The Linguistics of History 9781474468909

Citation preview

The Linguistics of History

Marx was fond o f quoting Heraclitus: panta rei, all things move. This is the one truth we seek to recapture when we write history. We know that our version, being set into words, is itself false. A. J. P. Taylor There is no reason to renounce traditional modes o f speech, yet we have to free words o f the metaphysical meanings ascribed to them. Leszek Kolakowski It was all so unimaginably different And all so long ago. Louis MacNeice History is too serious to be left to historians. Iain Macleod

The Linguistics of History

Roy Harris

Edinburgh University Press

© Roy Harris, 2004

Transferred to digital print 2013 Edinburgh University Press Ltd 22 George Square, Edinburgh Typeset in 11 on 13 pt Ehrhardt by Hewer Text Ltd, Edinburgh, and

printed and bound by CPI Group (UK) Ltd Croydon, CRO 4YY A CIP record for this book is available from the British Library ISBN 0 7486 1930 5 (hardback) The right of Roy Harris to be identified as author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

Contents

P reface 1 2 3 4 5 6 7

vii

Language and the historian History and the literate revolution History as palimpsest Historicism and linguistics History and ordinary language The autonomy of history The truth about history

1 34 68 101 137 168 198

R eferences Index

229 239

Preface

The reader should be forewarned not to expect in the following chapters any detailed study of the language of historical accounts, or of the rhetoric of the great historians, or of the controversies which surround the definitions of some of the key terms historians have used. The stylistic devices, compositional techniques and vocabulary used by Herodotus, Livy, Gibbon, Macaulay et al. are not my primary concern. Had I written that kind of book, its title would have been The Language o f Historians. Nor do I envisage as linguistics of history the kind of interdisciplinary enterprise that treats language as ‘both input and output to historical scholarship’ (Downes 1994), although my conclusions do have a bearing on various questions that arise for any such project. What I think of as the linguistics of history concerns the assumptions about language that historians have made in constructing their accounts of the past. It is clear, for example, that many historians of the present century make very different linguistic assumptions from their predeces­ sors in previous centuries. It is also clear that such differences are reflected in taking very different views of what history is, and how the task of the historian should be construed. My main thesis will be that, throughout the Western tradition, the basic options in philosophy of history have been determined by the basic options in philosophy of language. Thus I shall attempt to present an argument of the following general form: a type-A philosophy of language sponsors a type-A philosophy of history, while a type-B philosophy of language sponsors a type-B philosophy of history. My own linguistic and philosophical assumptions underlying the inquiry are derived from adopting an integrationist theory of language and communication, the relevant details of which I shall set out in due course. My reason for proceeding in this way is that I believe that integrationism offers an original and powerful account of how language and history are related, bringing to the fore considerations which have vii

Vlll

THE L I N G U I S T I C S OF HI S TOR Y

been virtually ignored in all philosophy of history hitherto. I further believe that an integrationist approach makes it possible to pinpoint the factors that were decisive in establishing this pattern of dependence between views of history and views of language in the development of Western civilization. In other words, I am advancing a historical thesis as well as a philosophical thesis. Some historians have made (some of) their linguistic assumptions quite explicit. Marc Bloch, for instance, declares that, for the historian’s purposes, the first tool needed by any analysis is an appropriate language; a language capable of describing the precise outlines of the facts, while preserving the necessary flexibility to adapt itself to further discoveries and, above all, a language which is neither vacillating nor ambiguous. (Bloch 1954: 130)

Bloch here specifies certain linguistic requirements for the successful writing of history, and implies that unless these requirements are met history will be distorted. Whether they can be met in principle, and if so how, are difficult questions, to which Bloch does not supply any very clear answers; but it is nevertheless evident that he believes he has identified an important linguistic challenge that the historian must face. To take another example, according to E. H. Carr, ‘The very use of language commits the historian, like the scientist, to generalization’ (Carr E. H. 1987: 63). What Carr had in mind was the presence of general terms such as war and revolution in the traditional historian’s vocabulary. From this Carr drew certain conclusions about the relationship between history and science: in particular, he thought it followed that Aristotle’s definition of history as being concerned only with particular persons and events was untenable. (This is plainly a linguistic argument: unfortunately, it is not only a bad argument but also bad linguistics.) Similarly, when F. R. Ankersmit posits that ‘Narrative language is autonomous with regard to the past itself (Ankersmit 1994: 36) he is making a contentious linguistic claim in order to bolster an even more contentious claim about historical writing. But in most cases the linguistic assumptions that historians make about their enterprise are not stated as overtly as in the above cases; and it is with the concealed bulk of covert presuppositions that I shall be mainly concerned. Why do I think the linguistics of history of any more interest than, say, the linguistics of ornithology or the linguistics of cookery? Not simply because historians produce more varied and wide-ranging texts, but because history stands in the first rank of those conceptual supercate­

PREFACE

IX

gories that provide our modern bearings in all thinking about human affairs. It does so both at the lay and the academic levels of inquiry. Like other supercategories in the same league, such as ‘science’ and ‘art’, it is more often invoked than precisely defined. Like its peers, it finds favour with those who need magniloquent terms in which to couch weighty general pronouncements about the achievements of outstanding indiv­ iduals, of societies and of the human race as a whole. Like them, it is easily presented in personified or quasi-personified form, providing a basis for statements that are all the more impressive for teetering on the edge of nonsense. Like them, it is given to appearing in print with an initial capital letter, mysteriously awarded as if to confirm its supercategory status. Thus, just as ‘science’ often becomes ‘Science’, and ‘art’ becomes ‘Art’, so we are often invited to contemplate not ‘history’ but ‘History’. I think it is important for us to understand the linguistic basis of these master-concepts and make sure they are not what Bacon memorably called ‘idols of the market’. In the case of history, I know of no attempt to probe this question, which is what I have tried to do in the present book. That language and history are somehow interconnected no one seriously doubts. Historians treat linguistic evidence (in the form of documents and eye-witness testimony) as important for their under­ standing of what happened in the past; while many linguists treat languages as historical continua, open to change over time. But linguistics does not seem to be an area in which historians are very happy. Some historians have apparently been so bemused by the semantics of adverbs like now and then as to think there is no serious possibility of writing about the present at all. One of these announces solemnly that ‘the historian can hardly talk about the present because, by the time he has evidence to examine, it has become the past’ (Cannon 1980: 12). But at the opposite end of the scale come those who have convinced themselves by similarly spurious reasoning that language somehow deludes us into believing that the past is past. ‘For historians, in fact, past events are ever “present,” the tense in which we speak, write, listen, read, and of course, remember; and the deployment of past tenses is a linguistic strategy for giving the illusion of time passing’ (Kelley 1998: 11). Those beleaguered historians who misread this book as yet another sceptical attack upon their profession will doubtless be the first to point out that the sceptical author himself makes many historical claims. So he does. Such blatant self-incrimination might suggest that an ‘antihistorical’ reading of my argument must have missed something important. And it has. The catch 22 in writing a book of this kind is that even discussion of the issues requires reference back to what others

X

THE L I N G U I S T I C S OF HI S TORY

have written, and thus appears to engage the writer in the very same historical conspiracy that is being investigated. To claim that this invalidates the investigation ab initio is to employ a form of counter­ argument which charges the writer with ‘performative self-contradiction’; that is, committing the very same errors of which your opponent accuses you. A simple example, in the case of history, would be Bernard Williams’s contention that ‘the attack on [. . .] historical truth itself depends on some claims or other which themselves have to be taken to be true’ (Williams 2002: 2). Thus anyone who attacks the notion of historical truth emerges as pot calling kettle black. I anticipate that critics will be tempted to use some kind of ‘perfor­ mative self-contradiction’ argument against the main thesis of this book. So it may be as well to say in advance why I do not think such criticisms carry much weight. In the first place, ‘performative self-contradiction’ arguments in general, even when they are convincing, do not establish the validity of the position they are deployed to defend, but merely point out the weakness of attacking that position in a certain way. Thus, for example, even if Williams is right, it would not follow that we can now carry on as before, reassured that the notion of historical truth that came under attack is sound after all. In the second place, if critics are debarred under pain o f‘performative self-contradiction’ from conducting the debate in the historians’ terms, then the world is safe for historians, for ever and a day. But there is something deeply suspect about drawing up the rules of debate in this way. It is rather like insisting that no illegal organization can reliably be exposed by an infiltrator, because infiltrators are automatically party to and guilty o f the very same illegality as the organization which, under false pretences, they have managed to infil­ trate. If that rationale were sound where academic studies are concerned, all criticism of historians would be stifled at birth and the mere existence of the discipline would be its own justification. This book might be regarded as a sequel to one I wrote nearly a quarter of a century ago called The Language-M akers (Harris 1980). Language­ making and history-making I have always regarded as correlative and inseparable processes; for language invariably operates within some framework of beliefs about the way our bit of the world is and how it came to be that way. Anyone who doubts this should try the thoughtexperiment of attempting to make sense of this morning’s newspaper while suppressing all such beliefs. These frameworks are in part supplied by historians, who in turn rely on language to construct them. That is why there has to be an umbilical connexion between the accounts of the past that historians construct and their assumptions about the language they

PREFACE

XI

use in constructing them. This connexion is the basis of some of the longest-running controversies concerning the historian’s task. I have been selective in my discussion of topics and focussed mainly on issues connected with meaning and truth, rather than with issues con­ cerning ‘objectivity’ or ‘evidence’ or whether history is ‘scientific’ (although these too have an obvious linguistic dimension). This is because I take the core of any historical account to include or presuppose statements of the type: ‘event E occurred in place P at time T . (For example: ‘King Harold was killed at the battle of Hastings in 1066’, or ‘Pressure flaking was introduced in Europe during the Middle Solutrean’.) Historical accounts include many other types of statement as well; but unless a linguistics of history can at least explicate statements of this simple nuclear type, then it has no foundation - in my view - for dealing with anything more complicated. I am grateful to Robert Burns, Paul Hopper, Rosy Singh and Romila Thapar for directing my attention to publications which would otherwise have escaped my attention and to Peter Hacker, Steve Farrow and Paul Maylam for pertinent comments on various points. Parts of my argument were tried out on sceptical audiences at the University of London’s Institute of Historical Research and Dulwich College: from these I had valuable criticism. I have also profited from the discussion of some of these issues at the conference on language and history that was held by the International Association for the Integrational Study of Language and Communication at New Orleans in March 2002. R. H. Oxford, September 2003

CHAPTER ONE

Language and the historian

Oscar Wilde’s boutade ‘The one duty we owe to history is to rewrite it’ plays mischievously with two familiar but apparently conflicting ideas. One is the idea that history is something over and done with, and therefore cannot be altered. The other is the idea that history is something written down, and hence, far from being unalterable, can be altered with every rewriting, just as any tale can be expanded, embroidered or abbreviated (‘to cut a long story short’) each time it is told. These two ideas - the mutability and immutability of history, its rigidity and plasticity - need to be reconciled. But how? My case will be that this question cannot be tackled within the confines of philosophy of history, unless philosophy of history includes some philo­ sophy of language, or, more generally, philosophy of communication. The nucleus of my case is the claim that history-making is a complex process and that integrating the various components involved requires linguistic and other forms of communication between the parties concerned. My first move in setting out that case will be to put forward as uncontentious the following proposition: that the task of the historian, as traditionally conceived in Western civilization, would be impossible without certain facilities typically afforded by Greek, Latin and other European languages. We do not need to go into any detailed linguistic analysis in order to establish this general point. If in doubt, we can readily convince ourselves of its validity by considering the following sciencefiction scenario. Imagine that the first manned space ship from Earth to alight on Mars discovered a population whose language had no past tense, no words for ‘yesterday’ or ‘last year’, and no expressions for calendrical dates or time measurement of any kind. Into such a language it would be impossible to translate sentences of the kind that occur constantly in European history books, sentences stating that such-and-such an event occurred at such1

2

THE L I N G U I S T I C S OF HI S TORY

and-such a time in the past. In these circumstances it seems clear that, whatever other intellectual enterprises might flourish on Mars, the teaching of history (as we understand it) would not feature among them. There would be no jobs for historians in Martian schools or universities. And if this first expedition to Mars included among its members a missionary historian, whose ambition was to introduce Martians to the academic discipline of history, that terrestrial historian’s first task would have to be a reform of the Martian language. Nothing in this science-fiction story should be taken as implying that the Martians had no conception of history because the limits of their language prevented them from having one: that would be quite a different proposition. All the example is intended to demonstrate is that it is not impossible to envisage a people living ‘without history’. It has sometimes been claimed that on Earth there have been civilizations ‘without history’, although never, I think, that there have been civilizations lacking the necessary linguistic apparatus to construct historical accounts. As soon as the elders of the tribe can tell the rest what they think happened ‘a long time ago’ or even in their youth, they are already verbally equipped to be potential historians. The further question that this reflection prompts is whether people like my science-fiction Martians, although lacking the language to develop historical accounts, could nevertheless present such accounts, or their equivalents, in some other way. Could some alternative form of com­ munication serve the purpose? Could they, perhaps, articulate history solely in pictures, for example? I think the answer is fairly clearly ‘no’. For if Martian artists understood certain pictures as being pictures of what happened in the past, as distinct from depictions of present or future or imaginary happenings, it is hard to see how that understanding could be made explicit in discussion of their art. Questions like ‘Did what you have painted really happen?’ would be automatically precluded. So the hypothesis of a form of historical presentation entirely divorced from language seems a non-starter. Or rather, if Martians had developed some such form of their own, it seems doubtful whether it would be recognized by terrestrial historians as a form o f history. Granted that terrestrial historians need languages with adequate verbal equipment for discussing the past, my next move is to propose that they also need a semantics to go with it. In other words, a statement about the past (like a statement about the present or the future) has to be under­ stood as meaning something. And it will be judged accordingly, i.e. according to how it is construed as meaningful.

L A N G U A G E A ND THE HI S T OR I A N

3

A statement like ‘Caesar crossed the Rubicon’ or ‘King Harold was killed at the battle of Hastings’, if it is to serve the historian’s purpose, must be taken to have a meaning. The question is not only what meaning the words are taken to have, but how they come to have that meaning. And the two are intimately connected in ways that affect the historian’s whole enterprise. There are various possibilities to examine. Let us consider initially just three basic ways of looking at words and meanings. For convenience of reference I will call them (1) ‘reocentric semantics’, (2) ‘psychocentric semantics’ and (3) ‘contractual semantics’. I will first of all distinguish these three in very broad terms, ignoring the caveats that arise when considering the particular views of individual semantic theorists, who often combine two or all three in complicated ways. 1. R eocentric semantics. What I am calling ‘reocentric’ is an approach which regards the meanings of words as deriving ultimately from things in the external world. To a reocentric way of thinking it seems quite obvious that what fixes the meaning of the word horse is that there really are horses. Here ‘really are’ is to be understood as claiming that these fleshand-blood animals exist as part of the natural world, and that they exist independently of being recognized as such by human beings, and irrespective of what they might or might not be called in English or any other language. The basic assumption is that the word horse has as its primary function that of enabling us to discuss, talk about or refer to animals of just this kind. Thus the best way of explaining to someone what horse means is to produce a horse or horses for inspection. The word is regarded as ‘standing for’ the animal in question. Similarly, words such as eat, grow , soft and near stand for the corresponding action, process, property or relation; and such actions, processes, properties and relations are likewise conceived of as ‘really existing’ in the world (as distinct from being figments of our imagination or subjective constructs conjured up by our acquaintance with the words eat, grow , soft, near, etc.). Meanings, according to the reocentric view of the matter, are the links between particular patterns of sounds or letters and specific parts of reality. These correlations between words and reality are treated as so fundamental and indubitable as to render any argument superfluous. Language thus mirrors a world that actually exists. The sentence Grass is green mirrors the fact that grass really is green (whereas Grass is red mirrors nothing at all, except perhaps a state of affairs that could be imagined, but is manifestly at variance with the facts). These ‘real’ correlations are what make it possible, for instance, to enlighten someone who does not know what green means by saying that green is the colour of

4

THE L I N G U I S T I C S OF HI S TORY

grass (among other things), or by pointing to an appropriate colour sample. Reocentricity is the foundation of dictionary definitions such as ‘any ground squirrel of the genus Tamias or Eutamias, having alternate light and dark stripes running down the body’ for the word chipmunk (Concise Oxford D ictionary, 8th ed.), where what the lexicographer offers as the meaning of the word turns out to be simply a description of the animal. Reocentric theorists admit that not all words lend themselves to reocentric definitions of this kind, but those words which do are regarded as being in some way basic to our vocabulary, and the meanings of other words as being built upon that foundation. Thus while there is nothing for which the word although stands, we could not grasp what although means and how to use it unless we had already understood the meaning of words like grass and sentences like Grass is green. This view of meaning is sometimes also called ‘nomenclaturist’. Different languages (English, French, etc.) are seen as providing different nomenclatures, or sets of names, for the same independently given reality. The same animal that English speakers call horse is called cheval by speakers of French, P ferd by speakers of German, and so on. 2. P sychocentric semantics. What I am calling ‘psychocentric’ is a view very similar to the above, except that words are regarded as ‘standing for’ ideas in the mind rather than things in the external world. Thus, although doubtless horses exist, the meaning of the word horse is determined by what people think horses are, rather than by what they actually are. So on this view the meaning of the word horse could vary from person to person, depending on each individual’s notion of what a horse is. Similarly the meaning of the sentence Fish swim could vary according to what creatures different individuals classify as fish and how they regard what counts as swimming. Psychocentric semantics is sometimes regarded as superior to reocentric semantics because it makes it easier to explain why a word like unicorn is no less meaningful than horse, even though horses ‘really exist’ whereas unicorns do not. This is held to show that what gives a word its meaning is not the existence of a ‘real’ correlate, but simply that a certain idea is attached to it. On this view, our language is not a direct reflection of the world as it actually is, but rather of the world as we perceive or suppose it to be (whether rightly or wrongly). Confusingly, this psychocentric view of meaning is also sometimes called ‘nomenclaturist’. I shall try to avoid this confusion by using when necessary the term surrogational to cover both reocentric and psychocentric semantics, on the ground that what both have in common is their treatment of words as surrogates for something

L A N G U A G E A ND THE HI S T OR I A N

5

else: in one case for things and in the other case for ideas. A widely held surrogational view attempts to combine both reocentric and psychocentric positions, holding that words may simultaneously ‘stand for’ things in the real world as well as for our subjective ideas about them. But whether these two claims are ultimately reconcilable is questionable. 3. C ontractual semantics. What I am calling ‘contractual’ is a view quite distinct from either of the above (although it is often overlaid on both) because the meanings of words are regarded as depending ultimately not on how things stand in the external world, nor on any individual’s ideas about the external world, but on a collective agreement or convention which establishes a communal view of how words shall be used and how they are related one to another. Thus for the contractualist the meaning of the word horse is independent both of facts about horses in the real world, and of what may happen to be known or believed about horses by particular individuals, but is determined solely by the tacitly agreed way in which, for purposes of public communication in English, it is assumed that the word horse relates to - and is to be distinguished from - words such as cow , sheeps mare, fo a l, quadruped, etc. Similarly, for the contract­ ualist, the meaning of Grass is green depends neither on what colour grass actually is, nor on what colour it is perceived as being, but simply on what colour it is agreed shall be called green. This agreement is regarded as being somehow ‘built into’ the English language, or at least the English language as properly used by its native speakers in their intercourse with one another. Hence the meanings of words come from nowhere ‘outside’ but are ‘internal’ to the particular language in question. Theoretically, therefore, it would be unobjectionable to call green blue or horses sheep, provided the linguistic community sanctioned that linguistic contract. On this view, the question of whether grass ‘really is’ green just does not arise - or, alternatively, is meaningless - except within the framework already provided by a linguistic contract. The contractualist will point out that what makes it possible to talk about unicorns in English is the fact that the vocabulary of English includes the word unicorn, and the availablity of this word is unaffected by whether there ‘really are’ unicorns or whether people believe that there are. Thus in contractual semantics the stress is not on identifying what things or ideas words ‘stand for’, but rather on the collective consensus assumed to underlie communicational exchange within a given linguistic community. Whether this corresponds to the way the world ‘really is’, or to people’s beliefs about reality, makes no difference to meaning: in any case, these are separate and further questions that cannot even be addressed until one can first establish

6

THE L I N G U I S T I C S OF HI S TOR Y

under what linguistic contract the words in question have any meaning at all. For the contractualist, therefore, both the reocentric and the psycho­ centric theorists have got their priorities wrong. The surrogational question ‘What does this word stand for?’ makes no sense unless the word can first be identified as one specific lexical item in a complex verbal network constitutive of a particular community’s language. In contract­ ualist thinking, languages are often compared to currency systems. To ask ‘What does word x mean?’ is seen as rather like asking ‘What is a dollar worth?’. There can be no answer unless it is first possible to establish whether this is a question about the American dollar, the Canadian dollar, the Mexican dollar, etc. And even then, since currency values bear a fluctuating relationship to the ‘real world’ of goods and services, the question makes sense only when related externally to a specific time and market. Nevertheless, within the system the dollar will always be worth the same number of cents. That is to say, its internal value in relation to other units in the same system is what is fixed. Analogously, for the contractualist the meanings of words belonging to the same linguistic system do not vary, even though there may be inconsistencies between the ways those words are applied externally to features of the ‘real world’. I should perhaps repeat that in the work of particular theorists reocentric, psychocentric and contractualist strands of thinking are often tangled up, and that is precisely why I have separated them here in a deliberately bald fashion. The aim of the separation is to throw into relief an important general point about the connexion between semantic theory and history. The historian who adopts what I am calling a reocentric semantics will have a different view of historical accounts from one who adopts a psychocentric or a contractual semantics. In other words, reocentric, psychocentric and contractual semantics project interestingly different views of the historian’s task. In reocentric semantics, statements about what happened are answerable directly to what did in fact happen, it being assumed that the historian’s language makes available the resources for reporting it. The ultimate reocentric justification for using such expressions as the Roman empire and the battle o f Hastings has to be that the Roman empire actually existed and the battle of Hastings really took place. If not, then these expressions can only mislead and their use undermines any claim that the historian is giving a reliable account of the past. In psychocentric semantics, statements about what happened are answerable to what is perceived or believed to have happened, it being

L A N G U A G E A ND THE HI S T OR I A N

7

assumed that words have meanings anchored to the ideas in the mind of the historian. Thus the ultimate justification for using such expressions as the Roman empire and the battle o f Hastings has a different basis from that assumed in reocentric semantics. The concepts expressed may still be judged to have a certain validity in the context of the historian’s account, even if it is conceded that they are abstractions bearing no simple or verifiable relationship to actual events in the past. It may be, for instance, that all historians are mistaken about exactly when or where the battle of Hastings took place. But that does not automatically invalidate in toto the historical claim that King Harold was killed at the battle of Hastings. Finally, in contractual semantics, the ultimate justification for using such expresssions as the Roman empire and the battle o f Hastings is simply that in a particular language (in this case English) these are well-formed expressions sanctioned by collective usage, occupying a clearly identifi­ able place in the structural network of signifying units. No further validation is required. The question of what these expressions mean is quite separate from whether the Roman empire actually existed or whether the battle of Hastings ever took place. Comparing these three perspectives, we see that (1) under the auspices of reocentric semantics, the language problem for the historian is in effect simply merged with that of truth, whereas (2) psychocentric semantics interposes a problem about what goes on in the mind of the individual historian as language-user, while (3) contractual semantics relocates the problem as one that pertains not to the historian but, in the first instance, to the language the historian has chosen to use, and thus divorces meaning from truth altogether. Later in this chapter, and in subsequent chapters, I shall try to show in more detail how one or other of these three semantic approaches can be seen as underpinning historians’ views of their own discipline. But as an introductory example, which will also serve to illustrate how strands of reocentric, psychocentric and contractualist semantics may intertwine, and how they highlight serious - perhaps insoluble —problems about our use of language, I invite the reader to consider the following example of nineteenth-century wisdom: from the wisest Thought of a man to the actual truth of a Thing as it lies in Nature, there is, one would suppose, a sufficient interval! Consider it, - and what other intervals we introduce! The faithfullest, most glowing word of a man is but an imperfect image of the thought, such as it is, that dwells within him; his best word will never but with error convey his thought to other minds: and then between his poor thought and Nature’s fact, which is the Thought of the Eternal, there may be supposed to lie some discrepancies, some shortcomings.

8

THE L I N G U I S T I C S OF HI S TORY

Viewed from this linguistic perspective, the task of the historian might seem well-nigh impossible, for reocentric, psychocentric and contractual reasons combined. This is a vision of Homo loquens trapped in a maze of verbal deceptions, but the sources of these deceptions are clearly analysed into three semantic types. Words are here conceived of as imperfect intermediaries (1) between our thoughts and their verbal expression (psychocentrism), as well as (2) between our own thoughts and those of others (contractualism), and (3) between the form of our thought and ‘Nature’s fact’ (reocentrism). ‘Nature’s fact’ is, to complicate matters further, envisaged as an idea in the mind of God. In short, there are three levels at which error may (and to some extent must) beset the use of words. For (1) words may not match the thoughts they are intended to express, (2) words may not match thoughts in the minds of others, and (3) words will not in any case fully match the facts. The first of these three discrepancies raises a psychocentric problem (’How do I know that my words actually mean what I think I am saying?’). The second raises a contractual problem (’How can I be sure that my words mean what someone else takes them to mean?’). The third raises a reocentric problem (’How do I know that my words state what is really the case?’). All these, patently, are problems for the historian; especially as the writer in this particular example sees the ultimate communicational objective in reocentric terms; that is, as being to come as close as possible to ‘Nature’s fact’, which is itself a reality having divine sponsorship. Nature, the ultimate source of all non-illusory meaning, is equated with the Almighty’s creation; whereas men’s words are potential sources of semantic deception. The writer of this passage, it would seem, believes in God, and is doubtless acquainted with Plato’s mistrust of the world of appearances, with Aristotle’s distinction between history and poetry, and with Locke’s scepticism about the public linguistic contract. Failing memory shames me into confessing that I cannot now recall where I first read and noted down this quotation; but I suspect on stylistic grounds that it must have been somewhere in the voluminous pages of Carlyle, for whom, it has been said, history was ‘the true revelation of the divine, and the inspired historian was the prophet’ (Le Quesne 1982: 34). But the authorship matters little for the point I wish to make. Given the view of language there expressed, the historian would need to be able to penetrate a triple layer of linguistic confusion and difficulty. In the everyday world, words are anything but an adequate expression of thought, the linguistic contract is anything but reliable, and language is anything but a faithful mirror of reality.

L A N G U A G E A ND THE HI S T OR I A N

9

It must be admitted at the outset of this inquiry that few historians show any inclination to think that they have anything to learn from linguistics at all. In her chapter on ‘History and Other Disciplines’, which concerns history and ‘the fields with which it interacts’, Ludmilla Jordanova fails to find any room at all for the discipline of linguistics (Jordanova 2000: 5990). Similarly, in the first edition of The Nature o f H istory (1970), Arthur Marwick discusses the relationship of history to various neighbouring disciplines - ‘history and social science’, ‘history and geography’, ‘history and psychology’, etc. - but there is no section on ‘history and linguistics’. Thirty years on, however, in the second edition of his book, Marwick has got as far as admitting the existence of what he calls ‘philology’ and recognizes as ‘an important ancillary technique in the development of the discipline of history’ (Marwick 2001: 292-3). But how or why it is important to the historian is never explained, and from Marwick’s idiosyncratic definition of this discipline (‘the study of the structure and devices of language’) it seems clear that he has difficulty in dis­ tinguishing between philology and linguistics. These symptoms are all the more striking because in recent years there has been much talk by historians of a ‘linguistic turn’ in their own discipline. According to Brian Fay, the ‘full-bodied linguistic turn’ in history-writing dates from 1973 and the publication of Hayden White’s M etahistory (Fay 1998: 11). Similarly, Franklin Ankersmit credits White with introducing ‘a linguistic turn for historical theory’ (Ankersmit 1994: 8) and had himself in 1981 published a book N arrative Logic, with the ambitious subtitle A Sem antic Analysis o f the Historian s Language. Neither White’s book nor Ankersmit’s, however, has very much to do with linguistics or linguistic analysis as nowadays understood and taught by linguists. Both are much more concerned with what was traditionally called rhetoric. This term had earlier appeared in the title of an influential paper by J. H. Hexter, ‘The rhetoric of history’ (Hexter 1967), which predicted ‘an imminent paradigm shift’ for the discipline. Here we have an interesting case of historians anticipating history. The ‘linguistic turn’ was about to become a self-fulfilling prophecy. What did the prophesied paradigm shift deliver? According to Richard Evans: ‘the “linguistic turn” has given us discourses. And history is widely argued to be only one discourse among many’ (Evans 2000: 3). But to an outsider that hardly seems a revolutionary conclusion, unless previous historians had imagined that historical discourse was the only academic discourse in town. A more positive assessment came from John Toews, who wrote in the American H istorical R eview :

10

THE L I N G U I S T I C S OF HI S TOR Y There can be no question that this turn has enormously enriched our historical understanding of the complex ways in which meaning is constituted, transmitted, and transformed in the heterogeneous, compound, interrelated worlds of meaning we call culture. (Toews 1987: 906)

But when he went on to add ‘The history of meaning has successfully asserted the reality and autonomy of its object’ he must have caused some eyebrow-raising among his readers: for this sounds like the last message some of the enthusiastic turners would have wished to send out. On the contrary, they would have questioned whether there are any ‘real’, ‘autonomous’ objects at all for history to concern itself with. There is a tendency for historians to put ‘linguistic turn’ in scare quotes, as if doubting its appositeness or validity. But even some who do this seem strangely indifferent to the history of this expression or its origin. Georg Iggers, for example, can write a whole chapter on the ‘linguistic turn’ without even raising such questions (Iggers 1997: 118— 83). Others, while raising them, are by no means unanimous about where the expression linguistic turn came from, or what it means, or its connexion with ‘postmodernism’. Some are aware that it was borrowed from one of history’s subdisciplines, namely the history of philosophy, and appears in the title of a volume edited by Richard Rorty (Rorty 1967). As Marwick points out, the historical irony in this is that the philos­ ophers represented in Rorty’s volume are far from being sympathetic to those postmodernist views with which, for many historians, the ex­ pression linguistic turn is primarily associated. Marwick writes: On the central importance of language, I am in agreement with the postmodernists (and perhaps in disagreement with some of my fellow historians). However, against specious postmodernist claims to have instituted a ‘linguistic turn’, tearing the bandages from the eyes of those who naively saw language as a simple, uncom­ plicated medium of communication, I point out that historical method has its roots in philology and that questions of semantics and signification have preoccupied historians for generations [. . .]. (Marwick 2002: 11)

This is a bold attempt at a linguistic double-turn, hoisting the postmodernists with their own petard. But when a linguist looks for this alleged ‘preoccupation’ with semantics on the part of generations of historians it is difficult to find evidence for it at all. Nor does Marwick supply any. In history of philosophy, the irony is that, having adopted the expression linguistic turn, philosophers promptly fell out over when the turn in question occurred. According to some, it began in 1884 with

L A N G U A G E A ND THE HI S T OR I A N

11

Frege’s Foundations o f Arithmetic. Peter Hacker, however, sees the landmark statement as Wittgenstein’s pronouncement that ‘All philoso­ phy is a critique of language’ (Hacker 1996: 36-8; Hacker 1997: 66-71). So far we have waited in vain for anyone to claim ‘All history is a critique of language’; but the day cannot be far off. The expression linguistic turn was not coined by Rorty. He got it from one of the members of the Vienna Circle, Gustav Bergmann. But not all philosophers - and even fewer historians - who happily took over this expression seem to have read Bergmann’s book Logic and R eality (1964) in which it first appears. Had they done so, they might have realized that it there stands in opposition to epistem ological turn. Thus we come to yet another irony in the story: for the position maintained by those historians who have actually taken (what they describe as) the ‘linguistic turn’ is nothing if not an epistemological position. Quite a different account, however, is also available. In H istory in P ractice Jordanova explains the expression linguistic turn by saying that ‘the word linguistic was used because of an indebtedness to Ferdinand de Saussure’ (Jordanova 2000: 85). (Saussure was, almost certainly, a source Bergmann had never read, and possibly never heard of. And if he had, Saussure’s views on language would have been anathema to him.) Jordanova follows her Saussurean explanation by a garbled account of the Saussurean distinction between signifiant and signifie, concluding: This gives language a more independent existence than is generally assumed. [. . .] Fundamental categories are in language rather than in the physical world. [. . .] it follows, just as inexorably, that what can be known about the past is limited. (Jordanova 2000: 85)

But how this follows she never elucidates. Saussure also features prominently in Beverley Southgate’s account of how the ‘linguistic turn’ challenged the old model ‘whereby historical truth is simply dependent upon correct identification and verbal descrip­ tion’ (Southgate 2001: 78). Although he acknowledges (in a footnote) Bergmann’s role in introducing the expression linguistic turn, Southgate seems oblivious to the incompatibility between Bergmann’s view of language and Saussure’s. His discussion of the linguistic turn more or less equates modern linguistics with Saussurean structuralism. That would already be bad enough as history of ideas. But worse follows. Southgate identifies the Saussurean notion of a language as an ‘auton­ omous structure’ with conceiving of language as ‘itself determining the world as it is experienced by us’ (Southgate 2001: 76). Wherever this idea

12

THE L I N G U I S T I C S OF HI S TORY

came from, it could hardly have come from any serious examination of the work of Saussure. Oddly enough, the appeal to Saussure seems to come from Hayden White himself, who declares that ‘the Saussurian theory of language as a sign system’ offers ‘the best immediate prospects for a fruitful revision of the central problem of intellectual history, the problem of ideology’ (White 1987: 190). I say ‘oddly’ because White, like many of his followers, appears to have only the vaguest idea, derived from second- or third-hand sources, of what Saussure’s ‘theory of language’ actually was. One is left wondering exactly how historians have managed to construct for themselves such views of the historical ‘turn’ in their own discipline, and what was ‘linguistic’ about it. We find such curious claims as: ‘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 history’s language that makes this possible remains undisclosed: or why the ‘different perspective’ does not reduce to the banal observation that time has passed between the historian’s writing and the events reported. The writer just quoted 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 funda­ mental importance for philosophy of history. Although, as mentioned above, his book N arrative Logic has the subtitle A Sem antic Analysis o f 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. However we interpret it, the expression linguistic turn seems less radical, more comfortably middle-class, than linguistic revolution. In 1956 one Oxford philosopher had edited a volume entitled The R evolution in Philosophy (Pears 1956). Two years later, another Oxford philosopher explicitly rejected the claim that any such revolution had occurred (Warnock 1958: 160-1). But neither suggested substituting turn for revolution. Given Bergmann’s role in the story, I am attracted by Robert Burns’ suggestion that turn may originally have been a translation or caique of the German words K ehre or Wende. But it is difficult to escape the impression that many historians themselves were and are unsure about what the ‘turn’ was. At best, there was the merest

L A N G U A G E A ND THE HI S T OR I A N

13

nod in the direction of a linguistic problem. The problem itself was never tackled. Yet there is a problem, which for centuries Western historians managed to ignore; namely, that their contribution to human knowledge was itself language-dependent (not just dependent on the adoption of particular narrative forms or rhetorical devices). The first scholar to draw attention to this more fundamental problem 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). To adapt a much later dictum of Wittgenstein, what Humboldt was saying was that the limits of the historian’s language are the limits of the historian’s world. But among historians no contemporary of Humboldt’s seems to have taken that point seriously. It was a shrewd observation, but its time had not yet come. What worried Humboldt was that the words the historian uses are rarely free from connotations of which the historian may not be fully aware. These connotations, in Humboldt’s view, ‘will give rise to false­ hood and uncertainty’. Nothing, he claimed, is rarer than ‘a narrative which is literally true’ (Humboldt 1821: 58). Behind this claim doubtless lay an implicit comparison between the language of history and the language of the natural sciences. Only in the latter domain, it might seem, can we hope to find a terminology circumscribed by exact definitions and free from the pervasive connotations that, in Humboldt’s view, are inimical to the statement of literal truth. Just as our language, according to Humboldt, is already pervaded by connotations of which we are not necessarily conscious, so, according to some philosophers, it is likewise pervaded by unconscious historical assumptions. Arthur Danto (sometimes — although somewhat oddly, given his philosophical stance - recognized as a ‘precursor’ of the linguistic turn in history: Fay 1998: 11) writes: It is impossible to overestimate the extent to which our common ways of thinking about the world are historical. This is exhibited, if by nothing else, by the immense number of terms in our language, the correct application of which, even to contemporary objects, presupposes the historical mode of thought. Should there be a people somewhere in time who truly thought unhistorically, we would know this by the fact that communication with them would be marginal, vast regions of our language being untranslatable into theirs. (Danto 1985: xv)

