases? We can perhaps overlook as of strictly minority import the fact that euro was already established in the speech of at least some English-speakers as the name of a species of kangaroo, even if that may tend to colour their apprehension of the word for the currency. Of more general significance is that in languages where, as in lexicogrammatical, -o- routinely combines one classical root with another, euro seems (designedly?) to be a truncation of some longer word that has, bafflingly, never been established. Euro-what, exactly? In others this question will not arise. Toolan focuses on a semantic issue of this sort: in many (British) English texts euro tends to collocate with words reflecting the fact that a very frequent context for its use is discussion of the vexed question whether or when the United Kingdom should adopt the currency in question. In English, therefore, although presumably not or not so strikingly in some other languages, the word is for now as much the name of a political problem as it is of a kind of money. That, of course, is because of the history of the euro and of specifically British attitudes to it. So is there ultimately any distinction to be drawn between the history of the euro and the history of euro?20 Toolan’s scepticism takes the form of calling in question the formulation ‘history of the euro’. For there is no ‘euro’ for there to be a history of, except what we can at different times understand to be the focus of a multiplicity of different uses of however many different linguistic entities we deem there to be that might be designated by the written form euro. Doubtless to the chagrin of those contributors who do not like the result, this introduction has, inevitably, treated their texts as subject to historical interpretation21 and recontextualisation in the light of the larger text in which they are embedded. Their sense of what has thereby been done to ‘what they wrote’ may be seen as a test case for one of the central theses of this collection as a whole. That thesis has it that there is no what that they wrote, over and above what subsequent interpreters, including themselves, may make of it on future occasions of interpretation. A language is not a device for encoding systematically recoverable semantic content, for one reason because there is no way such a ‘device’ could have been engineered, and for another because any such ‘content’ would either be literally ineffable or else itself linguistic in nature, and hence subject to decoding of its content, and so on in an infinite regress. And if a language is not such a device, then nor is the language of history. And if the language of history is not such a device, there is no support for

Language, history and Language and History

17

the view that historians’ statements, any more than anyone else’s, stand in an unchanging relationship to states of affairs that can somehow be established extra-linguistically. That language should be incapable of providing any such service to statement-makers can be seen from the difficulty of setting up a conceptually unified platform from which to make statements about the history of languages. The problem is that if languages are to have histories they and their component parts have to be thought of as objects of some kind, whose transformation through time constitutes the subject matter of the history. But because they are not objects of any kind, historians have to decide for themselves, unconstrained by any empirical facts of the matter, on the particular reifications they propose to entertain for their particular purposes, whether those purposes be pedagogical, political, pseudo-scientific or anything else. Setting this point in a broader context, Geoffrey Harpham (this volume) suggests that the history of linguistic thought itself ‘can in fact be considered as a series of attempts to determine the essentially linguistic, and with each new determination a new object is brought into focus’. Harpham discusses why such attempts should be so unremittingly made. His answer is that ‘the significance of language lies in the prestige it has long enjoyed as the defining characteristic of the human’; and he spotlights the trenchant resistance among some contemporary linguists to the linked ideas (i) that language as a human faculty should in principle be made to yield to evolutionary explanation and (ii) that other primates can be taught to use language. Thus is even the prehistory of language and languages distorted in the interests of an ideological programme. What non-human apes are very likely incapable of is engaging in the metalinguistic reflection that would allow them systematically to decontextualise the signs they use and by treating those signs as objects of intellectual contemplation and inquiry embark on the process of constructing what human apes know as languages and as histories of languages. One of the projects of integrational linguistics is to show how such second-order processes are founded in and arise out of our first-order linguistic experience. Especially relevant to that project is Stephen Cowley’s contribution to this volume. In ‘Bridges to history’ Cowley outlines a way of understanding how a child ‘comes into history’, i.e., enters encultured language-mediated life, and concludes that ‘language development consists primarily in taking steps towards acting in accordance with the belief that talk consists in acts whose sound structure enables us to achieve goals’. Cowley’s argument is that coming so to act may be explained without invoking the acquisition (before overt metalinguistic indoctrination starts to be imposed) of definite linguistic objects of any kind, let alone postulating genetically built-in equipment for acquiring a language. Accounts of how language is possible without a language will have their own intrinsic significance. But they may also point the way to grasping the role of language in constructing languages, the histories of languages, and history itself.

18 Nigel Love

Notes 1 That is not to say that particular language historians have not come up with their own definitions. See, for instance, Millar’s discussion of Meillet (this volume). 2 See Love (2002). 3 This is the crucial move behind the ‘strong tendency to view [a] language as an object (or an inventory of objects such as lexemes) rather than as verbal communicative behaviour’, discussed in this volume by Michael Walrod. On this point Geoffrey Harpham (this volume) identifies Saussure as the key historical figure in modern times, with his striking theorisation of the notion that ‘[a] language has a determinate nature as a concrete object’. 4 www.oed.com 5 As are roman vs italic and bold vs non-bold. 6 The term is used in the sense of Harris (1996); see especially ch. 12. 7 They are variant forms (‘strong’ and ‘weak’ respectively) of the past participle of the French verb fondre. 8 For recent discussion see, for example, Taylor (1990a,b). 9 Cf. Weigand’s comment (this volume) on the statement that Italian afferrare has undergone a metaphorical change of meaning from ‘grip, grasp’ to ‘understand’. The problem with this is that both meanings appear to exist side by side in the contemporary language. 10 See Weigand (this volume) for discussion of the need to base any viable account of semantic change on a recognition of synchronic semantic indeterminacy. 11 Walrod (this volume) doesn’t know what chick flick means. Do you? 12 Deumert (2003) is a comprehensive overview of this debate. 13 Milroy is an example; see e.g. Milroy (1992, pp. 169ff .). 14 Saussure (1922, p. 142). 15 See Kilpert (2002) for an interesting recent discussion of the role of prescriptivism in linguistic theorising. 16 See www.the-hutton-inquiry.org.uk 17 For amplification of this point see Love (2004). 18 Greek requires it to be transliterated. The orthographic rules of Welsh would prescribe ewro. Readers of a certain cast of mind may perhaps care to draw up a longer list of such caveats for themselves. 19 It is a moot point whether that doesn’t mean all languages whose speakers ever have occasion to refer to the currency in question. The French may still call pounds livres, just as they call London Londres, but in general the idea that one might devise one’s own name for a foreign country’s money seems scarcely more permissible nowadays than resisting the requirement that Ceylon should cede to Sri Lanka. 20 Cf. Austin’s question (Austin 1979 [1961]) whether ‘what is the meaning of the word rat?’ differs in any significant particular from ‘what is a rat?’ A whole philosophy of language, as reflected in modern linguistic semantics, has been founded on supposing that it does. 21 Given the late appearance of the volume some might think ‘historical’ here should be in bold italics.

2

The end of linear narrative? Reflections on the historiography of English Roger Lass

Erst kommt das Fressen, dann kommt die Moral. (Bertholt Brecht, Die Dreigroschenoper)

Introduction This is a somewhat paradoxical chapter from one who has spent the past three decades as a self-characterised ‘historian of English’, much of whose work has been fairly traditional narrative history (e.g. Lass 1992a, 1999a) or reflections on its construction (Lass 1997a). The genre recalls that of the ‘retraction’ at the end of the Canterbury Tales: it is to be taken with precisely that mixture of total seriousness and ambivalence appropriate to Chaucer. It is obvious that the ideal data-base for writing the history of a language would be something like (1) below – a set of attested ‘dialects’ (D1–3),1 each of which survives as a chronological sequence of text-corpora (t1–t4). The diagram then defines the geographical region over which the language is spoken and the temporal succession of its varieties. In (1) we have a continuous regional and temporal coverage.2 (1) t 1 t2 t3 t4 D1

D2

D3

20 Roger Lass Unfortunately, the survival-pattern of both Old English and earlier Middle English texts does not possess this ideal shape; it is often rather more like (2): (2) t 1 t2 t3 t4 D1

D2

D3

That is, some regional varieties are attested early and have a partial chronological sequence; some appear only at one period and apparently leave no direct descendants; and others appear in sequence, but late, without apparent ancestry. This means that at no point in the history is there a full t1–t4 sequence for any region. Any attempt at ‘vertical’ history becomes de facto ‘diagonal’ when the attestation is skewed or defective. This property of the history of English tends to be swept under the rug, or more charitably simply not noticed. What textbooks usually do is present apparent straight-line historical sequences which when unpacked may really be saying ‘some feature X in D1 at t1–2 “becomes” Y in D3 at t3–4’. This may be harmless if we know it is being done; but it is certainly something we should be aware of. Here is a toy model of the kind of conceptual problems this skewed attestation presents. Comparative and other evidence suggests that the history of the masculine/neuter a-stem genitive singular ending goes like this: *-os- ⬎ *-as ⬎ -æs ⬎ -es. The second-last stage is attested in Old English, particularly in eighth-century Northumbrian texts, and there is virtually no attestation outside of this dialect-cluster. However the ‘later’ form is richly attested, particularly in West Saxon. Below are four versions of the lines 1a–2a of a poem (usually called ‘Cædmon’s Hymn’) that appears in versions of Bede’s Historia ecclesiastica, either added to Latin texts or as part of Old English translations. The versions cover a period of some three centuries and are in two kinds of language: the first two Northumbrian, the second West Saxon. (3) (a) Northumbrian ii. Cambridge University Library, MS Kk.v.16 (M): 737 nu scylun hergan hefaenric-aes uard metud-æs maecti . . .3 ii. Public Library Leningrad, MS Lat. Q.v.I.18: mid 8th century nu scilun herga hefenric-æs uard metud-æs mehti . . .

Reflections on the historiography of English 21 (b) West Saxon ii. Oxford, Bodleian Library MS Tanner 10: 10th c. Nu sculon herigean heofonric-es weard metod-es mihte ii. Oxford, Corpus Christi Library MS 279: late 10th-early 11th c. nu we sculan herian heofonric-es weard metod-es mihte The Northumbrian versions also show other clearly ‘archaic’ features, for example, the verbal past plural ending -un and the unlowered -i in the word for ‘might’.4 The historical and comparative evidence tells us that these are the forms to expect in early texts, just as -on and -e are late developments. It appears therefore that the first group of texts shows forms ‘ancestral’ to those of the second group; but this is not the case at all. Because they come from a different language-cluster, they are merely cognate, no more ancestral in a true genealogical sense than their equivalents in Gothic or Old High German would be. Yet since they are archaic looking and actually older, and occur within a text-complex traditionally called ‘Old English’, the handbooks more or less allow the innocent to believe that archaic -un, -i, -æs in some sense ‘turn into’ the later -on, -e, -es. Of course, they almost certainly did in the ancestor of West Saxon, but the evidence does not survive. If D1 is Northumbrian and D3 is West Saxon, then the ‘change’ or ‘evolution’ of these forms is in fact being plotted diagonally on the pseudo-geographical map, and projected to a manufactured time-line which as well as removing the gaps, expunges the diagonality and appears ‘vertical’. Typology is silently transmuted into history. The results of the operation can be represented this way: (4) t 1 t2 t3 t4 {D1 + D2 + D3} = ‘History of OE’

Prima facie, two languages that do not share a common chain of transmission or have not been in extensive contact amounting to bilingualism cannot share a common history. No form in eighth-century Northumbrian language can be part of the history of a tenth-century West Saxon form, regardless of apparent ‘directionality’. This is because the main criterion for common history – lineality – is not met. Yet there is a strong (and not entirely unjustified) intuition that it is safe to talk about such phenomena as if they were ‘histories’; the problem is that the terms of discourse have been at best

22 Roger Lass dubious, and the phenomenology has been given an erroneous interpretation or at least limited interpretation. And this raises some difficult epistemological problems, which do not as yet have satisfactory solutions.

Factitious simplification As the last section suggested, reification is a potent temptation for historians: from a collection of text languages (Fleischman 2000), that is, utterances of geographically and temporally dispersed scribal witnesses, we extrapolate ‘a language’. And this allows us to organise witnesses which may have no relation except cognacy and similarity because of common ancestry (and in some instances contact) into pseudo-lineages, gathering these into one great superlineage which in our case we call ‘English’. And we tend to think, partly because of what we have done with the data, that such a lineage describes a ‘progressive’ or at least directional trajectory in time: the end-point is some kind of ‘standard’ language, or ‘English in general’, embodied primarily in the very language we use for writing our histories (cf. Milroy 2000). This error dogs the historiography of English from Ellis and Sweet to the present. I cannot think of one major history of the language that does not take this point of view:5 and indeed all one has to do is look at a set of familiar titles to see that this is what has been happening: Historische Grammatik der englischen Sprache, die englische Sprache, Growth and structure of the English language . . . In a sense, as we will see, it may have been procedurally necessary or at least almost unavoidable for this to have happened, but that does not mean that we can ignore it. There are four ill-formed concepts that play a dominant role in the historiography of English: Old English, Middle English, Early Modern English, and Modern English. That there are complex objects loosely corresponding to these terms is patent; but that they can be used plainly the way they are as the names of sequential ‘unitary’ language-states, without immense qualification, is not. As typological concepts they have some validity (see the rather queasy admission of this in Lass 2000 and the discussion below); as names of data-sources they do not, at least in the usual praxis. To begin the main argument, it is not quite proper at any point to say ‘in Middle English C was the case’, or ‘Old English X was [X]’. The first problem is this. The further back we go in time, the more invariant and simpler language states apparently become. Not just in the obvious practical sense that information becomes scarcer, but procedurally as well. From the comparativist fiction of the ‘dialect-free protolanguage’ to reifications like ‘English’, historians tend to adopt a manner of speaking in which the collection of materials whose ancestry we are reconstructing, no matter how complex or variable, always at some remove appears to go back to a single ancestral onbject, whose ‘story’ we then find ourselves telling. To take a phonological example: it would for instance seem unexceptionable to most historians of English to say that the vowels in ModE sap, cat, back ‘go back to ME a’ (mediated of course by the appropriate sound changes).6

Reflections on the historiography of English 23 There are good procedural reasons for talking this way; but we must understand what we are not saying. Except as a technical device, and within the framework of a particular kind of strategy of historical reduction, there is no such thing as either ‘ME a’ or ‘the vowels in ModE sap, cat, back’. To start with a paradox, the items whose mutual mappings we think we have established in some detail as the basis of etymological praxis are largely instrumental fictions. (But how can you tell a story without characters whose actions can be narrated?) It is important to make clear that characterisations of etymological categories, whether as italicised representations like ‘OE short a’ or apparent phonetic transcriptions like ‘OE [a]’ are necessarily coarse and imprecise. In any sophisticated sense, ‘OE short a (as in catt)’ means as much or as little as ‘ModE short a (as in cat)’. We speak as if we are dealing with single values, but of course we cannot be. Even just within modern England ‘short a’ covers a range from low central [ ] to low front [a] up through various [æ]-like and [␧]-like vowels; if we were to add South Africa and New Zealand to the set we would get values as high as [e], and parts of Scotland and Northern Ireland would give us low back [a] as well. Still, there is a categorical and lexical sense in which we can talk sensibly of ‘short a words’, and this is what (regardless of what we claim) we really must intend by such representations, or by symbols in phonetic brackets. This is a crucial point, because judging from the way history is written, more historians than not probably believe that something like ‘ME a was [a]’, is a legitimate thing to say.7 It is not. Each category is a cluster of variants (whose members in this case are unfortunately not available for inspection); we therefore cannot sensibly talk of ‘ME a’ without the complex stipulations implied by the comments above; and this is the sense in which categorical representations in historical or etymological narratives ought to be construed. Therefore all historical representations given particular mnemonic shapes are, like historians’ ‘transcriptions’, to be taken in a typological rather than strictly ‘phonetic’ sense. To put this in rather simple terms that will define the rest of this discourse, an object like ‘ME a’ is not an ‘individual’ but the name of a population of variants, treated uniformly as a matter of convenience and apparent procedural necessity. This of course holds a fortiori for larger-scale abstractions, in particular whatever is designated by language names like ‘Middle English’. There is also of course a deficiency of information at the distal end of the sequence and a surplus at the proximal, which is why all historians perform ‘end-matching’ simplification. a

Cylinders and funnels The ‘verticalisation’ described in first section might be an appropriate image for the traditional processing of simple attestation patterns. But the complexity of typical historical data requires, exegetically, a greater graphic and conceptual depth. Historians naturally tend to visualise the history of the

24 Roger Lass language as something like a ‘tube’: this metaphor suggests that the historiographical act consists of pouring in forms at one end of a transparent graduated cylinder, and looking to see what occurs at the various graduation marks (the points at which the surviving texts allow us to intersect the ‘stream’ of history), and then finally seeing what comes out of the bottom. This cylinder is ‘the history of the language’, and the very act of pouring the data in at one end and collecting the residue at the other typically gives history a ‘direction’. But the pattern of attestation is far more unruly and illshaped: a fact often noted but typically misdescribed. Here are two characteristic misdescriptions, carefully chosen from my own work: if I am going to retract, I certainly ought to be my own target. This is especially true because up until recently I was a member in good standing of the tradition I now have serious problems with, and the ease with which one adopts the stances of one’s predecessors ( gigantes erant in terra . . .) is one of the issues here.8 These two examples move from the kind of static description discussed in the previous section to history proper: As early as the twelfth century we begin, in non-northern texts, to get a scattering of ⬍o⬎, ⬍oa⬎ spellings for OE /a:/ words. The sermon In diebus Dominicis . . . written on the southern/midland border in the twelfth century shows gast ‘spirit’, lauerd ‘lord’ . . . as well as on ‘one’ . . . and the variant louerd; the Peterborough Chronicle for 1134 has a few ⬍o⬎ forms . . . and the Worcester Fragments . . . have mainly ⬍o⬎ . . . These spellings reflect a major change in progress, in which OE /a:/ rounded and raised to /ɔ:/ . . . Despite its variable implementation in diffferent texts and different areas, we can date it as coming to fruition in the late twelfth to early thirteenth century, beginning in the south-east and spreading northwards . . . (Lass 1992a: 46f ) Like many major structural simplifications in English, gender loss began in the north; there are fluctuations such as the tenth-century Lindisfarne Gospels, where, for instance, endung ‘ending’ appears as masculine (abstract nouns in -Vng are typically feminine), and stmn ‘stone’ is both masculine and neuter. By around 1200 the old system is in considerable disrepair in most dialects, and except in Kent the shift to natural gender was pretty well complete by the end of the century. (Lass 1992a: 107) These interpretations postulate two single developments (raising and rounding, loss of grammatical gender). They are visualised in the traditional way as unitary processes diffusing through space, with time as the vector. To anticipate, it is just as possible to view things another way: multiple developments in space, each with its own temporal profile, with space as the primary descriptive parameter. In this case we would have convergence

Reflections on the historiography of English 25 (at least partial, depending on the amount of contact there is between varieties), rather than diffusion: just because something appears first in one place and later in another place, this does not have to mean that it ‘travelled’ between them. Convergent development is not the usual choice, but there is no a priori reason for ruling it out, and as I will argue some good reasons for ruling it in. To begin with, the quality and quantity of our data-base shift massively over time. And (cf. the preceding section) the less we know, the clearer the picture. Here is an iconic (informal, pseudoquantitative) representation of one aspect of our data-base from the earliest attestation of Old English to the present: regional coverage: (5)

8th century

20th

9th c e n t u r y 10th c e n t u r y 11th c e n t u r y 12th c e n t u r y th 13 c e n t u r y 14th c e n t u r y 15th c e n t u r y 16th c e n t u r y 17th c e n t u r y 18th c e n t u r y th 19 c e n t u r y c e n t u r

y

The expansions and contractions of available material might be called respectively ‘funnels’ and ‘bottlenecks’. Here the sixteenth to eighteenth-century bottleneck represents the effect of growing standardisation and supraregionalisation on written dialect attestation; the nineteenth to twentieth century expansions are enabled by the growth of dialectology and the increasing collection of regional texts in the mid-nineteenth-century, and the development of descriptive linguistics as a discipline; the twentieth-century expansion includes growth in these areas, plus the advent of sound recording at the end of the nineteenth century. Configurations like this appear on other parameters as well: here is the picture for diversity of genre, which has a rather differently problematic shape. As standardisation takes hold, regional representation decreases; but at the same time the number of genres represented increases enormously, with growing numbers of personal letters, diaries, and eventually advertisements and finally sound recordings and e-mails becoming

26 Roger Lass part of our data-base: (6)

8th century 9 century 10th c e n t u r y 11th c e n t u r y th c e n t u r y 12 r 13th c e n t u 14th c e n t u r c e n t u 15th c e n t u 16th 17th c e n t u 18th c e n t u 19th c e n t u c e n t th

20th

y y r r r

y y y r

y r

u

y r

y

It is clear that if we force configurations like these into ‘cylinders’, we do indeed get the appearance of straight-line narrative – but at the expense of fidelity to the complexity of the data. This praxis is narratively ‘clean’, but empirically suspect. It forces immense quantities of heterogeneous and variable data into simple linear progressions that become apparently ‘thematic’ (the ‘trend’ has been loss of inflection, etc.), but where mechanisms and the essential non-linearity and variability of many of the processes are simply lost, because the shape of the representation does not allow them to surface. If we pour a funnel-shaped mass of objects into a cylinder, and inspect only what appears on the outside, we have no way of telling to what extent the process of ‘pouring’ has squeezed the data so that material from the ‘middle’ appears on the outside – which is all we can read. The only thing that will emerge from this procedure is ‘trends’ (if there are any); but the individual lineages will be distorted or destroyed. Another fallacy that the image will encourage is reading the characters that appear in vertical arrays on the surface as being in effect mapped into one another, where in fact they may very well be temporally successive phenomena from different sources (cf. first section). To sharpen the image, imagine that each letter or number in one of the rows in the two diagrams above is a linguistic character, and consider the number of pseudo-mappings that could potentially occur on the surface of the cylinder because of the ‘squeezing’ of the funnels into the cylindrical default format, and the appearance of items from ‘inside’ on the narrative ‘outside’. Judging from my experience as student, teacher and scholar, what a typical history of English might pour into the cylinder is a collection containing a sequence like: Cædmon’s Hymn (one of the WS versions), an Alfredian sample, Beowulf, Ælfric, Peterborough Chronicle (Final Continuation), Ormulum, Layamon, Ancrene Wisse, The Proclamation of Henry III, The Owl and the Nightingale, The Ayenbite of Inwyt, the Gawain Poet, Piers Plowman, Chaucer,

Reflections on the historiography of English 27 Caxton, Spenser, Shakespeare . . . While of course provenances would be noted, there would always be at least an implicit claim that the forms and structures in these ‘monuments’ (or whatever else is chosen) represent a lineage of sorts: ‘our’ English somehow is rooted among the languages of these texts, and various characters can be ‘traced’ through the sequence (e.g. development of ‘the vowel system’, changes in case, gender, tense and aspect marking, word-order). I taught this way for decades, using textbooks (including my own) that did the same; it never occurred to me that I was implicitly making statements of a kind that I never would have made in full consciousness as it were: for example that the inflectional system of Ancrene Wisse is in some sense ‘ancestral’ to that of the language in which I was teaching. Put plainly, the notion that a thirteenth-century Herefordshire linguistic system could be ‘ancestral’ to a twentieth-century London-based one is absurd; but the traditional praxis allows us to be implicit, and hence not apparently guilty of an absurdity we would all probably disavow if we had to say it out loud.

Lineality and convergence If two lineages with common ancestry at some remove show characters or historical developments in common, this is not necessarily a function of that ancestry. Whether it is or not is an empirical question – if not necessarily a soluble one, given the contingencies of survival and the limitations of historiographical method. There is a crucially important distinction in the historical sciences: that between similarities due to lineal relationship and similarities due to convergence. The long canines and shearing cheek teeth of placental carnivores are a heritage of common ancestry: there was a primitive carnivore ancestor that had them. The same character suite in marsupial carnivores is a convergence (independent innovation): there is no common ancestor with these characters. The commonalities stem from parallel ecological pressures and variant selection. The dental similarities of dogs and cats derive from a common ancestor that had them; they are homologous. The similar dentition of marsupial carnivores is analogous: they share no common ancestor that had such teeth with placental carnivores. One relation is historical, the other is typological: common phenotypy does not imply common genotypy, either in the technical biological sense or the looser linguistic sense. In the historiography of a language with a complex regional and temporal attestation, it is vital to distinguish analogies from homologies. Failure to make this separation can be seen most clearly perhaps in statements of the kind cited above: at least the issue of possible analogy is not invoked, and the developments are assumed to take place in a single lineage. Or that appears to the careful outsider what is being said, and indeed as far as I know that was more or less (if unreflectively) what I meant.9 ‘Middle English’ was taken as a single character field, and the only mechanism that seemed conceivable

28 Roger Lass for representing the variation and historical flow was diffusion within a single macro-lineage. This mistake was (and to some extent still is) common in biology as well, at least in more popular interpretations of evolutionary change. Consider for instance the development of modern horses: at one time (and in the popular imagination still) it was taken to be something like an ascending ‘ladder’: the equid lineage is a chain of organisational grades with a single temporal vector, from small animals10 with multiple toes to the current large ones with a single toe formed from the fusion of the middle digits. ‘The horse’ got bigger and lost toes. But a more careful examination of the fossil record shows something quite different: size fluctuated over the development of modern horses, and there is no single directional lineage. Rather than a chain or ‘ladder’ of related forms moving in a particular direction there is a ‘bush’, a complex mass of lineages, some convergent, some evolving in other directions. Modern equids are not the outcome of a linear development, but just the surviving end-branch of one of the complex configurations in the bush (for discussion and graphics (see Gould 1991b). There is in fact no such thing – except as an artifact of selecting and tracing one direction out of a complex of meanderings – as ‘the horse’.11 It is as much a reification as ‘English’. In an evolutionary sequence, the more evidence we have, and the more carefully we examine it, the less likely it is to come out looking linear.12 The most typical ‘completions of trends’ are contingent last survivors of complexes of lineages, rather than ‘goals’ the trends were in some way ‘aiming at’. There are two types of character transmission, both in biology and linguistics: vertical and horizontal. In biology vertical transmission (parent to offspring via the germ line) is the unmarked case, as it is to a lesser degree in language. But horizontal transmission (lineage to lineage) is not uncommon. A timely example in biology is the transmission of genes for antibiotic resistance from one bacterial lineage (‘species’) to another.13 In language of course horizontal transmission is better known under the name of ‘contact’. Both kinds of character transmission can lead to similarities between lineages; conservative vertical transmission in lineages stemming from a common ancestor is perhaps the most familiar (e.g. Germanic languages have a strong/weak verb distinction because their common ancestor had it).14 But convergence is not uncommon, though it can be problematical. The situation is made more complex by the fact that both kinds of character transmission can participate in convergence. There are probably at least three kinds of convergence:15 1

Primary convergence. This is the result of similar vertical transmissions in non-contacting lineages (e.g. the preterite ⬎ perfect shift in Yiddish and Afrikaans mentioned in fn 14).

Reflections on the historiography of English 29 2

Secondary convergence. This is convergence in appearance only, the result of horizontal transmission. A linguistic example would be the apparently independent loss of prefixes in the Continental Scandinavian languages, which is probably a function of contact with Low German.

Given considerable historical depth it is often difficult to tell type 1 from type 2; and it is not always possible to tell convergence from plesiomorphy (common retention of primitive characters): the historical record and constraints on reconstructive technique may not allow it. This difficulty is greater in historical linguistics than in biology, because there is no ‘hard’ medium of transmission like DNA: plesiomorphies in biology can be untangled from independent vertical or horizontal transmission by examining genes, but in language there is no similar option, no complete ‘embodied’ record of history. 3

Mixed or tertiary convergence. Both types of transmission are involved. For example, a character may be horizontally transmitted and trigger a development in the receiving lineage, or two or more lineages engaged in similar developmental processes may reinforce each other through continuous or periodic horizontal transmission, but with different results in the sporadically contacting lineages. I will argue that this maybe the correct characterisation for a number of familiar phenomena, such as the loss of gender at different rates and indifferent ways in various early regional Englishes (see the next section). In a densely contacting field this cannot be ruled out as a major kind of change, and the complexity of some MS stemmata may be a result of precisely this combination of independent development and self-reinforcing contact.

Large-scale convergence has been recognised for a long time as a major phenomenon in language change. The most famous (or should it be infamous) recognition is one variety of what Sapir called ‘drift’. In an attempt to explicate – among other things – various commonalities between English and German he wrote (1921: 172): The general drift of a language has its depths. At the surface the current is relatively fast. In certain features dialects drift apart rapidly. By that very fact these features betray themselves as less fundamental to the genius of the language than the more slowly modifiable features in which the dialects keep together long after they have grown to be mutually alien forms of speech. But this is not all. The momentum of the more fundamental, the pre-dialectic, drift is often such that languages long disconnected will pass through the same or strikingly similar phases.

30 Roger Lass On this rather romantic view, with its hint of essentialist Naturphilosophie, the convergences are caused by ancient immanent properties, the ‘genius’ of the original language, actualised only later, under different ecological conditions. I am not on the whole averse to the invocation of unobservables – no science can do without them – but I find this notion explanatorily vacuous and mystical, and will not assume that any convergences are due to a ‘pre-dialectic drift’.16 The problem is really how to recognise convergence in related lineages, and distinguish it from homology or horizontal transmission. In principle linguistic convergence in the simple sense is no different from the series of stages that marsupial and placental carnivores went through independently in developing their teeth: the histories must have been quite similar. The difference is that in biological evolution you can often reverse-engineer the adaptive motivations for changes: in language I do not think you can (Lass 1980). But there are typological story-varieties which repeat: it is not uncommon for highly inflected languages to lose inflection, even without contact (though it is perhaps easier with it: see Dury 2001). And there may be an independent likelihood of certain kinds of change sequences occurring convergently, such as repeating sequences of case loss in typologically similar languages. At least judging from Indo-European developments, certain case syncretisms seem to be more likely than others at early stages: e.g. loss of the more ‘specific’ local cases like ablative and locative first, with the dative or genitive taking over local functions (see the discussion in Lass 1991). So some convergences may be the results of what one might call clines of typological likelihood – nothing mystical like Sapirian ‘drift’, but still problematical.

Why linear history may be a map of convergences: the loss of gender in English Species are abstractions, not realities – are like genera. Individuals are the only realities. Nature neither makes nor breaks molds – all is plastic, unfixed, transitional, progressive, or retrograde. (Charles Lyell, Notebooks: Wilson 1970: 121) It is well known that English once ‘had’ grammatical gender, and then ‘lost’ it. But it did not in fact (unsurprisingly) ‘lose gender’ tout court after the OE period, but did it piecemeal, and most interestingly, in different ways in different varieties, and over long stretches of time. I will shortly look at two early text languages, one of which has about 90 per cent grammatical gender, and another which has about 9 per cent, as polar examples.

Reflections on the historiography of English 31 Let us define grammatical gender as the existence of different noun classes, manifested not only (if at all) in the formal structure of nouns (different declensions), but crucially in concordial relations. Gender is, aside from its classificatory function, a ‘tracking’ device that ties together nouns and their satellites and anaphors. On these criteria there is no doubt that Old English had a ‘complete’ (if variable) gender system, very like that of modern standard German. There is also no doubt that relics of this system (or better these systems, since ‘Old English’ is no more a monolith than later English) persist through a good part of the Middle English period, in different forms, in different varieties, and that by around the end of the thirteenth century none show anything but relics of the original system(s), and by the end of the fourteenth century English is genderless. But this process came about in a series of complex and differing ways; the traditional straight-line development that has been standard wisdom since the nineteenth century did not in fact occur. To return to the terminology I introduced earlier, not all gender-loss sequences are homologues. Here is a typical example of the received view. Richard Morris (1882: §40) notes that in Old English, which had ‘grammatical gender’, the categories are ‘marked by the termination of the nominative, and also by other case endings’. Later however, this system began to decay: §41, 1100–1250. Definite article ⬎ fe, feo, fæt instead of se, seo, fæt. ‘Some confusion is seen in the gender of nouns’ §42. 1250–1350. Article still preserves some inflections, e.g. f gen sg, m acc sg, but ‘Nouns exhibit much confusion in gender-words that were once masculine or feminine becoming neuter’. Summary (§67): ‘Grammatical gender went gradually out of use after the Norman Conquest, owing to the following causes:– (1) the confusion between masculine and feminine suffixes (2) Loss of suffixes marking gender. (3) Loss of case inflection in the masculine and feminine forms of demonstratives. This traditional account displays some perennial themes, as we can see from later comments. A major one appears to be unidirectional causality: the (prosodically induced) loss of inflections somehow ‘eroded’ the gender system, and provoked or enhanced the shift to natural gender. So Mustanoja (1960: 42) remarks that gender ‘ceases to exist as a result of the levelling of the final vowels under uniform -e, which eventually becomes mute. Thus the chronology of the loss of grammatical gender shows a certain correspondence to the decay of the OE nominal and pronominal inflections’. But this cannot be true. The final vowels of OE nouns play a much smaller part in providing

32 Roger Lass clues for gender than what the nouns collocate with: determiners, the strong adjective and anaphoric pronouns. In fact consonantal markers are probably more important – e.g. the masculine/neuter s-genitive, the -r- of feminine oblique determiners, the neuter -t; other indicators are the feminine vocalic genitive as long as it remains, and the personal pronouns.17 An OE noun in isolation is rarely recognisable as belonging to a particular gender: e.g. a termination in -u could mean feminine h-stem ( gief-u ‘gift’), masculine u-stem (sun-u ‘son’), or light neuter a-stem plural (scip-u ‘ships’). But with certain other markers (e.g. an s-genitive which in ‘classical’ OE at least excludes feminine), or in the presence of marked determiners or anaphoric pronouns, the identifications are relatively unambiguous. So the definition of ‘having gender’ here will not focus on the forms of words themselves, but, for simplicity in what is a purely exemplary discussion, on noun satellites. Thus if in some text a historical feminine like soul ⬍ OE smwol has a genitive in -e, this probably reflects the original feminine type; but if there are no concordial phenomena the text does not ‘have gender’, but merely different noun declensions. A genitive in -e in the absence of determiner concord is just a declension marker, no more an instance of grammatical gender than ModE plurals like cat-s, strat-a, ox-en, sheep-ø. So soul would require a recognisably feminine determiner or anaphor to count as feminine. But most important, the loss of gender is not apparently driven by phonology to any significant extent, but is as we will see at least partly a morphosyntactic ‘choice’, or better set of choices. Here is a preliminary example. The scribe known as ‘The Worcester Tremulous Hand’ (Worcester fragments, MS Worcester Cathedral Library 174, early thirteenth century: henceforth WF), shows what in the usual story-line could be called a typical ‘transition’ system. The original three genders are largely retained, and explicit concord allows identification; yet we can see what look like ‘the beginnings’ of the later system. Consider for instance the italicised forms in the following:18 (7) (a) ft soul-e hus ‘the soul’s house’ (f63v) (b) sæif feo soule soriliche to hire licame ‘the soul says sorrowfully to her body’ (f65v) (c) sæif fe soule soriliche to hire licame “(f64v) In (7a), the e-genitive marks soule as a historical (but not textual) feminine, whereas the determiner ft (a crossed thorn in the MS ⫽ fæt) marks hus explicitly as neuter (these happen to be the original OE genders). In (b), the determiner feo, while not an OE form, does suggest feminine (cf. hgo, sgo), which is confirmed by the anaphoric pronoun hire. In (c) the determiner is the genderless fe, but again hire says ‘feminine’.19 In the two examples below we find a non-feminine s-genitive as expected for ‘devil’, as well as feminine feo

Reflections on the historiography of English 33 agreeing with lore (OE lmr, f ), and non-feminine dative fen (⬍ m, n fæ–m): (8) (a) fæ-s deofl-es lore feo fe likede wel ‘the devil’s lore that pleased thee well’ (f65v) (b) cwemde fe-n deofl-e ‘pleased the devil’ (f66v) When he marks gender, The Tremulous Hand fairly regularly shows what would be ‘expected’. But of course there are many nouns that as far as we can tell could be any gender, since there are no clues: they occur in the plural, with no determiners, have no referring pronouns, and/or are marked with the uninformative fe. But every once in a while we see something new: while ‘soul’ is normally a feminine as expected, in one sentence it is marked by the anaphoric neuter hit. But let us look at this text in a less impressionistic way, in terms of a quantitative ‘gender profile’. One way of drawing up such a profile is to examine the clearest instances of grammatical gender: singular definite noun phrases with determiners. These will, by virtue of the forms of the determiners, give a clear picture of how significant gender is as a structural element. I divide these noun phrases into three categories: (a) matches (the form of the determiner indicates the historical gender of the noun, e.g. feo soule);20 (b) mismatches (the form of the determiner indicates a different gender from the one we think ‘original’, for example, ft soule); and (c) uninformative (the determiner carries no information about gender, e.g. fe soule). The profile for the Worcester Fragments is:21 (9) Match Mismatch Uninformative

81% 10% 9%

Why should this be thought of as ‘transitional’, and part of the story of the spread of a single innovation? One reason is that text-languages from other parts of the country, at earlier dates, show what appears to be a (virtually) ‘completed’ loss of gender; and these are located to the north and east of WF.22 For instance the earliest text in ‘proper’ Middle English, the Final Continuation of the Peterborough Chronicle (Oxford, Bodleian Library, Laud Misc 636, 1154: henceforth PC) is probably at least half a century earlier than WF, and shows almost no trace of grammatical gender. The profile for PC is: (10) Match Mismatch Uninformative

7% 2% 91%

34 Roger Lass A comparison of the two profiles shows an apparently coherent story of loss (cf. the numerous displays in Lass 1992a of precisely this form): (11) Match Mismatch Uninformative

WF

PC

81% 10% 9%

7% 5% 91%

To round out the picture somewhat, here is the profile of another roughly contemporary text, which is still at an early stage of gender loss, but from a different part of the country, in this case Kent (MS Oxford, Bodleian Library Digby 4 – Poema Morale, early thirteenth century): (12) Match Mismatch Uninformative

85% 11% 4%

The ‘transitional’ nature of Digby and WF can be further thrown into relief by looking at a text somewhat north and west of Digby, also early thirteenth century, which has much more extensive loss of gender: London, British Library, Stowe 34 (Vices and Virtues, hand A: V&V), SW Essex: (13) Match Mismatch Uninformative

37% 9% 54%

A geographical summary: (14) Match Mismatch Uninformative

Northants

NW Essex

Kent

Worcs

7% 5% 91%

37% 9% 54%

85% 11% 4%

81% 10% 9%

This then apparently legitimates a north-easterly source for the innovation, and diffusion westward and southward: how else could WF and Digby be in the state they are, and PC in the state it is, with V&V apparently ‘in between’ both geographically and typologically? But there are two questions: (a) does this tabular display really represent a ‘narrative’, and (b) do the numbers, in the absence of qualitative information, really tell whatever story there is to be told? The answer in both cases would appear to be a qualified ‘no’. Profiles of this kind are rather insensitive: there is another dimension, which strongly suggests that convergence is part of the picture. That is, while gender is lost in texts from various parts of the country, it is not always lost the same way. In the examples so far all I have shown is numbers, but not the

Reflections on the historiography of English 35 genders marked. Here are the figures for the first three texts profiled above, but this time with the particular genders the scribes choose to indicate. I have restricted myself here to singular definite gendered NPs in either subject or direct object function, to avoid syncretisms that would obscure precise gender choice: For example, I have omitted oblique objects, since these would make masculine and neuter potentially indistinguishable (the historical dative singular determiner was fæ–m for both masculine and neuter, so no reflex would allow us to distinguish the two, and we would have to invoke an extra gender ‘masculine/neuter’).23 (15) M F N

WF

Digby

PC

15% 43% 42%

59% 38% 3%

13% 0% 87%

It is clear from these texts that regardless of the degree of gender retention, which genders are retained in what proportions is a major characterising variable: the scribes appear to display distinct preferences for which genders to mark when they decide to mark any at all. The processes of gender loss (and conversely the shape of gender retention) are idiosyncratic, which suggests analogy, not homology. All three languages are apparently engaged in ‘the same process’, but the preferences are independent: WF clearly disprefers masculine, but makes no significant distinction between feminine and neuter, whereas Digby prefers masculine and feminine and disprefers neuter, which is the gender of choice for PC. While ‘loss of gender’ as a ‘trend’ may have diffused along westerly and southerly vectors, its details did not. Broadly, there are at least two major patterns of gender loss: 1

2

Reduction in the proportion of gendered forms, but with masculine, feminine and neuter still remaining and recogniseable, without the statistical profile of Old English (dominance of masculines) being retained. Digby PM is a good example of the later stage, WF of the early. Reduction as above, but virtually complete loss of feminine, which collapses with masculine into an uninflected ‘the’ gender, while neuter remains to some degree distinct. PC is a late example (late in the sense that the job is almost done).24

This kind of data allows us to say that one of the typological trends in the History of English writ large is loss of grammatical gender, and that the (North)East just about finished before the West and the far Southeast really started. But the details make it impossible for this loss to be a single development, or for any one of the sample texts to be ‘ancestral’, even in type, to any other. Rather a good deal of the history of English must be not a single linear progression of character profiles but a series of convergences on types, at least largely independent of each other and not part of the ‘same story’.25

36 Roger Lass It is rather as if there was a job to do, specified in terms of rather general parameters, and PC had got it done while WF was just starting, and various other text languages were further along, but in different ways. Why this loss occurred at all, or why the particular attested patterns were the ones chosen is opaque as far as I can see, and in any case an explanation is not part of my remit here.26 The question is what the mechanism behind typological convergence is: even though there is often considerable documentable horizontal transmission, this is not the answer, and we do indeed have true (if complex) convergence. I do not have an ultimate answer, but I do have at least two proximate ones. (1) ‘Toggling’. One profitable way of looking at the pictures at least some of the texts give us is to imagine the writers running two software packages on the same hardware, with a ‘toggle’ switch between the two. One of the packages ‘has gender’, and the other does not. Variation then is a function of which package is being run at a particular time. The gendered package of course has different structures in different lects, and the statistical controls on which package is to be run when vary from language to language. To illustrate how this toggling mechanism appears to work in one text, consider the clustering of gendered and non-gendered forms of the word ‘owl’ in one version of The Owl and the Nightingale (MS BL Cotton Caligula A ix, Language A); the forms occur in this order (the gender is always feminine):27 (16) cluster 1: ⫹gen, ⫹gen, ⫹gen cluster 2: ⫺gen cluster 3: ⫹gen cluster 4: ⫺gen, ⫺gen, ⫺gen, ⫺gen, ⫺gen, ⫺gen This image may appear to be somewhat forced; after all, what is the status of a one-member ‘cluster’, and is it in fact merely tendentious to name it this way? But I think the kind of pattern that emerges is of interest: gendered forms in this text (and in some others) often show a strong if not overriding tendency to appear in groups, though they do also occur as singletons. If there is a toggle mechanism its exploration is still in its infancy. Still, this is a suggestive way of looking at one type of transitional organisation. Whether or not this model is convincing is a matter for further research: this is just a heuristic beginning, and a great number of texts at different stages of the dismantling of gender will have to be analysed in this way before we can be sure whether such a model is worth pursuing. (2) Lexical Tagging. Many texts do not show this kind of clustering pattern, but rather an association of gender with particular lexical items, either absolute or nearly so. Thus in PC ‘folk’ and ‘land’ (as frequently in other texts) are preferential neuters, in Digby ‘heart’ is feminine. An alternative version is for a given lexical item to have two forms, one gendered and one not, but with both types of forms appearing in a more standard scattered alternation pattern, rather than clustering.

Reflections on the historiography of English 37 In any case, it seems clear that whether or not these are two distinct types of structure, the texts overall show, to borrow an old structuralist term, ‘coexisting systems’. Indeed, such an organisation is perhaps the most natural one in the circumstances.

Conclusion It is true that at one time every variety of English had grammatical gender; it is true that at present no variety does. It is also true that starting from Old English, arbitrary text samples at increasingly recent dates will tend to show progressively less gender, until the point of final disappearance in the sample. This is what has made the history of gender loss appear essentially linear. It is equally true that at any point in such a sampling procedure, no two texts (often not even by the same scribe) will show identical patterns of gender loss, or identical patterns of retention.28 This can be called the Paradox of Apparent Linearity; its solution as I have suggested is the conflation of space and time, and recognition of typological convergence as an important mode of change.29 The data suggests that the history of English is ‘areal’: it is not a single story but a set of roughly parallel similar stories, unfolding at different rates in different regions. So it is not really sensible to say that some change C began in the north and spread southwards: this is clearly not what happened. It is rather the case that northern varieties simply started their sequence earlier, though there was also (and the details have to be specified for each case) horizontal transmission as well. But we do not need to emphasise mechanisms of spread or connections between regions if we can justify the typological story. And this would give us a background against which to look at the differences in instantiation.30 At any rate there is a distinct need for a redefinition of ‘history of English’: rather than a linear development we must see it first as a smear of phenomenology analysable into separate partial trajectories in space and time, or perhaps better a single complex dimension of space/time. There is no ‘big picture’ except in the grossest typological outline; all of the text-languages conventionally grouped as ‘OE’ are more highly inflected than any of the text languages grouped as ‘ME’ or ‘ModE’. But before perhaps the mid to late fifteenth century, there is no text-language or cluster of text languages that stands as a proper lineal ‘ancestor’ to any present variety; all historiography of the language is based on only partly (and largely instrumentally) justifiable simplifying assumptions of the kinds I discussed in the earlier sections of this paper. Notes 1 The reason for the scare-quotes around ‘dialect’, particularly in a historical context, will become clearer later on. See also LAEME, vol. 1, ch. 1, Laing and Lass (2003) and Fleischman (2000).

38 Roger Lass 2 The diagrams and discussion below are an expansion of material to appear in Laing and Lass (forthcoming). 3 The MS uses both the two-letter sequence ⬍ae⬎ and the digraph ⬍æ⬎. 4 As well as regional features, such as ‘Anglian smoothing’ in Northumbrian maecti, mecti and ‘Palatal umlaut’ in WS mihte. This point will become significant later on. 5 Paradoxically, perhaps the most explicit espousal of directionality in this sense is the ‘backwards’ history of English (present-to-past) in Strang 1970. 6 I start from Middle English for convenience; if we went back to (WS) Old English the first and third forms would have ‘OE æ’ and the second ‘OE a’. 7 See Lass (1976: ch. 4) for an elaborate example of doing precisely this. 8 Cf. Gould (1991a: 158): ‘Independent thought has always been more difficult than borrowing, and authors of textbooks have almost always taken the easier way out’. But in all fairness it must be said that borrowing is often a function of belief, there is a ‘What oft was said but ne’er so well expressed’ factor, and some people get converted to ‘independence’ late in their careers, if at all. 9 Since the quotations come from work I did well over a decade ago I think I can safely consider myself an ‘outsider’ now. It would be rather sad if I were unable to. 10 The first of the lineage, Hyracotherium (formerly Eohippus), is traditionally said to have been the size of a fox terrier. For an amusing account of how this idea arose through a chain of plagiarisms, see Gould (1991a). 11 The other side of this error is the essentialist tendency to view ‘species as entities, not tendencies; things, not arbitrary segments of a flux’ (Gould 1987: 147). I will return to this issue below. 12 It is interesting that the one diagram in The Origin of Species (Darwin 1859: ch. IV) shows the development of life on earth as a complex branching configuration, not a linear progression. 13 Either by plasmids (small rings of DNA expelled from bacterial cells and taken up by others), or by the scavenging of DNA by other bacteria. 14 Except of course those Germanic languages that have no strong/weak distinction because they have lost the preterite and replaced it by the perfect, like Yiddish and Afrikaans. In this case the loss is a convergence, and leads to similar surface phenomenology with no historical connection. 15 This preliminary taxonomy is mine, and subject to revision. The distinctions are not to my knowledge made this way in any standard source. 16 For a treatment of large-scale convergence in which Sapir is taken rather more seriously, and an attempt to embed drift in modern typological theories, see Dury (2001). (Dury however sees contact, rather than drift or typological convergence as the prime mover in the development of so-called ‘Standard Average European’ languages.) To be fair, Sapir did make something of an approach to demystifying drift by giving it a psychological substrate, but the results were incoherent: for a critique see Lass (1987). 17 In Middle English the consonantal markers are in fact the main source of information about gender, at least in the determiner systems. 18 All data cited in this paper are from the corpus of tagged texts constituting the data-base for the Linguistic Atlas of Early Middle English (LAEME: Laing, Lass and Williamson, in preparation).

Reflections on the historiography of English 39 19 The increasing incidence of fe (which already appears as a variant of masc nom sg se in some ninth-century texts) marks the attrition of both the gender and case systems. 20 I make the simplifying (and sometimes counterfactual) assumption that historical gender is determinate. For a discussion of this problem and some cases where it is not, see Lass (1997b). 21 This profile and those below are based on samples of roughly 100 tokens per text, with percentages rounded off to give the simplest typological picture. 22 And loss of gender is ‘earlier in the North’ (in fact the Northeast, if Lindisfarne as usually believed is from Chester-le-Street); but the partial instability of the system in this tenth-century text is not a legitimate base for inference, because there is no lineal relationship between Lindisfarne and any ME text. 23 In these figures I have also included the ‘mismatches’, since the picture I want to present is that of overall choice of gendered determiners (without particular concern for whether the genders are ‘correct’) as opposed to uninformative fe types. I have omitted V&V because of a textual peculiarity that makes it difficult for this particular exercise: while it is extremely rich in feminine obliques, it appears to have almost no feminine nouns in subject or direct object position, and so has a very different syntactic profile, and for genuine description of choice would require oblique objects, which I have ruled out in this brief heuristic survey. A proper large-scale survey would of course be organised differently, and would include all syntactic positions. 24 This pattern is interestingly reminiscent of the route eventually taken by Dutch, much of Low German and most of continental Scandinavian: collapse of historical masculine and feminine into a ‘common’ gender, and retention of neuter. This is quite different from the Romance pattern, in which (broadly) masculine and neuter collapse into a single ‘masculine’ gender and feminine remains distinct. 25 What appears to be ‘the same’ phenomenon can in any case often have quite different underlying causes. For instance the appearance of OE æ as e in ME can – depending on region – be due to at least four quite independent changes: ‘second fronting’, ‘Kentish collapse’, Scandinavian borrowing (e.g. N eftir ⬍ OSc eptir) or independent raising in the North (wes for was). 26 For some discussion and literature on this topic see Millar (2002). 27 There may of course be triggers for toggling such as register, tone, topic, etc.; none of these appear to apply in this particular text. 28 Since in this preliminary and programmatic study I have not sampled all the extant texts this is of course a mildly risky statement: but the patterns I have looked at suggest it is unlikely to be false. It is at least vulnerable to disconfirmation. 29 Note by the way that claiming convergence does not in any way vitiate the uniformitarian assumption that by and large groups of cognate dialects constitute continua (Laing and Lass 2003). That is a different issue. It is possible in any case for continua to be split by innovating or conservative lect-clusters, or areas of unusually low population density. But even within a dialect continuum of the standard sort there are separations at the peripheries, and plenty of slack for convergence as well as diffusion. 30 If one would like to assign a larger role to horizontal transmission it is probably possible to do so. There is something about the whole picture of gender loss (and other developments as well) that does rather appear to resemble large-scale ‘areal

40 Roger Lass convergence’. Could it be that one way of looking at England during our period is as a kind of Sprachbund, with each text-language in a sense equivalent to a larger-scale ‘language’ as groups of these are spread across the map of Europe, or smaller regions like the Balkans or the circum-Baltic area? If one wanted to pursue this, the question would be whether inter-variety contact across England was strong enough, in the absence of a ‘centre’, to bring about convergence effects of the kind that typically result from long contact and prestige, for example those discussed in Dury 2001. I find it hard to believe that overall Northants and Worcs were ‘in contact’ in the way that Worcs and Herefords were; whatever contact there may have been was likely looser and less predictable, probably more like that of early mediaeval High German and Netherlandic. But this is a reasonable direction for research.

3

History and Comparative Philology Roy Harris

Saussure’s diagnosis of the reasons for the failure of Comparative Philology is blunt and uncompromising. The comparative method itself was vitiated from the very beginning by not taking history into account. According to the Cours de linguistique générale: The first mistake made by the comparative philologists was one which contains the seeds of all their other mistakes. Their investigations, which were in any case limited to the Indo-European languages, show a failure to inquire into the significance of the linguistic comparisons they established and the connexions they discovered. Comparative grammar was exclusively comparative, instead of being historical. Comparison is no doubt essential for all historical reconstruction. But in itself comparison does not warrant drawing conclusions. (Saussure 1922: 16–17) This passage comes from the introductory survey of the history of linguistics. In the final paragraph of the survey, the Neogrammarians are praised for having rectified that fundamental Comparativist error: ‘The achievement of the Neogrammarians was to place all the results of comparative philology in a historical perspective, so that linguistic facts were connected in their natural sequence’ (Saussure 1922: 18). The very same chapter gives what might be regarded as a mini-lesson in Comparative Philology, where the paradigms of the Latin noun genus and Greek génos are set out side by side and then compared to the corresponding Sanskrit forms. Curiously, however, the commentary on this comparison hardly squares with the remarks about Comparative Philology and history. Comparing the Greek and Latin forms, it is alleged, ‘tells us little’. But when they are set beside the Sanskrit forms all is revealed: ‘At a glance [Il suffit d’y jeter un coup d’œil] we now can see the relationship between the Greek and Latin paradigms’ (Saussure 1922: 15). Whatever happened to history? It may sound pedantic, but it needs to be said that, pace Saussure, ‘at a glance’ we can see no such thing. All that can be seen, from comparison alone, is a set of correspondences and lack of correspondences between three sets of

42 Roy Harris spellings. Whereas what Saussure obviously wants us to ‘see’ includes, for example, that an original intervocalic [s], retained in Sanskrit, fell in Greek but was rhotacized in Latin. He hastily adds, as if by way of afterthought, one essential proviso: that the Sanskrit paradigm must be assumed to represent the original forms. That, certainly, brings in history, albeit in a somewhat indirect, remote and hypothetical manner. But, unfortunately for Saussure’s example, it is nothing we can ‘see’, much less ‘at a glance’, and certainly not by mere inspection of just three word paradigms. It would not be worth taking Saussure or his editors to task merely for a clumsy piece of exposition. There is more to it than that. We see here in miniature a perfect example of the historical cleft stick in which much linguistic thought in the nineteenth century was trapped, and from which Saussure tried so hard, although not always successfully, to struggle free. ***

The two passages from the Cours quoted at the beginning of this chapter between them seem to present a stark and simplistic scenario: at Stage 1, linguists simply compared languages one with another, oblivious to the historical facts, whereas at Stage 2 it was realized that, when divorced from the historical facts, such comparisons were of little or no value. If this is history of linguistics, it is extremely crude history, if not downright false, and appears all the more so when one recalls the point of departure from which comparative Indo-European linguistics had set out. This is commonly assumed to be the famous paragraph from Sir William Jones’s Third Anniversary Lecture to the Asiatic Society in 1786. The Sanscrit language, whatever be its antiquity, is of a wonderful structure; more perfect than the Greek, more copious than the Latin, and more exquisitely refined than either, yet bearing to both of them a stronger affinity, both in the roots of the verbs and in the forms of grammar, than could possibly have been produced by accident; so strong indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists: there is a similar reason, though not quite so forcible, for supposing that both the Gothick and the Celtick, though blended with a very different idiom, had the same origin with the Sanscrit; and the old Persian might be added to the same family [. . .]. ( Jones 1786: 34) Jones’s rather cautious speculation, in which six branches of the IndoEuropean language family are tentatively identified, was later to be hailed as ‘the first known printed statement of the fundamental postulate of comparative linguistics’. We are told that ‘at this moment [sc. on 2 February 1786] modern comparative grammar was born’. Jones’s remarks reflect, it is said, ‘the flash of a prophetic inspiration’. These overenthusiastic assessments do

History and Comparative Philology 43 not bear serious examination; much less the idea that Bopp, Grimm, Rask and other European linguists at the beginning of the following century were just following up the comparativist ‘insight’ (as it has sometimes been called) vouchsafed to Jones. Had they been doing that, they could hardly have ended up as the benighted comparativists described by Saussure. For Jones’s conjecture, whatever we may think of it, was manifestly a historical conjecture: he was not in the least interested in comparison for comparison’s sake. Jones, who died in Calcutta in 1794, could hardly have anticipated that later generations would hail him as a ‘founder’ of Comparative Philology, much less as ‘the first modern linguist’. There was nothing very ‘modern’ about Jones, a well-educated product of Harrow and Oxford, who chose the law for a career, married the daughter of a bishop, and went out to administer British justice in India in return for a knighthood. His famous pronouncement about Sanskrit was not the result of years of scholarly research. When he made it, he had only just begun the study of Sanskrit and it was not with any academic purpose in mind. As a British judge newly arrived in India, he saw the practical services that could be rendered by compiling a digest of Hindu and Islamic law – a task for which a knowledge of Sanskrit was essential. His statement about the likely relation of Sanskrit to Greek and Latin is a report of his own initial reactions on being introduced to the study of that language. The ‘philologer’ it refers to is not the hypothetical spokesman for any consensus of current academic opinion, but Jones himself. He is simply noting what struck him immediately as a range of linguistic similarities so obvious as to have only one plausible historical explanation. That this should have struck Jones as obvious is itself a reasonable assurance that the general idea of linguistic relationship through a common ancestor was not novel as far as he was concerned. He does not sound in the least like a man excitedly crying ‘Eureka!’. In spite of this, a recent (British) encyclopaedia of linguistics can still assure its readers that ‘the notion of relatedness was first introduced by Sir William Jones’ – a sentiment perhaps admirable for its patriotism but hardly for its accuracy. The more general point relevant here is that if this idea was no novelty at the end of the eighteenth century, then it was certainly no novelty at the beginning of the nineteenth either. Whatever ‘revolution’ may have been ushered in by the Comparative Philologists, it was certainly not that. But there is an even greater irony in the history which gives Jones the Lord Mayor’s role of laying the foundation stone of Comparative Philology. Jones himself clearly had no great regard for philological studies as such. He held to a common eighteenth-century view (and one that still survives today) which considers languages to be worth studying only as a means of access to foreign cultures. In his inaugural discourse to the Asiatic Society in 1784 he makes this point perfectly clear: ‘I have ever considered languages as the mere instruments of real learning, and think them improperly confounded with learning itself’ ( Jones 1784: 7).

44

Roy Harris

This contrasts very markedly with Bopp’s statement in the preface to his Vergleichende Grammatik, where it is made no less clear that the languages analysed ‘are treated for their own sakes, i.e. as objects and not means of knowledge’. Thus Comparative Philology could hardly have had a less likely founder than Jones, or its establishment as an independent academic discipline a more improbable advocate. ***

The first question that arises, therefore, in assessing this chapter in the history of linguistics, is the historicity of the claim that it was William Jones who started the ball rolling by ‘founding’ Comparative Philology. And close upon the heels of that comes the question of whether the Comparative Philologists were concerned exclusively with comparison at the expense of history (as Saussure, for one, later claimed). In an article contributed to the Edinburgh Review in October 1851, Max Müller wrote: ‘Comparative Philology’ is a science of very modern date. It cannot trace its lineage much higher than the beginning of the present century, and it is scarcely received as yet on a footing of equality by the elder branches of learning. Classical Philology, in particular, as if afraid of the charms of a younger sister, has for a long time treated Comparative Philology with a kind of supercilious pride, and in several cases the hostile feeling of the two rivals has vented itself in open feuds and quarrels. Comparative Philology has, however, maintained its ground successfully against the prejudices and jealousies which everything new has to contend with. (Müller 1851: 125) The occasion of Müller’s article was the appearance of an English translation (by Eastwick) of Franz Bopp’s Comparative Grammar of the Sanskrit, Zend, Greek, Latin, Lithuanian, Gothic, German, and Sclavonic Languages, which Müller describes as being ‘universally considered as the classical work on this branch of Comparative Philology’. In this same article, Müller quotes the celebrated paragraph from William Jones in order to make the point that it was not Bopp ‘who first discovered the connexion of the Arian languages’ (Müller 1851: 141). So by the middle of the nineteenth century we already see the myth of Jones, founder of Comparative Philology, being propagated to the British general public. Müller repeats the quotation from Jones in the fourth of his highly popular Lectures on the Science of Language (1861). He adds that this discovery by Jones caused consternation among the learned. People were completely taken by surprise. Theologians shook their heads; classical scholars looked sceptical; philosophers indulged in the

History and Comparative Philology 45 wildest conjectures in order to escape from the only possible conclusion which could be drawn from the facts placed before them, but which threatened to upset their little systems of the history of the world. (Müller 1861: 163) The phrase ‘their little systems of the history of the world’ is particularly striking. Here we see one version of a theme that dominated a generation of historians imbued with the historicism of Ranke and his followers. At the same time we see the lines along which the myth is being developed. Jones is no longer just a British official with a practical interest in Indian antiquities, but the pioneer of a revolution in historical thinking. By 1874, Jones has become for Müller ‘a very great man – one in a million’ (Opening address to the International Congress of Orientalists). Almost contemporaneously with Müller’s article in the Edinburgh Review, there appeared in the North British Review (November 1851) an article by Alfred Strettell, likewise entitled ‘Comparative Philology’, which begins in strikingly similar terms: The comparative study of language is of quite modern date. It was hardly known in Europe thirty years ago; for that unscientific comparison of single words, without principle or analogy, which made itself so often ridiculous in the seventeenth and eighteenth centuries, does not deserve the name. (Strettell 1851: 163) It is interesting to note here the use of the term ‘unscientific’. For the claims of Comparative Philology to be a ‘science’ were, as later scholars were to realize, debatable. These claims were also intimately connected to the thesis that Comparative Philology was something ‘new’. That novelty, as the passages quoted indicate, was already, by the 1850s, a topos that had become part and parcel of the propaganda for the development of the subject. Its advocates never ceased assuring the general public that Comparative Philology was the ‘latest thing’ and a science to boot. They continued to do so with monotonous regularity for the best part of the century. However, Strettell’s version of history differs notably thereafter from Müller’s. The hero of Strettell’s version is not Jones (not mentioned once) but Humboldt. Humboldt does feature in Müller’s story, but not as a possible rival to Jones. For Müller, the author of Kosmos is merely one of those scholars who had recognized the importance of the comparative study of languages. He has no claim as founder of the discipline. We are dealing here with one of those episodes in intellectual life which can be – and has been – represented in quite different and almost contradictory ways. From the outset, Comparative Philology was a movement of which its own propaganda formed an essential element: in this respect, its practitioners showed that they had learnt at least one of the lessons of the

46 Roy Harris Enlightenment. According to some historians (who support the original claims of its nineteenth-century proponents) Comparative Philology was a revolutionary breakthrough not only in the study of language and languages but in modern understanding of the past and our relation to it. According to others, on the contrary, the Comparative Philologists simply developed ideas that were by no means new, by laboriously collecting much documentary evidence to support them. These two views – and some playing off of one against the other – are the basis of the usual accounts given nowadays in histories of linguistics. Neither is entirely convincing and behind both there lies a great deal of rather petty bickering about which scholars, universities and countries should be allotted the lion’s share of praise. As often in the construction of what passes for ‘the history of ideas’, there is an agenda which has much to do with self-justification, with placing oneself and one’s colleagues in the ‘proper’ line of descent, and also with current politics of both the academic and the national kind. It is significant that even the terminology of the language-family itself was not uncontroversial. Some preferred ‘Aryan’ or ‘Indo-Germanic’ (indogermanisch) to ‘Indo-European’: what was felt to be at stake can be gleaned from Whitney’s objection to the second of these that it ‘seems to savour of national prepossession’. ***

Another candidate for the historical role of founder or co-founder of Comparative Philology is Jacob Grimm. Grimm was born in Hanau in 1785, the year before Jones’s much-quoted address to the Asiatic Society. With his brother Wilhelm (b. 1786), Jacob is probably best known today as coauthor of Grimm’s Fairy Tales. Both brothers were also involved in the monumental dictionary of the German language (Deutsches Wörterbuch) which bears their name. It was begun in 1837, shortly after their dismissal from professorships at Göttingen for political protest against civil rights abuses, and was left unfinished when they died ( Jacob in 1863 and his brother in 1859). Continued by later scholars, the dictionary was eventually completed in 1960. It was Jacob Grimm’s Deutsche Grammatik, published between 1819 and 1837, which established his reputation as a philologist. More especially, the revised edition of the first volume in 1822, containing what later became known as ‘Grimm’s Law’, came to be looked on as a landmark in the establishment of Comparative Philology. The ‘law’ in question concerns a pattern of distribution of voiced, voiceless and ‘aspirate’ consonants in cognate words across certain Indo-European languages. Thus, for example, where a word in Sanskrit, Latin or Greek began with p, the corresponding word in Germanic languages would usually begin with f (Latin pater, English father; Latin piscis, English fish, etc.) But something more exact than this was needed for ‘scientific’ purposes.

History and Comparative Philology 47 Max Müller, who claimed to have coined the term ‘Grimm’s Law’, formulated it to his own satisfaction as follows: If the same roots or the same words exist in Sanskrit, Greek, Latin, Celtic, Slavonic, Lithuanian, Gothic and High-German, then wherever the Hindus and the Greeks pronounce an aspirate, the Goths and the Low Germans generally, the Saxons, Anglo-Saxons, Frisians, etc, pronounce the corresponding soft check, the Old High-Germans the corresponding hard check. In this first change the Lithuanian, the Slavonic, and the Celtic races agree in pronunciation with the Gothic. [. . .] Secondly, if in Greek, Latin, Sanskrit, Lithuanian, Slavonic, and Celtic, we find a soft check, then we find a corresponding hard check in Gothic, a corresponding breath in Old High-German. [. . .] Thirdly, when the six first-named languages show a hard consonant, then Gothic shows the corresponding breath, Old High-German the corresponding soft check. In Old High-German, however, the law holds good with regard to the dental series only, while in the guttural and labial series the Old High-German documents generally exhibit h and f, instead of the corresponding mediæ g and h. (Müller 1864: 199–200) The importance of these correspondences for Comparative Philology was clearly historical: that is, they were taken as showing that at certain points in the past major cleavages had arisen within the body of Indo-European. In other words, a regular interlinguistic difference at a later date was taken as reliable evidence of earlier (unwitnessed and undocumented) events in the linguistic past – in this instance, the splitting up of a single language into two or more separate languages. In this respect, linguistics became ‘scientifically’ comparable, at least superficially, to geology and astronomy – both sciences which accepted currently observable phenomena as reliable indicators of unobserved prior events. This was (part of ) what Max Müller had in mind when he proclaimed (in his second series of Lectures on the Science of Language in 1864): ‘There is a Science of Language as there is a science of the earth, its flowers and its stars.’ Unfortunately, however, the term ‘Grimm’s Law’ (whether Müller invented it or not) was a misnomer in more ways than one. In the first place, it went beyond what Grimm himself claimed. Grimm regarded such correspondences as tendencies rather than invariable regularities. It was not until the second half of the nineteenth century that theorists of the Neogrammarian school began to claim that ‘sound laws’ – if they were indeed laws – must have no exceptions. In the second place, as critics pointed out, the correspondences as stated by Grimm in 1822 do not in fact hold without a certain amount of juggling and allowance for exceptions. Third, they argued, the ‘law’ was in any case not Grimm’s but Rask’s, because Rask had drawn attention to the relevant correspondences earlier and Grimm had simply borrowed the idea from him.

48

Roy Harris

Perhaps a more telling argument against regarding Grimm as the founding father of Comparative Philology is that even if, as his supporters claim, he ‘went beyond’ Rask, it is doubtful whether he went in the right direction. For Grimm seems to have regarded all the correspondences encompassed in ‘Grimm’s Law’ as part of a single continuous cyclical process; whereas in fact some of them reflect linguistic changes many centuries apart, which are unconnected. A rather different failing is that Grimm’s work, as even his admirers concede, suffered from a ‘consistent failure to distinguish between letters and sounds’. The original title of the phonological section of his Deutsche Grammatik was ‘Die Lehre von Buchstaben’ (which Grimm altered to ‘Von dem Lauten’ only in the revised edition of 1840). He believed that the orthographic word Schrift expresses ‘eight sounds through seven signs’ – a calculation which requires sch to ‘stand for’ three sounds and f to ‘stand for’ [p] ⫹ [h]. The relevant point here is not Grimm’s own intellectual limitations, but rather those of this whole first generation of Comparative Philologists. As Saussure later put it, reading Bopp and Grimm takes us back to a time when the assumption was that ‘a language is inseparable from its alphabet’. Only when that separation had been clearly recognized does it make much sense to say that we are dealing with a ‘science’ of language. ***

Rasmus Rask, to whom Grimm acknowledged his indebtedness, was born at Brändekilde in Denmark in 1787 and was thus Grimm’s almost exact contemporary. He, in turn, has been hailed as ‘the father of historicalcomparative linguistics’, ‘le fondateur de la méthode du XIXe siècle dans le domaine de la grammaire comparée’, etc. In a brief academic career (he died in 1832 at the age of 44), he held chairs in literary history and oriental languages at the University of Copenhagen. He also published grammars of various languages, including Anglo-Saxon, Italian and Spanish. But his fame as a Comparative Philologist rests on his Undersøgelse om det gamle nordiske eller islandske sprogs oprindelse, completed in 1814 (when Rask was only 27) and published four years later. Previously (1811) he had made his academic mark by publishing an Icelandic grammar. Rask maintained that language was a more reliable indicator of cultural history than religion, laws and other institutions. It followed that if linguistic affinity could be established between two communities or peoples, this took precedence over non-linguistic affinities (or discrepancies). The issue that then presented itself was how best to establish linguistic affinity. Here Rask is often said to give priority to resemblances in grammatical structure, but he attached importance also to correspondences between words constituting core vocabulary. Interlingual correspondences between ‘letters’ thus became significant only if the ‘letters’ belonged to words that could be established as belonging to a presumed original lexical stock. The methodological difficulty here

History and Comparative Philology 49 (which Rask never resolved) is that the notions of correspondence and originality seem doomed to chase each other round in an endless circle, unless some independent criteria are available for establishing the latter. And this is itself theoretically problematic as soon as the philologist’s inquiries push back into a period before historical records. But it is doubtful whether Rask himself ever saw this. The lacuna is connected to Rask’s most famous ‘mistake’ – his refusal to admit Celtic as Indo-European. (He eventually retracted this.) But, as his compatriot Hjelmslev later pointed out, nowhere in Rask’s published or unpublished writings do we ever find a systematic statement of the theoretical basis of his work. ***

The third front-running candidate, Franz Bopp, was born in Mainz in 1791 and studied in Paris under the French Oriental scholars Chézy and Silvestre de Sacy. The publication in 1816 of Bopp’s Über das Conjugationssystem der Sanskritsprache in Vergleichung mit jenem der griechischen, lateinischen, persischen und germanischen Sprache has been claimed to be ‘generally regarded as marking the beginning of comparative Indo-European linguistics’. After four years in Paris, Bopp moved to London in 1817 and while there taught Wilhelm von Humboldt, then Prussian ambassador to the Court of St James. This contact stood him in good stead later, when Bopp moved to Berlin and was appointed (in 1821) Professor of Oriental Literature and General Philology, a post which he held until his death in 1867. The implications of his Conjugationssystem were followed up more comprehensively in his Vergleichende Grammatik (1833–1852). He also published a number of grammars of Sanskrit. In 1866 his Vergleichende Grammatik was translated into French by Michel Bréal. Of particular interest here is what Bréal saw as being the originality of Bopp’s contribution to Comparative Philology. In the introduction to his translation, Bréal writes: The originality of Professor Bopp’s first book [i.e. the Conjugationssystem] consists not in its presentation of Sanskrit as a language of the same family as Greek, Latin, Persian, and Gothic, nor in its precise definition of the nature and extent of the relation between that Asian language and those of Europe. These things had long since been discovered. The affinity between Sanskrit and Western languages is so manifest, it extends to such a large number of words and to so many grammatical forms, that it struck the first educated men who undertook the study of Indian literature. (Bréal 1991: 26) Bréal goes on to point out that even before Sir William Jones, a French Jesuit in Pondicherry, Cœurdoux, had raised the question of the relationship

50 Roy Harris between Sanskrit, Greek and Latin in a memorandum to the Académie des Inscriptions et Belles-Lettres in Paris, and identified similarities in words and grammatical structures. (This was in 1767, two decades before Jones’s lecture to the Asiatic Society.) So by the time Bréal was writing (1866), recognition of Jones’s historical claim to be the founder of Comparative Philology was presumably widespread enough to be worth questioning. In a celebrated put-down remark, Meillet said of Bopp that he had discovered comparative grammar while looking to explain Indo-European, much as Columbus had discovered America while looking for a passage to the Indies. What Meillet seems to have meant by this is that – contrary to later interpretations – Bopp was not interested in linguistic comparison as such, nor even in the principles underlying it, but simply in comparison as a means of getting back to ‘the origins’ of Indo-European. In this, said Meillet (in 1903), Bopp was still a man of the eighteenth century. Bréal comments more soberly on Bopp that ‘he does not set out to prove the common origin of Sanskrit and the European languages; rather, that common origin serves as a point of departure for, not a conclusion to, his work’. Here Bréal puts his finger on a point which later historians of linguistics seem to have missed or ignored. The fact is that not only Bopp but none of the putative founders of Comparative Philology ever set out to prove the common origin of the Indo-European languages. Rather, they took that for granted and contented themselves with drawing attention to examples which seemed to corroborate it. This is confirmed, as Müller points out (Müller 1851: 142), by Bopp’s own statement that ‘the establishment of a connexion of languages was not so much a final object with me as the means of penetrating into the secrets of lingual development.’ ***

We are thus led to a conclusion which may at first sight seem perplexing. For none of the reputed founders, to judge by their own published work, confronts the major foundational challenge that one might have supposed them to have been concerned with. Instead, they assume from the outset that descent from an unattested common language is unproblematic and obviously provides the correct explanation of the relationship between the various branches of Indo-European. So all that remains is to work out the details. And even though the ‘details’ depend so manifestly on the phonetics involved, none of them comes close to providing a convincing theory of sound change. Rask held that inherent differences in the anatomy of the vocal organs between one nation and another were responsible for the consistent phonetic correspondences that comparison revealed. Grimm even supposed that some of the phonetic developments in ‘Grimm’s law’ were due to the ‘craving for freedom’ that made itself felt in Germany in the Middle Ages. All of which brings us back by a circuitous route to the less romantic view that Comparative Philology did not mark a revolution in linguistic thinking

History and Comparative Philology 51 at all, despite the claims made by its supporters. What exactly were the new ideas it ushered in? Proponents of the less romantic view point out that it can hardly be maintained that the idea of comparing languages was new, for this goes back to Classical antiquity. Nor the idea that one language may be ‘descended’ historically from another, for this goes back to antiquity too (when it was commonly supposed that Latin was derived from Greek). The idea that linguistic relationships can be systematically investigated by ‘fieldwork’ (i.e. by collecting and collating empirical evidence from speakers of the languages concerned) seems at first sight more promisingly ‘modern’, until we recall (i) that Comparative Philology was essentially book-based and text-based, and (ii) that the idea of comparative linguistic ‘fieldwork’ was already current in the eighteenth century anyway. (In some respects the Comparative Philologists turned their backs on this, and reverted to the library-bound approach of their Classical colleagues.) Leibniz, for example, had recommended collecting samples of the Lord’s Prayer in many different languages, together with word-lists including numerals, words for parts of the body, kinship terminology, etc. Moreover, a considerable amount of this comparative ‘fieldwork’ was actually done, particularly in Russia, under the patronage of Peter the Great and Catherine II. Typical of this line of research was the publication in St Petersburg by G.S. Pallas of many such word-lists, under the ambitious title of Linguarum totius orbis vocabularia comparativa (1786–1789). But that was nothing for which the Comparative Philologists can take credit. Were their methods any more rigorous than those of their predecessors? In the passage from Strettell’s article on ‘Comparative Philology’ quoted earlier, there is a contemptuous reference to the ‘unscientific comparison of single words, without principle or analogy’ which passed for linguistic comparison in the seventeenth and eighteenth centuries. But this is another item of Comparativist propaganda. If we look without prejudice at Turgot’s famous article on etymology in the Encyclopédie, in certain respects, as one historian of linguistics has noted, Turgot’s grasp of the problems of etymology seems to be more penetrating and comprehensive than Rask’s. It is sometimes urged that where the Comparative Philologists made an advance was in insisting on comparisons of grammatical structure (rather than vocabulary) in establishing phylogenetic relationships. But this was no novelty either. It had been stated very clearly by Ludolf at the beginning of the eighteenth century. So what – and the repeated question now verges on the rhetorical – did Comparative Philology have to offer that was a historical ‘breakthrough’ in anything? Strettell in 1851 proposed the following answer: that if we are to comprehend Comparative Philology’s ‘tardy appearance amongst the number of the sciences’ we need to realize that Comparative Philology implies ‘a particular temper of mind’. This may be the right answer, although not necessarily for the reasons Strettell gave. ‘No nation, ‘ wrote Strettell, ‘has done so much as the English

52

Roy Harris

in the way of amassing materials for the comparative study of language.’ He goes on to expatiate on the opportunities for this work afforded by ‘our widely extended colonies and commerce’ and the services rendered by ‘merchants and missionaries, soldiers and civilians’, all of whom have exhibited ‘the spirit of intelligent and faithful observation which characterizes our nation’. But, he laments, when it comes to looking ‘either for a manual or an orderly collection’, then we have to look to ‘our German neighbours’, and particularly to works of Bopp, Grimm, Pott and Humboldt. The main body of Strettell’s article is devoted to an exposition of Humboldt’s posthumously published dissertation Über die Kawi-Sprache auf der Insel Jawa. As interpreted by Strettell, Humboldt’s main point is that, although humanity shares language in common, comparison shows that different nations have developed language in different ways. There are ‘higher’ languages and ‘lower’. There are nations that are ‘qualified to construct higher forms of language’. These include, it need hardly be added, such nations as the Germans and the British, whereas other languages became ‘petrified’, because ‘the national mind did not advance’. Similarly, Max Müller believed that there was a correlation between phonological complexity and mental progress and that the earliest speakers of proto-Indo-European might have had no ‘aspirate’ consonants at all. He held that ‘a very imperfect alphabet [i.e. range of sounds] will suffice for the lower states of thought and speech.’ Why was all this linguistic hierarchization important? Strettell’s answer is unequivocal: ‘the language of a people reveals the depths of individual character to an exercised and reflective mind, which even their practical works, their institutions and customs, cannot unfold’ (Strettell 1851: 165). Here, certainly, we seem to hear echoes of Rask. But there is more. If, therefore, the Christian missionary desires to follow the example of the great apostle, St. Paul, and to meet the heathen nations whom he wishes to convert to the faith of Christ, as St. Paul met the Athenians, upon their own standing ground, in order to destroy the falsehoods they have built up thereon without cutting all grounds of belief from under their feet, – so that he may apply wisely that one remedy for human sin and woe with which he is entrusted, – surely it is desirable that he should have, not a mere empirical acquaintance with their language, but such an acquaintance with it as will enable him to understand those hereditary modes of thought and feeling and contemplation, which discover themselves by the peculiar modes of expression in their language. (Strettell 1851: 165–166) What conclusions should one draw from all this? It would be absurd, patently, to write off Comparative Philology as a cultural byproduct of missionary zeal and the White Man’s burden in the nineteenth century. Nor do we have to suppose that the Comparative Philologists would all

History and Comparative Philology 53 have subscribed to supremacist doctrines of the kind represented in the mid-nineteenth century by Gobineau’s Essai sur l’inégalité des races humaines. (Pott, for one, had criticized Gobineau in his book Die Ungleichheit menschlicher Rassen (1856).) Nevertheless, in Strettell’s comments we see an aspect of Comparative Philology which is by no means irrelevant to the institutionalization of the subject as part of an educational system. It is interesting to note the role of national stereotypes in Strettell’s explanation of why Comparative Philology (a ‘Continental’ product) is important. The British are honest toilers in the fields of linguistic observation. But the Germans are the great systematizers: and without the systematizing mentality (the ‘particular temper of mind’) the honest toil can in the end avail very little. It is the notion of systematization which in the end holds the key to all the extravagant claims that Comparative Philology was a new ‘science’. The conflation of science with system is one of the academic occupational diseases of Renaissance and post-Renaissance culture. What the Comparative Philologists did was to set up, in the field of linguistic inquiry, correlations (whether on good grounds or on bad) that made it possible to ‘reduce’ as much as possible to simple formulae. (Thus ‘x’ in Language A corresponds to ‘y’ in Language B, and that in turn to ‘z’ in language C. So x ⫽ y ⫽ z captures in quasi-algebraic form a linguistic ‘fact’.) What is significant is that when Grimm corrected his earlier confusion of sounds with letters, little or nothing had to alter in the tables of ‘correspondences’. The more formulae of this type that could be conjured up, the better: for the more ‘scientific’ the subject becomes. And the more ‘teachable’. Learning off the formulae becomes much more like learning mathematical tables. It is not a question of what was later called ‘an algorithm for reconstruction’, but of an algorithm tout court. ***

What the Comparative Philologists do not seem to have realized was that interlinguistic correspondence formulae (of the kind typified by Grimm’s Law) are no more than that; that is, formulae. In themselves, they contain no historical information and capture no ‘facts’ other than the observations and assumptions (whether accurate or inaccurate) that went into their formulation. But this was not stated clearly until Meillet pointed it out in his Introduction à l’étude comparative des langues indo-européennes (a book dedicated, significantly, ‘à mon maître, M. Ferdinand de Saussure’). There Meillet wrote: What the method of comparative grammar yields is not a reconstruction of Indo-European as it was spoken: but nothing more than a definite system of correspondences between the languages historically attested. (Meillet 1903: 27)

54 Roy Harris (The italics here are Meillet’s, as they are in the following passage.) The comparative grammar of the Indo-European languages is in exactly the situation that the comparative grammar of the Romance languages would be in if Latin were unknown: the only reality it deals with are correspondences between the languages attested. (Meillet 1903: 23) He goes on to argue that if no documentation of Latin had survived, any attempt to reconstruct Vulgar Latin solely on the basis of comparing the Romance languages would inevitably be seriously flawed. The linguistic object (‘proto-language’) thus reconstructed would be illusory, a product of the extrapolations by which it was produced. The lesson for IndoEuropeanists, in his view, is clear: on ne restitue pas l’indo-européen. But that is exactly what the Comparative Philologists thought they were about. Meillet’s lesson can perhaps be illustrated by means of an analogy. We know (because the development happens to be well documented) that most of the many corner designs on Delft tiles from the seventeenth to the nineteenth century are all derived from one original pattern (the fleur-de-lys). But if the historical evidence were missing it would be impossible to determine with certainty what the original pattern had been simply by comparing all the surviving tiles, even though obvious ‘family resemblances’ can be detected among them. One might even group them into sub-families on the basis of resemblance; but this could be done in various divergent ways. Nor would it be obvious whether or not one any of the surviving tiles bore the original corner design. The only way to convert systematic comparison of the tiles into a historical hypothesis about their relationship would be to introduce assumptions based on the probability of certain design developments (e.g. that simpler designs are earlier and more elaborate designs later). Even then, one would risk ending up by reconstructing a hypothetical original design that never in fact existed, even though it satisfactorily ‘accounted for’ the features present on all known tiles. To put the point somewhat differently, a correspondence formula needs an interpretation before its validity can be assessed. We see this clearly enough in the case of bilingual dictionaries. They will give us, say, the correspondence ‘dog ⫽ chien’ or ‘chien ⫽ Hund’. But what that means is open to a number of interpretations, all of them more or less controversial, as translators and semantic theorists are well aware. Similarly, historical linguists nowadays admit that what Grimm’s Law ‘actually was’ (as one of them puts it) is not uncontroversially given, and depends on what assumptions the linguist is prepared to make about the constituents of the original Indo-European sound-system. But even this goes much too far for some, since arguing about conflicting views of what Grimm’s Law ‘actually was’ already presupposes that Grimm’s Law actually was something; that is, that the name ‘Grimm’s Law’ indeed designates some historical reality – event(s) or process(es) – and

History and Comparative Philology 55 it merely remains to identify this reality. (It is rather as if modern historians already had a prior assurance that something called the ‘French revolution’ actually occurred, but were hampered by the total destruction of contemporary evidence in their task of determining who revolted against whom, when, where, why and how.) The currently fashionable view among linguists endeavours to maintain some intermediate position between ‘naive realism’ (e.g. treating, say, ‘Grimm’s Law’ as the designation of some complex unattested occurrences in the evolution of the Indo-European languages) and what has been called ‘formulism’ (i.e. treating it simply as the term for certain correspondence formulae relating attested forms). The intermediate position chosen by any given linguist usually requires a great deal of hopping from one foot to the other in order to show that the compromise reached is not a case of trying to have one’s Comparative cake and eat it. In terms of this distinction (which is reminiscent of the opposition between the ‘God’s truth’ position and the ‘hocus-pocus’ position in post-war synchronic linguistics), the early Comparative Philologists were undoubtedly ‘naive realists’, in spirit if not in literal commitment. Their modern descendants, who shrink from being categorized as ‘naive realists’, often opt for a compromise which amounts to claiming that correspondence formulae like Grimm’s Law do designate historical realities, but unfortunately realities about which we can hope to know little more than what the formulae themselves suggest. They commonly fail to see that this is also a case of having the Comparative cake and eating it. The whole difficulty with Comparative Philology as a ‘science’ was – and is – that it requires belief in a lost ancestral language in order not only to ‘account for’ but even to identify the (relevant) correspondences, which are in turn taken as ‘evidence’ demonstrating the real existence of the ancestral language you had to believe in to start with. ***

Intrinsic to the Comparative Philologists’ historical thinking about language were the assumptions (i) that ‘languages’ are discretely identifiable entities which can safely be treated as consisting of decontextualized sets of forms and (ii) that languages multiply by a process of internal division or branching, whereby one (or more than one) subpopulation of an originally monolingual community adopts, for whatever reasons, forms which distinguish it (or them) collectively from other subpopulations. This over time gives rise to two (or more) discretely identifiable ‘language’ entities where originally there was only one. In the least complicated case, it is assumed that what will happen is that the ‘parent’ language continues to survive alongside its (single) ‘offspring’, the latter being spoken exclusively by that subpopulation whose usage collectively and uniformly deviates from that of the original monolingual community. This process may be repeated ad infinitum, with or without the survival of the ‘parent(s)’ at each stage.

56 Roy Harris Although this ‘family tree’ or Stammbaum model is usually associated with August Schleicher (1821–1868), who made it explicit and compared it to Darwinian models for the evolution of plants and animals, one may regard it as latent in the work of the Comparative Philologists from the beginning. For it is otherwise hard to make sense of the significance they attached to patterns of agreement and divergence across the various languages under comparison. In brief, it would not matter that languages A and B share certain features in common as against language C unless it were supposed that this reflected an earlier process by which one ‘branch’ of the family diverged from another (or from the main ‘trunk’). Various points about this model are worth noting. As already mentioned, it assumes the discreteness of individual languages (A, B, C, etc.), which in turn presupposes the concept of a single community sharing a common language. It assumes, in effect, that there cannot be a language occupying a position ‘in between’ or indeterminate with respect to the separate branches of the family tree. On the other hand, it does not in itself afford any means of distinguishing differences between languages from differences between dialects: in this respect, all correspondences are theoretically on a par. Furthermore, it follows that correspondences themselves do not reveal which is a ‘parent’ language and which its ‘offspring’. (Thus it was theoretically possible that one of the extant Indo-European languages was itself the ancestor of the whole family – and some had seen Sanskrit in that role – or even that the original language was still spoken somewhere by a linguistic community yet to be discovered.) It was not until the second half of the nineteenth century that this whole model was challenged by the publication of Schmidt’s Die Verwandtschaftsverhältnisse der indogermanischen Sprachen (1872), which put forward the alternative view that linguistic development does not proceed in the organic manner suggested by the tree-and-branch analogy, but after the fashion of waves radiating out from a central point. The waves in question are linguistic innovations, and they spread out (geographically) from the point at which they originated, taking time to do so and weakening progressively as they get further and further from the point of origin. Since different innovations may originate at different points, or the same innovation at different points and different times, and since their success in spreading will vary considerably, there is no room in the ‘wave’ model for the simple branching process that characterizes the ‘tree’ model. In fact, the basic idea of languages as discrete, homogeneous entities is challenged by the notion of a continuum of linguistic variation in which it is difficult if not impossible to say where one language or dialect ends and a different one begins. ***

Historians of linguistics often invite us to think of linguists as patient contributors to a seamless continuum of ideas; a continuum to be understood

History and Comparative Philology 57 only by reference to intellectual ‘predecessors’ and ‘disciples’, structured by its own internal ‘discoveries’ and ‘influences’, ‘phases’ and ‘turning points’. (We are told, for instance, that such-and-such a scholar paid no attention to the fall of Napoleon, or the Russian revolution, so engrossed was he at the time in the libraries of Paris or St. Petersburg in working out the grammatical rules of some language long since dead, or piecing together the fragments of some ancient text. The topos here is that ‘scholarship is timeless’.) But scholarship is no more timeless than any other human activity. What different generations or different societies pick out as focal points of linguistic inquiry and pursue as ‘scholarship’ are reflections of broader concerns. Views about language and languages are invariably part and parcel of a view of society and, more generally, of the human condition. To this Comparative Philology was no exception. It is often said that Comparative Philology grew out of the Romantic movement. Certainly Grimm and others were fascinated by folklore and the Middle Ages, as their publications attest. But Romanticism was not a movement which lasted in the way Comparative Philology lasted. Comparative Philology grew directly out of and was nurtured by the political consciousness of the time and its particular requirements, especially those of various (sometimes competing) forms of nationalism. Indo-European philology was always as much about nations and peoples as it was about languages. This is still evident from the way Comparative Philology was being taught in the early twentieth century. To cite an example at random, Cambridge University Press published in 1906 an Introduction to Comparative Philology for Classical Students in which the author ( J.M. Edmonds) thought it necessary to devote a section to ‘the Classification of Europeans by Race’. In this the ‘Scandinavians’ are presented as ‘a tall Northern long-headed race’, as distinct from the ‘Iberians’, who are a ‘short, swarthy, Southern long-headed race’, the ‘Celts’, etc; and the author goes on to pose the question as to which of these groups ‘were the actual Aryan race’. Doubtless today all this would be condemned as ‘politically incorrect’, if not factually inaccurate (although the former charge would almost certainly take precedence over the latter). The point is that in the climate in which Comparative Philology was born and thrived, far from being politically incorrect, it was regarded as a matter of national honour to be foremost in establishing Aryan linguistic credentials. That was why Sir William Jones was posthumously elected a founder of Comparative Philology. As F.W. Farrar put it in his lectures to the Royal Institution of Great Britain in 1869, Jones was the ‘Galileo’ of the subject – a subject otherwise (unfortunately) monopolized mainly by Germans. Farrar expressed the pious hope that a new generation of English Comparativists ‘may save England from the discredit of failing and lagging behind in the splendid torch-race which she, most undoubtedly, had the honour to begin.’ Establishing the historical pedigree of the European languages was inevitably an enterprise laden with potential political implications. Jones clearly realized this. As far as he was concerned, Sanskrit was simply one piece

58

Roy Harris

in a bigger jigsaw which, when assembled in its entirety, would show that Indian culture was ‘Western’ culture. (Thus the British in India – of which he was one – could be seen not as colonialist adventurers pillaging the riches of the East, but as restoring a far-flung, ancient outpost of Aryan civilization to its rightful place within the corresponding political boundaries of the modern world.) János Gulya suggested a quarter of a century ago that although comparison of the Finno-Ugrian languages got underway earlier than that of the Indo-European languages, it never ‘took off’ because its historical lesson was one that people did not want to hear. Thus whereas the work of Gyarmathi on the affinities of Hungarian with other languages preceded the earliest publications of Rask and Bopp by more than a decade, it got nowhere because it ‘did not point in a direction in which the Hungarian officials, scholars, or public of the time wished to go’. By contrast, it was clear enough to the Indo-Europeanists which was the right direction to follow. It is relevant in this connection to recall that Rask’s Undersøgelse was written as a prize essay. The organizers of the competition (the Danish Academy) were clearly committed to establishing the academic credentials of a ‘Nordic’ mother tongue. So in one sense Rask’s conclusions were already prescribed for him in advance. This is not to accuse Rask of fiddling the evidence; but the fact remains that what Rask had to say was what his sponsors wanted to hear. (Denmark did not grant Iceland a measure of home rule until 1874, and the country did not succeed in becoming entirely self-governing until 1918 – exactly a hundred years after the publication of the Undersøgelse. Politicians in Copenhagen clearly had an interest in what conclusions philologists could establish about the relations between Danish and Icelandic.) More generally, the ‘Indo-European hypothesis’ corresponded exactly to the direction in which Europe wanted to go. It provided in particular a historical ‘destiny’ for Germany, Britain and France in current world politics, and one which it would have been difficult to supply on any other ground. It called languages to witness that certain nations were properly regarded as the rightful heirs to and leaders of an ancient cultural tradition – that of the Indo-European or Aryan race. That was doubtless of some relevance to the relentless insistence that Comparative Philology was a new ‘science’. When the latest science ‘proves’ what you wanted it to prove all along, it is doubly welcome. What is perhaps most important in all this, from an integrationist point of view, is the conclusion that where linguistics is concerned ‘history’ is not a neutral concept. It carries a theoretical bias that can skew the whole enterprise in a certain direction. The Comparative Philologists’ notion of linguistic change over time would have made no sense at all without the framework provided by the language myth of the Western tradition. From this myth the Comparativists took without serious scrutiny the ancient notion of languages as codes of forms and meanings that bind communities of language-users into

History and Comparative Philology 59 socio-political units. They took it uncritically because it suited their agenda. For, given that notion, it becomes possible to compare the codes as such, to sketch out hypothetical ‘histories’ based on the assumption that formal and semantic likeness between codes implies degree of relatedness, and then project those hypotheses back on to the communities in question. This is history derived from the premise that, as Meillet once put it, ‘langue commune suppose civilisation commune’ (Meillet 1925: 17). Comparative Philology provides an object lesson in how to make a linguistic tail wag a historical dog.

4

Word-stories Etymology as history Christopher Hutton

In 1969, Donald and Doris Fisher opened a store in San Francisco that would, in the words of Eric Schlosser in Fast Food Nation, ‘sell blue jeans the way McDonald’s, Burger King and KFC sell food’ (2001: 97). In thinking about a name for the store, and in the spirit of the counter-culture of the day, they decided to honour the Vietnamese general and victor of Dien Bien Phu, General Vo Nguyen Giap – and the GAP was born. An alternative account might run as follows. In 1969 the Fishers named their store the GAP, which stood for ‘groovy and practical clothing’, GAP. Or what of this version? In 1969 teenagers were rebelling against the older generation and the clothing store was named to reflect that: GAP derives from ‘generation gap’. These etymologies of the brand-name GAP are arguably equally plausible, in the sense that there is no way one can distinguish between them by the application of linguistic analysis. What kinds of external evidence might count in weighing the plausibility of these three accounts? The minutes of a company meeting? A statement by a company PR official? A statement by someone who witnessed the moment of inspiration? This paper is concerned with etymology as historical narrative. Integrational linguistics denies that we can find order and system at the level of ‘phenomenon’ – that order is always imposed. But it does suggest that we can find order at the meta-linguistic level, as an investigation of the ways in which lay speakers and professional linguists seek to impose categorical order on linguistic phenomena. In this sense, integrational linguistics acts as an anthropology of meta-linguistic practices and beliefs. This however is complicated by the possibility of our questioning the nature of the distinction between language and meta-language. The question that arises in this paper is as to the uses of etymology, with etymology being understood as speculation about the origin and original meaning of words and signs of all kinds. To take a further example of etymological speculation, in the 1960s McDonald’s began tearing down the original restaurants with their golden arches, but the corporation was worried by the impact of this on the company’s image. A design consultant and psychologist, Louis Cheskin, was hired and he argued strongly against completely eliminating the arches, arguing that they were of great Freudian

Etymology as history 61 importance for consumers: ‘According to Cheskin, the golden arches resembled a pair of large breasts: “mother McDonald’s breasts”. The company followed Cheskin’s advice and retained the golden arches, using them to form the M of McDonald’s’ (Schlosser 2001: 97–98). The etymology of the ‘M’ in McDonald’s is effectively evolving post hoc in Schlosser’s account. If we imagine that the GAP company ran a campaign in which they maintained that GAP stood for ‘groovy and practical’ this might become the official etymology, even if we had evidence to suggest that this was not the origin of the brand name. Once a sign and an associated story enters the social domain, we cannot limit etymology to strict temporal sequence. In this sense, the existence of the sign precedes its own etymology. In some contexts, signs are given normative etymological meanings, and these may circulate alongside the sign itself. For example, the Basic Law, the so-called mini-constitution of the Hong Kong Special Administrative Region of the People’s Republic of China (HKSAR), states that the design of the HKSAR flag, which carries a design of five bauhinia flower petals, each with a star in the middle on a red background, ‘implies that Hong Kong is an inalienable part of the motherland. The five stars on the flower symbolize the fact that all Hong Kong compatriots love their motherland, while the red and white colours embody the principle of “one country two systems” ’ (Basic Law 1992: 87–88). What is the status of this reading of the flag, of which few citizens of Hong Kong are aware? This paper is concerned with the reading of Chinese characters as history, legend or myth. If ‘everybody is a linguist’ and ‘necessarily so’ (Harris 1998: 20), then what might an integrationist have to say about etymology and the academic linguist’s claim to descriptive rigor? In the context of integrationism, etymology might simply fall prey to the general critique of semantic determinacy. Etymological analyses are crucially based on determinate formmeaning correlations, and even from the point of view of conventional linguistics the etymological information provided by institutional authorities such as the Oxford English Dictionary often appears on close examination to be highly speculative. From an integrational point of view, there may be no fundamental difference in status between academic etymology and lay etymology. In the case of Chinese characters, a range of discourses is on offer. Chinese characters can be seen as the academic property of highly trained Sinologists, but they have also been a playground of the imagination for thinkers and intellectuals with little knowledge of Chinese. In the introduction to his Analysis of Chinese Characters, Raymond Blakney asked the rhetorical question: was it worthwhile for the ordinary foreign resident of China to learn Chinese characters? His answer was an unequivocal one (1935: 10): The Chinese have poured their life and soul into these written symbols; they never developed other sciences, but have written their reactions to life and things into this ordinary medium of their expression. It will be

62 Christopher Hutton readily admitted that the first step to be taken by the newcomer to China is the acquisition of the spoken dialect where he is to reside [. . .]. But if he will ever get at the fundamental characteristics of the Chinese peoples, he must study the much-revered characters, for these reveal the history, psychology and even the entire scientific bent of the people who made them. [. . .] The Chinese character may be said to be a key to the minds of the races of Eastern Asia, and they are no small portion of the entire human family. In their encounter with Chinese characters, Western intellectuals have struggled with the fundamental semiotic challenge they offered. Responses have ranged from the enthusiastic modernist embrace of Ezra Pound, for whom Chinese characters were vital moving signs, to the iconoclastic attacks by Protestant missionaries and modernizing Communist reformers (DeFrancis 1950). In general, the Chinese are presented as the people of the picture, their language the creation of a ‘visual rather than aural experience’ of the world (Lai 1980: 5). The Chinese writing system is said to ‘appeal mainly to the eye, and hence has more of orthography in its than it has of phonology, the main thing in the study of alphabetic languages’ (Chalfant 1906: 4). The obvious symbolic contrast is with the Jews, the ‘people of the book’ and the historical bearers of monotheism and the religious prohibition of images (Exodus 20: 4, Deuteronomy 4: 15–17). Etymologies are often stories, ‘just so’ stories, depicting primal or emblematic cultural scenes, and are held to retain historical or culturally fundamental information in compressed form: ‘Familiarity with Chinese modes of thought, methods of work, and social customs aids much in determining the significance of certain old signs.’ Thus ‘it is a tradition amongst the Chinese that the East Palace was the hall of audience and place for administering the laws. When we found an old symbol for “judge” composed of the elements “east” ( ) and ( ) “to speak” we see an appropriateness in the combination from the ancient custom of “judging” in the “east” ’ (Chalfant 1906: 3). Chinese characters can reveal culturally strange ideas: a common etymology of the character ‘to see’ is light ‘streaming from the eye, [. . .] based on a peculiar notion still held by the Chinese that when blind “no light comes out of the eyes” ( )’ (1906: 10). Using examples like these, it is often asserted that the study of Chinese characters gives an insight into the Chinese mind or Chinese culture. Given that this culture had or has its own logic and principles of association, the critique of Chinese etymology from the point of view of Western linguistics might be seen as ethnocentric. Western scholarship took a wide range of views of traditional Chinese linguistics. The Shuowen jiezi by Xu Shen was the subject of extensive comment. This work was compiled during the Eastern Han Dynasty, and completed in AD 121: ‘It was the first Chinese character dictionary and the first work of characterology in Chinese history’ (Yin and Rohsenow 1994: 341). This work not only formalized the organizational and semiotic principles of

Etymology as history 63 the modern Chinese dictionary, but also ‘had a significant impact on the Chinese language’ by checking the proliferation of variant characters (Creamer 1991: 2597). Chalfant (1906: 6, 19) rated the work highly as an etymological treatise, ‘the frequent flights of imagination’ made by the author notwithstanding. The debate about the etymological narrativizing of Chinese characters concerns crucially the question of how we determine what is a ‘flight of the imagination’. One clearly identifiable Western tradition can be termed Christian etymologizing, which began with the initial Jesuit confrontation with the Chinese language (Edkins 1876: viii). This tradition of reading Chinese characters and Chinese texts seeks to reconstruct the Christian message hidden or obscured by the passage of time and by the loss of the appropriate interpretative key. The philologist-translator’s task was to clear away the accumulated distortions in the text in order to reach the lost Christian – or at least monotheistic – message. This method of philological reconstruction came under attack from Orientalists who argued that Chinese texts should be seen on their own terms, that is, that the Christian reading did violence to the proper understanding of the text which required the application of objective philological methods to be understood (Watters 1870: 2–5). Recent publications in this Christian tradition thus represent the continuation of a missionary tradition of seeing the Chinese as originally monotheistic, and as having gradually declined to polytheistic idol worship. The ancient Chinese ‘served only one God, had no myths or idols, and kept a strict moral code’ (Kang and Nelson 1979: 2). The proof of this has been ‘stored in special characters that were in use hundreds of years before the first page of the Bible was written!’ (1979: 5). Among the key examples for this analytic technique are the characters for ‘boat’ chuan ‘to create’ zao , ‘devil’ gui and ‘tempter’ mo . In each case the character can be broken down into a number of atomic components. These components are then combined to create a story. This process of combination gives linear narrative structure to the non-linear character. Kang had found a missionary textbook where the character , meaning boat, had been analysed as follows: a vessel; eight; and mouth or person. A comment followed that, interestingly, Noah’s ark, the first great boat, had just eight passengers: Noah and his wife, with his three sons and their wives (1979: xii). Modern Sinologists and academic linguists would of course have little time for such curiosities. The nineteenth-century missionary scholar Joseph Edkins, whose etymologies were attacked in his time for lacking scientific rigour, drew the line at reading into the ‘boat’ character an allusion to the ‘Deluge’: ‘The day for this mode of explaining the Chinese characters has gone by’ (1876: viii). Nonetheless, it is interesting to push a little further and ask specifically what is objectionable about this methodology. If pressed, the linguist might argue that the interpretation is overdetermined by the target text (or underdetermined by the evidence). Given a number of Chinese characters it is likely that at least a portion of them, when broken

64 Christopher Hutton down in to their composite forms and their possible narrativized linear combinations, will in some way match up to some aspect of the story of Genesis. The problem is that the exact same argument can be used against any narrativized reading of Chinese characters. If we take the widely held view that Chinese characters are keys to Chinese history and culture, then we can expect that given a number of Chinese characters we will be able to discern a number of potential narratives, at least some of which will confirm to notions of early Chinese society, or can be read as encapsulating a primal cultural, social or mythic scene. Commentators differ in their dogmatism. F.W. Baller took a relaxed attitude to the exercise of the imagination in interpreting Chinese characters: ‘It is safe to say that as a good deal of imagination must have been needed to make up many characters in the first instance, so imagination must be allowed free play in determining the derivation of many of them now.’ If we allow our imagination this liberty, ‘Chinese characters may become pictures, and the mind may be hung round with them as a picture gallery’ (1913: 15). T.C. Lai suggests that if we assume that ‘the symbols are a way of communicating with ancestors and the records thereof, one may arrive at some understanding of them according to one’s imaginative ingenuity’ (1980: 7). A similar invitation, though more universal, is offered by Harbaugh (1998: [iii–iv]), whose study of Chinese etymology is based on the Shuowen jiezi . He notes that many etymologies are based on stories and are ‘speculative’ in relation to how ‘the elements combine to suggest a certain idea’. He notes that he ‘resisted the temptation to insert my own stories, though readers should by no means feel similarly bound’. Narrativizing characters is also a well-known memorization technique, and scholarly etymology is not always given priority over mnemonics (Keelan 1967: vii). A typical Chinese mnenomic device as learned by children is ‘wild goose is a bird by the river ’ (Yin and Rohsenow 1994: 95). Some well-known examples of the presentation of characters as representing primal scenes are an (‘peaceful’, ‘security’), gia (‘household’, ‘home’, ‘family’). The character an is seen as a ‘woman’ under a roof, and there is a general consensus that this character offers a narrative definition of ‘peace’ in a Chinese context: ‘After a man’s marriage, a family is formed, living in peace and contentment. The ancient form of this character depicts a woman sitting on kneeled legs in a room, indicating peace’ (Shi 1997: 475). Li (1993: 1) offers a more elaborate story: ‘In a silent house, with her hands folded on her bosom, a woman [. . .] kneels calmly on a straw mat as the ancient people used to do. Hence its original meaning was “stable”, “comfortable” and “safe” ’. Xie (2000: 318) sees in the character the image of a woman managing the home, a sign of the absence of war and natural disaster. Some accounts stress the presence of one woman, rather than several: ‘One woman in a house (An essential condition of peace in China)’ (Chalfant 1906: plate xxii);

Etymology as history 65 ‘Man conceived the idea that to attain peace he should have only one woman under the roof or confine her within the house’ (Ong 1980: 4). Wu (1990: 142) however casts doubt on the ‘feminized’ reading of this character (citing the authority of Li 1977: 57–58): The original meaning of this character is to be at home and inactive. Accordingly, the definition found in Shuowen is (to stay at home) (343). The present sense of ‘peace’ or ‘tranquility’ is derived by extension of this original notion of ‘to be at home’. And by further extension, ‘security’ or’ being safe’ is acquired. Often the case in the Chinese language, when is used as a component, it does not represent ‘woman’ exclusively; instead, it may represent a person or people. The meaning of the character is thus that the person is at home and therefore at rest or at peace, rather than out in the world. However, to treat the character as ‘person’ (and to reject the patriarchal reading of it as someone ‘in a kneeling posture befitting that of a female who held a low position in ancient society’ (Lai 1980: 38) puts into doubt a whole further set of character stories, for example, ‘woman’, read as ‘a woman holding a broom symbolising domestic work – hence a married woman’ (Lai 1980: 41)). Running counter to these historical speculations about early or ancient Chinese society is an equally historical vague tradition of assigning matriarchal readings, that is to see relics of a pre-patriarchal social order. This kind of historical speculation is encouraged by the presence of matrilocal and matrilinear social systems among present-day minorities in China. This reading can lead to the conclusion that the alphabet is aligned with patriarchy and oppression (Shlain 1998: 181ff.). The star witness for this tendency is the character xing, which in modern Chinese means ‘family name, clan, family, people’ and is classified under the ‘woman’ radical . Blakney offers a detailed historical narrative (1935: 68): refers to the very ancient times in China when polyandry was the marriage system. This system is still in vogue in Tibet and parts of Mongolia, where, presumably, the hard climatic conditions make it difficult for the sexes to maintain their numerical equality. The character is phonetic from but also ideographic in that it adds the woman radical, in indicating the derivation of the surname. In those days, it is said, children knew their mother, but seldom knew who their fathers were, and hence took their surname from the former. The ancient Chinese justified this system by appealing to the example of the three divine rulers, who, they said, got their surnames from their mothers. This same story appears in Ong and Tan (1980: 85): ‘The character , comprising (woman) and (born), literally means “born of woman”. It suggests that in some remote, forgotten era man, born of woman, got his name from his mother’.

66 Christopher Hutton The character

gia is commonly understood as a pig under a roof:

In distant ages past, peoples lived a nomadic life. Later, they came to know how to feed the wild animals they caught to meet their needs for life. As they also took the animals as a symbol of their property, they came to settle down. Such families would have animals fed in their rooms. This led to the creation of the character [. . .]. National minorities in the southwest areas of the country still maintain the practice of feeding pigs or other animals indoors. (Shi 1997: 490) Lai (1980: 58) interprets this character simply as a ‘pig in a hut – that is home’, noting that the pig ‘was probably the earliest domesticated animal’ and that the character was probably invented ‘in a fairly advanced stage of an agricultural society’. Looking back through Western readings of this character, a range of views can be found: Human dwelling, says the [Shuowen]. By extension, family. [. . .] The pigs live around the houses of the Chinese countrymen, and even enter in them, as well as the dogs. The street-cleaning and privy-emptying are left to these two animals (Wieger 1927: 173). Wilder and Ingram give the ‘pig under the roof’ story, but also note variant characters (1934: 76–77): The seal character of 100 B.C. shows that it is a picture of a pig, the upper line for the head, the left side showing the belly and legs, the right side showing the back and tail. In the Oracle bones this appears as a pig ready for sacrifice in the home, where ancestral offerings were made, ( J.M. Menzies). This explanation is not given in the first edition of this book. An early form shows three people under a roof, man, woman and child made a home (Chalfant XIX). Chalfant in fact describes this character as ‘origin obscure’, noting that the ‘pig under roof’ form has been the sole form since the Han Dynasty (c.100 BC): ‘Earlier forms show various objects under roof. The most reasonable is “three persons under roof” – a very early form which I have taken as the probable origin (1906: plate xix). Blakney also offers a sceptical reading (1935: 132–133): The character poqik is indeed famous, since it represents the Chinese home as complete when there exists a pig with a roof to cover him. That pigs do make their homes with humans among the agricultural folk of China is well known. In some very old forms, we have the dog represented rather than the pig. Both of these animals are truly domestic in the rural Chinese home and to the ancient Chinese lexicographers, it may have seemed that using

Etymology as history 67 the pig or the dog as the symbol of home would most fittingly designate the meaning intended by this character. However, the fact remains that any suggestion of human form is left out of the scene altogether. Blakney quotes a suggestion from the Shuowen that the phonetic is a remainder of a once used more complex phonetic and that this would explain its presence. Although the meaning of this character is not controversial (‘an abode or place in which one lives’, by extension family ‘since in China it was always a family that lived under one roof’, Wu 1990: 102), the interpretation of the co-component (the putative pig) has given rise to extensive discussion. Where is the pig exactly? ‘Some commentators say the pig is really outside the house’ (McNaughton and Li 1999: 109). An alternative account suggests that, once domesticated, ‘the pig brought man no domestic trouble and was allowed freedom to wander about in the house’ (Ong 1980: 7). Xie (2000: 316) argues that this character depicts a particular model of dwelling where the house is raised off the ground with the animals kept underneath, so that the basic or original meanings of this character include ‘household dependents’, ‘household’, ‘family’, ‘clan’ ( , , ). Ye (2000: 125) considers that the household would need to own a pig before it be called a family or . This debate can illustrate the tension between two tendencies in the interpretation of Chinese characters: the mytho-poetic or narrative tendency as against the anti-iconic, anti-ideographic tendency, the latter much more likely to propose that a sub-component serves a primarily phonetic function. Wu rejects the Shuowen’s reading of as derived from a phonetic : ‘Contractions in Shuowen can not be readily accepted as well-grounded for their so-called contractions were actually their original word or a complete misinterpretation of the component’ (1990: 102). The homely ‘pig under the roof’ gets short shrift (1990: 102–103): ‘If pigs were permitted to roam through homes, then why not chickens, dogs and other animals. Why was the character for chicken or dog or a combination of animal characters not chosen to represent those who lived together with man under the same roof?’ Wu’s analysis sees as created with as its meaning-bearing component (or ‘significant’). then comes to stand for as the ‘phonetic force’ in the character . For Wu, the story of the ‘pig under the roof’ – sacrificial offering, pre-modern garbage disposal device or domestic companion – is by implication a simple folk etymology, though we should note that the iconic account of the roof is not questioned. Many of the Christian etymologies given by Kang and Nelson are shared by this academic-popular tradition. A popular character for interpretation as linear narrative is xin , meaning ‘honesty’, ‘good faith’, ‘credentials’. This character ‘is formed by the combination of the characters of a man and words, which speaks well for the honourable truthfulness of the ancient Chinese’ (1979: 29). This reading has a substantial tradition. Blakney notes that the character ‘is a celebrated example of ancient Chinese idealism’ (1935: 91). Ezra Pound was also taken with this character which he read as ‘man standing by his word’ (1969: 193). In his presentation of Ernest

68 Christopher Hutton Fenellosa’s notes, Pound gives ‘[m]an and word, man standing by his word, man of his word, truth, sincere, unwavering’ (1936: 40–41). Drawing on the Shuowen, Wieger gives a little semantic depth to the term: ‘True words, and, by extension, the effect produced by these words upon others, truthfulness, faith, confidence; [. . .] a man and a word. In still more ancient form, a word and a heart; words coming from the heart and appealing to it’ (1927: 73). This character also has the meaning ‘arsenic’ (Wilder and Ingram 1934: 69). Wu (1990: 132), who demystifies the character , is indulgent of the ideographic reading of . Its original meaning was ‘envoy’ or ‘messenger’, and this was extended to the object used to pass on the message (‘letter’), and ‘by further extension, it is believed that the idea of “trust” came about because the word sent was supposedly sincere and trustworthy (sincerity) (Shuowen 93)’. It also has the meaning of ‘a token or promise or pledge’, which is contextualized by Wu in a practice of ancient China where the messenger from the king would carry half a jade piece, the other half of which was held by the feudal lord. Wu’s reading is that this character is that it is an Associated Compound of and , that is the man who brings the message or word to another. Boltz (1994) however takes the denarrativizing mode of criticism to its logical conclusion in his attack on the ‘man standing by his word’ reading of the character : ‘the simple juxtaposition of “man” and “speech” is suggestive of all kinds of meanings besides “true, honest”, e.g. “orate”, “cavil”, “harangue”, “lecture”, “monologue”, even “linguist”, to name only a few’. Historically, there is ‘no provision for characters to be formed as combinations of two or more constitutive elements none of which has a phonetic role’ (Boltz 1994: 148). The character is a or ‘shared phonophoric’, that is a member of the ‘set of Chinese characters all of which have the same phonophoric [‘soundbearing’] element, but differ in their other graphic components’ (1994: 180). This anti-ideographic stance has quite a long tradition in Western sinology, from Du Ponceau (1838), down to the discussion in DeFrancis (1984: 144–148). Frequently, however, scepticism about some narrativized readings of Chinese characters is inconsistently applied. For example, Chalfant (1906: 10–11) rejects the ‘pig under the roof’ etymology of , but explains the elements ‘work’ and ‘mouth’ in the characters for ‘left’ and ‘right’ as follows: ‘Why the sign (“labor”) should have been added to strengthen the idea of “left” is hard to say, unless the left-hand was once used to such an extent as to make it the important member for work. “mouth” was added to emphasize “right-hand”, possibly because that member was used in eating.’ Hager (1801: lxxi–lxxiii) is a good example of the selective sceptic, dismissing the assertions of Fourmont (1737: cap xvi): As to the general explications subjoined to the 214 elements in this work, though each one commonly retains some relation either directly or indirectly, when used in compositions, to its original import, it nevertheless would be improper to lay this down, with Fourmont, as a universal rule, much less to maintain, that the knowledge of the meaning of each of

Etymology as history 69 these elements will suffice to understand the groups they compose, and consequently the sense of every character. That this universal assertion of Fourmont is altogether groundless, I could shew by almost as many instances as there are composite characters; but of their number I shall produce one from his own book, where adducing the character kuang, which signifies folly, madness, it would follow that, according to his rule, this sense must be obvious, because the group is compounded of kiuen a dog, and vang a king. Amongst dogs reigns madness, he says; consequently in the coalescence if these two characters every person should know that kuang means madness. Hager then turns to Fourmont’s example of ‘island’, composed of ‘bird’ and ‘mountain’. This could mean ‘a flying mountain’, or a ‘mountainous bird’, or a ‘bird and mountain combined’, he objects. However, Hager accepts of the character ‘sea’ that it is appropriately combined, being made up of the elements ‘mother’ and ‘water’: ‘the seas being deemed by the Chinese the parent of all water’. However, he notes that ‘every one will see how many other meanings these two characters, when joined, may imply’ (1801: lxiii).

Is there a pig under the roof? The difference between academic etymology and lay etymological theorizing does not, arguably, amount to a difference of methodology. Although the professional etymologist may have a good knowledge of the language and sources, and may apply that knowledge in the pursuit of etymological readings, that erudition cannot in and of itself inform the decision as to whether apparent similarity of form and meaning can be seen as a result of a lost historical identity. In a similar way, no amount of legal expertise and experience and knowledge of legal theory (or linguistics) can decide whether a Bug’s Bunny T-shirt bought during an outing to the sea-side really is a souvenir or not (Harris 1981: 190–191). The lay etymologist however is often willing to embrace a historical scenario which current academic opinion or scholarly rationality rejects as a matter of instinct, for example, that Chinese characters encode aspects of the story of Genesis. Interestingly, whilst much discussion has focussed on whether there really is a pig under the roof, much less attention is paid – no doubt for sound philological reasons – to whether there is a roof over the pig. Is it odd to say that there is no pig under the roof, but that nonetheless the roof is still there? In many ways, the general popular-scholarly debate today is being conducted in much the same way as it was at the beginning of the nineteenth century. The scathing sceptic in one context is the credulous myth-maker in another. Herbert Allen Giles, the virulent critic of much late nineteenth century Western sinology and its etymologies, devoted a slender volume to showing how Masonic imagery can be found in Chinese characters and classical texts (1880: 18ff.).

70 Christopher Hutton Within Chinese scholarship, the study of oracle bones and tortoise shells ‘was, in effect, a termination of the traditional mode of study of Chinese characters [ ] that had held sway since the eighteenth century’ (Zhu 1998: 14). The etymologies of the Shuowen are now set against a vast body of work in Chinese paleography ( ), but they retain a central position. Even the analyses now viewed as dubious are accepted as having educational or cultural value, not least because of their ‘mnemonic value and insight into Chinese culture’ (Harbaugh 1998: [v]). But the readings of the characters are constrained by a pre-existing idea of what Chinese culture is. Thus were it a historical fact that an early branch of the primitive religion or protoFreemasonry did spread to China, or that Biblical myths had been encoded into Chinese characters, no academic or ‘serious scholar’ would reconstruct such a scenario. Academic etymologizing involves a systematic attempt to distinguish historically grounded or cognate resemblance from accidental similarity. Chinese characters are read in complex ways as having both a pictorial and a phonetic element, but the game is fundamentally the same in any etymologizing, that of restoring lost connections, erasing the passage of time, bringing what has become blurred back into focus or making what has been fragmented whole again. Is the Chinese word for ‘human being’ derived from a root meaning ‘close’, given the relation between the written Tibetan term for ‘kinsman’ and a related root meaning ‘close’? (Sagart 1999: 163). This is not ultimately a question that can be settled by the application of linguistic methodology. For the integrationist, etymologizing is primarily a meaning-giving activity, carried out by a whole range of ‘meaning-makers’. Etymologizing is thus not neatly to be distinguished from other forms of linguistic analysis. Any reduction of language to a fixed code involves complex judgments of similarity and difference, and the criteria by which we make judgments of similarity and difference are not given in advance of the problems which are to be addressed. We can learn to do linguistic analysis, just as we can learn to ‘see’ or ‘read’ a pig under a roof, or see first a pig under a roof and then – having been corrected by the etymologist – just a roof with a sound-bearing symbol under it. It is not nature that abhors a vacuum, rather linguists of all kinds, both lay and professional. Is the implication of this discussion that our meta-linguistic anthropology should remain neutral or relativistic in the face of etymological theorizing? Are all views of language equally valid cultural constructs? In many contexts etymologizing can be seen as a harmless or amusing ‘language game’. But speculation as to the true origins of words often forms part of the search for the lost unity of a language, and serves in a quest for the restoration of its historically authentic conceptual unity. The belief that social, linguistic or cultural phenomena exist in the present as corrupt or fallen versions of past organic wholes has been demonstrably destructive when activated through political ideologies.

5

Language: object or event? The integration of language and life Michael R. Walrod

Introduction There is a strong tendency to view language as an object (or an inventory of objects such as lexemes) rather than as verbal communicative behavior. This tendency has given rise to the metalanguage of linguistics, and the metalanguage in turn perpetuates the language-as-objects mindset. While acknowledging from the outset that the metalanguage is necessary in order for us to reduce the infinite complexity of the verbal communicative behavior to manageable categories, it is important to recognize that we do this for convenience, at times at the expense of accuracy. We are inclined to believe that our metalanguage reflects what is obvious about language, yet there are categories and dichotomies (e.g. noun, verb) that we assume for which there are no exact or approximate equivalents in the lexicons of other languages, particularly in preliterate communities. It follows that the categories and distinctions are not obvious to speakers of those languages, or are not considered important enough to have names. The active-passive dichotomy was once believed to apply universally, and it was employed in descriptions of voice in Tagalog. In retrospect we see that scholars were hindered in their understanding of Tagalog grammar by the use of this metalanguage. It took decades until other analyses were proposed and gained credibility, analyses more along the line of Rashidi’s (1997) treatment of Dari. It stands to reason that other metalanguage that we assume to be adequate may in fact be obscuring our understanding of linguistic reality. I will list several terms below that have some serious shortcomings. Although the shortcomings have been observed and criticized by others, the terms are still used quite consistently, and they continue to influence our thoughts. In fact, it seems to be impossible to even discuss this problem without attributing to the conventional (“inventory of”) symbols some independent existence. This is a philosophical conundrum, but one that may be overcome to the extent that we can remember that we are employing such metaphor and metalanguage for analytical and descriptive purposes. The lexemes themselves, and other “units,” actually exist in a dynamic and ever changing present, living

72 Michael R. Walrod in the collective consciousness and in the ongoing dialogue of the members of a speech community.

Misleading metalanguage Lexeme as object or unit This is a useful concept, but it does encourage us to continue to think in terms of an inventory of units that comprise the lexicon. Each unit is thought to have or contain a certain meaning or range of meaning. As such, each unit can be cited as an entry in a dictionary, and a canonical definition can be given, which many people consider to be prescriptive of its appropriate usages. We are tempted to think of the lexicographer as the curator of a sort of museum of objects that may be displayed and scrutinized. In actual fact, the lexicographer is more like a historian and a news reporter, since chronicling the latest lexical developments within a speech community is a never-ending task. The meaning of an utterance or text is not simply the sum of the meanings of the words. It is not the words that construct or rule the meaning of a communicative event. Rather, the words are subordinate to such events, and the histories of their usages in a speech community continue to modify and augment and reshape the lexicon, by virtue of modifying and augmenting the cognitive and neural relational networks of the speakers. Harrison (1997) has already debunked the myth of the “container” metaphor of meaning, but it warrants reiteration as preamble for the following sections. Evidence of the creative and ever-changing nature of the lexicon abounds. New lexemes/words are created, and old words are used with new senses. Apparently a new lexeme “chick flick” is now being used widely. I suggest it is new because the definition is still under negotiation. I would have assumed that it described the content as being “adults only,” but I overheard a conversation recently indicating that it refers to a movie of a sentimental nature that appeals to girls. And as the new term moves toward sedimentation and inclusion in the wider lexicon, the details yet being worked out are whether a man would attend such a film alone, or only if he were on a date. One of the best places to look for lexical innovation is in the speech of young people, particularly in the area of the modifiers they are using to express strong approval or disapproval. We have seen words like “neat” and “cool” and even “tubular” take on such senses, and be the word of choice for a period of time. Now the words “majorly” and “totally” have assumed similar roles in our North American English speech community. When we discuss lexical innovation we are not often talking of the invention of new concepts, but rather the expression of known ideas (or recombinations of ideas) with new terms or phrases. Sometimes the changes are not semantic, but are morphological or grammatical. The phrase “on the average” was the standard expression for at least the first half of my life. My dated edition of

The integration of language and life 73 Webster’s dictionary (1977) lists only the example “on the average,” but now the consistent expression is “on average.” Sentence as the core subject matter of linguistics This is worth mentioning not because many people currently believe that sentences are the core subject matter of linguistics, but because this view gained such wide acceptance for such a long period. Longacre for one has been vociferous in denouncing the notion (1983: ch. 8), and Gleason recently asserted that “texts constitute the core subject matter of linguistics” (1997: 185). However, this notion of the centrality of the sentence in linguistics has encouraged many to focus their analytical efforts on sentences in isolation, oblivious to the constraints of text and context. These errors concerning words and sentences as basic building blocks (units constitutive of meaning) have laid the foundation for other theory construction which is also flawed (see the next two sections). Sentence meaning versus utterance meaning This commonly accepted distinction suggests that a sentence-meaning is independent of a communicative context, but an utterance-meaning is not. In order to justify such a distinction, one needs to subscribe to the two errors cited above, namely the “museum piece” view of lexemes, and the focus on sentences as the core subject matter of linguistics. Thus the sentence could be examined in isolation, and “the meaning of a sentence is determined, at least partly, by the meanings of the words of which it is composed” (Lyons 1995: 32). Lyons does allow that “most language utterances, whether spoken or written, depend for their interpretation – to a greater or less degree – upon the context in which they are used” (1995: 4). But although he acknowledges the constraints of context on utterances, yet he sees sentences as units that can be examined apart from such contexts. One definition of sentence-meaning is that which one would ascribe if receiving an envelope containing one sheet of paper with only a sentence typed on it. But this begs the question of how one could ascribe any meaning at all to a piece of paper with ink marks on it. There is nothing intrinsic in that object which communicates semantic or propositional meaning to us. We can construe a meaning only because we recognize the conventional graphemic representation, and we can imagine the most likely context in which the sentence would be written or spoken. We draw on our data bank of cognitive experience and knowledge. This suggests that there is no tenable distinction between sentence meaning and utterance meaning, since both are dependent on a context in order to be meaningful at all. The context may be real and immediate, or a cognitive distillation and abstraction from previously experienced contexts. But the process of meaning-construal is the same. We process both by means of the neurocognitive relational networks

74

Michael R. Walrod

(Lamb 1999) that have been imprinted with the information received in our previous communicative experiences. To illustrate the difficulty of holding to the notion of sentence meaning as that which can be received by words and grammar apart from context, consider: “He was a converted bike rider.” Possible interpretations of this sentence are: He is a former bike rider who became something else, for example, a horse rider; He is a former horse rider who became a bike rider; He decided to use a car, a bus, or walk, rather than ride a bike; He decided to ride a bike rather than use a car, a bus, or walk; He was a bike rider who had a religious conversion experience; He was someone who rides converted bikes (i.e. modified); He was a converted bike rider (any of the above senses), but no longer is (i.e. he reverted or recanted). We could choose to compare the example sentence above with “Was he a converted bike rider?” in order to illustrate something about the effect of grammar on sentence meaning, but we would not have solved any of the ambiguities cited. For each of the proposed interpretations, we have imagined a plausible context in the real world that enables us to make sense of the sentence. And we have not even considered any possible metaphorical usages of such a sentence. The literal versus figurative dichotomy The dichotomy of literal versus figurative language is similar in kind to the misleading concepts discussed earlier, and it goes beyond them in obscuring our insight into language as communicative event or as ongoing dialogue of a speech community. Fiumara (1995) has observed our western propensity for focusing on spoken or written language, rather than on language as communicative event involving speaking-and-listening or writing-and-reading. This focus on the product or the expression is a decontextualizing move that is likely one of the consequences of literacy. In the Ga’dang speech community (a predominantly oral society in the northern Philippines) (Walrod 1988), the verb kun can be used appropriately to mean “say” or “do.” There is a similarity between verbal action and any other action. This is in contrast to our view, in which “say” is a unique verb that refers to producing an oral expression. The invention of writing allowed us to take an objectively distanced position from language, and consider it as the object of analysis. We speak of language being “reduced to writing” and this terminology is very appropriate. Writing is a great reduction compared to the infinite complexity of detail in

The integration of language and life 75 oral communicative events. Transcription made language more manageable, and led to lexicons and grammars. Just as written texts have a durable or canonical form, grammar and the lexicon came to be viewed in this prescriptive way. Perhaps nowhere is this more apparent than in the traditional view of metaphor. In Mark Johnson’s work on metaphor, he traces the development of the traditional view back to Aristotle’s assertion that it consists of giving the thing a name that belongs to something else. Johnson sees problems with this view: “First, the metaphoric transfer is located at the level of words, rather than sentences. . . . Second, metaphor is understood as a deviance from literal usage . . .” ( Johnson 1981: 5). He suggests that since Aristotle there has been a fatal separation of literal versus figurative, which influences our understanding of metaphor to this day. Aristotle’s view presupposes the traditional notions of words and sentences as described earlier, and also perpetuates these notions. It has even spawned a great deal of normative discourse over the centuries about the appropriateness of the use of metaphor, as well as its truth value if any. Johnson (1981: 12) refers to the linguistic theory of Hobbes as the “literal-truth paradigm” which he summarizes: 1

2 3

The human conceptual system is essentially literal – literal language (“words proper”) is the only adequate vehicle for (a) expressing one’s meaning precisely, and (b) making truth claims, which together make possible correct reasoning by the philosopher. Metaphor is a deviant use of words in other than their proper senses, which accounts for its tendency to confuse and to deceive. The meaning and truth claims of metaphor (if there are any) are just those of its literal paraphrase.