14

THE L I N G U I S T I C S OF HI S TORY

My science-fiction Martians may perhaps be taken as an illustration of this latter point. Where Danto begs a crucial question, however, is in his tacit assumption that their ‘thinking unhistorically’ would be the explanation of the deficits in their language. This is an assumption typical of what I am calling ‘psychocentric semantics’. (The corresponding assumption in (some versions of) contractual semantics would have it the other way round: ‘thinking unhistorically’ is explained by the structure of the language in which the thinking is articulated.) Danto continues: To endeavour, ourselves, to think unhistorically would require at least a linguistic restraint, for we would be committed to get on with a fragment of our vocabulary and grammar. We should, indeed, have to restrict descriptions to just those predicates which pass muster under empiricist criteria of meaningfulness. (Danto 1985: xv)

Here Danto’s psychocentric commitments become clearer: what he calls ‘empiricist criteria of meaningfulness’, and is keen to reject, are not far removed from those adopted in a strictly reocentric semantics, where for every meaningful word one needs to find something present in the external world to stand as its meaning (grass for the word grass, the colour green for the word green , etc.). So endeavouring to ‘think unhistorically’ would require a psychological suppression of all those linguistic habits and forms of expression for which one could not find a reocentric justification. Could we even think seriously about what happened yester­ day or what might happen tomorrow, given that yesterday is now gone for ever, and we cannot know what tomorrow may bring? And then the question arises (as presumably Danto intends that it should) of whether this Draconian suppression would leave us with much worth thinking about at all. There is another interpretation of the notion of linguistic limits, which puts the interface between language and history in a different perspective. Here 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 historian’s language happens to be. Given any language you like, there are also limits imposed by how its users understand a language to function. In particular, there are limits on how they understand their language to be related to the non-linguistic world. There are, for instance, communities in which it is believed that, according to their meaning, words may have direct physiological effects on the addressee (Calame-Griaule 1987;

L A N G U A G E A ND THE HI S T OR I A N

15

Harris R. 1990). Some such vestigial belief may perhaps underlie Western speech-acts such as cursing an enemy, or exorcism, or praying for a cure, but it does not play a major role in Western semantics. On the other hand, what does play a major role is the absence of any such belief: what we do not believe about the way language works is no less important to our self­ understanding than what we do believe. Thus the notion that the limits of the historian’s language are the limits of the historian’s world can be interpreted as meaning that the possible forms of history are in practice determined within limits set by a community’s view of what their language makes possible. Suppose, for instance, we found that Martians, while speaking a language well equipped with past tenses and dates, held the belief that what is said about the past, if said with the appropriate authority, actually alters what happened. This would put quite a different gloss on the limits of Martian. Thus, just as some religious sects believe in the possibility of retro­ spectively baptizing their ancestors, so the official historians of Mars are entrusted with the task of correcting unfortunate omissions or mistakes in the Martian past. That we may find such a belief quite absurd only reinforces the above point about linguistic limits. It is not just a case of recommending the would-be historian first of all to find a language with the right tense system, deictics, etc.; but, more crucially, having found such a language, to understand the way or ways in which that language makes engagement with the past possible. And that is a much more daunting intellectual task. No terrestrial historian, to the best of my knowledge, has ever attempted it. In spite of their reluctance to engage with linguistic issues, historians often show no hesitation in accepting at least one controversial semantic thesis. This is the claim 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 mean either the record of events, or events themselves’ (Shotwell 1910: 527). Thus, allegedly, in the former sense a history is what Gibbon wrote (his H istory o f the Decline and Fall o f 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 are some more recent examples. The first is from W. H. Walsh’s An Introduction to Philosophy o f History. According to Walsh,

16

THE L I N G U I S T I C S OF HI S TORY

it is a ‘simple and familiar fact that the word “history” is itself ambig­ uous’. The two meanings in question are: ‘(1) the totality of past human actions, and (2) the narrative or account we construct of them now’ (Walsh 1967: 16). A somewhat different version of the ambiguity is given by John Cannon in his essay ‘The historian at work’: ‘We use the same word to mean not only the events of the past, what actually happened, but also our analyses of and commentaries upon what happened’ (Cannon 1980: 7-8). Another example is from John Tosh’s The Pursuit o f H istory: ‘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). A decade later Bernard Bailyn states: The word ‘history’ has, I think, two meanings. One is simply what happened; that is, the events, developments, circumstances, and thoughts of the past, as they actually occurred. The other is history as knowledge o f what happened, the record or expression of what occurred. (Bailyn 1994: 7)

According to David Carr: ‘History’ in English (and its equivalent in most modern European languages) has two primary senses: (1) the temporal progression of large-scale human events and actions, primarily but not exclusively in the past; and (2) the discipline or inquiry in which knowledge of the human past is acquired or sought. (Carr D. 1995: 584—5)

As the above examples show, those who adopt the ‘semantic ambiguity’ thesis differ as to exactly what the two meanings of ‘the word history’ are; but that there are two they are agreed on. (Some add a third, but that is a complication I shall ignore for present purposes.) Many linguists would see nothing objectionable in the ambiguity thesis as such. Nevertheless it raises a fundamental issue in philosophy of language, and one which will serve to provide a first illustration of how and why there is room for a reassessment of linguistic assumptions which generations of historians have taken for granted. The issue in question is the ontological status of linguistic units, and on this linguists are deeply divided. What is it that, allegedly, ‘has two meanings’? It is not any individual utterance or inscription, not what someone said or wrote on any particular occasion, but ‘the word history’ itself. Nor are those who make this claim proposing stipulative definitions: on the contrary, they are treating ‘the word history’ as a linguistic object about which observable ‘facts’ are

L A N G U A G E A ND THE HI S T O R I A N

17

available. One of the writers quoted above goes so far as to say that the ambiguity of this word is a ‘simple and familiar fact’, as if it stood in need of no demonstration; another that the ambiguity is manifest ‘in common parlance’, and hence by implication, presumably, a matter of which we are all well aware. For these philosophers and historians, it seems, ‘the word history’ is something that actually exists in linguistic reality, with an identifiable form and two distinct meanings. In short, the existence o f‘the word history’ and its semantic ambiguity are themselves historical fa cts , and indeed the first facts to which they appeal. From these facts, the rest of their view of history can be seen as a logical progression. So here straight away we uncover a theoretical assumption that leads directly to the core of our problem; the assumption being that, for some historians at least, the realm of historical facts (facts available in principle to the historian) includes certain kinds of linguistic fact. What is inter­ esting from a linguist’s point of view is whether the definitions of ‘the word history’ listed above are to be understood as having a reocentric, a psychocentric or a contractual interpretation. All are possible, and all seem to presuppose that a word is a bipartite entity, having both a form and a meaning. Thus once the form of the word has been identified (in this case simply by citing the modern English spelling: h -i-s-t-o-r-y) it is taken as a legitimate question to ask whether it has one meaning or more than one. But the alleged meaning or meanings are then identified by citing various other English phrases, such as record o f events and totality o f past human actions, as if these latter phrases themselves encapsulated the meaning(s) in question. But that immediately raises more serious seman­ tic problems. Does it make sense to identify a meaning of one word (‘the word history’) by substituting some other words or words, unless our semantic theory can guarantee that the meaning of the verbal substitute is no different from the meaning of the original word in question? Where is that guarantee to come from? How can the identity in question be verified if it should be challenged? Whether they realize it or not, historians who propose definitions like those given above have already strayed far into areas of linguistic controversy. The whole jejune enterprise of trying to explain what history is by reference to what ‘the word history’ means is fraught with dif­ ficulties, of which the historians themselves seem as unaware as cross­ country hikers who, doggedly following their own recent map, do not realize they have walked into the middle of an ancient minefield. Now that is very relevant to the supposed linguistic turn that many historians today invoke. For the linguistic turn that Gustav Bergmann originally identified (Bergmann 1964: 7-9) boils down to taking a

18

THE L I N G U I S T I C S OF HI S TOR Y

philosophical question and substituting a linguistic question for it. (Thus ‘What is justice?’ is replaced by ‘What does the word ju stice mean?’. Cf. ‘What is history?’ and ‘What does the word history mean?’.) What is interesting, as is shown by my list of examples, is that this tendency to construe questions about the nature of history as presupposing answers to questions about the meaning of ‘the word history’ long antedates post­ modernism and has no necessary connexion with it. But that is not all. Linguists who take a radically different view of linguistic units and semantics will question whether there ‘really is’ any such word as ‘the word history\ let alone whether it ‘really has’ the meanings attributed to it, and would regard those who unquestioningly believe these to be matters of historical fact as victims of a naive reification. It might be urged in support of ‘the word history’ that its existence is attested in dictionaries, where its meanings are also listed, and sometimes its etymology. But the information contained in dictionaries can hardly be treated as summaries of incontrovertible linguistic truths. That inform­ ation, such as it is, is itself the product of sophisticated analyses conducted within the framework of a particular grammatical tradition. In short, citing the dictionary entry as historical evidence for the existence of ‘the word history’ is a kind of category mistake. It is rather like taking out your tape-measure in order to prove to someone that the centimetre really exists. To identify ‘the word history’ in the way that dictionaries do is already to engage in a deliberate process of metalinguistic abstraction and attribute a historical reality to the product. That can hardly be a sound basis for philosophy of history. Or if it is, it requires a great deal more sophisticated argument than either historians or philosophers have so far offered us. But surely, someone may object, the fact is that, irrespective of what theorists may say, we can all see several examples of the word on this page alone: so it clearly does exist, in defiance of all theoretical doubts. As an ontological argument, this fares no better. How we count and classify instances (whether of linguistic units or anything else) depends on the criteria we adopt for the purpose: a successful counting procedure proves nothing about the ontological status of what is counted. The Greeks could count the number of stars in a constellation without having any clear idea of what stars actually were. That we can take any random selection of dogs and sort them into dogs taller than eighteen inches and dogs of eighteen inches or less, and say exactly how many of each there are, does not prove the existence of two separate canine species, maxidog and minidog. The interesting question is why, in the case of words, people should be

L A N G U A G E A ND THE HI S T O R I A N

19

confused into thinking otherwise. The case of ‘the word history’ is only one example of this metalinguistic confusion. Doubtless there will be readers who think that there is no excuse for making such heavy going of all this, since talk of ‘the word history’ is no more than an innocent fa fo n de parler justified by common usage. But if it is indeed just a fapon de parler, then a manner of speaking about what? And why do so many historians and philosophers base their view of history on a mere fapon de parler? What seems clear from the quotations cited above is that we are not dealing with an innocent fapon de parler. We are dealing with an alleged semantic basis for tackling the question ‘What is history?’. And this semantic basis is in turn deemed to rest on a historical fact: a fact about a particular word. It is a case of history coming to its own semantic rescue. So the question ‘What is history?’ cannot be answered in any satisfactory way by appeal to existing dictionaries or to the lexical definitions preferred by particular historians. It has to be situated in a much more comprehensive context of linguistic inquiry. Accordingly, I shall now try to set out briefly what seems to me to have been the view of language adopted by the majority of traditional historians from Herodotus onwards. This provides a broad intellectual umbrella, under which reocentric, psychocentric and contractual semantics can all find room to shelter, along with their various offshoots and derivatives. The basic story can be broken down into three sets of propositions, as follows. 1. Senders and receivers. Linguistic communication is a transaction between senders and receivers of a message. The sender (in this case, the historian) takes the initiative of producing a text. An example of such a text would be, to take one chosen at random, Powicke’s The Reformation in England (Powicke 1941). The receiver is any reader of the text (in this example, any reader of The Reformation in England). 2. M essages. A message originates in the mind of the sender (in this instance, the late Sir Maurice Powicke). A message consists of thoughts or ideas. The message(s) Powicke wished to send to his readers will be contained in the text of The Reformation in England. 3. Codes. In linguistic communication, messages are expressed in codes. These consist in sets of sounds or written marks having established meanings. The message sent depends on the sounds or marks chosen from one particular code. In The Reformation in England, the code in question was (the British variety of) written English of 1940 or thereabouts. Generalizing from the above, we can say that if historian A wishes to convey to a listener or reader B some proposition, say, for the sake of

20

THE L I N G U I S T I C S OF HI S TOR Y

argument, that Columbus discovered America, A s communicational task is to find a language, known also to B y in which there occurs a combination of words appropriate to express exactly that thought. For instance, the sentence Columbus discovered America. If B has a knowledge of that language, then A can use this sentence to convey to B the desired message. (And so on for as many messages as the language in question allows.) It should be noted that it makes no difference whether what A claims is true. The sentence Columbus did not discover America is also available for communicational purposes, should A wish to use it. The truth or falsity of the message is, as it were, a matter of indifference to the language itself. Since both true and false statements alike are catered for, communication per se is situated at one remove from any set of Tacts’ (including inter alia any Tacts’ about the discovery of America). Communication thus becomes an autonomous or segregated domain, analysable in its own terms. This traditional model of communication is based on two underlying assumptions. One is the assumption that human communication is a process by which, when successful, thoughts or ideas are conveyed from the mind of one individual to the mind of another. I shall call this the assumption of Telementation’. The other assumption is that there exist systems of signs (including, specifically, languages like English and French) which, because they are common to two or more individuals, make successful telementation possible. This I shall call the ‘fixed-code’ assumption. According to this model, the languages we speak and write are codes which function as semiological grids. Through these grids our thought must pass and be shaped if it is ever to emerge in public and be transmitted accurately to others. These assumptions seem to some linguists, of whom I am one, to be fallacious. Put together, they make up what its critics call The language myth’. This myth has been discussed in detail elsewhere (Harris R. 1981; Harris R. 2002) and the deliberately provocative choice of the term myth has indeed provoked protests. They come, on one side, from those who think that linguistic communication does in fa c t work in the way described above (so this is no ‘myth’); on the other side, from those who think no one ever believed in this account anyway (so the ‘myth’ itself is a myth). Not being convinced by either set of objectors, I shall continue to use the term throughout the present discussion. By calling the story ‘mythical’ I mean that the communication process it conjures up is on a par with enduring popular fictions like the ‘social contract’ and the ‘spirit of the age’. Telementation is no more observable than telepathy; nor fixed codes of speech than fixed codes of morality.

L A N G U A G E A ND THE HI S T O R I A N

21

As I see it, this myth about the communicational process dominated Western thinking on language from at least the fifth century bc onwards. Historians did not invent it: they merely adopted it, in common with most of their contemporaries. Grammarians and philosophers developed various versions of it, but retained intact its basic structure as described above. As Locke later put it in one of the classic statements of the myth, men suppose their words to be marks of the ideas in the minds also of other men, with whom they communicate: for else they should talk in vain, and could not be understood, if the sounds they applied to one idea were such as by the hearer were applied to another, which is to speak two languages. (Locke 1706: III.2.4)

In other words, according to Locke, everyday communication relies on the principle that what a speaker says can, and normally will, be under­ stood by any hearer who speaks the same language. ‘Locke’s principle’, as we may call it, was implicitly reaffirmed two centuries later by Saussure. Saussure diagrammed speech communication as a circuit linking the brain of the speaker to the brain of the hearer. By means of this circuit, speaker and hearer can exchange thoughts in virtue of the fact that a concept in the brain of the speaker ‘triggers’ a sound pattern, the utterance of which stimulates the ear of the hearer, trans­ mitting the corresponding sound pattern to the hearer’s brain, where it activates the corresponding concept. This celebrated diagram and Saussure’s commentary on it (Saussure 1922: 27-8) became one of the foundational texts of twentieth-century linguistics. Thus a long-standing myth was upgraded overnight into a ‘scientific hypothesis’ (at least, for those linguists and psychologists who believed in it). The explanatory power of the myth is that, if taken as a true account of human communication, it guarantees, at least in the ideal case, identity between the original thought and its received interpretation, across whatever diversity of circumstances may separate sender from receiver. That is to say, delivery of the message transmitted from the mind of A to the mind of B is ensured by a code which identifies this message as the common possession of both. Message M is, by definition, the message in question if, and only if, both A and B are operating with the same code, which includes M as one possible message. The two basic components of the myth - the telementational fallacy and the fixed-code fallacy - are logically independent of each other, and neither is plausible in isolation; but together they prop each other up and combine to give the impression of rational cohesion.

22

THE L I N G U I S T I C S OF HI S TOR Y

Some combination of ‘telementation + fixed-code’ assumptions has informed countless discussions of human communication from Plato down to the present day. It carries conviction, superficially at least, because it appears to be grounded in ‘common sense’ and everyday experience. People know that they will not be able to read a German newspaper if they know no German; that if they want to tell someone that the kettle is boiling it is no use saying ‘Tomorrow is Tuesday’. All this seems to provide confirmation that understanding a verbal message is a matter of reconstructing, from the words uttered or written, what the speaker or writer had in mind when uttering or writing them, and that this would not be possible unless both sender and receiver of the verbal message were in possession of the system or code which enables this transmission to take place. I am not claiming that all historians in the Western tradition swallowed the language myth hook, line and sinker. Carlyle, for one, evidently did not. In the passage I quoted earlier he questions, in effect, not only surrogational semantics but the efficacy of the telementation process itself. I happen to agree with him on both points. But although I call this view of how linguistic communication works a ‘myth’, and thereby imply that I think it is mistaken, it makes no difference for purposes of my argument in this book whether it is mistaken or not. Why I nevertheless draw attention to it is that taking some such set of assumptions as gospel (whether mistakenly or not) has the effect of suppressing the whole question of the role language plays in the making of history. In short, the language myth automatically has certain consequences for the historian who accepts it. To these consequences I now come. (Paradoxically, as will emerge, the postmodernists who make such a song-and-dance about their linguistic turn are themselves dupes of the very language myth they think they have left behind.) The extent to which historians in the Western tradition have relied on fixed-code communication can be judged by the number of terms they never feel called upon to explain. When affirming, for example, ‘King Harold was killed at the battle of Hastings’, or ‘Napoleon was exiled to Elba’, they have rarely if ever seen it as part of their role to state the meaning of such words as battle and exile. This, if they thought about it at all, was considered a matter for grammarians or lexicographers, not historians. Confronting the crucial question of how the historian knows the right way to interpret a text, Gadamer makes the reliance on fixed-code semantics quite explicit: ‘It is a general presupposition that can be

L A N G U A G E AND THE HI S T O R I A N

23

questioned only in particular cases that someone who speaks the same language as I do uses the words in the sense familiar to me’ (Gadamer 1985: 237). It is obvious that Gadamer’s presupposition is a restatement of Locke’s principle: the only difference is that whereas Locke’s statement of it is psychocentric, Gadamer’s version is contractualist. If historians believe in Gadamer’s presupposition, it is small wonder that language in the service of history has for so long been virtually ignored. The question I wish to raise now is whether it has been ignored deliberately. For accepting Gadamer’s presupposition is tantamount to assuming that there is no problem about what history is; that must already be known to anyone familiar with the meaning o f‘the word history’. Thus it becomes possible to see why historians are so keen to insist on presenting their version of what it means as common knowledge. It is an uncharitable but inevitable suspicion that the ambiguity they ascribe to the word is one that historians have invented to serve their own purposes. To be sure, the ambiguity is registered in dictionaries. But it is not lexicographers who are to blame here. The trouble is that lexic­ ographers are reflecting what historians and philosophers themselves have claimed. In short, the appeal to semantic ambiguity is a blatant case where philosophy of history borrows directly from its own coffers. Claiming that ‘the word history’ is ambiguous is a convenient way of justifying certain assumptions that the professionals 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 the dictionary, 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 arena of possible dispute, so it does not 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 locate it in 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 fa cto 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 Toynbee’s work by refusing to count him as a historian.) Third, the supposed ambiguity conveniently allows two quite diver­

24

THE L I N G U I S T I C S OF HI S TOR Y

gent accounts of the same event to exist side by side without either having to sacrifice its status as history. In other words, it dispenses historian Smith from ever having to reach agreement with historian Jones, but protects both simultaneously as members of the historians’ profession. 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 is 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 ex­ clusively 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, in A. L. Rowse’s book The Use o f H istory (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; nor should that cleavage be smuggled in under the smokescreen of the supposed ambiguity of ‘the word history\ Fifth, recent critics of traditional historians have latched on to the ambiguity thesis and exploited it for their own purposes. It provides them with a handy short-cut to setting up their own concept of history as ‘discourse’. For example, Keith Jenkins complains that ‘the word history’ covers both ‘that which has been written/recorded about the past’ and also ‘the past itself (Jenkins 1991: 6). He proceeds without further ado to propose the following improved terminology, which is much more to his liking. ‘The writings of historians’ will be called historiography; ‘all that has gone on before everywhere’ will be called the past; while the word H istory (but with a capital H) will be retained for ‘the whole ensemble of relations’. Thus the word we started out with has allegedly been disambiguated by this proposed new trio of terms. But has it? For a start, classifying what historians do as ‘historiography’ (a classification I shall avoid for the reason here given) seems ill-advised unless the intention is to restrict their function to producing written texts. That

L A N G U A G E A ND THE HI S T OR I A N

25

doubtless suits Jenkins’ postmodern analysis of history very well, but it is quite misleading. What the disambiguation exercise does, as Jenkins is well aware, is set up the basis for claiming that there is no necessary connexion between what historians write (history as ‘discourse’) and what they purport to be writing about. The discourse and the past, Jenkins alleges, ‘float free of each other’. (Here we see one brand of contractual semantics in full flow: the meaning of the linguistic sign is divorced from any application to the external world.) But whatever kind o f‘floating’ this may be, the objective is evidently to isolate the ‘discourse’ as a selfcontained object of inquiry, and this smart epistemological move has no other basis than the supposed disambiguation of a lexical ambiguity. Miraculously, what began as two separate ‘meanings’ of a word have suddenly grown into two separate domains of investigation, each re­ quiring different methods; and this transformation has been effected at the cost of minimum argument (in fact, no argument at all) by presenting it as somehow following from, or implied in, our very acceptance of ‘the word history’. Sixth, it allows for the establishment of a branch of history which is nothing other than ‘history of history’. This is a convenient way of ensuring that only what historians themselves count as history receives official accreditation in the teaching of the subject, and only those who apply the approved ‘historical method’ are allowed any standing in the investigation of the past. In short, it makes the discipline of history a closed shop. We cannot even begin to get to grips with the linguistics of history until we realize that to accept all this lexical confabulation about ‘the word history’ —of which postmodern apologists are no less guilty than their conservative opponents - is already to buy into the language myth itself. It is erroneous to suppose that the ambiguity thesis is neutral with respect to philosophy of history. That is far from being the case. Once we accept it, we are already involved in taking philosophical positions. This begins to become obvious when we consider that if indeed there is such an ambiguity, and if we must therefore distinguish the meaning ‘history(l)’ from the meaning ‘history(2)’ - however these two are defined in detail it seems to follow that a phrase such as Roman history must also be ambiguous. That is, it will be ambiguous as between ‘history(l) of Rome’ and ‘history(2) of Rome’ (leaving aside any additional ambiguity con­ tributed by the adjective Roman). This in turn implies a relationship between the two meanings. According to the ambiguity thesis as usually stated, in one case (‘history(l) of Rome’) we would be dealing with what

26

THE L I N G U I S T I C S OF HI S TOR Y

the Romans themselves did, and in the other (‘history(2) of Rome’) with what historians said the Romans did. Which of these meanings is primary? We are often invited to see the former as taking priority over the latter, at least by those historians whose semantics is reocentric. In other words, a work which is a work of Roman history in the sense ‘history(2) of Rome’ qualifies as such in virtue of having as its content Roman history in the sense ‘history(l) of Rome’. Otherwise, whatever else it may be, it is not a work of Roman history in either sense. If this is indeed the semantic relationship between ‘history(l)’ and ‘history(2)’, the very notion of historicity is tied ab initio to the original events: these become historical ‘raw material’, independently of whether or how historians make use of this material. ‘What happened’ is thus fundamental, and what historians reported is an appendage. Now alternatively, it is possible to construe ‘history(2)’ as taking priority over ‘history(l)’. This, in effect, is the currently popular reversal of priorities that is called ‘postmodern’. The reversal causes apoplexy among reocentric historians, who cannot make sense of a semantics in which the history of Rome is in the first place the story historians tell, while what the Romans themselves did enters the picture only by courtesy of that story. They cannot, in short, accept that ‘history(l)’ is defined by reference to ‘history(2)’, while being entirely happy to accept that ‘history(2)’ is defined by reference to ‘history(l)’. When we compare these two possible interpretations, the comparison shows that, far from being philosophically neutral, the actual role of the ambiguity thesis is to set up a rigid framework within which the relevant philosophical positions can be located. Otherwise, there would be no point in invoking the alleged ambiguity of ‘the word history’ at all. At this point someone is likely to protest: ‘But historians are not linguists! Cannot they be allowed at least to use the English language like anyone else?’ Certainly, no one can stop them. But that does not absolve them from accepting any responsibility about how they use it. The claim that history, unlike the sciences, has no special language of its own, so that the historian for the most part makes do with the ordinary language of the day, has been championed in various places by a number of writers (Gardiner 1952; Berlin 1954; Frankel 1957; Walsh 1967). It is a thesis that managed to survive the advent of postmodernism, being accepted by no less a luminary than the architect of the linguistic turn himself, Hayden White. This survival is not accidental. It is actually the foundation of White’s claims about the methods the historian employs (and must employ) in order to construct a narrative. It is essential to White’s

L A N G U A G E A ND THE HI S T OR I A N

27

version of the linguistic turn, in other words, that history has no technical language of its own, and is consequently obliged to deploy exactly the same devices as those available to any writer of fiction. This is why it is legitimate to regard the historian’s account as a literary artifact, subject to exactly the same analysis as any other. If the thesis that history has to be written in ordinary language is accepted, then it becomes understandable that this language goes un­ noticed, appearing to be a translucent, almost invisible medium for reporting the past. Even more understandable is it that traditional historians have not felt disposed to engage with linguistics or philosophy of language, since the status of ‘ordinary-language-user’ seems auto­ matically to dispense them from any such obligation. I shall examine the ‘ordinary-language’ thesis in detail in a later chapter. There I shall argue that the claim that the language of history is ‘ordinary language’ itself stakes out a theoretical position in philosophy o f history. The ‘ordinary-language’ thesis is only one of a number of philosophical positions that cluster under the protective canopy of the language myth. These will come up for consideration later. But it seems useful at this point to take one or two preliminary examples to illustrate exactly how linguistic assumptions can impinge on philosophy of history. Arthur Danto, whom I quoted earlier in this chapter, does not hesitate to lay his linguistics of history on the table, and in so doing tries hard to distance himself from contractual semantics (or at least from the form he calls ‘Whorfian-Nietzschian’). He tells us: Historical language is external to historical reality, but in no different way than any language is, to which a referential function is attached. Referenda are not penetrated by the expressions which single them out for indication, and so are ‘there,’ independent of and unaffected by the fact that they are referred to: and this is so even if the structure of objects should itself in some way be assignable to language. That is, there may be a deep truth to the Whorfian-Nietzschian theory that grammar determines the way the world is going to be for those who use it, without this in any way changing the logical fact that reference will remain an external relationship between bits of language and bits of reality, even if the kinds of bits referred to vary from grammar to grammar. Reference makes realists of us all. Historical reference then is simply reference in a certain temporal direction relative to the referring expression itself: but that the referendum should be in this respect past does not make it more independent of and less unaffected by language, than were it instead to have been contemporary with the referring expression which falls causelessly upon it. (Danto 1985: 315)

I am not sure I understand everything in this passage, but one thing seems very clear. Anyone who holds that language works in the way

28

THE L I N G U I S T I C S OF HI S TOR Y

Danto describes is not going to present a theory of history in which the historian actually constructs historical reality. For that would be im­ possible, lying beyond the linguistic resources that any historian has available. This impossibility, for Danto, derives from a ‘logical fact’ about language itself - not just about the language of history. Whether Danto is right about this is not the issue; but his taking this view (if he is going to be consistent) constrains the kind of philosophy of history he is free to propose. A somewhat different example comes from an Oxford philosopher of history, Patrick Gardiner, who writes as follows in his book The Nature o f H istorical Explanation: 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)

We should note first of all how what I have called ‘the language myth’ informs the division of labour between historian and philosopher. Gar­ diner seems to assume that both occupy fixed professional positions in a community of intellectuals who share a common language. This is a language in which such expressions as Protestant Reformation and cap­ italist econom y have agreed meanings. This is not to say that Gardiner rules out the possibility that there could be dissent among authorities as to whether the economy in question was a capitalist economy, or when exactly the Protestant Reformation started. On the contrary, Gardiner’s hypothetical philosopher might well expect to unearth such discrepancies and to see them as bearing on the way different historians answer the question. But these would be differences about ‘the facts’, not about the meanings of the expressions. In short, ‘the question’ is identified ab initio in linguistic terms which are assumed to be comprehensible to philos­ opher 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.) It is no less important to note that the philosopher’s investigation is already conceived of as attempting to throw light on the relationship between (1) what the historian says and (2) what grounds the historian has for saying it. In

L A N G U A G E A ND THE HI S T OR I A N

29

short, typical historical discourse is presupposed to be grounded in a relationship between words and reality. Now whether, or to what extent, that may be the case is itself one of the open questions philosophy of history might 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 endorsing one particular stance within philosophy of history. Here we have a case where the philosopher of history is clearly misled by his own linguistic assumptions. The immediate source of Gardiner’s mistake is a familiar one that I have already discussed above: he is taking it for granted that ‘the word history’ has two meanings. Referring to ‘the ambiguity of the word “history” ’, he explains what he understands this ambiguity to consist in 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)

This passage is very revealing because it lays bare, albeit unwittingly, Gardiner’s underlying philosophy of language. It is evident that Gardiner has no qualms about endorsing a reocentric semantics in which meaning is explicated in terms of reference to items and events in the non-linguistic world. This is what explains Gardiner’s claim 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 long ago. In other words, the object of study is not just the verbal output of historians; and it would be absurd for the student to treat it as such (as apparently the literal-minded Keith Jenkins does (Jenkins 1991: 7) when he claims that to say one is reading history means ‘that history is quite literally on library and other shelves’). Thus,

30

THE L I N G U I S T I C S OF HI S TOR Y

ironically, the very example Gardiner chooses to illustrate the ambiguity of ‘the word history’ demonstrates the opposite: namely, that these two supposed meanings cannot be kept in watertight academic compartments. Once we grasp 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 attempting to define the role of the philosopher of history in this way amounts to adopting in advance a particular theory of ‘what history is’. A different but comparable example is to be found in the work of an even more distinguished Oxford philosopher of history, R. G. Collingwood. In The Idea o f H istory (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? [. . .] 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 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. (Colling­ wood 1946: 3)

Here we have no nonsense about ‘the word history’ having two meanings. But Collingwood is clearly committed to a semantic thesis of his own about ‘the word history'. He construes it psychocentrically as designating 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

L A N G U A G E A ND THE HI S T OR I A N

31

philosopher. But what is of interest to me is that the same semantic thesis 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 in the past’. That would be to misconceive altogether the task facing a philosopher of history. It is not part of my case to argue that Danto is right or wrong, or that what Gardiner says makes more sense than what Collingwood says, or vice versa. The point is, rather, that it is difficult to make sense of any of these positions until we see how the philosophy of history on offer meshes with the philosophy of language on offer. And this is not just a question of the adoption of (different) semantic theses about ‘the word history\ although that is certainly symptomatic. It would be a complete mistake to think of the difference as a quibble about the deployment of just one or two key terms, such as history and historical. It involves the whole question of how what the historian says is conceived as relating to whatever is deemed to lie outside language; how words relate to the world in general. Linguistic presuppositions are already embedded in the way historians and philos­ ophers of history formulate their inquiries. And until those pre­ suppositions are recognized as such, even the inquirers do not understand what kind of inquiry they are embarking on. It is futile to pluck decontextualized questions out of thin air and set up the task of answering them as the goal of a discipline. Inevitably, when the starting point chosen is the question ‘What is history?’, it is treated as a kind of superior 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 not successful, had a permanent value of its own. The strategy itself is a pretentious deception. The underlying reason why such questions are preferred is that, being already decontextualized, they lend themselves to analysis in purely linguistic terms, i.e. as reducible to, or treated as presupposing, lexical questions such as ‘What does the word history mean?’. And then recourse can be had to alleged historical evidence, such as commonalities of usage, or attestations in the works of acknowledged authorities, dictionaries, etc. So it emerges that, mirabile dictu, there is no problem after all: we knew what history meant all along, because that is a word belonging to a language we all know (even if philosophers and historians had to go to the trouble of pointing that out to us). In short, ‘What is history?’ is always - and was always - a bogus

32

THE L I N G U I S T I C S OF HI S TORY

question to start with. It is never asked except by those who think they already know the answer and want an excuse for expounding it. Genuine questions, by contrast, never occur except within the contextualized framework of a communication situation. They do not have timeless meanings. If we wish to make sense of genuine questions, we have to treat them not as calls for encyclopedic definitions or comprehensive treatises, but as queries arising out of and integrated into some specific pattern of activities. And the case of history is no exception to that. How did anything that happened in the past come to be recognized as ‘historical’ in the first place? Here we broach what is sometimes called, although not very felicitously, the problem o f‘historical consciousness’. I do not like that way of putting it, because I think that consciousness is a weasel word best avoided. But my argument will be that historical consciousness is actually a readiness to engage in certain forms of communication. To appreciate this we must recognize the foundation of history in the integration of different activities, of which the historian’s is only one. The integration in question is a socially articulated integra­ tion. There may perhaps be societies that have no historical consciousness at all. But even those societies that nineteenth-century anthropologists called ‘primitive’ often had stories about their own past, their ancestors and their origins, and recognized the recounting of such stories by specialist narrators as a distinctive form of discourse. If that is not a prima fa cie manifestation of historical consciousness, it is difficult to know what to call it. There is good reason for suspecting that the notion that early societies lacked a historical consciousness is, as Romila Thapar has demonstrated in the case of India, a fiction generated by imposing inappropriate Western standards of what counts as history (Thapar 2000: 155-72). Thapar has also proposed a general distinction between ‘embedded’ and ‘externalized’ forms of historical consciousness, and illustrated it by reference to the Itihasa-purana tradition in India. Such a distinction is very much in line with my own view that the integration of different communicational activities gives rise to different forms of historification. Thapar suggests that this distinction between ‘embedded’ and ‘external­ ized’ historical consciousness correlates in early India with the transition from a lineage-based society to a state system: ‘the existence of the state requiring its own validation encourages the creation of an externalized historical consciousness’ (Thapar 2000: 125). In the West, I wish to argue, the development that acted as integrational catalyst for the embryonic distinction between embedded and

L A N G U A G E AND THE HI S T OR I A N

33

externalized forms of historical consciousness was not the growth of the Greek city-states but the advent, throughout the Greek world, of writing as a mode of communication. It is with this sociolinguistic innovation that any serious analysis of history as a Western cultural institution must begin.

CHAPTER TWO

History and the literate revolution

Historical consciousness does not automatically furnish its conscious possessors with a ready-made theory of history. Not the least of the oddities about calling Herodotus ‘the father of history’ is that Her­ odotus clearly had nothing that could reasonably be called a theory of history (as distinct from his own reasons for undertaking his selfimposed project). But the same applies to many engaged in the flourishing history business that we have in the twenty-first century, where it has become an institutionalized activity, if not a family of activities. There is a more or less canonical list of ‘great historians’. There are journals, both popular and academic, devoted to history. History is one of the subjects regularly taught in schools and uni­ versities. It is now subdivided into many branches —political history, economic history, art history, etc. Books are written for the explicit purpose of teaching history. Television programmes are devoted to it. The tourist trade organizes ‘historical’ tours and engages lecturers with qualifications as historians to accompany them. But it would be absurd to imagine that everyone today engaged in the history business has a theory of history. Why would they need one, when they already have the institution itself as a going concern and are themselves lucratively employed in it? The modern institutionalization of history-making is the product of a long historical process. One of the landmarks in this process can be traced back to Aristotle’s P oetics, where, for the first time, a celebrated dist­ inction between history and poetry was formulated. This formulation was to remain influential for centuries to come in debates about the nature of the historian’s task. But its full implications have never been thoroughly analysed. The distinction between historian (historikos) and poet (poietes) is not in the one writing prose and the other verse - you might put the work of Herodotus into verse,

34

HI S T OR Y A ND THE LI TERATE REVOL UTI ON

35

and it would still be a species of history; it consists really in this, that the one describes the thing that has been, and the other a kind of thing that might be. (Poetics 1451a39-b4)

Seen from a modern perspective, this may look less like a distinction between history and poetry than simply a distinction between factual account and fiction. In order to see that it is more than that, it is important to realize that what lies behind it is the ancient quarrel between poetry and philosophy. This hinges specifically on the hostility to poetry expressed by Plato. According to Plato, the traditional Greek reverence for Homer and the poets was misplaced. Insofar as poetry and other mimetic arts are concerned with mere appearances, they offer no guide to true wisdom, but divert the seeker from serious inquiry into the under­ lying nature of things. We hear some people say that poets know all crafts, all human affairs concerned with virtue and vice, and all about the gods as well. They say that if a good poet produces fine poetry, he must have knowledge of the things he writes about, or else he wouldn’t be able to produce it at all. (Republic X.598d,e)

In Socrates’ view, these naive persons had been deceived, failing to realize that the works of poets ‘are at the third remove from that which is and are easily produced without knowledge of the truth’ (Republic X.598e, 599a). (At third remove because the poem, like a picture, is only a representation of a representation, being thus thrice removed from reality itself.) Plato’s controversial ‘banishment of the artists’ from his ideal state has been much discussed (Collingwood 1938: 46ff.; Murdoch 1977), but his attitude to history and historians is far less clear. Although Plato offers no explicit opinion concerning the function of the historian, his thinking has been interpreted, notably by Karl Popper, as being imbued with ‘historicism’. In the sense in which Popper uses this term, historicism involves a belief in laws of history. The central doctrine of historicism is that ‘history is controlled by specific historical or evolutionary laws whose discovery would enable us to prophesy the destiny of man’ (Popper 1962: 1.8). The historicist is inclined to look upon social institutions mainly from the point of view of their history, i.e. their origin, their development, and their present and future significance. He may perhaps insist that their origin is due to a definite plan or design and to the pursuit of definite ends, either human or divine; or he may assert that they are not designed to serve any clearly conceived ends, but are rather the immediate expression of certain instincts and passions; or he may assert that they have once served as means to definite ends, but that they have lost this character. (Popper 1962: 1.23)

36

THE L I N G U I S T I C S OF HI S TORY

In Popper’s analysis, Plato emerges as a reactionary maverick who distrusted the prevailing historicism of his day, rebelled against the doctrines of Heraclitus concerning flux, feared change, and tried to show that it can be resisted. This led him to develop (1) an epistemology which postulates, behind the changing appearances of the world, a sempiternal realm of unchanging Forms or Ideas, and (2) a sociology which takes as its ideal state a community governed by unchanging laws. If this analysis is on the right lines, it is easy enough to see why in Plato’s Republic there is no educational role for the teaching of history. History lessons might well inculcate the idea that the affairs of the world are in a constant and ineluctable process of change, which it is beyond the power of mortals to grasp or withstand. Plato would have had even greater cause for alarm if the history teacher had been given to Thucydidean musings to the effect that, human nature being what it is, what happened in the past will ‘at some time or other, and in much the same ways, be repeated in the future’ (Peloponnesian War 1.22). However, as Gilbert Ryle points out (Ryle 1966: 278-9), neither Plato nor Aristotle was averse to at least one kind of history as a pedagogic device in their own teaching; that kind of history being the history of ideas. Aristotle often prefaces his discussion of a topic with an account of earlier views, and in some of his works (Cratylus, Theaetetus, Sophist) Plato seems to adopt the same technique. The speculative history of political systems in Book III of Laws serves a similar purpose. It is introduced by an exchange of which the historicist tone is unmistakable: ATHENIAN: [. . .] But what about political systems? How are we to suppose they first came into existence? I feel sure that the best and easiest way to see their origins is this. CLINIAS: What? ATHENIAN: To use the same method that we always have to adopt when we look into a state’s moral progress or decline. CLINIAS: What method have you in mind? ATHENIAN: We take an indefinitely long period of time and study the changes that occur in it. (Laws 676a)

None of this would make much sense without three historicist assump­ tions. One is that changes in human affairs and in human understanding of the world do occur over a period of time. Second, we can trace such changes, at least in outline. Third, looking back on such changes can teach us lessons worth thinking about. In the light of this, it may seem surprising that the teaching of (some form of) history, by appropriately qualified teachers, does not find a place

HI S T OR Y AND THE LI TERATE REVOL UTI ON

37

in Plato’s ideal educational programme. But the explanation could be that in the ideal curriculum sketched in Plato’s Republic, no teaching of history is necessary. Being well informed about the past may be desirable in itself, but is not an essential qualification for directing the affairs of a just society. If this was indeed Plato’s assumption it can, to be sure, be questioned. No society exists in a chronological vacuum. Those who hold office have to make decisions in particular cases on the basis of available information. The relevant information may well include information about what happened in the past and beyond living memory. There may well be conflicting accounts of this. Disputes about ownership, inheritance and the like were familiar enough in Plato’s Athens. But nothing in Republic or Laws offers the slightest guidance about how the just ruler or the just community is to proceed in such matters. The relevant passages in Laws on the administration of the law contain all kinds of recommendations about setting up courts and juries, deciding penalties, and so on, but nothing that pertains to the assessment of evidence about what happened in the past. This is not even recognized as a special case. What is clear is that Plato’s alleged historicism is far removed from the historicism of modern thinkers who regarded knowing the history of one’s own civilization as a necessary condition for being civilized. In Plato we find nothing like the following passage, where the author directly invites us to try to imagine what everyday life would be like in a society in which no one knew any history. Imagination boggles, because it is only through knowledge of its history that a society can have knowledge of itself. As a man without memory and self-knowledge is a man adrift, so a society without memory (or more correctly, without recollection) and self-knowledge would be a society adrift. (Marwick 1970: 13)

Here a society ignorant of its own history is represented as being like a man suffering from amnesia. The lesson, evidently, is that among the primary duties of a public education system is one that is lacking in Plato’s ideal state; namely, to ensure that students learn the history of their own society. The metaphor of amnesia is interesting in this connexion for at least two reasons. One, which involves the interpretation of the enigmatic dialogue Timaeus, will be considered further below. The other is that Plato’s explanation of how we have any acquaintance with the eternal Forms at all is bound up with the doctrine of anamnesis, which assumes we have had direct intellectual experience of the Forms in a previous

38

THE L I N G U I S T I C S OF HI S TOR Y

existence (Phaedo 72e-77a). This experience was forgotten at birth but is rescuscitated when in life we perceive, through the senses, things that remind us of the Forms. Our amnesia is thus overcome and we recover our lost knowledge. Plato out-Poppers Popper. This is not just historicism but transcendental historicism. The institutionalization of history in the West is not to be confused with the establishment of practices such as record-keeping and preserving lists of kings, or ancestors, or officials, or decrees, or transactions. All these may offer valuable materials for the historian, but the historian qua historian stands outside those practices, even while engaging with them. It is the character of that engagement, and its public recognition as a separate enterprise, which mark the institutionalization of history. Here again Aristotle (not himself a historikos in his own sense of the term) is an invaluable witness for the way he proceeds to elaborate the distinction between history and poetry. He claims {pace Plato) that poetry ‘is something more philosophic and of graver import than history’ {Poetics 1451b). And the reason he gives for this surprising pronouncement is that ‘its statements are of the nature rather of universals, whereas those of history are singulars’ {Poetics 1451b). He explains this distinction between poetic statements and historical statements in overtly linguistic terms: By a universal statement I mean one as to what such or such a kind of man will probably or necessarily say or do - which is the aim of poetry, though it affixes proper names to the characters; by a singular statement, one as to what, say, Alcibiades did or had done to him. {Poetics 1451b)

The most obvious objection to this account is that many poems are ostensibly about historical figures and events, who are named as such. Aristotle immediately counters this difficulty with an astonishing piece of sophistry. He concedes that tragic poets ‘still adhere to the historic names’ but explains this away as part of their stock-in-trade: In Tragedy, however, they still adhere to the historic names; and for this reason: what convinces is the possible; now whereas we are not yet sure as to the possibility of that which has not happened, that which has happened is manifestly possible, otherwise it would not have happened. {Poetics 1451b)

Aristotle must have been having a bad day at the Lyceum if the surviving text of the P oetics at this point accurately reflects his thinking. What we are being asked to believe, in other words, is that the poets used the historic names simply in order to lend conviction to characters and

HI S T OR Y A ND THE LI TERATE REVOL UTI ON

39

actions that might otherwise strain credulity. This seems an unlikely story. Much more likely is that the poets and their audiences drew no clear distinction of the kind that Aristotle wanted to draw. They did not recognize what would nowadays be called the difference between myth and history. And in some ways, it has to be admitted, a refusal to recognize that difference with respect to what happened or may have happened in the long distant past is not only a defensible but a sophisticated position. Gilbert Murray’s astute comment on Aristotle’s use of the term mythos is relevant here. He points out that in Aristotle’s hands its meaning is almost reduced to what nowadays would be called ‘plot’ (Murray 1920: 13). Murray also refers to various other details that suggest that Aristotle was out of sympathy with the religious basis of Greek drama. If Murray’s points are accepted, one is led to a view of the Poetics which makes it, in effect, a revolutionary text for its time. Far from being a summary of received wisdom about the practice of Greek poets, it proposes a rationalization - in psychological terms - not only of the traditional stories that the poets reworked, but o f their reworking. It is in this context that Aristotle’s theory of the historian’s task falls into place. (He never discusses it elsewhere.) History exemplifies the most basic type of narrative, by comparison with which more ambitious literary narratives stand to be defined. Accordingly, what Aristotle is trying to do in his Poetics is to make out a case for saying that poetry has to be taken seriously by the philosopher, not for the negative reason that Plato had adduced (namely, that poets perpetuate unreliable stories against which the honest citizen must be warned), but for the positive reason that poets are trying to represent universal truths about human nature. This is a shrewd ad hominem turning of tables against Plato, who had claimed that knowledge is only of universals, whereas concerning particulars we can hope for no more than warranted belief, which is fallible and always open to revision. (It may turn out, after all, that everyone hitherto had got the date of the seige of Troy wrong, or that a rough guess was the best that could be expected. But no one - in that sense - could get the geometrical definition of a triangle wrong. Nor could it be held that a rough approximation to triangularity was the most geometry could hope for. Here ex hypot he si the seige of Troy is a particular, whereas the triangle is a universal.) If this is the right way to look at Aristotle’s strategy in his P oetics, it raises various questions. One is whether this idealized account of the poet’s purpose actually corresponds to the poetic praxis of Aristotle’s age. But more relevant for our present purposes is whether, in order to make

40

THE L I N G U I S T I C S OF HI S TOR Y

his case for poetry, Aristotle has not, either deliberately or unwittingly, done the historian less than justice. Fortunately, we do not have to waste time worrying about the kind of writer Aristotle has in mind when he refers to the historikos, because he cites Herodotus as an example. And those familiar with the Histories of Herodotus will immediately complain that it is absurd to regard the historian as concerned only with particulars. In fact, there are at least two complaints here. One is that Herodotus explains straight away as his motivation not the accurate establishment of particular facts but something far more ambitious: Herodotus of Halicarnassus, his Researches are here set down to preserve the memory of the past by putting on record the astonishing achievements both of our own and of other peoples; and more particularly, to show how they came into conflict. (Histories 1.1)

There will always be room for sceptics to object that historians are not to be trusted on the subject of their own aims. And this objection cannot be lightly brushed aside. On the other hand, there is no more cogent reason to prefer Aristotle’s account of what the historian is trying to do, particularly since that is obviously tailored to fit Aristotle’s own agenda. So I propose to see how far we can get by taking Herodotus at face value. Herodotus could not have fulfilled his stated aims simply by producing a sequence of what Aristotle calls singular statements. Herodotus takes it for granted from the start that his reader will not be content with a chronological list of events, but will expect him to provide explanations for the events in question. Thus, for example, he feels called upon to explain why Miletus could afford to be relatively independent in its dealings with Cyrus. The explanation he offers is ‘the general weakness of the Hellenic peoples at that date, and particularly of the Ionians, who of all the Greek races had the least power and influence’ (Histories 1.143). Not much of an explanation, someone may object, because the indepen­ dence of Miletus is part of the evidence for saying that the Hellenic peoples were weak at that time. That objection misses the point. The point is that weakness, like power and influence, is a psychological concept, even when translated into military or economic terms. The assumption is that communities, like individuals, do not - except in desperate circumstances - attempt to do things they fear they are not strong enough to do. To take a different example, Herodotus explains Cambyses’ cruel and seemingly irrational crimes committed against his own kin by suggesting that he was suffering from epilepsy (Histories 3.33). This is pure speculation, but again it shows Herodotus’ sensitivity to

HI S T OR Y A ND THE LI TERATE REVOL UTI ON

41

demands for explanations of human behaviour, and these explanations, although applied to particular instances, invariably take the form of universals (‘Epileptics act in this or that way’). Examples could be multiplied. Evidently the historian is no less concerned with generalizations about human nature than the poet. For explanations conform to the same implied syllogism and the same kind of universal major premise: People with such-and-such characteristics, or in such-and-such circum stances, will behave in such-and-such a manner. What this shows is how feeble Aristotle’s attempt to distinguish poetry from history is. Anyone who wanted to turn Aristotle’s case upside down need only point out that, by Aristotle’s own reasoning, it ought to follow that history is poetry par excellence. Or that all poets should give up poetry and try their hand at history. For the fact that historical figures did behave in the way they did surely adds credence to any explanation: ‘that which has happened is manifestly possible, otherwise it would not have happened’. The speciousness of Aristotle’s reasoning is brought home even more conspicuously when we consider the way in which both Herodotus and Thucydides put carefully composed speeches into the mouths of their characters. These were never conceived as - and could not have been verbatim records of speeches actually made. Thucydides himself draws our attention to this. I have found it difficult to remember the precise words used in the speeches which I listened to myself and my various informants have experienced the same difficulty; so my method has been, while keeping as closely as possible to the general sense of the words that were actually used, to make the speakers say, what, in my opinion, was called for by each situation. {Peloponnesian War 1.22)

This directly confirms the general point made above; the historian, no less than the poet, relies on principles of verisimilitude which in the end reduce to universals. In other words, the assumption always is that individuals in such circumstances would be likely to act in such a way. M. I. Finley points out that here Thucydides puts a noose round his own neck: There is no way to get round the incompatibility of the two parts of that statement. If all speakers said what, in Thucydides’ opinion, the situation called for, the remark becomes meaningless. But if they did not always say what was called for, then, insofar as Thucydides attributed such sentiments to them, he could not have been ‘keeping as closely as possible to the general sense of the words used’. It is also worth remembering that he never claimed to make generals, for example, conduct themselves in a manner called for by each situation, as distinct from what they actually did. (Finley 1972: 26)

42

THE L I N G U I S T I C S OF HI S TORY

The observation has a direct bearing on what Aristotle says in his Poetics. It is not simply that Thucydides’ speeches are ‘poetic’ construc­ tions intruding into an otherwise historical narrative. On the contrary, they are an intrinsic part of that narrative, for they are made to provide the psychological grounding for the actions of the main characters. So here the Aristotelian distinction between poetry and history breaks down in any case. But that is not all. What emerges as suspect is Aristotle’s claim that the poet, as opposed to the historian, is concerned with capturing universal truths about human nature. For if that were so, one would expect the poet to concentrate on very common patterns of human behaviour; whereas, on the contrary, the poet’s tale is frequently one of quite extraordinary deeds and circumstances. An epic in which nothing unusual ever happened would be a very boring poem indeed. In brief, it is not just that Aristotle’s distinction between poetry and history is unconvincing, but that both the account of poetry and the account of history resulting therefrom are independently implausible. However, the fact that Aristotle made a botched job of it must not distract our attention from the history-making process in which he is involved. It was not what Cambyses actually did (or may have done) to his relations that - for Aristotle - made history; nor that Herodotus sub­ sequently reported it. Aristotle integrated what Herodotus wrote about what Cambyses did into an intellectual programme of his own. One of the requirements of this programme was that the way Herodotus wrote about Cambyses (and others) should be a paradigm case illustrating the difference between history and poetry. But what prompted Aristotle it is important to note - was no concern for discovering the truth about the past (in this instance, about Cambyses’ behaviour). Nor even about whether Herodotus’ explanation of Cambyses’ behaviour was plausible. Aristotle’s motives sprang from his own engagement with the philosophy of Plato - something of no direct concern either to Cambyses or to Herodotus, neither of whom had, or could have had, any awareness of the role in which Aristotle’s theory retrospectively placed them. The example illustrates my more general thesis. What matters, if I am right, is not whether certain events did or did not take place, nor whether any subsequent account of them was or was not correct, but whether the events in question and the account in question can both be integrated into a view of human intellectual activity which identifies ‘history’ as a separate category with its own characteristics. Given this integration, then the events in question and the account in question have conjointly been ‘historified’, i.e. are then accorded the status of being ‘historical’. What historification amounts to, in practice, can only be assessed, for

HI S T OR Y AND THE LI TERATE R EVOL UTI ON

43

any particular society and any particular time, through examining the criteria by which that category is held to be distinguished from others. The crucial question, in any given instance, concerns the basis on which its institutionalization rests. There are, in principle, many possibilities. Aristotle, in my view, is a pioneer in the development of historification in the West because he is the first to propose explicit criteria which set up ‘history’ as a communicational enterprise sui generis. In retrospect, doubtless, Aristotle’s criteria look extremely unsatisfact­ ory. But of course they would; because we now judge them with what we call, hubristically, ‘historical hindsight’. Could something be salvaged from Aristotle’s theory if we assumed that what he had in mind was not really the difference between historian and poet (although that is what he says) but rather what we would nowadays see as the distinction between history and myth? And could this confusion - if confusion it were - be explained simply by pointing out that the major poets of ancient Greece (Homer, Hesiod, Aeschylus, et al.) were also those whose work dealt with mythological themes handed down from the remote past, whereas the historikos (if Herodotus and Thucydides are typical examples) dealt with more recent events and their immediate antecedents? Could the differ­ ence boil down to a difference between stories going back to times beyond recall (and therefore totally unverifiable) and, on the other hand, stories about a more recent past that was not beyond research and recovery? As historians nowadays constantly remind us, the Greek term historia, when Herodotus used it, meant something like ‘inquiry’ rather than ‘history’. The first point to be made in this connexion concerns the reason why Socrates and Plato objected to the Greek myths. It was not, as far as one can see, that the myths as such are unverifiable. When Socrates discusses the nature of piety with Euthyphro (Euthyphro 6ff), his reservations are based on the fact that the stories told about the gods are hard to believe and repugnant to one’s sense of morality and decorum. For example, Zeus is said to have castrated his father for swallowing his sons. Socrates says he finds it difficult to accept that. And do you believe that there really is war among the gods, and terrible enmities and battles, and other such things as are told by the poets, and other sacred stories such as are embroidered by good writers and by representations of which the robe of the goddess is adorned when it is carried up to the Acropolis? Are we to say these things are true, Euthyphro? (Euthyphro 6b,c)

44

THE L I N G U I S T I C S OF HI S TORY

If these things are true, then the gods are unworthy role models. The criticism was by no means new. In an earlier generation, Xenophanes had likewise criticized Homer and Hesiod for having ‘attributed to the gods everything that is a shame and reproach among men’ (Kirk, Raven and Schofield 1983: 168.) Nor, says Socrates, can piety be defined as that which is pleasing to the gods (as Euthyphro holds), because according to the myths different gods are apparently pleased by different things. (Impiety, it must be remembered, was the charge brought against Socrates, on which he was eventually condemned to death.) Similarly in Republic the objection to Hesiod’s account of Uranus’ behaviour and how Cronus punished him for it is that even i f it were true it is not a story to be passed on to the young. The battles of gods and giants, and all the various stories of the gods hating their families or friends, should neither be told nor even woven in embroideries. I f we’re to persuade our people that no citizen has ever hated another and that it’s impious to do so, then that’s what should be told to children from the beginning by old men and women; and as these children grow older, poets should be compelled to tell them the same sort of thing. We won’t admit stories into our city - whether allegorical or not - about Hera being chained by her son, nor about Hephaestus being hurled from heaven by his father when he tried to help his mother, who was being beaten, nor about the battle of the gods in Homer. The young can’t distinguish what is allegorical from what isn’t [. . .]. {Republic 378c,d)

In Phaedrus Socrates disclaims any interest in even trying to provide rational explanations for the more fanciful myths, as - we gather from his remarks - some of his contemporaries prided themselves on doing. Thus he will not be drawn on what he thinks of the legend that Orithuia was carried away by Boreas. ‘I am still unable, as the Delphic inscription orders, to know myself; and it really seems to me ridiculous to look into other things before I have understood that’ (Phaedrus 229e-230a). But if that order of priority is to be strictly adhered to, it seems to rule out historical inquiry too. However, as commentators have often noted, Plato himself is fond of introducing myths and allegories into the dialogues in order to establish assumptions and perspectives that it would be difficult to justify by rational argument. It is interesting to note that on one such occasion, the story of Atlantis in Timaeus, the question of veracity does arise, and is even insisted on at some length. Critias introduces it by claiming that the story is ‘a very strange one, but even so, every word of it is true’ ( Timaeus 20d). It is a story that was vouched for by Solon, ‘the wisest of the seven sages’ and a friend of Critias’ great-grandfather, who in turn told it to

HI S T OR Y AND THE LI TERATE R EVOL UTI ON

45

Critias’ grandfather. Socrates immediately says he has never heard of this story and asks whether it really happened. Critias responds to this by giving details of the circumstances and even the day on which the story was first told to him, and how it had been brought back from Egypt by Solon, who had it from Egyptian scholars. When Critias has told the story in outline, Socrates declares it a fitting offering to the festival of the goddess whose day it is. ‘And of course the fact that it’s no made-up story but a true account is no small matter’ ( Timaeus 26e). In short, Socrates seems to welcome the story of Atlantis as a serious contribution to the history of Athens. Whether this response is to be understood as a piece of Socratic sarcasm I find it hard to judge. It seems significant that the anecdotal circumstances of the original story, as told by Critias, include Solon’s admission that the Greeks were almost completely ignorant of the history of their own people. The scholars of Egypt chide him on this score: ‘Ah, Solon, Solon, you Greeks are ever children. There isn’t an old man among you.’ On hearing this, Solon said, ‘What? What do you mean?’ ‘You are young,’ the old priest replied, ‘young in soul, every one of you. Your souls are devoid of beliefs about antiquity handed down by ancient tradition. Your souls lack any learning made hoary by time.’ ( Timaeus 22b)

It then turns out that the story of Atlantis explains this historical ignorance. The cataclysmic flood that drowned Atlantis also swept away most of the Greek population, leaving only a few shepherds and herdsmen in the mountains. Greek history was lost because no records were preserved. Furthermore, according to the Egyptian, such floods recur cyclically. Each time, Greek civilization is destroyed and has to begin again. [. . .] no sooner have you achieved literacy and all the other resources that cities require, then there again, after the usual number of years, comes the heavenly flood. It sweeps upon you like a plague, and leaves only your illiterate and uncultured people behind. You become infants all over again, as it were, completely unfamiliar with anything there was in ancient times [. . .]. (Timaeus 23a,b)

This is something much closer to the notion of collective amnesia, but how to take Plato’s point is far from clear. Ostensibly, a historical explanation (the flood) is being provided for a current state of ignorance. Yet this historical explanation itself is evidently fictitious. Plato is the first writer to mention Atlantis: it was not part of traditional Greek mythology. Given Plato’s attitude towards the poets, a profound irony pervades the whole scene of Socrates accepting straight away this previously unknown story.

46

THE L I N G U I S T I C S OF HI S TORY

What does seem relevant is that the story of Atlantis is produced in order to fulfil a desideratum that Socrates himself has introduced. He has said he would like a story illustrating the ideal state in action, for he feels that his own description of the ideal state is static and lifeless. He compares its inadequacy with that of a picture of an animal standing still: what we want is a picture of the animal ‘in motion or engaged in some struggle or conflict’ ( Timaeus 19b). For only then can we see what the animal is capable of. Critias’ story of Atlantis is a response to this demand. But that is not all. What Critias claims is that we do not have to imagine or speculate how Socrates’ ideal state might perform in response to some challenge; for in the past the Greek state did perform in such a manner, as the war against Atlantis illustrates. Critias tells Socrates: W e’ll translate the citizens and the city you described to us in mythical fashion yesterday to the realm of fact, and place it before us as though it is ancient Athens itself. And we’ll say that the citizens you imagined are the very ones the [Egyptian] priest spoke about, our actual ancestors. The congruence will be complete, and our song will be in tune if we say that your imaginary citizens are the ones who really existed at that time. (Timaeus 26c,d)

This is a blatant appeal to a ‘golden age’ scenario; Socrates’ account of the ideal state is now being presented as a description of a state that actually existed in ancient Greece, of which the memory was subsequently obliterated. But what is the point of that elaborate fiction? Whatever it is, it undoubtedly chimes with Aristotle’s later contention, already examined above, that the poets use historical names because that gives their work greater credibility: ‘that which has happened is manifestly possible, otherwise it would not have happened’. The story of Atlantis, as pre­ sented by Critias, is just such a case, and its ‘authenticity’ is allegedly guaranteed by the wisdom of Solon and the ancient records of the Egyptian priests. If this is historical ‘authenticity’, then it is the last thing that either Socrates or his puppet-master Plato would ever seek or rely on. Yet here Socrates apparently accepts it. Why? Because the whole point of the Timaeus (and presumably of Plato’s abandoned Critias also) was to demonstrate that philosophers can come up with stories as fascinating as those told by the poets. Stories are valuable not in direct proportion to the amount of the past that they preserve (which we can never be sure of anyway) but insofar as they enable us to focus imaginatively on enduring questions of philosophical importance. #

#

#

HI S T OR Y AND THE LI TERATE REV OL UTI ON

47

The above background is interesting and relevant, but it does not provide a direct answer to the question of where Aristotle’s differentiation between history and poetry came from. That answer is contained, I think, in the very terms that Aristotle employs to describe it. The distinction between universals and particulars is not a historical distinc­ tion but a linguistic distinction. It relates directly to Aristotle’s logic and its concomitant philosophy of language. The linchpin of Aristotle’s linguistics was the difference between what would nowadays be called proper names and common nouns. S ocrates is a proper name designating one particular individual, whereas man is a common noun, applicable to many such individuals. What is the basis of this distinction? What makes it possible for words like man to apply to many individuals, regardless of the irreducible differences between them? Plato’s theory of Forms or Ideas was a bold attempt to solve at one stroke both that linguistic problem and the more general problem of how it is possible to have any knowledge of universal truths (given that the experience of each individual one of us is very limited). Cornford sums up very concisely the relevant difference between Plato and his teacher Socrates on this question: Socrates, who was not concerned with any system of Nature, confined himself to the attempt to define moral terms, such as ‘Just’. Plato (who was concerned with ontology), accepting the Heracleitean Flux as applied to sensible things, saw that the subject of a Socratic definition could not be any sensible thing, since such things are in perpetual change and cannot be known; so he said that it must be a separate entity, to which he gave the name ‘Form’, and that the group of sensible things bearing the same name partake of that Form. The underlying assumption here is that every common name must have a fixed meaning, which we think of when we hear the name spoken: speaker and hearer thus have the same object before their minds. Only so can they understand one another and any discourse be possible. (Cornford 1935: 9)

The basic point might be put even more succinctly. Both Socrates and Plato saw that the fundamental question for philosophy was one of linguistic epistemology: ‘Do we know what we are talking about?’. If we don’t, then discussion is futile and no philosophical method such as the Socratic question-and-answer method will ever get us any further. Plato saw this much more clearly than Socrates. But he was convinced that the Socratic ‘linguistic’ technique of inquiry was right, and sought to justify it. This was what led him to the doctrine of Forms. Plato saw that until the question of linguistic epistemology was tackled, any sophist could set up as a teacher and claim to be propounding ‘knowledge’.

48

THE L I N G U I S T I C S OF HI S TOR Y

Plato, in brief, is trying to solve precisely the semantic problem that the language myth poses. How can we be sure that the linguistic code incorporates any fixed meanings at all? Plato’s answer is his theory of eternal Forms. He postulates that (at least some) words get their meanings by ‘standing for’ or ‘representing’ things outside language altogether; and when these things in turn are assumed to be sempiternal realities, the desired guarantee of semantic stability is at last in place. This has an obvious parallel to the doctrine that pictures get their meanings by ‘standing for’ or ‘representing’ the subject they depict. Both words and pictures are ‘signs’ (in some sense of that elusive term). As Cornford observes: ‘With his usual disregard for precision, Plato uses all the common words for “signify”, “mean”, “indicate”, indiscriminately’ (Cornford 1935: 307). The parallel (or conflation, if we reject it) is highly relevant to Plato’s mistrust of both poets and painters, and it is constantly invoked, particularly in the discussion of the arts in Republic. Aristotle too relies on a fixed-code semantics, but his linguistic epistemology is subtly different from Plato’s. This is because he rejects Plato’s view that the Forms exist eternally and independently in a realm of their own. Once that epistemological prop goes, there is no longer any question of common nouns having meaning by ‘standing for’ some entity which lies outside the world of everyday experience. So how do they come to mean what they do? And how can we be sure that words like man mean the same for everybody? The fundamental problem, clearly, is the same as it was for Plato: do we know what we are talking about? Or is philosophical discussion doomed for ever to be a dialogue des sourds? Aristotle attempts to solve this problem by making a move that is no less questionable than Plato’s. He claims that we all perceive the world identically, or as he puts it, we all have the same ‘affections of the soul’ (De Interpretatione 16a). This in turn forces him to treat words as mere tokens (symbola) of affections of the soul. So different languages are just different sets of vocal tokens for exactly the same mental entities (i.e. ‘affections’ of the human soul). On this view, there is no philosophical problem about communication: we can carry on without worrying about whether we know what we are talking about. We know what we are talking about because a benevolent providence has somehow ensured that there is no discrepancy between human beings as to what they are talking about or how (i.e. by means of symbola covering exactly the same range of meanings). It is precisely on this linguistic epistemology that the foundation of Aristotle’s universal logic (the syllogism) is laid. The syllogism collapses once that epistemol­ ogy is questioned.

HI S T OR Y AND THE LI TERATE R EVOL UTI ON

49

A fo rtio ri Aristotle’s theory of history collapses along with it. It can no longer be assumed that there is an identically perceived world out there, that historians describe in a set of carefully chosen statements about particulars, thus avoiding all the problems surrounding universals. Nevertheless, what emerges unmistakably from Aristotle’s discussion is (1) that he lives in a society where the institutionalization of history is up for grabs, because people are unsure about what exactly the function of the historian is, and (2) that the question of how that institutionalization is justified is already posed in linguistic terms, as a question about how the historian qua historian uses language. What do we learn from this that is relevant to the philosophy of history? First of all, we learn that Aristotle comes up with a theory of history because he needs one. The answer to Collingwood’s question ‘Why does Plato write as if Herodotus had never lived?’ may be not, as Collingwood suggests, Plato’s anti-historicism, but simply that Plato did not need a theory of history. But Aristotle did. It is noteworthy that Aristotle does not start from any abstract question such as ‘What is history?’ or try to examine in detail the characteristics of particular historical texts. His need for a theory of history arises in a very specific context, which is itself reflected in the features of the theory. In fact, the theory, such as it is, deals only with those aspects of history that are relevant to that context. The need for such a theory arises because Aristotle is trying to integrate his account of poetry with his more general account of language and logic. Poetry has to be explained as a category of discourse that is sui generis. Poets have to be defended against the (Platonic) charge that they spread tales that are false. These are the twin requirements, and they are both linguistic requirements. The historian is brought into the picture as a necessary foil to the poet. Aristotle’s strategy is to contrast the historian’s treatment of the past with the poet’s. What would be a serious criticism of the historian (for instance, that an incident he describes never took place, or that it took place somewhere else than where he locates it, or involved other participants than those he names) is not eo ipso a serious criticism of the poet. The reason for this is that poetry and history involve different uses o f language. It is not that one of them pays attention to the truth while the other ignores it, but that the poet is concerned with a different level of truth from the historian. The poet’s truths are general truths about human nature, whereas the historian’s truths are particular truths about human deeds. Aristotle is no more concerned with the question of how the poet arrives at truths about human nature than he is with the question of how

50

THE L I N G U I S T I C S OF HI S TORY

the historian arrives at truths about particular human deeds. That, in the context of his argument, is an irrelevance, and consequently his theory of history has nothing at all to say about the problem of historical evidence, which was to vex later theorists considerably. To sum up, we learn that Aristotle does not first elaborate a theory of history and then look for a theory of language to back it up. The boot is on the other foot. His theory of history is tailored to fit a space made available by his theory of language (the space available for a corpus of particular statements about past events). It does not seem to worry Aristotle that the works of individual historians may not actually fit this allocated space very well. (He might have argued that, just as the failure of particular officials to carry out the duties of an office does not affect the way the office is defined, so the fact that particular historians lose sight of, or are diverted from, their proper task does not alter the nature of the task itself.) The interesting question that arises is whether making a theory of history fit a theory of language is just a peculiarity of Aristotle’s approach to philosophy of history. Or is it one instance of a pattern that recurs throughout the Western tradition? I shall go on to argue that indeed it is, and to inquire exactly how that pattern has been maintained down to the present day. The way Aristotle conducts his discussion of history makes it clear that he thinks of the historian as one kind of writer. The explanation of this aspect of the institutionalization of history, as well as its significance, require us to consider a wider social context in the ancient world. And that in turn demands a still wider context in which a global distinction is drawn between ‘literate’ and ‘pre-literate’ societies. Such a distinc­ tion, however, turns out to be less straightforward than one might have supposed. In his book The Literate Revolution in G reece and its Cultural Con­ sequences, Eric Havelock proposed to investigate ‘the material conditions surrounding a change in the means of communication between human beings, social and personal’ (Havelock 1982: 7). The change in the means of communication that Havelock had in view was the advent of (Greek) writing. He added: Underlying the analysis, and for the most part unstated but perceptible, lies the possibility of a larger and more formidable proposition, that the change became the means of introducing a new state of mind - the alphabetic mind, if the expression be allowed. (Havelock 1982: 7)

HI S T OR Y A ND THE LI TERATE R EVOL UTI ON

51

The Literate R evolution was the third in a series of books - the first two being P reface to Plato (1963) and Origins o f Western L iteracy (1976) - in which Havelock explored the relationship between the introduction of writing in Greece and the development of what was later recognized as the Greek genius for abstract analysis. Havelock’s pioneering work had a frosty reception from the more conservative wing of Classical scholarship, but it eventually jolted even his critics into paying more serious attention to the conditions of literacy in the ancient world than they had paid hitherto. Meanwhile a parallel hypothesis was being taken up on a larger scale by scholars in other fields, notably anthropology. In 1963 Goody and Watt published a paper in which they argued that the development of logic was ‘a function of writing’, inasmuch as it focussed attention on the separation of words, their possibilities of combination and recombination, and the manipula­ tion of syllogistic patterns. This was followed in 1977 by Goody’s landmark book The Domestication o f the S avage Mind, in which studies of contemporary pre-literate communities in Africa were brought to bear on the question. In 1982 W. J. Ong published O rality and L iteracy, in which he argued the general thesis that ‘writing restructures thought’. A few years later appeared R. K. Logan’s book The Alphabet E ffect (1986), with its subtitle The Im pact o f the Phonetic Alphabet on the D evelopment o f Western Civilization. According to Logan, the alphabetic mind that originated in Greece was responsible inter alia for the emergence in Europe of ‘systematic science’. This was seen by Logan as contrasting historically in a significant way with the absence of any comparable development in China, where a non-alphabetic form of writing held sway for thousands of years. Common to all these works was the general idea that the adoption of a certain form of written communication correlates with the development of a certain mental outlook and of certain cognitive skills. Which was chicken and which was egg remained less clear. Critics seized upon this as a theoretical weakness. William V. Harris, who dismissed Goody’s thinking as ‘woolly and grandiose’, said in his book Ancient L iteracy that the reaction of a historian to claims ‘such as those of Havelock and Goody is likely to be a desire for detail’ (Harris W. V. 1989: 41). But then it becomes radically unclear whether the details - at this distance in time support the Havelock-Goody thesis or the opposite interpretation that analytic ‘predispositions were already flourishing in Greece when people began to write’ (Harris W. V. 1989: 41). Logan admits that it would be ‘simplistic’ to conclude that the difference between writing systems alone explains ‘the emergence of abstract science’ in Europe and not in China.