But all of this discussion rests on the notion that the units of the lexicon have a correct and canonical form, and can be used appropriately to produce sentences with literal meanings. But Kant and Nietzsche, and more recently Johnson, Lakoff, and Gibbs, have all done work that challenges some of these traditional views. Gibbs in particular (1994) has shown the unlikelihood that there are separate processes of production or interpretation at work in the cases of literal versus figurative language. He provides some empirical studies that support the claims of Ricoeur that a single theory of language/interpretation will suffice for both text and metaphor (Reagan and Stewart 1978: ch. 10).

An integrational neurocognitive approach The problems inherent in the traditional metalanguage discussed earlier are resolved in an integrational neurocognitive view of language. The notion of linguistic units, such as morphemes and lexemes, is a convenient fiction for purposes of analysis and description. A more accurate

76 Michael R. Walrod view of language is that of interactive communicative events involving neurocognitive relational networks rather than linguistic objects, units or symbols (Lamb 1999: 381–3). It is like phenomenological reduction to speak of linguistic units. How else could we make the data manageable? Just as an artist reduces the infinite detail of physical reality to fairly broad brush strokes, reflecting that which is perceived to be more significant, so the linguist selects strategic “junctures” in the neurocognitive relational networks (used in text production and interpretation) to identify as units. We need such metalanguage in order to talk about language, else we would be restricted to paraphrase. We need to be able to reduce, to generalize, and to categorize. Yet our theory is more in tune with reality if we recognize that the so-called units of the various strata of language (e.g. phonemes, morphemes, or lexemes) are categories for heuristic and descriptive purposes, rather than objects that can be accessed in mental files or inventories. The fact that we can express these so-called units in written form is not an evidence of their reality as objects, but rather is an evidence of our ability to reduce the infinitely complex reality of verbal communication to symbols and categories, the brush strokes of our art. The following concepts are useful for countering some of the weaknesses of traditional metalanguage. 1

2 3

4 5 6

An integrational perspective, compatible with Malinowski’s (1935) insight that “. . . isolated words are in fact only linguistic figments, the products of an advanced linguistic analysis. The sentence is at times a self-contained linguistic unit, but not even a sentence can be regarded as a full linguistic datum. To us, the real linguistic fact is the full utterance within its context of situation” (reprinted in Nye 1998: 254). This perspective views texts (discourse, dialogue) as the most natural units or instances of language. “Inevitable contextual embeddedness” (Toolan 1996: 3) not only of words and sentences, but also of texts. Language and life are integrated. Neurocognitive relational networks with redundant/flexible interconnections. Recent research in neuropsychology (using Magnetic Resonance Imaging), suggests that this is a good way to think about what happens in the brain during communicative events (text production or interpretation). Continuous creation, modification, and augmentation of the relational networks, as new expressions or texts are produced or encountered within communicative events. Ongoing dialogue. Such communicative events are embedded within the ongoing dialogue of a speech community. Emergent text-level and communication-level meaning, rather than seeing words or even phrases as the “containers” of meaning.

These concepts affirm the malleability of the lexicon, which in turn undermines the notions of a distinct sentence meaning, as well as the literal-figurative

The integration of language and life 77 dichotomy. The literal versus figurative dichotomy is a dubious consequence of literacy. It has contributed to our metalanguage and our misunderstandings. It has encouraged us to think of language in terms of decontextualizable units or constructions that are the objects of our study. Rather than viewing language as having discrete literal and figurative categories, it is best to view literal and figurative as the opposite poles on a scale or continuum of creativity in text production and processing. The same mechanism is in use. It is the creation of textual meaning, in which the larger context gives significance and meaning to the particulars. It is not the sum of the parts that gives meaning to the whole. Rather it is an interactive process involving the neurocognitive relational networks of speakers and hearers or writers and readers, in which there is emergent meaning that is a text-level or communication-level phenomenon. Neither the meaning nor the truth of an utterance can be evaluated by examining the literal meaning of the parts, or proposing a literal paraphrase. “Literal meaning” is a construct of the lexicographer or grammarian as historian. It does not take into account the effect on history of novel texts and creative metaphorical action. This metaphorical action would not be possible or comprehensible if not for language-as-communicative-event. As such, texts are produced which typically follow conventional and sedimented ways of expression, but become more creative and non-conformist as needed to communicate desired meaning (cf. Radney 1999: 101–4). If a novel expression is succinct or artistic, or spoken by someone important, it may become conventionalized and lexicalized. The lexicon is not meaningless or inconsequential, but it is not a monolithic fixed entity that can be drawn upon such as a dictionary or a mental file cabinet. This is surely a metaphorical way of thinking about it and representing it. In any utterance, the conventional linguistic signs (the lexicon) contribute much to the total communication event, but they are not equal to or entirely constitutive of the communication event. The lexical units interact within a text or communication event as a whole. Although there are interpretive constraints based on the history of the usages of the words and phrases (the bits and pieces that we recognize as being the same as some we have previously experienced), the immediate context can and often does take precedence over the historical conventions in guiding the interpretive process. If that results in a somewhat non-conventional interpretation of certain lexemes or phrases, then that will form some new cognitive history which will influence future utterances within the ongoing dialogue of a speech community. The emergent text-level meaning may be produced by speaking/writing (“using the grammar and lexicon”) in ways that are conventional or sedimented, or in ways which are more novel or creative. At the conventional and sedimented end of the scale, people react to it as literal usage. At the novel and creative end, people react to it as metaphorical. But it is the same process of communication or meaning creation. This is not to say that a text has a

78 Michael R. Walrod single meaning, but rather that it has multi-faceted and comprehensive and complex meaning that is a text-level function. “Words” are as much constrained by texts and communication events as vice versa, both at the moment of the event, and subsequently.

Conclusion An integrational approach (Toolan 1996) focuses on texts/utterances as situated communicative events rather than as decontextualizable objects. In Fiumara’s terms, we need “. . . an outlook on life and language which assumes their reciprocal interaction” (1995: 7). This is very compatible with a neurocognitive approach which recognizes communication as process and event rather than thing, and recognizes the continuous restructuring – acquisition, augmentation, and adjustment (Lamb 1999: 10) – of our cognitive relational networks as we experience and interact in the ongoing dialogue of a community. My point here is to comment on the nature of linguistic research in the academy. Phrases and clauses/sentences are frequently presented for purposes of illustration and analysis, in complete isolation from the context where they were found. Or they are created context free for the same purpose. If we could reach a consensus about which of the proposed definitions of “converted bike rider” is the most acceptable one, it would be only because we have some intuitions about which context would be the most likely one in which we would hear such a phrase in our own speech community. By the same token, discussions of the acceptability of certain strings of speech, or of the meanings of certain lexemes, rely on our ability to imagine and consider possible social and linguistic contexts, based on similar situations we have experienced. We experience language in its inevitable contextual embeddedness, in which the so-called units of language derive their meaning and significance from the text and context, as much as they contribute to it. And the contribution they make to the text/context is only because of the linguistic events previously experienced and retained in memory, in the neurocognitive relational networks of members of the speech community.

6

Indeterminacy of meaning and semantic change Edda Weigand

Introduction I will take as my starting point the conclusion arrived at by Peter Koch in an article published in the volume on ‘Historical Semantics and Cognition’ (1999) which contains structural and cognitive studies. Koch admits that the ‘paths of semantic change’ outlined by him may have been used but do not have to have been used at all. ‘We have to realize that [the mode of presentation] is totally artificial in relation to what really happens in language change’ (Koch 1999: 296). Such a summary is more or less valid for most approaches to historical semantics. Paths of semantic change are designed on the basis of hypotheses which empirically remain without genuine verification. These hypotheses refer to a notion of language which can be characterized as a myth far apart from what really happens in language use (cf. Harris 1981). Now the question is: what really happens in language use and in language change? According to Hermann Paul (1966: 32[§16]), ‘the main cause of language change is normal human speech activity’. It is human dialogic interaction in which language is used and in which over time the verbal communicative means and their meanings change. The verbal communicative means are integrally combined with other communicative means, mainly cognitive and perceptual ones, and their use is dependent on historical conditions. These, in the end, cultural conditions make up the complex whole of the dialogic action game in which human interaction takes place. It is the complex which has to be considered our starting point and all variables have to find their integral place in it (Weigand 2002a). Human beings at the centre of the action game negotiate meaning and understanding dialogically. Such a basic principle of negotiation presupposes the assumption of meaning indeterminacy and variability. The possibility of change across times is thus included. The next question to be addressed is: what is described in historical semantics? The state of the art in historical semantics is strongly biased by cognitive linguistics. Both cognitive as well as structural linguists pick out single words and consider them as fixed correlations of form and meaning. In structural linguistics meaning is defined by a set of features, in cognitive linguistics meaning has blurred edges and is represented as a prototypical concept.

80 Edda Weigand In both cases, language is considered to be a set of signs or concepts. Such a symbolic view has nothing to do with language use and rests on the presupposition that language is grounded in cognition. Why should human beings have developed language as an expression of thought? Thinking does not necessarily need verbal expressions. For a differentiated dialogic exchange of thoughts, however, verbal expressions are needed. Language therefore is grounded in communication which is far more than just a configuration of concepts. Cognition plays its role in communication but a partial integral role. Isolating cognition opens the door to speculation of all sorts. As has been demonstrated by recent neurological experiments on mirror neurons, human beings are from the very beginning dialogic beings (Weigand 2002c). They are oriented in their abilities towards other human beings. The view of human beings as the symbolic species therefore isolates one ability which cannot be separated from other abilities. It has to be replaced by the view of human beings as the dialogic species, the species which is able to do far more than symbolizing, namely to integrally use different abilities in dialogic action and reaction, among them the ability of symbolizing. Structural and cognitive approaches however start from single signs and describe semantic change by correlating their defined meanings at different times. These correlations are classified by terms such as narrowing versus broadening, amelioration versus pejoration, metaphoric, metonymic, etc. and illustrated by a few single examples. It is indeed the case that we can observe examples of narrowing and broadening meaning. The question however of when a set of examples establishes a semantic regularity has to be raised. On the one hand, we cannot indicate where the regularities begin and where they end. On the other hand, regularities are postulated in both directions. Semantic change is thus mainly established by classifications made by linguists without taking account of what is really going on and of what has been changed in language use. Among the consequences of this are a concept of polysemy as diachronically developed and a tendency to discover semantic regularities in playing down the surrounding irregularities. Designed in this way, historical semantics as an artificial methodology cannot claim to be the testing ground for other models. It is simply one methodology besides possible others. The testing ground for models can only be the object-of-study itself, that is, the object selects the model. ‘What really happens in language change’, that is the genuine issue for a concept of historical linguistics which brings together human speech activity and the parameter of time. This issue is avoided in traditional approaches insofar as they deal with questions on a classificational level, that is, with questions that exist in the linguist’s mind not in real language use. Whereas structuralism has had its merits in clearly separating issues such as ‘langue’ versus ‘parole’ and synchrony versus diachrony, we are now in a position to address the complex and consequently discover the relationships of the parts. Human speech activity means dialogic interaction on the basis of competence-in-performance (Weigand 2001). The scientific interest of

Indeterminacy of meaning and semantic change 81 synchrony which aims at explaining how understanding in dialogue becomes possible is closely connected with the scientific interest of diachrony which aims at explaining how human speech activity allows for diachronic change. Language change happens in synchrony. Diachrony therefore has to be traced back to synchrony, and the conditions of why this is possible have to be elaborated in a theory of language which can account for dialogic speech activity. Human beings interact in dialogic action games, that is, they negotiate meaning and understanding in dialogue. During time the way we negotiate changes: different communicative means may be used and different meanings negotiated. The actors are human beings, individuals, and they depend on their abilities. Human behaviour is always open to change, it follows rules, conventions or regularities as far as they are available. It goes beyond rules when rules come to an end. Whereas the term of creative language use implies deliberate new use, the term of language change refers to intentional as well as unintentional change. The possibilities of creative use and language change are more or less constitutive for language use. We accept without problems new ways of use such as Indonesia – just a smile away. In order to discover what really happens in language change, we have to refer to authentic texts and justify our analyses as far as possible by large corpora. In order to understand the authentic text, knowledge of the cultural and historical background of the action game is needed. Let us first consider in more detail how semantic change is dealt with in a few so-called mainstream approaches of linguistics: the structural, the cognitive, the pragmatic-inferential and the semiotic approach. Second, let us consider semantic change in a holistic view which is not restricted to a notion of language as a myth and in this framework discuss central theses such as the thesis of semantic regularities, the thesis of indeterminacy of meaning and the thesis of a network of ways-of-use. Third, language use as a network of phrases will be illustrated by analysing a few examples which clearly demonstrate the complexity of our object.

The language myth in historical semantics The structural approach The structural approach is mainly characterized by the concept of the sign and the dogma of language as a sign system. I will briefly mention some problematic points: ●

The level of description is the level of competence alleged to underlie performance. It is gained by total abstraction from constitutive features of language use and therefore cannot be considered to be the basis for performance. The break between artificial competence and natural performance cannot be bridged.

82 ● ●

●

Edda Weigand Meaning is defined and understanding presupposed. The correlation of form and meaning is fixed according to a pattern or code. The pattern is the same for the speaker and the interlocutor. Language is restricted to verbal means.

The list of problematic points could easily be continued. Applying structural methodology to historical semantics implies picking out single signs from texts of different times and defining their meaning. These defined meanings are then compared and differences classified such as broadening or narrowing or as a change from the concrete to the abstract. Let me illustrate the structural technique with a few examples. Blank (1999: 72), for instance, postulates the usual metaphorical change from ‘grip, grasp’ to ‘understand’ for Italian afferrare: (1) It. afferrare ‘to grip, to grasp’ ⬎ ‘to understand’ He simply ignores that there are different uses side by side such as: (2) afferrare una spada afferrare l’occasione, il momento afferrare un’idea

‘to grasp a sword’ ‘to seize the opportunity, the moment’ ‘to get an idea’

In order to postulate a metaphorical change, it would be necessary to verify the emergence of a new meaning empirically more precisely. Based on a series of classifications Blank (p. 74) arrives at the hypothesis that polysemy is generated diachronically, a hypothesis which is very often put forward in traditional studies but lacks clear empirical evidence. Krefeld (1999), in another structural study, uses mainly items of a specific kind, namely, words which name things, that is, names. In general, human beings predicate when using words in the utterance. It is a specific feature of proper names to name or identify a person or a thing. By reducing the vocabulary of a language to names or words which are restricted to their use as designations, for instance, of body parts such as bras, main, pied, jambe, the author ignores other uses and thus receives examples which can be said to have a rather defined meaning. In this way, linguistics is reduced to a science of designations and language to a convenient object for structural methodology. To sum up, the structural approach is determined by a notion of language as a myth which is constructed in the mind of linguists. Historical semantics in such a methodological framework deals with artificial abstractions and classifications and has lost its empirical object of language use. The cognitive approach The cognitive approach, for instance, by Sweetser (1990) and Langacker (1999), is mainly characterized by a notion of language based on a configuration

Indeterminacy of meaning and semantic change 83 of concepts (Weigand 2002b). The most problematic point of this approach lies in the earlier mentioned presupposition that language is grounded in cognition. In a general way, language certainly is grounded in cognition but such a general thesis remains trivial and lacks precision in order to become analytically helpful. What is language? A list of verbal items and rules of combining them, as is assumed in structural approaches? Or a system of configurations of concepts, as cognitivists assume? As has become manifest in recent research, for instance, on mirror neurons, cognition is an integrated human ability which cannot be separated from other abilities. When cognitivists stress the view that language is grounded in cognition, they think of human beings using language for the purpose of subjectivity and self-expression. We admit that human beings are individuals, but they are social individuals, not hermits. Language has been developed because it is necessary for effective communication not for thought. A separate cognitive view on language therefore remains an interesting game for the cognitivist, another game of language as a myth with strange terms, for instance, Langacker’s ‘trajector’ and ‘landmark’, but without relevance for human speech activity. In cognitive approaches the configuration of concepts is typically seen as a configuration of prototypes (e.g., Geeraerts 1997). It might be right that words taken as single items evoke prototypical complexes but communication cannot be described as configuration of prototypes. Communication is based on the use of words in utterances. The ways-of-use are multiple, regular and irregular, motivated and unmotivated, conventional and creative, or as Sinclair (1991: 109f.) has called it based on free as well as idiomatic choice. Applying cognitive theses to historical semantics results in speculative descriptions which lack solid ground. Langacker (1999: 172), for instance, withdraws any claim to explain actual semantic change and admits that his analyses ‘have not been based on serious historical investigation’. Insofar as cognitivists aim at empirical verification, they gather examples which are mostly single items and describe them by ignoring counter examples which contradict their method. Sweetser (1990: 28) thus starts from the methodological stance of metaphor and in support of this stance offers ‘several historical puzzles’ which according to her are resolved by metaphorical change. In this way, she derives ‘intellectual understanding’ from ‘physical holding’ using the example of Latin comprehendere ‘seize’ which is considered to be the ancestor of French comprendre ‘understand’. This, however, is only half the truth insofar as in Latin comprehendere can already mean ‘understand’. The same can be demonstrated by another example: Latin capere cannot be taken as starting point of ‘physical holding’ for a change to ‘intellectual understanding’, expressed by Italian capire. Like comprehendere capere is also used in Latin in a variety of ways, to mention only consilium capere or a phrase like mens mea capit quae sit beata natura. The question, therefore, of whether the intellectual is metaphorically derived from the physical, is not at all settled. To sum up, the prototypical view is based on an interesting feature, namely the idea that meanings are not defined. This idea can be judged to be a

84

Edda Weigand

first right step but it still has to be embedded within the framework of language use. The pragmatic-inferential approach The pragmatic-inferential approach, for instance, by Traugott and Dasher (2002), is mainly based on Grice and NeoGricean pragmatics (e.g., Levinson 1995, 2000), that are, again on knowledge and cognition. It introduces an aspect of language use, namely that speakers use inferences in communication. The insight that speakers mean more than they say can be considered another step in the right direction but it is again only part of the whole. Pragmatics cannot be reduced to inferences. There is no alternative to action. Nor can the whole be grasped by a mix of parts, by an addition of knowledge and action (e.g., Fritz 1998: 9), or even by a ‘variety of theories associated more or less directly with Construction Grammar and Cognitive Linguistics’ (Traugott and Dasher 2002: 6). Parts find their role and can be adequately described only in the whole framework. Action integrally includes knowledge. ‘We cannot do everything at the same time’ might be right to some degree but cannot be accepted as an excuse for isolating parts. We have to have an overview of everything at the same time from the very beginning if we want to arrive at an acceptable view of the complex whole. Otherwise we arrive at a strange mix as hypothesized in Traugott/Dasher’s model of the Invited Inferencing Theory of Semantic Change (p. 38). Language use as a natural phenomenon is not the same as an abstract system or code put to use, as Traugott/Dasher and other orthodox linguists seem to assume who still draw on language as an abstract system (p. 26). Even if Traugott’s inferential approach is mainly characterized as a cognitive model of pragmatics, she emphasizes the importance of empirical verification by diachronic corpora which allow us to see how words are used in texts of different times. In this respect, Traugott rightly reminds linguists that their object is language-in-use not configurations of concepts. Nevertheless, her inferential model cannot be considered to be a genuine interactive model because of its emphasis on self-expression and subjectivity and its neglect of a consistent action theoretic basis in general, even if she also deals with intersubjectivity. A favourite example in Traugott’s analyses is in fact. Due to her method, which addresses only part of pragmatics, her results cannot completely satisfy. Sentence adverbs can be fully understood only with respect to a dialogic action theory. They express many more possibilities than ‘a denial of one or more manifest assumptions’ as indicated by Traugott (1999: 183). To sum up, making inferences certainly represents an important communicative means which however has to be integrated with other communicative means into the framework of the action game. Furthermore, reference to diachronic corpora is an inevitable empirical means of analysis even if the corpus cannot be equated with human speech activity.

Indeterminacy of meaning and semantic change 85 The semiotic approach There remains an approach to be mentioned which is often referred to as the final solution of the crux which other approaches are confronted with, the so-called invisible hand approach, for instance, by Keller (1997). The mysterious term attracts attention because it claims to explain the inexplicable. One might first assume that the metaphor refers to processes which are caused by chance, based on probability, at least not on a system of rules. However, the invisible hand approach is clearly constructed on the basis of rules within a semiotic framework, that is, within a compositional and rule-governed artificial system. Keller starts from a notion of ‘natural’ language as a code of the third kind, that is, as a set of substitution rules. Rules of the third kind are explained by invisible hand processes. What he means can, for instance, be illustrated with the example of a footpath which is causally and unintentionally created as a short cut by many pedestrians. We have to pose the question: Why do we need rules of the third kind in order to explain causally created events? The invisible hand approach seems to consider language change as an unintentional change of systems. Certainly, language change often happens unintentionally, but it happens intentionally as well. For instance, with the German phrase Das ist Schnee von morgen ‘That will be water under the bridge’, a new idiomatic utterance was intentionally created, or with the German word Wendehals ‘turncoat’ a new concept has been intentionally expressed. Used by prominent representatives in the mass media, such creative expressions very likely become conventional. The change from creative to conventional use cannot be described as an unintentional change of systems. Other examples are Anglicisms, for instance, in Italian election day in Maggio, or new ways of use coming into existence on the internet. Synchrony is diachrony, for instance, in the use of the new adverb zielführend: (3a) Wir werden diese Entscheidung zielführend fortsetzen. ‘We will continue this decision in a goal-directed way.’ There is the conventional use of the verb: etwas zum Ziel führen ‘to lead to the goal’ from which the adverb has been intentionally derived. One might also create a new adjective: (3b) eine zielführende Entscheidung, eine zielführende Zusammenarbeit ‘a goal-directed decision’, ‘a goal-directed cooperation’ To sum up, the semiotic approach of an invisible hand process represents another variant of the language myth. In my opinion, it is irritating that an artificial semiotic construction can be thought of as explaining natural language use and change. Where there is no clarity in the essence, mystifying expressions may attract the researcher’s interest. Obviously, there is still much confusion among linguists about their object.

86

Edda Weigand

Language change as a change in human speech activity Having shortly reviewed a few mainstream approaches to historical linguistics, all dealing with language as a myth, let us now draw some conclusions in a more constructive way. Even if modern linguistics has been dominated by the dogma of language as a sign system for approximately a century, real linguists always knew in their hearts that outside of scientific construction there is no such thing as language as a separate system of signs. Language as a natural phenomenon is an integral part of a complex human ability, the ability to communicatively interact by speaking. Having explicated my view on what is going on in human speech activity in, for instance, a contribution to the volume on ‘The language myth’ (Weigand 2002a), I will not go into detail again. I will mention only the main points: ● ●

●

●

●

Language use is always language use by human beings. Human beings interact in cultural and historical frameworks, that is, in dialogic action games. Human beings are purposeful beings. Communicative purposes and needs are therefore the key concepts for addressing human interaction. There is no alternative to considering language use as language action. Human interaction is carried out by different communicative means, verbal, cognitive and perceptual ones, all used integratively together. Linguistics therefore has to be humanised and can no longer be the science of language only. The fundamental linguistic interest, in my view, is directed towards explaining how and to what end interaction takes place. An approach to language as it is used in Dialogic Action Games has to be based on authentic texts. What is going on in the action game can be fully explained only by insiders of the Action Game.

Applied to historical linguistics, some restrictions have to be made in principle: Historical action games occurred in the past, insiders or native speakers can no longer be asked. Empirical evidence is necessarily gained from written data. Moreover, text tradition depends on ‘the happenstance of what has remained despite fire, pillage, and neglect’ (Traugott and Dasher 2002: 44). A representative diachronic text corpus therefore remains an illusion. There is however no other way than using authentic examples from historical texts, that is, focusing on ways-of-use and not on single items, and describing them, if possible, with respect to corpus analytic criteria, that is, mainly frequency of use in certain text types. Sometimes we can draw on comments in dictionaries and grammars. Moreover, we not only have the texts, we have historical and cultural knowledge from which we can at least partially construct the framework of the action game.

Indeterminacy of meaning and semantic change 87 Let us now try to give an answer to the crucial question of what really happens in language change using the model of the Dialogic Action Game as a model which deals with human speech activity. I will first concentrate on the hypothesis of semantic regularities in language change which seems to play a major role in present-day research. Having clarified what the term regularities implies, I will then advance two basic assumptions about words in language use which can give us an explanation for semantic change: the assumption of meaning indeterminacy and the corresponding assumption of networks of use. So-called semantic regularities The state of the art in historical linguistics is strongly influenced by the notion of semantic regularities, for instance, in the work by Traugott (1999), Traugott and Dasher (2002), Sweetser (1990), as if there was no debate on crazy rules. I will not dwell on this issue again, but focus only on the notion of regularities. Regularities are hypothesized by linguists who pick out single items, define their meaning and compare them in different times without verifying their hypotheses by corpus analytic criteria. To give an example by Sweetser (1990: 28): (4) katalambánein ⬎ ancient Greek ‘grip physically’

katalambánein modern Greek ‘understand cognitively’

On the one hand, ancient Greek katalambánein not only means ‘to grip physically’, on the other hand, modern Greek katalambánein has also retained a reading of ‘physically gripping’. katalambánein is dealt with by Sweetser in a similar way as the already mentioned Latin comprehendere: disturbing uses are simply ignored in order to maintain the desired regularity of a change from body to mind. Why however should the way human beings change their use be regular? Concerning form, there might be rules, rules of phonological change. Establishing regularities in semantic change however is far more complicated. First, the question of what is the meaning of words in language use should be clarified (Weigand 1998). Moreover, the fact that changes might go in both directions is irritating. It might be a change of narrowing as well as of broadening meaning or, according to Sweetser (1990: 35), a change from body to mind as well as from mind to body regarding hearing and vision, for instance: (5) Latin intendere ⬎ Old French entendere ⬎ Modern French entendre ‘direct one’s ‘understand’ ‘hear’ attention to’

88 Edda Weigand Even if to some degree regularities in the sense of tendencies can be observed, there are, on the other hand, so many irregularities. Where do regularities stop, where irregularities begin? These are essential questions which should worry proponents of a regular change. They seem to have bothered Traugott and Dasher (2002: 1) insofar as they indeed reduce the term regularity to tendencies. Their results thus remain rather non-committal. When tracing back words of the cognitive area, we mostly arrive at roots of the perceptual area, as Kronasser (1968: 188f.) remarks. We have however to go back to reconstructed forms, for instance, tracing back fir-stmn to a root meaning ‘vor etwas stehen’, ‘to stand in front of’. In these proto-times, it is very likely that human beings considered concepts of the mind as perceptual concepts, that is, that they did not strictly separate both areas, as can be seen with the Greek examples thumós relating mind to fumes or psyché relating mind to breath. Kronasser’s comment in this respect that it is often difficult to distinguish between the meanings of these words is very instructive. The separation has been made by linguists dividing the simple concrete area and the more difficult cognitive one. Indeterminacy of meaning Approaches based on semantic regularities establish fields of meaning such as body and mind or space and time and mostly consider them as fields of our world or so-called reality. The meaning of lexical signs thus is defined by reference to ontology. Sweetser’s remark that the observed regularities are motivated ‘within a cognitively based theory’, not within ‘an objectivist semantic theory’, does not take a clear stance: on the one hand, it is obvious that meanings are not objectivist referential entities, on the other hand, it remains unclear whether the ‘cognitively based theory’ is again based on ontology (Sweetser 1990: 1f.). In my view, which is confirmed by modern physics, there is no ontology independent of human beings. Reality exists in the eye of the observer. It is therefore not ontology which determines meaning but human beings’ thinking. Areas like space and time, even body and mind, are close together. Who of us can say what human perception and understanding of the world was like in pre-historic times? Human memory also confirms that the way of grasping concepts occurs in an approximative, not defined way. We orientate ourselves towards profiles or towards prominent characteristics and do not memorize concepts by means of a list of defining features. Human memorizing is characterized, in my view, by a combination of indeterminacy and routine. If we call up single words, their meaning remains indeterminate. If we use these words in utterances, we use phrases, multi-word units, which predicate more precisely and which are memorized as a whole as routines. The cognitive view accepts meaning indeterminacy by referring to the notion of prototypes. The process of ‘subjectification’ which plays a major role in cognitive studies, even if not always the same (cf. Traugott 1999 versus Langacker 1999), can also be hypothesized only on the basis of meaning indeterminacy and variability.

Indeterminacy of meaning and semantic change 89 Let us consider some examples which usually are referred to in order to illustrate the theses of semantic regularities and metaphorical change. First, a change from space to time is postulated, for instance, by Haspelmath (1997) or Traugott (1999). Haspelmath classifies prepositions such as after, before, from – to as indicating time or space and establishes a metaphorical regular transfer from space to time. Such a classification however completely results from Haspelmath’s method of description and is not based on a representative corpus, as Haspelmath (p. 16) himself admits. Why should from – to refer either to time or space and not to both? Checking Collins Cobuild Dictionary, it becomes immediately clear that after, as well as the other mentioned prepositions, are used in spatial as well as temporal contexts. It is the NP or the whole utterance which decide whether space or time is meant, not single prepositions. Why do we, according to Haspelmath (p. 141), need artificial categories of ‘abstract motion’, introduced by Langacker (1990), to deal with examples such as (6) The road leads from Minsk to Smolensk. The poplar is after the oak. instead of simply assuming that the meaning of these prepositions, if taken separately, is not defined with relation to space or time but allows for meaning indeterminacy? Why indeed should we isolate prepositions as such? Words are used in phrases. Especially prepositions display their meaning only in phrases, namely prepositional phrases. In the phrase or the larger context of the utterance it becomes clear whether space or time is meant. The study by Haspelmath and other studies which refer to him clearly demonstrate the limits of the orthodox view which is a methodology without a natural object. One might object that after, before, etc., are items of natural language; single items however do not make up natural language use. The study by Sweetser (1990) is also based on the thesis of regular metaphorical change. The cognitive meaning of ‘understand’ is derived not only from physical ‘seize’, as we have already seen, but also from perceptual vision and hearing which are bound to the concrete eye and ear. Thus Sweetser hypothesizes a semantic change for ancient Greek klúo from ‘hear’ to ‘heed’ even if she has to admit (p. 35), as already mentioned, that there is also a change in the opposite direction, from the cognitive to the physical. Moreover, she has to concede that the empirical facts of the earliest texts in Indo-European tradition, the ways-of-use of klúo in Homer’s Iliad, demonstrate a close link between hearing and heedfulness: ‘I was unable to find a single instance referring simply to physical sound reception’ (p. 41). How then can an original physical meaning of ‘hear’ be justified? The empirical facts are to be interpreted as a strong argument for meaning indeterminacy from the very beginning instead of postulating constructed changes. This leads to an interesting fundamental question: Why do we have to postulate different meanings of a word? How are readings established? Let us

90

Edda Weigand

illustrate this question with an interesting Italian example. Why, for instance, do we assume that avvertire has different meanings such as ‘to inform, admonish, warn, notice’. At least the first three of these meanings are very close together, the differences can be left to indeterminacy. In communication avvertire is always used in phrases and gets its meaning as part of these phrases. Thus it seems reasonable to suppose that the polysemy of the single word is hypothesized by linguists. Polysemy might be central for specific theories, such as Traugott/Dasher’s Invited Inferencing Theory (2002: 16), but not for language as a natural object. Meaning indeterminacy can even tolerate contrary meanings. It is not at all the case that contrary meanings get out-of-use. Speaking a foreign language we know that they exist because sometimes they raise serious problems for our understanding. I mention only a few Italian examples, for instance, richiamo, ribattere che, which can be used in the same syntactic context with quite different, even opposite meanings, richiamo for ‘call’, ‘retreat’, ‘reproach’, ‘attraction’, ‘reference’, ‘appointment’, etc., ribattere for ‘hit back’, ‘reject’, ‘refute’, ‘insist’, ‘reply’, etc. A very clear example is Italian intendere meaning ‘to mean’ as well as ‘to understand’. For native speakers these items do not seem to pose problems. The reason probably is that they have a better understanding of the whole linguistic and cultural context. Indeterminacy of meaning, that is, flexibility and variability of meaning, is a necessary precondition for language change. In his criticism of fixed codes, Roy Harris has already taken the decisive step towards introducing the concept of indeterminacy (e.g. Harris 1981: 10). How could change be possible on the basis of defined meanings? Indeterminacy does not mean arbitrariness. It is compatible with the concept of prototypes. Communicating, however, is not equivalent to profiling and combining prototypical concepts, as Langacker (1999) describes the cognitive basis. In dialogue human beings negotiate meaning and understanding and this is possible because meaning is an open changeable concept. It is the correspondence of meaning indeterminacy and determination in phrases which makes up language-in-use. Network of ways-of-use The example hérma Let us now describe in more detail language use and language change in terms of a network of phrases. Studying ways-of-use in Indo-European languages reveals new insights into language as a network of phrases. We had already the case of Latin comprehendere and capere, which from the beginning of textual tradition exhibit a complex system of ways-of-use. The same is true of other Indo-European examples. Janda (2006: 23ff.), when analysing, for instance, the Greek word hérma, demonstrates that we have different

Indeterminacy of meaning and semantic change 91 ways-of-use with different meanings from the very beginning of textual tradition with Homer’s Iliad: (7) hérma – ‘ship’s chain’ – ‘ear chain’ – ‘people supporting the city’ – ‘magic bonding through pain’ Very interestingly, there is no specific primary meaning which could be postulated as initial nor is there polysemy which is generated diachronically. From this example we can derive important insights: ●

●

Words are not symbols with a literal defined meaning. There is no root, no single word at the beginning literally meaning, for instance, ‘bind physically’. Words are always embedded in the context of the utterance and gain their meaning in this context as part of a complex multi-word unit (cf. Weigand 1998). Polysemy is not generated diachronically. There are different ways-of-use from the very beginning. There is no metaphorical change from concrete ‘bind’ to cognitive ‘bind’. Thus the starting point for diachronic development is not one literal meaning as assumed in Traugott and Dasher’s schema (2002: 12): (8a) A comprehendere ‘seize’

⬎ ⬎

冦AB冧 comprehendere ‘seize’ ⫹ ‘understand’

(⬎ B) ⬎

comprendre ‘understand’

Instead, the starting points are different ways-of-use, that is, not single words but phrases, with different meanings: (8b)

冦AB冧 comprehendere ‘seize’ dextram alcis comprehendere ‘understand’ alqd animo comprehendere (mente, sensibus, cogitatione)

(⬎ B) ⬎

comprendre ‘understand’

For comprehendere, an important point is noted in the Latin dictionary (Menge-Güthling), a point which is totally ignored by Sweetser (1990: 28) and others who only draw on single items: Having the meaning ‘understand’, it is mostly necessary for comprehendere to be used in a phrase which contains animo, mente, etc. Thus again we have a strong argument

92 Edda Weigand

●

●

in favour of meaning indeterminacy of single items and more determinate meaning arising from ways-of-use. Interactive use and negotiation of meaning are the causes of change, not the grammar of an abstract language system. Even if we start with meaning indeterminacy and complex multi-word units, change is effected by human beings in their interaction either by using words and phrases with a somewhat different meaning or by creating new expressions. Individual changes sometimes become conventional by increased frequency of use. To reserve the term change only for a change in the system, that is, for conventional change, misses the point insofar as there is no such system (cf. Love 2002: 27). It is however not the link in the mind established by iconicity of the sign which leads to frequent patterns. Explanations like ‘Things that go together go together’. are too simple to explain anything. It is not frequency itself which influences structure and causes language change. Frequency is frequent use by human beings. Frequent use confirms conventions but the question of why a specific use has become frequent has still to be settled for every individual case. For most cases, it is due to historical chance.

Let us now turn to two further examples, the English to apprehend and to fall, and analyse their ways-of-use in more detail in order to arrive at some final conclusions. The analysis can only be a first preliminary attempt at demonstrating some basic principles of word meaning. An exhaustive and formally correct analysis on the basis of large corpora goes beyond the possibilities of this chapter and is left to future lexicographic work. The example apprehend The example apprehend is a good example of medium complexity from which we can gain first insights into how words function in language use. It is a verb with different meanings changing over time and belonging to different semantic areas: two meanings, close together, ‘arrest’ and ‘seize’ to the concrete physical area, other meanings, again rather close together, to the cognitive area: ‘understand’, ‘fear’, ‘expect’. The readings ‘fear’ and ‘expect’ are restricted to formal use and are becoming out-of-date. The first issue to be tackled with regard to apprehend refers to the cognitive issue of a metaphorical change from body to mind. Following Sweetser and others, one might assume a literal and primary meaning of ‘arrest, seize’ from which the meaning ‘understand’ would have been derived. There is however no empirical evidence, in this case either, for a view of polysemy as being diachronically evolved. From the very beginning of textual tradition, we have different waysof-use. It is the phrase in which meaning indeterminacy becomes determinate: (9) (Collins Cobuild Dictionary) We often fail to apprehend the real nature of change.

Indeterminacy of meaning and semantic change 93 Now let us take a closer look at the meaning ‘fear, anticipate’ and ask where it comes from. According to the cognitive view, one might presume the following sequence of changes: (10) physical ‘seize’ ⬎ cognitive ‘understand’ ⬎ cognitive ‘understand with fear’ By textual tradition, there is however no evidence for such a sequence of change. In ancient Greek we have lambánein, katalambánein with various ways-of-use from the very beginning, among them the physical and the cognitive and, very interestingly, even one which is closely connected to ‘fear’, with phóbos, chólos or the like as subject, meaning something like ‘fear seizes sb’. The same can be assumed for Latin apprehendere even if not indicated in the dictionary. Consequently, we cannot say what the root meant in the protolanguage; the question of whether physis and psyché were already separated in the mind of the users remains open. Language and meaning are dependent on the way human beings perceive the world. On the one hand, there is indeterminacy of meaning, on the other hand words are used in phrases from the very beginning. Even in phrases indeterminacy of meaning remains to some extent as, for instance, in an example from the Oxford English Dictionary (OED) which demonstrates that we do not only ‘grip physically’ but also ‘cognitively’: (11) (Fielding: Tom Jones, 1749) A triumphant question to which he had apprehended no answer When we investigate in more detail phrases with apprehend, meaning something connected with ‘fear’, it becomes obvious that apprehend ‘fear’ does not directly evolve from ‘fear seizes sb’ but from uses such as ‘apprehend something to be feared’, that is, the element of fear derives from a specific object-NP. Many examples can be found in the Helsinki Corpus, the OED and the British National Corpus (BNC) specifying the object-NP as damage, infringement, violence, breach of peace, etc., for instance: (12) (Helsinki Corpus, 1570–1640) describe yf I had not apprehended an ymminent daunger It is the object-NP which differentiates the use of ‘understand’ from ‘fear’ as it was the ablative animo or mente which differentiated comprehendere ‘understand’. There are however cases in which the object-NP is semantically more empty, which seem to be perfectly understood by native speakers but pose problems to non-native speakers, for instance, in (13) (OED Sir T. Browne, 1643) Which makes me much apprehend the ends of those honest Worthies.

94 Edda Weigand Let us go one step further in semantically emptying the context and consider ways-of-use of apprehend without any object-NP. Here the question arises of how do we know what a phrase like apprehend that . . . means. Does it mean ‘understand that’ or ‘fear that’? How can we distinguish these two meanings? (14) . . . the plaintiffs paid the money unwillingly and only because they apprehended on reasonable grounds that without the permit which could not . . . (BNC) (15) . . . moneys were paid over by the plaintiffs to avoid the apprehended consequence of a refusal to submit to the authority . . . (BNC) (16) . . . if nothing was to be apprehended from their claws . . . (BNC) Even the adjective apprehensive does not tell us much more about the emergence of a new way-of-use. The Helsinki Corpus contains only two occurrances for apprehensive, one of them combined with danger: (17) (1640–1710) # that none are more apprehensive of danger or more fearfull of death than this sort of men (1640–1710) It is certain Godfrey grew apprehensive and reserved: for meeting me on the streets, after some discourse of the . . . Now the question arises of whether the element ‘fear’ becomes a meaning element of apprehend. One might think of Traugott’s explanation of lexical words as a sponge which gradually take in meanings from the syntactic context. Such an explanation however sticks to single lexical items and grammar of an abstract system and ends up in an empirical problem. Regarding the Helsinki Corpus, one has to take into account that it is not a representative corpus. Strong hypotheses therefore cannot be based on it. Nevertheless, the examples clearly confirm the view of basic conceptual indeterminacy of meaning. Different close meanings which could be analytically distinguished are comprehended in one indeterminate concept. These different meanings might be differentiated by the syntactic context or might come to the fore by frequency of use. The single word apprehend however does not take in a specific meaning element from the syntactic context. Without specific context, solely relying on frequency of use, the risk of misunderstanding remains. The example fall Let us finally have a look at an example which is characteristic of natural language in all its complexity, the example fall. To describe it goes far beyond structural and cognitive possibilities. It demonstrates once again that

Indeterminacy of meaning and semantic change 95 language use is not a rule-governed compositional object but an object of its own, the complexity of which is caused by a mix of regularities and irregularities, conventional routines and free creative uses. It is by no means an exceptional example, on the contrary, its complexity is typical of how language is used. I do not intend to give an exhaustive lexicographic analysis. Using examples from the Collins Cobuild Dictionary which is based on a large representative text corpus, I am going to highlight some essential theoretical points which demonstrate basic lines for the description of words-in-use. There is the single item fall which is characterized by an extreme degree of meaning indeterminacy that goes so far that it is very difficult to outline it. The orthodox view would have to accept a quasi-infinite polysemy. Nevertheless, it is outlined as a range of meanings and not an all-encompassing arbitrary area. As competent speakers we know how to handle this multiplicity of meaning possibilities, a fact which is indeed surprising. From the very beginning, fall is used in phrases with a more precise but not at all determinate, defined meaning. Dictionaries consequently take account of this fact by specifying the context, especially the subject-NP. Collins Dictionary distinguishes fall in the context of hair, of a blow or a weapon, of a building, of a piece of clothing or a curtain, of rain, snow, of an organization, of an idea, of barriers, of someone’s guard, of a sound, of words, of a value, of feelings, of silence, of light and shadow, of darkness and night, etc. These multiple ways-of-use constitute different multi-word units with a variety of meanings which are not at all differentiated by simple features such as ‘physical’ versus ‘cognitive’. Adopting the Collins Cobuild technique of meaning representation, for instance, in the area of the physical, the meaning of ‘move’ versus the meaning of ‘hang’ can clearly be distinguished: (18) big drops of rain fell his wavy, reddish hair falls to his shoulders

‘move’ ‘hang’

Prepositions are also an important element in the phrase by which meaning becomes more specific. In the Collins Dictionary, there is a long list of collocations with various prepositions which clearly demonstrate that the meaning of these multi-word units cannot be compositionally gained from the parts, for instance: (19) to fall on a specific day This year Easter Day falls on March 30th. to fall into a particular classification, grouping My work really falls into three parts.

‘be’ ‘come’

to fall into the hands of ‘become controlled’ The party was falling into the hands of extremists.

96 Edda Weigand to fall about laughing ‘be amused’ When he complained that it was unfair, they fell about laughing. to fall apart ‘go to pieces’ It was something to do, it kept you from falling apart. to fall back in confusion ‘withdraw’ They fell back in confusion, surprised by the direction of attack. to fall back on easier solutions ‘resort’ Often you give up and fall back on easier solutions. to fall behind with ‘not produce sth due’ Unfortunately, we have fallen behind with the payments. to fall down ‘be in bad condition’ The house was cheap because it was falling down. to fall down of an argument ‘fail’ In that particular instance the argument falls down. to fall for somebody He fell for her the moment he set eyes on her.

‘desire’

This list of specific collocations can easily be continued, also with other prepositions such as to fall off, to fall out, to fall over, to fall through, to fall to, etc., or with collocations of the type fall ⫹ Adj such as fall sick, fall short, flat, asleep, etc. Moreover, similar uses have to be taken into account such as (20) he has fallen a victim to hysterical exhaustion to fall in love, into disuse, into despair There is no empirical evidence for a semantic change from body to mind, only uses referring to the physical (21) large masses of rock are constantly falling into the sea she lost her balance and would have fallen if . . . they were waiting for the bridge to fall and uses referring to the cognitive (22) the regime had fallen Greater London will fall to Labour the barriers fell and we could see the fear in his eyes There are examples such as (23) the brave young men who have fallen in the struggle yet more began to fall of hunger and exhaustion

Indeterminacy of meaning and semantic change 97 which might be interpreted by cognitivists as dead metaphors. However taking account of the fact that there is no empirical verification for the primacy of physical fall, the distinction between literal and figurative meaning becomes meaningless. This conclusion however does not affect the possibility in principle of going deliberately beyond the literal as, for instance, with utterances such as The dice has fallen. This multiplicity of ways-of-use demonstrates the vast complexity of language use in the lexical area and unmasks a rule-governed sign oriented description as a myth. Questions of why specific ways of use die and others emerge, for instance, (24) to fall in such a rage ⬎ to fly into such a rage remain in the dark. There is no reliable empirical evidence in the corpora for justifying regular semantic change in a strict sense. Even Traugott and Dasher (2002: 11) admit: ‘Without some independent evidence such as changes in morphosyntactic distribution, we have no principled way to demonstrate that a linguistic change has taken place.’ In my view, from the very beginning, language as a whole has been a complex network of phrases which has mainly developed due to historical chance as has been convincingly demonstrated by Hundsnurscher (1988) with the German example GRÜN. Hundsnurscher checked all uses of this seemingly simple German adjective, from old high German times over medieval times to present day German, and did not find any clear evidence for regular changes. The different ways-of-use of GRÜN did not evolve from a basic literal meaning ‘colour’ but from different ways-of-use from the very beginning.

Conclusions and future perspectives Having highlighted a few points of our complex object I am now going to draw some conclusions and to indicate the direction which future research should in my opinion take. First and foremost, linguists should never forget that their object is at least in part an empirical one. Words play their role in dialogue as parts of multi-word units or ways-of-use which have to be studied in detail and more precision by lexicographers on the basis of large corpora. If historical semantics takes account of its real object, human speech activity, quite another picture emerges than that presented by present-day mainstream approaches. Instead of looking for regularities, whatever they may be, we have to start from texts with their complex networks of ways-of-use. Communication does not mean transfer of a pattern (Weigand 2002a) but negotiation of meaning and understanding in dialogic interaction. Change happens in this interactive process of negotiation between different interlocutors.

98

Edda Weigand

Indeterminacy and flexibility of meaning is a precondition for negotiation and change. The general principle of conceptual meaning indeterminacy opens up the way for different uses or collocations. It is this correspondence of meaning variability and concrete uses which makes language so efficient as a network of partly similar, partly different uses. Some of these multiple and changing possibilities are confirmed by use during time. Conventions thus result from frequent use. It is meaning indeterminacy which represents the point where change comes into existence. Synchrony is thus the beginning of diachrony and has to be explained in a way that allows for diachronic change. Change may be intentional or unintentional, may follow some analogy and therefore show some regularity but in the end it is due to historical chance. We should not feel forced to give a definite answer to the question of why specific uses become conventional and others not. The reasons are multiple as there are multiple variables which make up the complex. Even if we cannot go back to the beginnings, we have to take account of the fact that it is not ontology but human thinking on which meaning is based. We do not know much about pre-historic ways of thinking. What we nowadays separate in some respect, body and mind, may well have not been separated then. Assuming indeterminacy of meaning in correspondence to ways-of-use, the issue of polysemy vanishes to a large extent. Moreover, we do not communicate with verbal means only, we use other abilities as well, cognitive and perceptual ones. As far as polysemy does not yet disappear in the phrase, it is mostly clarified in the framework of the dialogic action game. Thinking about language should start from reflecting on the object, that is, authentic language use. Even if we are no longer insiders of action games that occurred in the past, there is nevertheless no other way than using examples from diachronic corpora. In the future, it will therefore be necessary to compile more and larger diachronic corpora for different languages. I am aware of the fact that we cannot explain the complex whole in all details – even the intent seems a contradiction to me – but we can falsify certain hypotheses made in the past and put forward others which are more likely. Only by making a continuous attempt at improving our understanding of the complex object of human speech activity can we develop a more and more adequate methodology for it.

7

On the cusp Antoine Meillet as a sociologist of language Robert McColl Millar

Introduction The sub-discipline sociology of language is generally associated with a particularly fruitful symbiosis formed between sociology, historical studies and linguistics in the 1950s and the 1960s in the United States in particular. In this form it is associated with the work of scholars such as Heinz Kloss, Charles A. Ferguson, Einar Haugen and (later) Joshua A. Fishman. But discussions of the relationship between sociopolitical structures, historical processes and language use had already surfaced in inter-war Europe and may have helped to inform and develop the more renowned later discussions. This similarity should not blind us to the fact that the ideologies which underpinned these earlier discussions were often not equivalent to the more liberal ones adopted after the war. This essay will discuss a particularly pointed example of these convergences and divergences to be found in the work of Antoine Meillet. In her essay on Meillet’s Aperçu d’une histoire de la langue grecque, Morpurgo Davies (1988) discusses why a comparativist working essentially on materials dealing with less well-covered Indo-European languages should have chosen to write what is arguably the first true ‘history’ of the Greek language, the classical language par excellence, incorporating both extra- and intra-linguistic analyses and insights, at a time when comparative Indo-European Linguistics and Classical Philology were generally considered to be antithetical to each other. She suggests (Morpurgo Davies: 243) that ‘. . . the Aperçu is in a sense a roman à these. It is meant to illustrate some principles but these are not spelled out in the text; the reader must extrapolate them from the narrative’. We will return later in this chapter to what on this occasion the principles were; her general comment lies at the heart of what is to follow, however: Meillet was, it might be suggested, developing a theoretical position on the relationship between language, history and society by implication, therefore. In the aftermath of the First World War, Meillet published Les langues dans l’Europe nouvelle in two strikingly different editions, of 1918 and 1928. This book attempts both to describe and analyse the linguistic situation in Europe after this catastrophe. But it also has an apparent ideological underpinning which, following Morpurgo Davies, is ‘not spelled out’. It is the purpose of

100 Robert McColl Millar this essay to extrapolate these general theoretical and ideological principles which Meillet appears to wish us to imbibe by osmosis. It will discuss Meillet’s work as an early example of linguistic sociology, comparing its findings with similar ideas found in other of his works. Throughout, Meillet’s views (explicit and implicit) will be placed in the context of both the linguistic politics of France and political and social developments throughout Europe.

Antoine Meillet as scholar Antoine Meillet (1866–1936) is particularly renowned as a comparative linguist. Although he is, perhaps, most recognized as an authority on the historical development of the Slavonic languages and of Armenian, he is said to have known at least one representative of each of the subfamilies of Indo-European (Krippes 1988). As Morpurgo Davies (1988) has suggested, perhaps the two most striking elements in his comparative historical linguistics were the willingness (previously largely uncountenanced) to attempt to harmonize extra- and intra-linguistic features, and to attempt ‘histories’ of languages and language families which did not avoid discussion of what we would now term modern sociolinguistic patterns. A less celebrated element of Meillet’s scholarly work was an interest in the contemporary sociological patterns of language use in Europe. In the period during and immediately after the First World War, Meillet published on a variety of linguistic and ethnic problems associated with the conflict and its solution, paying particular attention to central and eastern Europe (Meillet 1918a, b), culminating in Les Langues dans l’Europe nouvelle. A number of scholars have commented on this work, in particular Perrot (1988), who establishes a definite prejudice on Meillet’s part against the Magyar language. No-one appears either to have analysed this work as a whole, or fully enquired into what it tells us about the rest of his scholarship.

The ideology of language in its French context In order to understand Meillet’s views on the sociology of language, we need to contextualize these findings in the linguistic and cultural history of France. Although Meillet was a scholar with a prodigious and creative intellect, much of his work, including that under discussion here, demonstrates that he accepted many of the prevailing views of language held in France at the time. During the medieval period, the Romance dialect of the Ile-de France area had assumed a status which rendered it socially superior to the other dialects of northern (and later southern) France in the view of speakers. By the early modern period, French (as we can now begin to call it) gained singular importance as the sole language of governance and law throughout the kingdom. In the course of the seventeenth century, often under extreme stress from external sources (itself a stimulus, as Crowley [1989] has suggested, for an authoritarian seizure of the standardization process), the concept evolved

Antoine Meillet as a sociologist of language 101 among arbiters of French usage that ‘[t]he “best” French [was] the best because it [was] spoken by “the best people” ’, as Lodge (1993: 165) has pointed out. In this sense, French was not fundamentally different from many other languages, most notably English, in its standardization process. It is the second phase of the development, which Lodge dates (1993: 178) between 1660 and 1789, which differs significantly. Lodge suggests that the general view among arbiters of taste evolved that ‘[t]he “best French” [was] the best because it [was] the language of reason and clarity’. This view was certainly found among writers of other languages from that period. What makes French practically unique, however, is that these arbiters and their successors were able, under revolutionary circumstances, to demand agreement over this proposition from a wide range of literate French speakers, and also by many commentators from abroad (Schiffman 1996: chapter 4). Indeed, by the beginning of the nineteenth century, an ideology of French as a near-perfect instrument both for national and international interactions had developed, as Balibar (1985), among others, has suggested. This sense of special purpose and perceived inherent and constructed grace was inevitably intertwined with the mission civilisatrice which France claimed as its primary purpose in developing an overseas empire during the late nineteenth century. Throughout that century, actual knowledge of the standard form of French among citizens of France developed alongside the exponential growth of national education. Part of the educative process was, it could be argued, the compulsory military service through which all males were, by the end of the century, expected to pass. As well as providing a pool of troops trained to attack the external aggressors who were assumed to be present outside and (to some extent) inside the borders of France, this service led to an increasing level of knowledge of the national language among members of the – particularly rural – population who might previously have been largely unaware of its national importance, a point emphasized by Bourdieu (1991: 48–9) towards the end of the twentieth century. The sentiments of prestige and importance associated with the language might actually have been encouraged by the feeling current among many French thinkers as the nineteenth century progressed that the nation itself was failing to maintain its position in the ‘concert of Europe’, as discussed by Azéma and Winock (1970), Paraf (1981) and Taithe (1999), among others. In the first place, the growth of the British Empire and the influence of the United States of America had led to a geographical spread in the use of English which was associated with an elevation in status for that language. Moreover, the French nation had passed through a series of external and internal crises, including the loss of Alsace and parts of Lorraine and the Dreyfus affair, which had, in the view of French nationalists, lessened the nation’s standing. By the beginning of the First World War, French society, outwardly so confident, had endured a period of intense soul-searching and much conflict. That conflict, as might be expected, only added to the pressures felt by the opinion-forming parts of the French population. It

102 Robert McColl Millar would not be an exaggeration to suggest that the country had been bled dry of resources and, strikingly, of young men. The army, previously the heart and pride of the Republic, had been decimated at Verdun in 1916 and had reached the point of collapse the following year. Moreover, the world which the war partly brought into being was confused and confusing to many. Although, in theory, the Wilsonian dream of a world of nations brought about by the self-determination of peoples was an ideological descendant of some of the proposals of the more far-sighted ideologues of the French Revolution, the peculiarly bourgeois nature of the French Republican establishment (an alliance of small landholders, shopkeepers, artisans and professionals) could not but be concerned by the world which their liberal precepts had brought into being. The importance of academics within French culture, many of whom subscribed to a particularly narrow metropolitanism and prescriptivism in language, must also be borne in mind. No matter their origin or political background, many scholars at that time (as well as later) would not have been sympathetic to any movement which led to the downplaying of urban culture and language in favour of one which emphasized the rural or traditional (Grillo 1989; Charle 1993). In this sense the French intellectual attitude to the national culture was strikingly at odds with the views of many intellectuals from similar backgrounds in the countries to the east of them. Although normally equally urban, many scholars emphasized the importance of the rural and ‘pure’, in a truly romantic sense, against the more enlightened and urbanist mores of France (Smith 1971: 63). Yet these precepts of linguistic democracy (as well as, ostensibly, the protection of linguistic minorities) were enshrined in many of the constitutions of the states created by the post-war settlements. At the time at which the second edition of Les langues dans l’Europe nouvelle was published (1928), most of the new states of central and eastern Europe had at least the semblance of civil, middle class representative government (although a number were already moving towards various forms of authoritarianism). The main fear for many middle class observers of the events which followed the Russian Revolution and the end of the First World War was that Bolshevism would spread west: a not unreasonable fear given the brief exercise of Soviet power in Hungary and Bavaria in that war’s aftermath, as well as the considerable instability which Italy suffered during the period. It can be assumed that these fears would not have been allayed in the Paris of the time, where a great many Russian émigrés of a variety of political and social backgrounds had congregated; Meillet’s position as a Slavophile should be borne in mind in this context. From the point of view of academe, it might have seemed that the established order – social, cultural and linguistic – was gradually being replaced by one in which the great mass of the people might seize power in all of these spheres (as well as the crucial political one) in the place of intellectual and social elites. Many no doubt welcomed this; many more did not, however. It is in this climate – intellectual, social and political – that Meillet wrote his Les langues dans l’Europe nouvelle.

Antoine Meillet as a sociologist of language 103

Editions of Les langues dans l’Europe nouvelle Les langues dans l’Europe nouvelle is available in two editions, one published in 1918, the other in 1928. Although, in his introduction to the second edition, Meillet (1928: iv) remarks that there are few divergences between the two editions, there are in fact considerable differences between the two in terms of layout, illustrated in the following schema of chapter headings: Chapters in Meillet 1918 1

2 3 4 5 6 7 8 9 10 11 12 13 14 15 20

Chapters in Meillet 1928

Les langues qui se parlent en Europe

Les familles de langues Portée pratique de la classification des langues par familles Langues et races Langue et nation Langue et civilisation Le renouvellement des langues Unification et différenciation Extension des langues communes Différenciation des langues communes Les dialectes Les langues savantes Necessité des langues nationales Fixation des langues littéraires en occident Le Français littéraire Les langues nationales dans l’Europe occidentale

1 L’époque préhistorqiue 2 L’extension des langues indoeuropéennes 3 Les langues indo-européennes en Europe 4 Les langues non indo-européennes à l’est de l’Europe 5 Les familles de langues 6 Portée pratique de la classification des langues par familles 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

16

Les langues nationales de l’Europe orientale

17

Les langues nationales en Autriche-Hongrie

22 23 24 25

Langue et race Langue et nation Langue et civilisation Le renouvellement des langues Unification et différenciation Extension des langues communes Différenciation des langues communes Les dialectes Les langues savantes Necessité des langues nationales Fixation des langues littéraires en Occident Le français littéraire La situation linguistique suivant les regions de l’Europe La situation linguistique en Europe occidentale Origines de la situation linguistique dans l’Europe orientale Les langues nationals de l’Europe orientale La situation linguistique dans la bande centrale de l’Europe La region de la Baltique L’ancien Empire austro-hongrois et les deux grandes nations slaves de l’Europe centrale (continued)

104 Robert McColl Millar Continued 18

Les langues nationales dans la presqu’ile des Balkans.

19

Les langues nationales dans l’ancién empire russe

21

Réaction contre l’isolement linguistique De l’emploi des grandes langues de civilisation comme langues secondes Le latin et les langues modernes Le groupe slave Les essaies de langues artificielles Conclusion

22 23 24 25

26

La region balkanique

27 Le yiddisch 28 Les langues dans l’Union des Républiques Socialistes Soviétiques 29 Les inconvénients des langues nationales 30 Réaction contre l’isolement linguistique 31 De l’emploi des grandes langues de civilisation comme langues secondes 32 Le latin et les langues modernes 33 Le groupe slave 34 Les essaies de langues artificielles

The two editions are actually strikingly similar in layout, following (with one exception) the same linear chapter pattern (the one exception, Chapter 20 in both, is, arguably, placed in its correct position in the second edition, along with equivalent discussion of the linguistic situation in western Europe). There are a number of occasions where they diverge, however. In the first place, the second edition gives more information about linguistic relationship and historical development, while the first appears to leap in in media res. There is a strong suspicion that, intended as the book was for an educated layperson, this is a correction of an important lacuna in the earlier version. Moreover, the second edition deals in greater depth with the situation in central and eastern Europe, discussing, for instance, the Baltic region and Yiddish separately from other related issues. On this occasion this is more important for our purposes since the ten years of experience after the end of the Great War had singled out these particular topics as of especial importance in the post-war settlement. In the case of the Baltic States, it would have been a brave person indeed who would have predicted their independence at the end of the War as having anything other than a temporary reality; strikingly, Meillet had written an essay (1918b) at the very end of the Great War which predicted ‘problems’ in this very area. It is this benefit of hindsight which makes the second edition much more useful for our purposes. When Meillet was writing the first edition, all was conjectural; indeed, some points might even have been taken as wishful thinking, given that, even as late as the early summer of 1918, German forces were attacking in France. As a French patriot, Meillet would, inevitably, have given a more pro-Allied spin to the material then than he did in 1928, by which time the situation in central and eastern Europe appeared to have settled down.1 Because of the inclusiveness of the second edition, reference is

Antoine Meillet as a sociologist of language 105 made in this essay solely to that; nevertheless, a flavour of those (relatively small) divergences with the first edition will be given through references throughout this essay to articles which Meillet published on these topics towards the end of the Great War.

Sociology of language in the work of Meillet It must be recognized from the outset that Meillet himself appears to have been in two minds about the purpose of his book: whether it should be seen as a scholarly work, discussing these matters in a measured way, or whether it should be taken as an impassioned piece of polemic. Sometimes a reader might also suspect that there are actually two books rather than one included within the same covers: one which deals with the languages of Europe as linguistic systems, the other which deals with the sociology of language in Europe. Many of the ideas which are discussed in the following are not presented by Meillet in the order in which they are given here. Often Meillet repeats himself; he may not have established a fully formed idea of his model. Moreover, there are occasions when personal prejudices break through his argument and create ambiguities. This is particularly the case in his discussion of Belgium and Provence, where his obvious belief in the universality (and suitability) of the French language is in conflict with sociolinguistic and (in the case of Belgium) political developments in these territories. Nevertheless, despite many misgivings about his central thesis, his idea of a dynamic relationship between competing language varieties and the societies in which they are found is helpful in any discussion of the nature of language and society. Meillet’s essential contribution to our understanding of what we mean by language or dialect is to see them as being part of a Hegelian dialectic between other theses and antitheses: urban vs. rural, universal vs. provincial, and, most importantly, in the distinction between civilization or culture and a lack of these. Since many of these concepts are in the eye of the beholder, I will return to what he appears to have meant in my discussion of the distinctions and terminology he uses. At the heart of his typology is his term langue de civilisation, regularly used throughout his work, which could be translated either as ‘language associated with a given civilization’ or ‘language associated with civilization’; in fact, Meillet appears to intend both interpretations. He states (1928: 146) that Les grandes langues de civilisation forment un lien entre des hommes dont la langue courante est en partie autre. Ce sont les langues des élites intellectuelles qui conservent les doctrines acquises dans chaque groupe de l’humanité, qui inventent des idées nouvelles, qui transmettent la science et qui la font avancer. Par suite, elles fournissent à toutes les langues diverses et souvent de familles diverses, appartenant à un même groupe de civilisation, un fonds commun de mots savants et de manières de s’exprimer. Leur influence linguistique est immense.