52

THE L I N G U I S T I C S OF HI S TORY

He refers to the reservations expressed by the great Sinologist, Joseph Needham, about the idea that the Chinese ideographic system had been ‘a powerful inhibiting factor’ on the development of modern science in China. But he thinks Needham underestimated ‘the subliminal or hidden effects of rendering scientific words in an ideographic script’ (Logan 1986: 55). Logan mentions Havelock, but acknowledges Marshall McLuhan as the formative influence on his own thinking about literacy. The Chinese had indeed been described by McLuhan, a quarter of a century earlier, as being ‘tribal, people of the ear’ (McLuhan 1962: 38). This astonishing ethnocentric generalization is only partly mitigated by McLuhan’s admission that ‘tribal’ people are often more sensitive in various ways than ‘civilized’ people. Civilization, in McLuhan’s opinion, is often crude by comparison with the ‘hyperesthesia of oral and auditory cultures’ (McLuhan 1962: 38). These conclusions were the product of McLuhan’s studies of com­ munication in Western countries. In the course of a number of public­ ations from the 1950s onwards, he had developed a critique of the mass media and their impact on society (The M echanical B ride, 1951; The Gutenberg Galaxy, 1962; Understanding M edia, 1964; The Medium is the M assage, 1967). It was printing, rather than writing, that attracted his attention. (The subtitle of The Gutenberg Galaxy is The Making o f Typographic Man.) Typography, in McLuhan’s view, was ‘the architect of nationalism’. It was ‘no more an addition to the scribal art than the motorcar was an addition to the horse’ (McLuhan 1964: 158). According to one commentator: McLuhan assumes the stance of a sophisticated Luddite, distinguishing himself from his machine-breaking predecessors by the way in which he points out that the discovery of printing was the original sin from which all the subsequent woes of industrial civilisation are derived. (Miller 1971: 11)

At the basis of McLuhan’s approach, according to the same comment­ ator, is the assumption that methods of human communication vary in effectiveness and effect according to the number of sensory channels employed. The larger the number of senses involved, ‘the better the chance of transmitting a reliable copy of the sender’s mental state’. McLuhan believed that ‘the spoken word answers these requirements more faithfully than any other medium’ (Miller 1971: 9). Two points seem to me worthy of note here. One is that the traditional Western language myth emerges as the basis of all McLuhan’s thinking. McLuhan’s version of this myth is an unorthodox one, and based in part

HI S T OR Y AND THE LI TERATE REVOL UTI ON

53

on his eccentric misreading of Saussure (Gordon 1997: 323-38). But he accepts without question that communication is a telementational process: the message must go from Als mind to i?’s. Once this is taken for granted, the key question automatically becomes: ‘By what means is this trans­ ference from one mind to another best effected?’. And it seems obvious to McLuhan that not all means can be equally ‘good’. Cultures diverge, in McLuhan’s view, according to the means of communication they favour, and these means in turn are to be judged by the quality of the transference they make possible. The second point is that McLuhan echoes Plato in thinking that viva voce communication is superior to written communication, whereas in the centuries separating Plato from McLuhan the opposite assumption had often been made. It might at first seem odd that McLuhan should opt for sound as the best channel of transmission. But by speech McLuhan understands rather more than mere vibration of the vocal chords and corresponding stimulation of the auditory nerve. He claims: ‘The spoken word involves all of the senses dramatically’ (McLuhan 1964: 81). Speech in this sense may include facial expression, gesture, posture and other natural concomitants of face-to-face interaction. As Miller puts it, for McLuhan ‘the spoken word activates the entire human sensorium and thereby underwrites the accuracy with which the spoken message rep­ roduces the mental state to which it supposedly corresponds’ (Miller 1971: 9). By comparison, writing is seen as a poorer, attenuated, more artificial, ‘cooler’ form of communication, which is further removed from the actuality of the mental state which it expresses. (Cool and its antonym hot became, in McLuhan’s idiosyncratic vocabulary, technical terms for describing forms of communication.) In McLuhan’s version of the language myth, there is never any clear account of exactly what human thought is like in its pure, unadulterated form; and hence it becomes - conveniently for McLuhan - impossible to evaluate his central claim that thought is more fully ‘represented’ in certain forms of communication (notably, speech) than in others (notably, writing). Plato’s reasons for preferring speech to writing were somewhat dif­ ferent, although they are not entirely unconnected to McLuhan’s. Plato’s scepticism about writing attacks the very ground on which writing is often seen as having a great advantage over speech: that it provides a permanent record of what was said or agreed. This is one aspect of what Goody called ‘the storage function’ of writing, which ‘permits communication over time and space, and provides man with a marking, mnemonic and recording device’ (Goody 1977: 78). As regards the recording of speech, Plato held

54

THE L I N G U I S T I C S OF HI S TORY

the storage function to be an illusion. For once words are divorced from the speaker and the circumstances in which they were uttered, they are no longer a record o f anything, but a new text into which everyone is free to read their own interpretation, regardless of whether they know or under­ stand anything about its origin or the thinking behind it. Plato here puts his finger on a point which seems to me of paramount importance for the historian: far from writing being a ‘record’ of what was originally said, writing is itself an instrument of decontextualization (and ipso fa cto of potential falsification). In Phaedrus Socrates argues: When it has once been written down, every discourse roams about everywhere, reaching indiscriminately those with understanding no less than those who have no business with it, and it doesn’t know to whom it should speak and to whom it should not. {Phaedrus 275 e)

According to Socrates, those who think they can leave written instructions for an art, as well as those who accept them, thinking that writing can yield results that are clear or certain, must be quite naive [. . .]: otherwise, how could they possibly think that words that have been written down can do more than remind those who already know what the writing is about? {Phaedrus 275c,d)

Along this line of thinking, committing anything serious to writing, particularly when it affects the lives of others, is an irresponsible act: if Lysias or anybody else ever did or ever does write - privately or for the public, in the course of proposing some law - a political document which he believes to embody clear knowledge of lasting importance, then this writer deserves reproach, whether anyone says so or not. For to be unaware of the difference between a dream-image and the reality o f what is just and unjust, good and bad, must truly be grounds for reproach even if the crowd praises it with one voice. {Phaedrus 277 d,e)

It scarcely needs to be pointed out that Socrates himself committed nothing to writing, but devoted himself entirely to viva voce teaching. In this sense, he is one of the last great representatives of orality in the ancient world. (Some later philosophers were also reluctant to write, but whether this was on principle it is difficult to say.) What Socrates argues can thus be seen as a spirited defence not only of his own practice but of a view of education which was doomed not to survive in the West. How can Socrates’ antipathy to writing be reconciled with the fact that his disciple Plato left a considerable corpus of written texts, which have

HI S T OR Y A ND THE LI TERATE REVOL UTI ON

55

continued to be read and studied down to the present day? There is certainly an irony in the fact that the teachings of Socrates, the champion of orality, are known to us only in written form. It is difficult to resist the conclusion that Plato, by the time he died, had seen that resistance to writing was a losing battle. He therefore undertook the task of preserving Socratic wisdom for posterity in that written form which approximated most closely to the actual Socratic m ethod; namely, the dialogue. But eventually even Plato surrendered to the power of the new medium, virtually abandoned the dialogue form, and conceded that laws must not only be set down in writing but accompanied by an explanatory commentary: The real job of the legislator is not only to write his laws, but to blend into them an explanation of what he regards as respectable and what he does not, and the perfect citizen must be bound by these standards no less than by those backed by legal sanctions. (Laws 823a)

Earlier in the same text, as Ryle points out, we have already had the recommendation that any ‘unwritten discourses’ that may be of value for educational purposes should be preserved in writing (Laws 81 le; Ryle 1966: 26). The answer to any sceptic who doubts whether there ever was a ‘literate revolution’ in antiquity lies here, in the life and work of Plato, who knew all the arguments against writing but nevertheless capitulated to fo r ce majeure. (A parallel in our own day would be the futility of mounting a philosophical campaign for the abolition of television.) Plato, although an ‘idealist’, was also a philosopher pragmatic enough to realize that history was not on Socrates’ side. It does not need a great deal of imagination to see where the institutionalization of history fits into the changed intellectual scene brought about by the literate revolution. The writings of Herodotus and Thucydides —as it happens —emerge as exemplars of the first ‘serious’ texts to achieve emanicipation from verse (where verse is recognized as the perpetuation of a basically oral form of expression). History is thus placed in the forefront of the new communicational world that Plato so deeply mistrusted. The debate between Classicists who accept and Classicists who reject the ‘alphabetic mind’ hypothesis has been in large part a debate over the wrong issues. Or, to be more exact, it has been fought out on the wrong terrain. (How many Greeks were literate? How literate were they? For what purposes did they employ writing? Or different forms of writing? To be sure, it is relevant to distinguish between records that indicate

56

THE L I N G U I S T I C S OF HI S TORY

ownership, business accounts, contracts, lists of rulers, and so on. The focus on such questions is an attempt to keep the debate within the bounds of traditional ‘scholarship’, because these are questions to which scholarly research into surviving documents can attempt to supply answers, however incomplete.) But behind all this lies a much more basic question, which leaves traditional scholars either stranded or already partisan. The question is: ‘What makes some statements true and others false?’. Both Socrates and Plato realized this, although it had almost dropped out of sight by the time the professional Greek grammarians assumed guardianship of the entrance to education. The establishment of grammar as an ‘art’ (techne) in its own right, distinct from logic on the one hand and rhetoric on the other, belongs to a later phase in the literate revolution. But from the beginning, as the etymology of the word gramm ar itself indicates (Greek gram m a, ‘letter’), the teaching of grammar presupposed literacy in its teachers and hence literacy as a social fact of life. It was part of a pedagogic programme. Its nucleus was a phonetic taxonomy of the letters of the alphabet and a doctrine of the ‘parts of speech’, which, once definitively established, continued more or less unchanged for more than two thousand years. The early expansion of the ‘parts of speech’ system and its terminology can be traced from an original binary distinction between onoma and rhema (Robins 1997: 42). The details can be disputed. But what was to become the standard set of parts of speech already appears in something close to its canonical form in the earliest extant Greek grammar, that of Dionysius Thrax (Lallot 1989). It survived the Middle Ages and the Renaissance to provide the foundation of the famous Port-Royal grammar in the seventeenth century (Arnauld and Lancelot 1660). It was still going strong when, in the early years of the twentieth century, English schoolchildren were taught grammar from manuals such as E. A. Abbott’s How to Tell the Parts o f Speech (1912). Abbott’s definition of the noun as ‘the name of anything’ is the sedimentation of centuries of grammatical doctrine. His first grammar lesson for the pupils in his class consists simply of three instructions: 1. Tell me the names of some persons; such as John, Mary. 2. Tell me the names of some places; such as London, Middlesex, England. 3. Tell me the names of some things that you can see, feel, hear, or smell; such as apples, soldiers, cat, sky, air, thunder, gas. (Abbott 1912: 15)

The pupils then learn that these names, already familiar to them, together with many more like those, are ‘nouns’. In brief, the technique being

HI S T OR Y AND THE LI TERATE REVOL UTI ON

57

taught is how to group certain words together under a certain meta­ linguistic label, on the basis of what they are names o f in the everyday world. Until the pupil can grasp this relationship between the name and what is named, and classify a random selection of words accordingly, there is no question of proceeding to tackle more difficult cases, where words fulfil more complicated tasks. Understanding these more compli­ cated tasks and the analysis of word combinations presupposes under­ standing the naming function of single words likz jo h n , M ary, apples, etc. (Thus Abbott’s definition of the pronoun, ‘a word used for a noun’, makes little sense unless we already understand what a noun is, and in what sense another word could be a substitute for it.) For centuries before Abbott wrote, the primordial status of naming had been guaranteed by the Biblical account of Adam’s naming of the animals, which Christian grammarians took as a veridical account of the origin of (human) language. And out of the ground the Lord God formed every beast of the field, and every fowl of the air; and brought them unto Adam to see what he would call them: and whatsoever Adam called every living creature, that was the name thereof. And Adam gave names to all cattle, and to the fowl of the air, and to every beast of the field. (Genesis 2: 19-20)

These verses set out the founding charter for semantics in the Christian era; and that foundation was regarded not as a myth but as a historical event (albeit revealed by divine dispensation). It actually happened. Furthermore, in the Biblical scenario there is absolutely no doubt about the relationship between the names and what is named. The creatures, created by God, already exist before they are named. They are brought to Adam. The name is a superimposition, devised by Adam. The naming does not affect the creature in any way, but merely provides a means by which its existence may be brought within the compass of human speech. In short, the meaning of such a name is guaranteed solely and simply by its correlation with the antecedently given creature. This is the semantics of the language myth at its most basic, and the Biblical version is un­ compromisingly reocentric. First the creature: then the name. There is no suggestion that Adam sometimes gave different species of animal the same name; or that God sometimes caught him out by presenting the same animal twice. On the contrary, the unstated assumption is that the Adamic nomenclature matched the different species. Here we have the notion of one-to-one correspondence between language and reality in its most primitive form.

58

THE L I N G U I S T I C S OF HI S TOR Y

At various times in the Western tradition there was considerable dispute about exactly how many parts of speech there were. Divine authority had even been invoked to resolve the issue (Michael 1970: 51). But a feature common to all the versions that grammarians proposed from time to time was the notion that at least some words were names of people (Socrates), animals (horse) and things (tree), while others designated actions (run), properties (small), etc. In short, the basic idea was that parts of speech could be distinguished according to how words related to the world, even though there might be different views about which words fell into which categories, particularly when the form of the word was also taken into account. Nevertheless, the longevity of some of the traditional definitions of the parts of speech says a great deal about the linguistic thinking that underlay them. The assumption was, evidently, that words exist to facilitate commu­ nication about what is going on in the world. Human beings need to ask questions about the world, make statements about it, give instructions relating to it, and so on. Nothing, then, could be more useful than to have words that directly correspond to items and natural categories of the world itself, or at least to those parts and features of the world that human beings had most frequent occasion to talk about. The proper name is a clear example: if we wish to say something about Socrates, it helps to know that his name is Socrates. It helps to have general words for ‘tali’, ‘short’, ‘bearded’, etc. if we wish to say something about his physical appearance; and general words for mental abilities if we wish to say something about his prowess in philosophical debate. What is required in order to say something about Socrates will also be required mutatis mutandis for saying something about Plato, or the Acropolis, or anything else we might select as a topic for comment. Pursuing this line of thought leads gradually but inevitably to the conclusion that the language best suited for human purposes will be one in which, so far as is possible, there will be an isomorphism between the structure of the world and the structure of our discourse about the world. Thus, for instance, a simple sentence like Sheep eat grass can be seen as comprising three parts of speech, each of which corresponds to an element in the real world. Sheep is the name of a certain kind of four-legged animal. Grass is the name of a certain kind of vegetation. Eat is a verb designating the activity of ingesting food. What could be more natural than to see the combination Sheep + eat + grass as an obvious way (although not the only possible way) of capturing an observable feature of the world, viz. that the world contains certain animals (those called sheep) who feed on (eat) a certain kind of vegetation (called grass)} What more

HI S T OR Y AND THE LI TERATE REV OL UTI ON

59

could be demanded of such a sentence? Its credentials as a way of stating something about reality are impeccable. For sheep exist, and grass exists, and sheep do indeed eat grass. Given that human beings seem for the most part to manage fairly well in verbal communication about the world, its sheep, its grass, etc., their success in these communicational enterprises leads on to the further supposition that an isomorphism must, at some level, actually obtain between their language and the world. Plato’s theory of Forms is an early, and in some respects rather crude, mystical attempt to explain how communication works on this basis: it assumes that behind each general term in the language there lies the invisible, permanent Form which anchors its meaning for all speakers and hearers. Aristotle adopted what may be regarded as a more secular, down-toearth version of isomorphism. He rejected the belief that the Forms exist independently in some realm of their own beyond the everyday world of our acquaintance; but he substituted for this belief the assumption that the human mind stores images or impressions of what there is in the world, and that these are identical for everyone (De Interpretatione 16a). In brief, he held that the world, as observed, is the same for all observers, who deploy the same mental categorizations in analysing and under­ standing it. Any grammarian making a similar assumption will be led to see the parts of speech as no more than verbal classes reflecting general concepts. That, precisely, was the conclusion reached by one of the most celebrated Roman grammarians. ‘What else is a part of speech,’ asks Priscian, ‘but a word indicating a concept, that is a thought?’ (Michael 1970: 49—50). From Aristotle’s belief in the uniformity of human understanding of the world it is only a short step to assuming that different languages merely provide different words for what is ultimately the same reality, regardless of which language you happen to have been brought up to speak. (This, we may note in passing, also provides a theoretical explanation of how it is possible to learn a foreign language. For foreign languages are simply alternative sets of vocal labels for the same features of reality. It also solves the telementation problem, or rather dispels it. For there is no such problem if it is true that all members of a linguistic community start out with the same conceptions of reality and the same verbal equipment for expressing them.) This view of meaning simplifies the historian’s task enormously. For if human beings all see the world alike, that rules out the awkward possibility that our ancestors saw it differently from us. And if words are just names of items in the world as perceived, then a truthful report of

60

THE L I N G U I S T I C S OF HI S TORY

what happened in the past is just a matter of selecting the words that correspond in the historian’s language to the words our ancestors would have used if describing those same events. The historian’s account thus becomes a post-dated substitute for eye-witness testimony. And what better account could one ask for, given those linguistic assumptions? To be sure, there may be conflicting eye-witness testimonies. But if so, who better than the impartial historian, looking back from a safe distance, to reconcile such conflicts or select one version as most probable? An apologia of roughly this form (but with the underlying semantics unstated) has survived in the Western classroom down to the present day. An Aristotelian linguistics is one in which many historians still put their unquestioning trust. Why? Because it offers them in advance all the guarantees that cannot be extracted from the past itself. Does all this have anything to do with the alphabet? Would it not apply equally if Greek writing had not been alphabetic, but syllabic or logographic? Less sceptical about writing than Plato, Aristotle assumed the written word to be a valid, albeit conventional, device for representing visually (but arbitrarily) the sounds of the corresponding spoken word. This assumption was also accepted by many Greek and Roman authorities on language. The proof is that at least some of them felt they had to address the question of whether the letters of the alphabet were adequate to capture pronunciation accurately. Quintilian, for example, expresses doubts about whether the Roman alphabet has enough letters to do this (Institutio Oratoria I.iv.7ff.), but seems to think that any such deficiency can be made good by appropriate alphabetic reforms. In the light of these assumptions, I would like to propose a rather different account from Havelock’s of ‘the alphabetic mind’ that emerged in Graeco-Roman antiquity. It was a mind, in my view, inclined to accept - or at least disinclined to challenge - the notion that, just as letters of the alphabet are (or ideally should be) isomorphic with basic sounds of spoken discourse, so words of a language are (or ideally should be) isomorphic with basic units of non-verbal reality. In both cases, it is correspondence between two series of units that in the end makes it possible for units of one series to represent or substitute for units of the other, and thus make human discourse possible. But this crucial relationship depends on the fact that human beings are able to invent or construct series of units (such as alphabetic letters) and put them in one-to-one correspondence with other series as required. Seen in this light, letters are simply surrogates for sounds. For any given sequence of sounds in speech, one can substitute a

HI S T OR Y A ND THE LI TERATE REV OL UTI ON

61

corresponding sequence of letters. In short, the alphabetic mind tacitly accepts the alphabet as a surrogate system of signs. Whatever is said and heard can thus be ‘represented’ (= recorded) by a visual surrogate in the form of a series of marks. The surrogate survives even when the original has ceased to exist. It thus prolongs the life of the original, just as the statue survives long after the statesman is dead. The transition from oral society to literate society in the ancient world cannot be understood until this fundamental relationship is grasped. The alphabet comes to be construed as a m odel for other forms of understanding. The alphabetic mind instals this surrogational model at the centre of its understanding of the way language and reality interrelate. But not only language and reality. The notion of the alphabet as a cognitive model can be taken further than this. And it was taken further in ancient Greece, by none other than Socrates himself (at least, according to Plato). In one of Plato’s last works, Philebus, Socrates offers the alphabet as a model for solving the most fundamental riddle in the process of understanding. This is the riddle posed by the fact that ‘whatever is said to be consists of one and many, having in its nature limit and unlimited­ ness’ (Philebus 16d). We must not, according to Socrates, ‘grant the form of the unlimited to the plurality before we know the exact number of every plurality that lies between the unlimited and the one’ (Philebus 16d). Protarchus then protests that he is not quite sure what Socrates means. Socrates replies: SOCRATES: What I mean is clear in the case of letters, and you should take your clue from them, since they were part of your own education. PROTARCHUS: How so? SOCRATES: The sound that comes out of the mouth is one for each and every one of us, but then it is also unlimited in number. PROTARCHUS: No doubt. SOCRATES: Neither of these two facts alone yet makes us knowledgeable, neither that we know its unlimitedness nor its unity. But if we know how many kinds of vocal sounds there are and what their nature is, that makes every one of us literate. {Philebus 17 a,b)

It is worth interposing a comment at this point on the modern notion of ‘literacy’. Much has been written in the past quarter of a century by Classical scholars about literacy in Greece and Rome; but a considerable amount of what has been written is vitiated by the tacit adoption of a concept of literacy that derives from modern systems of education, and in particular from the political goal of teaching as high a percentage of the population as possible to be proficient in the basic ‘three Rs’. Nothing

62

THE L I N G U I S T I C S OF HI S TORY

remotely like this anachronistic conception of literacy obtained in anti­ quity, as Socrates’ reply to Protarchus makes clear. Literacy, for Socrates’ generation, is not to be equated with the skills of reading and writing: those are merely consequences or byproducts of literacy. (It is not even clear that, for Socrates, you have to be literate in order to read and write at least up to a certain point. You could ‘pick up’ the practice by imitation without understanding it; just as, we are told elsewhere in Plato, the slave doctor can ‘pick up’ procedures for treating illness without really under­ standing medicine.) Literacy, rather, is for Socrates a matter of under­ standing the basis on which the letters of the alphabet correlate surrogationally with the sounds of speech. This point becomes clear later in the same dialogue when Socrates lectures the dimwitted Protarchus on the subject: SOCRATES: Just as someone who has got hold of some unity or other should not, as we were saying, immediately look for the unlimited kind but first look for some number, so the same holds for the reverse case. For if he is forced to start out with the unlimited, then he should not head straight for the one, but should in each case grasp some number that determines every plurality whatever, and from all of those finally reach the one. Let us again make use of letters to explain what this means. PROTARCHUS: In what way? SOCRATES: The way some god or god-inspired man discovered that vocal sound is unlimited, as tradition in Egypt claims for a certain deity called Theuth. He was the first to discover that the vowels in that unlimited variety are not one but several, and again that there are others that are not voiced, but make some kind of noise, and that they, too, have a number. As a third kind of letters he established the ones we now call mute. After this he further subdivided the ones without sound or mutes down to every single unit. In the same fashion he also dealt with the vowels and the intermediates, until he had found out the number for each one of them, and then he gave all of them together the name ‘letter’. And as he realized that none of us could gain any knowledge of a single one of them, taken by itself without understanding them all, he considered that the one link that somehow unifies them all and called it the art of literacy. (Philebus 18a-d)

It is perhaps superfluous to point out that Socrates’ attribution of this feat to Theuth is totally anachronistic and supported by nothing in Egyptian (hieroglyphic) writing. This underlines the meretricious ease with which the Greeks tended to equate the invention of writing with the invention of the alphabet. But that equation itself is characteristic of the ‘alphabetic mind’. More important still: here we see Socrates, the champion of orality, apparently already in thrall to ‘alphabetic’ thinking about the nature of speech. #

#

#

HI S TOR Y AND THE LI TERATE REVOL UTI ON

63

As Havelock was well aware, the emergence of the alphabetic mind would have had profound implications for the distinction between history and poetry. Havelock goes so far as to say that, contrary to traditional opinion, and in spite of the received Greek view of the matter, Herodotus was not ‘the father of history’. ‘The true parent of history was not any one “writer” like Herodotus, but the alphabet itself (Havelock 1982: 23). What are we to make of this? A generation before Havelock, Goody et al.y a similar claim had been lodged by H. G. Wells, not on behalf of the alphabet but on behalf of writing in general. The difference is important. For the Greeks, with their alphabetic script, were by no means the first people to leave written records of the past. In his Short H istory o f the World, Wells wrote: The invention of writing was of very great importance in the development of human societies. It put agreements, laws, commandments on record. It made the growth of states larger than the old city states possible. It made a continuous historical consciousness possible. (Wells 1946: 49)

It is this last assertion that demands attention. Here there is no question of the alphabet, which seems to have been a rather late development in the evolution of writing systems. So if Wells is right, then, pace Havelock, the fact that Greek writing happened to be alpha­ betic is merely a coincidental fact when it comes to elucidating the Greek notion of history. Any writing system, alphabetic or not, is in principle capable of making ‘a continuous historical consciousness possible’. In other words the alphabetic mind, it seems, is only one variety of the more general scribal mind. So the question narrows down. Does alphabetic writing sponsor some particular form of historical consciousness? Or does the alphabet per se have nothing to do with the way a society becomes aware of its past through writing? In trying to formulate satisfactory answers to these questions, a great deal depends on what we take the alphabet to be. And this was often dodged both by supporters and by critics of the alphabetic hypothesis. It is another instance of how their academic battle was fought on the wrong field and with the wrong weapons. For Havelock, it is clear, the alphabet has something to do with the representation of sounds. But what exactly is less clear. The original system of letters was borrowed by the Greeks from the Phoenicians; but the Greeks then elaborated it to suit their own purposes. Havelock presents what has been described as a ‘romantic’ view of what happened.

64

THE L I N G U I S T I C S OF HI S TOR Y

According to the romantic view, the Greeks took a foreign set of letters, which was not a true alphabet, and ‘made’ it into an alphabet by adding letters for vowels. They, not the Phoenicians, were the analytic geniuses. Atomism and the alphabet alike were theoretic constructs, manifestations of a capacity for abstract analysis, an ability to translate objects of perception into mental entities, which seems to have been one of the hallmarks of the way the Greek mind worked. (Havelock 1982: 82)

An essential feature of this achievement, in Havelock’s view, was that it made possible the transcription of languages other than Greek: in fact, according to Havelock’s enthusiastic account, ‘the new system could identify the phonemes of any language with accuracy’ (Havelock 1982: 85). It is at this point that Havelock’s story begins to lack credibility; for ‘phoneme’, in its modern linguistic sense (where it is not to be confused with a phonetic unit: Crystal 1997: 287-8), was not a concept that Greek grammarians or philosophers showed any awareness of at all. Attributing the discovery of the phonemic principle to the Greeks on the evidence afforded by Greek alphabetic writing is like crediting the inventor of the kettle with discovering the principle of the internal combustion engine. If the Greeks were unaware of the phonemic principle they could hardly have made it the basis of the Greek alphabet. Can they at least be credited, as has sometimes been suggested, with achieving an ‘unconscious phonemic analysis’ (Robins 1997: 17)? But what kind of cognitive feat is an ‘unconscious phonemic analysis’? It hardly helps the romantic case to say that the Greeks hit upon the phonemic principle without realizing it. That will not do for at least two reasons. First, it contradicts the romantic story that the alphabet is a work of analytic genius. Second, the phonemic principle is not the kind of thing it is possible to hit upon without realizing it, any more than it is possible to grasp the Newtonian law of gravitation without realizing that you have done so. That is precisely why we attribute a grasp of the law of gravitation to Newton, and not to his many predecessors in human history who had been hit on the head by falling apples. When we strip away the anachronistic talk about ‘phonemes’ and ‘unconscious phonemic analysis’, what is left? We are left with the far less romantic story that the Greeks devised a system of alphabetic writing in which the individual letters were taken to correspond to indivisible sound segments in speech. This is already evident from the discussion in Plato’s Cratylus, where the same terms (gramma, stoikheion) are used indifferently to refer both to letters and to the corresponding sounds. This is far from

HI S TOR Y AND THE LI TERATE R EVOL UTI ON

65

being the work of any analytic genius. What it amounts to is a simplistic retrojection of the writing system on to the sound system - the reverse of what the romantic account likes to suppose. The Greeks never got beyond this in their thinking about speech and writing. Once we realize that limitation, it throws a quite different light on how the Greek ‘alphabetic mind’ conceptualized history. (And this must not be confused with the quite separate question of how the first Greek historians viewed their task as writers.) The crucial point is that if you believe (either rightly or wrongly) that letters stand surrogationally to sounds as numerals do to numbers, you have no option but to accept that a correct transcription of what is said functions like a correct calculation of what is measured. If the calculation was correct yesterday, it is correct today. Similarly, if the transcription was correct yesterday, it is correct today. Only if some relevant change has occurred in the interim are yesterday’s figures to be discarded by the mathematician or the surveyor. But in the case of speech there is ex hypothesi no relevant change to take into account. For what was said is over and done with, however much the speaker may deny today what he said yesterday. In fact, having in writing what he said yesterday is the surest guarantee that he will not be tempted to try to deny it today. In short, two truths are to be distinguished in the case of the written testimony. (1) The truth about the words X uttered. (2) The truth about what X said. This is precisely the dichotomy that exercised Plato. Alphabetic writing merely makes it conspicuous in a way that nonalphabetic scripts do not. In other scripts that distinction is blurred, simply because there is no prima fa cie correspondence between scriptorial signs and individual sounds. What are the implications for the historian? What did not happen in the Greek case was any immediate devaluation of oral testimony. (As has often been pointed out, Herodotus and Thucydides relied heavily on what informants told them.) What did happen was something more subtle, but in the long term far more influential. The alphabet introduced a previously unknown standard of accuracy in the recording of one aspect of human affairs; but it was - for the Greeks - a very crucial aspect, since it involved precisely that manifestation of logos which made human affairs human: language. Thucydides, for one, clearly recognizes the difference between the truth about the words uttered and the truth about what was said; for he draws our attention to it and opts for the latter fa u te de mieux. But in that apology there is already the implicit admission that it would have made better history if he could have reported the actual words uttered. Thucydides is manifestly aware of his role as history-maker.

66

THE L I N G U I S T I C S OF HI S TOR Y

In order to grasp the shift of criteria involved here, it must be borne in mind that in a world where photography is unknown, the only kind of pictorial record comparable with writing would be an on-the-spot artist’s drawing of what he saw happening before his eyes. The evidential value of a written record of speech depends on the assumption that what was written down does indeed accurately and fully record what was spoken. But, given a surrogational theory of the alphabet, there is in principle no reason why that accuracy cannot be achieved. Whereas, on the contrary, it is always open to question whether the pictorial artist has recorded everything there was to be seen. No two artists will draw exactly the same sketch of an event, whereas two scribes can produce identical transcriptions of a speech. This is a vital point in assessing the status that writing has for the alphabetic mind. It sets standards that cannot be approached in any other domain of representation. Alphabetic writing is thus conceived of as an exact technology. It possesses a systematicity that is intrinsically superior in its own sphere to that of the artist’s impression; for the latter, based on sight not on sound, has to cope with more dimensions and has no fixed units to structure it. The artist’s impression is, as a much later generation would say, ‘analog’ not ‘digital’. The Greeks had no corresponding distinction, but if they had had, they would undoubtedly have recognized the superiority of the digital. That - not the ‘storage function’ of their script or its social utility —was the key to their literate revolution. Writing is important for a society’s view of the past because it introduces a new concept of historical evidence. In a pre-literate society, witnesses cannot be resurrected from the grave in order to testify: in a literate society, they can. Citation is resurrection. That, I think, is what Wells had in mind when he asserted that writing makes a continuous historical consciousness possible. There is no longer any complete disjunction between present and past. The limitations of human memory are over­ come. Once writing is accepted as a reliable form of communication for other social purposes, there is no reason why the historian should not accept it too. That is why, before leaving the topic of Greek literacy, I would like to emphasize again how remarkable and how devastating Plato’s scepticism about writing is for any theory of history which takes it for granted that the historian is a writer. It makes no difference whether that historian’s sources are oral or written. If the sources in question are written, then the historical account is a double falsification, since what was falsified in the first place by writing it down is now itself subject to a second process of

HI S T O R Y AND THE LI TERATE REVOL UTI ON

67

the same order. If, on the other hand, the sources are oral, then the historical account is still a falsification, in virtue of the fact that what was said becomes decontextualized by the process of writing itself. So history that is written history cannot in either case escape the implications of Plato’s condemnation. The point is all the more worth emphasizing inasmuch as it is some­ times argued that Herodotus founded ‘scientific’ history by himself setting about reducing diverse oral testimony to a single account, which he then wrote down. Thus, according to Collingwood, by a refined technique of ‘cross-questioning’ informants, Herodotus managed to ‘attain knowledge in a field where Greeks had thought it impossible’ (Collingwood 1946: 28). A less euphoric description would be, perhaps, that Herodotus managed to deceive both himself and the Greeks into believing that the decontextualization of oral testimony could, if pursued with persistence, yield a higher-order truth amenable to definitive statement in written form. From the linguistic perspective I am adopting, the latter description would be nearer the mark. History, as it came to be institutionalized in Europe, embraced all the question-begging involved in reducing the oral to the written; and not just adventitiously but necessarily so. To question writing would henceforward be to subvert the whole tradition from Herodotus onwards.

CHAPTER THREE

History as palimpsest

To summarize so far: there is a case to be made out for saying that in Europe the institutionalization of history was a byproduct of writing. Literacy holds the key to the linguistics of history. To some readers this may seem to be rushing noisily through an open door. That is obvious, they will say. If Greek society had never undergone the literate revo­ lution, there would be no such historical tradition as we now trace back to Herodotus, Thucydides, et al. I agree, it is obvious. Yet its being obvious has not led to asking some equally obvious questions that follow from it. The question I now wish to pursue is how, given the linguistic basis of that institution, historians found themselves encouraged to envisage their task in a certain way and also obliged to confront certain problems. These problems did not alter in any major respect for the next two thousand years and more. This is not to say that historians all had similar aims, outlooks and methods; far from it. But the criteria by which their work could be recognized as historical remained constant. That is why Gibbon and Macaulay could (legitimately) regard themselves as Herodotus’ successors. They were workers still working, as it were, for the same institution. The basic terms of their contract of employment were (1) that the historian produce a written prose text of some length in a consecutive and dignified style, and (2) that the text focus on presenting a truthful account of past events. Both of these are linguistic requirements, but only the first is also a literary requirement. The interpretation of the second depends, in the end, on the view the historian takes of the relationship between the world of events and the language available to describe it. The view taken by Aristotle and his generation was based on what I earlier called ‘reocentric semantics’. This, at least, is the sense in which Aristotle explains the function of words in the introduction to his Sophistical Refutations.

68

HI S TORY AS P A L I M P S E S T

69

It is impossible in a discussion to bring in the actual things discussed: we use their names as symbols instead of them; and we suppose that what follows in the names, follows in the things as well, just as people who calculate suppose in regard to their counters. (Sophistical Refutations 165a6—10)

Aristotle’s comparison was taken up in the seventeenth century in Hobbes’ much-quoted observation that ‘words are wise mens counters, they do but reckon by them, but they are the mony of fooles’ {Leviathan I.iv). What Aristotle says about bringing the actual things we are discussing into a discussion points to what he clearly thinks of as one of the most important surrogational functions of words: namely, that the verbal surrogates make it possible to talk of things in their absence. In order to say something about Socrates, it is not necessary for Socrates to be present, because the name Socrates ‘stands for’ him. If we say ‘Socrates is clever’, we are using his name; but we are saying something about Socrates, not about his name. And if for the name S ocrates we substitute the name P lato, we are now saying that Plato is clever. This is how ‘what follows in the names, follows in the things as well’. This point is of quite basic importance for the historian, who is typically in the situation of writing about individuals who are perman­ ently absent: that is, dead. If words were not able to function as reocentric surrogates in the way Aristotle describes, the historian’s task might seem to be impossible. Here we can put our finger on the crucial feature that makes such a philosophy of language so attractive to the traditional historian, at least on first inspection. Although they had no dictionaries, and although they did not think that every single thing in the universe - or even every class of things - had a name of its own, the Greeks did not doubt that the vocabulary of the Greek language provided enough words to describe and discuss both the world of nature and the world of human affairs. Aristotle leaves us in no doubt about his reocentric definition of truth: ‘to say of what is that it is, and of what is not that it is not’ {Metaphysics 101 lb26—7). Transpose this into the past tense, and you have a working brief for the historian: ‘to say of what was that it was, and of what was not that it was not’. The Greek view of words was in practice doubly surrogational, in that they thought of written words as surrogates for spoken words, the latter in turn being surrogates for the things spoken of. This is very relevant to understanding Plato’s objections to writing, which are based on the contention that written words are not adequate surrogates for spoken words.