106 Robert McColl Millar Elles sont nécessairement peu nombreuses: il ne saurait y avoir beaucoup de grandes civilisations originales; et, en général, chaque grande civilisation n’a qu’une langue principale pour s’exprimer. Meillet sees a langue de civilisation – particularly a great one – as a means of transmission of ideas and innovations between people in different times and places. Another central element of his critique of the linguistic ecology of Europe after the Great War is that these mediators of civilization will necessarily be few in number. From this position, a langue de civilisation is the product of an educated elite. This is, of course, correct for almost all Ausbau processes. It is informative to see what he makes of the development of ‘national’ languages for newly established states, such as Estonia (1928: 202–3): Avant la guerre, l’Estonie n’avait pas l’autonomie qu’avait gardée en quelque mesure la Finlande dans l’empire des tsars. L’este n’y était qu’un groupe de parlers locaux et n’était pas une langue de civilisation. Il y avait une aristocratie de langue allemande, et l’Université de Dorpat (en este Tartu) a été longtemps de langue allemande, avec tous les caractères d’une université allemande. Le gouvernement tsariste l’avait russifiée. La classe cultivée ne comprenait donc que des hommes formés à la culture allemande ou, dans les derniers temps, à la culture russe. C’est maintenant seulement que commence à se créer une classe d’hommes cultivés en este. Meillet is suggesting, in fact, a contest between languages over domain use. The fact that one of the languages is ‘new’ but in competition with major languages with considerable literary heritages can be related to his comment (Meillet 1928: 239) that ‘. . . les nouvelles langues de civilisation de l’Europe sont banales et n’enrichissent pas vraiment la littérature. Ceux qui les emploient ne peuvent, sous peine de rester des gens inférieurs, se dispenser de bien posséder au moins l’une des grandes langues de civilisation: à vouloir se dispenser d’un effort, ils gagnent d’en devoir faire deux’. The assumption is made that a language can only be termed such when it is ‘fixed’ by an elite and used in a variety of different functions. Only then can it be mediated to the masses: ‘Mais la commodité, médiocre et provisoire, qui est conquise par la masse est compensée par une gêne pour les élites.’ In other work (Meillet [1928] 1948: 229), discussing Roman civilization, Meillet comments that ‘Le caractère dominant de tout ce qui est romain est l’unité, – une souple unité que sait échapper à un schématisme rigide’. Despite these worries about the apparent proliferation of these national languages in Europe, Meillet presents a dynamic model of the manner in which a collection of dialects can become a langue de civilisation (1928: 208): ‘Le slovaque n’est qu’un dialecte du groupe tchèque; or le tchèque a un passé et s’est donné au xixe siècle une langue de civilisation’, although it is worth noting that the Czech language’s superiority over Slovak (a mere dialect) is

Antoine Meillet as a sociologist of language 107 expressed largely in terms of its earlier establishment as a langue de civilisation; indeed, it is significant that he uses the word donné ‘given’ to describe this category, implying an external, elitist, force at play. He suggests which domains are vital to the development of a langue de civilisation (1928: 146–7) Ce n’est pas son rôle dans le commerce, ce n’est même pas son rôle dans la politique qui donne à une langue l’essentiel de son influence; c’est l’usage qui en est fait dans les choses de l’esprit. . . . Si l’hébréo-phénicien a eu de l’influence, ce n’est pas parce que les Phéniciens ont crée un grand empire colonial et ont fait un important commerce dans la Méditerranée et jusque sur les rives de l’Océan: de ce grand mouvement commercial il ne reste linguistiquement presque rien; c’est parce qu’une tribu de l’arrière-pays a eu un grand mouvement religieux et que des Hébreux ont compose le recueil qui est devenu le ‹‹Livre›› par excellence, la Bible, que les mots des langues européennes sont pleins de sens bibliques. Again we find the assumption that it is not through its use in trade, or because it is associated with a political elite, that a language can be associated with a civilization. Rather, Meillet suggests that it is an intellectual elite – particularly when associated with a new idea or creed, particularly when expressed through the authority of the book – which will assure that language’s spread and survival. This argument is certainly questionable for the modern era (a point reflected in Meillet’s opinions on the importance of the development of governmental/bureaucratic domains for a standard language discussed hereafter); it may have some justification in the past, however, although even on this occasion it is worth noting that the spread of Judeo-Christian ideas and scriptures was facilitated both by their translation into the first ‘world language’ and by the presence of a largely unitary state in the Mediterranean world. This elite function is not confined to the development of literature, however. Another term which Meillet employs which develops his central thesis is langue savante, which might be translated as ‘language for thinking’ as well as ‘language of thinkers’. He sees this term as representing linguistic relationships from an historical perspective (1928: 82): Il y a eu dans le Midi de la France, au xiiie siècle, une civilisation brillante, dont le provençal était la langue usuelle. Elle a peu duré. Durant le temps assez court où elle a pu se développer, elle a eu une poésie lyrique originale, mais assez peu variée; elle n’a pas crée une prose. La langue savante était alors le latin; le provençal a été une langue poétique, non une langue de civilisation complète. Quand une prose savante s’est constituée dans la France du Nord, la civilisation propre du Midi était brisée. Jusqu’au xixe siècle, la tradition n’a pas été reprise. Les parlers du Midi n’ont servi durant de longs siècles qu’à l’usage courant et familier; ils n’ont pas été employés à exprimer des idées générales; ils n’ont pas été des instruments

108 Robert McColl Millar de civilisation. Et quand, au xixe siècle, on s’est remis à écrire certains parlers du Midi, les seuls ouvrages notables ont encore été des ouvrages poétiques. La presse est de langue française à Toulouse, à Montpellier, à Marseille. Quand, pour répondre à cette objection, Mistral ne trouvait à citer que les almanachs provençaux, il confirmait en réalité l’absence d’une prose provençale servant à exprimer des idées. Here the association that the language state has is not primarily with a poetic literature, but rather with what we might now term administrative and scientific prose as well as with philosophical inquiry. This view is supported by his comment (1928: 135) that Quand se fondent les universités de l’Europe occidentale, le latin est partout la seule langue. Lorsqu’un écrivain comme Dante s’adresse au grand public, il écrit en italien; lorsqu’il s’adresse aux lettrés, il écrit en latin. A discrepancy might be construed as existing between this interpretation and the views advanced earlier about the centrality of literature in language development. It is important to stress that elite custody for these developments is still implied, however. Beyond this, what is interesting about this model is that, again, it is dynamic. Meillet can certainly envisage the possibility that Provençal might have become what he would probably term a langue de civilisation. When the langue savante in what is now France was Latin, there was still this potential. Another term Meillet employs, langue littéraire ‘literary language’, shares much with both of the preceding. By littéraire he does not appear to mean merely that a language is used for literary purposes (which would contradict somewhat his views on langue savante discussed previously). Instead, he sees a literary language as one used by the elite. Thus Flemish is not considered to be a langue littéraire, since it is usurping the rightful position – in his view – of French in Flanders (and, by implication, in the whole of Belgium) (1928: 183): La Belgique est donc un pays où existent deux langues officielles ayant des droits égaux: le français et le flamand. Théoretiquement rien de plus juste. Mais, pour les Wallons, tout gain du flamand représente une diminution de la langue littéraire, qui est le français. D’ailleurs les éléments français du pays éprouvent une répugnance bien naturelle à apprendre une langue d’intérêt tout local, comme le flamand qui n’a, hors de la Belgique et de la Hollande, aucun rayonnement. Flamandiser une université comme celle de Gand, c’est lui retirer son caractère universel pour en faire une institution provinciale, et la décision, qui parait au premier abord naturelle, d’attribuer au flamand l’une des universités d’Etat a heurté les sentiments de beaucoup de professeurs et étudiants; elle n’a même pas pu être réalisée d’une manière complète. Dans la mesure où elle est réalisée, elle aboutit naturellement à chasser les étudiants étrangers.

Antoine Meillet as a sociologist of language 109 Thus what appears ‘natural’ and ‘just’ can actually be seen as a move away from the ‘literary’ and ‘international’ towards the provincial. We can assume that a langue littéraire has supra-national functions (at least when that language is French), and is associated with the haute bourgeoisie. Its creation, he suggests (1928: 227–8), is not the work of state builders: L’albanais même a été pourvu d’une langue écrite. C’est la dernière des langues indo-européennes qui ait été fixée par écrit, la dernière où ait instituée une langue ‹‹littéraire››. Mais cette langue écrite est une création en partie artificielle, comme était l’Etat albanais qu’on a essayé d’instituer de 1912 à 1914, et qui a si misérablement échoué. . . . Les Albanais . . . sont pour la plupart incultes; et leur pays, montagneux, pauvre, se prête mal à développer une civilisation; il n’y a en Albanie aucune tradition de culture propre. An implication of this is that, in Meillet’s view, the attainment of language status is not a matter of linguistic incomprehensibility with other systems; rather it is a societal attribute. Since – at least in his opinion – the idea of the various groups of people who speak Albanian being seen as a ‘society’ with a ‘culture’ is not possible, Albanian is therefore not a true language, a point to which we will return.2 This insight appears to be only partial, however, since Romanian is treated as a fully fledged langue littéraire (1928: 208): ‘Les Roumains ont une langue littéraire, bien dévelopée, et qui, appartenant au groupe roman, est de plainpied avec les grandes langues de l’Europe occidentale.’ It is probably no mistake that Romanian is both a Romance language and was heavily influenced by French in its modern development ( Joseph 1987: 100). It can therefore be established that, in Meillet’s view, more recently planned languages can share in the combined history of their more illustrious sisters; particularly if these sisters tend to be descendants of Latin (or another western European language). Moreover, the connection with the ‘language of the people’ is emphasized by Meillet (1928: 195–6) in his discussion of the development of literary Russian: La fraîcheur de la littérature russe du xixe siècle tient pour beaucoup à ce contact entre la langue de tout le peuple et la langue de la littérature. Le charme de la langue d’un Tolstoï provient de son caractère idiomatique, de la justesse de ses nuances, de son naturel. Et ce n’est pas un hasard que le genre où ces mérites d’une langue littéraire proche de la langue courante ont le plus d’avantages, le roman, soit celui où a particulièrement brillé la littérature russe du xixe siècle. That this very ‘freshness’, this very closeness to the language of the people, which he criticizes elsewhere, is what he celebrates here perhaps suggests that

110

Robert McColl Millar

what Meillet truly interprets as a langue is a language which produces cultural artefacts he finds aesthetically pleasing himself. Nevertheless, in the same passage he also comments that: ‘La Russie où, depuis Pierre le Grand, les classes dirigeantes se tournaient vers l’Europe, a été la première à s’émanciper de la langue ancienne. Le vieux slavon des premiers traducteurs slaves est fondé sur un parler de type méridional, assez éloigné du russe’. Thus the language of the people is only fully acceptable if interpreted by a highly Europeanized, implicitly anti-clerical, elite. A further term which he employs – langue commune ‘common language, language held in common’ – could be interpreted as an umbrella concept for many of the ideas described in the preceding. As a term, it has a prior existence in the works of a number of French-language scholars. Perhaps his most striking definition of this term is also amongst his shortest (1928: 244): ‘L’unité de langue commune est une force immense pour ceux qui la possèdent’. Yet his elaboration of this view is not without a degree of ambivalence. In the first place, he seems to offer a linguistic interpretation of what a langue commune should comprise,3 as he illustrates (1928: 141) in a positive sense for Serbo-Croat. He uses the same arguments negatively in his discussion (1928: 232–3) of the smaller Slavonic varieties used in the former Tsarist Empire: Quant au petit-russe, on peut regretter les procédés autoritaires par lesquels la bureaucratie du tsarisme a empêché le ruthène littéraire de recevoir aucun développement en Russie: l’unité linguistique doit procéder du libre choix des hommes, et elle n’est durable que si les sujets parlants y viennent par leur propre volonté, et en considération des avantages qu’ils y trouvent. Mais la constitution d’une langue commune petite-russienne n’était ni nécessaire ni utile. Les langues slaves de civilisation ne sont déjà que trop diverses; il convient de ne les diviser davantage que dans le cas où les parlers locales sont par eux-mêmes arrivés à se différencier de manière complète. Tel n’est pas le cas pour le éprouver aucun dommage à utiliser le grand-russe comme langue commune: la structure générale des deux groupes de parlers est la même; les différences ne portent que sur des détails; un grand-russe et un petit-russe du peuple, employant chacun leur parler local, n’ont besoin pour se comprendre que d’un petit effort. En 1920, on parlait couramment grand-russe dans les villes d’Ukraine. Faire du petitrusse la langue commune, comme le fait le gouvernement ukrainien, c’est imposer aux populations urbaines un idiome fondé sur le parler des paysans, c’est-à-dire abaisser la civilisation. Par la masse des individus qui le parlent, par l’originalité de sa littérature, par l’importance des ouvrages qu’il a servi à rédiger, le russe littéraire commençait à s’imposer au monde; on l’étudiait; il devenait l’une des langues principales de l’Europe. En s’isolant, les sujets de

Antoine Meillet as a sociologist of language 111 langue petite-russienne perdent pour eux-mêmes la bénéfice de ce progrès; et ils le ralentissent, ils l’arrêtent dans la mesure de leurs forces, pour le grand-russe; car la petite-russe est la partie à bien des égards la plus richement dotée des pays de langue russe, et, si elle cesse d’avoir pour langue commune le grand-russe littéraire, ce sera pour celui-ci une perte grave. Le dommage sera double, et il n’y aura de profit pour aucun Slave. Par leur nombre, les Petits-Russes, qui sont plus de trente millions, en Pologne et dans l’ancienne Russie, peuvent prétendre à avoir une langue écrite propre: beaucoup de populations moins nombreuses et moins cultivées en ont une. Mais la raison d’être d’une langue séparée n’est pas dans le nombre de ceux qui emploient un type de parlers; elle est dans le degré d’autonomie de ces parlers. Or, les parlers petits-russes diffèrent trop peu de grand-russe pour empêcher les Petits-Russes de prendre leur part des avantages qu’offre la langue commune fondée sur le grand-russe. Here lack of linguistic distinctiveness – that Ukrainian is not a separate linguistic system when compared to Russian – is complemented by sociological arguments; most significantly for his general thesis by the fact that the urban population of the Ukraine were essentially Russian speaking, and thus were in danger of having their ‘civilization’ overthrown or even debased by the language of a peasantry. It is significant that Meillet recognizes that ‘l’unité linguistique doit procéder du libre choix des hommes’, before adding a significant mais and expressing his own authoritative (and authoritarian) opinion that this is ‘pointless’ on this occasion. In general, Meillet appears to hold that a langue commune is part of a dynamic move away from the local and particular towards the general (1928: 117): Dans une région où coexistent un parler local et une langue commune de civilisation, les habitants qui, soit par leur situation sociale, soit par leur degré avance de culture, soit par les relations qu’ils possèdent au dehors, sont tenus de connaître la langue commune sont ceux qui ont un prestige. Ils emploient volontiers la langue commune dont la possession prouve leur supériorité. Et, en vertu de la tendance qui pousse les hommes à se rapprocher des classes considérées comme supérieures et à les imiter, tout le monde tient à connaître cette langue commune. La population tend ainsi à devenir bilingue. Le parler local ne sert plus qu’aux relations de famille, aux rapports privés. Il s’emplit d’éléments étrangers; il se vide de son originalité. Il n’a bientôt plus aucun intérêt, pas même celui de maintenir une tradition à laquelle on tient; car il devient peu à peu un calque grossier de la langue commune, avec d’autres procédés linguistiques, mais presque sans caractère propre. Le parler local est inutile le jour où toute la population, connaissant la langue commune, est bilingue. Les jeunes n’éprouvent plus alors le besoin de connaître le parler

112 Robert McColl Millar local; s’ils l’ont pratiqué dans leur enfance, ils l’oublient l’âge mûr venu. Il en est des patois comme des costumes locaux que les vieux continuent à porter, quand déjà leurs enfants les ont abandonnées et ont adapte le costume habituel dans la ville voisine. Un jour vient où, avec la mort du dernier vieillard qui le savait, le parler s’éteint. C’est ainsi qu’on a vu mourir sur les bords de la mer Baltique les parlers slovinces éliminés par l’allemand; il y a quelques années déjà, on ne trouvait plus dans chaque village que quelques personnes âgées qui les savaient encore; on ne sait s’il en subsiste aujourd’hui aucun. Il en va de même du breton dans beaucoup de localités. As more recent scholarship such as Gal (1979), Fishman (1991) and Macafee (1994) demonstrate, the typology of domain loss for lesser-used languages which Meillet describes has a considerable basis in fact. Yet inherent in Meillet’s view is the idea of superiority, not only through greater levels of communication, but also through the societal importance of one variety over another – implicitly because that variety is associated with ‘progressive’ and hierarchical developments. The sadness which post-war linguists express over language death is matched here by, at best, an entirely neutral stance which might even represent a degree of satisfaction. Meillet would argue that this sense of commune should normally be associated with originalité, with civilisation, a point which he supports through his discussion of developments in Ancient Greece (1928: 148–9), a primary and innovative focus of his historical linguistic scholarship. On these occasions we see suggested a range of competing reasons for the establishment of a langue commune: associations with a literary/philosophical tradition, with imperial expansion and with a hierarchical structure of government. Meillet’s particular ideological position becomes clearer when he discusses the modern age, significantly, perhaps, in relation to French and France (1928: 174): La langue commune française est donc une langue traditionnelle, littéraire, aristocratique, et qui ne peut être maniée d’une manière courante que par des personnes ayant un degré très élevé de culture. Elle a été créée par le travail d’une élite intellectuelle et d’une élite sociale. C’est une sorte d’idéal dont les Français se rapprochent plus ou moins sans qu’aucun arrive à le réaliser. This is in marked contrast to the actual political structure of Modern France, whose culture he explicitly dismisses: Actuellement la France a une organisation politique démocratique, et tous les citoyens y ont accès à une certaine culture, qui est souvent faible, qui même pour la plupart de ceux qui passent par l’enseignement secondaire est très médiocre.

Antoine Meillet as a sociologist of language 113 (Meillet appears, as Vendryes (1937) hints, to have been a rather conservative republican who took a somewhat cynical view of the internal affairs of the country while remaining a staunch French patriot.) Meillet does admit (1928: 175) that the situation in France is ‘extrême’. But he points out that: . . . toutes les langues communes qui ont commencé de se fixer à l’époque de la Renaissance présentent, à des degrés divers, une situation comparable. Comme le français, et en partie presque autant, l’anglais, l’italien, l’espagnol, le portugais, l’allemand, le polonais sont des langues traditionnelles, créées par des élites pour des élites, qu’on ne peut parler et écrire qu’au prix d’un sérieux apprentissage et dont la pratique suppose une forte culture. Perhaps centrally from his point of view: ‘Dans l’Europe occidentale et en Amérique, des langues aristocratiques servent d’organes à des démocraties qui s’embourgeoisent.’ A language is therefore envisaged as having an existence of its own, in some way separate from the people who speak it. A langue commune, Meillet suggests, can only flourish if its nature and acculturative processes are decided upon by an essentially non-democratic elite who can control its use even among the mass of its speakers. Whilst grudgingly accepting the democratic order which appeared to have triumphed across Europe at the end of the First World War, Meillet suggests that it is only these ‘aristocratic’ (a highly loaded word for a French person to use) arbiters who can protect a language from the ‘feebleness’ it receives culturally from the majority of its speakers. At its heart, Meillet (1928: 105) sees the growth of a common, elitist language in the place of localized varieties as being a necessity to guarantee the transmission of culture, as well as to facilitate the establishment of government over extensive areas. As he puts it, ‘Les grandes forces collectives agissent au profit de l’unité de langue’. He uses (perhaps predictably) the example of France. In a passage too lengthy to quote here, Meillet (1928: 105–7) creates a litany of devotion to the standard, national, language. It is ‘la langue de l’administration; c’est la seule langue qui puisse s’employer avec les agents de l’Etat’. Moreover, it is ‘l’unique langue admise dans les contrats que sanctionne la puissance publique’; ‘la langue de l’école’ as well as ‘la langue de l’armée’. Thus the French language is in some ways representative of historical continuity for the nation, both in its Revolutionary solidarity of army and school, and in its appropriateness and elitism. It is also the language of modern technology: ‘[l]a langue des chemins de fer’. The koiné function of the common language is also stressed: ‘[l]a croissance des villes impose l’emploi de la langue commune. Les grandes villes et les localités industrielles tirent souvent de très loin et de régions diverses, en partie de l’étranger, l’afflux de population dont elles ont besoin pour se développer.’

114 Robert McColl Millar As well as a view based on ‘civilization’, a utilitarian viewpoint is emphasized. Again, language status is implied through the presence of prose in that language. There are contradictions in his viewpoint, but these are contradictions which have been maintained in theoretical debate over the standardization process up to the present time, as seen in the work of Kloss (1978) in comparison to Joseph (1987). In sum, he describes his view of what a langue is in the following terms (1928: 240–1): En se donnant une langue de civilisation propre, une nation s’isole des autres. Elle devient comme une cellule particulière de la civilisation universelle. A la littérature, cet isolement ne nuit pas en général. Chaque langue a ses moyens d’expression, ses nuances. Chacune peut fournir à un écrivain doué le moyen de faire une œuvre originale, et qu’il n’aurait été possible de composer en aucune autre. Si Mistral avait écrit en français, il n’aurait pas été le poète qu’il a été; sa Mireio vaut parce qu’elle exprime la Provence même, et c’est la langue provençale qui a permis une telle œuvre; en la permettant, elle en a d’ailleurs limité la diffusion. Mais l’originalité d’un dialecte est limitée; les possibilités d’expression d’une langue demeurée ou redevenue inculte sont peu variées. Quand un poète ou deux ont tiré les effets qu’elle comporte, la mine est épuisée. Ce n’est pas un accident que Mistral n’ait pas eu de véritable successeur. De plus, cette originalité ne vaut guère que pour la poésie, ou du moins pour la littérature d’imagination, où le pittoresque a une place notable. Si un Provençal s’avisait d’écrire en son dialecte un ouvrage qui aurait une portée philosophique, il priverait de la possibilité de le lire la plupart des Français et le reste du monde. Et les quelques détails d’expression originaux qu’il y pourrait employer n’ajouteraient guère à la valeur de son exposé. ‘Originality’ is a quality he therefore associates with the acquisition of the status of language, at least when the variety involved has a lengthy literary history connected, again, to philosophical inquiry. This is in contrast to his view of standard languages which have developed more recently: D’ailleurs les nouvelles langues nationales qui s’organisent sont faites par des professeurs, des instituteurs, des journalistes. Elles calquent les éléments abstraits des grandes langues de civilisation bien plus qu’elles ne mettent en œuvre l’originalité des parlers sur lesquels elles reposent. Leur valeur est pédagogique et politique plutôt que littéraire. La saveur propre de chaque parler en est absente, ou du moins s’y fait peu sentir. Jusqu’à présent la littérature n’a guère profité de la plupart des créations ou des résurrections de langues écrites, et, de toutes, c’est le provençal, dont le rôle pratique a été le moindre, mais le ‹‹rendement›› poétique le plus grand.

Antoine Meillet as a sociologist of language 115 Again, a dichotomy is established between political or pedagogical purposes on the one hand and literary or philosophical on the other. What happens when, in Meillet’s opinion, these criteria are not fulfilled, but a language has still taken on an official capacity, will be dealt with in the following. At the other end of his scale lie a number of terms referring to local language varieties: parler ‘vernacular’, patois ‘local variety’ and, to a lesser extent, dialecte ‘dialect’.4 Meillet perceives this set of non-standard varieties as having a historical basis. In the first place, initial unity is replaced by diversity at a highly local level due to societal breakdown (1928: 129–30). As with other elements in his model, a progressive connection with the history and societal structure of a given area is stressed. Thus these initially highly localized variants eventually crystallize into a larger unit with wider political and cultural associations. His statement (1928: 130) that ‘Il y a en Allemagne des dialectes parce que la féodalité y a laissé des Etats ayant une autonomie; in n’y a presque pas de dialectes en France, surtout dans la France du Nord, parce que la royauté centrale a été puissante de bonne heure’ equates the maintenance of dialect with the maintenance of feudalism.5 Modern, progressive, ideas will gradually shape a language towards unification. Thus a hierarchy can be set up whereby a langue is interpreted as the result of a unification process (which Meillet often describes as fixée), a dialecte of a similar, but lesser, process, whereby unity is achieved in a more geographically circumscribed area, perhaps due to feudal obscurantism, and – we can assume, given the definitions given for langue commune above – with a rather more limited function, and a parler or patois of a highly limited language variety both in terms of geographical extent and function. According to Meillet, parler can also represent a more modern function – it appears to be his assumption that the written dialectes would fade away in the modern age (an assumption with some basis in fact). In one sense, a parler would represent the linguistic situation on the ground if we ignore (or at least disregard) the association a given parler might have with a langue commune, a point he makes most forcibly in his discussion of the Slavonic dialects of southern Europe (1928: 131–2). In a sense he is correct, since we could argue that languages in the modern era have been as much ‘imagined’ as are nations: a langue commune is, he would claim, built on the back of the parlers. Particularly illuminating is his discussion of Macedonian, and his idea of parlers de transition ‘transitional vernaculars/dialects’. The fact that this seems to contradict his idea of langue as a product of ‘aristocratic’ elitism is highly noteworthy. His initial ideas are only really sustainable in those circumstances where there has been considerable societal stability and continuity. For most of the new languages of Europe, this is essentially an impossibility; indeed, he veers close to a form of linguistic democracy when he notes (1928: 149–50) that Il est commode d’avoir pour langue commune un idiome qui soit analogue au parler couramment employé dans l’usage local et familier. L’effort à faire par un Français du Nord pour s’assimiler le français

116 Robert McColl Millar commun est peu sensible; chacun aperçoit de suite les systèmes de correspondances linguistiques qui lui permettent de passer de son parler local au français commun. Ce français commun est pour lui la norme idéale d’une langue dont son parler maternel n’est qu’un autre aspect. Yet Meillet also presents a darker interpretation of the idea of parler. In the modern era, he seems to suggest that the use of a parler is a marker of lack of progressive views and of a lower social background (1928: 108–9): . . . à un linguiste qui lui demandait s’il savait le patois de son village, un paysan vosgien répondait, il y a déjà vingt ans: ‹‹Oui, mes parents m’ont parlé patois; et ce n’est pas le meilleur service qu’ils m’aient rendu.›› This passage is particularly interesting because of its use of constructed dialogue. As Tannen pointed out (1989: 133), an effect of this type is often used in reported conversation because it ‘creates involvement by both its rhythmic, sonorous effect and its internally evaluative effect’. It may also allow the middle class observer to place his or her own views, if not prejudices, in the mouth of an apparently ingenuous working class rural inhabitant. If anything, this is even more pervasive here because of the idea of the anonymous scientist supplied by the phrase un linguiste. The problem, as Meillet perceived it, was that, in the post-revolutionary Zeitgeist of Europe, and, in particular, in the upsurge of nationalism towards the end of the nineteenth century, speakers of parlers, or users of languages such as Dutch in Belgium (1928: 150), who were, at best, ‘demi-cultivés’, were no longer willing to ‘surrender’ their native language varieties in favour of an elitist langue commune. As Meillet points out (1928: 200): Dans la grande bande qui va de la Finlande à la Grèce, la division linguistique est extrême, et la situation trouble presque d’un bout à l’autre. Depuis les traités par lesquels s’est terminée la grande guerre, il y a là un grand nombre d’Etats indépendants, dont chacun a une langue officielle distincte. Cette langue repose sur les parlers de la majorité de la population, qui, dans une notable partie des cas, est surtout rurale; dans les villes, la population a en partie, et parfois en majorité, une autre langue ou d’autres langues. Souvent ces langues officielles ne s’écrivent couramment que depuis peu, ne se sont fixées sous la forme actuelle qu’au cours du xixe siècle et sont inconnues au dehors. La langue de l’état est celle d’une majorité plus ou moins forte; elle n’est pas la langue du pays entier. Il y a des minorités. Souvent, ces minorités sont d’anciennes classes dominantes; les gens qui les parlent ont conservé une aisance; et surtout le niveau de leur culture est élevé, supérieur a celui de la majorité; leur langue est parfois une grande langue

Antoine Meillet as a sociologist of language 117 de civilisation, officielle dans un pays voisin. Il a donc fallu tenir compte des minorités. C’est pour ces cas qu’a été constitué le droit des minorités qui a crée des situations linguistiques. Thus the old leadership of an area are rendered minorities; the aristocratic/authoritarian order passes away. Discussion In summary, therefore, Meillet’s implicit and explicit views can be presented in the following manner. He sees a dynamic model of the social relationship between language varieties essentially based upon a polar distinction of absolutes. At one end are parlers which, in his opinion, lack any connection with civilisation. At the other end are grandes langues de civilisation, which achieve this eminence not only by their number of speakers or even their native speakers’ association with a ruling elite; rather, these languages develop through the work of a self-perpetuating, oligarchic intellectual elite (in his terms, an ‘aristocratic’ elite) capable of the production of works of originality and aesthetic importance expressing the ‘genius’ of that language. Another term for this might be langues savantes. They are by their nature standardized. In between these opposites are a number of stages. In the past, it was possible to talk in terms of langues vulgaires; probably these should be associated with the functions he ascribes to dialectes in the modern period. These language varieties are capable of literature in their own written form, but will not be used either as the language of administration or the language of intellectual debate. When he is in favour of their use (largely in the past), dialectes are seen as capable of developing into langues officielles, associated with a given ethnic group; in particular, these should have an essentially urbanbased culture dominated by a socially prestigious, western European style, intellectual class. When he is less in favour of this happening (and, as we have seen, he is particularly worried about this development when it comes to encroachment upon the position of the French language), he tends to question whether these new national languages are anything more than part of a process whereby semi-educated people of lower middle-class origin and a largely illiterate peasantry work in harness in opposition to the (former) ruling classes and their language. Because of this origin, Meillet would claim that the language involved must inevitably suffer from a number of weaknesses when it comes to the expression of originality in thought. As he rightly points out, at the time he was writing this was less of a problem in western Europe than it was in the central or eastern parts of the continent. Even when the process is carried out and completed in sociological terms, it may result merely in the development of a ‘mediocre’ standard language, with the educated classes being compelled to use another language in their intellectual discourse; this, Meillet would claim, places such people at a disadvantage in

118 Robert McColl Millar the apparently Darwinian struggle for the expression of thought. What it means for the majority of the people of a territory that the educated minority should speak another language by preference is not treated at all by Meillet.

Conclusion Meillet’s Les langues dans l’Europe nouvelle is a work of considerable importance in our understanding of the development of the sociological interpretation of language use between the wars. Primarily this is due to his appreciation of the dynamic relationship between historical processes and social patterns in the development of what is meant by language. The work is also fascinating in its portrait of the views of a member of the French middle classes of considerable intellectual ability when faced by the changes through which Europe passed in those years. Echoes of this dynamism he promotes can be found, whether consciously or not, in the works of Kloss (1929, 1952, 1978 in particular), where Sachprosa is established as the most important determinant in the development of a standardized language, in the work of Haugen (in particular 1966), where the elitist attitude to language development is at least partially applauded, Sommerfelt (1938) and Joseph (1987), where particular stress is given to literary acculturation as a focus for standardization. The French tradition, of which Meillet was a part, continued well into the post-war period in the work of Cohen (for instance, 1955), among others. What most, if not all, of these later scholars lack is the near-triumphalist views of urban and urbane linguistic superiority which Meillet embraces. It is not, of course, that Meillet was necessarily ‘wrong’ in his appreciation of the social and ideological background of the actors who brought about the development of a standard form of a given language; rather it is that he seems largely uninterested in what non-elite parts of the population might desire in order to achieve linguistic (as well as, presumably, political) democracy. The fact that the book demonstrates some discomfort with this view is, in fact, to Meillet’s credit. Meillet’s work can therefore be seen to stand on the cusp between the previous authoritarian and prescriptive discourse of the French linguistic tradition and a newer sociology of language which embraced diversity and at times actively encouraged, and engaged in, language development.

Acknowledgements I am grateful to the British Academy (Grant SG-35239) for support which made this research possible.

Notes 1 Of course, that would not have removed all ideological bias from Meillet’s work. Given his pro-Slav stance, evinced in his sometimes highly overt dislike of

Antoine Meillet as a sociologist of language 119

2

3 4

5

Hungary and Magyar, and the French ‘little entente’ with Czechoslovakia and Poland, among other nations, it is inevitable that Meillet should discuss these nations and their languages favourably. When discussing Modern Greek katharevousa elsewhere, Meillet (1920: 251–2) makes much of the status of that literary language as a langue artificielle, blaming lower government officials and school teachers – implicitly members of the lower middle classes – for this artificiality. Elsewhere in his work he comments that ‘Toute langue commune est le résultat de l’extension d’une langue dominante au-delà de ses limites premières’ (Meillet [1913] 1920: 13). Meillet uses the term langue when referring to varieties of this type only when discussing matters in the past when there was no official, cultured language. Even then he uses terms such as langue vulgaire ‘language of the vulgar, the unlettered’ (1928: 158–9). To someone not from France, it may seem strange that Meillet’s apparent admiration for aristocratic values should not translate into an admiration of feudal regionalism; it is worth remembering, however, that one of the single most vital reforms of the period of Louis XIV was the breaking of the old feudal aristocracy and the creation of a new who were essentially creatures of Versailles and of the crown. Whilst similar events took place in, for instance, the British Isles, they were never so clearly absolutist in their tendencies (Mettam 1988: chapter 2).

8

‘The grammatical being called a nation’ History and the construction of political and linguistic nationalism John E. Joseph

Introduction Languages are historical products, therefore any attempt to derive enlightenment from them about the reality they purport to represent is illusory. This is something we have known for an awfully long time – it is, in fact, the main theme of Plato’s Cratylus – yet have shown a remarkable capacity for ignoring. Kelly (2002) offers a thorough account of how medieval linguistic thought, built upon Platonic foundations, nevertheless finds direct reflections of the divine order of the universe in the grammar of Latin, exactly what Plato, to his evident disappointment, found could not be done. In Joseph (2000) I have traced some of the ways in which modern linguistics continues the quest to discover natural order within language, using various strategies to erase the traces of the wilful human input into language that is principally what we mean when referring to it as a historical product. This centuries-long continuity in the suspension of disbelief over the naturalness of language is at the heart of what Roy Harris calls the ‘language myth’. I have disagreed with Harris over the actual continuity of particular elements he identifies as constituting that myth ( Joseph, 1997, 2003), and over whether we are really dealing with the clear-cut distinction between illusion and truth that his use of the term ‘myth’ implies, but I have also taken pains to stress that his powerful expositions of the topic have been among the high points of late twentieth (and now early twenty first) century writings on language. One thing on which I think he and I agree is that language ‘myths’ have been a central component of culture and merit study in their own right, study which stands to illuminate the development of virtually every aspect of human experience. This view has also been shared by some very prominent historians and social anthropologists over recent decades, who have shaped the contemporary discourse about the origins of the modern concept of national identity, the historical events which surrounded them, and the role played by theories of the nature and function of language and languages. Although this discourse has taken the form of debates among the participants, making it easier to discern what separates them than what unites them, part of their communal

The construction of nationalism 121 project has been to shift the understanding of nationalism from an ‘essentialist’ toward a ‘constructionist’ mode. We shall see that they have sometimes accomplished this by shifting the essentialist burden onto the national language, turning it into something that can be taken for granted as what members of a nation share and what the rest of nationalism can be constructed upon. But we shall also examine more recent work which has considered how a national language is itself constructed as part of the nationalist project. Those of us whose principal interest is in the development of theories of language need to attend to this discourse, because the evidence suggests not that any one of the triad – historical events, nationalism and theories of language – shaped the other two, but that the shaping was mutual and interwoven. The present paper explores their interaction over the course of the nineteenth century, and the changing interpretation of this interaction over the past four decades of enquiry into linguistic nationalism.

The nature of national identities ‘Nation’ is an inherently ambiguous word, used sometimes in its etymological sense of people linked by nativity, birth, as when one speaks of the Hebrew nation or the Cherokee nation. More often it is used in its extended sense of an expanse of territory, its inhabitants and the government that rules them from a single, unified centre – the British nation, for instance. When the etymological and extended senses of nation coalesce, the term ‘nation-state’ is sometimes used. Thus Ireland (Eire) would count as a nation and a nationstate, whereas the United Kingdom is only a nation in the extended sense, comprising as it does at least four nations in the etymological sense, the English, Northern Irish, Scots and Welsh. Scotland, Wales and others of their type are sometimes called ‘nations without states’. There is a problem in that the two basic senses of ‘nation’ can never really coalesce. For them to do so, no one but members of the nation-by-birth would inhabit the national territory, and no member of the nation-by-birth would live outside the territory. Such a perfect mapping constitutes the ‘ideal’ of the nation-state – a dystopian rather than a utopian ideal for anyone but the most rabidly puristic nationalist. In the modern world, affirmation of belief in the nation-by-birth has been strongest whenever a political nation has perceived itself to be under threat from ‘outsiders’, either because immigration has made the population visibly diverse, or because of imperial or colonial dominance (see further Joseph, 2004, chapter 5). The constructionist turn in sociology, social psychology, linguistic anthropology and political science has affected the analysis of national identity at least as much as any other form of identity. Indeed, the repeated reconfigurations of national boundaries in the wake of the two world wars, the reorganisation of the USSR and Eastern bloc countries in 1989–1991, and the recognition of sub-national entities in Western Europe in the 1990s have all contributed to a strong awareness of the fluidity and arbitrariness of nationality. Although

122 John E. Joseph this awareness has not destroyed an underlying belief in ‘real’ national identity as something imposed on us by birth or early circumstances and remaining essentially unchanged thereafter, it has undoubtedly helped fuel the analytical trend among scholars to treat such beliefs as mythical, and to strive instead to understand identity as something we construct and negotiate throughout our life. A consistent theme within studies of national identity over the past four decades has been the central importance of language in its formation. As we shall see, a number of prominent historians, sociologists and political scientists have argued that the existence of a national language is the primary foundation upon which nationalist ideology is constructed. Others, however, have paid more serious attention to the evidence compiled by linguistic historians showing that national languages are not actually a given, but are themselves constructed as part of the ideological work of nationalism-building.

When did linguistic nationalism begin? As with many ‘doctrines’ which represent the articulation of something that has already been being put into practice for some time, it is debatable where the beginning of nationalism in general, and language nationalism in particular, should be located. Modern scholarly opinion as surveyed in this chapter places it anywhere from the late eighteenth to the late nineteenth century. But even if it is true that nationalism underwent a sea change at some point in the last 250 years, it did not arise from nowhere. Modern nationalism certainly exhibits important continuities with national identities that extend all the way back to the beginning of recorded history. The Old Testament records the oral traditions of the Hebrew nation concerning its origins, beliefs, relations with neighbouring nations, reduction to slavery and estrangement from its homeland, followed by its return to the homeland as prelude to its golden age. It was not written merely as an historical chronicle, but also to manifest and assure the nation’s ongoing existence. Developments in nationalism in the eighteenth, nineteenth and twentieth centuries were all interpreted via their refraction through the biblical texts, the common base of European culture across national and social divides. Nations make their first appearance in Genesis 10. The chapter lists the names of the sons of Shem, Ham and Japheth (the three sons of Noah), together with the places where they dwelt, sometimes with precise specification of borders. Each of the three sets concludes with a passage like the following: By these [seven sons and seven grandsons of Japheth] were the isles of the Gentiles divided in their lands, every one after his tongue, after their families, in their nations. (Gen. 10:5)

The construction of nationalism 123 Land, tongue, family . . . nation. All laid down in the Book of Genesis – according to believers, by the hand of God Himself. Genesis 10 is a genealogical interlude between the story of the Flood (Gen. 6–9) and the account of how the descendants of Noah were subsequently spread across the world (Gen. 11). At the beginning of Gen. 11 we revert to a time when ‘the whole world was of one language, and of one speech’ (Gen. 11:1), and the whole tribe of Noah, journeying westward, finds a plain in the land of Shinar and dwells there. They decide to build ‘a city and a tower, whose top may reach unto heaven’, plus one thing more: and let us make us a name, lest we be scattered abroad upon the face of the whole earth. (Gen. 11:4) Implicit here is the belief that, unless they have a shared name – which is to say, a national identity – they will inevitably scatter. An identity must be constructed for the nation to cohere, for its members to be mutually interdependent and form cities, rather than disperse, each to seek his own patch – a dispersion into rural space that in time would come to be characterised as ‘natural’, in opposition to the ‘artificial’ formation of urban spaces. The ancient empires of the Mediterranean basin were well aware of the nations they overlay. In more modern times, English nationalist sentiments are obviously present in Shakespeare’s history plays from the end of the sixteenth and start of the seventeenth centuries – but to call them ‘nationalist’ is, arguably, anachronistic, if the whole concept of nationalism as a general doctrinal position does not appear until two centuries later. There is wide agreement that the American Revolution of 1776–1781 and the French Revolution of 1789–1793 were cardinal events in establishing the modern concept of nation as a political reality. But in a book that may be said to have got the contemporary scholarly discourse on nationalism under way, Elie Kedourie (1926–1992) identified the crucial change as having taken place at the start of the nineteenth century, triggered by the Napoleonic aftermath of the French Revolution. His book begins with an intentionally provocative opening sentence: Nationalism is a doctrine invented in Europe at the beginning of the nineteenth century. [. . .] Briefly, the doctrine holds that humanity is naturally divided into nations, that nations are known by certain characteristics which can be ascertained, and that the only legitimate type of government is national self-government. (Kedourie, 1960, p. 9) Most prior work on nationalism, including comprehensive studies by Deutsch (1953) and Shafer (1955), had focussed on its twentieth-century manifestations while assuming that the nation itself, as a social structure, had

124 John E. Joseph existed in its modern form at least since the Renaissance, with nationalism as its inevitable ideological accompaniment. Moreover, having become the basis of political and social organisation world-wide, nations and nationalisms doubtless always would exist – unless, of course, Marx was right, and the nations of the world would fall one by one, like ripe apples, into communist internationalism. Karl Marx (1818–1883) had not invented the constructedness of nations. He was already able to quote from Thomas Cooper (1783–1839), writing in 1826, that ‘The moral entity – the grammatical being called a nation, has been clothed in attributes that have no real existence except in the imagination of those who metamorphose a word into a thing [. . .]’ (in Marx, 1955 [1847], §3a). Unsurprisingly, Marx interpreted that reification of the concept of nation in class terms, as a means whereby the bourgeoisie protects and maintains its interests. The existence of nations was, like religion and capitalism, merely a necessary phase in the historical development of mankind toward socialist perfection. The fact that Marx’s analysis was tied to a revolutionary programme aimed at bringing a quick end to those less-than-perfect phases probably made it more difficult than it might otherwise have been for non-Marxists (especially anti-Marxists) to accept the key notion that the concept of the nation was a historical product. A powerful non-Marxist case to this effect was made in 1944 by Hans Kohn (1891–1971), who argued that nations are a modern concept dating back not earlier than the mid-eighteenth century, and that ‘Nationalism is first and foremost a state of mind, an act of consciousness, which since the French Revolution has been more and more common to mankind’ (Kohn, 1944, pp. 10–11). A difficulty with Kohn’s argument lay with its being grounded in an essentialist dualism between a ‘voluntaristic nationalism’, characteristic of England and France, versus the ‘organic nationalism’ of Germany and Central European nations – tied to the empiricist philosophical tradition of the former and the rationalist one of the latter. His positive portrayal of voluntaristic nationalism and criticisms of organic nationalism played well to the wartime audience, but lost relevance when, after the war, the key dualism came to be that of Marxist anti-nationalism versus any nationalism at all. Deutsch (1953) tried to fill the void in typical modernist fashion by reimagining nationalism in social science terms, starting with his redefinition of a people as a ‘community of social communications’, and seeking a quantitative methodology for pinning down what nations really are – a desire that, today, seems almost poignant in its hopelesness. Kedourie (1960) presented a more purely constructionist view than Kohn’s, replacing nationalism as ‘act of consciousness’ with nationalism as doctrine, unambiguous in its conventionalism, and pushing its beginnings forward by a few decades. By placing the concept within a historical context in which Marx simply did not figure, he made it possible for political scientists, historians and other scholars to treat nations and nationalism as historical contingencies without having others automatically classify their

The construction of nationalism 125 work as partisan. As we shall see, some of the most important work to benefit from this shift would begin by disagreeing vehemently with Kedourie on various particulars, though still acknowledging his key role in setting the discourse in motion, and for drawing attention to a thinker who, whenever one takes nationalism to have begun, was one of its most original and robust theoreticians, Johann Gottlieb Fichte (1762–1814).

Fichte on language and nation General Napoleon Bonaparte seized control of the French government in 1799. In 1803 he became President of the Italian Republic as well, and in 1804 the French senate and people voted him their emperor. Over the next six years he expanded his empire to include most of Europe. It was in this period that German Romantic thinkers, many of whom had previously heroworshipped Napoleon as the embodiment of the possibilities of the human will, now had to come to grips with the fact of having their own country defeated by him and themselves becoming his imperial subjects. From this experience arose the argument that such imperial rule was unjust, because it is natural for each nation to rule itself. But what were the ‘natural’ boundaries of a nation? This was the key question, to which the answer seemed obvious to anyone in this period when the predominant definition of ‘nation’ was the extended one focussed on the expanse of territory. The natural boundaries were the geographical obstacles, the seacoasts together with any mountain ranges or great rivers cutting the nation off from easy reach of its neighbours. But by that answer, there was nothing in principle to prevent ‘Europe’ being conceived of as a ‘nation’ rather than an empire composed of nations. None of the natural barriers within it was insurmountable (apart from the English Channel). In particular, what most concerned the German Romantics was that there was certainly no great land or water barrier defining their nation as distinctive from those of their neighbours to the east or west. If the right of the German nation to autonomy was to be maintained by something more fundamental to the Romantic mind than mere historical difference, something non-geographical yet plausible as a primordial, ‘natural’ boundary had to be identified. One solution might have been to hark back to religious affiliation, on which the whole pre-modern edifice of dynasty had been defined. But all Europe was officially Christian, and however strong were the doctrinal differences within Western Christianity, particularly those separating Protestants from Roman Catholics, Germans in particular could not overplay them without weakening Western unity in the face of ever-present fears of Orthodox Slavs to the east. Moreover, in the wake of the Enlightenment, mainstream European thought was grounded in the secular. Arguments built upon a religious foundation had the air of belonging either to a bygone age or to an increasingly specialised domain of theology.

126 John E. Joseph The most convincing answer was formulated by Fichte in an 1806 ‘address to the German nation’, in which he argued that what defines a nation most clearly is its language. The first, original, and truly natural boundaries of states are beyond doubt their internal boundaries. Those who speak the same language are joined to each other by a multitude of invisible bonds by nature herself, long before any human art begins; they understand each other and have the power of continuing to make themselves understood more and more clearly; they belong together and are by nature one and an inseparable whole. (Fichte, 1968 [1808], pp. 190–1) In the context within which Fichte wrote, language was by no means the obvious candidate to stand as the defining characteristic of nations. Most of the languages of Europe were understood to have descended from a common ancestor tongue, with the differences being merely the historical by-product of different subgroups of the original tribe having settled in different parts of the continent, separated by the geographical obstacles that were assumed to be the natural, primordial boundaries of nations, and remaining relatively isolated for long stretches of time. Fichte turned these traditional views on their head. From this internal boundary [of language], which is drawn by the spiritual nature of man himself, the marking of the external boundary by dwelling place results as a consequence; and in the natural view of things it is not because men dwell between certain mountains and rivers that they are a people, but, on the contrary, men dwell together – and, if their luck has so arranged it, are protected by rivers and mountains – because they were a people already by a law of nature which is much higher. Thus was the German nation placed – sufficiently united within itself by a common language and a common way of thinking, and sharply enough severed from the other peoples – in the middle of Europe, as a wall to divide races not akin [. . .]. (Ibid.) Fichte’s writings are given the principal credit for rousing Germans to rise up against Napoleonic rule. The point of view he espoused was by no means just a political one, though; it resonated so loudly just because it accorded so well with the idea system of German Romanticism generally. Neo-Platonic in character, it was oriented toward the realm of eternal ideals, and located reality not in the world of mere superficial appearances and historical contingencies, but in the permanent, unchanging essence of things. In the case of a nation, its essence exists in purest form in its founder, and that essence that was in him persists throughout the whole history of the nation, providing the

The construction of nationalism 127 base of its language, culture, way of thinking and intellectual and artistic achievements. However, mixture with other nations means dilution of this essence. Such a whole [as the nation defined by language], if it wishes to absorb and mingle with itself any other people of different descent and language, cannot do so without itself becoming confused, in the beginning at any rate, and violently disturbing the even progress of its culture. (Ibid.) This particular aspect of Romantic thought, which follows logically from its founding principles, would lead to the development of ‘scientific racism’ from the mid nineteenth to the mid twentieth century, with unspeakably horrible consequences. Whether any of those writing in this period foresaw such developments is a matter of interpretation and debate, but in the case of Fichte, one can be quite confident that his intention was to rescue the German nation, language and culture from what seemed at the time like an overwhelming domination by the French, with little if any thought that one day his own countrymen might invoke his equation of absorption with confusion as part of a rationale for genocide.

Renan and the Kedourie–Gellner debate Halfway between Napoleon and Hitler occurred an event which put France in a position very like the one Germans had felt themselves in seven decades earlier. Between 1863 and 1871, Prussia, led by Otto von Bismarck, unified the nation of Germany under its own hegemony, through a series of victorious wars against Denmark, Austria and France. The culmination of the FrancoPrussian War with the siege of Paris in 1870–1871 was a defining moment for modern nationalism in a number of respects: it ended with the proclamation of the German Empire – modern Germany as we know it – and the Empire’s annexation of Alsace-Lorraine, territories which had repeatedly shuttled between French and German rule, where the local dialects were Germanic but the political allegiance of the populace was strongly to France. Paris continued to resist the new German Empire after the rest of France had surrendered, and for two months it was ruled by the Commune, a loosely organised proletarian ‘communist’ government that was finally crushed by the new French national provisional government formed in the wake of the treaties with the Prussians. These events had a tremendous impact on the French psyche, comparable to that of Napoleon’s victories on Germans at the start of the century, which produced Fichte’s writings on nationalism among many others. Fichtean arguments about language defining a nation naturally were the mainstay of Germany’s justification for the annexation of Alsace-Lorraine. So powerfully had this way of thinking shaped the modern European conception of

128 John E. Joseph nationalism that even Frenchmen who believed wholeheartedly that Alsace-Lorraine must be French could not find an obvious way to counter the linguistic argument. It was finally a linguist, Ernest Renan (1823–1892), who produced a new conception of nationalism in response, and it was that conception that would become the basis for the Wilsonian principles by which the twentieth-century world map was redrawn at Versailles in 1919. Renan’s 1882 address ‘Qu’est-ce qu’une nation’ is universally cited as a landmark. His conception of the nation starts from the Romantic idea of a shared âme (a word that means both ‘mind’ and ‘soul’), as one would expect from someone whose approach to language, mind and race had crystallised in the 1840s under the influence of Herder (see Joseph, 2004, p. 44). But he passes beyond the Romantic when he breaks this âme down into its component parts: a heritage of memories, plus a will to continue validating that heritage of memories. A nation is a soul, a spiritual principle. Two things that are actually one make up this soul, this spiritual principle. One is in the past, the other in the present. One is the common ownership of a rich legacy of memories; the other is the present-day agreement, the desire to live together, the will to continue validating the heritage that has been inherited jointly. (Renan, 1882, p. 26, my translation)1 The nation, in other words, exists in the minds – the memories and the will – of the people who make it up. This is the conception that Anderson (1991, p. 6) would return to in defining the nation as ‘an imagined political community’. The ‘legacy of memories’ Renan pointed to would dominate future philosophical and academic attempts to analyse national identity. The other element, the collective ‘will’ of the people, would however have the deepest political impact, starting at Versailles. It has continued to be the assumed basis for the legitimacy of the political nation up to the present time. Renan would turn up at the heart of the first great debate in the contemporary discourse on nationalism, which took place between Kedourie and Ernest Gellner (1925–1995). The two were friends, and each acknowledged the other as having played a key role in the formulation of their conflicting views of the nature of nationalism, which differed on two fundamental points. The first is that, for Gellner, Kedourie’s view of nationalism as ‘a doctrine invented in Europe at the beginning of the nineteenth century’ (p. 123) turned it from the natural, necessary, universal historical development it was generally assumed to be, into something ‘utterly contingent, an accidental invention, a by-product of the scribblings of a set of thinkers in one particular historic situation’ (Gellner, 1997, p. 10, italics in original). Gellner credited Kedourie with awakening him from ‘dogmatic slumbers on this point – until I read his book, I continued to assume, or at least not to criticise with lucidity, the “naturalness” view of nationalism’ (ibid.). But while taking Kedourie’s point that nations are not an inexorable historical development for

The construction of nationalism 129 all people everywhere, Gellner rejected the further conclusion that nationalism is just an ideological accident that would never have come about if Kant and Fichte had not happened to write what they did. Gellner set himself the task of identifying the factors that made certain peoples historically prone to nationalism and others resistant to it, and soon lighted upon the very one Fichte had pointed to, the possession of a common language. As a result, contemporary scholarship on nationalism and national identity has tended, following Gellner, to assume language as a foundational factor, a tendency helped along by a general ‘post-structuralist’ ethos that sees all social structures as linguistic constructs. The Kedourian alternative, in which language is downgraded from primordial binding force of the nation to just one of several ideological sites within nationalist rhetoric, would find echoes in the arguments of those post-structuralists too wary of essentialisms to assign language or any other single factor a foundational role. Gellner (1964, chapter 7) argued that nationalism was best understood as the result of the uneven way in which modernisation had spread, causing massive economic and social changes, disrupting traditional lifestyles and motivating people to move from the countryside into cities. Traditional village and tribal structures upon which social organisation had been based no longer functioned and had to be replaced, and what was available to replace them in the urban context was language and language-based culture, especially print culture. Modern education, funded by the state, grew up around the printed word, and functioned as an institution for creating new social hierarchies based upon literacy and standards of language. But the new hierarchies engendered new tensions, as people struggled to retain old privileges under the new regime. Ethnic alliances took on a new importance in this struggle, and from the new ethnic awareness nationalist movements developed, ‘inventing’ nations where, in reality, they did not exist. In later work, Gellner (1973, 1983) reformulated this theory to take account of certain facts it could not explain. One of these had to do with the central role he had assigned to language: it would lead one to predict that nationalisms would not arise in the absence of a recognised national language, yet there were plenty of examples of that happening, for instance in the Arabic-speaking world and Hispanophone Latin America (as well as the English-speaking world, where separate American, Canadian etc. sub-varieties are recognised, but not as distinct languages). Moreover, relatively stable nations had formed around a multiplicity of languages, as in the case of Switzerland. Gellner therefore shifted the focus away from language, and ever more onto the institutional structure of the public education system and its role in defining and maintaining a culture within which nationalism as a political principle is embedded and enacted in a wide range of ways. Gellner’s second major dissent was from Kedourie’s Kantian conception of the nation as being modelled upon the Romantic ideal of the Individual. For Gellner, the nation is social in its constitution from top to bottom. In support of this he famously invoked Renan’s (1882, p. 27) view that ‘The existence of

130 John E. Joseph a nation is – pardon my metaphor – a daily plebiscite [. . .]’,2 as well as his description of the mental constitution of nations as being based not simply on shared memories, as was commonly assumed, but on shared forgetting, the putting aside of differences among groups constituting the nation, while also ceasing to remember that there was a time when they were not united as a nation (see next section). There is a certain irony in Renan’s now being so widely remembered for these rather modernist-sounding statements, given that he was one of those nineteenth-century linguistic thinkers who developed the essentialist view of language to its highest point. In his celebrated early work on the origin of language, Renan followed the German Romantic view, memorably articulated by Humboldt (see Joseph, 2004, p. 45) that the structure of languages must have been already fully fixed at the moment of their creation (Renan, 1858, pp. 105–6). Primitive man, Renan believed, created language effortlessly, like the child that he was (ibid., p. 98), not by employing his will, but by letting language flow spontaneously and naturally from the structure of his faculties, physical and mental (ibid., pp. 92–3). In general Renan adheres closely to the views of Herder, but he rejects Herder’s view that reflection was the key to the origin of language, and returns instead to something like the Epicurean idea of the language proceeding from the body – more precisely, from the ethnic body (see Joseph, 2004, p. 43). Like Fichte and Humboldt, Renan believed that ‘The mind of each people is in the closest connection with its language [. . .]’ (Renan, 1858, p. 190).3

Anderson’s ‘imagined communities’ and Billig’s ‘banal nationalism’ The conjunction of Renan and Gellner would figure very prominently in Benedict Anderson’s influential defintion of the nation as an ‘imagined political community’: It is imagined because the members of even the smallest nation will never know most of their fellow-members, meet them, or even hear of them, yet in the minds of each lives the image of their communion. Renan referred to this imagining in his suavely back-handed way when he wrote that ‘Or l’essence d’une nation est que tous les individus aient beaucoup de choses en commun, et aussi que tous aient oublié bien des choses’. With a certain ferocity Gellner makes a comparable point when he rules that ‘Nationalism is not the awakening of nations to self-consciousness: it invents nations where they do not exist’. (Anderson, 1991, p. 6)4 As with the ‘discovery’ of a national language, a crucial part of this inventing or imagining of a nation is the creation of a belief that the nation has not been invented. Its invention must, in other words, be forgotten. For if invented,

The construction of nationalism 131 the nation might be perceived as merely artificial, arbitrary, contingent in character, thus making its validity seem very shallow indeed. Instead the myth must be made that the nation is a natural entity, with a deep-rooted authenticity that is being rediscovered. If the nation in question has not existed as a nation during the whole of recorded history, then the myth (or more usually, the complex of myths) will be extended back into pre-history as far as needed to establish its claim to legitimacy. Anderson goes on to explain that the nation [. . .] is imagined as a community, because, regardless of the actual inequality and exploitation that may prevail in each, the nation is always conceived as a deep, horizontal comradeship. Ultimately it is this fraternity that makes it possible, over the past two centuries, for so many millions of people, not so much to kill, as willingly to die for such limited imaginings. (Ibid., p. 7) Both of the primary organisational structures which preceded the modern conception of the nation, the religious community and the dynastic realm, were ‘vertical’ rather than ‘horizontal’ in arrangement. Authority flowed downward from God to the supreme human authority, whether religious or secular, and outward from there to the rest of the community. A hallmark of modern thought was that these vertical hierarchies came to be seen as mythical, serving the interests of those at the top and oppressing those at the bottom. And so they have come to be replaced, in part, by the horizontal nation of which every citizen is, in a sense, an equal member. The fact of their inhabiting a contiguous territory becomes essential, overriding differences of religion, culture, class and so on. Yet how then to motivate people to fight to the death on behalf of the nation – often against other members of their own religion, for example? This is why the new mythologies were required. Relying heavily on Seton-Watson’s (1977) account of nationalism as being built upon linguistic difference, Anderson ascribes the construction of national myths, starting in the Renaissance, to a shift from the idea that a particular script–language offered privileged access to ontological truth, precisely because it was an inseparable part of that truth. [. . .T]he search was on, so to speak, for a new way of linking fraternity, power and time meaningfully together. Nothing perhaps more precipitated this search, nor made it more fruitful, than print-capitalism, which made it possible for rapidly growing numbers of people to think about themselves, and to relate themselves to others, in profoundly new ways. (Anderson, 1991, p. 36) These profoundly new self-conceptions found a ready-made template from which to work: the national languages, which, Anderson believes, emerged in

132 John E. Joseph the sixteenth century ‘as a gradual, unselfconscious, pragmatic, not to say haphazard development’ (ibid., p. 42); ‘in their origins, the fixing of printlanguages and the differentiation of status between them were largely unselfconscious processes’ (ibid., p. 45). The accuracy of these views will be brought into question in the next section. Nationality is not necessarily the identity people will be most willing to die for. Regional and local identities matter, as do social class identities, racial, religious and sectarian identities. Even linguistic identity can be an end in itself, though it tends to get transmuted into quasi-racial terms. Given their importance in defining who individuals believe they really are, one would expect identities to be founded in every case on an extremely profound basis, such as whole libraries of texts recording thousands of years of cultural tradition. This has usually been the case with the older organisational structures of religious communities and dynastic realms, but modern structures such as the nation are typically founded on much shallower, often purely symbolic grounds. Anthony Smith in particular has emphasised how much of the effort of nationalism construction is aimed at reaching back to the past in the interest of ‘ethno-symbolism’ (see e.g. Smith, 1998, chapter 8). Michael Billig, who collaborated with Henri Tajfel on early work establishing the modern understanding of identity within social psychology, has made a number of crucial expansions upon Anderson’s position. The term ‘imagined community’ might suggest that the nation ‘depend[s] upon continual acts of imagination for its existence’ (Billig, 1995, p. 70). Instead it is the case that the original ‘imagining’ is reproduced – Billig takes the term from Bourdieu (1991) – sometimes through purposeful deployment of national symbols, but mostly through daily habits of which we are only dimly aware or even unaware. Examples include the national flag hanging in front of the post office, or the national symbols on the coins and banknotes we use each day. Billig introduced the term banal nationalism to cover the ideological habits which enable the established nations of the West to be reproduced. It is argued that these habits are not removed from everyday life, as some observers have supposed. Daily, the nation is indicated, or ‘flagged’, in the lives of its citizenry. Nationalism, far from being an intermittent mood in established nations, is the endemic condition. (Billig, 1995, p. 6) This idea was perhaps implicit in Anderson’s invocation of Renan on the necessity of ‘forgetting’, but by not drawing out the implications, Anderson leads readers to associate nationalism strictly with what Billig calls the ‘passionately waved flag’, and to ignore ‘routine flags’, like the one hanging limp in front of the post office, which operates to reproduce banal nationalism precisely because it is a ‘forgotten reminder’ (ibid., p. 8) – its significance is ‘forgotten’ by the observer, yet remembered in the depths of his mind. Billig’s

The construction of nationalism 133 point is that studies of nationalism have perversely paid attention to the strongly asserted nationalism that is typical only of a small minority of people, and ignored the banal nationalism that is part of everyone’s everyday life (strong nationalists included). What is more, he argues that this is part of an ideological pattern in which ‘our’ nationalism (that of established nations [. . .]) is forgotten: it ceases to appear as nationalism, disappearing into the ‘natural’ environment of ‘societies’. At the same time, nationalism is defined as something dangerously emotional and irrational: it is conceived as a problem, or a condition, which is surplus to the world of nations. The irrationality of nationalism is projected on to ‘others’. (Ibid., p. 38) In his view, which owes more to Bourdieu than to Tajfel, ‘an identity is to be found in the embodied habits of social life’ (ibid., p. 8), including, as we shall see in the next section, language. One further aspect of national identity that will be highlighted in this paper is not explored in any depth by Billig, although he points toward it by quoting Said’s (1983) insistence that nations are ‘interpretive communities’ (borrowing Fish’s concept; see e.g. Fish 1980) as well as imagined ones, because what has to be created is not only a concept of the nation but an entire history, based on a particular interpretation of recorded events. In fact, as pointed out in Joseph (2004, chapter 1), identities are not simply a matter of what their possessors (or would-be possessors) project, but of how such projections are received and interpreted. As a group of sociologists have asserted, National identities are not essentially fixed or given but depend critically on the claims which people make in different contexts and at different times. The processes of identity rest not simply on the claims made but on how such claims are received, that is validated or rejected by significant others. (Bechhofer et al., 1999, p. 515) Nor, I would add, can we neglect the identities that others project onto us. However, it is important to note that, of all the claims which people make and receive about a national identity, none is more crucial or powerful than the claim that the identity is in fact fixed and given, is imposed on us by birth and does remain essentially unchanged thereafter. From the constructionist point of view, the mistake of an essentialist analysis lies in its failure to see past the myth that is embedded within the identity under study. At the same time, constructionists must take care to avoid a potential mistake of their own, by dismissing the ‘myth’ as mere fallacy, and therefore unworthy of analytical attention in its own right. It is a cultural construct that ultimately cannot be separated from the national identity as a whole.