70

THE L I N G U I S T I C S OF HI S TORY

Aristotle’s definition of truth is a classic statement of what later philosophers called the ‘correspondence’ theory. It echoes what Socrates says in Plato’s Cratylus, where he persuades Hermogenes to agree that statements ‘that say of the things that are that they are, are true, while those that say of the things that are that they are not, are false’ (Cratylus 385b2). But where does a correspondence theory of truth come from? It cannot even get off the ground unless, from the start, the assumption is made that there is already established a fixed correlation between features of reality and their verbal surrogates. This is where surrogationism leans heavily on the support provided by the Western language myth. The reocentric surrogationist assumes that reality comes divided up in such a way that appropriate verbal surrogates can unambiguously be allocated to the individual parts and a linguistic ‘code’ thus constituted. One word for gold, another for copper, a third for lead, and so on. Otherwise, to say that a statement is false (‘This is gold’ when actually the metal in question is copper) would be no more than a vague expression of personal disagree­ ment with the speaker. According to correspondence theorists, falsehood is a much more serious matter than that. For either a metal is gold or it is not. The matter is not something to be decided by the speaker. Nor by the hearer. Not, at least, if they are interested in the truth. What truth requires, for the reocentric surrogationist, is matching ‘things as they are’ with the right words; the right words being those that ‘stand for’ precisely those things and the assertion that they are ‘as they are’. Thus if someone points to Socrates and asks ‘Who is that?’, all we have to do in order to give a truthful answer is reply ‘Socrates’; i.e. supplying the name of the individual in question. Nothing further is required, because according to the surrogational account of words, there is nothing fu rth er to require. By producing the right verbal surrogate, the name S ocrates, we have done all that was asked. Similarly, if asked ‘What colour is that?’ we speak the truth if we give the name of the colour. Should someone ask, ‘But what is it about the word Socrates that makes it a truthful reply to the question “Who is that?”?’, the sole and sufficient reocentric answer is that it is truthful because, as a matter of fact, that is the name of the person in question. And similarly that green (or blue, yello w , etc.) is the name of the colour we were asked to identify. Truth is no more than that (at least, in the simplest cases; but the more compli­ cated cases are in the end all based on the simple ones). That is why when, in Plato’s dialogue (Cratylus 383b), Cratylus makes the opening claim that Hermogenes’ name is not H ermogenes, that initial pronouncement is both startling and deliberately challenging. It appears to call in question the whole basis of reocentric semantics as generally

HI S TOR Y AS P A L I M P S E S T

71

accepted in Greek culture. For if Hermogenes’ name is not H ermogenes (as everyone - including Hermogenes - had supposed hitherto), then what about all other names? An earthquake tremor goes through the whole edifice of language. Since I have referred so often to Plato, perhaps I should say at this point that Gadamer’s influential account o f‘the emergence of the concept of language in the history of Western thought’ (Gadamer 1985: 366-97) seems to me unsound. It is based on a misreading of Plato’s Cratylus. This is not a side-track, but an issue very relevant to understanding how history became institutionalized in the Western tradition. According to Gadamer, ‘Greek philosophy more or less began with the insight that a word is only a name, i.e. that it does not represent true being.’ This he contrasts with a hypothetical earlier view in which ‘the intimate unity of word and object was so obvious that the name was considered to be part of the bearer of the name, if not, indeed, to substitute for him’ (Gadamer 1985: 366). So, if Gadamer is right, Greek philosophy begins with a rejection of reocentric semantics. If, on the other hand, I am right, Greek philosophy begins by embracing that doctrine (and then developing it into logic on the one hand and history - including ‘natural history’ - on the other). Gadamer finds evidence for his view in Plato. He writes: Plato’s intention seems quite clear to me - and this cannot be emphasised sufficiently in view of the way that the Cratylus is constantly misused in the treatment of the systematic problems of the philosophy of language: in this discussion of contemporary theories of language Plato wants to demonstrate that no objective truth (alethia ton onton) can be attained in language, in language’s claim to correctness (orthotes ton onomaton) and that without words (aneu ton onomaton) being must be known purely from itself (auta ex heauton). (Gadamer 1985: 368)

Nothing, it seems to me, could be further removed from ‘Plato’s intention’. On the contrary, Plato’s (and Socrates’) philosophical pro­ gramme has to be understood as seeking by every possible means to discredit the Sophistic view, associated in particular with Protagoras, that speaking the truth is an illusion. If Gadamer were right, then the whole Socratic method of pursuing the truth by means of question and answer would be an absurdity, because ‘no objective truth [. . .] can be attained in language’. The intelligent philosopher would just keep his mouth shut and observe the world around him. (Which is what, according to legend, Cratylus was eventually reduced to doing.) But that, manifestly, was not the credo of Plato’s Academy, which was nothing if not a talking-shop.

72

THE L I N G U I S T I C S OF HI S TORY

Gadamer’s gross decontextualization of Plato in turn obscures what Aristotle was trying to do when he assigned to the historian the task of describing ‘the thing that has been’ (Poetics 1451b). Here the historian is enrolled on the side of those who think that telling the truth is not an illusion. And the historian’s task is nothing other than to recount (in words) the truth about past events. When the historian sets out to give a true account of past events, what is being promised - according to the reocentric view - is an account in which there are no earthquake tremors of the Cratyline sort, because the right names and descriptive terms are correctly matched to the events in question. All this assumes that the language the historian is using has the names and other verbal equipment required. For if the right words are not available, there is no question of giving such an account. Truth, in short, makes its own linguistic demands on verbal surrogates. We can no more say of grass that it is green in a language that has no colour words than Plato could have expected his pupils to pay in US dollars. It will be useful to have a general term for any historians of any period who, explicitly or implicitly, accept such a view of language and truth. I shall call them ‘reocentric historians’. If Caesar never crossed the Rubicon, then a reocentric historian ought not to tell us that he did. Likewise, if he did cross the Rubicon, a reocentric historian ought not to tell us that he did not. The demands on the linguistic account are thus set by ‘the reality’ of Caesar and what he did: it is up to the historian to try to meet those demands verbally. Given Caesar’s career, there must be at least one correct description available for everything in it, a word for every kind of weapon that Caesar ever carried, for every piece of armour that Caesar ever wore, for every kind of food that Caesar ever ate, and so on: and the historian who adopts a correspondence theory of truth is implicitly committed to producing such descriptions. In this sense, the great majority of Western historians have been reocentric historians, whether they realized it or not, and even if they sometimes fell short of the reocentric ideal. For the historian, reocentric semantics thus poses a linguistic challenge (to find the words to capture the truth), but in compensation has one great advantage. It provides an easy answer to the crucial question: ‘How is it possible to write about the past in the language of the present day?’. The reocentric reply runs as follows. Since words get their meanings by standing for the various things, persons, places, actions, qualities, etc. that really exist, then provided that the world does not change, the same words will continue to have the same functions. Thus, for example, the word gold means the same today as it meant yesterday and a hundred years ago,

HI S TOR Y AS P A L I M P S E S T

73

because the metal gold is still the same metal and the colour gold is still the same colour. Similarly, in the domain of human affairs, such words as man, woman, child, hunger, thirst, love, hate, kill, and thousands more remain semantically stable because what they stand for does not alter from one generation to the next. Human affairs take the form of a continuum. Only, therefore, in those respects in which today’s world is not a simple continuation of the world of the past does it become necessary for the historian to seek any other terms than those provided by the language of the present day. Thus, for instance, there is no problem about saying today that a certain individual was born or died a thousand years ago, because birth and death are still landmarks in human lives that have not altered. But if the individual in question held an office long since abolished, or followed a trade no longer practised, then today’s language may have no words for that office or that trade. In such cases, the historian must either revive an old word and explain to readers what it once meant, or else resort to descriptive approximations. No such linguistic makeshifts, however, are required in cases where people are still doing today what their ancestors always did, entering into similar relations, facing familiar problems, experiencing like emotions and entertaining the same beliefs. How does reocentrism carry over from discourse about the material world to discourse involving moral judgments? Are these not subjec­ tive? Do not standards vary from age to age? Is it not anachronistic for a modern historian to describe the emperor Tiberius as a vicious tyrant? Not if we believe that human nature has remained essentially unaltered, and that our vocabulary of moral approval and disapproval still gets its meaning from correspondence with the same unchanging roster of human virtues and defects. For the reocentric surrogationist there is no problem about presenting this as a valid philosophical position. It is exactly the position adopted by no less a philosopher than Hume. The final sentence, it is probable, which pronounces characters and actions amiable or odious, praise-worthy or blameable; that which stamps on them the mark of honour or infamy, approbation or censure; that which renders morality an active principle and constitutes virtue our happiness, and vice our misery: it is probable, I say, that this final sentence depends on some internal sense or feeling, which nature has made universal in the whole species. (Hume 1777: 137)

Thus just as liquid and solid, black and white, hunger and thirst are the same for all humanity, so are the recognition of good and bad in human

74

THE L I N G U I S T I C S OF HI S TORY

conduct. And hence it is possible in all languages to find words desig­ nating the same moral assessments. The very nature of language guides us almost infallibly in forming a judgement of this nature; and as every tongue possesses one set of words which are taken in a good sense, and another in the opposite, the least acquaintance with the idiom suffices, without any reasoning, to direct us in collecting and arranging the estimable or the blameable qualities of men. (Hume 1777: 138)

This is reocentric semantics at its most intransigent. In effect, it gives the historian carte blanche not only to tell us what our ancestors did, but to pass judgment on their characters as well. As a linguist, I detect various difficulties with these reocentric assump­ tions, but I do not intend to pursue them here. What I do wish to draw attention to is the fact that these assumptions already have a built-in historical dimension. Thus, for example, although at first sight there appears to be no appeal to history in the reocentric definition telling us that the word gold means what it does in virtue of being the name of a particular metal with certain properties, that definition conceals a thesis about change over time. It is a thesis to the effect that the word gold has meant, does mean and will mean the same for as long as the correlation with the metal in question remains unaltered. So it is small wonder that historians over the ages have been so ready to accept a reocentric view of language: it already butters their bread for them. Furthermore, by adopting this view of language the historian is led almost inevitably to adopt a palimpsest conception of history for purposes of self-justification. I borrow the term palimpsest from Orwell, and by a palimpsest conception I mean the notion that, over the course of time, earlier accounts of the past are erased and new accounts written over them. The rationale for these substitutions has to be that the later accounts are more truthful, either because new evidence has come to light, or because old evidence has been better interpreted. Thus eviden ce, and the inferences drawn from it, becomes a primary focus of attention: for that is what backs up the reocentric historian’s claims to truth and thus distinguishes history from myth and fiction. Aristotle said nothing about the problem of historical evidence at all: it was left to his heirs to tackle. Emphasis on truth eventually comes to dominate definitions of history. It takes pride of place already in the second century ad in Lucian’s treatise on how to write history, which Donald Kelley summarizes in the following terms:

HI S TOR Y AS P A L I M P S E S T

75

The poet, on behalf of his art, commands his material; the orator may shape it to his argumentative ends; but the historian, out of respect for the facts, must follow it and worship only Truth. [. . .] In short - the maxim honored at least nominally from Lucian to Ranke and beyond - ‘the historian’s sole task is to tell the tale as it happened.’ The ‘noble dream’ of modern scientific historiography was thus already part of the idea of history embodied in the Greek canon [. . .]. (Kelley 1998 : 47)

That ideal was reaffirmed in the famous opening sentence of Voltaire’s article Histoire in the E ncyclopedic: ‘the account of facts presented as true ((dom es pour vrais); as opposed to fable (fable), which is the account of facts presented as false’. It is the same emphasis on truth that allows history to be regarded as science. For in science a similar palimpsest conception emerges: early and inadequate accounts of the natural world are gradually superseded by more accurate accounts, based on better evidence. A whole section of the E ncyclopedic article is devoted to histoire naturelle, which, according to Voltaire, is nothing other than ‘une partie essentielle de la Physique\ None of this would be remotely plausible without the essential underpinning provided by a reocentric philosophy of language, accepted by historians and scientists alike. Likewise, where history is seen as differing from the natural sciences, the difference often hinges on the fact that the evidence the historian is forced to rely on may be unverifiable reports from the past. This eventually led to a retreat from truth to credibility or probability as the historical desideratum. One of the landmark texts in this shift is Hume’s discussion of the limits on what he is prepared to believe about the reign of Queen Elizabeth. But suppose, that all the historians who treat of England, should agree, that, on the first of January 1600, Queen Elizabeth died; that both before and after her death she was seen by her physicians and the whole court, as is usual with persons of her rank; that her successor was acknowledged and proclaimed by the parlia­ ment; and that, after being interred a month, she again appeared, resumed the throne, and governed England for three years: I must confess that I should be surprised at the concurrence of so many odd circumstances, but should not have the least inclination to believe so miraculous an event. I should not doubt of her pretended death, and of those other public circumstances that followed it: I should only assert it to have been pretended, and that it neither was, nor possibly could be real. You would in vain object to me the difficulty, and almost impossibility of deceiving the world in an affair of such consequence; the wisdom and solid judgement of that renowned queen; with the little or no advantage which she could reap from so poor an artifice: All this might astonish me; but I would still reply, that the knavery and folly of men are such common phenomena, that I should rather believe the most extraordinary events to arise from their concurrence, than admit of so signal a violation of the laws of nature. (Hume 1777: X.ii.99)

76

THE L I N G U I S T I C S OF HI S TORY

It is interesting to note that underneath the case that Hume sets out there is a concealed linguistic argument which runs likes this. Queen Elizabeth is the name of an English monarch. English monarchs are human beings, not gods or superhumans. We do not apply such a name to any who can die and then come alive again. So the story of Queen Elizabeth’s death and resurrection is self-contradictory. Given that the name identifies the monarch in question, we already know that certain things could not have happened in her reign. We need not bother with alleged historical evidence, for we are dealing with semantic impos­ sibilities. This is the clinching argument, not the one that Hume ostensibly offers. For if it were merely a matter of Hume’s preference for belief in human ‘knavery and folly’ as the probable explanation of the testimony, that would throw equal doubt on all historical testimony. To erect socalled ‘laws of nature’ into semantic criteria of truth is a typically reocentric move. Another snag with truth that historians were slow to recognize is that truth is only a necessary but not a sufficient condition for the traditional historian’s tale. It is not enough that all the particulars of the story be accurate. The historian’s story must also be complete. Nothing relevant must be concealed or omitted. (For otherwise it would be possible to select a set of true particular statements which nevertheless gave in toto a misleading account of the course of events.) But what is ‘relevance’? And how is it to be objectively determined? Relevance is not a notion that is easy to relate to truth. It sits altogether uncomfortably alongside truth as construed in reocentric versions of the correspondence theory of truth. It seems to require an overview from a higher vantage-point; or else a pre­ emptive decision about the significance of the tale itself. Descartes raised a simple but devastating objection to the notion that a historian could ever tell ‘the whole truth’: And even the most accurate histories, while not altering or exaggerating the importance of matters to make them more worthy of being read, at any rate almost always omit the baser and less notable events; as a result, the other events appear in a false light [. . .]. (Descartes 1637: 1.114)

Thus, according to Descartes, even i f the historian’s statements are all true individually, and even i f nothing has been suppressed in order to conceal awkward facts, nevertheless the very process of selection and omission automatically affects the truth of the acccount. It is the question

HI S TOR Y AS P A L I M P S E S T

77

of relevance again. Did Caesar shave on the day he crossed the Rubicon? We are not told, presumably because there is no evidence. But even if there were, it would doubtless be dismissed as ‘irrelevant’ to the historian’s tale of the momentous event. But who can be sure? Perhaps Caesar was superstitious and would have retreated back to Gaul if the razor had drawn blood. The detail, although trivial, illustrates how difficult it is to determine relevance in particular cases. How much more difficult would it be for historians to come up with a gen eral theory of relevance. The point I wish to emphasize here is that the very notion that in history there is a ‘whole truth’ to be told is a surrogational notion. It can be traced back from Descartes at least as far as Cicero. In De Oratore II.xv.62, Antonius declares that the first law of history (prima lex historiae) is that the historian must tell nothing but the truth. Then he immediately adds that in addition the historian must withhold no truth. This places squarely on the historian the responsibility to discover the whole truth and tell it. Ignorance, indolence or turning a blind eye will not do. But this again presupposes that the whole truth is out there to be discovered, and that, once discovered, it can be reported without further ado. It is as if the past was a realm of items, each ready labelled, waiting for the historian to come along and read the label. (‘Ah! Here is what is called a “revolution”. This looks like a case of “regicide”. The “feudal system” over here.’ And so on.) Since truth and relevance seem to appeal to different criteria, one would expect to find that a recurrent problem for the reocentric historian arises over defining the boundaries of the factual. And that is precisely what we do find. The problematic move is typically the move from historical report to historical explanation. The historian who may appear to be on sound ground when simply describing and dating certain events (reading their labels) often plunges into quicksand when it comes to explaining them. It is possible to argue that here there are no ‘facts of the matter’, but only interpretations of facts, or speculations. When the frontier between report and explanation cannot be adequately policed, reocentric truth-claims are automatically called in question. The reocentric historian, as defined above, believes or simply takes it for granted that the relationship between past events and the language currently available to describe them is such as to give rise to no serious doubts about the adequacy of the language for the task in hand. Reassurance on this score is provided by reocentric semantics itself: it is assumed that language could not function communicationally if words

78

THE L I N G U I S T I C S OF HI S TORY

did not correlate to reality with recognizable fidelity, or if they acquired their meanings in some quite different way from ‘standing for’ things other than themselves. Given this assumption, reocentric historians are required to take responsibility for their language in a particular way. They undertake, implicitly, to use words ‘correctly’; that is, each word in accordance with whatever it may be that it stands for, and not otherwise. Qua historians, they cannot opt out of this linguistic responsibility, or decide to accept it only when it suits them; for upon it depends the professsional status of their account of the past. In effect, they enter into a communicational relationship with their readers which (as Aristotle saw) differs signifi­ cantly from that offered by poets, priests, politicians and many others engaged in public communication. It is in this sense that the historian’s view of language determines the historian’s view of history. If the terms of the linguistic engagement are not reocentric, that automatically affects the truth requirement. Once that requirement is removed or differently interpreted, the result is a different conception of history. Aristotle is far from being the only Western thinker to have realized this. The necessity of the interconnexion is brilliantly captured, albeit in fictional form, in Orwell’s Nineteen E ighty-Four, where a ruthless state apparatus is dedicated to expunging reocentric history. In the Ministry of Truth, the hero of the novel works in the Records Department. His job is to ‘rectify’ (i.e. rewrite) newspaper reports previously published. Day by day and almost minute by minute the past was brought up to date. In this way every prediction made by the Party could be shown by documentary evidence to have been correct, nor was any item of news, or any expression of opinion, which conflicted with the needs of the moment, ever allowed to remain on record. All history was a palimpsest, scraped clean and reinscribed exactly as often as was necessary. (Orwell 1949: 35)

The linguistic implications of this policy are far-reaching. The Records Department is in the business of closing the gap between fact and documentation. It does so by ensuring that all undesirable documentation is destroyed or rewritten. Once the gap is closed, the past becomes a projection from the present, and has no independent existence. It is no coincidence that the state which sets up a Ministry of Truth also plans a reform of English. The reformed language (Newspeak) will constitute a means of thought censorship by which politically incorrect ideas can no longer be verbally expressed at all. One programme is a logical extension of the other. The objective is to reverse the dependencies implicit in

HI S TOR Y AS P A L I M P S E S T

79

reocentric thinking and bring reality under the control of language, which is in turn controlled by the state. Those postmodernists who claim that the past is no more than a linguistic construct come straight out of Nineteen Eighty-Four. Orwell’s own thinking about language was pervaded by reocentric assumptions, as we see from essays such as ‘Politics and the English language’ (Orwell 1946). His invention of Newspeak is a surrogationist parable from beginning to end. Likewise, the palimpsest as a metaphor of history is surrogationist through and through. In the case of Newspeak, the twist is that the language, by exerting a mental stranglehold on its speakers, inexorably reshapes their understanding of reality. In the case of the palimpsest, erasing the original text is both concealing the truth and, at the same time, blocking access to knowledge. For the transmission of knowledge becomes equated in Western education with the transmission of text. In Orwell’s novel, the state does not wish to usher in the era of post-literacy that Wells envisaged in The Time M achine: far from it. But there would be no point in the state rewriting the texts from the past unless the texts represented both truth and knowledge. Thus in Nineteen EightyFour the themes of past, present, truth, literacy, freedom and knowledge interlock in a way that makes no sense unless a certain view of language is accepted without question. Ironically, Orwell’s land of Big Brother is, at one level, a critique of the extent to which surrogationist assumptions have corrupted the West’s basic conceptions of history and politics. In this Orwellian retrospect, it seems almost self-evident that the Greek view of language determines the Greek view of history, via a certain concept of truth. Why is this not acknowledged by modern historians? Partly because modern historians may be reluctant to admit the implica­ tions of this thesis for their own work; but also because the view that modern scholars take of Greek antiquity tends to focus on a more obvious aspect of the literate revolution. This concerns the identification of history-writing as a specific literary genre. Thus we find endless studies pointing out how the historikoi were conscious of and indebted to the rhetorical practices of their day. According to Kennedy’s analysis of Herodotus’ text, ‘the four characteristics of rhetorical consciousness are all to be found in Herodotus’ (Kennedy 1963: 45). The same scholar concludes that Thucydides ‘was quite aware of the theories of sophists and professional rhetoricians about oratory’ (Kennedy 1963: 48). So history-writing, it seems, was ab initio marked by conformity to patterns of literary production already established. This kind of identification is what Aristotle is already protesting against

80

THE L I N G U I S T I C S OF HI S TORY

when he claims that it would have made no difference if Herodotus had written in verse. He does not, it should be noted, make the symmetrical claim that Homer would still be a poet if the Iliad and the O dyssey had been prose compositions. That asymmetry reveals another side of Aris­ totelian linguistic thought. While Aristotle the logician recognizes that in the end what matters is that words must be unambiguously connected to reality, Aristotle the rhetorical and poetical theorist admits that how words are used to impress the audience matters no less. Exactly how these two functions of language are related was a problem Aristotle never solved. Nor did anyone else in antiquity. The ‘solution’ that eventually became enshrined in the curriculum of the medieval universities involved academic surgery. Logic and rhetoric were treated as quite separate disciplines, each with its own metalanguage. Here, as so often in the Western tradition, we see the partisan division of academic territory taking priority over any serious attempt to deal with issues that cross pedagogic boundaries. But where that leaves history is another question. How far back does the recognition of history as a distinct branch of expertise go? All we can say is that at some time between Homer and Hesiod, Clio was probably appointed its Muse. And the traditional emblem associated with Clio was the scroll. This is as good evidence as one is likely to find that the origin of history as a cultural super­ category was —or was conceived as being —intimately bound up with writing. In a footnote to his translation of Cicero’s O rator, H. M. Hubbell observes: ‘History was regularly regarded in antiquity as a branch of rhetoric, much to the disadvantage of history’ (Hubbell 1962: 333). The occasion of this remark is Cicero’s inclusion of ‘histories’ (historiae) as a type of composition. Later in the same treatise, Cicero describes the historia as requiring ‘a narrative in an ornate style, with here and there a description of a country or a battle’ (Orator xix, 66). He adds: ‘It has also occasional harangues and speeches of exhortation.’ It is interesting, although possibly embarrassing, to note how well this weathers as an account of the work of much later historians. But it would be doing Cicero less than justice to say that he fails to distinguish between the account and the events recounted. On the contrary, he distinguishes quite meticu­ lously between the former (historia) and the latter (res gestae). Thus Hubbell mistranslates when he renders Cicero’s sententious observation on our knowledge of the past as:

HI S TOR Y AS P A L I M P S E S T

81

To be ignorant of what occurred before you were born is to remain always a child. For what is the worth o f human life, unless it is woven into the life of our ancestors by the records of history? (Orator xxxiv, 120)

What Cicero is actually referring to is not histories (historiae) but awareness of the deeds of one’s ancestors (memoria rerum veterum). Here he recommends the orator to study the Liber Annalis by his friend Atticus, which is not a historia in the manner of Herodotus but a chronology covering the outstanding events of the past seven hundred years. This is not the place to enter into the learned debate which has exercised Classicists as to whether annalists counted as historians. Some authorities in antiquity seem to have drawn such a distinction. But, as one critic pertinently remarks, Modern scholars have given their own twist to this debate, and used the term ‘annalists’ for historians such as Valerius Antias and Claudius Quadrigarius, whom they consider to have been simple-minded, uncritical, and mendacious. Since their works do not survive, the quality of these historians is difficult to judge. In any event, the modern habit of disparaging particular historians by the use of the term ‘annalists’ has no basis in the ancient evidence, and should be avoided as unhelpful and misleading. (Cornell 1996: 99)

Laudable and judicious though these sentiments may be, the above passage itself exemplifies the complete muddle today’s historians have made of the terms of their trade. F or here the luckless Valerius Antias and Claudius Quadrigarius (neither with extant works to his credit) are already classified by the modern historian as ‘historians’, when the ostensible (historical) issue to be determined is whether they were historians at all. It would be difficult to cap this as an example of history shooting herself in the foot. It is not, however, an isolated example. An even more blatant one occurs in the following passage from Rosalind Thomas’s Oral Tradition and Written R ecord in Classical Athens: One can often judge the accuracy and changes in an oral tradition by comparing it with the very much more detailed accounts of Herodotus and Thucydides. This does not assume the historians always provide a reliable and historically accurate account, since they may restructure or transform the information they use. But in the cases I discuss, the account of Herodotus or Thucydides is the one generally accepted by historians as most reliable [. . .]. (Thomas 1989: 7)

The double-take one experiences on reading this paragraph is prompted by the quasi-Orwellian double standards shamelessly invoked in defence of historians. The unreliability of certain historians’ written

82

THE L I N G U I S T I C S OF HI S TORY

accounts is redeemed by the assumed but unproven reliability of others’. Thus all, miraculously, escape whipping. Plato would have had apoplexy had he survived to read such casuistry. The phrase historically accurate is particularly worthy of note. It could doubtless be argued that the historikoi of antiquity are themselves to blame for this unsatisfactory state of affairs. If so, their medieval successors fare no better. The impact of Christianity on philosophy of history was to qualify the emphasis on truth and to reinforce the importance of evidence. But, under the aegis of Christianity, evidence tends increasingly to be equated not with a simple-minded reocentrism (the word gold means ‘gold’, and anyone can see what gold is) but with reliable authority, and authority in turn to require approval by the Church. This did not, however, elevate history to the status of revelation. We see this in Bede’s Ecclesiastical H istory which begins with a long list of his impeccable sources, terminating with the caveat: And I humbly beseech the reader, that if he shall find anything set down otherwise than truth in this that I have written, he will not impute it unto us, as the which have endeavoured with all sincerity to put in writing to the instruction o f our after comers such things as we have gathered by common report, which is the true law of history {quod vera lex historiae est). (Historia Ecclesiastica Gentis Anglorum, Praefatio)

This is something of a retreat from the position stated centuries earlier by Antonius in De O ratore, where the prima lex historiae was said to be reporting the truth. I take this retreat to reflect a religious view according to which it is not given to mere mortals by their own investigations to know the truth about the past, except insofar as the Church or divine inspiration may authenticate it. That seems borne out if we contrast what Bede says about the utility of history with what Livy says. Both hold that history is a repository of lessons. The difference is that for Bede ‘knowing of things done or spoken by worthy men before your time’, or of evil deeds done by evil men, will move the reader ‘more earnestly to follow after the things he knoweth to be good and acceptable to God’. Accord­ ingly he exhorts King Ceolwulf to have this history ‘more widely published, both to the instruction of yourself and also to the edifying of such other whom the authority of God hath committed unto your governance’ {Historia Ecclesiastica Gentis Anglorum, Praefatio). But this is a circumlocutory way of saying that in order to read history aright, the reader needs to be a Christian. According to Donald Kelley:

HI S TOR Y AS P A L I M P S E S T

83

Medieval authors did have a recognizable idea of history, though it had more to do with rhetoric and the liberal arts than with higher forms of knowledge such as philosophy and theology, this last, of course, being the primary guide and control for all the other disciplines. In fact history did not appear among the seven liberal arts except as an auxiliary (appendentia artium, in the phrase of Hugh of St. Victor) or a genre related to the study of grammar [. . .]. In general, history was defined in commonplace terms stressing truth, utility, and, secondarily, aesthetic or literary value, but all of these classical qualities were reinterpreted in the context of Christian faith, piety, doctrine, and the different attitudes towards what was ‘worth remembering’ (digna memoria). (Kelley 1998:

100- 1)

Thus medieval writers repeat all the sacrosanct formulae about needing to preserve the ‘memory’ of the past. In other words, they dodge the question - just as Herodotus and Thucydides had done - of the difference between account and memories, even when they invoke it. In 1247 Matthew Paris was instructed by Henry III ‘to write an accurate and full account of all these things and commit them indelibly to notable writing in a book, so that their memory shall in no way be lost to posterity’ (Clanchy 1993: 101). But King Henry suggested no way of ensuring that the account be ‘accurate and full’, much less of guaranteeing that posterity would somehow know it was. Kelley observes: History is often regarded as a form of memory, but this nice identification creates more problems than it solves. We each of us are familiar with our own memories and those of others - until we forget - but what does ‘memory’ mean in a collective sense? And just where does this memory reside? In the remains and records of human activity or in the act of historiographical poesis that gives shape and meaning to these traces o f the past? In any case memory is selective; it is not a simple process of recording or retrieval but a way of commemorating or memor­ ializing certain phenomena within some system of meaning, which itself changes in time. (Kelley 1998: 11-12 )

Kelley here states a twentieth-century point of view and puts a twentieth-century question. But for medieval historians the problem of memory is invisible, because they are operating within a surrogational philosophy of language. The wording of the historical account itself embodies the ‘memory’ of the past, since words are direct surrogates of reality. This is what disguises, for the surrogational historian, the enormous gap between anyone’s memory and what actually happened. That gap is reduced by collapsing the difference between memory and tradition. The latter is often treated as a kind of collective ratification of the former. So

84

THE L I N G U I S T I C S OF HI S TORY

we end up none too sure whether the historian’s job is (1) to establish what happened, or (2) merely to report (accurately, if possible) what was traditionally believed to have happened (fama vulgante in Bede’s words, donne pour vrai, in the memorable formula of the E ncyclopedic), or (3) to try to ensure that posterity will be able to distinguish one from the other. Ideally, presumably, all three. Three layers of truth, in short, each at one metalinguistic remove from the next. But it was far from clear to medieval minds bemused with the idea that the highest form of truth was divine revelation, and obsessed with convictions of human fallibility, how all these desiderata could be captured simultaneously. Nor was it clear that any medieval historians were prepared to address the problems involved. Aquinas seems to have held that the certainty attached by Christians to stories of the Virgin Birth and the Resurrection was not the result of better historical evidence but an effect of divine grace (Burns 2000: 6). How did this predicament arise? Because historians adopted basically the same linguistic epistemology as their Greek predecessors, and con­ tinued to do so down to modern times. But in this acceptance, from at least Aquinas onwards, lay the long-dormant seeds of controversy. Bishop Usher’s renowned dating of the creation of the world marks the last desperate attempt to reconcile history with Christianity in its traditionally received written form. The discovery of civilizations earlier than Usher’s primordial terminus a quo meant that either history or theology had to yield. History refused to yield, so the Bible had to be linguistically reclassified as poetry. The alternative would have been to affirm the historicity of the Bible and reclassify secular history as tradition. But that alternative would have required a new linguistic epistemology that no one was in a position to advance before the emergence of twentieth-century structuralism. And by then it was too late. For most of the many centuries separating Livy from Gibbon, the only credible alternative to a plainly reocentric semantics was its twin psycho­ centric version. The difference between reocentric and psychocentric semantics does not entail altering the priority as between the word and what it stands for: the latter still takes precedence in any explanation of meaning. All that has changed is the account of ‘what it stands for’. It is important to note that adopting a reocentric semantics does not mean restricting the world of ‘things’ to material objects, even though such objects may often be taken as the paradigm examples of what a word stands for. Varro, for instance, gives a classically reocentric explanation of etymology. It hinges on the fact that every word is identified for Varro by its form and by two features: ‘from what thing and to what thing’ it is

HI S TOR Y AS P A L I M P S E S T

85

applied (a qua re et in qua re vocabulum sit impositum) (De Lingua Latina V.2). But these ‘things’ can be qualities or abstractions, as is evident from Varro’s first example. This is the abstract noun pertinacia (‘obstinacy’). If obstinacy is a ‘thing’ (res), it is certainly not a thing of the same kind as a rock or a tree. So it would be misleading to think of reocentric semantics as committed to a philosophical position which later thinkers would call ‘materialism’. The materialist philosophers of antiquity (Democritus, Epicurus, et al.) were certainly influential in their atomistic view of the universe, and insofar as Greek thinking about sounds and letters was manifestly atomistic, they may be thought to have made an impact there too. But these philosophers never, as far as I know, elaborated a linguistic doctrine to the effect that words were first or solely or ultimately ‘imposed’ on isolated aggregates of atoms. And if obstinacy can be a ‘thing’, there seems no reason why the realm of things should not include compliance, compromise, hopes, expectations, and even thoughts. William James once observed that ‘thoughts’ and ‘things’ are names for two different sorts of object ‘which common sense will always find contrasted and will always practically oppose to each other’ (James 1904: 1). So does not such a liberal extension of the realm of ‘things’ as Varro entertained inevitably blur any distinction between reocentric and psychocentric semantics? No. What matters is not whether ‘things’ include mental items of various kinds, but whether the meaning of a word is determined by the thing itself, or by some intermediary concept which links word and thing. Why, it might be asked, cannot these two accounts be combined, so that the meaning of the word is underwritten jointly both by the realworld existence of its correlate and also by corresponding ideas in the minds of those who use these words? One uncompromising answer was given in the nineteenth century by John Stuart Mill, to the effect that psychocentric semantics was plainly a mistake. According to Mill, it is clear that the word sun is ‘the name of the sun, and not the name of our idea of the sun.’ For names are not intended only to make the hearer conceive what we conceive, but also to inform him what we believe. Now, when I use a name for the purpose of expressing a belief, it is a belief concerning the thing itself, not concerning my idea of it. When I say, ‘the sun is the cause of day,’ I do not mean that my idea of the sun causes or excites in me the idea of day: or in other words, that thinking of the sun makes me think of day. I mean, that a certain physical fact, which is called the sun’s presence (and which, in the ultimate analysis, resolves itself into sensations, not ideas) causes another physical fact, which is called day. (Mill 1884: I.ii.l)

86

THE L I N G U I S T I C S OF HI S TORY

M ill’s answer is based on the telementational view of communication enshrined in the language myth. Whether we accept this or not, there is still a problem about trying to combine reocentric and psychocentric semantics. For people’s ideas about the sun may not only differ but conflict. In the end, one must either opt for the reocentric solution and discount erroneous beliefs, or for the psychocentric solution and ignore the validity of the beliefs. Aristotle came out in favour of a reocentric solution (which he needed in order to ground logic in reality). Many centuries later, Locke argued strongly in favour of the psychocentric alternative. On the whole, modern science is committed to reocentric semantics. (On this view, the meaning of the word sun is being continually redefined as scientists gradually discover more and more about the sun.) How about the past? In this perspective, one linguistic dilemma for the surrogational historian is whether to write a reocentric history or a psychocentric history. (Not that any traditional historian ever posed the problem in those terms.) The advantage of the former course is that it aligns history linguistically with the exact sciences. The advantage of the latter is that it gives the historian a powerful communicational role as maker and interpreter of tradition. Trying to have the best of both worlds, many distinguished historians succeeded only in falling between two stools. How this may come about is already illustrated by the historians of antiquity. The following is Thucydides’ comment on the instability of the Greek world in the fifth century bc : So revolutions broke out in city after city, and in places where revolutions occurred late the knowledge of what had happened previously in other places caused still new extravagances of revolutionary zeal, expressed by an elaboration in the methods of seizing power and by unheard-of atrocities in revenge. To fit in with the change of events, words, too, had to change their usual meanings. What used to be described as a thoughtless act of aggression was now regarded as the courage one would expect to find in a party member; to think of the future and wait was merely another way of saying one was a coward; any idea of moderation was just an attempt to disguise one’s unmanly character; ability to understand a question from all sides meant that one was totally unfitted for action. Fanatical enthusiasm was the mark of a real man, and to plot against an enemy behind his back was perfectly legitimate self-defence. Anyone who held violent opinions could always be trusted, and anyone who objected to them became a suspect. (Peloponnesian War 3.82)