134 John E. Joseph

De-essentialising the role of language: Hobsbawm and Silverstein Although several years older than either Kedourie or Gellner, Eric Hobsbawm (b. 1917) turned his attention to nationalism some two decades after they had set the parameters of the current discourse. For Hobsbawm, a committed Marxist until 1991, the discourse of nationalism, including the prominent role assigned to a national language, encodes other, deeper concerns, and it is a mistake simply to take the discourse at face value. No one disputes that when the modern concept of nation began to be asserted in the late eighteenth century, it was for political reasons, but when justifications were put forward on the basis of a people’s natural right to self-determination, it was never just from hostile foreign powers that autonomy was being proclaimed, but at least as much from the ruling class of the people’s own country. [W]hat characterized the nation-people as seen from below was precisely that it represented the common interest against particular interests, the common good against privilege, as indeed is suggested by the term Americans used before 1800 to indicate nationhood, while avoiding the word itself. Ethnic group differences were from this revolutionarydemocratic point of view as secondary as they later seemed to socialists. Patently what distinguished the American colonists from King George and his supporters was neither language nor ethnicity, and conversely, the French Republic saw no difficulty in electing the Anglo-American Thomas Paine to its National Convention. We cannot therefore read into the revolutionary ‘nation’ anything like the later nationalist programme of establishing nation-states for bodies defined in terms of the criteria so hotly debated by the nineteenthcentury theorists, such as ethnicity, common language, religion, territory and common historical memories [. . .]. (Hobsbawm, 1990, p. 20) As for national languages, Hobsbawm agrees with earlier students of nationalism, culminating with Anderson, on their central importance within the discourse. But whereas Anderson takes the national language as a given, furnishing the foundation on which the rest of national identity can be constructed, Hobsbawm insists that the national language is itself a discursive construction. National languages [. . .] are the opposite of what nationalist mythology supposes them to be, namely the primordial foundations of national culture and the matrices of the national mind. They are usually attempts to devise a standardized idiom out of a multiplicity of actually spoken idioms, which are downgraded to dialects [. . .]. (Ibid., p. 51)

The construction of nationalism 135 No one who has studied the history of any national or standard language (unless for partisan purposes) has come up with a different conclusion than this one. But historians of nationalism had not in general attended to the work of linguistic historians in the way Hobsbawm did; and as for the linguistic historians themselves, they were seldom aware of the broader implications of their own findings. Indeed, I do not think any linguist has ever provided quite as apt and succinct definition of the standard language as Hobsbawm does: ‘a sort of platonic idea of the language, existing behind and above all its variants and imperfect versions’ (ibid., p. 57). A ‘mystical identification of nationality’ then occurs with this idea of the language, an identification Hobsbawm believes is much more characteristic of the ideological construction of nationalist intellectuals, of whom Herder is the prophet, than of the actual grassroots users of the idiom. It is a literary and not an existential concept. (Ibid., p. 57) Here I cannot fully agree: for while it may be historically true that the national/standard language is the property of nationalist intellectuals rather than of ordinary users during the period when it is initially being constructed, this ceases to be the case once it enters the educational sphere, and once education is widespread. The linguistic ideology then becomes common national property, as least as likely to find firm belief among the working classes who do not control it as among the upper classes who do. Indeed, in a subsequent chapter of his book Hobsbawm will place great stress on the fact that enthusiasm for linguistic nationalism has historically been a phenomenon of the lower middle class. The classes which stood or fell by the official use of the written vernacular were the socially modest but educated middle strata, which included those who acquired lower middle-class status precisely by virtue of occupying non-manual jobs that required schooling. (Ibid., p. 117) These are also the people who become the mainstay of nationalism – not just by active flag-waving on symbolic occasions, but daily in the banal ways pointed to by Billig, including their use of ‘proper language’ and their insistence on its norms, for instance in conversation with their own children. Hobsbawm believes that ‘national identity’ in the sense we usually think of it really goes back to Victorian shopkeepers and clerks who envied the sort of class belonging enjoyed by both the upper classes, with their clubs and aristocratic titles, and the workers, who could locate their identity in socialism. If they already lived in a nation-state, nationalism gave them the social identity which proletarians got from their class movement. One might suggest that the self-definition of the lower middle classes – both that

136 John E. Joseph section which was helpless as artisans and small shop-keepers and social strata which were largely as novel as the workers, given the unprecedented expansion of higher education white-collar and professional occupations – was not so much as a class, but as the body of the most zealous and loyal, as well as the most ‘respectable’ sons and daughters of the fatherland. (Ibid., p. 122) In other words, although their real identity was that of a social class, they masked it for themselves and others in a nationalistic guise. And the mask was double-sided: in their obsession with ‘speaking properly’ as a mark of respectability, they were contributing to the linguistic construction of their nation. Gellner had already suggested that, even if it is true that nationalism began as an ideology at the start of the nineteenth century, something transformative occurred with the events of 1870–1871 and their aftermath. With Hobsbawm this later period becomes the truly cardinal one, as for the first time ideological notions about nation and language, heretofore restricted to intellectuals and the government elite, spread down through the general populace, eventually even reaching the working class. Hobsbawm points to one further development in this period that would have dramatic consequences. Prior to about 1880, the claims of a group of people to constitute a ‘nation’ would have been taken seriously only if their population met a certain unstated threshold. But from that time onwards, any body of people considering themselves a ‘nation’ claimed the right to self-determination [. . .I]n consequence of this multiplication of potential ‘unhistorical’ nations, ethnicity and language became the central, increasingly the decisive or even the only criterion of potential nationhood. (Ibid., p. 102) This seems to conflict with evidence we have seen of much earlier arguments using language to define the nation, notably by Fichte. What Hobsbawm leads us to consider, however, is the possibility that we read Fichte and others of his period through post-1880s lenses, finding implications that Fichte and his contemporaries would not have entertained, and that reflect the concerns of the later age that has effectively defined nationalism for us. Moreover, we perhaps overestimate the influence Fichte and his fellow intellectuals exercised on their countrymen, of whom, after all, a rather small proportion were actively engaged in these debates. One development that certainly altered the intellectual climate at the start of the modern period was the rise and dissemination of belief in human evolution, associated in particular with the name of Charles Darwin. Among the many effects that Darwin never could have foreseen was that evolution

The construction of nationalism 137 theory was used to construct a ‘scientific’ basis for belief in racial differences of an intellectual and moral order. As such ideas diffused into popular culture, they subtly and gradually made it seem more and more like ‘common sense’ that ethnic differences were fundamental in nature, so that it was naturally right for distinct nations to define distinct states. One of the problems, though, as Hobsbawm points out, is that ethnic differences are not always so easy to detect on the physical level, at least not reliably (see Hobsbawm, 1990, pp. 65–7). Where language differences correspond with ethnic differences, they appeared to provide a more objective basis on which to draw dividing lines – this despite the insistence of leading linguists that language did not have any direct historical link with ethnicity. Evidence was in fact readily available to anyone that the two do not have to go together, so long as they had ever encountered someone bilingual (and it is difficult to imagine that they might never have done so). But again, such was the desire to construct national difference that what appeared to help would be seized on, and anything contradictory ignored. Whether or not one is prepared to go as far as Hobsbawm does in locating class-based factors underlying linguistic nationalism, his work has had an undeniably salutary effect in counteracting Anderson’s a-prioristic approach to language within identity. Michael Silverstein has recently launched a similarly-spirited critique of Anderson’s use of ‘language in modeling the cultural phenomenology of nationalism’ (Silverstein, 2000, p. 85). His critique, which relies heavily on his somewhat idiosyncratic reading of the linguistic ideas of Whorf, culminates in an assertion that Anderson has mistaken discursive for ‘real’ linguistic nationalism: [. . .] Anderson seems to mistake the dialectically produced trope of ‘we’ness for the reality. He seems not to see that the dialectical workings of political processes that construct the sharable space of realist reportage in standardized language are the facts to be characterized and explained. (Silverstein, 2000, p. 126) The regime of language on which such a dialectic depends is a frequently fragile sociopolitical order, seething with contestation that emerges from actual plurilingualism, heteroglossia, and like indexes of at least potentially fundamental political economic conflict. Such a regime of language is, however, energized and in a sense maintained by the ritually emblematized trope of ‘we’-ness. It seems to have taken in Anderson, who buys the trope as a transparently imagined ‘reality’. (Ibid., pp. 128–9) Again, it would be hard not to agree with Silverstein’s critique insofar as it implies that Anderson takes language too much for granted. That is, in order to explain his major variable, the construction of national identity, Anderson uses national languages as though they were a constant – when in fact they

138 John E. Joseph are just as much variables, constructs, ‘imagined communities’ as the national identities they are invoked to explain. In other words, Anderson’s constructionist approach to nationalism is purchased at the price of an essentialist outlook on languages. It seems a bargain to the sociologist or political scientist, to whom it brings explanatory simplicity (not to mention ease). But to Silverstein, as to Hobsbawm, it is a false simplicity. National languages and identities arise in tandem, ‘dialectically’ if you like, in a complex process that ought to be our focus of interest and study. Silverstein goes a step further, however, to assert that the only ‘real’ facts are the ‘political processes’ and ‘political economic conflict’ which underlie the discourse through which the national/standard language is battled into existence. The ‘we’-ness on which the imagined national community is built is but one ‘trope’ produced out of this discourse. The fact that it is ‘ritually emblematized’ leads to the illusion that it is really real, when in truth it is only a figure of speech. What this means is that, contra Anderson, identity constructed in language is not the true locus of nationalism. Where nationalism really exists is in politics and economics, and what we see in language is only the reflection of that real nationalism. Anderson, in effect, has mistaken the image in the mirror for the thing reflected. Hobsbawm does not go this far. On the contrary, he is alert to the danger of ‘reduc[ing] linguistic nationalism to a question of jobs, as vulgar-materialist liberals used to reduce wars to a question of the profits of armaments firms’ (Hobsbawm, 1990, pp. 117–18). Silverstein, in contrast, is close to a vulgarmaterialist reduction when he insists that ideologies of language are merely a reflection of what is real, and have no reality in themselves. In doing so, he perpetuates the very error he has already criticised another aspect of in Anderson, namely an overly strong distinction between linguistic and political ‘reality’. Anderson recognises that the two are functionally intertwined, but treats them as fundamentally different in their internal nature, language being a coherent given, and political identity a construct. Silverstein recognises that their internal natures are more alike than Anderson assumes, but he denies that they are functionally intertwined except in the relatively trivial sense that the one reflects the other. Here I believe Anderson has it right. Silverstein’s mistake, to borrow his own terms from the first quote earlier, is to assume that what he calls ‘we’-ness is a ‘dialectically produced trope’ rather than part of ‘the dialectical workings of political processes’ themselves. For this assumption demands a strict and transparent division between what is ‘in language’ on the one hand and what is ‘political’ on the other. In the absence of such a division – and in my view that division could only be illusory – Silverstein’s relegation of ‘we’-ness to the category of mere ‘trope’, on which this part of his critique of Anderson depends, is nothing more than an axiomatic declaration, and an unjustified one. This ‘we’-ness, and the national identities and imagined communities founded upon it, are neither more nor less real than ‘the dialectical workings of political processes’ or ‘political economic conflict’, because they are in fact an inseparable part of them.

The construction of nationalism 139 Comments by Silverstein elsewhere in the chapter lead me to suspect that he might wish to invoke here a principled difference between ‘standard’ languages that involve political construction in the way I have suggested, and ‘non-standard’ languages or dialects, which have not been politically constituted in the same way. Although I myself accepted the existence of this difference when setting out to problematise it in Joseph (1987), I have ultimately ceased to be convinced that any language or dialect, standard or non-standard, can be constituted by anything other than a form of the same political processes (see Joseph, 2000). But even if one accepts the distinction, the ‘we’-ness of which Anderson and Silverstein write is clearly and unambiguously a matter of political construction. The fact that it overlaps with a first-person plural pronoun shared by non-standard dialects does not somehow remove it from the political sphere, either by making it ‘natural’ or by making it ‘metaphorical’. It does, as Hobsbawm and Silverstein correctly recognise, contribute to an essentialising of national identity. As I argue in Joseph (2004, chapter 4), that essentialism is – like Harris’s ‘language myth’ – an important fact that we need to analyse and explain, hopefully without letting it seep into our explanation. Insofar as Anderson’s treatment of language in a quasi-essentialist vein opens the way for such seepage, Silverstein has made a useful contribution to stopping it.

Conclusion There is a general tendency for groups of people who occupy contiguous territory and see themselves as having common interests to develop, over long stretches of time, ways of speaking that are distinctive to them, marking them out from groups who either are not geographically adjacent to them or else are perceived as having different, probably rival interests. In other words, language tends to mark out the social features on which national belonging will come to be based – but it is only a tendency, because it also happens very frequently that the same way of speaking is shared by people with very different interests (religious ones, for instance), and that markedly different ways of speaking exist among a group of people who nonetheless see themselves as part of the same nation. The ideology of national unity has favoured (particularly in Kohn’s ‘organic’ nationalism, but not exclusively there) a view that nations are real because those within them share a deep cultural unity, and this has co-existed with a widespread – indeed, nearly universal – belief that deep cultural unity is the product of a shared language. This is what Fichte meant by the ‘invisible bonds’ by which nature has joined those who speak the same language, and it is one of the insights behind the Sapir–Whorf hypothesis. Again, it cannot be more than a tendency, since it is certainly not the case that people who identify themselves as belonging to the same culture or nation think identically. Political disputes exist within every polity, and civil wars have riven many nations. Yet language is central to the habitus, to use Bourdieu’s

140 John E. Joseph term: the fact that we spend our formative years attending so long and hard to the task of learning words and their meanings from those around us produces the result that we acquire tastes, habits, ways of thinking from them that will endure into adult life without our remembering their provenance. The language does not somehow transmit culture to its speakers – rather, the language is the text through the constant interaction with which the older speakers transmit culture to the younger. At the same time, it must be remembered that the young will want an identity of their own, and will attain it first of all by resisting the imposition of culture upon them by their elders. Culture is thus in a continuous process of transformation, sometimes more blatantly than at other times, and despite the huge investment members of certain cultures make into assuring continuity (an investment which certainly can slow down transformation). The concepts of the ‘national’ and ‘standard’ language are precisely examples of such investment, seen as essential to the creation, maintenance and transmission of national and cultural unity. When transformation occurs, it takes the form of new ‘readings’ of the handed-down text that is the language, or by the substitution of one such text for another, not by the creation of a wholly new language. In addition to being the ‘text’ of cultural transmission, the language is the principal medium in which texts of national identity in the more usual sense will be constructed. It is not the only such medium, nor the only powerful one, as Billig’s exposition of ‘banal nationalism’ has shown. But the particular concepts which constitute a national identity correspond to words in the national language, embodied in ‘sacred texts’ of the nation such as a constitution or key works of the national literature, including the national anthem. Again, a postRomantic belief in a deep-seated link between the national language and the national literature – embodied in a widespread view that literary (especially poetic) language cannot be translated (see Joseph, 1998) – further reinforces the view that the language is absolutely essential to what the nation is. As universal education is adopted throughout the nation, inculcation of the standards of ‘correct’ language assumes a central role, overtly for reasons of a duty to maintain the culture. However, as Hobsbawm has shown, such is the force of the language-culture-nation-class nexus that, especially for the upwardly mobile members of the lower middle class, being a ‘proper’ citizen and member of the community is inseparable from using ‘proper’ language. Insofar as nations are not the historical essences they purport to be, but are constructs which inevitably involve arbitrary and even capricious divisions and classifications, when a nation wants to control who can live in it, vote in it and enjoy state benefits, language can appear to be the most obvious test for deciding whether particular individuals belong to the nation or not. Most nations no longer have laws based upon ‘racial’ classification – which are rarely easy to apply in any case – yet many do require cultural qualifications to be met, which are likely to include language either overtly or indirectly.

The construction of nationalism 141 Each of these factors has reinforced the others in giving the national language the force of a cultural-historical ‘ethno-symbolic’ myth as suggested by Smith (see p. 132 earlier). Within each, too, there is a contradiction or a caveat that has periodically pendulum-swung to prominence, such that the loss of belief in the national language and all it stands for is always there in potentia, and must come to the fore at least periodically. The question that is unanswerable for now is whether national languages, together with the nations to which they are attached, represent a historical phase that is now on a course of decline heading for eventual disappearance, to be replaced with ‘glocalisation’ – a term that has been coined to denote the combination of globalisation with the resurgence of local, sub-national sites of belonging – or whether they are an invention that has proven too useful for human social organisation to be given up.

Notes 1 ‘Une nation est une âme, un principe spirituel. Deux choses qui, à vrai dire, n’en font qu’une, constituent cette âme, ce principe spirituel. L’une est dans le passé, l’autre dans le présent. L’une est la possession en commun d’un riche legs de souvenirs; l’autre est le consentement actuel, le désir de vivre ensemble, la volonté de continuer à faire valoir l’héritage qu’on a reçu indivis.’ 2 ‘L’existence d’une nation est (pardonnez-moi cette métaphore) un plébiscite de tous les jours.’ 3 ‘L’esprit de chaque peuple et sa langue sont dans la plus étroite connexité [. . .].’ 4 The quotations are from Renan (1882, p. 9) and Gellner (1964, p. 169). The Renan quote may be translated: ‘The essence of a nation is that all the individuals have many things in common, and also that they have all forgotten many things’.

9

How to make history with words Roy Harris

Oscar Wilde once remarked: ‘The one duty we owe to history is to rewrite it.’ That epigram tacitly acknowledges the uneasy relationship between two ways of making history. The more obvious consists in actually doing something, or taking part in events, then or later deemed worthy of public remembrance, such as fighting in the First World War, or inventing the internal combustion engine, or being the first human being to set foot on the moon. This kind of history-making makes history of the kind the Romans called res gestae. But history is also made by the persons and processes involved in recording memorable events for posterity. The people who specialize in making history in this second way are commonly known in Western culture as ‘historians’. Historians do not necessarily witness personally or participate in the events they give an account of. Occasionally, exceptional individuals can claim to be history-makers on both counts. Julius Caesar would be an example. He was not only the first Roman general to conquer Gaul, but wrote an authoritative account of this achievement in his Gallic Wars. What exactly is the connexion between these two kinds of history-making? This is generally reckoned to be a question that falls in the domain of philosophy of history. I wish to argue that integrationism has a serious contribution to make towards philosophy of history. This may seem a highly presumptuous goal. That is the more or less unanimous view of all the academic historians with whom I have broached the matter; and I have no reason to think that view would not be shared by the majority of their colleagues. Most historians seem fairly content with the way their discipline conducts itself, and see no reason to change it, in spite of mavericks who occasionally announce the end of history, or deny that the Gulf War ever happened. History is so well entrenched as an academic subject that it can afford to shrug off alarmist pronouncements of this kind. Historians I have encountered show little inclination to think they have anything to learn from linguists at all, let alone an esoteric group called ‘integrationists’, whom they have never heard of. Yet, oddly enough, in recent years there has been much talk by historians of a ‘linguistic turn’ in their own discipline. We find papers with titles like ‘The use of language in the writing of history’ (Ankersmit 1989). The expression linguistic turn is itself borrowed

How to make history with words 143 from one of history’s subdisciplines, namely the history of philosophy. But when a linguist looks for evidence of this new development in the second half of the twentieth century, the historians’ ‘linguistic turn’ turns out to look in retrospect like no more than a slight wiggle in an otherwise perfectly straight road. We find such curious claims as the following: ‘The language of history permits the historian to see the past from a perspective different from that of the historical agents themselves’ (Ankersmit 1994: 57). But what it is about language that makes this possible is never explained: nor why the difference does not reduce to the fact that the world has moved on in the time between the historian’s writing and the events reported. Similarly in the late 60s we find J.H. Hexter discussing the ‘rhetoric of history’ as if he had been the first to notice that the language of historians is not the language of physicists. I have found one philosopher of history, Franklin Ankersmit, who claims that the philosophy of history is ‘of fundamental importance for linguistics’ (Ankersmit 1981: 7). He seems less inclined to admit the reverse – that linguistics might be of fundamental importance for philosophy of history. In fact, although his book Narrative Logic has the subtitle A Semantic Analysis of the Historian’s Language, it seems to be almost exclusively concerned with the ways historians use metaphor. Ankersmit goes so far as to dismiss a linguistic approach to discourse analysis as incapable of throwing any light on what he calls, somewhat mysteriously, the ‘fundamental secrets of the narratio’ (Ankersmit 1981: 213). In view of his obvious reluctance to engage with relevant work in linguistics, Ankersmit’s position cannot fail to strike a linguist as awkward, to say the least. This is hardly a ‘linguistic turn’ by historians but the merest nod in the direction of a linguistic problem. The problem itself is never tackled; at least, not in any way that would satisfy an integrationist. ***

Yet there is a problem, which for centuries Western historians managed to ignore; namely, that their contribution to human knowledge was itself language-dependent. As far as I know, the first person to draw attention to it was Humboldt at the beginning of the nineteenth century. Humboldt argued that if the historian’s task is to present an account of ‘what actually happened’, then perhaps the greatest obstacle in the way of attaining that goal was language itself (Humboldt 1821: 58). What Humboldt argued, roughly, was later encapsulated by Wittgenstein in the famous dictum: ‘The limits of my language are the limits of my world.’ That was eminently true, Humboldt saw, of the world of the historian. But among historians no contemporary of Humboldt’s seems to have taken that point seriously. Now the notion of linguistic limits and its relevance to the historian can be interpreted in various ways. The crudest interpretation might be illustrated by the following science-fiction scenario. Let us imagine that in Byebyeland they speak a language in which there is no way of referring to what we call ‘the past’. The grammar of Byebyespeak has no past tense, its vocabulary has no words for

144 Roy Harris ‘yesterday’ or ‘last year’. In Byebyeland they have no calendars or clocks, no way of identifying dates. In short, there is no way of translating into Byebyespeak a sentence like King Harold was killed at the battle of Hastings in 1066. At this level, linguistic limits are clearly operative. The limits of Byebyespeak preclude the kind of work in which Livy, Bede, Gibbon, Ranke et al. engaged. Whatever other intellectual enterprises might flourish in Byebyeland, there would be no jobs for historians in its schools and universities. At least, not for historians of the kind we are familiar with. And if one day a missionary historian from the planet Earth alighted from a space-ship and set about introducing the scholars of Byebyeland to the discipline of history, that historian’s first task would have to be the reform of their language. Now it does not take an integrationist to point this out. But there is a more interesting interpretation according to which the limits of one’s language are not construed as constraints imposed by the tense system or the vocabulary of English, or Hopi, or whatever the language in question happens to be. Given any language you like, there are also limits imposed by what its users understand a language to be. In particular, there are limits on how they understand their language to be related to the non-linguistic world. It is here, it seems to me, that integrationist thinking can make a significant contribution to philosophy of history. Let me now try to outline, very briefly, how I think an integrationist philosophy of history would be radically different from some of the philosophies of history that are currently on offer. By way of illustration I am going to look at two accounts of what the purpose of philosophy of history is supposed to be. Both, as it happens, are by Oxford philosophers. Before I do that, I need to make an excursus into the question of what the word history means; because, as I shall try to show, some philosophies of history are themselves based on what, from an integrationist point of view, are highly dubious semantic premisses. I earlier distinguished between two modes of making history, and gave Julius Caesar as an example of a history-maker who made history in both modes. The distinction between those two modes, as I formulated it, must not be confused with the claim often made by historians and philosophers that the word history itself has two distinct meanings. To take a typical example, we are told: ‘The word “history” is used in two senses. It may be either the record of events or the events themselves’ (Shotwell 1910: 527). Thus, allegedly, in the former sense a history is what Gibbon wrote (his History of the Decline and Fall of the Roman Empire), and in the latter sense the history is what actually happened (which in Gibbon’s case would be events that occurred during the period his history covers). Similar claims continue to be repeated throughout the twentieth century. Here is a more recent example: The word history carries two meanings in common parlance. It refers both to what actually happened in the past and to the representation of the past in the work of historians. (Tosh 1984: vii)

How to make history with words 145 What worries an integrationist about such claims is, in the first place, that they are based on this reification of a linguistic entity called ‘the word history’. But leaving that worry on one side for the moment, there is another reason why the alleged ambiguity of this word should be treated with suspicion. The suspicion is that professionals have invented the ambiguity to serve their own metahistorical purposes. It is not lexicographers who are to blame here: dictionaries do list these two meanings. The trouble is that lexicographers are reflecting what historians and philosophers themselves claim, and these claims have to some extent affected general educated usage. If I am right, then, this appeal to semantic ambiguity is actually a case where philosophy of history borrows directly from its own coffers. Claiming that the word is ambiguous is a convenient way of justifying certain assumptions that the interested parties wish to take for granted. The most important of these concerns the validity of distancing the historian qua historian from the historical events. By appealing to semantics, it is made to look as if that relationship is somehow guaranteed in advance: it resides in the language itself, not in any theory about history that someone is putting forward. It is a way of taking that relationship out of contention, removing it in advance from the area of possible dispute. So it doesn’t have to be argued for. Second, it is also a convenient way of establishing the proprietary rights of the historian over a certain domain of discourse. This too does not have to be argued for if we can somehow build it in to the meaning of the word history itself. Thus history does double duty as the name of the historian’s business, and the name of the historian’s product. With these proprietary rights established, it becomes possible to exclude certain things from the realm of history because de facto they are not dealt with by historians; and likewise to deny that someone is a historian on the ground that what that writer deals with is not history. An interesting example of this was the attempt by some academics to marginalize the work of Toynbee by refusing to count him as a historian. Third, the supposed ambiguity conveniently allows two quite divergent accounts of the same events to exist side by side without either having to sacrifice its status as history. In other words, it still counts as history, not despite the fact historian A says one thing and historian B another; but because they do. In other spheres of intellectual activity, this is known as a licence to print your own money. Fourth, it makes room for a thesis to the effect that there are both an art of history and a science of history. This was a popular distinction in the nineteenth century. The ‘scientific historian’ was supposed to be exclusively concerned with establishing the truth about what happened. Consequently, it was said, ‘his uncoloured picture of the past will never rank in literature beside the splendid distortions which glow in the pages of a Michelet or Macaulay’ (Shotwell 1910: 528). In other words, the distinction between two types of historian provided an excuse for gross bias in history-making, although that is not how historians like to put it. But the validity of that distinction survived well into the twentieth century. It was still going strong, for example,

146 Roy Harris in A.L. Rowse’s book The Uses of History (1946), which has a chapter on ‘History as science and art’. The issue was brought to a head, according to Rowse, by Bury’s declaration in his inaugural lecture at Cambridge that: ‘History is a science, no less and no more.’ Against this, Rowse urges that ‘even in the realm of historical method, there is a non-scientific element that is just as important’ (Rowse 1946: 94). Now there may or may not be a case for allowing historians their personal prejudices. But it should not be justified by separating art from science; and nor should that cleavage be smuggled in under the smokescreen of the supposed ambiguity of the word history. In short, there is every reason to be wary of basing philosophy of history on dubious semantics. Let me now try to show in detail how that sponsors philosophies of history that an integrationist would not wish to accept, and then how integrationism supplies a theoretical basis for the rejection of such accounts. ***

My first example is an Oxford philosopher of history, Patrick Gardiner, who writes as follows: We may distinguish between questions asked within history and questions asked about it. Historians answer the first sort of question, philosophers answer the second sort. Thus an historian may try to answer the question: ‘Was there a connexion between the Protestant reformation of the sixteenth and seventeenth centuries, and the emergence of a capitalist economy?’ A philosopher, on the other hand, would not be interested in the truth or falsehood of the answer given by the historian to this question, but he would be very interested in the kind of evidence an historian might bring forward to substantiate his claim, in the criteria he uses for deciding whether or not a connexion existed. (Gardiner 1952: x) An integrationist would find this approach unacceptable for several reasons. First of all, the division of labour between historian and philosopher is stated in such a way as to beg in advance the kind of question that an integrationist would be interested in investigating. Gardiner seems to assume that historian and philosopher occupy fixed positions in a community of intellectuals who share a common language. This is a language in which such expressions as Protestant Reformation and capitalist economy have agreed meanings. In short, the very terms in which the explanation of philosophy of history is couched presuppose acceptance of what, for integrationists, is some version of the ‘language myth’. This is not to say that Gardiner rules out the possibility that there could be dissent in the community of scholars as to whether all capitalist economies have this or that feature, or when exactly the Protestant Reformation started. On the contrary, Gardiner’s hypothetical philosopher

How to make history with words 147 might well expect to unearth such discrepancies and to see them as bearing on the way different historians answer the question. But that is not the point. The point is, first, that ‘the question’ in question is identified ab initio in linguistic terms which are assumed to be comprehensible to philosopher and historian alike, not to mention their readers. (Otherwise it would be a bizarre example to take as the starting point for an explanation of how the historian’s interests differ from the philosopher’s.) But also, second, and no less important, the philosopher’s investigation is already conceived of as a linguistic investigation. The philosopher attempts to throw light on the relationship between (a) what the historian says and (b) what grounds the historian has for saying it. In short, typical historical discourse is assumed to be based on a relationship between words and facts. Now whether, or to what extent, that may be the case is precisely one of the open questions an integrationist philosophy of history would seek to address. But it cannot be properly addressed if the philosopher is already taking it as a foregone conclusion. In other words, Gardiner does not seem to see that, in trying to explain what philosophy of history is in general, he is actually taking up one particular position within philosophy of history. That is to say, he is confusing his subject with its metatheory. He cannot see the wood for the trees. Now how does this come about? It seems to me that here we have a case where the philosopher of history is misled by his own philosophy of language; that itself provides an interesting gloss on the dictum ‘the limits of my language are the limits of my world’. The source of Gardiner’s mistake is that he is already taking it for granted that the word history has two meanings. How do we know that? He says so. He refers to ‘the ambiguity of the word “history” ’, and what he understands this ambiguity to consist in he explains as follows: when we use the word we may, on the one hand, be taken to be referring to the past events, activities, thoughts, and so forth about which historians write – when, for example, we say that Jones’s researches are leading to the discovery of much of the history of the last few years, or when we remark that history repeats itself: on the other hand we may be taken as referring to the books, pamphlets, or lectures historians produce to describe what has happened – when we state, for instance, that Smith is reading history. (Gardiner 1952: ix) From an integrationist point of view, this passage is very revealing because it lays bare, albeit unwittingly, Gardiner’s underlying philosophy of language. It is evident that the writer of this passage has no qualms about endorsing what integrationists would call a surrogational semantics. That is to say, meaning is explicated in terms of reference to items in the nonlinguistic world. And that throws light on the puzzle we noted earlier of why it should

148 Roy Harris be assumed that the main thing the philosopher is concerned with is the historian’s evidence. At the same time, it is interesting to note how Gardiner condemns himself out of his own mouth. His example of the second meaning of the word history is saying that Smith is reading history. As an Oxford tutor, Gardiner seems curiously oblivious to the fact that all Oxford students are said to ‘read’ the subject they are studying. One ‘reads’ chemistry in the laboratory, not just in the library: it involves doing experiments, not just going through books by distinguished chemists. Analogously, Smith’s reading history might well include examining ancient artifacts in a museum, or visiting the site of a battle fought centuries ago. In other words, the object of study is not just the output of historians; and it would be absurd for the student to treat it as such. Thus, ironically, the very example Gardiner chooses to illustrate the ambiguity of the word history demonstrates the opposite: namely, that these two meanings cannot be kept in watertight academic compartments. And that, for an integrationist, throws wide open once again the issue of how to approach philosophy of history. Once we see that the distinction between questions asked ‘within history’ and questions asked ‘about history’ is itself predicated on the assumption that the word history is ambiguous, we see that it is circular to try to define the role of the philosopher of history in this way, because that amounts to endorsing in advance a particular theory of what history is. I have one further point to make about Gardiner, but in order to dispel the possible impression that I am pursuing some personal vendetta here, let me turn to an even more distinguished philosopher of history, R.G. Collingwood. In his book The Idea of History (1946) Collingwood sets out a quite different account of philosophy of history. According to Collingwood, the question the philosopher is interested in is ‘How do historians know?’. Gardiner, as we have seen, leaves it open whether historians know anything at all. Collingwood, on the contrary, takes it for granted from the start that historians do acquire knowledge of the past, knowledge of events they could not possibly have experienced themselves. Collingwood writes: How do historians know? How do they come to apprehend the past? Conversely, it is the historian’s business, not the philosopher’s, to apprehend the past as a thing in itself, to say for example that so many years ago such-and-such events actually happened. The philosopher is concerned with these events not as things in themselves but as things known to the historian, and to ask, not what kind of events they were and when and where they took place, but what it is about them that makes it possible for historians to know them. (Collingwood 1946: 3) Collingwood’s reference to the past as a ‘thing in itself’ has an unmistakably Kantian ring to it. But the point I wish to draw attention to is a rather

How to make history with words 149 different one. The passage I am quoting from Collingwood continues: Thus the philosopher has to think about the historian’s mind, but in so doing he is not duplicating the work of the psychologist, for to him the historian’s thought is not a complex of mental phenomena but a system of knowledge. He also thinks about the past, but not in such a way as to duplicate the work of the historian: for the past, to him, is not a series of events but a system of things known. (Collingwood 1946: 3) Here we have no nonsense about the word history having two meanings. But Collingwood is clearly committed to a semantic hypothesis of his own. For him, the word history designates a particular form of intellectual inquiry; or, as he likes to put it, a ‘special form of thought’. And that is why it is of interest to the philosopher. It is the same semantic hypothesis that allows Collingwood to declare roundly that ‘history did not exist’ in the days when Sumerian civilization flourished, even though the Sumerians were clearly aware of their own past, kept records, and so on. For Collingwood, there is no way the word history means simply ‘what happened’. That would be to misconceive altogether the task facing a philosophy of history. ***

How, you may ask, would an integrationist philosophy of history differ fundamentally from either of the two approaches I have discussed briefly here? First, let me stress that this is not an argument about the meaning of just one or two key terms, such as history and historian. It involves the whole question of how language is conceived as relating to whatever is deemed to lie outside language; how words relate to the world in general. The integrationist lesson for the philosopher of history is fundamentally the same as the integrationist lesson for the linguist. First ask yourself what linguistic presuppositions are embedded in the way you formulate your inquiry; because until you have addressed that initial question, you do not understand what kind of inquiry you are embarking on. It is no use plucking decontextualized questions out of thin air and setting up the task of answering them as the goal of a discipline. The favourite starting point for philosophers of history is the question ‘What is history?’. And it is treated as a kind of Socratic question, on a par with ‘What is justice?’ or ‘What is love?’. As if such essentialist questions had some timeless meaning, so that the search for answers, even if unsuccessful, had a permanent value in itself. What is wrong with that approach is that questions never occur except with the framework of a discourse. They do not have timeless meanings. If we wish to make sense of them, we have to treat them as elements integrated into some wider pattern of activities. And the case of history is no exception to that.

150 Roy Harris Let me go back for a moment to what is perhaps, from an integrationist point of view, the most astonishing of Gardiner’s contentions: according to him, the philosopher is not interested in the truth or falsehood of the historian’s claims. This is rather like saying that a philosophy of witchcraft is not concerned with whether or not witchcraft works, but merely with why witches are convinced that it does. But, more disastrously, it is in effect a refusal to consider the communicational dimension of history. In an integrationist philosophy of history, it is impossible not to be interested in the truth or falsehood of what the historian claims, if the historian in question is. For truth is a metalinguistic concept and, as such, intimately bound up with all the other assumptions that pervade the ways historians use linguistic resources to establish the validity of the particular communicational enterprise in which they are engaged. At the centre of the integrationist contribution, as I see it, is the following thesis: all Western theories of history, from Greek antiquity onwards, are based on theories of communication – specifically, on theories of language. And this is necessarily so, because the Western concept of history is the product of a certain tradition of literacy. It depends upon taking a certain view, endemic in that tradition, of the relationship between writing and speech, and of the relationship of both to whatever lies outside language altogether. An integrationist approach to history has many implications. It affects not only the justification of the tools the historian uses, not only the so-called questions of ‘historical method’, not only our assessment of what historians have tried to do in the past, but, more importantly, the ways in which we do and should teach history in schools and universities today. For instance, an integrationist would point out that not the least of the paradoxes involved in calling Herodotus ‘the father of history’ is that one thing Herodotus never had was a theory of history. The people who had a theory of history were those who later looked back to Herodotus and gave him that title. In fact, the question ‘What is history?’, at least as it is asked today, could never have occurred to Herodotus at all. As Classical scholars constantly remind us, when Herodotus wrote in the fifth century BC the word historia did not even mean ‘history’, but something more like ‘inquiry’ or ‘investigation’ in general. Thus a modern translator renders the opening sentence of Herodotus’ text in the following terms: ‘Herodotus of Halicarnassus, his Researches are here set down . . .’ So it would not even have made sense to ask Herodotus (if per impossibile we could have asked him) ‘What makes your researches historical?’. For Herodotus and his generation, a question like ‘What is history?’ had no communicational function. There were no activities such a question might serve to integrate. That, in integrationist terms, is what it is to lack meaning. A fortiori, Herodotus had no theory of history. That realization is as good a starting point as any for an integrationist philosophy of history. We have to ask: under what circumstances, in what

How to make history with words 151 kind of context, does such a question become meaningful? The weakness of Western philosophy of history can be traced back to its failure to address that question. The failure is all the more conspicuous in that, once the question is posed as an integrationist poses it, the answer is readily available. The question that would have made no sense to Herodotus certainly made sense to Aristotle, born a hundred years later. Aristotle is the first thinker in the Western tradition of whom we can say with any certainty that he had a theory of history. And we can say it with certainty because he tells us himself what that theory is. For Aristotle the question ‘What is history?’ plays an integrational role in philosophical inquiry into the modes of discourse. It is not a question about the past. It is a question about the use of language. Aristotle is quite clear about this. However misguided we may think Aristotle’s answer to the question ‘What is history?’ is, that does not detract from its value as evidence for the integrationist thesis that Western theories of history are based on theories of language. On the contrary, perhaps because Aristotle’s answer now seems unacceptable, we see all the more clearly that his theory of history is no more than an extension of his theory of language. But no less important, from an integrationist point of view, is the context in which Aristotle raises the question. Aristotle left no treatise on history as such. But he did leave a very important treatise on poetry. It is, in fact, as far as we know, the first Poetics that any Greek philosopher had attempted. And it is in the context of this endeavour to say what poetry is that the question arises. According to Aristotle, one thing you can say with certainty about poetry is that it is not history. And the reasons Aristotle adduces for making this claim are linguistic reasons. Now Aristotle was a reocentric surrogationist, and it therefore comes as no surprise to find that Aristotle’s philosophy of history is exactly what we might expect a reocentric surrogationist to advance. It is a philosophy which depends on regarding reality as independently given, and the function of language as being essentially to make true statements about that reality possible. That, of course, is a view which many historians in the Western tradition subsequently adopted, either covertly or overtly. ***

When we come to consider more recent historical metatheory, the schools which seem to share – or ought to share – most common ground with integrationism are the ‘constructionists’ and the ‘narrativists’. Let me take the constructionists first. Integrationists will doubtless agree with constructionists that history is a construct created here-and-now by the historian. Integrationists might even endorse Croce’s dictum that ‘every true history is contemporary history’ (Croce 1960: 12). But constructionists such as Croce and Oakeshott (1933) based their position on sceptical philosophical arguments about the past (Meiland 1965), not on linguistic considerations. Integrationists do not have

152 Roy Harris to go along with the thesis that there are no ‘past events’. There is no linguistic reason why they should. As for the narrativists, I have already mentioned one member of this school, Ankersmit. Others are Mink (Historical Understanding, 1987) and Hayden White (Tropics of Discourse, 1978). Barthes and Lévi-Strauss are sometimes included as narrativists, because narrativism, or at least the version sometimes called ‘high narrativism’, has come to be associated with the thesis that history is a literary production. But that, it seems to me, confuses the issue. It is one thing to focus on the notion that history requires a story to be told, and ask what kind of story it is, how its telling differs from other (non-historical) kinds of story that might be told, etc. It is quite another thing to conclude that the historical story exists in some kind of non-referential vacuum, identified merely by its stylistic features. The integrationist can sympathize with the first of these positions, but hardly with the second. For that would deprive the story of any serious integrational function in public discourse. According to Hayden White, ‘history has no stipulatable subject matter uniquely its own; it is always written as part of a contest between contending poetic configurations of what the past might consist of’ (White 1978: 98). (It is important not to miss in White’s use of the term poetic the oblique allusion to Aristotle’s distinction). Mink, however, goes further and claims: ‘We cannot refer to events as such, but only to events under a description’ (Mink 1987: 199). More linguistic problems here. Is Mink claiming that events do not have names, only descriptions? If so, many linguists would wish to take issue with him. Or is he claiming, à la Russell, that names are disguised descriptions anyway? If so, many linguists would take issue with that too. One possibility is that Mink wants to claim that there is no straightforward way of ascertaining, for example, whether the battle of Hastings actually took place, because in the world of events there was no given item or set of items which constituted that battle. By referring to it as the battle of Hastings the historian is already invoking a tacit description, but failing to make clear what that description is supposed to include. Worse still, I suppose, is that if it should turn out that, as far as the historian is concerned, the expression battle of Hastings by definition included the event of Harold’s death, then the statement ‘King Harold was killed at the battle of Hastings’ would emerge semantically not as a historical statement but as a tautology. Again, we are confronted with linguistic issues; but, it seems to me, not clearly identified as such by the theorist, let alone resolved by proposing the adoption of a specific philosophy of language. Mink’s theory of history is not good theory because it is not derived from any clearly stated principles. Here integrationists could at least help Mink to see what the linguistic problem actually is. What is worrying Mink is a puzzle which arises from

How to make history with words 153 adopting a surrogational semantics. In his essay ‘Narrative form as a cognitive instrument’ he is bothered by the fact that we can have quite different descriptions of the same event, all of them true as far as they go. But then he asks: Under what description do we refer to the event that is supposed to sustain different descriptions? It seems that the ordinary use of the term ‘event’ presupposes both an already existing division of complex processes into further irreducible elements, and some standard description of each putative event; then, to say that there are different descriptions of the ‘same’ event is to say that they are selected from or inferred from that standard and preeminent description. (Mink 1987: 200) For an integrationist, it is clear that the whole historical conundrum and Mink’s analysis of it are generated by tacitly adopting the presuppositions of surrogational semantics. What Mink evidently fails to see is just that. He wants to have events structured in such a way that their components answer in an unambiguous fashion to the language we use to refer to them. Or at least to some ‘standard’ language. Then the world would be safe for historians. Where Mink goes astray, for the integrationist, is in supposing that there must be some unique and independently specifiable portion of what really happened which will act as the basis or anchor for all subsequent talk of the ‘same event’. Instead of asking ‘Under what description do we refer to the event that is supposed to sustain different descriptions?’ an integrationist would ask ‘Under what conditions are this description and that description contextualized as descriptions of the same event?’ In short, the question that needs to be addressed is not about hypothetical ‘standard’ descriptions but about the metalinguistic implications of the phrase same event. And those will depend on the specifics of that episode of communication. It is at this linguistic level that we get to grips with the mechanisms of history-making with words. They hinge, according to the integrationist, rather crucially on such matters as whether the telling is in speech or in writing, because these are two related but fundamentally different communicational enterprises. Where writing makes a difference is in the integrational functions it makes available. A literate community is aware of the possibility of communicating across generation after generation. Both past and future come within its communicational reach. The price to be paid, however, is that what is written in the present automatically becomes permanently answerable to what was written by one’s ancestors as well as to what may be written by one’s descendants, who in turn will probably have access to what one’s ancestors wrote as well.

154 Roy Harris Linguistic acts do not have a special temporal status of their own, which somehow puts them outside the sequentiality of the rest of our existence. That is what integrationists call the ‘principle of cotemporality’. Yesterday’s document survives today, but it cannot be rewritten: it can only be altered. Yesterday’s events never come round again for a re-run. Writing is, by virtue of the principle of cotemporality, encapsulated in time in a way that speech is not. A historical narrative is the kind of narrative wherein these temporal implications begin to cast their shadow over what is written. Historians typically refer back to earlier accounts, to sources: poets on the whole do not. Not, at least, until poetry starts to become contaminated by history, as it does when literate poets compose for a literate society. Virgil was such a poet: Homer was not. ***

One final question. Does this mean that the integrationist philosopher of history is in effect committed to the view that history-making is just a form of discourse, to be analysed, in the end, like any other form of discourse? I think not. The expression ‘just a form of discourse’ is not one I would rush to use. It seems somehow to imply approval of those theorists in antiquity who conveniently subsumed history under rhetoric. The possibility that the analysis of history-making will eventually be reduced to stylistics (which does, indeed, seem to be the view endorsed by at least one school in modern philosophy of history) is not a prospect for which I can muster much enthusiasm. For the integrationist, as I see it, there is no such thing as ‘just a form of discourse’, whether in speech or writing, or however else articulated. Discourse is not an activity added on to life as an optional extra. Communication is intrinsic to life, at least as the majority of human beings live it. All forms of communication have integrational functions in the life of the society that sponsors them: they articulate relations between individuals and groups of individuals that would otherwise be impossible. Collectively, they integrate a plurality of human beings into a community. The question for an integrationist philosopher of history is how those forms of discourse that are accounted ‘historical’, for whatever reason, are integrated into the culture of those societies that recognize them. The crucial idea that integrationism brings to philosophy of history is the idea that the kind of history that historians make (the kind of history that is traditionally made with words after the event, sometimes long after) involves, like language-making itself (of which it is indeed one form), engagement in a non-stop process of contextualization and recontextualization. The notion that at any point in time we can, if we wish, step back from, or outside, this process and identify the established facts of history is rather like a diachronic counterpart to Saussure’s conception of la langue, with its parallel assumption that at any given point in the ongoing temporal process we can in principle stop and draw up an inventory of the current

How to make history with words 155 linguistic facts. Both are historical illusions generated by deeply flawed philosophies of language.

Acknowledgements I am grateful to Robert Burns and Paul Hopper for directing me to relevant publications which would otherwise have escaped my attention, and to Paul Maylam for comments on an earlier draft.

10 Talking about what happened Talbot J. Taylor

In the context of a volume on ‘language and history’, certain questions seem to me worthy of our consideration. An obvious question is that which forms the topic of historical linguistics. Another is the question of the origin and evolution of human language. I am not going to touch on these questions here. Instead, I will begin with the avowedly simplistic premise that history is a matter of ‘what happened’. Of course, historians do not merely report or describe what happened. They also ● ● ●

●

●

explain why it happened, analyze the conditions that led to its happening, trace the relationships between what happened at one time and what happened at another, propose principles or forces which are said to direct the course of historical events, and more . . .

Still, keeping to the most general level, it seems fair to say that history is a matter of ‘what happened’. And, of course, this applies to both the ontological and the epistemological senses of the term ‘history’. In other words, in addressing questions of history, we are addressing questions about historical events themselves and also questions about what historians and others assert about historical events. At this point I ought to say something about the sorts of questions about ‘language and history’ that particularly attract my own attention. However, I admit straightaway that the points touched on in this chapter are already very well-known, both to linguists and historians. My intention here is simply to foreground certain themes in preparation for the main part of my discussion. First, it would seem appropriate to begin with a question that concerns the place of language in historical happenings. If we take history in its most general and commonsense terms to be a matter of ‘what happened’, we can

Talking about what happened 157 ask “How was language part of what happened?” After all, historical events very often involve language. ● ●

● ●

Churchill was told in advance that the Germans were going to bomb Coventry. The American naval captain John Paul Jones shouted across the water to the British captain that he would never give up his ship. General Douglas MacArthur declared to the Philippine people “I shall return.” President John F. Kennedy informed the Berlin people that he was a jelly doughnut.

Naturally, then, the historical events and circumstances addressed by historians often involve language. And this raises a question: Given that the work of historians begins with descriptions and statements of what happened, how should they address the roles that language plays in those historical events? Of course, there are also historical events – with or without the participation of humans – which do not involve language. It seems obvious that these two – events with and events without language – should not be treated identically, that is, that the former pose an extra layer of issues and questions not raised by the latter. If we take this difference as a given, then what should the historian do to account for it? When language is present, how should the historian approach the role that it played in ‘what happened’? A second question, related to this first, is one that can be – and, indeed, increasingly is – raised about language and history. This question concerns talking about historical events, whether this talk (or writing) occurs at the same general time of the events themselves or at a later time. This question can be put as follows: What effect does the way that we characterize an historical event have in how we treat that event? I have cautiously put that final clause – “how we treat that event” – in very vague terms. But some theorists today are less cautious and would put this question in a more specific form: How and to what extent is what-we-say-about an historical event ‘constitutive’ of the event? Or: how does what-we-say-about-it ‘construct’ the event? Personally, I am skeptical about the manner of speaking in which language or a particular use of language is said somehow to be ‘constitutive’ of an historical event, or to ‘construct’ that event. I cannot go into my reasons for this skepticism here (see Taylor 1997: 19–21), and, in any case, I sympathize with the more general idea which I take to motivate such claims. For, it does matter how people characterize what happened – or rather, it can matter. It can matter to the event’s participants or to those who observed its occurrence. It can matter to those who hear, read, or think about what happened at a later time. It can matter to the people – historians or others – who are characterizing the historical event in those terms. And it can matter more than one might naively assume that the verbal designation of an event could matter. At the same time, we must not overlook the fact that it can matter in what would seem to be an endless variety of ways. This is one of the reasons why

158

Talbot J. Taylor

I am unhappy with speaking of language as “constituting” or “constructing” the event. Those are catch-all terms that do no justice to the variety and complexity of ways that how an event is spoken or written about can matter to various people in various circumstances. “Constructing” and “constituting” make how the verbal characterization matters sound too consensual, too homogeneous, even too context-independent. Now consider a third kind of important question about language and history. What account should historians take of the various kinds of rhetorical relationship that hold between the intellectual discourse of “History with a capital H” and the lay discourse of ordinary members of the relevant culture who state (as we say) “in their own words” what happened? This general question about lay and intellectual discourses about history raises a number of more specific issues. But let us focus on one in particular: What difference does it make if the history-writing is produced by historians who live in and are members of a different culture from those who were part of what happened? If we grant the general point that how an event is spoken about matters to various constituencies, then what account should be taken of the differences between how event E was characterized at the time that it occurred and how today’s historian characterizes it? To connect this back to the preceding question: Is the way that today’s historian reflexively “constructs” the historical event different from the way it was reflexively “constructed” at the time, by the members of the culture in which that event took place? Does that difference matter, and how in particular should the historian take account of that difference in her own discourse? ***

I have raised three different-but-related questions about the conjunction of language and history. The question on which the rest of this paper will focus incorporates aspects of all three of these initial questions. Before stating this question, it would help first to bring to mind the kind of historical circumstances whose occurrence is typically reported in utterances using verba dicendi, that is, “verbs of saying” or speech act verbs. In other words, I am speaking of historical events which might be reported by saying that: ● ● ● ● ● ● ● ● ● ●

S promised to H to do X S threatened to do X S reported to H that P S mentioned that P S warned H that P S okayed H’s proposal S thanked H S invited H to P S congratulated H S boasted that P

Talking about what happened 159 ● ● ●

S admitted that P S insisted that P S lied that P

In some languages there are also nouns which correspond to their verbs of saying: thus, for example, in English one can speak of a promise, a threat, a report, a lie, a boast, a warning, etc. In sum, I am speaking of historical events 1 2 3

in which language plays an integral part, the cultural significance of which, under one theorized manner-ofspeaking, is at least in part ‘constructed’ by the verb of saying, and which present interesting rhetorical problems for those who would represent such events in intellectual discourse.

How should the historian (or the linguist or the discourse analyst, for that matter) characterize such events? For events of this type, how should he say ‘what happened’? This question should be of triple interest to historians and any others who reflect on the theoretical implications of historical discourse. In addressing the question I have just raised, certain factors must be taken into account. In the first place, we must keep in mind that verbs of saying are not linguistic or cultural universals. In spite of some arguments (Wierzbicka 1996) claiming that there are in fact a few universal verbs of saying, it is abundantly clear from ethnographic studies (e.g. Stross 1974, Silverstein 1985, Verschueren 1985, Agar 1994) that different languages – different cultures – have vastly different inventories of such terms. Language 1 may have many more verbs of saying than Language 2. Language 2 may have no way of expressing the meaning of one or more of Language 1’s verbs of saying. The set of verbs of saying in Language 2 may not be isomorphic with the equivalent set in Language 1, with a variety of gaps, mésalliances, asymmetries, and faux amis that are detectable only on comparative analysis. Furthermore, it is no less clear that one and the same languaculture – to borrow Michael Agar’s (1994) terminological blend – will change its verbs of saying over time: change, that is, both the inventory of verbs and their meanings. This relativism of verbs of saying suggests a complication for the historian: the historian’s languaculture today may not have in its inventory the verb of saying that a member of the contemporary languaculture would have used to report what happened. In this case we must ask: how should the historian characterize the event? Should he use one of his own culture’s verbs of saying even though it was not part of the languaculture’s inventory at the time of the reported event? Or should he use the verb of saying that the culture’s own members would have used at that time, even though this verb of saying does not exist in the historian’s present-day languaculture? And what are the implications of each of these options? There is a second important factor that we must take into account in considering the question of the use of verbs of saying to characterize historical

160 Talbot J. Taylor speech events. Verbs of saying have important and culture-specific significance. Much of the time when, in English, we report what happened between a speaker and a hearer, we do so using a verb of saying: that is, we say something like S promised to H to do X; then H insulted S; whereupon S confessed to H that P, and H replied that Q, and so on. The metalinguistic framing of such reports can both draw on and influence cultural understanding. Without going too far down the “language-constitutes-experience” road, we can still make sense of the claim that much of our understanding of ‘what happens’ in interpersonal communication is filtered through our language’s verba dicendi. At the very least we have to accept that we talk about speech events using verbs of saying (and their related nouns) and that it is by means of such talk that we inform others of, negotiate responses to, raise questions about, express objections to, discuss the implications of, and provide a justification for speech events – as well as make comparisons and other kinds of connections between two or more speech events. In other words, our verbs of saying are tools whose employment is crucial to the methods by which we in our culture attempt to negotiate the interpersonal significance of what happened in a given speech event or set of speech events. In considering the cultural significance of verbs of saying – and the differences between one languaculture’s verbs of saying and another’s – some thought should be given to what is involved in being a competent user of a given verb of saying. For example: ●

●

●

If you know what it means in English languaculture to say that S promised to H to do P, then you know that, all things being equal, H will now take S to be committed to do P. If you know what it means in Chinese languaculture to apologize (Chinese word: daoqian), then you know that the person who has been apologized to – the apologizee – has the right to determine if there has been any harm done. If you know what it means in American adolescent culture to ask someone for a date or to snitch on someone, then you know that . . .

One way of putting this general point is that being a competent user of a verb of saying is not just a matter of knowing some atom of meaning or definitional equivalence, but rather of being familiar with a whole web of implications, from the potential to the inevitable, which competent members of the culture are inclined to draw from its use. This means that even if between two languages there is a conventional translational pair of two verbs of saying – say, English ‘apologize’ and French ‘demander pardon’ – one should not assume that the use of each expression in its home languaculture has the same significance. While what is termed the two expressions’ ‘lexical meanings’ may conventionally be said to be ‘the same’, still, the web of implications that the use of each draws on could be very different. This also applies, of course, to the pairing of the “apologize” of renaissance English with the “apologize” of American English today.

Talking about what happened 161 A further implication of this line of discussion is that simply using another culture’s or another time period’s own verb of saying in reporting a speech event is not enough to avoid the problems raised here. For the reader of the report will need much more background, cultural information, etc. if she is to understand the cultural significance of that report – what that report would mean (in the much fuller – cultural-implicational – sense of that term) to the members of the relevant languaculture. Consider, for example, saying that ‘Harald Harbraba gave his word to William that P.’ One has to know a lot more about medieval culture to understand the implications – cultural significance – of such a claim. We might at least begin by asking what it meant to give one’s word in Harald’s time and in what ways this is similar to or different from what the implications might be of giving one’s word in contemporary American culture. Rather than continue this somewhat unstructured list of the various kinds of anxieties that one might have in using verbs of saying to report historical speech events, I would like to shift the focus to the consideration of the arguments of a leading school of linguistic anthropologists, grouped around Michael Silverstein, which has claimed that there is crucial theoretical importance to the question of what social scientists should make of folk linguistic expressions. In the course of field work, linguists, like other anthropologists, spend a great deal of time listening to people talk about what they are doing. The resulting data form a corpus of speech about speech, a “meta-corpus” as it were, that consists of speech at the same time that it seems to talk about, or characterize, speech as a meaningful social action. In reply to our queries or spontaneously, people will utter descriptive statements about who has said or can say what to whom, when, why, and where, just like statements about who can give presents of certain kinds to whom, when, why, and where. But talking about “saying” is, for better or worse, also an example of “saying”; and such metalanguage, for the analyst of culture, is as much a part of the problem as part of the solution. As is readily apparent, all our efforts to differentiate “conscious native models” from “anthropologist’s models,” or “ethnotheories” and “ideologies” from “objective social reality,” are attempts to come to grips with the metalanguage vs. language relationship, or its more general form, (meta-) language vs. action. (Silverstein 1981: 1) ***

Perhaps the central concept in Silverstein’s extensive theory of language is that of ‘metapragmatics’: If we take pragmatics to be the phenomenon of signal usage in communicative situations, then metapragmatics is the metalinguistics that

162

Talbot J. Taylor describes such pragmatics. Not surprisingly, metapragmatic usage in a language frequently centers on seeming . . . functional descriptors of what people use the language for, constituting a lexical set of verbs of saying (verba dicendi). (Silverstein 1985: 133)

Silverstein and his colleagues have published widely on the kinds of metapragmatic features that they have identified in the discourses of different cultures and on the kinds of communicative functions that these metapragmatic features can be seen to serve. Arguing from a theoretical model that Silverstein has developed over the past three decades – combining Peircean semiotics with Prague School functionalism and Whorfian ethnolinguistics – he and his colleagues argue that the metapragmatic features of discourse, in their view, extend far beyond the verbs of saying that first come to mind. On the premise that metapragmatic forms all signify something about what is or was an ongoing communicational situation, Silverstein and his colleagues include in the class of metapragmatic features a wide variety of indexical forms that includes, for example, first- and second-person pronouns, demonstratives like this, that, and those, tense and aspect indicators, prosodic structure, and more. But naturally, I will confine my attention in most of what remains to what Silverstein and his colleagues have to say about verbs of saying. It is noteworthy that, unlike the vast majority of orthodox linguists, Silverstein does not treat the metapragmatic function of language as a supplementary add-on to language’s more common-and-garden communicational functions. On the contrary, Silverstein takes the metapragmatic function to be essential to language: in particular, he sees it as necessary for the coherence and interpretability of interactional discourse. Loosely put, his claim is that it is only because we talk about contextualized linguistic activities – and conceptualize them accordingly – that those activities can serve as effective means of communication. In a 1993 paper, he states: Without a metapragmatic function simultaneously in play with whatever pragmatic function(s) there may be in discursive interaction, there is no possibility of interactional coherence . . . (. . .) Understanding discursive interaction as events of such-and-such type is precisely having a [metapragmatic] model of interactional text. (Silverstein 1993: 36) In other words, Silverstein claims that discursive interaction is only possible because, among other things, speakers and hearers conceive of what-is-going-on as consisting in events of particular types; they impose a metapragmatic model on their and their interlocutors’ contributions to the developing interaction. The ‘smoothness of coherent discursive interaction’ would be impossible without this kind of metapragmatic regimentation

Talking about what happened 163 (Silverstein 1993: 48). One of Silverstein’s colleagues at the University of Chicago, John Lucy, puts this point even more starkly, declaring that “reflexive ((or metapragmatic)) activity is essential to language use.” (Lucy 1993a: 18) While verbs of saying – which exist in every language – are just one of the many kinds of metapragmatic feature that populate discourse, nevertheless, according to Silverstein they have a special importance. For it is by means of its verbs of saying that a community’s conceptualization of the eventstructure and purposivity of linguistic activity is constructed – in other words, their metapragmatic model of what happens in verbal interaction and how one can bring it about. This conceptualization amounts to what Silverstein calls a ‘cultural ideology of language’ or a ‘folk metapragmatics’. In turn, Silverstein holds that how a culture conceptualizes linguistic activity influences the ways in which that culture’s speakers conceive of contributing to verbal interactions, the ways in which they do in fact contribute to those interactions, and the interpretive means by which hearers make sense of their contributions. In other words, a culture’s metapragmatic ideology not only influences how its members conceptualize language-in-use but also how they participate in its on-going practices: in particular, how they integrate acts of speech into their interactional participation. Furthermore, it is because a culture’s verbs of saying serve to construct its members’ folk metapragmatic ideology that Silverstein says that these verbs give linguists, anthropologists, and others an empirical entree into the conceptual understanding of language that each society of speakers brings to bear on the activity of actually using it . . . . In other words, there is a necessary relationship between the way in which metapragmatic constructions of languages code the pragmatics of speaking, and the ideological and cognitive strategies that speakers employ in culturally-conceptualized situations of speaking, i.e, in their [purposive use] of language. (Silverstein 1985: 138) This view of folk metapragmatic ideology is related to a more general theoretical premise underlying the work of Silverstein’s school. Linguistic interaction is taken to be a dialectic between signs and interactional context, but, crucially, a dialectic which is mediated by language ideology. For it is a culture’s metapragmatic ideology which informs its members’ purposive use of signs to respond to and contribute to ongoing interactional circumstances. “Ideology and praxis reinforce each other . . .” (Silverstein 1979: 210). It is clear, then, that Silverstein and his colleagues attribute great importance to folk metapragmatic discourse, as to its constituent terms and expressions, such as its verbs of saying. In this case, one may be surprised to learn that Silverstein and his colleagues are at best ambivalent about the role

164 Talbot J. Taylor that metapragmatic expressions should have in the study of language and discourse – that is to say, in the subfield of linguistics called pragmatics. [H]ow fraught with danger is our taking at face value any statements by participants about various pragmatically-meaningful action . . . (Silverstein 1981: 20) Why, one wants to ask, would it be “fraught with danger” to take at face value a participant’s claim, say, that he had promised to do something? Or that he had apologized? Or that he had been teasing? Or that he had been complaining? Silverstein’s answer is that native participant testimony is a kind of ‘metabehavioral interpretation’, that is, an “ethnosociology” with a folk-metapragmatic component (Silverstein 1976: 52), and the view from this perspective does not necessarily correspond with “what the natives are really doing” (Silverstein 1976: 23). His support for this skeptical judgment is that “natives’ understanding of their own systems of linguistic usage frequently conflicts” with the understanding of their behavior that can be developed from (a more objective) linguistic perspective (Silverstein 1979: 208). In a 1976 paper he laments: I think that every fieldworker has had such experiences, where a careful sorting out of kinds of pragmatic effects ultimately just cannot rely on the metapragmatic testimony of native participants. (Silverstein 1976: 49) What are Silverstein’s objections to folk metapragmatics, in particular to verbs of saying? He spells out these objections in a series of articles published in the late 1970s and early 1980s in which he criticizes John Austin’s theorizing of speech acts and of the various kinds of ‘acts’ and ‘forces’ which Austin conjures up to explain how we do things with words. Silverstein claims that, from the native folk-ideological point of view, performative formulae such as “I promise you that P” seem to accomplish, to ‘do’, some specific predicated transformations of the social relations and other contextual understandings in the situation of speech . . .” (Silverstein 1979: 209) For example, folk ideology takes the utterance of the performative formula ‘I promise you that . . .’ to place the speaker “in some new understood relationship of obligation to the addressee.” The utterance of the performative formula – given the satisfaction of certain prerequisites – is, in and of itself, the ‘doing’ of a kind of conventionally recognized act. This is the act of promising. The uttering of various other performative formulae – again, given the satisfaction of the relevant contextual conditions – constitutes the

Talking about what happened 165 performance of other kinds of illocutionary acts, such as apologizing, baptizing, ordering, betting, warning, etc. Austin calls these kinds of communicational acts ‘illocutionary acts’. These acts are the product of the illocutionary forces which, Austin claims, the performative formulae have when used in context. However, Silverstein argues that the concept of the illocutionary act, like that of illocutionary force, is an illusion. There are no such things as illocutionary acts or forces. Instead, these so-called acts and forces are mythical objectifications of folk metapragmatic discourse. In other words, to take a step back for a moment to this chapter’s overarching theme, Silverstein claims that the kinds of speech acts which are commonly reported in historical discourse – promises, apologies, insults, giving one’s word, warnings, etc. – are merely the illusory figments of folk imagination; moreover, the folk imagination of a particular languaculture at a particular historical moment in time. In other words, not only are verbs of saying culturally dependent and variable, they also stand for phenomena which are only the illusory figments of a culture’s folk-linguistic imagination. If this is true, then it is clear that there is even more justification for the historian or linguist to be wary of using verbs of saying in reporting historical speech events: [L]inguistics has characteristically taken at face value a native ideology that objectifies “force” in language itself, concretizing this force in terms of propositional-structure-as-usual. Social anthropological studies have characteristically analyzed native ideology as though it were an accurate “scientific” picture of the relation of language form to social context. Both of these . . . are expectable in terms of a broadened understanding of the bases of ideology that Whorf proposed, especially as such broadened understanding explains the tendency to assimilate our own “scientific” views to the source from which they have emerged, our own European folk ideology of language. (Silverstein 1979: 204) However, before we can weigh up the implications of Silverstein’s argument, we need to look more closely at its grounding in Silverstein’s analysis of how Austin was fooled into believing in the existence of illocutionary acts and forces. “Where,” Silverstein asks, “do Austin’s ideas come from about the nature of these ‘acts’ and the ‘forces’ and ‘effects’ they embody?” (Silverstein 1979: 210). Their origin lies in the ‘objectification’ of “particular ways of reporting ‘what happened’ in an event of using language” (Silverstein 1979: 211) – in particular, in the “projection of cryptotypic selective categories of lexical forms in typical metapragmatic discourse” (Silverstein 1979: 210). The following is an abbreviated explanation of Silverstein’s reasoning. Silverstein says that native metapragmatic ideology has a tendency to “rationalize the pragmatic system of language . . . with an ideology of

166 Talbot J. Taylor language that centers on the surface-segmentable items” of propositionally interpreted sentences (Silverstein 1979: 208). Thus, native metapragmatic ideology tends to locate the pragmatic effectiveness of language-in-context in particular words, phrases, and similar surface-segmentable items of sentences. For instance, native ideology focuses on the surface-structure similarity (i.e. cryptotypic relationship) between indirect quotations like a b

“He said that he will buy the house,” and “He promised that he will buy the house”

Both of these sentences describe a speech event, and, because the framed construction “—— that he will buy the house” is the same in both cases, native ideology rationalizes that “promising is really saying . . . plus something else” (Silverstein 1979: 213). In other words, it identifies a relationship of hyponymy between the two lexical forms say and promise and concludes that . . . this relationship reflects a real inclusion relationship at the level of what happened. (Silverstein 1979: 213) It is therefore merely a matter of following native ideology when Austin reasons that an utterance like “I promise that I will buy the house” consists in the rhetic act of my asserting a particular propositional content – that I-will-buy-the-house – plus the illocutionary act of promising. Note that the reasoning proceeds from a perceived cryptotypic relationship at the surface level of utterances to an assumed pragmatic relationship at the level of the communicational act itself. Second, native ideology focuses on the surface structure analogy between the metapragmatic reports c

“He promised that P,”

on the one hand, and, on the other, d

“He joked that P” “He complained that P” “He answered that P” “He explained that P” “He threatened that P” “He boasted that P” “He exclaimed that P”, and so on.

Talking about what happened 167 Note that while “He promised that P” could be a report of the utterance of a grammatically related performative formula “I promise that P,” the reports under (d) could not. There are no performative formulae in English such as “I explain that P” or “I joke that P” or “I answer that P” or “I threaten that P” or “I boast that P” to relate in the same way to the third-person reports under (d). Nevertheless, on the basis of the perceived cryptotypic pattern between “He promised that P” and the other third-person metapragmatic reports under (d), native ideology suggests that – like promising – joking, complaining, explaining, threatening, boasting, etc., are also kinds of illocutionary acts: pragmatic add-ons to the rhetic act of expressing a propositional content. And, by extending this reasoning to similar analogies between other kinds of metapragmatic reports, native ideology also concludes that teasing, insulting, whining, appealing, asking for, bragging, justifying, thanking, emphasizing, etc., etc., are further kinds of illocutionary forces that an utterance might have. The fact that, in contrast with “He promised that P,” these others do not have related explicit performative formulae is explained away by assuming that the performance of these illocutionary acts is always ‘indirect’. That is, their illocutionary force has to be deduced from the contextual utterance of a non-explicit sentence. According to Silverstein, it is this kind of folk-ideology-guided reasoning that leads Austin to declare that every utterance is a laminated form involving a phonetic act, a phatic act, a rhetic act – these last three making up the locutionary act – an illocutionary act and perhaps also a perlocutionary act. But Silverstein’s view is that these types of acts are nothing but mythological beasts. Complaining, explaining, joking, teasing, etc.: these are not real acts that an English speaker – or any other kind of speaker – ever produces in verbal interaction. Silverstein grants that it is important to English speakers to speak of these acts, and even to design their contributions to linguistic interaction on the premise that such acts regularly occur. But this should not lead the theorist to construct pragmatic theories which attribute a reality to these ideological objectifications of the metapragmatic discourse of English languaculture. Silverstein concludes that “speech-event names” do not “delimit a principled area for ‘scientific study’ ” (Silverstein 1979: 210). There is no reason why an ideology that grows piecemeal from various metapragmatic formulations of a language should . . . give adequate analytic insight . . . (Silverstein 1979: 214) Silverstein has a different explanation of the pragmatic properties of communicational interaction and of the pragmatic function of performative formulae such as ‘I promise that P’. This is not the place to go into the details of his pragmatic model. In any case, it is not relevant to the general point that I am trying to raise here concerning what use historians, linguists and anthropologists should make of native metalinguistic characterizations.

168

Talbot J. Taylor

However, it is worth looking briefly at what Silverstein says elsewhere about the folk metapragmatics, not of English, but of Chinook, a Native American language formerly spoken near the mouth of the Columbia River, in the northwestern United States. Again, I will not have space to give even a brief sketch of Silverstein’s insightful analysis of the various verbs of saying used in Chinook narratives. Needless to say, the metapragmatic forms characterize their framed speech reports in ways that have little if any correspondence to the ways that the familiar English verbs of saying do. But, what is most relevant in the present context is Silverstein’s claim that Chinook has no metapragmatic forms that characterize an utterance as an assertion. [I]n Chinookan metapragmatic coding . . . there is no explicit metapragmatic concept of referring-and-predicating as [a purposive function of language] akin to that embodied in our metapragmatic formulae such as “He asserted that . . . ,” “He stated that . . . ,” – or even “He explained that . . .” Nor is there any text-pragmatic evidence that . . . truth-functional semantic propositionality . . . is of significance to the Chinookan conceptualization of language use as a significant activity. (Silverstein 1985: 169) Nevertheless, this does not lead Silverstein to conclude that Chinook utterances never assert propositional contents or represent states-of-affair in truthfunctional terms. On the contrary, Silverstein maintains that propositionality is implicit in the Chinook narratives and in the Chinookan speech events reported in those narratives (Silverstein 1985: 169). However, propositionality is never a focus of Chinookan metapragmatic discourse; nor is it therefore of significance to the Chinookan conceptualization of language use, since, as we have seen, that conceptualization is derived from the culture’s folk metapragmatic discourse. So, these two examples – English and Chinook verbs of saying – appear to illustrate opposite possibilities. In the case of English, there is a wealth of verbs of saying, but these do not correspond to the real pragmatic properties of English language-in-use (although those verbs have misled speech act theorists into thinking that they do). On the other hand, although Chinook does have verbs of saying, they are not translation equivalents of the English ones. But, even more important is the fact that Chinook does not have a verb of saying that represents a pragmatic feature that really is characteristic of Chinook language-in-use – namely, the production of utterances that assert propositional content. In both cases, a realistic interpretation of their inventory of verbs of saying gives a false picture of the pragmatic properties of language-in-use. Furthermore, given Silverstein’s general claims ●

●

that metapragmatic discourse influences speakers’ conceptualization of the properties of discursive interaction and, just as important, that this in turn influences how they design their own participation in discourse,

Talking about what happened 169 we can only conclude that English and Chinook speakers conceive of the properties and possibilities of language use in very different ways and that, on the basis of those conceptions, English and Chinook speakers purposively contribute to the ongoing production of discourse quite differently as well. In other words, to put this in more general terms, both the conceptualization of discursive interaction as well as the pragmatic techniques that communicators use to integrate their verbal behavior into particular interactions are culturally variable. Moreover that variation is, at least in part, the product of – that is, are discursively constructed by – the particular properties and patterns of ‘folk’ metapragmatic discourse in their community. However, notwithstanding this wide range of cultural variability, Silverstein suggests that there are pragmatic properties of discursive interaction which remain unaffected by the cultural vagaries of ‘folk’ metapragmatic discourse and which are only discoverable by the analytical methods of comparative linguistic research. Chinook speakers do assert, even though they never talk about it, and English speakers do not in fact produce many of the kinds of speech act that they frequently talk about in reflexive discourse. ***

In an earlier paper (Taylor 2001) I discussed the different positions taken by philosophers and cognitive psychologists regarding the use of what are called ‘folk psychological’ expressions: expressions such as S believes H, S intends to do X, S wishes that P, etc. I opposed one group of theorists, the so-called Eliminativists to two other groups, the Realists and the Anti-Realists. The Eliminativists argue that the expressions of folk psychological discourse should be eliminated from scientific discourse about the mind. On the other hand, the Realists and Anti-Realists insist on the retention of these expressions in scientific discourse. The Realists and Anti-Realists part company, however, on the issue of the phenomena referred to by folk psychological expressions. The Realists claim that there are real, cognitive phenomena referred to by expressions like ‘belief’, ‘wish’, and ‘intention’, while the Anti-Realists maintain that those expressions have no referents: they seem to stand for something but do not really. Nevertheless, the Anti-Realists argue that because folk psychological expressions have a crucially instrumental function in interpersonal discourse, the Eliminativist’s project needs to be resisted. Daniel Dennett, a leading anti-realist, argues that what he calls ‘the intentional stance’ – the lay psychological ideology embodied in folk psychological discourse – is essential to human activity and that to do away with it would be to do away with our ability to understand others and ourselves. And yet surely that ability is part of what cognitive science needs to explain. In other words, eliminating folk psychological expressions from cognitive scientific discourse would mean that cognitive science had no access to the role that self-understanding plays in human psychology. Nevertheless, Dennett does not intend this defense of folk psychological discourse to be taken as an

170 Talbot J. Taylor argument in favor of basing the terms of cognitive science on those of folk psychological discourse – for those expressions of folk psychological discourse do not in fact refer to anything. On the contrary, while the folk psychological ideology is essential to human psychology and therefore to the subject matter of cognitive science, it should not be taken as a fundamentally accurate view of that subject matter. It is a misleading basis for the scientific theorizing of the human mind. It might appear therefore that Michael Silverstein’s position on folk metapragmatic discourse is analogous to that of the anti-realist Dennett on folk psychological discourse. Just as Dennett claims that ‘the intentional stance’ – the folk psychological ideology – (i) is essential to human psychology but (ii) is not a firm basis for a scientific appraisal of human psychology, so also Michael Silverstein claims that a culture’s metapragmatic ideology (i) is essential to language use in that culture but (ii) is not a solid basis for a scientific appraisal of the true pragmatic properties of their discursive interactions. Without dismissing the practical importance of folk psychological concepts and theories to human behavior, Dennett feels that cognitive science ought to free itself from the cultural prejudices and idiosyncratic reasoning that it embodies. Similarly, Silverstein feels that a scientifically based pragmatics needs to look both at how language ideologies influence discursive behavior and at the pragmatic realities that are obscured by those culturally specific ideologies. Where does this leave the historian? What should the historian – or the linguist, for that matter – do with the ‘folk’ reports of what happened, in particular when the reports in question characterize what happened in terms of a particular kind of speech event? Should such reports – and the psychological and interactional conceptions they embody – be eliminated from professional historical discourse? Should professional historical discourse refrain from using verbs of saying and speech-act names? Or should the use of those items of a culture’s reflexive linguistic inventory be treated as an accurate way for the historian to communicate to her audience what really happened? Or should the historian take the anti-realist view, acknowledging the instrumental effectiveness of those expressions for the members’ own construction of cultural self-understanding and so also cultural praxis, all the while insisting that those expressions are unsuitable for re-employment in the construction of professional histories? What relationship should there be between cultural self-understanding and the professional understanding that is the goal of historical discourse? I realize that this question is hardly a new one for the historian, although I think the particular problem that verbs of saying pose to the history of events-of-speech has not received as much attention in historical discussions as it deserves. At the same time, I believe that modern linguistics remains deeply ambivalent about the role that folk metadiscourse should play in the theorizing of language and discourse. If we look into the foundations of modern linguistic theorizing, we will find, I believe, that this ambivalence

Talking about what happened 171 has been an ever-present source of rhetorical anxiety and that it has motivated a variety of theoretical acrobatic stunts to counteract its influence. The foundational distinction between ‘etic’ and ‘emic’ properties of language is an easily recognizable example of this. As both a reaction to and a continuing source of nourishment for this anxiety has been the tendency for linguistics to proceed as if the problem doesn’t really exist or is at least only a background topic that can be ignored in front-line theorizing and description. It is time that this topic were brought once again into the discursive foreground of language theory.

11 Part of the meaning/history of euro Integrational corpus linguistics Michael Toolan

Is it possible to write an integrational linguistic history of the word euro, contributory to an account of how the United Kingdom joined or did not join the euro and the EU exchange rate mechanism in the early years of the twenty-first century? And can such a history-writing use corpus-linguistic methods? How might one begin to go about these tasks? In Signs, Language and Communication, p. 32, Roy Harris notes with understandable integrationist approval Collingwood’s argument that the historian too often misconceives the past; that misconception Collingwood called ‘substantialism’. This is the assumption that the acts and events that are the stuff of historiography are acts and events that flow from, or around, or occur to, entities that themselves exist before, after, and independently of those events. They are entities outside the historiography and those communicators who participate in that historiography. This appears to be a sub-type of what has come to be known as foundationalism or essentialism; it is a thing-orientedness that is as deeply rooted in linguistic thought as in historiography: there are ‘things’ (Paris, the monasteries, the village, the family, Ireland – all of these invite scare-quote diacritics) that subsisted in the past, and endured, and the historian’s task is to describe and explain what happened to them. Everywhere you look, certainly in the practices of the everyday world, you see confirmation of this substantialism, beginning with the titles of books and essays: a history of New Orleans, a history of baseball, and so on. The tell-tale word here being of. How rare, by comparison, is the book title, and concomitant presentation of content, which declares: New Orleans: what it was like before it became what it is now or Baseball today: a description of its antecedents Not the catchiest titles, I grant; but why do we almost never see such formulations? The reasons are perhaps pragmatic, a triumph of expediency over precision or theoretical correctness. For one thing, Baseball today: a description

Integrational corpus linguistics 173 of its antecedents, might give the impression that chapters will proceed in reverse chronology, moving from baseball in the 1990s back to the 1980s, then to the 1970s. This is not such a terrible idea, since it would highlight causation, always at the core of history, narrative, and the law. In effect such a reverse-chronological telling would repeatedly explain (or attempt to) why this happened and why that was done. But as is well known we privilege outcomes over causes, we very much like the idea of beginning at the beginning and going on until we reach the end; not least because we share the common sense that our own lives have begun and are advancing towards a close. We also privilege the individuation and persistence of things over time (or, more accurately, regardless of time): property, including personal property; and specificity of reference; objectification; and segregation. One thing at a time, a place for everything and everything in its own place. In our collective Western subscription to and enforcement of a modus vivendi and modus operandi, it makes pragmatic sense, unquestionable common sense, to write, read and teach histories of things, in which we postulate that those things are relatively stable entities, persisting in part if not in whole over time. This goes along with an objectification of history itself, as a thing. Hence the default history-book title is A History of X, history being here understood as a thing that attaches to X, containing a chronological account and explanation of all the things that have happened to X. But such objectification or substantialism also shows up in the indefinite article usage, as when the doctor asks me if I have a history of asthma. I am not supposed flippantly to reply that yes, it’s right here in my briefcase; I am supposed to concede that, as I am asthmatic from time to time, I do indeed ‘have’ a ‘history’ of asthma. As soon as something is denoted by specific or definite reference, it is a long way towards being able to claim a history. How does all this relate to the euro? One of the things of interest about the euro is that, in a sense, prior to the 1990s, it had no history – and is therefore apprehended by us today in a different way to that in which we look upon the British pound or German mark. At the same time, in light of the textualist inflation of recent years, according to which nothing of social or cultural significance can occur without accompaniment by a sea of written and spoken and recorded commentary, available to most of us via the mass media (again to an unprecedented extent), the euro is accruing to itself a historiography at an unparalleled rate. There are few people within the borders of the European Union who live entirely outside the market economy and who therefore have no need to have any dealings with money. Coins and currency touch almost everyone, and so everyone wants to have their say, their ten cents worth, on what the introduction of the euro means and entails. And perhaps no community has had more to say on these matters, singly and collectively, even though we might still not ‘go in’, than the British. Nor does it seem to be the kind of issue where, once you’ve said your piece, all urge to debate is spent. People return to the topic continually, as to an old grievance or the picking of a scabbed

174

Michael Toolan

wound. Perhaps the sense, in Britain, of an extended span of prevarication, of uncertainty of political will and judgement, has fuelled this serial contentiousness. Within the British broadsheet press, a relatively small group of journalists has returned to the topic to offer speculative commentary, over and over again; to cite an example from among the nay-sayers, William ReesMogg offered one of his ten cents worths in The Times of 7 January 2002, in a commentary characteristically entitled ‘Why I wouldn’t put my money on the euro’. Or, as part of the same process of commentary but approximately a year later (28 January 2002), we find the Guardian’s Patrick Wintour writes a report headlined ‘Gloom dominates as June deadline looms’, with the telling sub-heading: ‘In the second of our reports marking 12 months of the single currency we ask if British voters will ever be given their say’. Over the past few months, while the public discourse in Britain – among the ordinary voters whose decision will be crucial if a referendum transpires – has sputtered along at a relatively low level of sophistication and detailed analysis, one major contrast has gradually shifted in a sensible direction. That is the contrast between 1)

2)

thinking of ‘going into the euro’ as chiefly a matter of dropping the pound and adopting a uniform European currency, with all that this entails in terms of ease of travel and work in more than one EU country. understanding that ‘going into the euro’ is more importantly a matter of going into the ERM, of transferring monetary policy decisions (interest rates and money supply) from the present working coalition of the Bank of England committee and the Treasury, to a EU-wide authority or authorities, principally the European Federal Bank based in Frankfurt, the ECOFIN committee of the European Commission and Council of Ministers. (The Stability and Growth Pact, designed to advance the provisions of articles 103 and 104 of the TEU, promoting balanced national budgets, whose effect is to disbar on pain of financial penalty any member state from going into budget deficit to the level of 3% of GDP or more.)

Gradually, the latter consideration has become the focus of attention rather than the former. There is a linguistic analogy to the position of those who still concentrate on the minor consideration numbered (1). They are very much like those people who advocate embrace of English as an international lingua franca on the grounds that all those ‘excessive’ translation and interpreting costs that are otherwise involved in the multiple exchanges of internationally relevant speeches and documents are, at a stroke, done away with; this is a shallow efficiency argument, segregational nomenclaturism at its most naked. Rather more defensible is the argument, in relation to the fostering of English as a lingua franca for the EU region, that a political community needs a shared language in which to develop a democratic and participatory discourse. But even here, one wants to see the dog wagging the tail and not the other way about: the linguistically

Integrational corpus linguistics 175 diverse communities should be pressing for the development of arrangements in which greater mutual dialogue can develop, rather than people falling into the delusion of thinking that language planning and policy (even imposition) can precede and usher in any major shift in identities. In the linguistic case, analogous to the euro case, the careful and considered sense that it better meets a diversity of speakers’ interests that they all use English for certain communicational and expressive purposes – perhaps, for example, because the business at hand is in some sense an English-medium activity – needs to underpin individual and collective decisions to adopt English. In neither case can the massive upheaval involved – the moving away from one set of arrangements and affiliations and the moving towards another alternative constellation, all the tinkering with cultural and political roots and equally laborious graftings – be motivated by something so trivial as an urge to ‘cut out the middle man’, the interpreter or money-changer. It is commonly claimed that communication by way of linguistic signs succeeds because language users have come to some sort of agreement as to what individual signs mean or signify. It is no mere quibble about the use of the verb agree, however, if the puzzled student of language asks when the agreement (or any of the seemingly innumerable agreements) was made. At first glance, British domestication of the euro, should it happen, promises to be a great deal more consultative – extraordinarily consultative. It would appear, then, that the place and signification of the euro in British history and culture will be established by means of procedures rather far removed from the normal case for the signs that comprise la langue, as conceived by Saussure, who stressed that the individual by herself was powerless to create or modify those signs (31). This is Saussure’s ‘fixed arbitrariness’ of the signs in the system, beyond the reach of individuals and any voluntarist change. On the other hand, any referendum-mediated consultation of the British people, clearly, is not in practice a submission of a circumscribed body of evidence to the 40 million or so voters akin to the presentation of a case to a jury in court, with each juror empowered to make his private and secret decision. All kinds of hierarchy of influence and recommendation operate, and we know that the government will only run the referendum when it is sufficiently decided for or against, and is sufficiently confident that it has persuaded the majority of the voting electorate in effect to vote with it. It is hardly more than received wisdom these days to acknowledge the forces of discursive constructivism in the shaping of any identity (national, class, gender, etc.) of interest. All of this, in turn, is predicated on what the historian Patrick Joyce has summarized as the post-structuralist conception of language not as the mirror of a world external to it, but as a conventional and arbitrary structure of relations of difference, the eventual shape of which is produced through cultural and power relations. (‘The end of social history’, Social History, 20 (1995) reprinted in Tosh, ed. Historians on History, 274)

176 Michael Toolan That summary seems not entirely self-consistent (the ‘conventional and arbitrary’ sitting awkwardly with that which is ‘produced’ through power), but let that pass. What is of interest here are the myriad ways in which the discourse of and about the euro is contributing to the ‘eventual’ shape (sc. the status, the valeur) that the euro will take in British affairs (always with the integrationist proviso that any ‘eventual shape’ is continually pro tem, rather than fixed and permanent). The euro’s eventual shape, for me or anyone else, on the day after the UK referendum will yield place to its new shape the day after that, and so on. Nevertheless the shape or shapes that UK voters regard the euro as having when they enter the polling booths and vote accordingly will obviously be critical to their voting decisions, and the ongoing discursive contestation, evaluation, and calibration of it (every person a sage, grabbing a different part of this monetary elephant), is therefore of special interest. What I am particularly interested in is the history of the euro, or even simply the history of the word euro, in British political debate and public opinion. And for the remainder of this chapter I want to raise some difficulties generated by this local task of history-writing. The scope of the difficulty is evident, indirectly, in the commentary and analysis provided by Roy Harris, in Signs, Language and Communication, on the plotted shifts in the visual and verbal presentation of Hazell and Hawley’s Darkee toothpaste (which became, in English, Darlie toothpaste) in South-East Asia in the late 1980’s and early 1990’s (Harris 1996: 209–223). His commentary shows how complex the account needs to be, even for something as minor as a commercial toothpaste, as the latter went through a process of verbal and visual refashioning, in ways clearly calculated to satisfy or placate several interested groups, who used at least two languages and included both consumers and stockholders. And of course Harris’s account, as historiographer or semiologist, is by integrational definition just one account, at one time, for particular purposes. How much more complex and resistant to orderly structuralist analysis, then, is the euro, whose value and signification have already taken many forms. How does one ‘constrain the data’? One way is to chart usages and collocations in some influential public arena such as a national newspaper. That, in essence, is what a good deal of critical discourse analysis does: attempting a corpus linguistic quantification and interpretation of patterns of usage, collocation, etc., of a nodal or focussed-upon word, phrase, or construction. But corpus analysis is controversial – for no linguists more so than for integrationists. I will submit, however, that the accumulation and inspection of large corpora, as a reflexive analytical activity on a par in some respects with dictionarywriting and history-writing, cannot be dismissed out of hand. Rather it must be first recognized as another, relatively recently developed, strategy in the rich repertoire of metalinguistic activities available to linguists and nonlinguists alike. The interesting question is not whether we should have and consult corpora; Google.com is one such, that many of us use regularly, and it would be bizarre to proscribe such use. Rather the interesting

Integrational corpus linguistics 177 questions concern how we should use corpora: under what assumptions and with what expectations? To begin with, habeas corpus (You have the body: deliver it to us), cannot apply to linguistic studies. Individuals do not experience communication as a body, animate or otherwise, their own or someone else’s, self-contained, well-defined, separated from all the other activities that make up our lives. So to talk of ‘consultation of a corpus’ as if this were on a par with a medical examination is misleading if it leads us to forget that no corpus is just there to be consulted; rather, it has been fashioned, often with much effort, and, crucially, always for one or more purposes. The motivated or interested nature of corpus-creation applies as much to the Cobuild Bank of English and the British National Corpus as to the corpus I will introduce shortly. (As the large corpora grow yet larger – 400 million words and more – there is an increased temptation completely to disregard framers’ purposes, to pretend that this ‘ocean’ of data is just there, as-yet untouched by human intervention, innocently available to be ‘fished in’ by the enterprising linguist.) The larger the corpus, so the thinking runs, the greater its representativeness, reliability, authority, and neutrality, all breathing human passion and axe-grinding far above; and up to a point this is importantly true. But in a much more foundational sense, everything involved in the design and use of these corpora is informed by purposes, objectives, prioritized questions concerning the nature and structure of the language, and of real or imagined users’ needs, and the setting aside of countless alternative potentially pertinent questions. In all these respects corpora are essentially no different from dictionaries. This remains the case even as corpora are promoted as antecedent to, and often the best basis for, dictionary-making. It may not be easy to specify the purposes of the Bank of English corpus (some are still emerging), but it makes no sense to suggest that it was embarked upon without a purpose. And because the Bank of English, and its various sub-corpora, is language (a text, discourse) used for a purpose, there is a sense in which it is actually just like all other language use and not different in kind from it: it does not stand outside (near-) contemporary English, but is part of it. It is an enormous metalinguistic enterprise, differing chiefly in degree from the little metalinguistic enterprises that people invest in every day. For related reasons, integrationists are equally sanguine about the ontological status of linguistic data and linguistic facts, regarding both as precipitations from a particular kind of metalinguistic scrutiny of communicative activity rather than subsisting in the activity and existing prior to the scrutiny. But none of these qualifications, in my view, requires integrationists to turn away from metalinguistics and discourse analysis (and it is hard to see how a turning away from all their imaginable forms would be possible). And indeed there has been some attention to matters of discoursal political correctness from integrationists or those sympathetic to integrationism (Harris, as noted earlier [1996: 209–223] on the mutation of Darkie toothpaste into Darlie toothpaste, and Cameron, 1995, on verbal hygiene).

178 Michael Toolan Nor are integrationists alone in broaching foundational difficulties in practising corpus linguistics or critical discourse analysis (CDA). A recent exchange of views in the journal Applied Linguistics, between Widdowson, de Beaugrande and Stubbs, on the methodology and non-impartiality of corpus linguistics of whatever type, is relevant in this respect. Widdowson, for example, argues: ‘Although textual findings may well alert us to possible discourse significance and send us back to their contextual source, such significance cannot be read off from the data’ (Widdowson 2000: 9). Integrational linguists would go further, and put the word data in scare quotes; in first-order linguistic communication there are no such things as data, data being the analyst’s term for those interpretively decontextualized selections, from the flow of interaction, that they have decided to cast as evidence or topic of discussion. To write ‘a history of’ something, as indeed to write a dictionary or encyclopedia entry for something, is clearly to perform some kind of interpretive abstraction from the continuous sea of experience. Abstraction can evidently be of many kinds, and can be felt to be rather more or rather less removed from the particularities on which it is based. Narrative summary interests me in these respects, as a kind of abstraction which appears to involve rather less of an interpretive transformation or re-presentation of the original (where summary is thought of as a short version of a longer, fuller original). This seems fundamental to historiography. For a recent study (Toolan 2003), I prepared a small corpus of uses of the phrase politiquement correct (and its closest inflected or derived forms) to be found in the Le Monde online archive of articles published between June 2000 and June 2001. The aim was to produce, as a minimum, a summary of ‘how the phrase le politiquement correct was chiefly used in Le Monde in the late 1990s’, with a view to using that in making suggestions about how the phrase was used and understood and what leverage it did or did not have in French political discourse, more generally in the late 1990s. Mutatis mutandis and pari passu, something similar should be possible with respect to the phrase the euro. Clearly there are many layers of abstraction, interpretation, summary, sampling and classification involved here. But we do these things all the time, and circumstances require us to do them all the time, requiring us relatedly to bring closure at local points of arbitration, where we have to decide whether something is something we are in favour of or not, is good or not, fair or not, established or not, grammatical or not, to be used (written, said . . .) or not, and so on. What, then, might be the integrationist view of such exercises, using a corpus, summarizing, and so on? When George Wolf and his co-authors built up a database of New Orleans residents’ use or acceptance of various forms of their own family names, this was a kind of corpus-based linguistics. While corpus analysts themselves are keen to insist that their work is corpusbased but not corpus-bound, the distinction remains questionable, particularly since the implications of corpus-boundedness need fuller elaboration. Similarly Hayley Davis’s gathering of informants’ responses to presented

Integrational corpus linguistics 179 words and phrases constituted a corpus, too. In neither case, it can be argued, was the point of view of the actual language-user lost in the aggregative corpus design and corpus analysis. But the user’s viewpoint can be attended to by many different means. It seems reasonable to say that a significant part of a Daily Telegraph-reader’s experience of the euro, as a sign with an evolving history, subsists in his reading about the euro in that daily newspaper – and perhaps particularly in his reading of various commentaries therein on the UK’s relations with the euro, on the desirability or otherwise of joining the ERM, and so on. Furthermore it is assumed that the valeur and signification of the euro, for writers and readers of the Telegraph, is by no means fixed and unchanging but, on the contrary, critically unstable and open to revaluation and renewal (anti-substantialism, in the spirit of Collingwood), and perhaps particularly importantly so, since so much of a political and economic nature is at stake. Integrationist-minded corpus-analysis cannot be entirely congruent in its assumptions with those that normally underlie orthodox historical linguistics, corpus linguistics, and translation studies. For instance, there can be no invoking and privileging of the synchronic over the diachronic: just as there is no stable état de langue for any speaker or at any point in time, so there is no determinable état for the euro sign, no representable map or featureinventory or relations-network, associative and syntagmatic, which might capture ‘what euro means’ for British speakers, as a group at one point in time, or even for one British speaker outside particular occasions of interaction. And yet we do plan for future events, make fixed deals at t1 about what things at t2 would mean. People say things such as the following: ‘If I had to vote for or against adoption of the euro in 12 months with the domestic base rate at 3% and if there were then no actual or foreseeable major world crises, and if [etc., etc.] . . . . I would vote yes.’ That is a current and specific determination, of course, not an anachronistic report of a future one: cotemporality cannot be escaped. It is not the case, however, that Britons are debating whether ‘euro’ or euro should come into existence at all. At the Maastricht Treaty and by implication in earlier agreements also, the UK government agreed on our behalf that this should happen. Clearly what has been particularly debated in the United Kingdom over the last couple of years (indeed, since Labour returned to power in May 1997) is not whether euros and cents and a unified ERM should displace previous arrangements across a coalition of EU states: it would make little sense for people in Britain to pretend that somehow none of that had come to pass, and that the local currency one used in France continued to be francs and centimes, that in Germany was still marks and pfennigs. Still, as far as namings are concerned, there seems no compelling reason why such regional or national namings of the euro might not emerge, like regional dialectal terms for pound or dollar. Indeed, relative to those national communities now using the euro, differences may already and unsurprisingly be emerging as to just how to pronounce the words euro and cent

180 Michael Toolan (German options include [oiro] and [juro], [tsent] and [sent]) and as to whether these inflect in the plural. When Heinrich Pfandl from the University of Graz used the Linguist internet discussion list to canvass subscribers on these matters (reported on 16 January 2002), concerning his speculation that in establishing the cent as the hundredth part of a euro, the EU might be forcing Europeans to wrench their phonologies to fit an American English procrustean bed, he got a vast range of responses, reflecting regional and national specificities. One respondent, for example, reported as follows on the situation in France: There is still a fluctuation/variation between ‘cent’ (pronounced [sent] or even [sents] to avoid homonymy with the word for ‘hundred’), ‘eurocents’, ‘centimes’, ‘eurocentimes’, ‘centimes Euro’. Actually, directives in France ask the public to use (in cheques, for example) the old term ‘centimes’ to avoid confusion with ‘hundred’. Indeed, a number of commentators have suggested that there is no compelling reason why hundredths of a euro should not be called Pfennigs in Germany, Groschen in Austria, penniä in Finland, centimos in Spain, and p (or is pence making a comeback?) in the United Kingdom. And such national and even regional diversity could go further. For example, there seems no overriding reason, rooted in financial or logistical grounds, why English euro banknotes should not be printed in the size, colour, and with on one side (if not both) an identical range of pictographic displays as are currently found on the five, ten, and twenty pound notes. Nor is there any overriding reason why, in England, the currency might not return to being called pounds – just as some Americans routinely call their dollars bucks, Canadians have their loonies, and so on. Mutatis mutandis, similar local accommodations could apply in Scotland, Wales, and other nations within the European Union. Only the principle that all euros are identical in value and freely interchangeable, whatever the local name and format, is, presumably, sacrosanct. The other kinds of convergence, of name and printed notes, are matters of policy and collective aspiration, and clearly only partial rather than total. But even if such diversity of name and form signifiers flourished, no doubt there would be pressures towards convergence on a single name, for the purposes of international banking, with euro preferable to equivalent terms (such as pound, mark, franc) on policy grounds. Anything other than the most cursory sketch of what I take to be the theoretical foundations of history – accounts of the past – is impossible here. I have space only to suggest that integrational principles assume that with every passing moment we apply ourselves communicatively to our ongoing activities and situations (concomitantly making, modifying and renewing our language) as we think best fits. Or by and large we do, not ignoring bouts of irrationality, destructiveness, fantasy, obsession, addiction, fear and animosity. But the latter are far from the norm, and ordinary people do not use

Integrational corpus linguistics 181 and trust language as if such bouts were the norm. An important part of that everyday sensible cooperativeness that integrationism postulates is the assumption that we do not fully understand our present circumstances. Nevertheless we can usually cope with those circumstances, make sense of them, make plans, and generally rely upon an implicit model or narrative of what is going on in our lives and what it roughly means (but not what it definitively or conclusively means). If so much is true of our relation to our own present day, so much the more is it true of our relation to the past, which, in similar fashion, we can make sense of as best we can with all available and deemed-relevant resources and indications. And certainly if we follow Harris in regarding languages as ‘second-order social constructs of an intrinsically open-ended, incomplete and variable nature’ (Harris 1993: 321), then it follows that a corpus-driven account of dominant tendencies in the use of the euro in parts of the British print media, a commentary on what the valeur and signification of the phrase might be (assuming that it was treated or made a sign by those composing the texts in which it occurred), a very local and partial history of a selected group of ‘moments’ from the process that a language is, will be equally open-ended, incomplete and variable. Whether it is ill-advised or pointless is another matter, and depends in part on whether there is any sense or profit in making situated determinations about intrinsically indeterminate and motile phenomena. Everyday experience seems to answer that question resoundingly in the affirmative: at least pro-tem, for present purposes, interim determinations seem to be the rule not the exception. This is what doing linguistics, or doing history, in practice invariably entails. This argument for the value of corpus analysis of euro usage as reflection and projection of its shifting contemporary history is supported by linguists’ renewed interest in a post-Firthian study of collocation. Here the pioneering work has been done by Sinclair: he has emphasized ‘the idiom principle’ that underpins much language production and interpretation, against the categories-and-combinatorial-rules assumptions of early and still-dominant transformational-generativism. Where the latter proposed that language fluency entailed control of a substantially innate syntactic repertoire, that governed the combination and processing of a huge inventory of free-standing lexical items, the collocational or idiomatic approach collapses any sharp lexicon and syntax division (questions any assumption that these are two distinct ‘components’ in the user’s head), and rather attends to the many ways in which a language’s commonest vocabulary, in corpus-recorded actual usage, reveals patterns of co-occurrence and co-predictability – these patterns being of a kind that are multi-word but not particularly sensitive to syntactic or phrasal categories. Instead, the recurrent collocational behaviour of many common lexical items seems to be a reflection of semantic considerations, not syntactic ones. And, equally importantly, such semantic collocational clustering is usually quite relative and not absolute: it is a matter of, for example, a lexical phrase like with/to the naked eye tending to occur whenever

182 Michael Toolan the word naked occurs, in some significant minority of all cases, and for the larger construction strongly tending to carry or implicate a particular semantic stance. Important discussions of the complex of issues involved, but with clear divergences of emphasis on particular issues, include: Sinclair (1991, 1996), Francis (1993), Moon (1998), Hunston and Francis (2000), Stubbs (2001a), Hoey (2002) and Teubert (2001). In Stubbs’ words, ‘All the most frequent words [of English vocabulary] have strong phraseological tendencies. That is, they have a strong tendency to collocate with restricted sets of words.’ (2001a: 60). They display idiomaticity without being so fixed as to be idioms (Moon 1998) and therefore entail a more effortful learning than would be required if they approximated monosemic single words in the way that idioms do. These extended strands of lexis need to be treated, in their use and processing, more in the way that polysemous words were hitherto assumed to be treated. Furthermore, the idiomaticity may often be genre- or register-sensitive. Sinclair points to one of the strongest potential implications of accepting that meaning is largely phrasal, with several words co-selected in the course of discourse-production: ‘The idea of a word carrying meaning on its own [can] be relegated to the margins of linguistic interest, in the enumeration of flora and fauna for example’ (Sinclair 1996: 82). The consensus of these views is not quite that set phrases are the rule in language activity, but certainly that multi-word strings, reflective of an underlying semantic prosody, are a significant part of what demonstrably counts as knowing a language well. And as Joseph (2001) has pointed out, a corollary shift in conceptualization of language competence is required to match this new picture: a language faculty in the mind, storing lexical items in some sense separately from each other (to be combined via parameterized principles, or by minimalist or other means) looks less plausible. Instead knowing a language and how to use it would seem to be rather closer to other repertoires of knowledge that seem simply part of memory (also adumbrated in Toolan 1996: 176–179 and passim, where it is argued that a proper valuation, in linguistic theory, of the role of each individual’s memory and imagination, allows a defensible place to mentalism). One of Firth’s examples, cited by Joseph, is the collocability of night with dark and vice versa, Firth suggesting that as a result part of the meaning of night is that it collocates with dark. Similar to this is Stubbs’s example of assault and its common collocate, physical. In such collocations, it is argued, the meaning is distributed among or across the items involved, it is collectively evoked or instantiated – and that is part of the reason that, despite one or more of the terms often apparently logically superfluous (nights are typically dark, assaults are typically physical), speakers persist in using the multi-word phrasing. Incidentally it is intriguing to find, in his description of Orwell’s 1984, that the new mechanized and standardized public English deployed in that futurist regime, is said by Firth to cause ‘a prevention of originality’ (Firth 1957[1950]: 188). It is curious that Firth uses the term

Integrational corpus linguistics 183 prevention, which now at least has a negative prosody ( prevention of crime, of terrorism, of tooth decay), in conjunction with what is evidently a desired attribute, originality. By way of a very preliminary probing of patterns of collocation, and thus indirectly of meaning, of the word euro in parts of the British print media, I created four small corpora of news and comment stories (not sport where, as noted earlier, references to Euro competitions evidently intend no reference to the currency, monetary union, or even the European Union more generally). These corpora comprised large samples (between a third and one half) of all stories, in The Independent and the Daily Telegraph respectively, of news or comment items in which euro figured in the story headline, for the period from 1 October to 31 December in 2000 and 2002. Each sample, containing roughly 50 distinct stories, amounted to approximately 18,000 words. Thereafter I used Wordsmith Tools, a corpus linguistic analysis package authored by Mike Scott (University of Liverpool), to prepare a concordance of all occurrences of euro, and of the most frequent collocates, and of the multi-word clusters into which the word enters, for each of the four samples. My guarded assumption is that, taking the four samples collectively, concordancing and collocate-identification of the combined corpus may give some indications of the contextual ‘meanings’ associated with euro in these two newspapers in those two three-month periods (and, putatively, across the period of time that separates the two periods from which samples were taken). Additionally, I was interested to see whether there might be some slight collocational and clustering differences between the two newspapers’ samples, and even between the two samples, two years apart, from the same newspaper. The picture that Wordsmith produces, of the most frequent collocational partners of euro in these corpora, suggests that, in The Independent, among the closest lexical company that euro keeps are pro and joining, Britain, anti, referendum and join (preceding euro), and entry (following the item euro). And equivalently, in the Telegraph corpus, the closest lexical company (before euro) is pro, joining and join and (after euro) zone and sceptic. Does this mean anything? I suggest that if the logic of the idiomatic-phraseological approach to grammar and meaning proposed by Sinclair, Hunston and Francis, Stubbs, and Hoey is adopted, it could mean a great deal. Because the lexical-phraseologists’ claim is that where certain words or lemmas co-occur on a significant frequency of occasions (such as 10 per cent of occurrences of one of those words), we are often justified in arguing that meaning is carried by all the words in the phrase jointly, with all the words co-participating in what proficient users know to be one of the senses commonly associated with the designated node word of the phrase (this phrase often not constituting a wellformed phrase by traditional grammatical standards, and its node word similarly often not the grammatical head of the phrase). If one were to prepare a non-traditional lexical guide to the language (as displayed, with all its limitations and misrepresentations, in a controlled written corpus), one would want to record the frequent apparent meaning of the node word when

184

Michael Toolan

collocated with each of its noted collocates, as among the most important ‘meanings’ of the word in question. Now applying these assumptions to study of a new word such as euro, and its recurrent lexical company, is clearly risky in that one must do so without any certainty that current patterns or collocates will prove to be longlasting. There is potential contrast then, with the way that, apparently, we have been speaking and writing of nights as dark nights, or of deriv(ing) as frequently linked to and co-interpreted with benefit (or a near equivalent: pleasure, satisfaction, comfort), for many generations. But the difference is chiefly that between attempting a longer-term history and a short-term one, although study of euro may also be affected by the fact that the word is a much more specialist lexical item – outside the notional ‘common core’ vocabulary of English – than those which lexical-phraseologists usually study. In this final section of the chapter, I will present some of the kinds of configured information (‘data’) that corpus analysis of euro-usage can give rise to, and in the process I will touch further on the difficulties that persist – in particular the difficulty of deciding what to make of such information. I began by creating four relatively small corpora, of articles gathered from two British newspapers in the final quarter of the years 2000 and 2002, in which the word euro appeared in the headline of those articles (and might therefore be expected to figure also in the body of the article, perhaps also lying close to the main topic of the article). Compared to the 82 stories in the Independent in Oct–Dec 2000, 64 stories had euro in the headline in the Independent in Oct–Dec 2002; in the Daily Telegraph during the earlier period there were 101 stories with a euro in the headline, and 55 in the 2002 period. (By contrast with both of these newspapers, the tabloid Sun had 95 such stories, in both the 2000 and 2002 quarters.) In the earlier samples, a sizeable minority of the Telegraph euroheadlined stories considered the then-much-debated proposal to establish an EU defence force (dubbed by the press a Euro-force or Euro-army), while there were only a very few stories in the Independent in Oct-Dec 2000; no such stories or usage arise in the euro-headlined stories in the final months of 2002. Such frequencies may suggest that whatever else may be said about the word, it is already less newsworthy in 2002 than it was only two years earlier. Furthermore, in the 2002 samples by contrast with the 2000 ones, as many as half the headlines use the form euro with nothing to do with the European Union or the EMU in mind, but in sport story headlines, where Euro seems to be intended simply as a clipping of European. Thus a story headed VAN NISTELROOY OUT OF EURO DECIDER (The Independent, 15 Oct 2002) is about a soccer star forced out of a European match through injury. And the tendency to use euro in headlines, in reference to sport and not EU fiscal and currency arrangements, is seemingly more pronounced in the tabloids, such as the Sun.

Integrational corpus linguistics 185 Here, by way of a beginning, is a display of the collocational patterning that accompanies occurrence of euro in the 36,000-word sample of stories, which also had euro in the headline, from the Independent: Collocational frequency patterns for euro in the Independent, 2000 ⫹ 2002 sample N L5

L4

L3

L2

1 2 3 4 5 6 7 8 9 10 11

THE AND IN TO A IF

TO OF BRITAIN REF’DUM WE

OF THE EURO IS THE ON PRO THE IS JOINING OF AND WOULD THE A ENTRY A JOIN ANTI HAS WORDS TO BUT THIS AGAINST WAS NOT FOR FOR HE IN ABOUT AS BE THAT TO MR MEM’SHIP IT

THE OF BRITAIN TO AND BE THAT

L1

Centre

R1

R2

R3

R4

R5

THE A OF TO THAT HAVE BE

THE OF THAT A IN IS THAT HE

THE A TO IS AND HE

SAID

This Wordsmith-derived display is based on 272 concordance lines from the sample, in each of which euro appears as the node word. This ‘map’ indicates that, for example, pro is the second most frequent word (after the) to appear immediately to the left of euro in the sample, while anti is the fifth most frequent word in the same position (L1), and that neither pro nor anti is frequent in any other position. It is worth bearing in mind that only the first few rows are likely to be of significance: that, for example, the fact that joining is in the third row, thus the third most frequent collocate appearing two words to the left of euro means it is a much more regular co-selection than anti which, though it appears immediately to the left of euro, is only the seventh most frequent item to do so. (joining [word] euro occurs 15 times, anti-euro occurs only 5 times; and pro-euro has 15 occurrences). And here is a display of the collocational patterning that accompanies occurrence of euro in the n-word sample of stories, collected on the same basis as for the Independent, and across the same periods, in the Daily Telegraph: With all the details taken into consideration, the lexical-phraseological approach might justifiably suggest that the corpus evidence in the Independent and the Telegraph in turn suggests that the euro was construed by writers in these newspapers in the years 2000 and 2002 as something to be pro or anti or against, as something that one might join, and as something with a zone and skeptics, and associated with Britain and a referendum. At that time and in those newspapers, to mention the euro frequently (in a significant minority of cases) meant mentioning Britain-joining-the-euro or referendum-on-the-euro and, the further speculation would run, these were the recurrent, quasi-propositional

186 Michael Toolan Collocational frequency patterns for euro in the Daily Telegraph, 2000 ⫹ 2002 sample N

L5

L4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

THE THE A TO THAT BRITAIN TO A FOR OF BRITAINAND OF BRITISH IS IN

L3

L2

L1

Centre

R1

R2

R3

TO THE THAT ON SHOULD WITH AGAINST OF

OF THE EURO ZONE THE THE JOINING PRO IS BE IN ON ON AND NOT TO JOIN OF SCEPTIC IS IT TO A WAS IN A THE 2000 IT A WORDS AND THAT ANTI THE THAT BY FOR IN WAS IS IN WOULD WORDS OF WITH WILL TO AND HAS HAS AGAINST BUT AND A HE MR MPS IT AS REFERENDUM SCEPTICISM BY

R4

R5

THE A THAT WOULD BY BE TO IT BE IN

THE AND A TO WORDS IN THAT

scenarios often associated with the use of the form: these were its standard public construal. Using Wordsmith to identify the most frequently occurring four-word clusters in the corpora confirms the same picture: the six most frequent clusters, in The Telegraph corpora, with frequency indicated, are: 1 2 3 4 5 6

TO JOIN THE EURO SHOULD JOIN THE EURO THE EURO WOULD BE ON JOINING THE EURO BRITAIN SHOULD JOIN THE THAT JOINING THE EURO

9 7 6 6 6 5

while the five most frequent in The Independent are: 1 2 3 4 5

REFERENDUM ON THE EURO TO JOIN THE EURO OF JOINING THE EURO LAUNCH OF THE EURO THE LAUNCH OF THE

6 5 5 4 4

(Care is needed in interpreting such clusters: here items 2 and 5 in the Telegraph overlap, as do 4 and 5 in the Independent.) In making the earlier claims about this ‘local’ history of euro I do not assume that these observations are proof against counter-evidence, or that they capture the main elements of the meaning of euro for the writers and editors involved. But I would also suggest that these tendencies in the co-occurrence of these written forms are not without significance. Some

Integrational corpus linguistics 187 corpus analysts might argue that part of their significance lies in the fact these are not patterns that stare every reader of the newspapers in the face, but only such as emerge under the agency of the analyst. On the other hand, these collocations (e.g. of join with euro) are ones that the assiduous and regular reader of the Telegraph or Independent will have encountered, and to that extent this kind of corpus analysis is aligned with, not removed from, the everyday experience of a sizeable number of ordinary British readers. So the multiple instances of usage of euro subjected to classification here are treated as samples from a multiplicity of situationally and generically related transactions (involving newspaper and reader), and from these transactions, it is postulated, shifting values and significations emerge. It will be objected that whether particular words are disproportionately frequent co-textual partners of euro may be or contribute to only a very small part of the value or history of the word, of ‘what happened’ to euro for British people circa 2000. And the objectors will be right. The immediate verbal contexts within which euro was used, in those newspapers in those months, is only a small part of what, from occasion to occasion, readers may have taken as communicationally relevant in their determination of what the word variably signified and entailed, on all these distinct occasions. But I take identification of adducible evidence of just a small part of the history as a useful step, which attempts an integration of its own. The attempted integration is of two ideas: Harris’s second basic axiom of integrational semiology, that ‘the value of a sign is a function of the integrational proficiency which its identification and interpretation presuppose’, and Firth’s emphasis on the importance of habitual collocation to ‘one level of meaning’ (among, it is implied, several levels). To adopt Firth, one of the meanings of the euro [in the Telegraph/Independent, 2000–2002) is its habitual collocation with join (and pro, anti, zone, etc.). The habitual collocation is nothing like so dominant, or unchallenged by counter habits, as in Firth’s you silly ass example. But then euro is a key term in a vast domain of fiscal policy and national identity and culture, where uncertainty and revision are still uppermost. In these respects it and its collocates are still unlike the almighty dollar, the dear old pound, and similar idiomatized phrases, which encapsulate and project, in little, particular histories of those currencies.

12 Language and prehistory Geoffrey Galt Harpham

What we might call the thought of language has been for over a century a most important concept for many thinkers working outside the discipline of linguistics. Indeed, the twentieth century has with justice been called the “century of language,” a period during which the concept of language has enjoyed an absolutely singular preeminence not only as an object of fascinated study by specialists, but also as a kind of philosopher’s stone capable of grounding or anchoring other fields far removed from linguistics. Indeed, the thought of language has been more than just another concept, and has almost appeared to be a partially hidden master-concept whose features, if they could be determined, might solve a host of problems within other disciplines, problems unsolvable within the disciplines themselves. The belief in language as a master-concept has had a profound impact on the intellectual history of the past century, and has survived even the remarkable crumbling or erosion of confidence that has afflicted the discipline of linguistics itself over the past generation. Perhaps the two discourses or areas of academic inquiry most indebted to the thought of language are the study of ethics and the study of ideology. Both try to make good on the Enlightenment claim to identify universal principles, and both generate, in pursuit of this claim, problems they cannot solve. Ethics posits the law but has never been able to define or locate it beyond dispute. Attempts to do so by referring to reason, power, communal norms, or some set of especially worthy virtues expose themselves to challenges they cannot answer: how can you say that X is the principle of principles, the most virtuous virtue; on what basis is your account grounded? The history of ethical discourse is a sequence of challenges in which some new principle, denied by the current paradigm, rises up to assert its claims, only to suffer the same fate in its turn. Still, the effort continues to find a ground or adequate explanation for the distinctive quality of universal imperativity that distinguishes the modern form of the ethical from other kinds of law. And from the beginning of the modern era in ethical thought, theorists have attempted to explain their mysterious subject by turning to language. How do we make ethical decisions; how does the law operate; how does the law bind us all? Since Hume,

Language and prehistory 189 the answers to these questions have been sought in one place: language. So, for example, if we posit some intimate relationship between the ethical law and language, then we know why the ethical law transcends all lesser laws, why it pervades and regulates all human affairs. It does so because it is established somehow in language – which binds all humans. Theorists have posited all sorts of relationships between language and the law – as a “guide to virtue,” as a force for virtue, as a rebellious force undercutting all oppressive laws, as the critical factor in community, or as the structure of the unconscious – but in a movement that has been gathering force steadily especially for the past century, progress has been made in the field of ethical theory by theorists discovering some new relationship between language and the law. In the other distinctively modern intellectual field, the study of ideology, the genealogy of the association is even more direct, since the inventor of the term, Destutt de Tracy, was working on a massive Grammar at virtually the same time he was composing his masterpiece, Projet d’éléments d’idéologie (1801–15). Although the meaning of this term changed radically, first with Marx and Engels, where it became “false consciousness,” and then with subsequent theorists who treated it like the background presuppositions that filter through a culture and constitute common sense, the link between language and ideology has remained amazingly constant and indeed is reaffirmed with each new theoretical advance, whether by Volosinov, Gramsci, Raymond Williams, Louis Althusser, Stuart Hall, Ernesto Laclau, or Slavoj Zizek. For only some such link can explain how ideology pervades all cultures, everywhere, how it structures thoughts without our permission, how it provides the terms by which we judge everything, describe everything, know everything, perceive everything. It does all these, ideology theorists agree – and it is perhaps the only thing they agree on – because ideology is borne by the medium of language, determined by language, or functions as a language. There is, of course, an opposition. Chomsky, for example, argues that the nature of language gives us direct access to a human nature characterized by freedom and autonomy, and so constitutes objective evidence that we should not be influenced by ideology, which always works against freedom and autonomy. But Chomsky agrees with his opponents on a deeper premise, that we can know ideology by examining language. Why has modernity, including postmodernity, so consistently turned to the concept of language as the solution of its problems? What’s the attraction? I think that the ancient prestige enjoyed by language as the defining characteristic of the human species is the secret. This identification of language as the mark of the human difference can be located in the Bible, where language is given to man so that he may name the beasts. But it has been reaffirmed in an entirely secular spirit by a long tradition of philosophers and others, who have affirmed that language is a human characteristic, the cornerstone of the humanistic tradition. To know man, in this tradition, you must know language.

190 Geoffrey Galt Harpham If the roots of modern humanism can be traced in the Bible, the Biblical tradition is clearly more interesting than we may have thought. In fact, it is far more interesting than I have begun to indicate. The Gospel according to John begins by associating the Word not with Man but with God, in a verse whose rhetorical complexity seems to strain to say the unsayable and even the unthinkable: “In the beginning was the Word, and the Word was with God and the Word was God. He was in the beginning with God; all things were made through him, and without him was not anything made that was made” ( John 1: 2–3). By positioning the Word both inside God, as his essence, and outside, or “with” God, this passage manages to suggest with extraordinary compression two more themes that have been carried forth in the modern conception of language, that language is transcendent, even metaphysical, and that it is a tool or implement, a thing. Language is, in other words, nonhuman in two senses, one higher and one lower than man. “According to the Bible,” then, language represents divinity, the human difference, and thingness or instrumentality. The word language does not occur in Genesis or the Gospel of John, and what I am suggesting is that the Biblical contribution to the thought of language should be considered to encompass all three – essence, humanity, and instrumentality – and all three together. This is, at least, the way that the thought of language has been constructed in modernity, as a complex of moving parts, a variable ensemble in which the human is enfolded by two forms of inhumanity, and defined by the relations established among them. One of the many curiosities in the concept of language is that language is not the singular form of the plural languages, but rather the larger theoretical category in which the others are gathered: the singular is greater than the plural, and, although they are both speculative theoretical constructs (a surprisingly difficult proposition to convince people of ), they are different in kind as well. Languages are coherent even if they do not somehow contain all three of these factors. One can speak of the grammar of a natural language without summoning up the human and two kinds of the inhuman; one can write the history of Portuguese or Russian in a more or less empiricist way, without such conceptual overtones. But what really defines modernity is the aspiration to universality, and so the most characteristic expression of the modern spirit is a deep interest in language as such, the very nature of language, language alone. And language alone is necessarily marked by all three. It is not difficult to isolate these three elements in what passes for theoretical common sense outside the discipline of linguistics. While it has, for the most part, dropped the suggestion that language is divine in origin, modern thought, in the work of Heidegger, Paul de Man, Michel Foucault, Jacques Lacan, and others, regards language as a superior force to which mankind is subject, even the “slave,” as Lacan says. “Language is more powerful than we,” Heidegger says in this spirit; “and therefore weightier” (31). Or as the anthropologist Derek Bickerton puts it, language is “the direct root and cause of all those mental characteristics that distinguish us from other creatures.”

Language and prehistory 191 It “puts its own spin on the way we execute biological imperatives”; indeed, it “created our species, and created too the world that our species sees” (156, 157). Especially since Saussure, we have also learned to regard language as something with which things are made, a tool whose features can be described. The cultural-intellectual movement that goes by the name of modernism was, in some accounts, launched from Saussure’s discovery, on the basis of which it began to take a linguistic turn that spread from linguistics to the arts, to philosophy, anthropology, psychoanalysis, literary theory – a dizzying series of turns from which we haven’t yet stopped spinning. Beneath Saussure’s propositions, and thus beneath much of the intellectual history of the past century, is the belief that language is a “concrete object” with a determinate nature. And, third, I have already mentioned the continuation of the long humanistic tradition that identifies humanity with the gift of language, a tradition continued in our day by Chomsky and Jackendoff in linguistics, and by many others in other disciplines, prominently including Jacques Lacan, Michel Foucault, and others for whom the redescription of the human subject as a discursive construct became an article of faith, if not of war. It is interesting, I think, to trace the evolution of positions taken in the Bible in modern thought, but far more compelling is the hypothesis that each of these three elements – essence, humanity, and instrumentality – is a kind of requirement in any modern theory of language. Demonstrating this fully would take far more time than I have here, but I can point to some intriguing indicators. Take, for example, the discourse of Saussure, which I have described as the most emphatic statement of the proposition that language has a determinate nature as a concrete object, something not to be confused with metaphysics, humanity at large, or the organic world in general. For Saussure, language is the sign system, whose operations determine the form of our utterances and in many ways override our individual agency. For providing an account of language as a thing susceptible to rational investigation, Saussure is widely considered to be not just the founder of modern linguistics, but perhaps the most authoritative theoretical voice in the ethos of modernism and postmodernism. It is one of the freaks of intellectual history that Saussure continues to be the fountainhead of wisdom on the subject of language for so many of those outside the discipline of linguistics, while within linguistics, his stature is more diminished even than that of Freud for psychoanalysts. Lacan, Laclau, de Man, Derrida, Zizek and many others refer to the sign system as the authoritative theoretical paradigm for the study of language, chiefly for ideological or political reasons. He is particularly compelling because his account of systematicity corresponds with certain rationalist tendencies within cultural modernism, and his notion that the sign is a differential entity without positive content accords so nicely with a postmodern emphasis on representation over ontology. He has been remarkably useful in a theoretical climate dominated by antihumanism. But how solid is that rock?

192 Geoffrey Galt Harpham We must begin by noting that the place of the human in Saussure is not, after all, precisely empty, for the system does require human speakers, in the same sense that a wave requires water. And the content of the sign is “psychological,” lodged in the brain. But perhaps the most telling evidence for the unspoken presence of the human can be found in the explanatory metaphors by which Saussure explains how language “chooses,” “interprets,” and “decomposes” its elements. At such moments, language seems to possess superhuman powers, exercising an agency denied to mere language-users. But this agency is plainly modeled on human agency: God does not “interpret” or “choose.” And so, through what Saussure calls “indispensable” metaphors, the “missing” elements of the superhuman and the human – God and Man – are smuggled back into a discourse that is explicitly designed to banish them. Still, these elements are unstressed in Saussure, and this lack of emphasis or visibility creates a kind of tension in the concept to which Chomsky’s revolution may be considered to be the most massive response. Chomsky’s correction takes the form of attempting to identify a series of formal computations and manipulations in the brain that produce the grammatical system, and then to characterize these as the kernel of human nature itself, from which point he mounts a series of normative arguments about ethical rights and responsibilities, and the optimal form of governance. He tries, in other words, to establish linguistics not just as a human science but as a science of the human, with no conceptual overtones at all. Chomsky insists on an absolute distinction between humanity and all lower forms of life. In the course of a commentary on Rousseau, he says, for example, that the thought processes of beasts are subject to “mechanical” explanations: the beast is “merely an ingenious machine, commanded by natural law. Man’s freedom and his consciousness of this freedom,” on the other hand, “distinguish him from the beast-machine” (1987: 145). Humans share with primates certain elements of a conceptual system that involves object identification and the capacity to attribute intent to other organisms, but the capacity to conceive of a “discrete infinity” – the most obvious examples being the infinite number of natural numbers and the infinite number of sentences that can be generated by a grammar – is, Chomsky says, an exclusively human trait. When, at some point in the distant past, the conceptual system hooked up with the computational capacity, the result was an ability to formulate an infinite number of new sentences over an unbounded range. And this new capacity, Chomsky says, produces “a totally new organism” living in “a total new world” (1982: 21, 22). And so, while language is an evolved capacity, it constitutes “the essential difference between man and animal” (1966: 3). Like Descartes and Rousseau, Chomsky believes that humans are “metaphysically distinct from non-humans,” and that language proves it (1979: 92). The pursuit of political goals based on the features of human nature revealed by language – freedom and autonomy – will lead, he anticipates, to a world in which “animal nature is transcended and human nature can truly flourish” (1975: 134).