The tirade continues in this vein. The remarks about words changing their meanings may be profoundly ironical. Whether they are or not, what is interesting about them in the present context is Thucydides’ reocentric

HI S TOR Y AS P A L I M P S E S T

87

explanation of semantic change. Words change their meanings in order to keep abreast of events. Thucydides clearly disapproves of the events in question, so the implication is that the ‘new’ meanings he describes are debased meanings, reflecting the deplorable state of a degenerate society. The linguistic logic of this condemnation calls for comment. From a strictly reocentric point of view, it is unjustified. For, according to Thucydides’ own account, the ‘new’ meanings reflect the reality of the times more faithfully than the ‘old’ meanings. The reality itself may be lamentable, but that is another matter. From a reocentric perspective, the meaning of words is determined by what /$, not by what ought to be. For Thucydides, however, the need to pronounce a moral verdict takes priority over linguistic coherence. So he condemns those who use the ‘new’ language, even while implicitly conceding that they have no option, if indeed the ‘new’ language does reflect the current state of affairs. This is rather like deploring the fact that a devalued currency is now in circulation, and at the same time criticizing those who use it for contributing to inflation. Implicitly, the ‘right price’ to pay for anything is to be assessed by the old currency, prior to devaluation. But currencies are not devalued for no reason: their value reflects current economic conditions. Similarly, in reocentric terms, words do not acquire new meanings for no reason, since their meaning was always determined by their relationship with things outside language. It is here that the language of Thucydides the moralist starts to clash with the language of Thucydides the reporter. We see him beginning to slide from a reocentric to a psychocentric rationale. For the moralist (in Greece, or in any other place or period), a psychocentric view of language will always prevail over a reocentric view, because on the reocentric view there is nobody to blame. Moralists need to have someone to blame and someone to praise. But words can hardly be blamed for merely reflecting the status quo any more than a five-pound note can be blamed for not being worth what it was worth ten years ago. Switching to a psychocentric view means that what determines the meaning of the word is no longer some neutral, external ‘thing’ but the speaker’s own ‘idea’. Once this move is made, it becomes possible to hold speakers accountable for the uses and abuses of words. Just as a mistake in arithmetic or measurement is attributed to the individual making the calculation, so the misapplication of a term is to be laid at the door of the person who misapplies it. Thucydides had never read Locke; but if he had been able to, he would doubtless have approved Locke’s dictum that

THE L I N G U I S T I C S OF HI S TORY words, in their primary or immediate signification, stand for nothing but the ideas in the mind o f him that uses them, how imperfectly soever or carelessly those ideas are collected from the things which they are supposed to represent. (Locke 1706: III, ii, 2)

This would suit Thucydides the moralist down to the ground; for it is evident that any ‘imperfections’ or ‘carelessness’ in what Locke calls the ‘collection’ of ideas are also the responsibility of the user. Thus, to take up one of Thucydides’ own examples, if I apply the word cowardice to the behaviour of someone who is merely acting circumspectly, that is my fault. I cannot blame the word, or anyone else’s use of the word. For I am applying it in accordance with my own idea of what constitutes cowardly behaviour. Theoretically, it is possible to espouse either a reocentric or a psycho­ centric semantics; the surrogational historian qua historian must opt for one or the other, being committed in advance to giving a truthful account of past events. The question is: whose truth is to be told? The problem with Thucydides and other ‘moral historians’ is that they want to tell two or more versions of the truth at once, but often fail to realize that this undermines the credentials of the historical narrative itself. A comparable Roman case is Livy. He sits on the fence when it comes to accepting or rejecting traditions from the dim, distant past. Such traditions as belong to the time before the city was founded, or rather was presently to be founded, and are rather adorned with poetic legends (poeticisfabulis) than based upon trustworthy historical proofs (incorruptis rerum gestarum monumentis), I purpose neither to affirm nor to refute. It is the privilege of antiquity to mingle divine things with human, and so to add dignity to the beginnings of cities; [. . .] But to such legends as these, however they shall be regarded and judged, I shall, for my own part, attach no great importance. (Ab urbe condita, Praefatio, 6-9)

This is as clear a statement as one could wish of the thesis that history begins when reliable evidence (monumenta) is available. But Livy chides those of his readers who skim impatiently through the historical account in order to reach what is said of their own generation. For there, he suggests, the historian will run into controversy. Personally, he prefers to concentrate on the less recent past: that I may avert my gaze from the troubles which our age has been witnessing for so many years, so long at least as I am absorbed in the recollection of the brave days of old, free from every care which, even if it could not divert the historian’s mind from the truth (a vero), might nevertheless cause it anxiety. (Ab urbe condita, Praefatio, 5)

HI S TOR Y AS P A L I M P S E S T

89

So for Livy, it seems, the chronological scope of trustworthy history is bounded by an upper and a lower limit. The terminus a quo is the time before which there are only unverifiable legends. The terminus ad quern is the time beyond which disputes between living individuals about ‘what actually happened’ blur the picture. In this latter respect, an example of Livy’s historical ‘tact’ is sometimes taken to be his discussion (Ab urbe condita, 4.xx) of the seemingly trivial issue of whether or not Cornelius Cossus was a military tribune when awarded the spolia opima in 437 b c . A simple reocentric matter of fact, one might have thought. But Augustus thought otherwise, and according to one modern commentator ‘it would have been out of the question to refute Augustus, who had political reasons for wanting Cossus not to have been a military tribune’ (Briscoe 1996: 878). So Livy was well aware that the historian’s ‘facts’ about the past sometimes have to be accommodated to present political realities. It is all the more interesting that Livy shows no sign of admitting that the notion of an uncontroversial historical terrain stretching between two historically established termini is itself either a convenient fiction or else a position in philosophy of history that stands in need of supporting argument (which Livy, evidently, has no intention of providing). In his hands, the problem is concealed behind a bland linguistic assumption that the Latin of his own day is perfectly adequate for purposes both of describing and of judging what happened in the past. It is on this basis that he can claim to offer the study of history as a repository of lessons for his contemporaries. What chiefly makes the study of history wholesome and profitable is this, that you behold the lessons of every kind of experience set forth as on a conspicuous monument; from these you may choose for yourself and for your own state what to imitate, from these mark for avoidance what is shameful in the conception and shameful in the result. (Ab urbe condita, Praefatio, 10)

In this profession de fo i we see no hint of any realization by Livy that the language used by the historian in the telling of the tale may itself retrospectively impose certain interpretations on ‘the facts’; or that things may have appeared differently to those who took part in the events described; or even that antagonists are always likely to have different accounts of their conflict. It is as if he thought of time as eventually imposing a closure on debate about what happened. That illusion of closure is all of one piece with treating the language of the historian as being itself adequate to cope with any changes brought by the passage of time.

90

THE L I N G U I S T I C S OF HI S TOR Y

It is unfair, someone may object, to tax Livy in particular with succumbing to this illusion, when so many of his successors did likewise. But no surrogational historian had any other option. Given the institu­ tionalization of history that Livy and others inherited, its rationalization required a certain view of language. With respect to temporality, it had to be a view that allowed the historian’s language to be reocentrically ‘objective’ or disengaged from the events told in the tale. There was only one philosophy of language that ever appeared to afford any such guarantee. All commentators agree on how rapidly Livy’s account became widely accepted as the ‘official’ history of Rome. It was impressively written and intensely patriotic. Not until the nineteenth century and Niebuhr’s magisterial Romische G eschichte was the ‘complete discrediting of Livy’ (Marwick 1970: 36) eventually brought about. However, condemning Livy for his ‘unscientific attitude’ (Foster 1919: xxx) towards his sources is a move which, apart from being anachronistic, misses a more important point. In Livy we see how easily the new ‘historical consciousness’ brought about by the literate revolution could be translated into uncritical respect for written accounts per se. If Livy accepted what he found in Polybius, Claudius Quadrigarius and Valerius Antias, it was because he belonged to a generation already accustomed to regard the written word as being, in general, soberly authoritative and, in the best cases, definitive. It was just such prestige to which his own history aspired, and which his own generation accorded it. When Orwell applied the term palimpsest to the history-making of his fictional totalitarian state, he was envisaging a process by which the original truth was erased and falsehood ‘written over’ it. But this is simply the reverse of the process envisaged by the reocentric historian, in which earlier accounts that were incorrect or incomplete are gradually erased, and more reliable, fuller accounts replace them. That improvement over time supplies the main raison d'etre for the continuation of the historian’s profession. Once we see this, we realize that reocentrism in the Western tradition is locked into a certain view of historical progress by the succession o f historians themselves. Which is precisely why Orwell’s tale is calculated to shock. Language, for the reocentric surrogationist, is not supposed to work in the way Orwell’s Ministry of Truth forces it to work; on the contrary, language is supposed to provide a guarantee that we can, through words, obtain a reliable purchase on reality and inform one another about it. The notion that by relying on verbal report we might easily be systematically deluded about the past is a notion we owe to

HI S TOR Y AS P A L I M P S E S T

91

Orwell’s fertile imagination. In formulating this idea, Orwell had before him the practical examples of official propaganda produced in Hitler’s Germany and Stalin’s Russia. But the idea itself is no more than a simple reversal of the priorities inscribed in the philosophy of language that reocentric historians had accepted for centuries. Orwell also appreciated the difference between merely reocentric falsification and the more dehumanizing psychocentric variety. This is the point of the brainwashing scene where, under torture, Winston Smith is forced not only to say that four fingers are five but to see them as five. The analogue of this in the case of the reocentric historian is the metaphysical conviction that Collingwood called ‘substantialism’. Once the historian is conditioned to thinking of words as surrogates for things, there is a standing temptation to reify all the descriptive terms available for discussing the past. Collingwood writes: [. . .] already in Thucydides the historical point of view is being dimmed by substantialism. For Thucydides the events are important chiefly for the light they throw on eternal and substantial entities of which they are mere accidents. The stream of historical thought which flowed so freely in Herodotus is beginning to freeze up. As time goes on this freezing process continues, and by the time of Livy history is frozen solid. A distinction is now taken for granted between act and agent, regarded as a special case of substance and accident. It is taken for granted that the historian’s proper business is with acts, which come into being in time, develop in time through their phases, and terminate in time. The agent from which they flow, being a substance, is eternal and unchanging and consequently stands outside history. In order that acts may flow from it, the agent itself must exist unchanged throughout the series of its acts: for it has to exist before this series begins and nothing that happens as the series goes on can add anything to it or take away anything from it. History cannot explain how any agent came into being or underwent any change of nature; for it is metaphysically axiomatic that an agent, being a substance, can never have come into being and can never undergo any change of nature. (Collingwood 1946: 42-3)

One of Collingwood’s best examples is Livy. He points out that Livy’s Rome is not just a city, or a people, or a republic, or an empire, but something over and above all these, which continues indefinitely in and through these particular phases or manifestations, fulfilling its own destiny. ‘Rome is the heroine of his narrative. Rome is the agent whose actions he is describing’ (Collingwood 1946: 44). Least of all, for Livy, is Rome just a collection of genetically related individuals who happened to succeed one another in time and live in roughly the same part of the world.

92

THE L I N G U I S T I C S OF HI S TORY

Collingwood, in my view, understates the case. This is perhaps because he does not see that what he calls ‘substantialism’ is the product of a certain philosophy of language, when projected on to the big screen of history. In order to write a reocentric history of Rome, you have to believe that to the word Rome there corresponds a unique entity in reality, which has its own existence and properties, and which is the centre and source of all those events that can properly be called ‘Roman’. Otherwise you would be constantly plagued by doubts about whether what happened fell within your brief or not. For many things may happen which only accidentally happened here rather than there, now rather than later, and recorded rather than unrecorded. Livy’s Rome is a relatively simple type of case. The historical straitjacket of substantialism is forced not only on kingdoms, empires, races and dynasties but on religions, philosophies and even single ideas. They become, once captured by the surrogationist, not just participants in history but even responsible for shaping the course of events. They acquire a momentum of their own, sponsoring ambitious theories about human development in the past and in eras to come. To substantialism on the grandest scale we owe the mind-set of historical thinkers like Hegel, Marx and Toynbee, where truth yields to prophecy as the name of the game, and what actually happened can be ignored or highlighted, depending on the prophet. The most basic problem about truth that surrogational history faces is encapsulated in its unresolved dilemma about proper names. To realize how deeply this cuts, one has only to imagine what would be left as the province of history if all proper names were systematically excised from historians’ accounts. Even if dates survived intact, K ing H arold was killed at the battle o f Hastings in 1066 would end up as A king was killed at a battle in 1066. But this, clearly, would be true of any king who happened to be killed in any battle in 1066. The reocentric historian claims to be telling us more than that. That claim stands or falls by the identification of the king and the identification of the battle. Those identifications are made by introducing proper names. The proper name is the linguistic device that anchors the historian’s account to the truth about what actually happened at a particular time and place. Harold was the name of the king and Hastings the name of the place. So far, so good. The difficulties begin when we start to inquire into the linguistic credentials of the names themselves. Was Harold Harold’s name, or is it just the historian’s name for Harold? If Harold existed, he

HI S TOR Y AS P A L I M P S E S T

93

presumably called himself by some name, even if we are not sure what. Hastings, on the other hand, never called itself anything, but was always called something by people living there. It does not take much reflection on such questions to conclude that the claim K ing Harold was killed at the battle o f Hastings in 1066 already takes too much for granted where names are concerned. We can neither accept nor reject such a claim unless we are clear about how these names are being used and what assumptions lie behind their use. For none of us now alive can have any first-hand evidence about whether such a king ever existed. We can indeed have first-hand evidence about whether there is a place now called Hastings:; but what relationship that bears to the place where King Harold fought remains open to inquiry. In brief, the historian who asks us to accept without further ado any such statement about Harold or Hastings is already asking too much; certainly more than is warranted by any linguistic considerations adduced. Many of us, maybe, feel disinclined to press the historian further about proper names because we do not question our own beliefs about them. (This is part and parcel of the ‘ordinary-language’ theory of history.) However, unless we can be made to feel such an inclination in Harold’s case, we shall have to rest content with failing - or refusing - to distinguish history from myth and fiction. For all that distinguishes the two uses of the name Harold is the presupposition that on one reading Harold actually existed (and took part in the events retailed by the historian), whereas on another reading there was no such person. Was/is there such a person as King Harold, or Jesus Christ, or Homer, or Osama bin Laden? The answers to such questions are not obvious. They require us to match up a proper name to some unique individual. That requirement has led to much controversy. For no one has satis­ factorily explained how, even in theory, such a matching could be achieved. On the subject of names, modern linguistics is no further forward than the Book of Genesis. A well-known attempt to tackle the problem in the early twentieth century was Russell’s proposal that proper names are substitutes or abbreviations for descriptions. In his Philosophy o f L ogical Atomism (1918) Russell wrote: You can take, say, all the things that Livy has to say about Romulus, all the properties he ascribes to him, including the only one probably that most of us remember, namely, the fact that he was called ‘Romulus’. You can put all this together, and make a propositional function saying ‘x has such-and-such proper­ ties’, the properties being those you find enumerated in Livy. There you have a

94

THE L I N G U I S T I C S OF HI S TORY propositional function, and when you say that Romulus did not exist you are simply saying that that propositional function is never true, [. . .] i.e., that there is no value of x that makes it true. [. . .] You see, therefore, that this proposition ‘Romulus existed’ or ‘Romulus did not exist’ does introduce a propositional function, because the name ‘Romulus’ is not really a name but a sort of truncated description. It stands for a person who did such-and-such things, who killed Remus, and founded Rome, and so on. It is short for that description; if you like, it is short for ‘the person who was called “Romulus” .’ (Russell 1918: 242-3)

It is no coincidence that Russell chose as his example the name of a figure from the remote past, about whom little is certain and conflicting beliefs might be held. This, it seems clear, is very pertinent to the problem of the surrogational historian. Nor can it be avoided unless the historian decides to write only about the events of his or her own generation. Does Russell’s explanation of proper names succeed in getting Livy (or any other historian) off the surrogational hook? The first point to note is that Russell’s argument is itself a surrogational argument. Its strategy is to turn the issue of proper names into a psychocentric rather than a reocentric matter. What Russell claims, in effect, is that what can be done with proper names depends not on who or what their bearers are but on what is known or believed about them. Thus, earlier in The Philosophy o f L ogical Atomism, Russell had discussed the word P iccadilly in the following terms: We, who are acquainted with Piccadilly, attach quite a different meaning to that word from any which could be attached to it by a person who had never been in London: and, supposing that you travel in foreign parts and expatiate on Piccadilly, you will convey to your hearers entirely different propositions from those in your mind. They will know Piccadilly as an important street in London; they may know a lot about it, but they will not know just the things one knows when one is walking along it. (Russell 1918: 195-6)

Here Russell’s acceptance of a telementational theory of communica­ tion is quite explicit. (Propositions are located in the mind, words ‘convey’ them from one mind to another, etc.) But, unlike many thinkers in the Western tradition, Russell evidently detects a conflict between the tele­ mentational mechanism and the possibility of fixed linguistic codes. This is because Russell, like Locke, believes that words uttered ‘stand for’ ideas in the mind of whoever utters them, and then again, at the terminal point of transference, for ideas in the mind of the hearer. So if the hearer attaches different meanings to these words, it follows that the speaker will fail, communicationally, to convey the desired message.

HI S TOR Y AS P A L I M P S E S T

95

It further follows that the only circumstances in which communication will be successful are those in which both speaker and hearer share exactly the same knowledge of Piccadilly. How this might be possible Russell does not explain. Presumably the optimal set of conditions would be one in which A talked to B about Piccadilly while both were actually walking down Piccadilly. But even this does not take into account any possible divergence between what A already knows about Piccadilly and what B already knows. How does Russell’s argument carry over to cases like the name Romulus? Presumably, someone who had never met Romulus in person (known him ‘by acquaintance’, as Russell terms it) could never know just the things about him that were known by someone who had. But this makes things doubly difficult for Livy and other historians writing long after the age of Romulus. For even if there were an unbroken chain of testimony extending from those who personally knew Romulus down to Livy, it would necessarily be subject to all the instances of miscommunication resulting from the intervening disparities of knowledge between those informants who were links in the chain. Thus the notion that the name Romulus ‘stands for’ the same person from one end of the chain to the other turns out to be a linguistic illusion. Even if Russell’s general argument is accepted, it is relevant to point out that he trips himself up when he concludes that the name may be regarded as ‘short for “the person who was called ‘Romulus’ ” ’. Livy gives us no guarantee that no one else ever had the name Romulus. The Romulus Livy is interested in was the traditional founder of Rome, and no other. We could, perhaps, treat ‘the person who was called “Romulus” ’ as itself short for ‘the person who was called and whom I am calling “Romulus” ’. Thus Livy would be saved one embarrassment, but only at the expense of incurring another. For if the name in Livy’s account is always to be understood as meaning ‘the person who was called and whom I (Livy) am calling “Romulus” ’, the question that immedi­ ately arises is: ‘How do you, Livy, know who it was you are calling “Romulus”, or who it was your ancestors thought they were calling by that name, or whether there was any such person?’. This in turn leads on to a third difficulty. If the name, as Russell contends, stands for a truncated but inexplicit description in which various properties feature, we would need to know exactly what Livy knew about Romulus in order to understand correctly any proposition of Livy’s concerning Romulus. Although Livy tells us various things about Romulus, it would be rash to suppose that this amounted to the sum total of Livy’s knowedge on the subject. Even if it were, how would we know?

96

THE L I N G U I S T I C S OF HI S TORY

Worse still, perhaps, is that if we assume this to be the case, and thus interpret Livy’s use of the name as ‘short for’ something like ‘the person who, inter alia, was said to have founded Rome’, where inter alia covers all the other things Livy says about him, then any statement by Livy to the effect that Romulus was said to have founded Rome turns out to be a tautology, viz. ‘The person who, inter alia, was said to have founded Rome was said to have founded Rome.’ From all this it would seem that, far from rescuing Livy, Russell’s analysis impales him more firmly on the surrogational hook than ever. More generally, the historian who uses proper names to identify figures from the distant past is thereby conveying propositions about the past that we will almost inevitably misunderstand: we would in any case never be sure whether we did or not. Furthermore, and for the same reasons, the historian can never be sure of understanding the proper names occurring in the sources utilized. Thus history emerges as a palimpsest in which successive scribes have never known exactly what they were erasing. Is there a surrogational theory of proper names that holds out any better prospect for the historian? An alternative proposed by those who reject ‘descriptivist’ accounts like Russell’s has come to be known as the ‘causal theory of proper names’, although the term is somewhat misleading (Wettstein 1995). The main idea is that proper names are instituted not through description but by an original act of baptism or voluntary imposition, and are continued thereafter by a chain of communication long after the circumstances attendant upon the original imposition of the name have been forgotten or obscured. This proposal has obvious implications for historians. Donnellan writes: It is instructive to look at the use o f proper names in historical contexts if only to see why so many philosophers who discuss proper names appeal to examples of it. In general, our use of proper names for persons in history (and also those we are not personally acquainted with) is parasitic on uses of the names by other people - in conversation, written records, etc. Insofar as we possess a set of identifying descriptions in these cases they come from things said about the presumed referent by other people. M y answer to the question, ‘Who was Thales?’ would probably derive from what I learned from my teachers or from histories of philosophy. Frequently, as in this example, one’s identifying descriptions trace back through many levels of parasitic derivation. Descriptions of Thales we might give go back to what was said, using that name, by Aristotle and Herodotus. And, if Thales existed, the trail would not end there. (Donnellan 1972: 373)

HI S TOR Y AS P A L I M P S E S T

97

Thus far, in Donnellan’s account, nothing has been said that Russell would be forced to disagree with. Nor does it bring much relief to the historian; for, in mid-stream, Donnellan switches from focussing on the individual who originally bore the name to invoking a ‘presumed referent’. Donnellan seems to be espousing an unprovable theory of etymology, in which words are assumed to have an ancestry that can be traced back in unbroken succession over the centuries, via innumerable and unknowable acts of communication, to some unique source. It is as if he were talking about some ancestral heirloom, handed down from generation to generation, sometimes lost to view, sometimes misidentified, but always emerging eventually as the same original object bequeathed to posterity. Donnellan continues: The history behind the use of a name may not be known to the individual using it. I may have forgotten the sources from whence I got my descriptions of Thales. Even a whole culture could lose this history. A people with an oral tradition in which names of past heroes figure would probably not be able to trace the history back to original sources. Yet, for all that, they may be telling of the exploits of real men in the past and they may possess knowledge of them and their deeds. (Donnellan 1972: 373)

Thus it emerges that, according to Donnellan’s theory of proper names, what we think we are saying about Thales is not necessarily what we are really saying about Thales, who may really have been someone quite different. We need follow Donnellan’s Alice-inWonderland account no further; for this already reveals the un­ compromisingly reocentric foundation of his etymological speculations. According to his story, few if any historians can possibly know whether they are telling the truth or not. In fact, what they know, unknowingly, may be concealed from them by the very names they use in recounting it. So some of these historians may be telling the truth, even if they do not realize it; because, unknown to them, there actually were real people answering to the names appearing in their historical account. Others may not be telling the truth, because, unknown to them, there were no real people answering to the names used. In what sense the ‘answering’ could be construed still remains obscure. But whatever it may be, history emerges as a palimpsest in which truth occasionally appears, more by luck than judgment. This is a theory of proper names which leads us straight back to Adam and the animals in the Garden of Eden. The primordial act of naming is an archetypal labelling process, pure and simple. And no one, not even Adam, even though he enacts it, understands what relationship that

98

THE L I N G U I S T I C S OF HI S TORY

establishes between the name and what it stands for. Adam, evidently, would not have made a reliable historian. Whether or not we agree with Donnellan’s account of proper names, it can obviously be extended to names in general. I do not need to know how, why, or under what certain circumstances certain flowers were first called daffodils in order to decide whether to call certain flowers daffodils myself. I may do that on the basis of copying others who called apparently similar flowers by that name, and who may in turn have been copying others. In theory, none of us - in my local network of people known to call some flowers ‘daffodils’ —may know whether they are daffodils or were originally rightly so called. (’Even a whole culture could lose this history.’) So just as, according to Donnellan, one may not know whether one is speaking the truth about Thales or not, so one may not know whether one is speaking the truth about daffodils. I take this argument to be a reductio ad absurdum of any philosophy of language which incorporates a reocentric semantics. The case of proper names is thus only one example of the dense nexus of linguistic difficulties encountered by any historians or philosophers of history who are committed to some form of reocentric semantics (and among the more prominent figures falling into this category I would include Hume, Condorcet, Kant, Hegel, Marx, Comte and John Stuart Mill). None of those just mentioned explicitly disavows reocentrism or explains what alternative philosophy of language he is adopting. None questions the assumption that beyond language there lies an independent realm of reality, which provides the basis for words as names, and can accordingly be reported through the linguistic codes available. The disadvantage is that historians committed to a view of linguistic communication as an intrinsically code-based telementational enterprise in effect are submitting their accounts of the past to be judged in the same way as other tales that may be told. The historian is a teller of tales, like any other teller of tales. The historian’s story and the poet’s story are on a par. They make no different demands on our linguistic abilities. They do not even appeal to different forms of the imagination; for it taxes the imagination no less to recreate unwitnessed events that actually happened than to do the same for unwitnessed events that did not. King Harold may well have been killed by an arrow at the battle of Hastings. Or perhaps that is a fabrication. From the words alone, it is impossible to tell. Is not the statement ‘King Harold was killed at the battle of Hastings’ on all fours with ‘Count Roland blew his horn at the battle of Roncevaux’? In both cases, whoever is telling the story, whether a historian or not, is

HI S TOR Y AS P A L I M P S E S T

99

supposedly trying to convey a certain idea to the minds of an audience or readership. In both cases, the assumption is that the success of that communication depends on the common acceptance of the same verbal code by both sender and receiver. For without that assurance there can be no guarantee that the message will be rightly understood. The historian’s statement ‘King Harold was killed at the battle of Hastings’ has to be understood by reference to a linguistic code in which the verb kill is not synonymous with the verbs rescue, save or ransom, and many more. So the problem becomes to distinguish the historian’s message in principle from that of any other teller of tales. This is where a corres­ pondence theory of truth comes into its own. Put very crudely, the historian’s essential claim then becomes: ‘My tale, as linguistically formulated, is true’, where truth requires conformity to an independently given reality which lies outside language. And here the emphasis falls not on the historian’s knowledge, or honesty, or good intentions, but on the linguistic formulation adopted. Unfortunately, as both Russell’s and Donnellan’s argument about proper names demonstrates all too clearly, a strict application of the correspondence theory of truth leaves the historian in the position of someone who does not know - often cannot know - whether the tale is true or not. It therefore comes as no surprise to find that, sooner or later, historians in the Western tradition would find it prudent to rest their case not on truth but on credibility or probability. Condorcet, Hume, Buckle and Mill were among the leading lights in this shift of ground. Mill seems to have thought that, after all, history could be rescued for science by the application of statistics: The facts of statistics, since they have been made a subject of careful recordation and study, have yielded conclusions, some of which have been very startling to persons not accustomed to regard moral actions as subject to uniform laws. The very events which in their own nature appear most capricious and uncertain, and which in any individual case no attainable degree of knowledge would enable us to foresee, occur, when considerable numbers are taken into the account, with a degree of regularity approaching to mathematical. (Mill 1884: VI, xi, 1)

The point here, again, is not whether Mill was right or not, but that for someone with M ill’s philosophy of language (expounded in detail in the opening chapters of the same work), this was as good a case as could be made out on behalf of history. By using such dismissive phrases as ‘Alice-in-Wonderland’ I have already indicated my own attitude towards certain theories of proper names. But my argument does not depend on accepting that view. And

100

THE L I N G U I S T I C S OF HI S TORY

while the discussion presented above could be developed into a general scepticism about our knowledge of the past, whether I propose to develop it thus is what the reader must wait and see. All I have tried to show so far is that a surrogational philosophy of language promotes a surrogational philosophy of history, based on a certain view of the correlations between words and a chosen non-verbal domain; that on this basis the historian’s goal is identified as presenting a veridical account of some selected segment of the past; that such a goal is superficially congenial to the historian, since it appears to be in principle attainable with the linguistic resources available; but that in the end this conception of the task leaves the historian with linguistic problems that are insoluble within the surrogational framework. Before proceeding further I shall consider one major challenge to surrogational theory that emerged in the twentieth century. It was the first serious challenge for more than two millennia.

CHAPTER FOUR

Historicism and linguistics

If I am right in supposing that a philosophy of history has to have the backing of a philosophy of language, it would be reasonable to expect that when a radically new philosophy of language appears, that will auto­ matically sponsor in due course a radically new philosophy of history. In this chapter I wish to argue that just such an expectation is confirmed by one of the major intellectual upheavals of the twentieth century. Surrogationist assumptions about language were not questioned in any systematic and penetrating way until the advent of the movement known as ‘structuralism’. Historians and philosophers of history took little notice of it at the time. Although the foundational text, Saussure’s Cours de linguistique generate, was published in 1916, it was not until after World War II that there belatedly emerged a so-called ‘structuralist’ philosophy of history. In order to realize how deeply structuralist thinking affected ideas about language it is necessary to appreciate that it was not until the nineteenth century that languages had been investigated in detail as historical phenomena. It would be no exaggeration to say that the main condition for setting up a ‘science’ of linguistics was the recognition that Greek, Latin, English, French and all other tongues are not ‘timeless’ objects, but are no less subject to continuous change than the com­ munities they serve. Structuralism can scarcely be understood except as a reaction (over­ reaction?) against the prevailing historicism in linguistics. Saussure himself gave credit to the linguists of the Neogrammarian school for the establishment of the historical approach: 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)

101

102

THE L I N G U I S T I C S OF HI S TORY

But this historical perspective was something far removed from that of the traditional historian. What it involved is revealed by Saussure in the same chapter, where the paradigms of the Latin noun genus and Greek genos are set out side by side and then compared to the corresponding Sanskrit forms. The forms in question, as given by Saussure (in Latinized transcription) are: Latin: genus, generis, genere, genera, generum Greek: genos, geneos, genei, genea, geneon Sanskrit: ganas, ganasas, ganasi, ganassu, ganasdm

According to Saussure, comparing the Greek and Latin forms ‘tells us little’. But when they are set beside the Sanskrit forms all is revealed: ‘At a glance [II suffit d’y je te r un coup d’oeif] we now can see the relationship between the Greek and Latin paradigms’ (Saussure 1922: 15). But, 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 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. But that is nothing we can ‘see’ either, much less ‘at a glance’, and certainly not by mere inspection of just three word paradigms. In short, the historical message which is so evident to Saussure is no more obvious to the untrained eye than the historical message contained in successive layers of soil or rock is apparent to the observer who is not a geologist. In fact, treating mere sets of linguistic forms as historical evidence involves a subtle shift in the application of the concept ‘histori­ cal’. The new perspective is by no means directly comparable to that of the political historian, who relies on documents from the past for information about past events. Linguistic forms, even when attested in writing, are not in that sense ‘historical documents’, although Saussure and other linguists repeatedly insisted on that equivocation. The point of departure for the introduction of this new perspective on language was commonly held 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

H I S T O R I C I S M AND L I N G U I S T I C S

103

refined than either, yet bearing to both of them a stronger affinity, both in the roots of 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 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 ‘insight’ (as it has sometimes been called) vouchsafed to Jones. 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) encyclopedia of linguistics can still assure its

104

THE L I N G U I S T I C S OF HI S TORY

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 modern linguistics. 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.’ This contrasts very markedly with Bopp’s statement in the Preface to his Vergleichende Grammatik of 1833, 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 indepen­ dent academic discipline a more improbable advocate. Here, in short, we have meta-meta-history (at three removes from the original, as Plato said of the artist’s picture). It is a revealing example of how scholars set about constructing historical accounts of their own discipline designed to safeguard its claims to autonomy and sometimes its national identity. In the case of comparative philology, that process of historification was already in full swing by the middle of the nineteenth century. In an article contributed to the Edinburgh R eview in October 1851, Max Muller 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 every thing new has to contend with. (Muller 1851: 125)

H I S T O R I C I S M A ND L I N G U I S T I C S

105

The occasion of Muller’s article was the appearance of an English translation (by Eastwick) of Franz Bopp’s Comparative Grammar o f the Sanskrit, Zend, Greek, Latin, Lithuanian, Gothic, German, Sclavonic Languages, which Muller describes as being ‘universally considered as the classical work on this branch of Comparative Philology’. In this same article, Muller 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’ (Muller 1851: 141). So already by 1851 we see the historicist myth of Jones, founder of comparative philology, being propagated to the British general public by a German scholar in Oxford, anxious to establish his own credentials in the British academic world. Muller repeats the quotation from Jones in the fourth of his highly popular Lectures on the S cience o f 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 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. (Muller 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 Muller ‘a very great man - one in a million’ (Opening address to the International Congress o f Orientalists). Almost contemporaneously with Muller’s article in the Edinburgh R eview, there appeared in the North British R eview (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’. The claims of comparative philology to be a ‘science’ were, as later scholars

106

THE L I N G U I S T I C S OF HI S TOR Y

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 fo r 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. Even today the term ‘comparative philology’ survives in the titles of posts in more conservative universities, as if the antiquated terminology could ensure the survival of a form of linguistic inquiry which has long ceased to offer the world anything new. However, Strettell’s version of history differs notably thereafter from Muller’s. The hero of Strettell’s version is not Jones (not mentioned once) but Humboldt. Humboldt does feature in Muller’s story, but not as a possible rival to Jones. For Muller, 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 move­ ment 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 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 retrospectively 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

H I S T O R I C I S M A ND L I N G U I S T I C S

107

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 was 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 co-author of Grimm's Fairy Tales. Both brothers were also involved in the monumental dictionary of the German language {Deutsches W orterbuch) which bears their name. It was begun in 1837, shortly after their dismissal from professorships at Gottingen for political protest against civil rights abuses, and was left unfinished when they died (Jacob in 1863 and his brother in 1859). It was Jacob Grimm’s Deutsche Grammatiky 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 / (Latin pa ter, English father:; Latin piscis, English fish , etc.) But something more exact than this was needed for ‘scientific’ purposes. Max Muller, who claimed to have coined the term ‘Grimm’s Law’, formulated it to his own satisfaction as follows: I f 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/, instead of the corresponding mediae g and b. (Muller 1864: 199-200)

108

THE L I N G U I S T I C S OF HI S TORY

The importance of these correspondences for comparative philology was clearly historical: 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 astron­ omy - both sciences which accepted currently observable phenomena as reliable indicators of unobserved prior events. This was (part of) what Max Muller had in mind when he proclaimed (in his second series of Lectures on the Science o f 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 Muller 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 hold without a certain amount of juggling and allowance for exceptions. Thirdly, they argued, the iaw ’ was in any case not Grimm’s but Rask’s, because Rask had drawn attention to the relevant comparisons earlier and Grimm had simply borrowed the idea from him. 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 S chrift expresses ‘eight sounds through seven signs’ - a calculation which requires sch to ‘stand for’ three sounds and / 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

H I S T O R I C I S M A ND L I N G U I S T I C S

109

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 Brandekilde in Denmark in 1787 and was thus Grimm’s almost exact contemporary. He, in turn, has been hailed as ‘the father of historicalcomparative linguistics’ and ‘le fondateur de la methode du XIXe siecle dans le domaine de la grammaire comparee’. 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 Undersegelse om det gam le 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 discrep­ ancies). The issue that then presented itself was how best to establish linguistic affinity. Here Rask is often said to give priority to comparison of grammatical structure, but he attached importance also to correspon­ dences between words constituting core vocabulary. Interlingual cor­ respondences 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 (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 linguist’s inquiries push back into a period before written 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. #

#

#

110

THE L I N G U I S T I C S OF HI S TOR Y

A third front-running candidate, Franz Bopp, was born in Mainz in 1791 and studied in Paris under the French Oriental scholars Chezy and Silvestre de Sacy. The publication in 1816 of Bopp’s Uber das Conjugationssystem der Sanskritsprache in Vergleichung mit jenem der griechischen, lateinischen, persischen und germanischen S prache has been claimed to be ‘generally regarded as marking the beginning of compara­ tive 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 (1833-52). He also published a number of grammars of Sanskrit. In 1866 his Vergleichende Grammatik was translated into French by Michel Break Of particular interest here is what Breal saw as being the originality of Bopp’s contribution to his discipline. In the introduction to his trans­ lation, Breal 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. (W olf 1991: 26)

Breal goes on to point out that even before Sir William Jones a French Jesuit in Pondicherry, Cceurdoux, had raised the question of the relation­ ship between Sanskrit, Greek and Latin in a memorandum to the Academie des Inscriptions et Belles-Lettres in Paris, and identified similarities in words and grammatical structures. (This was in 1767, two decades before Jones’ allegedly landmark lecture to the Asiatic Society.) In a celebrated put-down remark, Meillet said of Bopp that he had discovered comparative grammar while looking to explain IndoEuropean, 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-

H I S T O R I C I S M A ND L I N G U I S T I C S

111

European. In this, said Meillet (in 1903), Bopp was still a man of the eighteenth century. Breal 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 Breal 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 fo r granted and contented themselves with drawing attention to examples which seemed to corroborate it. This is confirmed, as Muller points out (Muller 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 develop­ ment.’ 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 historical 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. Never far beneath the surface of linguistic studies in the nineteenth century is another historicist conviction: that linguistic evolution, proper­ ly interpreted, reveals deep truths about the mental ability of different peoples. According to Strettell, writing in 1851, Humboldt had already shown in his posthumously published dissertation Uber die K aw i-S prache a u f der Insel Ja w a 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

112

THE L I N G U I S T I C S OF HI S TORY

British, whereas other languages became ‘petrified’, because ‘the national mind did not advance’. Similarly, Max Muller 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 depths of individual character to an exercised and reflective mind, which even their practical works, their institutions and customs, cannot unfold’ (Strettell 1851: 165). So we have not only a new form of historical evidence, but one which surpasses any available to the traditional historian. 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 o f 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 their expression in language. (Strettell 1851: 165-6)

It would be silly to conclude from this that comparative philology was a cultural byproduct of missionary zeal and the White Man’s burden. Nor do we have to suppose that the comparative philologists would all have subscribed to supremacist doctrines of the kind represented in the mid­ nineteenth century by Gobineau’s Essai sur rinegalite des races humaines. (Pott, for one, had criticized Gobineau in his book Die Ungleichheit m enschlicher Rassen (1856).) Nevertheless, in Strettell’s comments we see an aspect of comparative philology which is by no means irrelevant to the establishment of the subject as part of an educational system. It is interesting to note the role of national stereotypes in Strettell’s explana­ tion 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. What the comparative philologists never seem to have realized was that interlinguistic correspondence formulae (of the kind typified by Grimm’s

H I S T O R I C I S M A ND L I N G U I S T I C S

113

Law) are just that: no more than 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 much later in his Introduction a f etude com parative des langues indoeuropeennes (a book dedicated, significantly, ‘a mon maitre, M. Ferdinand de Saussure’). There Meillet wrote: What the method of comparative grammar yields is not a reconstruction of IndoEuropean as it was spoken: but nothing more than a definite system o f correspondences between the languages historically attested. (Meillet 1903: 27)

(The italics here are not mine but 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 lindo-europeen. But that is exactly what the comparative philologists thought they were about: a programme of historical recovery. If not way ahead of Ranke (whose first work, with its famous dictum about showing wie es eigentlich gew esen, was published in 1824), the comparative philologists may be said to have shared his objective and demonstrated howy at least in the fie ld o f language, it was possible to achieve it. Nowadays linguists admit that what Grimm’s Law ‘actually was’ (as one of them puts it) is not uncontroversially given, and depends on what assumptions one 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; i.e. that the name ‘Grimm’s Law’ indeed designates some historical reality event(s) or process(es) - and it merely remains to identify this reality. (It is rather as if political historians already had a prior assurance that something called the ‘French Revolution’ actually occurred, but were

114

THE L I N G U I S T I C S OF HI S TORY

hampered by the total destruction of contemporary evidence in their task of determining who revolted against whom, when, where, why and how.) The face-saving view nowadays adopted by many linguists endeavours to maintain some intermediate position between ‘naive realism’ (e.g. treating, say, ‘Grimm’s Law’ as the designation of some complex un­ attested series of 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 it both ways at once. In terms of this distinction, the early comparative philologists were un­ doubtedly ‘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 trying to have it both ways at once. 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. This is a form of historicism that stands history on its head. Intrinsic to the comparative philologists’ historical thinking were (1) the assumption that ‘languages’ are discretely identifiable entities which can safely be treated as consisting of decontextualized sets of forms and (2) that languages multiply by a process of internal division or branching, whereby one (or more than one) subpopulation of an originally mono­ lingual 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.

H I S T O R I C I S M A ND L I N G U I S T I C S

115

Although this ‘family tree’ or Stammbaum model is usually associated with August Schleicher (1821-68), 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 historical 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 IndoEuropean 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 in some remote part of the globe 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 Verwandtschaftsverhaltnisse der indogermanischen Sprachen (1872), which put forward the alternative historical hypothesis 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. The basic idea of languages as discrete, homogeneous entities is challenged by the notion of a continuum of linguistic variation

116

THE L I N G U I S T I C S OF HI S TORY

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 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. Com­ parative 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 P hilology fo r 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’, and so on. 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 anthropologically 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

H I S T O R I C I S M A ND L I N G U I S T I C S

117

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 (Farrar 1873: 27) - 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’ (Farrar 1873: 30). 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 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.) Janos Gulya suggested a quarter of a century ago that although comparison of the Finno-Ugrian languages got under weigh 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 Undersegelse was written as a prize essay. The organizers of the competition (the Danish Academy) were clearly committed to establish­ ing the academic credentials of a ‘Nordic’ mother tongue. So in one sense Rask’s historical 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 Undersegelse. 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

118

THE L I N G U I S T I C S OF HI S TORY

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 people wanted it to prove all along, it is doubly welcome. The comparative philologists, in common with their successors, the Neogrammarians, were far more interested in the comparison of forms than the comparison of meanings. It was not until the 1880s that Michel Breal coined the term semantique and not until 1897 that he published the first full-length treatise on the subject, his Essai de semantique. Breal viewed semantics as a science of the psychological laws governing the evolution of languages. It was to form a counterpart to the study of sound change, which Breal regarded as governed by physiological laws. Until this ‘official’ recognition of semantics as a branch of linguistics, comparative and historical grammar had been underpinned, as regards the meaning of forms, by nothing more rigorous than traditional etymol­ ogy. Here we encounter a variety of confused but often ingenious speculations over the centuries separating Plato from Horne Tooke. Common to these —insofar as they had anything in common —was the notion that the ‘true’ meaning of a word was its original meaning. This is a basically historicist thesis. It carries the implication that semantic change is a deviation or degeneration that masks the rationale governing the ‘proper’ use of the word. It also implies that by going back in time, and stripping away successive layers of semantic change, the philologist can uncover this rationale. Nineteenth-century historical linguistics did not need to subscribe to this view, but it still needed etymology in order to underwrite its reconstructions of past linguistic evolution. For without continuity of form and meaning, historical ‘explanations’ of how languages develop collapse (just as, in political history, historians cannot afford to suppose that there are any real ‘gaps’ in the sequences of events they are claiming to connect). So as regards languages a curious situation arises in which what survives theoretically is the bare assumption that somehow mean­ ings too have a history of their own, but no one is very clear about what exactly it consists in. The meanings are in some sense continuous, but nevertheless subject to change over time as a result of factors that cannot be predicted, but might in principle include anything and everything

H I S T O R I C I S M A ND L I N G U I S T I C S

119

connected with the use of the word in question. Breal’s semantics is best seen as a belated attempt to bring some kind of classiflcatory order into this etymological chaos. It was to be swept away by Saussurean semantics, in which etymology played no role whatsoever in determining the meaning of a word. What is perhaps most important in all this 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 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 premiss 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. The story about enduring things that remain the same even while undergoing all kinds of change is one I have already touched on in discussing Collingwood’s fallacy o f‘substantialism’. The notion that ‘the same language’ which started off as Latin eventually ended up as French is a classic example of this historical fallacy, as Saussure eventually saw (though he did not describe it in Collingwood’s terms). Rather than subscribe to this fallacy, Saussure’s mature proposal for the future of his discipline was to replace ‘historical linguistics’ as understood in the nineteenth century by ‘diachronic linguistics’. One of the great ironies in this chapter of ‘the history of linguistics’ is that many of Saussure’s successors, including Levi-Strauss, thought that diachronic linguistics (linguistique diachronique) was just an alternative term preferred by Saussure for what had previously been called ‘historical linguistics’. Nothing could be further from the mark. What Saussure proposed was that the notion of a language changing over time be replaced by that of a chronological succession of ‘linguistic states’ {etats de langue), each independent of its predecessors or successors. Synchronic linguistics would be concerned with the etats de langue as such. Diachronic linguis­

120

THE L I N G U I S T I C S OF HI S TORY

tics would be concerned with the particular events linking successive etats de langue. According to Saussure, when a linguistic change occurs, for whatever reason, it provokes a reorganization of the system, and this reorganization cannot be predicted. The consequences may be negligible or profound, depending on the case. Such changes are never systematic in the sense that it is possible to say that when change a has occurred, change b will follow. Nor are they systematic in the sense that they are internally co­ ordinated with one another. Each is an isolated occurrence and unfore­ seeable. Thus for Saussure there are no diachronic ‘laws’ governing linguistic evolution (Saussure 1922: 131). This structuralist conclusion ran against the grain of mainstream nineteenth-century linguistics, where pursuit of the claim to have linguistics recognized as a science depended crucially on establishing so-called ‘laws’ of sound change. Here we see re-enacted within the narrow province of language an old historicist debate about the ‘regularity’ of human affairs. Saussure was certainly aware of the views of Comte and possibly those of Mill, who was greatly impressed by the unsuspected regularities that statistics seemed to reveal (Mill 1884: V l.xi.l). Although he never denied the so-called ‘regularity’ of sound change, Saussure evidently regarded it as a case where linguists had been deceived by historical hindsight into taking detectable patterns to reflect forces operative within the language. What Saussure took from his predecessors, and in particular from the Neogrammarians, was the notion that sound change operates ‘blindly’, i.e. without regard to the meanings of words or grammatical distinctions. This became, in Saussure’s hands, the doctrine of the ‘arbitrariness of the linguistic sign’, which he set up as the ‘first principle’ of linguistics (Saussure 1922: 100-2). It is but a short step from claiming that form has no necessary connexion with meaning to claiming that the lexical and grammatical distinctions that a language imposes on its speakers have nothing do with anything outside language either. They are all ‘internal’ and ‘arbitrary’, varying from language to language. Saussurean structuralism was widely interpreted in this sense; that is, as denying that words derive their meanings from an independently given reality. What the word gold means would owe nothing to the existence of the metal gold. The existence or non-existence of the metal would be an irrelevance to the semantics of the word. And mutatis mutandis the same would hold for all other words. The effect of this denial is to liberate language from any dependence on the

H I S T O R I C I S M AND L I N G U I S T I C S

121

non-linguistic world. It implicitly rejects the picture conjured up by the nomenclatorial exploits of Adam in the Bible. There was no pre-existing set of separate species for Adam to ‘name’. If Adam had been a structuralist, he would have said to God: ‘Look, Lord: stop assuming that your acts of creation have produced the only possible classification of creatures. I can produce as many different classifications as I want, just by introducing names for them. And I can also introduce names for creatures you haven’t yet produced at all.’ If God then said to Adam: ‘That is ridiculous! How can there be names for creatures I have not created?’ Adam the structuralist’s reply might have been as follows: ‘Words are not labels. By creating words, I can analyse and classify not just your creations but all imaginable creations. If I restrict myself to just two words, then that automatically divides all actual and possible creatures into two classes. If I create forty-nine words, then there are forty-nine classes. But that has nothing to do with you r activities. My words create their own values, and those values are determined by nothing other than the similarities and contrasts between them. Once they are introduced, I have no further control over them. They have a life of their own. And neither you nor I, Lord, can do anything about it.’ If the rebellious role of Adam in the above scenario is transferred from a mythical individual to the linguistic community as a whole, then an incipient structuralist philosophy of language is already up and running. Adam the structuralist was the world’s first linguistic heretic. Although Saussure initiated this kind of thinking about language, whether he would have gone along with all its implications is a matter best left for experts on Saussure to settle between themselves, if they can. What is more important in the present context is how that way of thinking impacted on philosophy of history. The leading figures in structuralist philosophy of history are usually taken to be Roland Barthes and Claude Levi-Strauss. Some commenta­ tors see the originality of their approach as consisting in a kind of dehistoricization of history. Thus according to Hugh Rayment-Pickard, Unlike the dynamic system of say Hegel, which evolved in time, the structuralists saw the narrative systems of history as timeless diagrams. Thus Levi-Strauss described the narrative of history as an ‘a-temporal matrix’ detached from historicity; and Barthes has argued that in narrative ‘temporality only exists in the form of a system’ and that ‘true’ time is merely ‘a reality effect’ of language. (Rayment-Pickard 2000: 276)

122

THE L I N G U I S T I C S OF HI S TORY

This assessment, a sceptic might say, takes both Barthes and Levi-Strauss more seriously as philosophers than they deserve to be taken. I shall begin with Barthes. His essay ‘Le discours de l’histoire’ (1967) sets the stage for all his subsequent pronouncements about history. It is an essay which, beneath its thick veneer of linguistic terminology, sets out to undermine what Barthes regarded as ‘bourgeois’ history; that is, the (alleged) monopoly by French right-wing historians o f‘official’ accounts of the past. In order to do this, Barthes has to present the whole enterprise of writing history in the Western tradition as being less a search for ‘the truth’ about the past than an effort to present partisan versions of the truth; and this in turn casts cumulative doubt on whether there actually is any objective truth to be reported. In Barthes’s essay, apart from the ritual references to historians of antiquity (Herodotus, Thucydides, Xenophon, Caesar, Tacitus), the only writers mentioned by name are the following: one from the Middle Ages (Joinville), one from the Renaissance (Machiavelli), one from the seven­ teenth century (Bossuet), one from the eighteenth (Michelet), and four from the nineteenth (Thiers, Thierry, Blanc and Fustel de Coulanges). It would be hard to select a group of more resolute axe-grinders to represent Western history. But drawing attention to the existence of biassed historians hardly constitutes a ‘structuralist’ argument per se. Barthes puts them all into a Procrustean linguistic bed, in which historical narrative is subject to a ‘linguistic’ analysis. In this analysis, the chief - virtually the sole - analytic instrument deployed is the Jakobsonian ‘shifter’ (a word that Barthes leaves untranslated, and thus endowed with all the exotic mystique of the foreign borrowing). Although Barthes did not tell his readers this, Jakobson had cribbed both the concept and the term from Jespersen, who had first discussed it in 1922. But since Jakobson’s essay had only recently (1964) been translated into French, it sounded to Barthes’s French audience like the very latest transatlantic wisdom. In fact, there was nothing either new or ‘structuralist’ about it. A ‘shifter’ was nothing more than what had traditionally been recognized as a deictic, defined by Jespersen as a word ‘whose meaning differs according to the situation’. The most conspicuous class of such words, in Jespersen’s view, were personal pronouns. Thus the meaning of I depends on who is speaking, while the meaning of you depends on who is being addressed. Other deictic devices recognized as such in traditional grammar were verb tenses and temporal or locative adverbs {now, then, here, there, etc.). But Barthes writes as if Jakobson were the first linguist to reveal the phenomenon of deixis to an astonished world.

H I S T O R I C I S M AND L I N G U I S T I C S

123

The argument of Barthes’s essay is shallow. It emerges that, in the end, all he has to say is that historians, like other writers, utilize a variety of deictic devices in order to frame and present their narrative. These devices include adopting impersonal constructions to mask the fact that the account being presented is actually the view of the writer. Barthes also points out that the amount of space that a historian devotes to recounting different events does not match the supposed length of ‘real’ time over which the events took place. But none of this is specific to the writing of history. It applies pari passu to myth, romance and occasional reports of all kinds. Barthes apparently wishes us to conclude from this ‘linguistic’ demonstration that historians are not to be trusted: their accounts are artful constructions, not objective reflections of a given reality. As he puts it, the classes of ‘content units’ (unites du contenu) and their sequence are ‘the very same that have seemingly been discovered in fictional narrative’ (Barthes 1967: 161). This, according to Barthes, shows us that: in the nineteenth century, at the privileged moment when history tried to establish itself as a genre, it came to treat ‘purely and simply’ relating the facts as the best proof of those facts, and to set up narrative as the privileged signifiant of the real (du reel). (Barthes 1967: 166)

It takes a little time to recover one’s breath after being told that the institutionalization of history as a genre belongs to the nineteenth century. But worse is to follow for anyone familiar with the language of structural­ ism. Signifiant is a Saussurean technical term. But it is deployed here in a way that reveals Barthes’s failure to understand it. In Saussurean terminology, the signifiant, as one facet of the linguistic sign, belongs to la langue, not to parole; whereas historical narration belongs to parole (or discours, as Barthes prefers to call it). There is no sense to be made of the claim that narrative is a signifiant of reality or real events. In a structuralist philosophy of language, reality is not part of any linguistic sign or combination of signs: it lies ‘outside’ language altogether. By the same token, narration does not ‘signify’ anything: it is not even a sign, but is just one possible way, among many others, of organizing parole. In another no less confused essay written about the same time, ‘Introduction a Panalyse structural des recits’ (1966), Barthes claims that time as presented by the narrator is a linguistic mirage: temporality is only a structural class of the narrative (of discourse), just as, in the language, time exists only in the form of a system; from a narrative point of view, what we call time does not exist, or at least exists only functionally, as an element of a semiotic system: time does not belong to discourse properly so called, but to the

124

THE L I N G U I S T I C S OF HI S TOR Y

referent; the narrative and the language recognize only a semiological time; ‘real’ time (le ‘mat temps) is a referential ‘realist’ illusion [. . .]. (Barthes 1966: 184)

Hayden White chose a trendy quotation from Barthes as the epigraph for his volume of essays on historical writing entitled The Content o f the Form (1987). It reads: ‘Le fait n’a jamais qu’une existence linguistique.’ What a linguistic existence would be, divorced from any other kind of existence, we are never told. Bearing all this in mind, it is easy to see why Rayment-Picard construes Barthes as propounding a revolutionary philosophy of history. For if real time is an illusion, the whole basis of the traditional historian’s account of the past is called in question. Without that mooring, the events recounted are themselves no more than projections conjured up by the narration, just as in poetry and other forms of fiction. The trouble is that nowhere does Barthes produce any structuralist argument for drawing such a conclusion. And the reason why he does not produce one is that there is none. All Barthes has managed to do (if we wish to take a charitable view of his linguistic meanderings) is reaffirm, with the help of a structuralist vocabulary, the position of those theorists of antiquity who treated the writing of history under the rubric of rhetoric because it featured standard rhetorical techniques. It does not follow, one need hardly say, that ‘real’ time and ‘real’ events are illusory, any more than it follows from the fact that a portrait takes the form of patches of paint, laid down with skill by the painter, that the existence of the sitter is an illusion. True, we may not be able to tell just from inspection of the painted image whether or not this is a portrait of a real person or some character of the painter’s imagination. Or whether it is partly real and partly imaginary. But it was laughable to suppose in the first place that careful inspection of the paint on the canvas would somehow reveal this. Barthes’s crass error was to suppose that a structuralist analysis of the linguistic sign, as proposed by Saussure and his successors, because it rejected the reocentric account of words, licensed the conclusion that history was a branch of fiction. It does nothing of the kind. But the point I wish to draw attention to here is that the very conclusion Barthes draws, fa r from being structuralist in spirit, is actually a reactionary reaffirmation of surrogationist thinking about language. By opposing ‘real time’ to narrative time, one implicitly posits a pre-linguistic sequence of events to which the narrator is ultimately answerable, and by reference to which any narration can be judged. Fiction then emerges (as for Aristotle) as the null possibility in which no such events ‘really’ took place. #

#

#

H I S T O R I C I S M AND L I N G U I S T I C S

125

The case of Levi-Strauss is even more curious. Rayment-Pickard describes him as ‘an unlikely figure to have risen to significance in the philosophy of history’ (Rayment-Pickard 2000: 277). ‘Unlikely’ is an understatement. The reason why history was of any interest to LeviStrauss at all was that his chosen field of investigation in cultural anthropology was ‘myth’: and myth, in the Western tradition from antiquity onwards, stands opposed to history. So to be a serious student of myth almost automatically requires one to have a theory of history, if only to have some basis on which to define myth by contradistinction. (It is worth recalling that Barthes also climbed aboard the bandwagon by calling his first collection of essays (1957) on mass culture M ythologies.) Levi-Strauss’s structuralism was Saussure transmitted via the un­ reliable hands of Jakobson, whom Levi-Strauss encountered as an exile in New York during World War II. Before the war, structuralism seems to have made no impression on Levi-Strauss at all. In 1945, however, he published a remarkable and controversial essay in which he claimed that, as a discipline, anthropology must follow the lead taken by linguistics. Linguistics, in his view, was unique among the social sciences in develop­ ing a proper scientific method. The method to which he alluded was phonemic analysis (developed not by Saussure but by Jakobson and Trubetzkoy upon a Saussurean basis). Levi-Strauss wanted a universal method which would yield a determinate analysis of all myths, world­ wide, in just the same way as phonology (in his view) yielded a determinate analysis of sound systems for any language anywhere. This is why he chose to call his approach ‘structural anthropology’, a term deliberately calqued on structural linguistics. What attracted Levi-Strauss even more was the notion that structural linguists had shown an apparently reliable way of analysing patterns of linguistic behaviour o f which the speakers themselves were not fu lly aware. If such a method could be developed in anthropology, this meant that anthropologists were no longer condemned to rely on direct questioning of members of the society under investigation. A traditional story could be treated as evidence in its own right, irrespective of what informants took its ‘meaning’ to be. (Cf. Saussure’s assumption that linguistic forms in themselves constituted historical documentation.) The informant might not be conscious of its structure or of the underlying significance of particular features. Structural analysis held out the promise of getting ‘behind’ the superficial details and interpretations supplied by those who provided the anthropologist with first-hand ‘data’. Levi-Strauss was so excited by this prospect that he began to regard all components of culture as ‘languages’ (i.e. as structured systems or langues

126

THE L I N G U I S T I C S OF HI S TORY

in the Saussurean sense), even inventing terminology to match. Thus, for the analysis of cuisine, he proposed a unit called the gustem e, by analogy with phoneme. Chinese cooking could accordingly be explained as based on its own set of gustemes, just as the Chinese language was based on its own set of phonemes. What Levi-Strauss never explained was exactly how an anthropologist could set about identifying the gustemes of a culture in a manner parallel to the methods employed by the linguist to identify phonemes. (The fatuity of concepts like the gusteme was doubt­ less masked at the time by the fact that scholars with far more plausible credentials as linguists were proposing no less ambitious theoretical units, such as the behavioreme. It might even be that some of them envisaged a total analysis of ‘culture’, in which both phonemes and gustemes were subvarieties of behavioremes.) It is against the backdrop of structuralism run amok in the heady postwar years that Levi-Strauss’s view of history has to be set. LeviStrauss adopted with enthusiasm Saussure’s distinction between synchro­ nic and diachronic studies. Disappointingly, it was a distinction that he, like many others, never understood. The reasons for this I shall come back to below. But in Levi-Strauss’s case, it led to a conundrum for anthropology: is it possible for the investigator to describe a present state of society without reference to its past, or to reconstruct the past without reference to the present? In his essay ‘Histoire et ethnologie’ (1949), later incorporated into A nthropologie structurale (1958), he defines the problem as follows: we can formulate the problem of the relationship between the anthropological sciences and history as follows: Either anthropology is focused on the diachronic dimension of phenomena, that is, on their temporal order, and thus is unable to trace their history; or anthropologists attempt to apply the method of the historian, and the time dimension escapes them. The problem of reconstructing a past whose history we are incapable of grasping confronts ethnology more particularly; the problem of writing the history of a present without a past confronts ethnography. That is, at any rate, the dilemma which has too often halted the development of these sciences in the course of the last fifty years. (Levi-Strauss 1958: 3)

In order to grasp the point Levi-Strauss is making, it is necessary to understand two things. The first is that in the above passage his American translator has rendered by anthropological (and, by extension, anthrop­ ology and anthropologist) what appears in the original as ethnologique. The second is that on the preceding page Levi-Strauss has given an explan­ ation of what he understands the difference to be between ethnologie and ethnographie. The latter is the observation and analysis of human groups

H I S T O R I C I S M AND L I N G U I S T I C S

127

considered 4dans leur particularity and aiming at rendering as faithfully as possible of the life of each one. The former, on the other hand, is a study which makes a comparative use of the data acquired by the ethnographer. It here becomes evident that Levi-Strauss assigns a priority to ethnog­ raphy over ethnology, in the sense that ethnology is impossible until ethnographers have done their work. At the same time, he cannot see - or professes to find difficulty in explaining - how ethnographers can write the history of a present that has no past (Thistoire d’un present sans passe’). What is odd about Levi-Strauss’s proclaimed structuralism is that Saussurean linguistics had upheld the validity of doing ju st that: i.e. analysing a linguistic state in its own terms without reference to any antecedent state. This - and nothing else - is the theoretical implication of the term synchronic. Later in the same essay, Levi-Strauss puts his foot in it again. Even the analysis of synchronic structures, however, requires constant recourse to history. By showing institutions in the process of transformation, history alone makes it possible to abstract the structure which underlies the many manifestations and remains permanent throughout a succession of events. (Levi-Strauss 1958: 21)

Here again we have the specifically Saussurean term synchronic, and only a few paragraphs earlier Saussure’s Cours de linguistique generate has been hailed as a landmark in the structural linguistics that Levi-Strauss is holding up as a model to his fellow anthropologists. If further evidence were needed of Levi-Strauss’s deep confusions about structuralism, the following passage should suffice: From words the linguist extracts the phonetic reality of the phoneme {la realite phonetique du phoneme); and from the phoneme he extracts the logical reality of distinctive features. And when he has found in several languages the same phonemes or the use of the same pairs of oppositions, he does not compare individually distinct entities. It is the same phoneme, the same element, which will show at this new level the basic identity of empirically different entities. We are not dealing with two similar phenomena, but with one and the same. (Levi-Strauss 1958: 20)

None of this has any grounding in the Cours de linguistique generate, where on the contrary ‘phonetic entities’ have no place in linguistic structure and it is impossible for the linguist to equate structural elements across languages. According to Levi-Strauss, it is a mistake to think of the historian as reporting facts about the past. In La pensee sauvage he writes:

128

THE L I N G U I S T I C S OF HI S TORY

the historical fact is no more given than any other; it is the historian, or the agent of historical development {Fagent du devenir historique), who constitutes it by abstrac­ tion, and as though threatened by an infinite regress. (Levi-Strauss 1962: 340)

What process of abstraction is this? And abstraction from what? LeviStrauss explains on the same page: Each episode in a revolution or a war breaks down into a multitude of individual psychic movements: each of these movements translates unconscious evolutions, and the latter break down into cerebral, hormonal or nervous phenomena, referring back in turn to the physical or chemical order of things. (Levi-Strauss 1962: 340)

This must rate as one of the worst arguments ever mounted in the philosophy of history. On this reasoning, no one can avoid the guillotine of abstraction. The passer-by who tells you that he has just witnessed a road accident is no less guilty than the historian of abstracting from ‘a multitude of individual psychic movements’. If so, so be it. But that is what we otherwise call noticing what goes on around you. To appreciate how far both Levi-Strauss and Barthes are from developing a genuinely structuralist critique of history it is necessary to go back to Saussure. One reason why the Cours de linguistique generate is a key text in philosophy of language is that it contains the first published attack antedating Wittgenstein’s —on the surrogationist notion that a language has as its basis a ‘nomenclature’ of some kind; that is, a list of correlations between names and what they name. (E.g. arbre correlates with ‘tree’, cheval with ‘horse’, etc.) Such nomenclatorial correlations, Saussure points out, can be given two interpretations. On one interpretation the correlation is between a word and a thing or things, while on another interpretation it is between a word and an idea or ideas. Saussure recognizes, in short, what I have been calling the difference between reocentric and psychocentric semantics. In the former case arbre, cheval, etc. are envisaged as standing for some category of ‘real’ items in the physical world (trees, horses, etc.), whereas in the latter case they stand for the idea of a tree, horse, etc. Saussure rejects both interpretations. The reasons for Saussure’s rejection, however, are somewhat different in the two cases. Saussure is a ‘mentalist’. He has no objection to the contention that a linguistic sign effects a link between an idea and a sound pattern (where both are understood to be items located in the mind), but he refuses to accept that there is any necessary connexion in this link or that there is any direct link between a sound pattern and some object or objects in the external world. What he has against both the reocentric and

H I S T O R I C I S M A ND L I N G U I S T I C S

129

the psychocentric versions of surrogationism is that in both cases each correlation between the word and what it stands for is construed as a separate one-to-one pairing established completely independently of any other. (Thus the link between arbre and ‘tree’ is quite independent of the link between cheval and ‘horse’: the two are not connected in any way.) For Saussure, this betokens a fundamental failure to understand what a language is. A language (langue) - as distinct from speech {parole) - is not a random aggregate of independently existing linguistic signs, but a structured whole. None of its parts (least of all forms and meanings) exist except as parts of that whole and the way it is internally structured. It is in this sense that Saussure is usually described as a ‘structuralist’ (which is not, however, a term he ever applied to himself). In Saussurean linguistics, it is axiomatic that the structure of the whole determines the parts, and not vice versa. Thus, for Saussure, although words and what they mean are mental items, the mind of the individual has no control over them: they are known to the speaker only as parts of a structured system that cannot be altered at will. This is where Saussure differs fundamentally from a psychocentric theorist such as Locke, who attributes to the individual speaker the responsibility for ‘collecting’ the ideas attached to each and every word uttered. Although verbal signs are, in Saussure’s terminology, ‘arbitrary’, neither the form nor the meaning is arbitrarily chosen by the speaker. The result is a theory of language according to which a message has to be conveyed by opting for a particular system {langue) in which to express it, just as a financial payment has to be executed by choosing a currency in which to pay it. The system and its structure cannot be modified ad hoc to meet the needs or wishes of particular users, because any alteration of one part would automatically have repercussions on the whole. In any case, the notion that such ad hoc alterations are possible is, from a Saussurean viewpoint, as inadmissible as supposing that one person could, by his or her own financial efforts, unilaterally alter the fact that a hundred cents make a dollar. A language is thus conceptualized as a holistic network of linguistic signs, which do not exist except in virtue of their values within it. The value {valeur) depends solely on contrast with co-existing signs. A language is not a set of surrogates for anything given in advance, or for anything ‘outside’ itself, but a system of internally constituted values. The important point to grasp is that Saussurean structuralism does not reject any o f the basic assumptions o f the Western language myth. What it does is impose a very rigorous and untraditional interpretation of the

130

THE L I N G U I S T I C S OF HI S TORY

notion of a fixed code. Strictly speaking, every time a linguistic change occurs, however minimal, the result is a new language, because a new holistic structure results. The relevance of this theoretical point has profound implications for the historian. The other key idea of Saussurean structuralism that must be taken into account is the distinction between ‘synchronic’ and ‘diachronic’ linguis­ tics (Saussure 1922: 114—40). For here time enters the theoretical picture in a particularly awkward and problematic way. It is difficult to reconcile the holistic conception of linguistic structure outlined above with the traditional notion of linguistic change. According to the latter, the mere fact that new words, pronunciations and constructions gradually appear, while old words, pronunciations and constructions gradually die out, does not invalidate the notion that we are still dealing with ‘the same language’. On the contrary, even if Shakespeare’s English is sometimes hard to follow today, few audiences will doubt that it is still ‘English’ (doubtless because they can still ‘understand’ a great deal of it, or think they can). In short, ‘a language’ is traditionally conceptualized as open to change, at least up to the point where further change blocks comprehension. Thus, to take one case from which Saussure draws many examples, most linguists do not contest that French evolved by a continuous and long-drawn-out transition from the Latin originally spoken in Gaul by Roman invaders (Ewert 1933; Pope 1934). However, a knowledge of French manifestly does not suffice to read a Latin text. Nor does a knowledge of Latin suffice to read a French text. So although Latin is acknowledged as the historical ‘ancestor’ of French, and although many individual changes can be identified and even dated, it still remains unclear at what stage a slowly evolving Latin ‘became’ French, if it ever did. The general problem of cumulative change is well known to philoso­ phers under the rubric ‘sorites paradox’. It is traditionally posed in terms of the question of how many grains of sand make a heap. But this also has a direct chronological parallel: a person does not grow ‘old’ overnight, but as one day succeeds another, eventually youth gives way to old age. The question of French succeeding Latin is a typical sorites enigma. The structuralist way of dealing with the sorites paradox is ruthless and uncompromising: any structural change inaugurates - brings into existence - a new structure. {How a change is brought about is a quite different problem; but that is not the issue here, where we are considering the effects of change.) So the structuralist position is the polar opposite of what I will call the ‘evolutionist’ position, which allows that a language

H I S T O R I C I S M AND L I N G U I S T I C S

131

may accommodate an indefinite number of small linguistic changes over a period of time but nevertheless remain itself. In terms of the sorites paradox, this is equivalent to maintaining that there is no definite point at which grains of sand become a heap. Two grains of sand are already a small heap and eventually a thousand grains of sand are a big heap. But it is the same heap (provided the addition of new grains of sand has been continuous). Perhaps it will be thought that I am inventing an ‘evolutionist’ position that no linguist would ever dream of maintaining. On the contrary: here are a few examples. French is ‘merely the modern form of the spoken Latin introduced into Gaul’ (Ewert 1933: 2); or ‘une des form es modernes de Fidiome des anciens Romains’ (Wartburg 1934: 4); or V est du latin, mais profondem ent transformed (Dauzat 1944: 5); or, more bluntly still, ‘nothing other than Modern Latin’ (McWhorter 2002: 17-18). Anyone who can believe that has been totally bemused by the sorites paradox, quite apart from being very unobservant. (For Latin still survives - written, read and even spoken - in the present century, albeit in straitened circumstances which bear little resemblance to those which obtained when Latin was the common tongue of Europe.) Where does a structuralist theory of language leave the traditional historian? Perhaps the first point to be noted is that no structuralist tenet makes it impossible for the historian to refer to and describe past events as such, provided the language the historian uses has the linguistic signs required. However - the structuralist will say - what we are not entitled to assume is that the values of the signs deployed for that purpose in the historian’s present account reflect the values of the signs found in any source documents, unless indeed those documents represent the same etat de langue. This caveat echoes Humboldt’s worry about the language of the historian, voiced in his paper to the Prussian Academy in 1821. But whereas Humboldt’s misgivings were expressed only in the vaguest of terms, and lacked any theoretical backing, Saussure’s holistic analysis of la langue provides a very precise explanation of why values cannot be ‘transferred’ from an earlier linguistic system to a later one. The historian may, of course, be mistaken about this, or even convinced that such words as revolution, monarchy or clergy mean exactly the same today as they did in the sixteenth century. From a structuralist point of view, any such conviction would be a historical illusion. Or, more precisely, a double deception: for it would presumably be based on the fact that certain words in a sixteenth-century text looked recognizably like certain modern forms. But this supposed resemblance itself conceals the fact that, structurally,

132

THE L I N G U I S T I C S OF HI S TORY

actual spelling or pronunciation is no indication of valeur. Between one etat de langue and the next, there is by definition no identity of forms or meanings. In short, whether the historian is aware o f it or not, the semantics of any historical account is the semantics of one and only one etat de langue. It cannot somehow merge lexical values across a diachronic divide. The structuralist analysis of language thus places the traditional historian in a dilemma. To reject that analysis out of hand would either be tantamount to reaffirming some kind of faith in surrogationism, which is patently inadequate to provide the semantic support that traditional historical accounts need, or else it would imply acceptance of some other philosophy of language that historians find more congenial. Both LeviStrauss and Barthes fall into this no-man’s land. Neither is willing to revert to a pre-Saussurean view of language; but neither is willing to accept the most radical implications of Saussurean thinking. In philosophy of history, the thinker who comes closest to taking a genuinely structuralist position is Colling wood. This does not seem to be due to the influence of Saussure, since in Collingwood’s most detailed discussion of language, Chapter XI of his book The Principles o f Art (1938), Saussure is not even mentioned. Moreover, in that discussion Collingwood is concerned to establish a distinction to which Saussure pays no attention at all: the difference between what Collingwood calls ‘intellectualized’ language and what he regards as the more basic function of words, which is to express emotions. Nevertheless, as well as criticizing such surrogationist philosophers as Hobbes and Locke, and attacking the historical fallacy of ‘substantialism’, which is itself a surrogationist notion, Collingwood goes further. In The Principles o f Art, he argues that it is a mistake to think that we can trace the history of Western art as far back as Classical Greece. The Greeks, he maintains, had no concept of art because they had no word for it. This, although Collingwood shows no sign of realizing it, is exactly what Saussure might have argued. According to Saussure, one of the cardinal errors in traditional Western thinking about language is to suppose that our language merely supplies words for ideas already established. On the contrary, claims Saussure, ‘in itself, thought is like a swirling cloud, where no shape is intrinsically determinate. No ideas are established in advance, and nothing is distinct, before the introduction of la langue’ (Saussure 1922: 155). That is to say, the absence of some form of linguistic expression is not a merely contingent or fortuitous omission, as surrogationists had traditionally assumed. Within a structuralist frame­ work, the word and the idea are head and tail of a single coin: one cannot