Language and prehistory 193 How successful is this radical effort to contain language entirely within the domain of the human? According to him, the case is self-evident, but feeling this way, he has not made it in much detail. The argument is more exposed to air and light in the form that it takes in the work of one of his leading advocates, Stephen Pinker. In his entertaining book The Language Instinct, Pinker argues that Chomsky is basically right when he says that language is instinctive with us; the case only requires a bit of supplementation in the form of the assertion that, like other aspects of Homo sapiens, the language instinct is the product of evolution: it progressed from simple to complex forms to the point where grammatical structures are now as natural to us as web spinning is to spiders. Pinker notes, but professes not to understand Chomsky’s resistance to this obvious point, which is articulated at a rhetorical and conceptual level designed to appeal to the common reader. One of them knows something the other doesn’t – but which one? Before we can answer this question, we have to pursue the argument a bit, through Pinker and beyond, for he, too, omits many of the details that might make a difference. Pinker claims that the language faculty evolved very slowly within the human brain, in obedience to the law of natural selection. Even in its most primitive forms, he says, speech must have provided a highly efficient way of sharing complex kinds of information. The evolutionary advantage begins small, he says, in the “first traces of language,” but grows, through the processes of natural selection, ever larger (352). What about these “first traces of language?” When – and what – were they? Indifferent to particulars, Pinker suggests that they might have begun to appear as early as Australopithecus afarensis, the famous “Lucy” fossil, from about 3.2 million years ago; or perhaps even earlier, from the time of the human-chimp split, about 5–7 million years ago (352–53). And here we may be able to see why Chomsky so vigorously resists the evolutionary argument, and why the affable Pinker, who seems to think he is merely helping Chomsky out by putting his magnificent but incomplete theory on a firm evolutionary footing, providing a component that Chomsky really ought to have provided himself, may actually be Chomsky’s worst nightmare. Chomsky has not demonstrated a sustained interest in chimp nature or chimp rights. Primates are suffering unjustly all over the world, but the only ones he gets exercised about are those with a language module in their brains. For him, the thought that language had first traces that might be retained by today’s primates would constitute a decidedly unwelcome distraction. Worse than distracting, from Chomsky’s perspective, would be Pinker’s functionalism, which stresses the evolutionary advantage represented by the ability to share information, rather than the capacity for individual “creativity” that grounds Chomsky’s anarchist-libertarian political stance, or the instinct for “freedom” that anchors his philosophical arguments. If Pinker does not register these concerns, the reason may be that he has no evident philosophical or political agenda; although he insists that speech is restricted to the human species, there is, for him, no cost in tracing language

194 Geoffrey Galt Harpham back to pre-human origins, nothing worrisome in relinquishing the notion of “Human Uniqueness,” nothing at stake in conceding that language “for all we know . . . could have had a gradual fade-in” (347, 346). As long as “the modern adult language instinct” is granted to be “unique in the modern animal kingdom,” that’s all he cares about (366, 342). His cavalier attitude obscures, however, the true damage of the evolutionary argument to Chomsky’s account of language, a damage that emerges even more clearly in the work of others who have pushed this argument more aggressively, with more detail concerning those first traces. Their work clarifies the real, if unrecognized and even unintended, effect of Pinker’s intervention, which is to detach language from Chomsky’s normative presuppositions by loosening the human grip. If, as Chomsky contends, sophisticated grammar defines language, then human beings appear to be in sole possession of language. But an inquiry into the evolution of language complicates this position considerably. First, by introducing evolution, Pinker actually introduces a host of questions whose solution can only be non-scientific. By asking us to imagine, for example, the processes by which noises could become increasingly complex in their organization, Pinker embarks upon a speculative narrative about prehistory and the “missing link,” a narrative that takes us out of science and into philosophy and even into mythology. Second, Pinker forces us to ask what primates can actually do in the language line, and the answers to this question threaten to undercut the most essential premise of Chomskyan linguistics, that human being, human rights, and human political and ethical norms are predicated on the exclusive human possession of grammar. A good deal of recently accumulated – and, to many, controversial – evidence suggests that even in the absence of a computational capacity, primates are capable of grasping rudimentary grammatical distinctions. One of the most aggressive apostles of primate intelligence, Sue Savage-Rumbaugh has claimed to identify a number of rudimentary grammatical distinctions in the “language” of her prize bonobo chimp Kanzi, which is produced by tapping on a “lexigram,” a keyboard with dozens of symbols. These constitute, she argues, a “cognitive substrate” to language, not a fully developed grammar but a preliminary structuration best regarded as “proto-grammar” that must have existed in the common ancestor to humans and chimps (Savage-Rumbaugh and Lewin 165). According to her, Chomsky is right when he says that the language capacity is innate, but wrong when he insists that the structure of language is pre-wired only in the human brain (163). For Chomsky, as we have seen, innate means unique to humans; for Savage-Rumbaugh, who holds human beings to be “African apes, of an unusual kind,” innateness signifies the biological continuity of the human species with its primate ancestors (280). Others have made even more aggressive claims for the primate command of proto-linguistic skills, based on gesture, or visible bodily action – a cultural, rather than a genetic endowment, at least during Pinker’s “first traces.” Humans have only been able to speak for a few hundred thousand years,

Language and prehistory 195 a very short time, as Philip Lieberman points out, for something as complex as grammar to be hardwired into our brains. Grammar, or its component parts, must have been loaded into the system much earlier, during the long run-up to anatomically modern humanity when we were gesturing with increasing expressiveness and urgency. It makes perfect sense, from this point of view, that some of the most impressive proto-linguistic accomplishments yet recorded should come in the area of hand signals. One chimp, the legendary Washoe, reportedly learned over two hundred signs in American Sign Language; and while this claim has been disputed, it is generally conceded that some chimps have mastered the use of signs to a certain extent. And this small encroachment on the exclusive human possession of language, including grammar, makes the evolutionary argument even more compelling, and compelling in a slightly different way, than Pinker seems to want to realize. According to Pinker, “the first steps toward human language are a mystery” (351). He is willing to leave it at that, but surely the evolutionary argument itself obliges us to inquire into that mystery, to construct some account of the origins of language, the straining and groping of primitive organisms for the regularity and repeatability that secure the possibility of meaning. One who has made the attempt, Robin Dunbar, argues that one of the very earliest forms of proto-language is primate grooming. When primates pick the burrs, bugs, and scabs from each other’s fur, he says, they are not simply providing a cleaning service. Nor is the utility of grooming confined to the production, in the groomees, of endogenous opiates, enhancing as these are to one’s quality of life. What grooming really does, according to Dunbar, is establish an elementary social bond that mitigates the threat of predation and facilitates the development of a sense of common cause in order to ensure the survival of the group. In other words, grooming is a primitive form of communication that meets a number of individual and group needs. As the group grows in size, physical grooming must be replaced by “vocal grooming,” and gradually, with the advent of Homo sapiens, expressive noises replaced the older form of grooming almost entirely, even though the purpose, the creation of a shared social bond, remained the same. Dunbar puts the date for the emergence of language proper at about 500,000 BP. To understand why primate groups became so large, we need to imagine a vast process that began about five million years ago, and this will require us to posit one of those species-narratives whose charming whimsicality, like an animated cartoon where corporeal transformation is the rule, makes evolutionary psychology one of the few academic disciplines capable of reaching a wider audience. We begin at the point when one strain of the great Old World apes, their small and fragmentary groups happily grooming and gibbering away in the rain forest, abandoned their protected life in the trees and established a new, knuckle-walking way of life, first on the edge of the forest, then in the woodland, and then in the open country. A mere 1.8 million years

196

Geoffrey Galt Harpham

ago, Homo ergaster (closely related to Homo erectus) appeared, fully upright if not precisely graceful. Standing in the sun, it shed much of its body hair, keeping only the protective covering on the head and shoulders. Less hair meant greater efficiency in cooling, which enabled the species to become more mobile (Dunbar 107–08). But the crucial factor was the growth, at just this time, of the primate brain: after hundreds of thousands of years of little or no growth, the brain suddenly became 50 per cent larger; over the next 1.3 million years or so, it nearly doubled once again. Then, with the first appearance of Homo sapiens half a million years ago, the rate of growth underwent a further rapid increase until roughly 250,000 years ago, when brains attained modern levels of 1350–1400 cc. Bigger brains, Dunbar argues, enabled bigger groups, and bigger groups required their members to possess the language their brains could sustain. But how did our brains grow? According to one recent explanation, the brain was able to grow because it “borrowed” tissue from the gut, which could only shrink if it began to process meat, rather than leaves and fruits. In pursuit of meat, our ancestors would have spent less time in the trees and more on the ground, becoming ever more dependent on the two-legged life and on each other. The shift to a diet composed primarily of meat, or, as Leslie C. Aiello ambiguously puts it, “animal-based products,” entailed a more mobile and cooperative style of living, and larger and more firmly bonded groups required more complex forms of interaction and communication than had been required in the shelter of the trees. And so, Aiello says, dietary change, and the accompanying morphological and cognitive developments it gave rise to, preceded and propelled the development of language. Language comes out of the mouth because meat goes into it. Meat-eating entails other changes as well, including a longer period of infancy and prolonged mother-infant food sharing. Complex communication, including deception and secrecy, would become not an option but a necessity, a matter of survival. Low as these are, they represent social intelligence of a high order, and social intelligence, Aiello says, “is a fundamental prerequisite for the origin of fully developed modern human language” (29). One major plot element in this evolutionary dream-narrative remains missing. Chimps cannot speak, according to many (but not all) researchers, because their brains are smaller than ours; but even if their brains grew, their large and clumsy faces, with the larynx high in the throat, could not produce language. Such faces are needed to chew leaves, roots, and tubers; but meat can be bitten off and chewed by smaller mouths and teeth – the kind seen in Homo ergaster, the first meat-eater. Once we change diets, we can do with less face than before, and a small lower face becomes an evolutionary advantage. Unlucky prey: their flesh is literally food for thought, helping to make the body and the mind that can hunt it. But lucky predator, whose diet not only nourishes the body but also enlarges the mind and helps create the vocal tract that enables it to communicate with other minds so it can hunt more efficiently, and do other things as well.

Language and prehistory 197 With our cunning lips, our supple and flicking tongues, our low larynxes, we can begin to project produce complex phonemic sequences; we can cluck about our neighbors and hoot at their eccentricities. We can deceive and cheat each other. We can reflect on past dining experiences and anticipate future meals, the difficulties, pleasures, and meaning thereof. And with these and related reflections, the episodic character of archaic life is broken; life becomes a single, continuous thing. We are in a position to acquire natures, rights, destinies, ideologies, ethics – identity in general. We are on the way to forgetting or denying our humble origins, and to becoming Chomskyan man, destined (he insists) to transcend animal nature entirely. Thus the “first traces” that Pinker invokes are evident in primate capacities, and this suggests that language – broadly defined, as it must be – cannot be considered the limit of a “metaphysical” difference. What differences there are between chimps and human beings are the consequences of an enormous evolutionary process, with its contingencies, its many “mistakes,” its dependency upon such events as an immense meteor striking the earth and wiping out the dinosaurs, permitting the development of mammals – its sheer improbability and overwhelming flukiness. And it is this fumbling and experimental process, and not a metaphysical design, that eventually succeeded in transforming the “missing link” into modern man. Once considered the Rubicon between man and beast, language might just as well, under the pressure of the evolutionary argument, be considered the bridge. The ancient triad of God, Word, and man is now being forced to accommodate a fourth term, Nature, which seems not so much to join the others but to contain them, its mechanisms of organic self-modification capable of eliminating all metaphysical distinctions. Chomsky’s identification of language with the human succeeds, then, in producing an integrated entity susceptible of rational investigation, the faculty in the brain that is responsible for our seemingly effortless acquisition of grammatical structures and rules. This faculty is, moreover, freed from the conceptual overtones that have disturbed others theories. There is, in Chomsky, no trace of God, and no suggestion that language can be reduced to thingness, to something separate from human being, like a tool or implement. The only problem, really, seems to be that, as Pinker’s helpful supplement, properly developed, reveals, the containment of language entirely within the evolved figure of homo sapiens has the effect of decomposing the categories of both language and the human. It is as if Chomsky’s unwillingness to situate this core humanity in relation to transcendence and thingness has produced, as a consequence, an unstable and unanchored concept. Part of the charm of The Language Instinct is that Pinker seems cheerfully unaware of the possibility that, by introducing the dimension of prehistory to Chomsky’s theory, he has destroyed it. But the point is that Pinker is absolutely right to think that language as a human faculty demands an evolutionary explanation. Nor does it follow from this fact that language cannot possess normative implications. In his widely

198 Geoffrey Galt Harpham read book The Moral Animal, Robert Wright argues that evolution has been driven by factors that, at first glance, strongly appear to cut across the natural grain, and to moderate if not cancel out the impulses to self-assertion and self-protection that we might associate with a “state of nature.” Language in particular, Wright says, most likely evolved through a set of self-reinforcing “positive feedback mechanisms” of the sort that create evolutionary advantages for intelligence in all species. Such mechanisms involve the selective transmission of nongenetic “cultural” information from animal to animal, and can actually accelerate evolution. The most pertinent example of such a mechanism is “reciprocal altruism,” the capacity, once scarce and fragile but now instinctual across the human species, to forge mutually beneficial relationships. The conceptual power and memory required to forge these relationships take intelligence, and those that have what it takes will thrive at the expense of those who do not. In both the long and the short runs, such moral or nearly moral emotions as affection and empathy work on behalf of the species’ deeper interests. Even in the very earliest stages, Wright argues, language would facilitate and concretize such emotions, and so promote the evolution of a “morality gene” that provides human beings with a highly advantageous faculty of sympathetic fellowfeeling. On this basis, evolution as a whole can be said to be not a blind striving, but a purposive, even spiritually suggestive, progression, one with “religious” overtones. An evolved language faculty can, in other words, still be considered a moral force in the world, but this force does not derive from the brain’s capacity to generate grammatically correct sentences, but from the basic fact of communication, the expression of mutual interest, or mutuality in general, an expression of which primates are fully capable. One of the most mysterious aspects of language is its apparently infinite availability to observation and description, when, in fact, language is a word for that for which there are no words, a chaos not worth thinking about. Not worth thinking about but excellent to think with, the unqualified term language commends itself to us primarily as a proxy for a host of concerns, questions, and anxieties – our position in the order of things, our rights and obligations, our relation to law, our beginnings and ends, the relation between faith and knowledge, the scope of our agency, the extent of our responsibilities, the laws of our being – that otherwise frustrate the desire for a formulation that would satisfy rationalist sensibilities. What we talk about when we talk about language are things that cannot be spoken of in a literal, direct, rational manner. Remarkably elusive as an object of knowledge, language has nevertheless served as a primary vehicle through which normative, ethical, even religious ideas have entered the field of scientific knowledge. Indeed, as I have tried to demonstrate in a larger project in which I have been engaged (Harpham 2002), the intellectual history of the past hundred years consists largely of the displacement of the otherwise undiscussable onto language. As a concept, language is amenable to this displacement because it has no content that is all its own. Like a Saussurean sign, it functions in the system

Language and prehistory 199 without having any “positivity,” any real object or necessary referent. The history of linguistic thought can in fact be considered as a series of attempts to determine the essentially linguistic, and with each new determination, a new object, claiming the name of language, is brought into focus. This may be why linguists from von Humboldt to Chomsky are likely to preface or to conclude their discussions by acknowledging that even specialists are “just beginning” to understand the object, or that their discipline is still “immature.” After all this time, a linguist might become discouraged to think we are still at the beginning. I am arguing that we ought to be discouraged, even to the point of giving up the search for the true nature of language. We ought, in short, to turn our attention to other things that can be observed and described, things about which nontrivial falsifiable statements can be made. I look forward eagerly to tracking the progress of an “integrationalist” approach that would begin by abandoning the very term language as an impediment to knowledge.

13 Bridges to history Biomechanical constraints in language Stephen J. Cowley

Learning to talk What Wittgenstein calls (1958, §242) ‘agreements in judgements’ are necessary to learning to talk. By coming to act consistently with such agreements, babies develop biomechanical constraints necessary to encultured life. Although grooming (e.g. Hinde, 1983; Dunbar, 1996) shows many primates consistently assess status, kinship, what it is allowed and so on, they do so independently of local behavioural dispositions or ‘habitus’ (Bourdieu, 1971, 1991). Human perceptual ecology, by contrast, is irreducibly historical. As infants get caught up in a matrix of conduct, by the fourth month, they begin to participate in culturally appropriate activity. Using primate biology, they find themselves adapting to historically derived practices. Below, I argue that these capacities for consistent behaviour enable babies to enter history. To show how we enter history, I focus on a 14-week-old who does what her mother wants. In so doing, I take an inclusive view of culture that brings out how ‘meaningful’ behavioural patterns permeate the child. To act as the mother wants, I argue, is a culturally-based judgement. Since it defies classic nativist or empiricist explanation, I advocate a ‘distributed’ view. This enables me to stress that phylogenetic and ontogenetic history allows the infant to exploit two brains. At three months, ‘dual control’ provides babies and caregivers with a new way of contextualizing. Thanks to something like the motivational/ affective systems posited by Trevarthen and colleagues (e.g. Trevarthen, 1979, 1998; Trevarthen et al., 1999), infants are prompted, at times, to act appropriately. Biomechanical constraints thus develop in their interests and allow them to meet and, later, anticipate social expectations. Finally, I suggest that required motivational/affective systems derive from selection pressures on an infant-in-a-niche (Laland et al., 2000). Their primate equivalents may have changed as, in pre-historical time, caregiver activity became a cognitive resource. Today, the systems ensure new-borns act as if knowing that social life was shaped around talk. By 14 weeks, social biases and associative learning allow behaviour consistent with historically based judgements. As culture is implanted, infants learn to contextualize, and by

Biomechanical constraints in language 201 the end of the first year act in ways that can be analysed in terms of sounds, acts and goals. For Western adults, they speak ‘first words’.

The social life of babies Behaviourism, genetic epistemology, information-processing and activity theory all examine how babies learn. The theories, however, downplay early social life. Rather than ask how culture permeates babies, they are presented as coming to know and/or learning about ‘the world’. Appeal is made to Vygotsky’s (1978) ‘lower’ cognitive functions or abstract concepts like object, stimulus, response, learning, accommodation, assimilation etc. It is hastily assumed that conditioning and/or inference-making limits development to processes within the organism. Unfortunately, this downplays social learning, active perception and specific biases (Karmiloff-Smith, 1993; Mehler and Dupoux, 1997; Rochat, 2001). Further, as Thelen and Smith (1994) observe, the view throws little light on what develops or how changes arise. Indeed, if cognition is exclusively within the skull it is hard to account for how babies adapt to individual persons. As Rochat (2001) stresses, we fail to ask why change occurs. While developmental psychology and linguistics make little of social adjustments, work by Spitz (1965) and Bowlby (1969) generated a major research tradition. Today, all agree that, by the third month, social events dominate infant life (see, Schaffer, 1977; Bullowa, 1979; Kaye, 1982; Bråten, 1998; Nadel and Butterworth, 1999). Together with caregivers, baby bodies and voices enact and evoke practical understanding. Since events can be explained (by lay persons), they have affective and cultural content (e.g. Bateson, 1979; Trevarthen, 1988; Cowley, 2003). Social development leads human babies into varied cultural worlds. Further, given neural selection (e.g. Johnson, 1997; Edelman and Tonini, 2000), sound and movement sculpt infant brains (Deacon, 1997). Early development, then, exploits behavioural content that, in some way, arises as activity is orchestrated between baby and caregiver. Utterance-activity, or ‘language in movement’ (Piaget, 1962) plays a major role in infant life.

Experience and experiencing social behaviour In a human world, interaction exploits culture, bodies and persons. An infant soon begins to use affective and cultural content in her own interests. Using prepared learning (Wilson, 2000) she sets off predictable activity that relies on biases like those allowing newborns to mimic tongue protrusion and mouth opening (Melzoff and Moore, 1997; Kugiumatzakis, 1999). At first, social behaviour lacks cultural content. As Rochat (2001) notes, the newborn deals with events at its body’s boundaries. Then, around the sixth week, there comes a change ‘almost as great as birth itself ’ (Stern, 1977). In the West, infants become able to have fun. In the ensuing ‘intersubjectivity’

202 Stephen J. Cowley (Trevarthen, 1979), like boxers or dancers, babies and caregivers co-ordinate their doings with joy and finesse. As with other kinds of dialogical activity, they become able to initiate and respond (e.g. Linell, 1990) or to act both retrospectively and prospectively. Recently, it has been suggested that this body-world activity is primate imitation or mimesis (see, Donald, 1991) that goes a long way to making us human (e.g. Bråten, 1998; Nadel and Butterworth, 1999). In this chapter I show how interpersonal dynamics draw on phylogenesis and ontogenesis to produce cultural conduct. Co-evolutionary processes allow babies to use two brains in fitting activity to encultured patterns in the world. Joint behaviour exploits affect-seeking bodies to link neural processes between persons in ways ‘explained’ by what we say, feel and do. Not only does behaviour link babies to cultural content but this also occurs publicly. When a mother wants a baby to smile, we see what she wants. Similarly, if seeking a baby’s attention, the activity persists until she gets it, gives up, or elicits it directly (e.g. by chucking the baby’s chin). Simply, as the baby adapts to caregiver judgements, cultural content shapes her use of adult acts, goals and sounds. From birth, interaction exploits affective and cultural events as a baby is persuaded to smile or, as described below, to hold her tongue. While encultured, the behaviour arises from meshing an adult’s display with a baby’s motorcentric activity. Intrinsic as it is to what the dyad does together, it is irreducible to kinds of act.1 Utterance-activity exploits affect to ensure that the baby’s doings are channeled by cultural expectations. Early social life is multimodal activity that forges joint experience. Further, given contrasting interests, different motives and capacities ensure that the behaviour becomes more orchestrated over time. Joint action locks the infant’s doings to caregiver beliefs and, as this happens, events adjust to local ways of explaining. What an infant does adapts to what precedes, accompanies and follows. As joint activity gains coherency (see, Cowley and Spurrett, 2003), the infant enters history. Early motorcentric activity is mainly biological. Drawing on much shared across phyla the child makes cultural use of patterns heard by species as different as humans, chinchillas and cotton tailed tamarins (Kuhl and Miller, 1978; Ramus et al., 2000). While using perceptual biases that prepare us for learning and/or action, the cultural setting shapes how these function. Sensitivity to rhythmic patterns thus reflects, not linguistic categories, but how primate hearing serves history-prone individuals. By analogy, we can ask how social life affects non-human vocalization. It seems, for example, that prepared learning underpins vervet ability to issue alarm calls (Cheyney and Seyfarth, 1990). Like humans, juvenile vervets use circumstances to adapt to norms: ABC learning2 allows activity to be gradually shaped by social controls. The synergy of cowbird song-learning is even more human-like in drawing on instants of dual control (West and King, 1996). Specifically, silent females assess the song of juvenile males by providing reinforcing ‘wing flaps’. These mark passages that males remember and use in mature

Biomechanical constraints in language 203 courtship. Interestingly, such bursts of song are most likely to elicit copulation (ibid.) only in an ‘appropriate’ geographical area. In vervets, cowbirds and humans, vocalization develops, in part, through dual control. Using ABC learning, it develops through the influence of circumstances and companions as well as within-organism mechanisms. Synergy enables vervets to learn alarm calls, cowbirds to make up sexy songs and, babies to discover appropriate conduct. Infants connect with history as social learning guides them into culture. In primates, perceptual biases guide infants to act in line with norms and circumstances. Humans, however, use encultured behaviour which, drawing on history, is irreducible to species specific biases. Rather than appeal to innate representations, therefore, I use Hutchins (1995), Clark (1997), Hurley (1998) and Rowlands (1999) to treat cognition as distributed. Bringing out how infants use vocal and visible patterns, I ask how babies come to enact agreement about what is judged. Oddly, I find, infants use experience to act in ways that can be ‘explained’ by acts, goals and sounds. Although a newborn’s first days are outside history, babies of 6⫹ weeks participate in intersubjective behaviour that shows precisely this kind of understanding. Clarifying, I show how biomechanical constraints develop through an epigenetic process that gradually tunes affective and intentional activity.3 Thanks to biases for affective and cultural content, a baby adopts consistent behaviour that manifests the appropriate forms of agreement.

Cultural mediation Before examining how shared behaviour arises, I distinguish exclusive and inclusive uses of ‘culture’. On the one hand, many appeal to mechanisms that, by definition, allow culture to be replicated across generations. This exclusive approach is found in the work of, say Dawkins (1982), Dennett (1995) as well as in Tomasello’s (1999) theory of how infants allegedly attend to the intentions of others. Placing culture outside the body, memeinfluenced views effectively identify it with systems, practices, artefacts and symbols. Culture can be also applied in an inclusive sense. Boyd and Richerson (1996), for example, define it as “information or behaviour acquired from conspecifics through some form of social learning”. It thus concerns both external entities and, crucially, behavioural and informational content. Culture is a set of patterns and frequencies that spread in time and space. In humans, of course, these can also be explained. Thus appeal to mechanisms must account for the dispositions that Bourdieu (1971) calls habitus and which, as Lizardo (2004) shows, ensure that social activity is jointly constituted by culture and cognition. Such a view is developed in Hutchins’ (1995) work on navigation in a US navy ship and Micronesian boats.4 Defining where a vessel is relies on external design and instruments as well as, say, discipline, ways of speaking, connect talk with world. Related inclusive views are

204 Stephen J. Cowley used to examine non-human culture. Rendell and Whitehead (2001), for example, document how whales and dolphins learn cultural behaviour patterns. For example, some Brazilian dolphins have worked symbiotically with humans for over 100 years. The forefathers of today’s fishermen established a tradition of sharing the catch in return for the dolphins’ ancestors help in trawling. Stressing variability of joint experience, inclusive approaches highlight culturally based understanding. Rather than a language instinct (Pinker, 1994), symbolic reference (Deacon, 1997) or socio-cognitive adaptations (Tomasello, 1999) weight falls on empirical study of how individuals act culturally. We ask how they come to exploit a matrix for consistent action that, once formed, allows activity to be perceived round goals, sounds, acts and structures. Since these are defined by social function, social judgements can begin with behavioural consistency. Where picking up wants or beliefs, to act consistently is thus to move to the edge of history. Since culture derives from (or is) behaviour, what matters is how infants become sensitive to what counts as significant. With Piaget (1952), Donald (1991) and others, entry to a historically based world is based on motorcentric activity.

What Nokhukanya shows us The inclusive view of culture allows us to examine how an infant’s doings come to be mediated by historically based events. As often occurs in mammals, this depends on how an individual adjusts to social activity. However, given primate biases, such adjustment leads a child to experiencing others as encultured. Further, for historical reasons, each social group is likely to have slightly different entry points to cultural life. Hereafter, this is illustrated by activity in a single child-caregiver dyad. Using an example from recordings of 14-week-old South African infants (see, Cowley, 2003),5 emphasis is given to how caregivers, infants and culture conspire to put development on locally appropriate developmental pathways. In the sample of Indian, White and Zulu dyads, encultured activity occurs first in the isiZulu speaking group. Not only do infants everywhere get upset but, as the behaviour derives from biology, it can be identified even by outsiders. There are, moreover, many differences in how adults deal with distress. In the sample, Indian and White South Africans tend either to fuss over the child or soothe it (e.g. by rocking the baby) while their Zulu counterparts are more likely to control ‘from a distance’. Avoiding touch, jiggling and blowing on the body, infants are more often (and earlier) encouraged to thula, or fall silent, by means of purely vocal and visible movements. If successful, the events become joint activity (henceforth, TH-activity) during which a caregiver silences the baby. The frames below show Florence getting the 14-week-old Nokhukanya to be quiet. They illustrate the baby’s ability to enact her mother’s want.6

(1)

(2)

(3)

Figure 13.1 Florence getting the 14-week-old Nokhukanya to be quiet.

206 Stephen J. Cowley In the first frame, Florence’s singing evokes or parallels the infant’s incipient crying. Immediately, she responds and makes Nokhukanya party to dramatic Zulu hand waving together with vocalization of, among other things, ‘njega, njega, njega’. In the second frame, the silent baby is rewarded with a beam of approval. Immediately, there follows ritual ticking off (using an index finger to prod in the direction of the child) and a warning about the father’s arrival. In the third frame, nine seconds after the first, as Nokhukanya re-orients to intersubjective contact this is acknowledged by her mother’s full-bodied expression. Nokhukanya falls silent. Rather than react to movements qua movements (TH-movements) the baby’s behaviour (henceforth S-behaviour) enacts understanding. Since a sceptic may deem the account ‘subjective’, I stress that while the events can be formulated many ways, the baby indisputably has a practical grasp of what her mother wants. Four sets of reasons back this ascription. First, it is consistent with what is usually required of a distressed baby. Second, among other things, Florence says no, no, no (njega, njega, njega). Third, those familiar with Zulu life see the want in hand-movements and facial expressions. Fourth, the ascription is supported by subsequent events: when Nokhukanya does fall silent, Florence rewards her with a smile and reinforces her actions with (what adults hear and see as) ritualized threats. Thus, all who view the video agree that, as Florence wants silence, Nokhukanya’s behaviour is appropriate. What must be clarified, then, is how baby behaviour achieves this coherency, or comes to adopt an encultured way of going on. Somehow, locking on to caregiver expectations, she fits behaviour to a historically derived pattern. Her body enacts a judgement about what, in the circumstances, her mother wants. Neither language nor universal signalling explains what happens. First, no-one thinks a 14-week-old baby understands words. Second, if babies heard ‘what was said’ in [njega] the pattern would be in general use. In fact, outside the Zulu sample, touch is (almost) always used to silence 14-week-olds. Verbal interpretation, then, does not explain Nokhukanya’s behaviour. Nor, it seems, does the baby use universal non-verbal signals. Even if some of the activity might fit such an analysis (perhaps, aspects of voice quality and/or some pitch contours), the events are unmistakably Zulu. As clarified below, Nokhukanya’s silence results from hand-waving and vocalization. Further, if TH-movements were conventionalized, their nature would be explicitly known. Yet, when asked, members of the group give no such explanation.7 Further, in the recording, the mother obtains silence in similar- not identicalways 3 times in five minutes. Nokhukanya’s S-behaviour is not set off by a determinate stimulus. Rather, the child participates in joint activity by meeting full-bodied Zulu TH-movements with silence. Finally, this view fits what happens in groups where silence is invoked by other modes of contact (i.e. touching and shaking). Although Nokhukanya uses encultured patterns, her brain-body system certainly uses resources that have been used in dealing with earlier (non-cultural) incidents.

Biomechanical constraints in language 207 Nokhukanya’s ‘act’ is motivated. She uses social learning about encultured activity that is meant to invoke silence.8 She exploits maturation, her current affective state, her mother’s movements and the Zulu setting. Nokhukanya’s entry into history arises as S-behaviour meshes with TH-movements. Using dual control and another body, she acts in line with how her mother’s behaviour is meant to be judged.

Cognitive internalism versus distributed cognition Nokhukanya’s intersubjective behaviour shows ‘aboutness’ as it fits what her mother’s body represents. Since this is inexplicable from both classic nativist and empiricist viewpoints, I stress that, in a lay sense, Nokhukanya ‘understands’ when to thula. Using principles like those allowing male cowbirds to learn song that enhances reproductive success, I suggest that both parties exploit previous encounters in learning to orchestrate what they do. The baby uses synergy for which its brain-body systems are prepared. Biology and culture conspire to ensure that utterance-activity falls under dual control. Synergy allows external features of the world to shape cognitive processes. Events mesh with sub-personal systems that, in real-time, sustain behaviour. For Nokhukanya’s activity to loop with her mother’s, she must use cognitive powers from outside her body. What I have elsewhere called ‘cognitive internalism’ (Cowley and Spurrett, 2002; Cowley, 2003)9 is thus an inappropriate for understanding the events. They can be explained by neither representational nativism nor general learning. The case against general learning is simple. Where learning is justified by stimulus-response or accommodation-assimilation, complex behaviour is not posited in the 14th week. Especially in relation to a cultural event, the general learning theorist struggles with the following: ●

●

●

●

●

What happens uses, not a movement, but a want manifest as encultured activity. Since infants from other groups do not learn to thula, the behaviour is culturally specific. Since the ‘meanings’ are complementary, the event cannot be explained by imitation. What Nokhukanya does is neither automatic nor mediated by a specifiable emotion. Nokhukanya’s activity is ‘strategic’ and ‘motivated’.

As Nokhukanya reacts to a want, what is perceived is holistic and irreducible to a single signal. Even though the baby falls silent without training, no bottom up account can explain what she has learned. This is a starker version of Quine’s puzzle (1960) of how ‘gavagai’ can be heard to identify rabbit. It appears intractable because, far from performing similar perceptual actions an animated Florence induces Nokhukanya to freeze. How, then, can a

208

Stephen J. Cowley

14-week-old act as if it knows what to do? Indeed, given the behavioural contrast, her knowledge can be based in neither imitation, mimesis, or, as Sroufe (1995) shows, ‘fear’. Further, to the embarrassment of the general learning theorist, TH-activity makes strategic sense. While natural selection is likely to favour juveniles who can be ‘motivated’ to fall silent, no such explanation is sufficient to explain what Nokhukanya does. While using subpersonal systems, her S-behaviour is plainly also set off by historically shaped events. Even if what she does is based in biology, general learning is not sufficient to explain what happens. Classic nativist views are immune to empirically-based refutation. As Matthews (1993) argues, such theories appeal to faith by presenting behaviour around epistemic problems that can only be solved by hypotheticaldeductive methods. To challenge classic nativism, then, I deny that, to thula appropriately, the baby needs representations. Indeed, any such view is bound to adopt proposals along the following lines: ● ●

●

●

●

She knows what makes a TH-movement a TH-movement. She knows (at least one) ‘function’ of TH-movement is encoding an I-WANT-SILENCE representation. She knows principles or rules for applying I-WANT-SILENCE representations to a SELF-representation. She knows how to use SELF-representation to generate an S-representation (i.e. one that triggers S-behaviour). She knows how to use an S-representation so that, within 9 seconds, she can produce the relevant S-behaviour.

None of this is plausible. First, it is unlikely that a 14-week-old represents TH-movement, I-WANT-SILENCE, SELF and S-behaviour. Second, there is no reason to believe that the child must know abstract principles or rules. Third, even if I-WANT-SILENCE, SELF and S-behaviour representations were ‘innate’, this cannot apply to the (typically Zulu) TH-movement. Either this is selected from innate alternatives or, perhaps, learned. Further, unless we say what exactly might be innate, we are floored, once again, by the general learning problem. The TH-movement is more likely to be a TH-movement, then, because of social function: it functions (with variations) because someone ‘wants silence’. Are we then to posit a universal baby who is born with a complete set of representations for invoking silence? That is unlikely. Further, even if a formula for identifying TH-movements is found, an explanation will be needed for how this evokes I-WANT-SILENCE while feeding output to SELF. Finally, the problem also applies to whatever causes S-representations to produce output that produces S-behaviour. The classic nativist assumes that babies process representations and thus rely on ways of relabelling events. In presenting a deflationary alternative, however, we must not undermine the insight that babies come to act autonomously. What must be clarified is how a child can act consistently

Biomechanical constraints in language 209 with calls for silence (howsoever made).10 We must explain how what sets off S-behaviour can be activity that demands silence. Somehow, without knowing, Nokhukanya comes to act with coherency in response to certain social moves. To proceed, rather than ask what is in-the-system, we can look at how learning and experience exploit bodies in the world (see also Clark, 1997, 2001; Rowlands, 1999; Cowley, 2003; Spurrett, 2003). We can challenge what Hurley (1998) calls the Input–Output view of mind.

The distributed perspective Instead of identifying the mental with brain processing, cognition may spread across brain, body and world. Reliance on ‘internal’ representations may be complemented (or replaced) by brain-body systems that learn, in part, from physical events. Even ABC learning can be directly shaped by what Dewey (1896) called ‘circular’ co-ordination. In more recent terms, brains can use dynamic patterns whose properties impact on neural and bodily activity. On this view, the child couples with the caregiver’s doings as, in real-time, each pursues sub-personally driven interests. The baby learns that, all things being equal, it is positive to connect S-behaviour with (so-called) TH-movement. In place of internal processing and learned habits, coherency arises because TH-activity is of strategic value for both parties. Far from lying at the end of a causal chain, the baby’s action contributes to a circuit of ‘scattered causes’ (Clark, 2001). Nokhukanya recognizes prospective behaviour on the basis of previous experience and, in the circumstances, this sets off retrospective action. Using perceptual biases and ABC learning, affectively driven activity thus takes on coherency. To come further in understanding the events, therefore, we must posit that the baby uses underlying motivating/ affective sub-systems. The nativist is correct that the infant draws on natural selection. However, rather than appeal to invariant representations, she may rely on affect seeking activity. Provided that we avoid the assumption that psychological terms name mental objects,11 we can see affective systems as the basis for Nokhukanya’s growing behavioural consistency. Thanks to learning, she allows another’s ‘want’ to evoke a form of dual control favoured in the environment. Using prepared learning and biases shaped by experience, this enables the want to be felt by both parties. Looping with her motivational systems, the baby is prompted to ‘go on’ as she does. Her falling silent couples with the mother’s wanting activity. In integrational terms, using past experience, she contextualizes the want to affect future experience. Joint activity gains coherency through action that relies on ABC learning as well as in-built biases and social retrospection. In the short-term, the coherency of the activity elicits reinforcement. The child’s doings evoke, first, her mother’s broad smile and, then, ticking off. Given its appropriacy, dual control gives the caregiver a new capacity to influence her babys’ affective/ motivational systems. Understanding the want makes Florence feel that the

210

Stephen J. Cowley

baby grasps what she means. In the mid term, then, she may well change how she reacts to the infant’s not falling silent. Thanks to the temporal embedding of human action, the infant’s contextualizing will shape her judgements. Just as TH-activity affects responding, in other words, Florence gradually becomes inclined to invoke silence by waving and vocalizing. Conversely, using ABC learning, Nokhukanya may learn to react to stereotyped expressions of the want. Consistently with Harris’s (1996) principle of co-temporality (what has happened impinges on this, now) utterance-activity comes to be shaped by biomechanical constraints. This happens, moreover, both in microtime and the slow domain of verbal thinking. Dual control loops utterance-activity with sub-personal systems that control social behaviour. The dynamics of utterance-activity are themselves motivating content. Nokhukanya’s affect influences her mother’s neural systems which, in turn, exert joint control over what happens. The baby directs attention at – and picks up on – what counts as TH-behaviour. By participating in the event, she enacts what it is that is judged. At 14 weeks, Nokhukanya is learning, literally, to assess and manage her mother. Without inner understanding, she is establishing new ways of contextualizing. Before asking how, biologically, this is possible, I must clarify how each influences the other’s contextualizing.

Distributed contextualizing Contextualizing occurs where, in the circumstances, an individual uses experience to act in ways that affect its future. While even insects contextualize, social vertebrates use synergy to ensure that contextualizing shapes both their doings and those of others. In mastering alarm-calls, vervets use social learning that brings experience of different circumstances into contact with what they do. In cowbirds, sexy songs use both circumstances and an individual’s salient wing-flapping experience. Apparently, Nokhukanya draws on circumstances and affective markers to align her behaviour to cultural expectations. Far from using TH-movements as a stimulus, she produces a realtime judgement. By acting as she does when she does, she does what underpins agreement in judgement. In other terms, she accommodates to what her mother seeks to achieve. Thanks to sub-personal systems, the baby exerts control over activity: in the circumstances, she acts to let the want prevail. Human dual control is managed in real-time. While like cowbirds in drawing on past experience and circumstances, Nokhukanya also relies on meshing what she does with concurrent events. Specifically, in a brief timeinterval, she detects movement, identifies a pattern and, immediately, adjusts her doings. While Florence does not control the baby, her movements are allowed to orchestrate events. Exploiting affect, the dyad also depends on shared social biases, ABC learning, and internal processes. Nokhukanya exploits the sound and movement to fit the cultural pattern because, in realtime, she draws on experience. Both parties, then, rely on fast-working subpersonal systems that produce coherency. This is possible because of how each

Biomechanical constraints in language 211 uses the other’s experience to make future. Human contextualizing arises as joint activity is embedded in larger time domains: synergy draws on individualbased learning, circumstances and also history. It is the historical dimension of dual control that characterizes distributed contextualizing. Bodily activity not only meshes but, as this occurs, it conforms to a local habitus. Distributed contextualizing allows the real time coherency of human action to spread across individuals and, therefore, space. An event gains consistency thanks to its feel, its timing and appropriacy. Briefly, distributed contextualizing features the following. ● ● ● ●

●

It occurs in real-time in the public domain. It occurs in a historical context (uses macrosocial constraints). It exploits (we assume) affective markers present in the other’s activity. It exploits shared experience of similar moments (however, what counts as similar is not necessarily common to the parties). It exploits a learned ability to anticipate what will happen.

Contextualizing later comes to rely on the ability to utter and interpret word-forms. This will require babies to develop both episodic memory and a capacity to hear and speak syllabic forms. Nokhukanya, however, has no need for this: she needs no more remembering than that which allows, say, a dog to find the path home. This is sufficient because the baby needs only couple movements to give and produce real-time feedback. Relying on retrospective activity, Nokhukanya uses dual control to probe and feel out her mother. While almost certainly using perception-action mechanisms like those found in mammals, birds and amphibians (see, Preston and de Waal, 2002), she also uses her mother’s feeling. Thanks to this affective regulation, her mother is moved to recognise baby’s learning. Affect thus ensures that experience unites with real-time sensitivity in ways that allow each to guide the other. It is this, I suggest, that allows the baby to apprehend – not just movements – but cultural patterns. Unlike vervets and cowbirds, her biomechanical resources come to draw on cultural constraints. In agreeing about the want, S-behaviour is made appropriate by TH-movements whose value is macrosocially constrained. Acting as a Zulu, Nokhkanya steps into history. Synergy allows her to draw simultaneously on macrosocial constraints, circumstances and sub-systems. While using historically derived patterns, these precisely fit the timing of microbehaviour. Elsewhere I show how events in microtime shape the interpersonal sense of talk (e.g. Cowley, 1994, 1998, 2001) and have parallels across the vertebrates (Cowley, 1997). Nokhukanya shows how babies mesh social activity in two time domains. Thanks to dual control, infant and caregiver synchronise, first, in microtime and, later, in the slower ‘tonic’ domain (Owings and Morton, 1998). To give her mother what she wants, Nokhukanya uses instantaneous perception/action and, slowly, adjusts to cultural expectations. Thus she holds silent, not just for an instant, but for (at least) several seconds. Eventually, by so doing, a domain of thought

212 Stephen J. Cowley and emotions opens up alongside affectively mediated activity. Joint action gains coherency from sub-agent systems that are more complex than those of cowbirds. Next, I turn to a model of such systems and, afterwards, ask how they could arise from natural selection.

Intrinsic motive formation Thirty years of research have allowed Trevarthen to develop hypotheses about the neural mechanisms that drive intersubjective behaviour (Trevarthen et al., 1999). His system of innate motive formation (IMF) recognizes that perception-action mechanisms allow infants to exploit social activity. Given these goals, no stance need be taken on how experience is recorded, patterns identified, or the biases’ embryogenetic origins. Highlighting intrinsic organization, he relies on cross-specific neural comparisons. In contrast, I ask how neural organization like that described in the IMF model can contribute to distributed contextualizing. Applied to Nokhukanya and Florence, while the areas in grey on the Figure 13.2 below show brains, the enclosed areas identify the peripheral bodysystems that manage action and perception. (The labels P A e S E identify perception, action, expression, sensation and expression respectively). Natural selection ensures that both persons access similar IMF systems that work by using mimesis (the m-system) to form and deal with higher motivations (the M-system). While small babies are (largely) restricted to functions based on mimesis, intersubjective behaviour allows higher-level M-systems to develop. In ontogenesis, joint activity is affected by social events that influence the baby’s IMF system. While explaining the newborn’s reciprocal, mimetic responses, the model also links with recent work on the neural substrate.12 Thanks to learning, mimetic mechanisms gradually change higher level motivations. We come to use mimesis together with affect and motives whose

Nokhukanya

Florence a

e

e

a

P

S

S

P

m

m M A

Neural systems

O

E Peripheral systems

M E

A

Peripheral systems

Figure 13.2 Intrinsic motive formation. Source: Based on Trevarthen, Kokkinaki and Fiammenghi (1999): 132. Notes M sub-system: biased to direct perception/action (micro temporal scale). M system: incorporates goals, judgements and beliefs (larger temporal scale).

Neural systems

Biomechanical constraints in language 213 results are explained by actions, wants and intentions. Using social learning the higher level systems later become integrated with word-forms and, in so doing, allow us to formulate explicit goals. In infancy, they function to permit basic sense-making that is consistent with such goals. As IMF components develop in tandem, we act both intuitively and in a less spontaneous domain. Redescibed in IMF terms, Nokhukanya uses motivation that derives from an adult’s affect-driven movements.13 Without a central mechanism (or SELF), Nokhukanya brings events under dual control by gearing to Florence’s activity. Without (inner) understanding she uses mimesis to respond affectively and develop experience based in utterance-activity. At the moment described, however, rather than let the m-system trigger agitated behaviour, Nokhukanya inhibits mimesis. Using experience and a nascent M-system, she freezes negative affect, and shows learning.14 Overriding her distress, the IMF system readies her for a smile and reprimand. Social biases, ABC mechanisms and affective response give her activity a special quality. Like an affectively charged robot, Nokhukanya couples one activity (S-behaviour), with another (TH-movement). Using inhibition, she contextualizes a want by using biomechanical resources. As a human, she uses real-time meshing and historically derived constraints to inhibit behaviour. Later, she will get better at feeling when silence is wanted and, once a motorcentric analogue develops, will exploit the mechanism to silence others.15

Motives and biomechanical constraints Nokhukanya can respond appropriately to her mother’s want because she is motivated to use distributed contextualizing. She has both accumulated knowledge of a pattern and, using her IMF system (or equivalent), is able to inhibit mimetic response. Instead of becoming agitated – of allowing her m-system to produce behaviour that resonates with her mother’s, a higher level system inhibits response. By thus falling silent she acts in line with her mother’s want. Given how this exploits Zulu hand and vocal patterns, Cowley et al. (2004) call this an early ‘sign of culture’. The joint activity of the parties is encultured because it fits what is expected, is closely timed and both affectively and historically appropriate. It is joint activity that legitimates the judgement that the baby understands, is obedient, or shows respect. Florence does not cause her daughter’s response. Rather, she triggers an intrapersonal sub-system that, in the circumstances, produces inhibition. Importantly, were the motive for distress greater, the baby would go on quite differently. In other words, as Wittgenstein advises us, we avoid any supposition that the action is (or requires) a ‘process’ of understanding. Metaphorically, the events are like the threatening and parrying of boxers in that the child no more chooses to meet the mother’s want than the mother herself chooses how to act. Silencing the child is, literally, part of a Zulu form of life. Without rules, the events show cultural regularity which has a triple

214 Stephen J. Cowley strategic function. The mother gets silence, the child approval, and, together, they perform a cognitive trick. Discovery of coherency shows that, in social life, different behaviour often serves a single social function. Without words, Nokhukanya acts in line with the belief that, sometimes, babies must be silenced without fussing or soothing. Even if the details are wrong, Trevarthen shows how synergy can transform social life and, in specific settings, contribute to distributed contextualizing. Specifically, it exploits a two-part motivational system that allows real-time adjustment to affectively modulated patterns. Utterance-activity is a cognitive resource which, while beyond the body, contributes to development. Sub-personal systems thus allow the child to use scattered causes in picking up on an abstract pattern. Nokhukanya exploits experience, current activity, circumstances, her mother’s goals and activity in the social environment. Their dual control allows her to use utterance-activity in binding her body to the social world. Given something like IMF, infants develop biomechanical resources. While learning from statistically based contextualizing, infants come to act strategically. For caregivers this is seen as involving generalizations and, for this reason, they act to encourage socially relevant learning. Just as might be expected, bonobos also elicit socially appropriate human behaviour (Cowley and Spurrett, 2003). It thus seems that, using inhibitory mechanisms and ABC learning, individuals can come to act in ways that shape affectively based social repertoires. By laying stress on motivational systems, this account overcomes difficulties of both classic nativism and naïve empiricism. Next, therefore, I will consider what kinds of selection pressures might have made human IMF systems sensitive to the scattered causes that drive, for example, TH-activity.

Niche construction and co-evolution While suggesting that each party uses IMF systems to couple with the other, the analysis extends Trevarthen’s view. Emphasising the effects of utteranceactivity, I claim that each person uses the other’s experience as a cognitive resource. Nokhukanya exploits a caregiver’s real-time motives by learning to inhibit m-system response. Florence, perhaps, makes sense of what her baby does and, especially if she puts this into words, can use it to change future interaction. While using perception-action mechanisms, distributed contextualizing provides a basis for coming to agree – and, later, disagree – about what it is that we judge. In a suitable environment, moreover, bonobos use similar mechanisms and effects (Savage-Rambaugh et al., 1998). Thanks to something like an IMF system, Kanzi exploits English and lexigrams in a ‘North American’ style (see, Cowley and Spurrett, 2003). Below, I use this finding to speculate about the phylogenetic underpinnings of talk. Infant activity, like that of other primates, draws on biases or prepared learning. While using species-specific adaptations, this is no reason to think that learning to talk exploits specific ‘language’ genes (or suites of genes).

Biomechanical constraints in language 215 Indeed, some may find it more likely that infants draw on phylogenetic mechanisms whose evolutionary origin is bound up with prospecting and retrospecting patterns in utterance-activity. In short, such mechanisms are likely to serve the rise of distributed contextualizing. Given that this varies across cultures, times and dyads, natural selection may well have promoted – not public forms – but, rather, underlying capacities. Rather, as seems to be the case for number systems, it seems unlikely that we have (or need) adaptations for grammar or meaning.16 If so, this provides grounds for returning to the traditional view that (most) grammatical constructions and verbal units are second-order historical entities. If cognition is distributed, however, this position nonetheless raises interesting questions about the role of language in both evolution and development. Rather than focus on symbol manipulation, what comes centre stage is how historically based conduct can serve in shaping underlying systems and, at the same time, establishing and maintaining cultural organization. Accordingly, I proceed by using co-evolutionary theory to ask how vertebrate abilities for dual control might lead to selection for distributed contextualizing. Natural selection does not rely only on phenotypes. As Dawkins (1982) stressed, selection also uses aspects of the environment. In beavers, for example it uses bodies, the forests that provide wood for dams and, of course, conditions that dams produce. Drawing on extra-somatic selection, beavers have become unusually social rodents with large teeth and flat tails. Crucially, while selection uses the world beyond the body, it also exploits epigenesis. Since human development has an evolutionary history, the embryogenetic basis of language may lie in biases and social learning. Not only may many ontogenetic facts depend on a child’s communicative infrastructure, but aspects of culture may use natural selection. In IMF terms, an infant may draw on (something like) motivational systems geared to dynamic patterns in human activity. Once distributed contextualizing starts, an encultured caregiver can guide the child to social complexity. If early joint activity relies on mimesis (or m-systems), by 14 weeks, cultural patterns are already insinuating themselves into infant-caregiver relationships. The biological basis of talk may thus lie in how, in phylogenesis, neural sub-systems come to control how infants participate in utterance-activity. On a co-evolutionary model, the object of selection is the organism-in-aniche (Laland et al., 2000). Genotypes adjust to environments that their bodies modify. The phenotype includes a ‘niche’ which, when constructed, affects reproductive success. Unlike in Lamarckian theory, the gene pool uses not acquired characteristics but individual variations. What is ‘selected’ is a species-constructed world or phenogenotype (ibid., 2000). Just as beaver brains, bodies and behaviour allow for the gnawing of trees that fall into lakes, over time, it led to dam-building and oversized teeth. Other characteristics drawing on niche-based resources helped beavers become resistant to socially-transmitted diseases, unusually monogamous and intolerant of running water (Sampson, 1997). Other examples of how an organism-in-a-niche

216 Stephen J. Cowley influences/is influenced by natural selection are found in earthworms, the Galapogos woodpecker finch (Laland et al., 2000) and in human pastoralists. Human populations with a history of herding have higher lactose tolerance than those lacking such traditions (see, Feldman and Cavalli-Sforza, 1989). Where milk is a nutritional resource, efficient digestion enhances reproductive success because those unable to use it have lowered fitness. In evolutionary time, herding populations come to have a higher concentration of lactose tolerance genes.

Co-evolution and the human IMF system Close examination of how Nokhukanya comes to fall silent gives reason to think that learning to talk uses the affect that regulates utterance-activity. Using IMF mechanisms, human infants are predisposed for distributed contextualizing. While what follows is speculative, I suggest that this depends on co-evolutionary processes. In so doing, I follow Laland et al. (2000) in pointing out that pressures depend on ‘how an organism obtains its resources or defends itself in an environment’. What happens reflects both selection pressures on individuals and niche-based constraints. Distributed contextualizing, I suggest, arose as selection favoured infants who used utterance-activity to obtain protection and resources. By thus shifting attention to underlying systems, it can be posited that the primate IMF systems contributed to the formation of biases used in managing and assessing multimodal activity. And, strikingly, even today, infant behaviour is managed around manual movements and vocal patterns. Before weaning, babies rely on relatively primitive utterance-activity. Natural selection, therefore, may have favoured tricks that benefit the infant-in-a-niche. Nokhukanya’s lesson may even be that IMF systems developed to allow infants to assess and manage parental behaviour by using cultural dispositions. While bonobos (and, other primates) also use IMF systems, humans are specialized in discriminating and producing vocalizations. The Figure 13.3 below sketches the general idea. I have used circles of various sizes to suggest Ancestral IMF

Bonobo IMF Human IMF

Figure 13.3 IMF: a co-evolutionary view.

Chimpanzee IMF

Biomechanical constraints in language 217 the extent to which IMF systems are likely to affect behaviour by juveniles of each species. My view is that an earlier form of IMF (shared by bonobo and human ancestors) gradually adjusted to the demands of utterance-activity. Given conflict between parties (an infant has 50 per cent of her mother’s genes), infants may have set off an assessment based arms race that changed adult cognitive resources. Especially when times are hard, an infant who monopolises resources is more likely to survive than siblings. Natural selection, then, may favour individuals whose behavioural assessments successfully shape social events. While assessment-management is a communicational universal (see, Owings and Morton, 1998), human utterance-activity allows prospective and responsive activity to coincide (e.g. Cowley, 1994, 1998). Further, alongside this relatively primitive behavioural co-ordination it also becomes organized in the relatively slow domain of verbal thinking and action. How we speak and move, it seems, has a symbiotic relationship with a community’s encultured sound patterns. Parental behaviour is managed both intuitively (thanks to an m-system) and around forms of conduct consistent with local values and beliefs (using M-systems). Once babies learn to participate in this dispositional matrix, they begin to use salient events as a basis for remembering past outcomes. Viewed in this way, it seems that natural selection built our kind of IMF system by relying on how perceptual and motor discriminations helped us become affectseeking primates who exploited social learning. The relevant biomechanical systems, therefore, became tuned to prosodic and visible subtleties. Independent as they are of word-forms, such changes could go back 2 million years or more. In our ontogenetic niche, IMF may have changed in tandem with how we assess others and, of course, cultural ways of controlling babies. Niche construction illuminates both species’ specific aspects of brain-body systems (IMF systems) and physiological capacities for using utteranceactivity as a cultural resource. Further, if infants live in a niche until (roughly) the end of the first year, this may explain why they forge syllable-based resources. As contextualizing develops, it focuses on ‘abstraction-amenable’ aspects of talk (Spurrett and Cowley, 2004). As infants are drawn to structured sounds they also find themselves producing ‘canonical syllables’ (Oller, 2000). By the end of the first year, culture permitting, adults treat infants as if they both understand and know ‘what to say’. Using capacities for active attention, this probably also uses prepared learning. Indeed, older infants become biased to hearing utterance-activity around culturally specific ‘sounds’ (Kuhl, 1998). Though not sufficient for language learning, this is an important step. Coming to engage with utterance-activity, then, is also a matter of adjusting an IMF system to the local environment. Learning to produce syllables shapes a brain-body system around distributed contextualizing. The child needs, not an adaptation for language, but ways of assessing and managing others that sustain encultured and word-based judgements. As this happens, the child moves from the edge of history to take up new roles in its community.

218

Stephen J. Cowley

On the edge of the historical present Having established how Nokhukanya becomes able to act with coherency, I outline why this may be necessary in learning to talk. Given restrictions of space, I stress the following: ●

●

●

●

●

By acting with coherency the baby induces major qualitative changes in a caregiver’s utterance-activity. By acting with coherency the child learns, in some circumstances, to act consistently. Thanks to intrinsic motives, change and coherency set off a dynamic process of distributed contextualizing. Distributed contextualizing nudges the infant towards local forms of behaviour. Discovering what social behaviour is ‘about’ leads to formation of functional categories that, in principle, map onto words and values.

Thanks to intrinsic motivations, utterance-activity promotes discovery of historically based constraints. Distributed contextualizing encourages the child to exploit consistent activity that maps onto local values. When Nokhukanya acts as if knowing that TH-movements are meant to invoke S-behaviour, her mother ‘feels’ her baby’s understanding. Activity that uses simple mechanisms changes how a caregiver acts. Based on past experience the infant begins to exploit utterance-activity as a cognitive resource. Ceasing to be purely affective, her activity now serves, among other things, to prompt and probe Florence’s wants. Thanks to intrinsic motives, the child finds that culture is, among other things, a matrix of dispositions. Distributed contextualizing allows the infant-in-a-niche to lock onto encultured patterns. Affective response enables her to assess and manage caregivers while gaining protection and resources. Indeed, this may be why humans use primate biases in perceiving sounds. As these are highly susceptible to cultural patterning, I speculate that intrinsic motive formation brought new synergy to vocal learning. Internal and external motives came together to ensure that utterance-activity ceased to be purely affective or, in other terms, fell under macrosocial constraints. Today, IMF based affective movement may be both necessary and sufficient to lead affectively-charged humans into a cultural world. Further, since biomechanical constraints serve the infant’s interests, we gain from conforming to our affective heritage. Rather as a child comes to exploit culture, encultured caregivers thus use the child’s ABC learning to persuade her to meet wants and needs. Adults gain from fitting their doings to children’s capacities for attuning to cultural patterns. Acting consistently can incline the infant to certain values and beliefs. Not only does its mimetic system tune to the caregiver’s but, as this happens, higher level constraints lead to adjustments in behaviour. Brains become profoundly biosocial.

Biomechanical constraints in language 219 Distributed contextualizing uses a caregiver’s adjustments to base practical knowledge on its functional properties. While Nokhukanya’s ability to thula is based in biology, it exploits cultural conduct. Without inner understanding, habitus draws her into acting consistently with goals. Later, similar processes help her discover local sounds. While based on affect, human dual controls use coherency to induce practical understanding. Long before activity is analysed, adult beliefs shape biomechanical constraints. Before pursuing this, however, I ask how language relevant capacities are shaped by attuning to what it is that is judged.

Biomechanical constraints in language: getting started Biomechanical constraints emerge as utterance-activity organizes early behaviour. Not only do past interactions influence current events but, using synergy, a child develops motorcentric routines to integrate activity with historically derived patterns. Given her predispositions, Nokhukanya gains practical knowledge of what her mother wants. In the terms above, she discovers distributed contextualizing. Over time, this changes her biomechanical constraints and, as a result, she learns to assess her mother’s doings and, immediately, cope with consequences. Utterance-activity comes to be used – not just affectively – but also to alter another’s immediate and mid term behaviour. While Nokhukanya uses inhibition, babies later use other mechanisms to control what they do. In short, by acting jointly, synergy enables a caregiver to guide the child in learning to assess and manage others. Nokhukanya makes a step toward acting in line with the belief that talk consists in acts whose sound structure enables us to achieve goals. This view depends on individual members of a group all being able to draw on a matrix of consistent behaviour. Conversely, by coming to act with coherency the child makes a step towards becoming a social player. Unknown to the baby, co-evolution has equipped her to use scattered causes in aligning with expectations. Drawing on Florence’s feeling/moving, therefore, subpersonal systems motivate Nokhukanya to act appropriately. Distributed contextualizing exploits social motives that, unlike physical causes, depend on word-mediated conduct. In discovering a social world, Nokhukanya integrates events with patterns allowing de facto verbal judgements. Even at three months her behaviour can be explained in terms of goals. At the end of the first year, it is heard around sound patterns and goal directed speech acts. Consistent behaviour feeds overinterpretation. When judgements reflect social facts, we use patterned behaviour/information to make construals. Especially if a baby acts in line with verbal ascriptions, adult responding reinforces this way of acting. The pattern, moreover, is consistently found in early development. Low-level cultural patterns encourage the expected and,

220 Stephen J. Cowley subsequently, aid the child to acting with coherency. Without hesitation, babies fit their doings to the matrix. As patterns repeat and generalize the baby later gains understanding of acts and consequences. Agreement about what is judged is thus necessary in learning to talk. When a baby ‘knows when to fall silent’, her activity can be used to encourage this kind of behaviour. The child can be trained to act in line with ascriptions to which it has no access. Further, many ascriptions fit the same events. Given the example, Nokhukanya may be encouraged to continue acting ‘like a good girl’ or, perhaps, to go on ‘showing respect’. While what is judged is coherency or real-time behaviour, how this will develop depends on judgements that reflect a higher cultural level. Social life develops from activity about activity. Biomechanical constraints use the reflexivity that arises from exploiting utterance-activity. In later years, activity about activity gradually becomes talk about talk. At three months, it suffices to act as if different vocalizations/ movements have the same functional value. The child can use this to enact social events with predictable outcomes. Indeed, anticipating reward may be necessary in learning to act prospectively. Thus, seeking to fulfill another person’s wants may precede having wants of one’s own. Relevantly, in the next weeks, Nokhukanya will reach for objects to put in her mouth. While also a way of achieving coherency by integrating information from different cognitive systems, this allows what is seen to be co-ordinated with complex manual activity. Early interpersonal integration thus gives way to contextualizing within the child. In Piaget’s (1952) terms, it allows for ‘secondary circular reactions’. Later, once she can relate to a familiar material environment, object-object co-ordination will give way, after nine months, to triadic behaviour. The baby will use gaze, for example, to contextualize wants or in using a caregiver to see if a situation is threatening. Finally, once wants couple with syllables, adults will hear these as ‘words’ (e.g. up, more, ball). As Nohkukanya learns to use distributed contextualizing she adjusts to cultural practices. Responding to wants allows her to probe others while giving off affective information. Judgements derived from a mimetic system are thus slowly reshaped by higher motivations. The encultured element in her intrinsic motives begins to decouple from affect driven mimesis. In Croce’s (1938) terms, her judgements cease to derive, immediately, ‘from an act that is being undertaken’. Instead, they come to be dominated by what he terms the ‘coscienza’ (conscience/consciousness) of the act.17 The baby becomes ever more reliant on historically specific ways of knowing. What she thinks binds with what, now, she says and does. As reliance on agreement in judgements and behavioural consistency diminishes, new changes arise. In time, she may, for example, come to give great weight to ‘literalness’ and agreement in definitions. Probably, she will follow the rest of us by coming to believe in the ‘reality’ of languages, minds and selves.

Biomechanical constraints in language 221 Appendix: Transcript/description of how Nohkhukanya ‘enacts her mother’s want’. The frames shown on the three images in the text are marked in italic. Frame

Time

1.

05.10

Not shown

06.00

Not shown Not shown

07.04

Not shown 2.

English

isiZulu

Smiles, points and moves right hand

08.13

Sings “ing na bel” “ing na bel” No no no [njega njega njega]

09.00

No

ngeke [nge.ke]

10.07

Not shown Not shown Not shown

11.00

3.

14.23

Not shown Not shown

15.00

12.23 13.00

16.00

Mother

I don’t want it No I’m going to get you

angifuni ngeka (mumbled) mina ngizothatha wena

Raises arms Waves hands twice; brings hands together in front of child ‘Zulu wave’ Responds, lowers hands and smiles ‘Waves’ both hands 3–4 times Moves forward; beams Ticks off twice and gentle prod Pouting mother open to child

Child Takes fingers from mouth, lifts arms Synchronizes and nearly touches hand Breaks gaze; cries Gazes again

Child relaxes

Starts to suck fingers Child removes fingers Child open to mother

Where’s your kuza baba father now? manje Your father’s ubaba coming wengane

Notes 1 While many take human life (and talk) to be grounded in acts, with HendriksJansen (1996), it is possible that, far from being the basis of our life, ‘act’ is a concept based on analysis of body-world activity. 2 The term ‘ABC learning’ is used by Karmiloff-Smith (1993) and Dennett (1995) to capture the within-organism-learning that uses neural networks sensitive to classical or operant conditioning. Other kinds of learning also occur. 3 In short, development incorporates both maturational and ‘external’ factors; it is not epigenetic in a strict biological sense (for discussion, see Hendriks-Jansen (1996)). 4 For Hutchins (1995) Culture is not any collection of things, either tangible or abstract. Rather it is a process. It is a human cognitive process that takes place both inside and outside the minds of people. It is the process in which our everyday cultural practices are enacted. I am proposing an integrated view of human cognition

222 Stephen J. Cowley in which a major component of culture is a cognitive process . . . and cognition is a cultural process. (p. 354) 5 By using this example I aim to give readers access to close analysis. In the current context, I include a transcription/description of the events in appendix A. 6 Hauser, et al. (2002) suggest that other primates are unable to produce behaviour that is “intentional in the sense of taking into account what others believe or want”. It is striking that, because of the cultural environment, 14-week-old Nokhukanya does this without any ‘inner’ understanding. 7 An elaborate explanation is that hands are used to silence the child because their movements are a stand-in for physical punishment. However, not only does this seem post-hoc but, in this instance, it plainly does not apply. 8 Elsewhere, this is called an ‘indexical’ function (Cowley et al., 2004; Spurrett and Cowley, 2004). This is justified because triggering behaviour correlates with a culturally defined similarity class. Here, I avoid the term, first, because it invites exclusion of the iconic and the symbolic and, second, because the events of, say, early tongue protrusion are already indexical (they use (noncultural) similarity classes). 9 It is beyond dispute that much behaviour has a biological basis. In referring to ‘classic’ nativism, I refer to the representational views of, above all, Fodor (1983). Arguments along similar lines are found in, say, Karmillof-Smith (1993), Thelen and Smith (1993) and Elman et al. (1996). 10 From a robotics point of view the problem can be solved by crude sensors and simple programming. It is analogous, for example, to systems that get Robovie (Ishiguro et al., 2001) to pick up and respond to a human’s pointing. The machine needs neither a central processor (a SELF representation) nor any knowledge of its own S-behaviour routine. 11 This position is taken both by eliminativists (see, Churchland, 1996), Dennett (1987, 1991) and Harris (1981, 1998). Fodor (1983) assumes that ‘real’ intentional states and linguistic forms embody, say, wants and verbs. 12 None dispute that this occurs. Further, the discovery of mirror neurons (e.g. Rizzolatti and Arbib, 1998) provides evidence that other primates have such abilities. What is controversial is how to account for microtemporal events that, following Ballard et al. (1997) can be said to occur at the embodiment level. While most accept that we use perception-action mechanisms (see, Preston and De Waal, 2002), Trevarthen’s dedicated IMF system is more controversial. The model stands in decided contrast not only with standard views of social learning (e.g. Stern, 1977; Kaye, 1982) but also representation based theories (e.g. Melzoff and Moore, 1997). 13 At 14 weeks, the infant is not yet able to reach for things and put them in her mouth. Sucking actions emerge when an object is placed in (or near) the infant’s mouth or, interestingly, in her hand. 14 Interestingly, she puts her fingers to the mouth, invoking an accommodatory schema. 15 Biomechanically, the most important outcome of such behaviour is that it leads to anticipatory control: if a social cue can trigger recognition of what is coming, the infant can gain a deeper sense of what social life is like. 16 Of course, we may still rely on biases akin to those which allow, say, pigeons and monkeys to draw on their own ways of representing quantity to gain practical understanding of human number systems (e.g. Gallistel, 1980; Boyson et al., 1996). 17 Contemporary history ceases to be that which “nacsce immediatamente sull’atto che viene compiendo” and, in so doing, becomes ‘come la coscienza dell’atto’.

References

Agar, M. (1994) Language Shock: Understanding the Culture of Conversation, New York: William Morrow. Aiello, L.C. (1998) ‘The foundations of human language’, in N.G. Jablonski and L.C. Aiello (eds) The Origin and Diversification of Language, pp. 21–34, Wattis Symposium Series in Anthropology. Memoirs of the California Academy of Sciences, no. 24 (San Francisco, Cal.). Anderson, B. (1991[1983]) Imagined Communities: Reflections on the Origin and Spread of Nationalism, London, New York: Verso. Ankersmit, F.R. (1981) Narrative Logic. A Semantic Analysis of the Historian’s Language, Groningen: Krips ReproMeppel. Ankersmit, F.R. (1989) ‘The use of language in the writing of history’. Reprinted in Ankersmit (1994). Ankersmit, F.R. (1994) History and Tropology. The Rise and Fall of Metaphor, Berkeley, CA: University of California Press. Austin, J.L. (1979 [1961]) ‘The meaning of a word’, in J.O. Urmson (ed.) Philosophical Papers, pp. 55–75, Oxford: Oxford University Press. Azéma, J.-P. and Winock, M. (1970) Naissance et mort, Paris: Calmann-Lévy. Balibar, R. (1985) L’institution du français. Essai sur le colinguisme des carolingiens à la république, Paris: Presses Universitaires de France. Ballard, D., Hayhoe, M., Pook, P. and Rao, R. (1997) ‘Deictic codes for the embodiment of cognition’, Behavioral and Brain Sciences, 20, pp. 723–742. Baller, F.W. (1913) The ABC of Chinese writing, Shanghai: China Inland Mission. [Basic Law] (1992). The Basic Law of the Hong Kong Special Administrative Region of the People’s Republic of China, Hong Kong: One Country Two Systems Economic Research Institute. Bateson, M. (1979) ‘The epigenesis of conversational interaction: a personal account of research development’, in M. Bullowa (ed.) Before Speech: The Beginning of Interpersonal Communication, pp. 63–77, Cambridge: Cambridge University Press. Bechhofer, F., McCrone, D., Kiely, R. and Stewart, R. (1999) ‘Constructing national identity: arts and landed elites in Scotland’, Sociology, 22, pp. 515–534. Bickerton, D. (1995) Language and Human Behavior (The Jessie and John Danz Lectures), Seattle, WA: University of Washington Press. Billig, M. (1995) Banal Nationalism, London: Sage. Blake, N.F. (ed.) (1992) The Cambridge History of the English Language II. 1066–1476, Cambridge: Cambridge University Press.

224 References Blakney, R.B. (1935) A Course in the Analysis of Chinese Characters, Shanghai, China: The Commercial Press. Blank, A. (1999) ‘Why do new meanings occur? A cognitive typology of the motivations for lexical semantic change’, in A. Blank and P. Koch (eds) Historical Semantics and Cognition, pp. 61–90, Berlin and New York: Mouton de Gruyter. Blank, A. and Koch, P. (1999) Historical Semantics and Cognition, Berlin and New York: Mouton de Gruyter. Boltz, W.G. (1994) The Origin and Early Development of the Chinese Writing System, New Haven, CT: American Oriental Society. Bourdieu, P. (1971) ‘Intellectual field and creative project’, in M. Young (ed.) Knowledge and Control: New Directions for the Sociology of Education, pp. 161–188, London: Collier MacMillan. Bourdieu, P. (1991) Language and Symbolic Power: The Economy of Linguistic Exchanges, J.B. Thompson (ed.) (tr. G. Raymond and M. Adamson), Cambridge: Polity Press. Bowlby, J. (1969) Attachment [Vol. 1 of Attachment and Loss], London: Hogarth Press. Boyd, R. and Richerson, P. (1996) ‘Why culture is common but cultural evolution is rare’, Proceedings of the British Academy, 88, pp. 77–93. Boysen, S., Bernston, G., Hannan, M. and Cacioppo, J. (1996) ‘Quantity-based inference and symbolic representation in chimpanzees (Pan troglodytes)’, Journal of Experimental Psychology and Animal Behavior Processes, 22, pp. 76–86. Bråten, S. (ed.) (1998) Intersubjective Communication in Early Ontogeny, Cambridge: Cambridge University Press. Bréal. M. (1991) (ed. and tr. G. Wolf) The Beginnings of Semantics: Essays, Lectures and Reviews, London: Duckworth. Bullowa, M. (ed.) (1979) Before Speech: The Beginning of Interpersonal Communication, Cambridge: Cambridge University Press. Bury, J.B. (1903) An Inaugural Lecture, Delivered in the Divinity School, Cambridge, on January 26, 1903, Cambridge: Cambridge University Press. Cameron, D. (1995) Verbal Hygiene, London: Routledge. Chalfant, F.H. (1906) Early Chinese Writing, Pittsburgh, PA: Carnegie Museum. Charle, C, (1993) ‘Academics or intellectuals? The professors of the University of Paris and political debate in France from the Dreyfus affair to the Algerian war’, in J. Jennings (ed.) Intellectuals in Twentieth Century France: Mandarins and Samurais, pp. 94–116, New York: Academic Press. Cheyney D. and Seyfarth, R. (1990) How Monkeys See the World: Inside the Minds of Another Species, Chicago, IL: Chicago University Press. Chomsky, N. (1966) Cartesian Linguistics: A Chapter in the History of Rationalist Thought, Lanham, MD: University Press of America. Chomsky, N. (1975) Reflections on Language, New York: Random House. Chomsky, N. (1979) Language and Responsibility, New York: Pantheon Books. Chomsky, N. (1982) Noam Chomsky on the Generative Enterprise: A Discussion with Riny Huybrechts and Henk van Riemsdijk, Dordrecht: Foris. Chomsky, N. (1987) ‘Language and freedom’, in J. Peck (ed.) The Chomsky Reader, pp. 139–155, New York: Pantheon Books. Churchland, P. (1996) The Engine of Reason, the Seat of the Soul: A Philosophical Journey into the Brain, Cambridge, MA: MIT Press. Clark, A. (1997) Being There: Putting Brain, Body and World Together Again, Cambridge, MA: MIT Press. Clark, A. (2001) Mindware: An Introduction to the Philosophy of Cognitive Science, Oxford: Oxford University Press.

References

225

Cohen, M.C.R. (1955) Cinquante années de recherches: linguistiques, ethnographiques, Paris: Klincksieck. Collingwood, R.G. (1946) The Idea of History, Oxford: Clarendon Press. Cooper, T. (1826) Lectures on the Elements of Political Economy, Columbia, MO: Doyle E. Sweeny. Cowley, S.J. (1994) ‘Conversational functions of rhythmical patterning: a behavioral perspective’, Language and Communication, 14, pp. 353–376. Cowley, S.J. (1997) ‘Conversation, co-ordination and vertebrate communication’, Semiotica, 115, pp. 27–52. Cowley, S.J. (1998) ‘Of timing, turn-taking and conversations’, Journal of Psycholinguistic Research, 27, pp. 541–571. Cowley, S.J. (2001) ‘Prosody and pedagogy in a democratic South Africa’, Southern African Linguistics and Applied Language Studies, 19, pp. 179–196. Cowley, S.J. (2003) ‘Distributed cognition at three months: caregiver-infant dyads in KwaZulu Natal’, Alternation, 10(2), pp. 229–257. Cowley, S.J. and Spurrett, D. (2003) ‘Putting apes (body and language) together again’, Language Sciences, 25, pp. 289–318. Cowley, S.J., Moodley, S. and Fiori-Cowley, A. (2004) ‘Grounding signs of culture: primary intersubjectivity in social semiosis’, Mind, Culture and Activity, 11(2), pp. 109–132. Creamer, T. (1991) ‘Chinese lexicography’, in F.J. Hausmann, O. Reichmann, H. Weigand and L. Zgusta (eds) Dictionaries: An International Encyclopedia of Lexicography, vol. 3, pp. 2592–2611, Berlin, New York: Mouton de Gruyter. Croce, B. (1938) La storia come pensiero e come azione, Bari: Laterza. Croce, B. (1960) History, Its Theory and Practice (tr. D. Ainslie), New York: Russell and Russell. Crowley, T. (1989) The Politics of Discourse: The Standard Language Question in British Cultural Debates, London: Macmillan. Darwin, C. (1859) The Origin of Species by Means of Natural Selection; or, The Preservation of Favoured Races in the Struggle for Life, London: Murray. Davis, H. (2001) Words: An Integrational Approach, London: Curzon. Dawkins, R. (1982) The Extended Phenotype: The Long Reach of the Gene, Oxford: Oxford University Press. Deacon, T. (1997) The Symbolic Species: The Co-evolution of Language and the Human Brain, London: Penguin. DeFrancis, John (1950) Nationalism and Language Reform in China, New Jersey: Princeton. DeFrancis, John (1984) The Chinese Language: Fact and Fantasy, Honolulu, HI: University of Hawaii Press. Dennett, D. (1987) The Intentional Stance. Cambridge, MA: MIT Press. Dennett, D. (1995) Darwin’s Dangerous Idea: Evolution and the Meaning of Life, New York: Simon Schuster. Deumert, A. (2003) ‘Bringing speakers back in? Epistemological reflections on speaker-oriented explanations of language change’, Language Sciences, 25, pp. 15–76. Deutsch, K.W. (1953) Nationalism and Social Communication: An Inquiry into the Foundations of Nationality, Cambridge, MA: Technology Press of the Massachusetts Institute of Technology; New York: Wiley. Dewey, J. (1896) ‘The reflex arc concept in psychology’, Psychological Review, 3, pp. 357–370.

226 References Donald, M. (1991) Origins of the Modern Mind: Three Stages in the Evolution of Culture and Cognition, Cambridge, MA: Harvard University Press. Dunbar, R. (1996) Grooming, Gossip, and the Evolution of Language, Cambridge, MA: Harvard University Press. Du Ponceau, P.S. (1838) A Dissertation on the Nature and Character of the Chinese System of Writing in a Letter to John Vaughan, Philadelphia, PA: M’Carty and Davis. Dury, R. (2001) ‘The history of the English language in the context of the history of the European languages’, in D. Kastovsky and A. Mettiger (eds) Language Contact in the History of English, pp. 69–96, Frankfurt: Lang. Edelman, G. and Tononi, G. (2000) Consciousness: How Matter Becomes Imagination, Harmondsworth: Penguin. Edkins, J. (1876) Introduction to the Study of Chinese Characters, London: Trübner. Edmonds, J.M. (1906) An Introduction to Comparative Philology for Classical Students, Cambridge: Cambridge University Press. Elman, J., Bates, E., Johnson, M., Karmiloff-Smith, A., Parisi, D. and Plunkett, K. (1996) Rethinking Innateness: A Connectionist Perspective on Development, Cambridge, MA: MIT Press. Feldman, M. and Cavalli-Sforza, L. (1989) ‘On the theory of evolution under genetic and cultural transmission with application to the lactose absorption problem’. in M. Feldman (ed.) Mathematical Evolutionary Theory, Princeton, NJ: Princeton University Press. Fenellosa, E. and Pound, E. (1936) The Chinese Written Character as a Medium for Poetry, San Francisco, CA: City Lights. Fichte, J.G. (1808) Reden an dieDeutsche Nation, Berlin: Realschulbuchhandlung. [English version (1968) Addresses to the German Nation (ed. G.A. Kelly, tr. R.F. Jones, G.H. Turnbull), New York: Harper Torch Books]. Firth, J.R. (1957[1950]) ‘Personality and language in society’, in Papers in Linguistics, 1934–1951, pp. 177–189, London: Oxford University Press. Fish, S. (1980) Is There a Text in This Class? Cambridge, MA: Harvard University Press. Fishman, J.A. (1991) Reversing Language Shift, Clevedon: Multilingual Matters. Fisiak, J. (ed.) (2002) Studies in English Historical Linguistics and Philology. A Festschrift for Akio Oizumi, Frankfurt: Lang. Fiumara, G.C. (1990) The Other Side of Language: A Philosophy of Listening, New York: Routledge. Fiumara, G.C. (1995) The Metaphoric Process: Connections between Language and Life, New York: Routledge. Fleischman, S. (2000) ‘Methodologies and ideologies in historical linguistics: on working with older languages’, in S.C. Herring, P. Van Reenan and L. Schløsler (eds) Textual Parameters in Older Languages, pp. 33–58, Amsterdam: Benjamins. Fodor, J. (1983) The Modularity of Mind, Cambridge, MA: MIT Press. Fourmont, E. (1737) Meditationes Sinicae, Paris: Quis Edidit. Francis, G. (1993) ‘A corpus-driven approach to grammar: principles, methods and examples’, in M. Baker, G. Francis and E. Tognini-Bonelli (eds) Text and Technology, pp. 137–156, Amsterdam: Benjamins. Fritz, G. (1998) Historische Semantik, Stuttgart, Weimar: Metzler. Gal, S. (1979) Language Shift: Social Determinants of Linguistic Change in Bilingual Austria, New York: Academic Press. Gallistel, C. (1980) The Organization of Learning, Cambridge, MA: MIT Press. Gardiner, P. (1952) The Nature of Historical Explanation, Oxford: Clarendon Press.

References

227

Geeraerts, D. (1997) Diachronic Prototype Semantics, Oxford: Clarendon Press. Gellner, E. (1964) Thought and Change, London: Weidenfield and Nicolson. Gellner, E. (1973) ‘Scale and nation’, Philosophy of the Social Sciences, 3, pp. 1–17. Gellner, E. (1983) Nations and Nationalism, Ithaca, NY: Cornell University Press. Gellner, E. (1997) Nationalism, London: Weidenfeld & Nicolson. Gibbs, R.W. (1994) The Poetics of Mind: Figurative Thought, Language and Understanding, Cambridge: Cambridge University Press. Giles, H.A. (1880) Freemasonry in China, Amoy: Xiamen. Gleason, A. (1997) ‘96 theses’ in A.K. Melby (ed.) The Twenty-third LACUS Forum 1996. Gould, S.J. (1987) Times’s Arrow, Time’s Cycle. Myth and Metaphor in the Discovery of Geological Time, Cambridge, MA: Harvard University Press. Gould, S.J. (1991a) ‘The case of the creeping fox terrier clone’, Bully for Brontosaurus. Reflections in Natural History, pp. 155–167, London: Hutchinson Radius. Gould, S.J. (1991b) ‘Life’s little joke’, Bully for Brontosaurus. Reflections in Natural History, pp. 168–181, London: Hutchinson Radius. Gould, S.J. (1991c) Bully for Brontosaurus. Reflections in Natural History, London: Hutchinson Radius. Grillo, R.D. (1989) Dominant Languages. Language and Hierarchy in Britain and France, Cambridge: Cambridge University Press. Hager, G. (1801) An Explanation of the Elementary Characters of the Chinese, with an Analysis of Their Ancient Symbols and Hieroglyphics, London: Printed for Richard Philips by T. Bensley. Harbaugh, R. (1998) Chinese Characters: A Genealogy and a Dictionary. . Zhongwen.com. Harpham, G.G. (2002) Language Alone: The Critical Fetish of Modernity, Routledge: New York. Harris, R. (1981) The Language Myth, London: Duckworth. Harris, R. (1993) ‘Integrational linguistics’, in A. Crochetière, J.-C. Boulanger and C. Ouellon (eds) Actes du XVème congrès international des linguists, pp. 321–323, Ste-Foy: Presses de l’Université Laval. Harris, R. (1996) Signs, Language and Communication: Integrational and Segregational Approaches, London: Routledge. Harris, R. (1998) Introduction to Integrational Linguistics, Oxford: Pergamon. Harrison, C. (1997) ‘Semantic specification/semantic emergence: against the container metaphor of meaning’, in A.K. Melby (ed.) The twenty-third LACUS Forum 1996. Haspelmath, M. (1997) From Space to Time. Temporal Adverbials in the World’s Languages, München, Newcastle: LINCOM Europa. Haugen, E. (1966) Language Conflict and Language Planning. The Case of Modern Norwegian, Cambridge, MA: Harvard University Press. Hauser, M., Chomsky, N. and Fitch, W. (2002) ‘The faculty of language: what is it, who has it, and how did it evolve?’ Science, 298, pp. 1569–1579. Heidegger, M. (1971) On the Way to Language (tr. P.D. Hertz), Cambridge, MA: Harper and Row. Hendriks-Jansen, H. (1996) Catching Ourselves in the Act: Situated Activity, Interactive Emergence and Human Thought, Cambridge, MA: MIT Press. Herbst, T. (1996) ‘What are collocations: sandy beaches or false teeth?’, English Studies, 4, pp. 379–393. Herring, S.C., van Reenen, P. and Schløsler, L. (eds) (2000) Textual Parameters in Older Languages, Amsterdam: Benjamins.

228 References Hexter, J.H. (1967) ‘The rhetoric of history’, History and Theory, 6, pp. 3–13. Hickey, R. (1999) ‘The phonology of gender in Modern German’, in M. Rissanen and B. Unterbeck (eds) Gender: Cross-linguistic Studies, Berlin: Mouton de Gruyter. Hinde, R. (1983). Primate social relationships: An integrated approach, Oxford: Blackwell. Hobsbawm, E.J. (1990) Nations and Nationalism since 1780: Programmes, Myth, Reality, Cambridge: Cambridge University Press. Hoey, M. (2002) ‘A theory of lexical primacy’, Talk delivered to the English Language Research seminar, University of Birmingham, November 2002. Hopkins, L.C. (ed. and tr) (1954) The Six Scripts or the Principles of Chinese Writing by Tai T’ung, Cambridge: Cambridge University Press. Humboldt, W. von (1821) ‘On the historian’s task’, History and Theory, 6 (1967), pp. 57–71. Hundsnurscher, F. (1988) ‘Über den Zusammenhang des Gebrauchs der Wörter. Eine methodologische Untersuchung anhand des deutschen Adjektivs GRÜN’, Poetica, 28, pp. 75–103. Hunston, S. and Francis, G. (2000) Pattern Grammar: A Corpus-driven Approach to the Lexical Grammar of English, Amsterdam: Benjamins. Hurley, S. (1998) Consciousness in Action, Cambridge, MA: Harvard University Press. Hutchins, E. (1995) Cognition in the Wild, Cambridge, MA: MIT Press. Ishiguro, H. Ono, T., Imai, M., Maeda, T., Kanda, T. and Nakatsu, R. (2001) ‘Robovie: an interactive humanoid robot’, Journal of Industrial Robotics, 28, pp. 498–503. [Web version at: http://www.mis.atr.co.jp/~kanda/industrial-roboticsrobovie.pdf] Janda, M. (2006) Hermes im Rahmen der indogermanischen Religionsgeschichte, Münster. Johnson, M. (1981) Philosophical perspectives on metaphor, Minneapolis, MN: University of Minnesota Press. Johnson, M. (1997) Developmental Cognitive Neuroscience: An Introduction, Oxford: Blackwell. Jones, W. (1993[1784]) Discourses Delivered at the Asiatick Society. Reprinted in R. Harris (ed.) British Linguistics in the Eighteenth Century, London: Routledge/ Thoemmes. Jones, W. (1993[1786]) Discourses Delivered at the Asiatick Society. Reprinted in R. Harris (ed.) British Linguistics in the Eighteenth Century, London: Routledge/ Thoemmes. Joseph, J.E. (1987) Eloquence and Power: The Rise of Language Standards and Standard Languages, London: Pinter Joseph, J.E. (1997) ‘The “language myth” myth; or, Roy Harris’s red herrings’, in G. Wolf and N. Love (eds) Linguistics Inside Out: Roy Harris and his Critics, pp. 9–41, Amsterdam, Philadelphia, PA: Benjamins. Joseph, J.E. (1998) ‘Why isn’t translation impossible?’, in S. Hunston (ed.) Language At Work: Selected Papers from the Annual Meeting of the British Association for Applied Linguistics held at the University of Birmingham, September 1997, pp. 86–97, Clevedon: Multilingual Matters. Joseph, J.E. (2000) Limiting the Arbitrary: Linguistic Naturalism and its Opposites in Plato’s Cratylus and Modern Theories of Language, Amsterdam, Philadelphia, PA: Benjamins. Joseph, J.E. (2001) ‘Rethinking linguistic creativity’, H.G. Davis and T.J. Taylor (eds) Rethinking Linguistics, pp. 121–150, London: Routledge.

References

229

Joseph, J.E. (2003) ‘Orthodox unorthodoxy’, Language Sciences, 25, pp. 99–109. Joseph, J.E. (2004) Language and Identity: National, Ethnic, Religious, Basingstoke, New York: Palgrave Macmillan. Joyce, P. (1995) ‘The end of social history’, Social History, 20(1), pp. 73–91, reprinted in John Tosh, (ed.), Historians on History: A Reader, London: Longman (2000), pp. 271–288. Kang, C.H. and Nelson, E.R. (1979) The Discovery of Genesis: How the Truths of Genesis Were Found Hidden in the Chinese Language, St Louis, MO: Concordia Publishing House. Karmiloff-Smith, A. (1993) Beyond Modularity: A Developmental Perspective on Cognitive Science, Cambridge, MA: MIT Press. Kastovsky, D. and Mettinger, A. (eds) (2001) Language Contact in the History of English, Frankfurt: Peter Lang Verlag. Kaye, K. (1982) The Mental and Social Life of Babies: How Parents Create Persons, London: Methuen. Kedourie, E. (1993 [1960]) Nationalism (4th ed.), Oxford: Blackwell. Keelan, F.X. (1967) Chinese Characters Explained, Taiwan: Kuangchi Press. Keller, R. (1994) On Language Change: The Invisible Hand in Language (tr. B. Nerlich), London: Routledge. Keller, R. (1997) ‘Kodewandel’, in R. Posner, K. Robering and T.A. Sebeok (eds) Semiotik. Semiotics, vol. 1, pp. 414–435, Berlin, New York: Mouton de Gruyter. Kelly, L.G. (2002) The Mirror of Grammar: Theology, Philosophy and the Modistae, Amsterdam, Philadelphia, PA: Benjamins. Kilpert, D.M. (2002) Language and Value: The Place of Evaluation in Linguistic Theory. Unpublished PhD thesis, Rhodes University, Grahamstown. Kloss, H. (1929) Nebensprachen, Vienna: Braumüller. Kloss, H. (1952) Die Entwicklung neuer germanischer Kultursprachen von 1800–1950, Munich: Pohl & Co. Kloss, H. (1978) Die Entwicklung neuer germanischer Kultursprachen seit 1800, Düsseldorf: Pädogogischer Verlag Schwann. Koch, P. (1999) ‘Cognitive aspects of semantic change and polysemy: the semantic space HAVE/BE’, in A. Blank and P. Koch (eds) Historical Semantics and Cognition, pp. 279–306, Berlin and New York: Mouton de Gruyter. Kohn, H. (1944) The Idea of Nationalism: A Study in its Origins and Background, New York: Macmillan. Koopman, W., van der Leek, F., Fischer, O. and Eaton, R. (eds) (1987) Explanation and Linguistic Change, Amsterdam: John Benjamins. Krefeld, T. (1999) ‘Cognitive ease and lexical borrowing: the recategorization of body parts in Romance’, in A. Blank and P. Koch (eds) Historical Semantics and Cognition, pp. 259–278, Berlin and New York: Mouton de Gruyter. Krippes, K. (1988) ‘Meillet, the researcher and the teacher’, Histoire Epistémologie Langage, 10, pp. 277–283. Kronasser, H. (1968) Handbuch der Semasiologie. Kurze Einführung in die Geschichte, Problematik und Terminologie der Bedeutungslehre, Heidelberg: Winter. Kugiumutzakis, G. (1999) ‘Genesis and development of mimesis in early imitation of early facial and vocal models’, in J. Nadel, and G. Butterworth (eds) Imitation in Infancy, pp. 36–59. Cambridge: Cambridge University Press. Kuhl, P. (1998) ‘Language culture and intersubjectivity: the creation of shared perception’, in S. Brjãaten (ed.) Intersubjective Communication in Early Ontogeny, pp. 63–88, Cambridge: Cambridge University Press.

230 References Kuhl, P. and Miller, J. (1978) ‘Speech perception by the chinchilla: identification functions for synthetic VOT stimuli’, Journal of the Acoustical Society of America, 63, pp. 905–917. LAEME ⫽ Laing, M., Lass, R. and Williamson, K. (in preparation) A Linguistic Atlas of Early Middle English, Edinburgh: Institute for Historical Dialectology. Lai, T.C. (1980). Chinese Characters, Hong Kong: Swindon Book Co. Laing, M. (2000) ‘Never the twain shall meet. Early Middle English – the east-west divide’, in I. Taavitsainen, T. Nevalainen, T. Pahta and M. Rissanen (eds) Placing Middle English in Context, pp. 97–124, Berlin: Mouton de Gruyter. Laing, M. and Lass, R. (2003) ‘Tales of the 10001 nists: The phonological implications of litteral substitution sets in thirteenth-century Southwest Midland texts’, English Language and Linguistics, 7, pp. 257–278. Laing, M. and Lass, R. (forthcoming) ‘Early Middle English dialectology: problems and prospects’, in A. van Kemenade and B.L. Los (eds) Handbook of the History of English, Oxford: Blackwell. Laland, K., Odling-Smee, J. and Feldman, M. (2000) ‘Niche construction, biological evolution and cultural change’, Behavioural and Brain Sciences, 23, pp. 131–175. LALME ⫽ McIntosh, A., Samuels, M.L. and Benskin, M. (1986) A Linguistic Atlas of Late Mediaeval English (4 vols), Aberdeen: Aberdeen University Press. Lamb, S.M. (1999) Pathways of the Brain: The Neurocognitive Basis of Language, Amsterdam, Philadephia, PA: John Benjamins. Langacker, R. (1990) Concept, Image and Symbol: The Cognitive Basis of Grammar, Berlin, New York: Mouton de Gruyter. Langacker, R. (1999) ‘Losing control: grammaticization, subjectification, and transparency’, in A. Blank and P. Koch (eds) Historical Semantics and Cognition, pp. 147–176, Berlin and New York: Mouton de Gruyter. Lass, R. (1976) English Phonology and Phonological Theory, Cambridge: Cambridge University Press. Lass, R. (1980) On Explaining Language Change, Cambridge: Cambridge University Press. Lass, R. (1987) ‘Language, speakers, history and drift’, in W. Koopman, F. Vander Leek, O. Fischer and R. Eaton (eds) Explanation and Linguistic Change, pp. 151–176, Amsterdam: John Benjamins. Lass, R. (1991) ‘Of data and “datives”: Ruthwell Cross rodi again’. Neuphilologische Mitteilungen, 92, pp. 395–403. Lass, R. (1992a) ‘Phonology and morphology’, in N.F. Blake (ed.) The Cambridge History of the English Language II. 1066–1476, Cambridge: Cambridge University Press. Lass, R. (1992b) ‘What are language histories histories of?’ in H.-H. Lieb (ed.) Prospects for a New Structuralism, pp. 243–272, Amsterdam: John Benjamins. Lass, R. (1997a) Historical Linguistics and Language Change, Cambridge: Cambridge University Press. Lass, R. (1997b) ‘Why house is an Old English ‘masculine a-stem’, in T. Nevalainen and L. Kahlas-Tarkka (eds) To Explain the Present: Studies in the Changing English Language in Honour of Matti Rissanen, pp. 101–109, Helsinki: Société néophilologique. Lass, R. (1999a) ‘Phonology and morphology’, in R. Lass (ed.) The Cambridge History of English Language III. 1476–1776, Cambridge: Cambridge University Press. Lass, R. (ed.) (1999b) The Cambridge History of the English Language III. 1476–1776, Cambridge: Cambridge University Press. Lass, R. (2000) ‘Language periodization and the concept “middle” ’, in I. Taavitsainen, T. Nevalainen, T. Pahta and M. Rissanen (eds) Placing Middle English in Context, pp. 7–42, Berlin: Mouton de Gruyter.

References

231

Lass, R. and Laing. M. (2004) ‘Did front rounded vowels remain in the Southwest Midlands?’, in N. Ritt, H. Schendl and D. Kastovsky (eds) Rediscovering Middle English Philology, Berlin: Peter Lang Verlag. Levinson, S.C. (1995) ‘Three levels of meaning’, in F.R. Palmer (ed.) Grammar and Meaning: Essays in Honour of Sir John Lyons, pp. 90–115, Cambridge: Cambridge University Press. Levinson, S.C. (2000) Presumptive Meanings: The Theory of Generalized Conversational Implicature, Cambridge, MA: MIT Press. Li Hsiao-ting (1977) A Brief History of Chinese Characters , Taipei: Lien Ching Publishing Co. Li Leyi (1993) Tracing the roots of Chinese characters: 500 Cases (tr. by Wang Chengzhi), Beijing: Beijing Language and Culture University Press. Linell, P. (1990) ‘The power of dialogue dynamics’, in I. Marková and K. Foppa (eds) The Dynamics of Dialogue, pp. 144–177, Hemel Hampstead: Harvester Wheatsheaf. Lizardo, O. (2004) ‘The cognitive origins of Bourdieu’s habitus’, Journal for the Theory of Social Behavior, 34, pp. 375–401. Lodge, R.A. (1993) French: from Dialect to Standard, London: Routledge. Longacre, R.E. (1983) ‘The grammar of discourse’, New York: Plenum. Love, N. (2002) ‘The language myth and historical linguistics’, in R. Harris (ed.) The Language Myth in Western Culture, pp. 25–40, Richmond: Curzon Press. Love, N. (2004) ‘Cognition and the language myth’, Language Sciences, 26, pp. 525–544. Lucy, J. (1993a) ‘Reflexive language and the human disciplines’, in J. Lucy (ed.) Reflexive Language, Cambridge: Cambridge University Press. Lucy, J. (1993b) ‘Metapragmatic presentationals: reporting speech with quotatives in Yucatec Maya’, in J. Lucy (ed.) Reflexive Language, Cambridge: Cambridge University Press. Luschützky, H.C. and Madelska, L. (2001) ‘European integration from a polycentric point of view. How stable is the euro?’, in C. Schaner-Wolles, J. Rennison and F. Neubarth (eds) Naturally!: Linguistic Studies in Honour of Wolfgang Ulrich Dressler Presented on the Occasion of his 60th Birthday, pp. 291–300, Torino: Rosenberg & Sellier [⫽ Linguistica 19]. Lyons, J. (1995) Linguistic Semantics: An Introduction. Cambridge: Cambridge University Press. Macafee, C. (1994) Traditional Dialect in the Modern World: A Glasgow Case Study, Bamberger Beiträge zur englischen Sprachwissenschaft 35, Frankfurt am Main: Peter Lang. McNaughton, W. and Li Ying (1999) Reading and Writing Chinese (2nd ed.), Vermont, Tokyo: Tuttle. Malinowski, B. (1935) ‘The translation of untranslatable words’. Reprinted in A. Nye (ed.) Philosophy of Language: The Big Questions (1998), Malden, MA: Blackwell. Marx, K. (1847) Misère de la philosophie: Réponse à La philosophie de la misère de M. Proudhon, Paris: A. Frank; Brussels: C.G. Vogler. [Anon. Engl. transl., The Poverty of Philosophy, Moscow: Foreign Languages Publishing House, 1955]. Matthews, P. (1993) Grammatical Theory in the United States from Bloomfield to Chomsky, Cambridge: Cambridge University Press. Mehler, D. and Dupoux, L. (1997) What Infants Know, Cambridge: Cambridge University Press Meiland, J.W. (1965) Scepticism and Historical Knowledge, New York: Random House.

232 References Meillet, A. (1903) Introduction à l’étude comparative des langues indo-européennes, Paris: Hachette. Meillet, A. (1918a) ‘La situation linguistique en Russie et en Autriche-Hongrie’, Scientia (Rivista di Scienza), Serie 2, 23, pp. 209–216. Meillet, A. (1918b) ‘Les langues dans le bassin de la mer baltique’, Scientia (Rivista di Scienza), Serie 2, 24, pp. 383–392. Meillet, A. (1920[1913]) Aperçu d’une histoire de la langue grecque (2nd ed.), Paris: Librairie Hachette. Meillet, A. (1925) La méthode comparative en linguistique historique. Oslo: Aschehoug. Meillet, A. (1928[1918]) Les langues dans l’Europe nouvelle (2nd ed.), Paris: Payot. Meillet, A. (1948[1928]) Esquisse d’une histoire de la langue latine (5th edn.), Paris: Librairie Hachette. Meillet, A. and Vaillant, A. (1924) Grammaire de la langue serbo-croate. Collection de grammaires de l’institut d’Etudes slaves 3, Paris: Librairie ancienne Honoré Champion. Melzoff, A. and Moore, M. (1997) Explaining facial imitation: a theoretical model. Early Development and Parenting, 6, pp. 179–192. Mettam, R. (1988) Power and Faction in Louis XIV’s France, Oxford: Blackwell. Millar, R.M. (2002) ‘After Jones: some thoughts on the final collapse of the grammatical gender system in English’, in J. Fisiak (ed.) Studies in English Historical Linguistics and Philology. A Festschrift for Akio Oizumi, pp. 267–292, Frankfurt: Lang. Milroy, J. (1992) Linguistic Variation and Change: On the Historical Sociolinguistics of English, Oxford: Blackwell. Milroy, J. (2000) ‘Historical description and the ideology of the standard language’, in L. Wright (ed.) The Development of Standard English 1300–1800: Theories, Descriptions, Conflicts, pp. 11–28, Cambridge: Cambridge University Press. Mink, L.O. (1987) Historical Understanding, Ithaca, NY: Cornell University Press. Moon, R. (1998) Fixed Expressions and Idioms in English: A Corpus-based Approach, Oxford: Clarendon Press. Morpurgo Davies, A. (1988) ‘Meillet, Greek and the Aperçu’, Histoire Epistémologie Langage, 10, pp. 235–251. Morris, R. (1882) Historical Outlines of English Accidence, Comprising Chapters on the History and Development of the Language, and on Word-formation, London: Macmillan. Müller, F.M. (1864[1861]) Lectures on the Science of Language. Second Series (4th ed.), London: Longman Roberts and Green. Müller, F.M. (1995[1851]) ‘Comparative philology’, Edinburgh Review, 94. Reprinted in Language and Linguistics vol. 3 (The Wellesley Series), London: Routledge/ Thoemmes. Mustanoja, T. (1960) Middle English Syntax. Part I: Parts of Speech, Helsinki: Société néophilologique. Nadel, J., Butterworth, G. (eds) (1999) Imitation in Infancy, Cambridge: Cambridge University Press. Nevalainen, T. and Kahlas-Tarkka, L. (eds) (1997) To Explain the Present: Studies in the Changing English Language in Honour of Matti Rissanen, Helsinki: Société néophilologique. Nye, A. (1998) Philosophy of Language: The Big Questions, Malden, MA: Blackwell. Oakeshott, M. (1933) Experience and Its Modes, Cambridge: Cambridge University Press.

References

233

Oller, J. (2000) The Emergence of the Speech Capacity, Mahwah, NJ: Lawrence Erlbaum. Ong Tee Way and Tan Huay Peng (1980) Fun with Chinese Characters; the Straits Times Collection, vol. 1, Singapore: Federal Publications. Owings, D. and Morton, E. (1998) Animal Vocal Communication: A New Approach, Cambridge: Cambridge University Press. Paraf, P. (1981) La France de 1914. Le passé et l’avenir nous parlent, Paris: Editions du Sorbier. Paul, H. (1920) Prinzipien der Sprachgeschichte (5th ed.), Halle [7th ed. Tübingen: Niemeyer, 1966]. Perrot, J. (1988) ‘Antoine Meillet et les langues de l’Europe: l’affaire hongroise’. Histoire Epistémologie Langage, 10, pp. 301–318. Pfandl, H. (2002) ‘On the grammar and pronunciation of euro and cent’, Linguist internet discussion list, 16 January 2002. Piaget, J. (1952) The Origins of Intelligence in Children, New York: International Universities Press. Piaget, J. (1962) Play, Dreams and Imitation in Childhood, New York: Norton. Pinker, S. (1994) The Language Instinct: The New Science of Language and Mind, London: Penguin. Popper, K.R. (1972) Objective Knowledge: An Evolutionary Approach, Oxford: Clarendon Press. Pound, E. (1969) (ed. and tr.) Confucius. The Great Digest; The Unwobbling Pivot; The Analects, New York: New Directions. Preston, S., de Waal, B. (2002) ‘Empathy and its proximal bases’, Behavioral and Brain Sciences, 25, pp. 1–20. Quine, W. (1960) Word and Object, Cambridge, MA: MIT Press. Radney, J.R. (1999) Confluence, Convergence, and Contextualization in Philosophy and Linguistics. Unpublished Ph.D. dissertation, University of Texas at Arlington. Ramus, F., Hauser, M., Miller, C., Morris, D. and Mehler, J. (2000) ‘Language discrimination by human newborns and by cotton-top tamarin monkeys’, Science, 288, pp. 349–351. Rashidi, L.S. (1997) ‘Functional passive rethought’, in A.K. Melby (ed.) The twentythird LACUS Forum 1996, pp. 69–73, Chapel Hill: LACUS. Reagan, Charles E. and Stewart, D. (1978) The Philosophy of Paul Ricoeur. Boston, MA: Beacon Press. Renan, E. (1858[1848]) De l’origine du langage, Paris: Michel Lévy, Frères. Renan, E. (1882) Qu’est-ce qu’une nation? Conférence faite en Sorbonne, le 11 mars 1882, Paris: Calmann Lévy. Rendell, L. and Whitehead, H. (2001) ‘Culture in whales and dolphins’, Behavioral and Brain Sciences, 24, pp. 309–382. Rissanen, M. and Unterbeck, B. (eds) (1999) Gender: Cross-linguistic Studies, Berlin: Mouton de Gruyter. Rizzolatti, G. and Arbib, M. (1998) ‘Language within our grasp’, Trends in Neurosciences, 21, pp. 188–194. Rochat, P. (2001) The Infant’s World, Cambridge, MA: Harvard university Press. Rowlands, M. (1999). The Body in Mind, Cambridge: Cambridge University Press. Rowse, A.L. (1946) The Use of History, London: Hodder and Stoughton. Sagart, L. (1999) The Roots of Old Chinese, Amsterdam, Philadelphia, PA: John Benjamins.

234 References Said, E. (1983) The World, The Text and the Critic, Cambridge, MA: Harvard University Press. Sampson, G. (1997) Educating Eve, Cassell: London. Sapir, E. (1921) Language: An Introduction to the Study of Speech, New York: Harcourt Brace & Company. Saussure, F. de (1922 [1916]) Cours de linguistique générale, Paris: Payot. Savage-Rumbaugh, S. and Lewin, R. (1994) Kanzi: The Ape at the Brink of the Human Mind, New York: John Wiley & Sons. Savage-Rumbaugh, S., Shanker, S. and Taylor, T.J. (1998) Apes, Language and the Human Mind, Oxford: Oxford University Press. Schaffer, H. (1977) Studies in Mother – Infant Interaction, London: Academic Press. Schiffman, H.F. (1996) Linguistic Culture and Language Policy, London: Routledge. Schlosser, E. (2001) Fast Food Nation: The Dark Side of the All-American Meal, New York: Perennial. Scott, M. (1999) WordSmith Tools, Oxford: Oxford University Press. Seton-Watson, H. (1977) Nations and States: An Enquiry into the Origins of Nations and the Politics of Nationalism, London: Methuen. Shafer, B.C. (1955) Nationalism: Myth and Reality, London: Gollancz. Shi Zhengyu (1997) Picture Within a Picture: An Illustrated Guide to the Origins of Chinese Characters, Beijing: New World Press. Shlain, L. (1998) The Alphabet versus the Goddess, London: Penguin. Shotwell, J.T. (1910) ‘History’, Encyclopaedia Britannica (11th ed.), pp. 527–533, Cambridge: Cambridge University Press. Silverstein, M. (1976) ‘Shifters, linguistic categories, and cultural description’, in K. Basso and H. Selby (eds) Meaning in Anthropology, Albuquerque: University of New Mexico Press. Silverstein M. (1979) ‘Language Structure and Linguistic Ideology’, in P. Clyne, W.F. Hanks and C.L. Hofbauer (eds) The Elements: A Parasession on Linguistic Units and Levels, Chicago, IL: University of Chicago Press. Silverstein, M. (1981) ‘The limits of awareness’, Sociolinguistic Working Paper No. 84, Austin: Southwest Educational Development Laboratory. Silverstein, M. (1985) ‘The culture of language in Chinookan narrative texts: or, on saying that . . . . in Chinook’, in J. Nichols and A. Woodbury (eds) Grammar Inside and Outside the Clause, Cambridge: Cambridge University Press. Silverstein, M. (1993) ‘Metapragmatic discourse and metapragmatic function’, in J. Lucy (ed.) Reflexive Language, Cambridge: Cambridge University Press. Silverstein, M. (2000) ‘Whorfianism and the linguistic imagination of nationality’, in P.V. Kroskrity (ed.) Regimes of Language: Ideologies, Polities, and Identities, pp. 85–138, Santa Fe: School of American Research Press. Sinclair, J. (1991) Corpus, Concordance, Collocation, Oxford: Oxford University Press. Sinclair, J. (1996) The search for units of meaning, Textus 9, pp. 75–106. Smith, A.D. (1971) Theories of Nationalism, London: Duckworth. Smith, A.D. (1998) Nationalism and Modernism: A Critical Survey of Recent Theories of Nations and Nationalism, London and New York: Routledge. Sommerfelt, A. (1938) ‘Conditions de la formation d’une langue commune’, pp. 42–48. Spitz, R. (1965) The First Year of Life. New York: International Universities Press.

References

235

Spurrett, D. (2003) ‘Why think that cognition is distributed’, Alternation, 10(1), pp. 292–306. Spurrett, D. and Cowley, S.J. (2004) ‘How to do things without words’, Language Sciences, 26(5), pp. 443–466. Sroufe, L. (1995) Emotional Development: The Organization of Emotional Life in the Early Years, Cambridge: Cambridge University Press. Stern, D. (1977) The First Relationship, London: Fontana. Strang, B.M.H. (1970) A History of English, London: Methuen. Strettell, A. (1851) ‘Comparative philology’, North British Review, 16. Reprinted in R. Harris (ed.) British Linguistics in the Eighteenth Century, London: Routledge/ Thoemmes. Stross, B. (1974) ‘Speaking of speaking: Tenejapa Tzeltal metalinguistics’, Explanations in the Ethnography of Speaking, Cambridge: Cambridge University Press. Stubbs, M. (2001a) Words and Phrases, Oxford: Blackwell. Stubbs, M. (2001b) ‘Texts, corpora, and problems of interpretation: a response to Widdowson’, Applied Linguistics, 22, pp. 149–172. Sweetser, E.E. (1990) From Etymology to Pragmatics. Metaphorical and Cultural Aspects of Semantic Structure, Cambridge: Cambridge University Press. Taavitsainen, I., Nevalainen, T., Pahta, T., Rissanen, M. (eds) (2000) Placing Middle English in Context, Berlin: Mouton de Gruyter. Taithe, B. (1999) Defeated Flesh. Welfare, Warfare and the Making of Modern France, Manchester: Manchester University Press. Tannen, D. (1989) Talking Voices: Repetition, Dialogue and Imagery in Conversational Discourse, Cambridge: Cambridge University Press. Taylor, T.J. (1990a) ‘Normativity and linguistic form’, in H.G. Davis and T.J. Taylor (eds) Redefining Linguistics, pp. 118–148, London: Routledge,. Taylor, T.J. (1990b) ‘Which is to be master? The institutionalization of authority in the science of language’, in J.E. Joseph and T.J. Taylor (eds) Ideologies of Language, pp. 9–26, London: Routledge. Taylor, T.J. (1997) Theorizing Language: Analysis, Normativity, Rhetoric, History, Oxford: Pergamon. Taylor, T.J. (2001) ‘Folk psychology and the language myth: what would the integrationist say?’, in R. Harris (ed.) The Language Myth in Western Culture, pp. 100–117, Richmond: Curzon. Teubert, W. (2001) ‘Corpus linguistics and lexicography’, International Journal of Corpus Linguistics, pp. 125–154. Thelen, E. and Smith, L. (1994) A Dynamical Systems Approach to Development and Cognition, Cambridge, MA: MIT Press. Tomasello, M. (1999) The cultural origins of human cognition, Cambridge, MA: Harvard University Press. Toolan, M. (1996) Total Speech: An Integrational Linguistic Approach to Language, Durham: Duke University Press. Toolan, M. (2003) ‘Le politiquement correct dans le monde francais’, Discourse and Society, 14(1), pp. 69–86. Tosh, J. (1984) The Pursuit of History, London: Longman. Traugott, E.C. (1999) ‘The rhetoric of counter-expectation in semantic change: a study in subjectification’, in A. Blank and P. Koch (eds) Historical Semantics and Cognition, pp. 177–196, Berlin and New York: Mouton de Gruyter.

236 References Traugott, E.C. and Dasher, R.B. (2002) Regularity in Semantic Change, Cambridge: Cambridge University Press. Trevarthen, C. (1979) ‘Communication and co-operation in early infancy: a description of primary intersubjectivity’, in M. Bullowa (ed.) Before Speech, pp. 321–347, Cambridge: Cambridge University Press. Trevarthen, C. (1998) ‘The concept and foundations of infant intersubjectivity’, in S. Bråten, (ed.) Intersubjective Ccommunication in Early Ontogeny, pp. 15–46, Cambridge: Cambridge University Press. Trevarthen, C., Kokkinaki, T. and Fiammenghi, G. (1999) ‘What infant imitations can communicate; with mothers, with fathers and with peers’, in J. Nadel and G. Butterworth (eds) Imitation in Infancy, pp. 127–169, Cambridge: Cambridge University Press. Vendryes, J. (1937) ‘Antoine Meillet’. Bulletin de la société de linguistique de Paris, 38, pp. 1–42. Verschueren, J. (1985) What People Say They Do With Words, Norwood: Ablex. Vygotsky, L. (1978) Mind in Society: The Development of the Higher Psychological Processes (ed. M. Cole, V. John-Steiner, S. Scribner, E. Souberman), Cambridge, MA: Harvard University Press. Walrod, M.R. (1988) Normative Discourse and Persuation: An Analysis of Ga’dang Informal Litigation. Special monograph issue no. 26, Manila: Linguistic Society of the Philippines. Wang Kuo-wei (1979) Definitions and Classifications of the Six Scripts According to Hsü Shen (ca. A.D. 58–147) and Leading Ch’ing Scholars. Unpublished Ph.D. dissertation, Georgetown University. Watters, T. (1870) Lao-tzu: A Study in Chinese Philosophy, Hong Kong: China Mail [London: Williams & Norgate]. Weigand, E. (1998) ‘Contrastive lexical semantics’, in E. Weigand (ed.) Contrastive Lexical Semantics, pp. 25–44, Amsterdam, Philadelphia, PA: John Benjamins. Weigand, E. (2001) ‘Competenza interazionale plurilingue’, in S. Cigada, S. Gilardoni and M. Matthey (eds) Comunicare in ambiente professionale plurilingue, pp. 87–105, Lugano: USI. Weigand, E. (2002a) ‘The language myth and linguistics humanised’, in R. Harris (ed.) The Language Myth in Western Culture, pp. 55–83, Richmond: Curzon. Weigand, E. (2002b) Rezension von Blank/Koch (1999). Beiträge zur Geschichte der deutschen Sprache und Literatur. Weigand, E. (2002c) ‘Constitutive features of human dialogic interaction: mirror neurons and what they tell us about human abilities’, in M. Stamenov and V. Gallese (eds) Mirror Neurons and the Evolution of Brain and Language, Amsterdam, Philadelphia, PA: John Benjamins. West, M., King, A. (1996) ‘Social learning: synergy and songbirds’, in B. Galef and C. Heyes (eds) Social Learning in Animals: The Roots of Culture, pp. 155–178, New York: Academic Press. White, H. (1978) Tropics of Discourse, Baltimore: Johns Hopkins University Press. Widdowson, H. (2000) ‘On the limitations of linguistics applied’, Applied Linguistics, 21, pp. 3–25. Wieger, L. (1927) Chinese Characters: Their Origin, Etymology, History, Classification and Signification. A Thorough Study from Chinese Documents (tr. L. Davrout), Ho-kien-fu: Catholic Mission Press [First English edition 1915].

References

237

Wierzbicka, A. (1996) Semantics: Primes and Universals, Oxford: Oxford University Press. Wilder G.D. and J.H. Ingram (1934[1922]) Analysis of Chinese Characters, Peking: College of Chinese Studies. Wilson, E. (2000) Consilience: The Unity of Knowledge, Knopf: New York. Wilson, L.G. (ed.) (1970) Sir Charles Lyell’s Scientific Journals on the Species Question, New Haven: Yale University Press. Wittgenstein, L. (1958[1953]) Philosophical Investigations (tr. G.E.M. Anscombe), Oxford: Blackwell. Wright, L. (ed.) (2000) The Development of Standrad English 1300–1800: Theories, Descriptions, Conflicts, Cambridge: Cambridge University Press. Wright, R. (1994) The Moral Animal: Why We Are the Way We Are: The New Science of Evolutionary Psychology, New York: Pantheon Books. Wu, T.L. (1990) The Origin and Dissemination of Chinese Characters , Taipei: Caves Books. Xie Guang-hui (2000) [Etymological dictionary of Chinese characters], Beijing: Beijing University Press. Xu Shen (121) [Explaining pictographic characters and analyzing composite characters], Hong Kong: Zhonghua shuju, 1972. Ye Jian-yuan (2000) [Ever-changing Chinese characters], Hong Kong: Mingpao. Yin Binyong and John S. Rohsenow (1994) Modern Chinese Characters , Beijing: Sinolingua. Zhu Min-shen (1998) [The Shouwen jiezi and the palaeographic study of ancient Chinese characters], Shanghai: Fudan daixue chubanshe.

Index

Agar, M. 159 Aiello, L.C. 196 Althusser, L. 189 ambiguity 145 analogical levelling 9 analogy (vs homology) 27 Anderson, B. 128, 130–132, 137–138 Ankersmit, F.R. 142–143, 151 Aristotle 75, 151, 152 Armenian 100 Austin, J.L. 18, 164–167 Azéma, J. P. 101 babies, social life of 201 Balibar, R. 101 Baller, F.W. 64 Bechhofer, F. 133 Belgium 108, 116 Bible 189–191 Bickerton, D. 190–191 Billig, M. 132, 135, 140 Blakney, R.B. 61–62, 65, 66–67 Blank, A. 82 Boltz, W.G. 68 Bopp, F. 43, 49–50, 52, 58 Bourdieu, P. 101, 133, 139, 200 Bréal, M. 49 Brecht, B. 19 Bury, J.B. 146 Celtic 49 Chalfant, F.H. 62, 63, 64, 66, 68 Charle, C. 101 Chaucer, G. 19, 26 Cheyney, D. 202 Chinese 2, 61–70 Chinook 168–169 Chomsky, A.N. 189, 191–194, 197, 199 Clark, A. 203, 209

cognitive internalism 207–209 cognitive linguistics 79, 82–84 Cohen, M.C.R. 118 Collingwoord, R.G. 148–149 collocation 181ff. communicative behaviour 71–78 comparative philology 3, 5, 99; historiography of 41–59 convergence 27–30 Cooper, T. 124 corpus linguistics 177ff. correspondence formula 5, 53, 54 cotemporality, principle of 154, 210 Cowley, S.J. 17, 201, 204, 207, 209, 211, 214, 217, 221 Creamer, T. 63 critical discourse analysis 176–178 Croce, B. 151, 220 Crowley, T. 100 Czech 106 Darwin, C. 38, 136 Dasher, R.B. 84, 86, 87, 88, 91, 97 Davis, A.M. 99–100 Davis, H.G. 178–179 Dawkins, R. 203, 215 Deacon, T. 201, 204 DeFrancis, J. 62, 68 Dennett, D. 169, 203, 221 Deumert, A. 18 Deutsch, K.W. 123, 124 dialect 3, 19–20; dialecte (Meillet) 13, 115, 117 dialogic interaction 79–80, 86, 87, 97 discursive constructivism 175 discursive interaction 162ff. distributed cognition 207, 209–210 Dunbar, R. 195–196, 200 Du Ponceau, P.S. 68

240 Index Dupoux, L. 201 Dury, R. 30, 38, 40 Early Modern English 22 Edkins, J. 63 electronic printing 6 Engels, J. 189 englisch 11 English 101; history and historiography of 19–40; as lingua franca 174 epistemology 1, 13, 22 etymology 4, 7, 23, 60–70 euro, euro 15–16, 172–187 Fichte, J.G. 125–127, 130, 136, 139 Finno-Ugrian 58 Firth, J.R. 182–183, 187 Fish, S. 133 Fishman, J.A. 112 Fiumara, G.C. 74, 78 Fleischman, S. 22, 37, font 6–10 Fourmont, E. 68 France 112–113; cultural history of 100–103; political history of 127–128 French 18, 83, 87, 101 Fritz, G. 84 Gal, S. 112 Gardiner, P. 146–148 Geeraerts, D. 83 Gellner, E. 128–130, 134, 141 German 11, 30 Germanic languages 4, 28, 38, 46 Gibbs, R.W. 75 Giles, H.A. 69 Gleason, A. 73 Görlach, M. 38–39 Gothic 21 Gould, S.J. 28, 38 Greek 11, 18, 99, 46, 50, 51, 87, 90, 99 Grillo, R.D. 101 Grimm, J. 43, 46–48, 50, 52, 53, 55 Hager, G. 68 Hall, S. 189 Harbaugh, R. 70 Harpham, G.G. 17, 18, 196 Harris, R. 5, 14, 18, 61, 69, 79, 90, 120, 139, 172, 176, 181, 187, 210, 221 Harrison, C. 72 Haspelmath, M. 89 Haugen, E. 118

Herodotus 150 Hexter, J.H. 143 Hinde, R. 200 historical constructionism 151 historical linguistics 1–17, 19–40; actuation problem in 9; macro- 7, 11; micro- 11; socio- 9 historical narrativism 151–152 historiography 1, 3, 14, 15, 172, 173 history passim; as linear narrative 3, 19–40; vs historiography 1 Hobbes, T. 75 Hobsbawm, E. 134–138, 140 Hoey, M. 182, 184 Homer 89, 91 Humboldt, W. von 130, 143 Hungarian 58 Hutchins, E. 203, 221 Hutton, C. 2 Icelandic 48 ideology 189 illocutionary act (Austin) 165–166 Indo-European 4–5, 42ff., 99–100 Ingram, J.H. 68 integrational linguistics, integrationism 17, 60, 75–78, 142, 144, 146–149, 153, 154, 172–187, 199 integrational semiology 187 invisible hand 11–12, 85 Italian 83, 90 Jackendoff, R. 191 Janda, M. 90 Johnson, M. 75 Jones, W. 42–45, 49, 57 Joseph, J.E. 109, 114, 118, 120, 121, 128, 130, 133, 139, 140, 182 Joyce, P. 175 Kang, C.H. 63, 67 Karmiloff-Smith, A. 201, 221 Kedourie, E. 123, 124, 128–129, 134, 139 Keelen, F.X. 64 Keller, R. 10–12, 84 Kelly, L.G. 120 Kilpert, D. 18 Kloss, H. 114, 118 Koch, P. 79 Kohn, H. 124 Krefeld, T. 82 Kronasser, H. 88

Index 241 Lacan, J. 190 Laclau, E. 189 Lai, T.C. 62, 64, 66 Laing, M. 37, 38, 39, 40 Lamb, S.M. 74, 76, 78 Langacker, R.W. 82, 83, 88, 89, 90 language: biomechanical constraints in 200–222; concept of 190ff. language and culture 200ff. language and history 1–17, 142–155 language myth 81–85, 146 language of history 16; see also philosophy of history langue (Meillet): artificielle 119; de civilisation 13, 105–106, 108, 117; commune 110–113, 115, 117; littéraire 108–109; officielle 117; savante 107–108, 117; vulgaire 117 Lass, R. 3, 22, 24, 30, 34, 37, 38, 39, 40 Latin 46, 50, 51, 54, 83, 87, 90, 109, 120 Levinson, S.C. 84 lexeme 71–72, 78 lexical innovation 72 lexicography 8, 11, 16, 72, 145 Lieberman, P. 195 Li Hsiao-ting 65 Li Leyi 64 lineality 21, 27 Linell, P. 202 linguistic change 2, 6–12, 79–98 linguistic codification 6, 10 linguistic consciousness 9 linguistic description 8 linguistic historiography 2, 3 linguistic history 9; see also historical linguistics linguistic metapragmatics 161ff. linguistic nationalism 120–141; origins of 122–125 linguistic normativity 9 linguistic prescription, prescriptivism 8, 13, 15 linguistics, history of 56–57 linguistic standardisation 25 literal (vs figurative) language 74–75, 76–77 Li Ying 67 Lodge, R.A. 101 Longacre, R.E. 73 Love, N.L. 18, 92 Lucy, J. 163 Lyell, C. 30 Lyons, J. 73

Macafee, C. 112 Macedonian 115 McNaughton, W. 67 Magyar 100 Malinowski, B. 76 Marx, K. 124, 189 Matthews, P.H. 208 Mehler, D. 201 Meiland, J.W. 151 Meillet, A. 12–13, 18, 50, 53–54, 59, 99–119 metahistory, historical metatheory 145, 151ff. metalinguistic reflection 16 metaphor 75 Mettam, R. 119 Middle English 20, 22, 23 Millar, R.McC. 12–13, 18, 39 Milroy, J. 18, 22 Mink, L.O. 152–153 Modern English 22, 23 modernism 191 Morris, R. 31 Müller, F.M. 44, 47, 50, 52 Mustanoja, T. 31 nation 121, 125–127 national identity 121–122, 128, 138, 140, 175 nationalism 13 national language 13, 134–135, 138 Nelson, E.R. 63, 67 neurocognition 75–78 Northumbrian 3, 20–21, 38 Nozick, R. 11 Oakeshott, M. 151 Old English 20–22, 30ff.; gender in 30–37 Old High German 21 Old Testament 122–123 Paraf, P. 101 parler (Meillet) 13, 115, 117 patois (Meillet) 115–116 Paul, H. 79 Perrot, J. 100 philosophy of history 14, 142–155, 156–171 philosophy of language 147 Piaget, J. 201, 204, 220 Pinker, S. 193–197 Plato 120 polysemy 80, 91

242 Index Popper, K.R. 1 Pound, E. 62, 67–68 protolanguage 4, 22

structural semantics 81–82 Stubbs, M. 178, 182 surrogational semantics 147, 153 Sweetser, E.E. 82, 83, 87, 88, 90

Quine, W.V.O. 207 Radney, J.R. 77 Rashidi, L.S. 71 Rask, R. 43, 47–49, 50, 51, 52, 58 reification 5, 22ff., 145 Renan, E. 127–130, 141 reocentric surrogationism 151 Romance 54 Romanian 109 Romanticism 57, 126–127 Rousseau, J.-J. 192 Rowse, A.L. 146 Russian 109, 111 Sagart, L. 70 Said, E. 133 Sanskrit 46, 49, 50, 56, 57 Sapir, E. 29–30, 38, 139 Saussure, F. de 18, 41–42, 48, 53, 154, 175, 191–192, 198 Savage-Rumbaugh, S. 194, 214 Schiffman, H.F. 101 Schleicher, A. 56 Schlosser, E. 61 semantic change 6–10, 79–98 semantic (in)determinacy 61, 79–98 sentence 73–74 Serbo-Croat 110 Seton-Watson, H. 131 Seyfarth, R. 202 Shafer, B.C. 123 Shakespeare, W. 123 Shi Zhengyu 66 Shotwell, J.T. 144, 145 Silverstein, M. 137–138, 159, 161–170 Sinclair, J. 83 Slavonic 100 Slovak 106 Smith, A.D. 102, 132, 141 social behaviour 201ff. sociology of language 99–119 Sommerfelt, A. 118 sound change 50 Spurrett, D. 207, 209, 214, 217, 221 Strang, B.M.H. 38 Strettell, A. 45, 51–53 Stross, B. 159

Tagalog 71 Taithe, B. 101 Tajfel, H. 132–133 Tannen, D. 116 Taylor, T.J. 14–15, 18, 157, 169 Tomasello M. 203, 204 Toolan, M. 15–16, 76, 78, 178, 182 Tosh, J. 175 Toynbee, A. 145 Traugott, E.C. 84, 86, 87, 88, 89, 91, 97 Trevarthen, C. 200, 201, 202, 212, 222 Ukrainian 111 Vendryes, J. 113 verba dicendi 15, 158–171 Verschueren, J. 159 Vygotsky, L. 201 Walrod, M. 7, 18 Weigand, E. 18, 79, 80, 82, 86, 87, 91, 97 Welsh 18 West Saxon 3, 20–21 White, H. 152 Whorf, B.L. 137, 139 Widdowson, H. 178 Wieger, L. 66, 68 Wierzbicka, A. 159 Wilde, O. 142 Wilder, G.D. 68 Williams, R. 189 Winock, M. 101 Wittgenstein, L. 143, 200 Wolf, S.G. 49, 178 Wright, R. 196 writing, Chinese 61–70 Wu, T.L. 65, 67 Xie Guang-hui 64, 67 Xu Shen 62 Ye Jian-yuan 67 Zhu Min-shen 70 Zizek, S. 189 Zulu 204ff.

eBooks – at www.eBookstore.tandf.co.uk

A library at your fingertips!

eBooks are electronic versions of printed books. You can store them on your PC/laptop or browse them online. They have advantages for anyone needing rapid access to a wide variety of published, copyright information. eBooks can help your research by enabling you to bookmark chapters, annotate text and use instant searches to find specific words or phrases. Several eBook files would fit on even a small laptop or PDA. NEW: Save money by eSubscribing: cheap, online access to any eBook for as long as you need it.

Annual subscription packages We now offer special low-cost bulk subscriptions to packages of eBooks in certain subject areas. These are available to libraries or to individuals. For more information please contact [email protected] We’re continually developing the eBook concept, so keep up to date by visiting the website.

www.eBookstore.tandf.co.uk