H I S T O R I C I S M A ND L I N G U I S T I C S

133

exist without the other, and neither can exist without the support provided by the entire linguistic system. In the case of Greek ‘art’, Collingwood’s linguistic case is based on the Greek word techne, which was commonly applied to a wide variety of forms of craft and expertise, from sculpture and architecture to hus­ bandry and medicine. From this linguistic evidence Collingwood draws the conclusion, which Saussure would certainly have endorsed, that the Greeks did not distinguish any particular group of skills or products that correspond to the modern European term art as used by ‘art historians’. A reocentric art historian, it need hardly be said, would see no difficulty here, since the actual Greek paintings, statues and other works that art historians discuss existed and in many cases still survive: so their identification poses no great problem. But this would be to miss Collingwood’s point, which is that the Greeks themselves did not regard these objects as art. Collingwood goes on to argue that it is naive to suppose we can get round the difficulty by going directly to Plato and other Greek sources who tackle aesthetic questions. For the way we read Plato is coloured by our assumption that he too is discussing ‘art’ when in fact he is not (Collingwood 1938: 6). A good example of such misunderstandings is Plato’s alleged ‘banishment of the artists’ from his ideal state, much discussed by modern commentators. This, as Collingwood points out, is a myth. Plato had nothing against ‘art’: his objections were aimed at one type of poetry (Collingwood 1938: 46ff.). Thus if Collingwood is right, surrogational historians who write whole books on the subject of Greek art and its development are engaged, wittingly or unwittingly, in a historical deception. Either they are confusing history with mere antecedence, which would be like treating the horse-drawn carriage as an early form of motor-car; or else they are crudely retrojeering a modern aesthetic into a culture where it has no place. If this holds for the history of art, will it not also hold in respect of all other branches of history where a parallel linguistic problem arises? Collingwood does not pursue the question. This is another indication that his thinking is only superficially structuralist. From a Saussurean point of view, on the other hand, there is no reason at all for treating the concept ‘art’ as a special case, or the Greek term techne as an isolated example. That would run counter to the whole logic of structuralism. Although Saussure himself never developed structural linguistics into a structuralist philosophy of history, it is not difficult to see how that might be done, or what the main thesis of such a philosophy would be. Given

134

THE L I N G U I S T I C S OF HI S TORY

that every etat de langue defines a unique conceptual system, it would follow that the historian’s task becomes possible only in cases where a single etat de langue embraces all the terms and concepts that are deployed in the historian’s account. It falls to the historian, as a propaedeutical requirement, to establish that this is so. For it cannot be taken for granted. That is to say, it cannot be taken for granted that the account does not just describe the past as if it were an extension of the present. (Rather as some medieval romances depicted the heroes of antiquity as feudal knights, or some productions of Shakespeare present Hamlet as wearing a dinner jacket and bow tie.) Reluctance to attempt any such demonstration is what partly explains the popularity among historians of the proposition that the language of history is just ‘ordinary language’, since that lets them off the hook, or at least appears to on first inspection. So it is to the question of history and ‘ordinary language’ that I next turn. Before leaving the topic of historical linguistics, however, it is worth pointing out that the bona fid es of the subject came under sceptical scrutiny as early as the late 1920s. At least, that is how I interpret the observations of Voloshinov (or whoever wrote Marksizm i filosofija jazyka). The author in question notes: abstract objectivism supplies the grounds for the reification of the linguistic form, for its becoming an element supposedly extractable in actuality and supposedly capable of an isolated, historical existence of its own. [. . .] History of language, then, amounts to the history of separate linguistic forms (phonetic, morphological, or other) that undergo development despite the system as a whole and apart from concrete utterances. (Voloshinov 1929: 79)

To spell out the implications more clearly, what is being suggested here is that in order to conceive of ‘a language’ as having a history, we have to adopt a certain theoretical viewpoint. That is a viewpoint which already legitimizes such reifications as the sounds, the words, the sentences, all tracing their own individual paths through time. Voloshinov’s penetrating remark anticipates the integrationist critique of diachrony, to which I shall return in Chapter 6. On surveying the historification of language studies that took place in the nineteenth century at the hands of the comparativists and the Neogrammarians one can hardly doubt that it is here (and not in structuralism or postmodernism) that are to be found the roots of the much maligned relativism that was later to cause such consternation and controversy among historians. Once Bopp, Grimm and Rask,

H I S T O R I C I S M A ND L I N G U I S T I C S

135

together with their colleagues and successors, had thoroughly historified academic conceptions of language, and these conceptions had been incorporated into the agenda of university courses in the humanities, it was no longer possible to treat language as a neutral, transparent medium in which facts about the past could be objectively reported. There was no going back to a position where the same Tacts’ could be stated in any language you liked, and at any time: languages could no longer be seen as providing alternative but equivalent vocabularies, or even stable vocabularies at all. On the contrary, the message from the linguists was that linguistic stability is an illusion. (But the message took a long time to sink in. Not until Saussure did anyone spell out the full theoretical implications.) Relativism is defined by Marwick (as applied to history) as: the view that there are no objective standards: in each age and culture history will be written differently, and, indeed, even within one culture different groups and different individuals will produce different types of history. (Marwick 2001: 293)

For good measure, he adds: ‘Marxists and postmodernists are relativ­ ists.’ The anti-relativist position (that Marwick personally prefers) claims that ‘there are consistent professional standards against which all history can be measured’ (Marwick 2001: 293). What sense it makes to apply to Herodotus or Livy the ‘professional standards’ that one might apply to Ranke or Toynbee Marwick never explains. Nor where these universal standards come from if they are not merely retrospective consecrations of criteria that Marwick and other modern historians would like to see applied. Marwick seems never to have taken on board what happened in the nineteenth century when linguists began to see the implications of the fact that languages themselves do not somehow fall outside the purview of history. For once it is accepted that languages themselves are historical products, subject to incessant fluctua­ tions of which the results become apparent only in hindsight, it becomes impossible to detach any given historical account from the language in which it was couched at the time and place when that particular language occupied the historical niche that it did. It does not follow at all that ‘there are no objective standards’; but merely that statements of standards have to be relativized to the language in which they are formulated, and that language in turn seen as reflecting the assumptions of a particular moment in history: specifically, of the linguistic community whose communicational purposes that language served. It is easy to see why that sounded like heresy to those historians

136

THE L I N G U I S T I C S OF HI S TORY

who thought their discipline had achieved, or was in principle capable of achieving, a definitive, God’s-eye view (sometimes called ‘scientific’) of the past. In short, it was not postmodernism but nineteenth-century linguistics that pulled out the historical rug from under their feet.

CHAPTER FIVE

History and ordinary language

Here we come back to the linguistic responsibilities of the historian. I am not suggesting that the ‘ordinary-language’ defence of history is an academic reaction against structuralism. (The original advocates of this defence seem to have had little or no grasp of linguistic structuralism or its implications. Nevertheless, their proclivities may have been affected by a certain intellectual climate of opinion to which they were, consciously or unconsciously, responding.) So let us begin by trying to situate the problem in very general terms. It is possible to claim that the historian has no special linguistic obligations over and above those incurred by any other writer, and that perfectly adequate accounts of what happened in the past can be given in the language current in the historian’s own day. It has even been argued by some philosophers that one of the historian’s strengths resides in keeping to ordinary language and avoiding technical jargon. This view is often associated with dissent from the proposition that history is a science. I shall not attempt to survey everything that historians or philosophers of history have said or implied about ordinary language, but I shall consider in detail the cases presented by two of the most energetic advocates of the ‘ordinary-language’ defence of history. In The Nature o f H istorical Explanation (1952), Patrick Gardiner writes: Language used for certain purposes requires a conceptual apparatus different from that required by language used for other purposes. Historical concepts like ‘revolution’ were not evolved to meet the descriptive needs of an expanding scientific system; they were developed to meet the requirements of those who wanted a short means of referring to a fairly common instance of the behaviour of human beings in society. The historian is concerned with human activities, and he is principally interested in those activities in so far as they have been found related to one another in social groups; to speak conveniently about the multitude of interrelated activities with which he finds himself confronted he uses short-hand

137

138

THE L I N G U I S T I C S OF HI S TORY

terms like ‘revolution’. When asked what such terms mean he points to less complicated events, like mob-riots, changes in government, and so forth, and says that it is to the occurrence of events of this type that he refers when he says that a revolution has happened. What he does not do is lay it down explicitly that any single event or set of events must be present if a social movement is to be called a revolution. He does not, to take a fantastic example, make it a necessary truthcondition of the statement ‘a revolution has occurred’ that at least 40 per cent, of the total male population should have appeared armed in the streets, shouting subversive slogans. To restrict the use of the term in this manner would be equivalent to eliminating its utility as a means of referring to a large number of social movements of varying types occurring at different times. And this is precisely what he wishes it to be able to do. (Gardiner 1952: 60)

Here we have a quite overt and specific linguistics of history being expounded. It seems evident that the hypothetical historian on whose behalf Gardiner produces this semantic apologia must be a surrogationist. For when asked what he means by the term revolution, instead of saying ‘fundamental change’, or ‘overthrow of established order’, or giving whatever other definition seems apposite, his preferred semantic tech­ nique is to ‘point to’ events such as mob-riots and changes of government. Why he thinks that a riot or a change of government is per se a ‘less complicated’ event than a revolution is not clear, nor how ‘pointing to’ simpler events throws light on the meaning of terms for more complicated events, nor in what sense the word revolution is ‘short-hand’ for anything. One begins to wonder about this historian and his view of language. The examples seem oddly chosen; for neither a riot nor a change of govern­ ment is in itself a revolution, and although both might in certain cases occur as pivotal events in a revolution, neither is a necessary condition. Furthermore, the question-begging coda ‘and so forth’ leaves it entirely unclear what else might count on a par with these. In short, Gardiner’s historian, when asked to clarify his use of the term revolution, evades the question and merely gestures in the direction of possible features of revolutions. (Surprisingly, he does not even list examples of past revolu­ tions, such as the French Revolution, or the Russian Revolution, or the Revolution of 1688.) This is like explaining one’s use of the term breakfast by pointing to packets of cornflakes, cups of coffee, and then saying ‘and similar things, but not necessarily any of these’. What the passage quoted reveals, in the first place, is not a description of any actual historian’s way of dealing with terminological queries, but rather a philosopher’s theory o f ordinary language. Intrinsic to that theory is the contrast between ordinary language and the language of science, as becomes apparent when Gardiner continues:

HI S TOR Y A ND O R DI NA R Y L A N G U A G E

139

The crux of the distinction between the historian and the scientist is as follows. The scientist frames hypotheses of precision and wide generality by a continual refining away of irrelevant factors. Things are otherwise with the historian. His aim is to talk about what happened on particular occasions in all its variety, all its richness, and his terminology is adapted to this object. That is the reason why terms like ‘revolution’ are left so vague and so open. They are accommodating terms, able to cover a vast number of events falling within an indefinitely circumscribed range. (Gardiner 1952: 60-1)

Here, however, we have what amounts to implicit self-contradiction. Gardiner’s account of the historian’s task resounds with unmistakable echoes of Aristotle. But if it is an accurate account - and the historian’s aim is not to generalize but to talk specifically about ‘what happened on particular occasions’ - it is hard to see why the historian needs terms like revolution at all. They would frustrate the historian’s purpose. For such terms, according to Gardiner’s linguistic analysis, are so vague as to blur any differences between particular events. To be told ‘a revolution has occurred’ is to be told nothing specific about what happened. Rather, the specifics of what happened are hidden behind a generalization. This is the opposite of what Gardiner’s historian is supposed to be doing. Whereas if instead we were told exactly what took place, we should then be able to judge for ourselves whether revolution was an appropriate term to apply. How does Gardiner get himself into this predicament? It is not because he happens to have chosen a bad example, but because his conception of the difference between ordinary language and scientific language is at fault. According to Gardiner’s philosophy of language, it seems that whereas the language of science is based on explicit definitions, which scientists continually refine, the language of the ordinary citizen is not. It is, in other words, futile to ask for an explicit definition of a term like revolution, because that is not how ordinary language works. In ordinary language, terms are not defined; they are just applied to individual cases, and the basis of their application may vary from case to case ad infinitum. What Gardiner does not seem to appreciate is that this makes the historian’s task harder, not easier. For it is tantamount to denying that ordinary language supplies any fixed code on which the historian and the historian’s readers can rely for purposes of communication. Whereas, I suspect, Gardiner intended the opposite: allowing the historian to use ordinary language and fend off calls for definitions was designed to exempt the historian from the rigorous linguistic constraints imposed upon scientists. In brief, Gardiner’s philosophy of language is a restate­ ment of Locke’s, with the curious twist that in Gardiner’s case the imprecisions of ordinary language are treated as virtues. As the basis for a

140

THE L I N G U I S T I C S OF HI S TORY

philosophy of history this is odd, to say the least. For it is hard to find examples of traditional historians who vaunt the vagueness of their statements as an excellence. On the contrary, those who claim to be telling the truth about the past do not regard imprecision as commendable in the least. Nor can this objection be met by the kind of reductio ad absurdum that Gardiner proffers. Calls for a historian who uses it to define a term like revolution are not calls for statistics about the number of armed men on the streets. Quite the reverse. If the historian can tell us how many armed men were on the streets and what subversive slogans they shouted, and what the outcome of this disturbance was, then the need for deploying any such term as revolution is much diminished; provided - and the proviso is important - Gardiner is right to treat the historian as a scholar who eschews generalization but is concerned with ‘what happened on partic­ ular occasions’. That many historians have been concerned with formulating generaliz­ ations, and regarded what happened on particular occasions chiefly of interest as grounds for formulating those generalizations, does not give Gardiner a moment’s pause. He sails on regardless. He concedes the point, but insists that these historical generalizations do not have ‘the status of scientific laws’. Historical generalizations have not the status of scientific laws, but this does not mean that we must deny their existence altogether and look for a different method of understanding history. Nor is it the case that in history we suddenly find ourselves faced with peculiar intractable entities that refuse to be ‘fitted in’. The difficulty in question is a reflection o f the language o f historical description, and once this is recognized it is seen no longer to be a difficulty but an entirely natural outcome o f the usage o f this language. (Gardiner 1952: 59. M y italics.)

In short, in order to understand history we first have to understand the language of history, and then we shall see how and why what the historian is doing is not to be confused with what the scientist is doing. So here we are back to Linguistic Square One. The historian’s language, once recognized for what it is, is allegedly seen to be not answerable to requirements imposed on the language of science. So what status do generalizations have in the language of history? All Gardiner is willing to tell us is that ‘they function frequently as guides to understanding’ (Gardiner 1952: 61. Gardiner’s italics). Guides to understanding what? Guides to understanding the facts? If so, it is difficult to see how they function differently from generalizations in science. The point I wish to emphasize here is not the tortuous way Gardiner

HI S T OR Y AND OR D I NA R Y L A N G U A G E

141

handles his own argument. More important for my present purposes is to observe how reliance on an inadequate philosophy of language traps him into taking an indefensible position in philosophy of history. It is worth noting that at no point does Gardiner himself define what he means by the expression ordinary language. This may be to be expected in ‘ordinary language philosophy’. But early in his book he does not hesitate to equate ordinary language with the language of ‘common sense’. As common sense may be contrasted with science, so ordinary language may be contrasted with the language used by scientists. The various branches of scientific inquiry represent investigations into different fields of phenomena. The scientist aims to go beyond the crude rough-and-ready generalizations that are a feature of common sense. (Gardiner 1952: 7)

It is this, according to Gardiner, that accounts for the divorce between ordinary language and the language of science: And while it is, of course, true that all science must proceed from the data of common sense, from the things of ordinary perceptual experience and from the fragmentary generalizations of commonsense inductive procedure, yet, as the observational instruments and experimental techniques of each branch of scientific inquiry evolve, so the original correlations are replaced by others. And, reflecting these developments, new descriptive methods become necessary, adapted to the accurate expression of the new discoveries and to their systematic representation. The language of common sense proves too cumbersome and imprecise for this purpose, the advanced sciences, for example, finding in mathematics the vehicle for the formulation of their theories and hypotheses. (Gardiner 1952: 8)

Here it becomes apparent that Gardiner’s theory of ordinary language is subservient to his more general philosophy of ‘progress’ in science. Specifically, we find mathematics taking over from ‘the language of common sense’ in a manner which suggests that - in some way not clearly explained - mathematics and common sense are at odds. No less puzzling is his claim that ‘of course’ it is true that science ‘must’ proceed from what he calls ‘the data of common sense’ (not a phrase often encountered either in ordinary language or in the language of science). To treat the deliverances of common sense as ‘data’ is to envisage a scientific programme which aims at developing a theory of everyday experience. But that, many scientists will be quick to point out, is not their objective at all. We do not send missions to the moon in order to verify the commonsense view that the moon is up there. Nor does a theory of photosynthesis ‘start from’ the commonsense observation that many plants are green. In short, even as a component of a more general

142

THE L I N G U I S T I C S OF HI S TORY

philosophy of science, Gardiner’s account of the difference between ordinary language and the language of science lacks credibility. As to whether statements in ordinary language are translatable into scientific language, or vice versa, and if so under what conditions, Gardiner is equivocal. He writes: Scientists and ordinary people often seem to be talking about different things when they produce explanations of events. Where the ordinary person talks about tables and chairs, the scientist discusses neutrons and electrons; where the ordinary person talks about minds and ‘nerves’, the psycho-analyst discusses the unconscious and traumata. And the suspicion arises that the scientist is discussing entities lying upon a different level of existence, that he is describing the inhabitants of a rarified world of which he is a privileged observer. (Gardiner 1952: 17-18)

But this, according to Gardiner, is something of an illusion. In many cases scientists and ‘ordinary people’ are ‘in fact talking about the same world in different ways’ (Gardiner 1952: 18). How Gardiner can be sure of this he does not say. Perhaps it is a reassurance provided by his common sense. In any case, there is evidently no room in Gardiner’s philosophy of language for the thesis that the use of different languages is eo ipso an indication that there are different worlds to be talked about. It might seem, therefore, that in Gardiner’s view ordinary language and the language of science are commensurable, since in the end the former is the basis of the latter. To take an elementary example, consider [. . .] the difference between a medical explanation of why a person has caught pneumonia and a commonsense one. The ordinary man may explain it by saying that the sufferer stood too long in the cold and leave it at that. But he knows also that there are many instances of people standing in the cold and not catching pneumonia. His generalization is, therefore, of the form: other things being equal, people who stand too long in the cold contract pneumonia. And the question remains whether other things are in fact equal in the particular case. The medical specialist, on the other hand, is in a position to provide a far wider and more comprehensive account of the conditions favouring the contraction of pneumonia in a given instance; and the possibility of exceptions or counter-cases to his diagnoses has been correspondingly reduced. Yet the differ­ ence in the position of the ordinary man and the specialist remains essentially one of degree. For the latter can never be certain that in the case with which he is dealing he has taken into account all the relevant factors, that there may not have been hidden peculiarities of which he has failed to take account; and the ordinary man [. . .] is not ‘wrong’ in his explanation - he is not denying anything the specialist may assert regarding the action of bacteria upon the bloodstream and so forth. He is merely not in a position to make explicit the type of factors that are made explicit

HI S TOR Y A ND OR D I NA R Y L A N G U A G E

143

when the specialist analyses the case; and he therefore leaves room for the tabulation of these factors by inserting a wide ceteris paribus clause. (Gardiner 1952: 16-17)

Gardiner’s example dodges the linguistic issue of communication altogether. It takes no account of the commonsense observation that if you do not understand what your doctor says about bacteria then that diagnosis tells you nothing about the world in which your pneumonia is causing you distress. To say that the doctor’s world and the patient’s world are one and the same is already to take the doctor’s point of view. It assumes that the expert also knows, qua lay person, what it is like to be ill. But this does not hold the other way round. There is an asymmetry between the doctor’s language and the patient’s language that cannot be bridged (except fo r surrogationists) by reference to ‘the same world’ common to both. These dubious assumptions about ordinary language and its relationship to ‘the world’ are also central to Gardiner’s view of history in quite a different respect. Teaching in Oxford during the heyday of ‘ordinary language philosophy’, Gardiner favours settling philosophical questions wherever possible by appeal to what one would ‘ordinarily’ say about the matter. The role of the philosopher, as he conceives it, is to dispel conceptual confusions, in particular those having their roots in language. His analysis of historical explanation is thus inscribed in a broader therapeutic programme. Those sceptical about any such programme will point out that it always begins at an unconvincing level of abstraction. In Gardiner’s discussion we are constantly referred to such idealizations as ‘ordinary life’, ‘everyday life’, ‘daily life’, ‘normal life’, ‘common life’, ‘common experience’, and to such shadowy agents as ‘the ordinary man’, ‘the ordinary person’ and ‘normal people’. The collectivity of these, presumably, constitutes the collectivity of those who follow what Gardiner calls ‘normal usage’, take part in ‘ordinary discourse’, and give ‘ordinary explanations’. All in ‘ordinary language’, based on ‘common sense’. Explaining what he takes the term common sense to mean (presumably in ‘normal usage’), Gardiner writes: ‘ “Common sense” may be conveniently regarded as a name for those skills used by human beings in making their way about the world’ (Gardiner 1952: 6). The nomenclaturist formulation is noteworthy: common sense is ‘the name of something. But even more noteworthy is that Gardiner says nothing about the problem that this account immediately raises; namely, that what

144

THE L I N G U I S T I C S OF HI S TOR Y

counts as common sense for some people in some circumstances may not be common sense at all for other people in other circumstances. And if ‘ordinary language’ is, as he claims, the language of common sense, then it would follow that ordinary language too varies accordingly. As soon as this is realized, confidence in ordinary language as a reliable or consistent guide to anything begins to seem questionable. Unless, that is, one can bring oneself to believe in a world-wide layer of common sense that somehow underlies all languages and all human activities. I do not think Gardiner is asking us to believe in anything so implausible. There is a simpler explanation. What probably lurks behind his disinclination to deal with the problem of linguistic variation is a tacit equation o f‘ordinary language’ with ‘standard language’. The latter is not a term Gardiner uses. Nevertheless, his constant insistence on what he describes as ‘normal’ usage has prescriptive overtones. He speaks o f‘rules for the use’ of words and of whether it is ‘correct’ to say this or that, but these remarks sit oddly with his insistence elsewhere that ordinary language, unlike the language of science, is irremediably vague and riddled with uncertainties. The basic reason for this tension is that, where questions of meaning directly affect his argument, Gardiner reserves the right to tell us ex cathedra what the expressions of ordinary language mean, and whether it would be ‘nonsense’ to say such-and-such. In short, like many other ordinary-language philosophers, he wants to be judge and jury when his own linguistic case comes to court. To take two specific examples, Gardiner discusses in some detail the statements ‘Napoleon’s actions were motivated by a will to power’ and ‘Richelieu’s policy was guided throughout by his aim to establish a centralized monarchy’. In both cases motives are attributed to the person in question. Gardiner seeks to establish that although statements of this type give explanations of behaviour, they are not causal explanations. He claims that here we have examples of metaphor, and that metaphor is a conspicuous feature of ordinary language. Ordinary language, for reasons both of economy and force, is packed with elliptical and metaphorical expressions. This is particularly true of expressions that are said to describe the ‘workings of the mind’. The phrase ‘the workings of the mind’ itself suggests a machine or engine. And we also speak of people being ‘occupied by certain thoughts’, ‘guided by certain considerations’, ‘governed by certain desires’, and ‘driven by certain impulses’. [. . .] It would be foolish to criticize expressions like these, to adopt towards them a tone of puritanical disapproval. From the point of view of ordinary communication they serve their purpose in a convenient and often vivid way. It is not the task of the philosopher to alter people’s linguistic habits or whittle away the richness and

HI S T OR Y A ND OR D I NA R Y L A N G U A G E

145

variety of everyday speech. But it is at least a very important part of his task to prevent the occurrence of philosophical confusions; and this can often be achieved by underlining metaphors when they occur and by pointing to the logical limits of figurative expressions. (Gardiner 1952: 120-1)

There is no indication here of what particular theory of metaphor Gardiner espouses, but he seems committed to some doctrine of a distinction between ‘literal’ and ‘non-literal’ meaning. Perhaps Aristotle’s doctrine, but there is no way of telling. Nor is it clear what is being claimed by saying that figurative expressions have ‘logical limits’. These seem to be something over and above the conventions of ordinary usage; but who or what imposes such limits is left unexplained. Gardiner proceeds to argue that to say someone is ‘motivated by a will to power’ could not possibly mean that the ‘will to power’ was a force that drove the individual ‘like the force that drives a locomotive’. But, pace Gardiner, it is hard to see why not. That would appear prima fa cie to be quite a plausible interpretation. According to Gardiner this would transgress ‘the boundaries within which it is safe to use the expression’. Safe for whom? And whose boundaries? Gardiner’s objection is evidently based on his own conviction that in our mental world, and in the mental world of such historical figures as Napoleon, there are or were, as a matter of fact, no such forces operative. But that is a different matter from what a language allows us to say: it relates to Gardiner’s personal beliefs about psychology, just as objecting to talk about ‘divine providence’ might be based on a refusal to accept the existence of a divinity. Nevertheless, people do talk about ‘divine providence’ and what Gardiner calls ordinary language allows them to do so. According to Gardiner, saying that Napoleon was motivated by a will to power involves a usage ‘considered to be correct’ if the historian can give instances of Napoleon acting in ways calculated to further his personal advancement, or produce supporting excerpts from Napoleon’s diaries or letters. But here Gardiner conflates a question o f‘correct’ usage with the evidence for claiming a statement to be true. Regardless of evidence, the historian can say this about Napoleon and few will object that the statement is linguistically improper. The tendency to confuse meaning with truth is endemic in surrogational thinking about language. Similarly, Gardiner rules out any interpretation of the metaphorical statement about Richelieu’s aim which suggests that in Richelieu’s mind there was some ongoing process that continued throughout part of his career and could be called ‘the aim of creating a centralized French

146

THE L I N G U I S T I C S OF HI S TORY

monarchy’. Such an interpretation, declares Gardiner, would be ‘wrong’ (Gardiner 1952: 122). But how we can tell this without knowing some­ thing more about the historian’s view of mental processes it is hard to see, or on what ground choosing to call such a process ‘the aim of creating a centralized French monarchy’ could be condemned as linguistically incorrect. Again, it is like the atheist accusing the preacher of bad English for saying ‘God loves us’. ‘Ordinary language’, in short, along with ‘normal usage’, ‘common sense’ and the rest, are for Gardiner convenient cover terms under which the philosopher of history is at liberty to present his own convictions about the empirical assumptions that it is legitimate for the historian to make. It is interesting to compare this with the position taken by Isaiah Berlin in H istorical Inevitability (1954). Here Berlin mounts a powerful and sustained attack on various forms of relativism and determinism in historical thinking. The first way in which ‘ordinary language’ enters his argument is that, according to Berlin, it harbours and perpetuates many potentially misleading ideas concerning historical processes. In particular, Berlin sees ordinary language as conducive to teleological thinking about history. Teleology is not a theory, or a hypothesis, but a category or a framework in terms of which everything is, or should be, conceived and described. The influence of this attitude on the writing of history from the epic of Gilgamesh to those enjoyable games of patience which Professor Arnold Toynbee plays with the past and future of mankind - and plays with exhilarating skill and imagination - is too familiar to need emphasis. It enters, however unconsciously, into the thought and language of those who speak of the ‘rise’ and ‘fall’ of states or movements or classes or individuals as if they obeyed some irresistible rhythm, a rising or falling wave of some cosmic river, an ebb or tide in human affairs, subject to natural or super­ natural laws. (Berlin 1954: 15)

Another idiom Berlin condemns is that which represents life as some kind of play or dramatic performance: To those who use this figure, history is a piece, or succession of pieces, comical or tragical, a libretto whose heroes and villains, winners and losers, speak their lines and suffer their fate in accordance with the text conceived in terms of them, but not by them. (Berlin 1954: 15-16)

From this it is not entirely clear whether Berlin regards such idioms as the source of teleological thinking or merely as the vehicle. He writes:

HI S T OR Y AND OR D I NA R Y L A N G U A G E

147

‘Historians, journalists, ordinary men speak in these terms; they have become part and parcel of ordinary speech’ (Berlin 1954: 16). Here we may note the repetition of the word ordinary and the implication that this insistence carries. It is not the fault of historians that these ways of speaking exist. But what is culpable, in Berlin’s view, is ‘to take such metaphors and turns of phrase literally’. This develops a theme familiar to Gardiner’s readers: ordinary language as the home of metaphors. At this point Berlin adds a long apologetic footnote: I do not, of course, wish to imply that metaphors and figures of speech can be dispensed with in ordinary utterance, still less in the sciences; only that the danger of illicit ‘reification’ - the mistaking of words for things, metaphors for realities - is even greater in this sphere than is usually supposed. The most notorious cases are, of course, those of the State or the Nation, the quasi-personification of which has rightly made philosophers and even plain men uneasy or indignant for over a century. But many other words and usages offer similar dangers. Historical movements exist, and we must be allowed to call them so. Collective acts do occur; societies do rise, flourish, decay, die. Patterns, ‘atmospheres’, complex interrelationships of men or cultures are what they are, and cannot be analysed away into atomic constituents. Nevertheless, to take such expressions so literally that it becomes natural and normal to attribute to them causal properties, active powers, transcendent properties, hunger for human sacrifice, is to be fatally deceived by myths. ‘Rhythms’ in history occur, but it is a sinister symptom of one’s condition to speak of them as ‘inexorable’. Cultures possess patterns, and ages spirits; but to explain human actions as their ‘inevitable’ consequences or ex­ pressions is to be a victim of misuse of words. (Berlin 1954: 15-16)

Here at last the hidden prescriptivism that has underlain the whole discussion emerges into the open. ‘Misuse of words’ is the cardinal sin. Perhaps metaphor itself is strictly a misuse of words, as Aristotle appears to have thought. One wonders in any case how the ‘ordinary man’ or even the ordinary historian is supposed to know when words are being misused and when they are not. Doubtless they should turn to the philosopher for an answer, for the philosopher somehow seems to know immediately and without consulting anyone else. Even the philosopher, however, can offer only the vaguest of advice to the general public: There is no formula which guarantees a successful escape from either the Scylla of populating the world with imaginary powers and dominions, or the Charybdis of reducing everything to the verifiable behaviour of identifiable men and women in precisely denotable places and times. One can do no more than point to the existence of these perils; one must navigate between them as best one can. (Berlin 1954: 16)

148

THE L I N G U I S T I C S OF HI S TORY

On this note of gloom the footnote ends. But it has revealed more of Berlin’s philosophy of language than the rest of his text put together, and his position can be seen to differ significantly from Gardiner’s in various respects. Although both subscribe to some ill-defined doctrine of ‘literal meaning’, Berlin declares that metaphors cannot be ‘dispensed with’, either in ‘ordinary utterance’ or in the sciences. If so, that removes much of the difference that Gardiner insisted on between the language of science and the language of history. And presumably there must be a reason why metaphor is indispensable in both cases. Berlin does not tell us what it is, but a reasonable inference would be that metaphor is indispensable because there are no non-metaphorical terms available for certain ideas. Its linguistic function, in short, is not that of mere embellishment, as Aristotle and others had assumed. It is an essential complement to ‘literal meaning’: without it both ordinary language and the language of science would be inadequate for their own purposes. This puts quite a different complexion on matters. For if metaphor is not just an optional extra to our linguistic resources, but a necessary component, this implies quite a different relationship between words and the world. Whenever we find a scientist or a historian insisting on a metaphor, using it repeatedly and consistently, we should perhaps start from the assumption that it is thus used because nothing else will do. That, at least, would be a more charitable assumption than supposing the writer to be too lazy to find the exact non-metaphorical form of ex­ pression, or else constantly concerned to give the account a spurious vivacity or heightened colour. All this presupposes that we are in possession of criteria that enable us to detect when an expression is being used metaphorically and when it is not. But neither Gardiner nor Berlin ventures to explain what these criteria might be. So how the ‘ordinary man’ is supposed to pick his way through a text in which metaphorical and literal meanings may stand side by side, or even arm in arm, it is difficult to see. Or does being an ordinary man, accustomed to ordinary language, carry with it the ability to recognize such subtle distinctions without even being aware of them? According to Berlin, we navigate the linguistic straits as best we can. But if that is a metaphor to be taken seriously, it would imply that we need a compass or a map that Berlin declines to provide. Berlin’s troubles with metaphor do not end there. For he affirms that indeed there are historical movements ‘and we must be allowed to call them so’. By whom, one wonders, is the permission granted? Similarly with collective acts, the rise of societies, their decay and so on. But if all these processes do take place and we must be allowed to call them what

HI S T OR Y A ND OR DI NA R Y L A N G U A G E

149

they are, then in what sense are we calling them something other than what they should be called? And in that case why is permission needed? A related conundrum is how to unpack the accusation that sometimes expressions are taken too literally, or ‘so literally’ that certain undesirable conclusions seem to follow. Is taking an expression literally a matter of degree? And if so must not metaphor also be a matter of degree? And in any case, what evidence is there that attributing causal properties, active powers, etc. to historical movements and forces is actually a result of taking such words as movement and fo r ce too literally (whatever that might mean)? How could one take the word movement too literally if indeed there was a movement, or the word fo r ce if indeed a force was exerted? That neither movements nor forces need be physical many of us will readily agree. But it is gratuitous to attribute any such misconception to historians: that would indeed be to suppose that historians thought they were using the language of physics. And even more gratuitous to assume that the language of physics defines the ‘literal meanings’ of such words. Here Berlin’s doctrine of literal meaning threatens to self-destruct. The whole of Berlin’s discussion of the issue is a telling example of the surrogationist entangled in a verbal web of his own making. It brings us back to the question raised earlier. Is ordinary language merely the vehicle for what Berlin condemns as ‘illicit reifications’ and related confusions, or is it actually their source, as Bacon held when condemning idola fo r i? Presumably the latter if we attach any weight (in pounds and ounces) to Berlin’s observation that a teleological approach to history enters ‘how­ ever unconsciously’ into the minds of historians who use such terms as rise and fa ll. But the very next moment we are being told that rises and falls of states, movements, individuals, etc. do actually occu r. So it can hardly be a linguistic confusion after all to believe that they do. If there is any confusion here, it lies not with the historian nor in the historian’s language, but in the clumsy analysis that the philosopher foists upon it. Having thus muddied the linguistic waters at the outset, Berlin leaves the question of ordinary language in abeyance until later in H istorical Inevitability. There he returns to the attack on determinism by trying to show that since determinism is not assumed in the way we ‘ordinarily’ talk about the world, it cannot be given much credence in history. If the determinist hypothesis were true and adequately accounted for the actual world, there is a clear sense in which (despite all the extraordinary casuistry which has been employed to avoid this conclusion) the notion of human responsibility, as ordinarily understood would no longer apply to any actual, but only to imaginary or

150

THE L I N G U I S T I C S OF HI S TORY

conceivable, states of affairs. I do not here wish to say that determinism is necessarily false, only that we neither speak nor think as if it could be true, and that it is difficult, and perhaps impossible, to conceive what our picture of the world would be if we seriously believed it; so that to speak, as some theorists of history (and scientists with a philosophical bent) tend to do, as if one might accept the determinist hypothesis, and yet to continue to think and speak much as we do at present, is to breed intellectual confusion. (Berlin 1954: 33. M y italics.)

This is an extremely ambitious case that Gardiner never even tries to put forward. The assumption is that the way we ‘think and speak [. . .] at present’ precludes the serious adoption (although not the idle contem­ plation) of any countervailing or incompatible theory. It is an argument which is only one short step away from the Whorfian thesis that our language determines the way we view the world. The irony is that here is one form of determinism enlisted to oppose another form of determinism. Whereas the Whorfian version presupposes that we cannot ‘get outside’ the conceptual system inscribed in our native language, the determinism that Berlin is attacking presupposes exactly the opposite. What credibility does such an argument command? The claim that we cannot (in ordinary language) talk about the world in one way, while actually believing it works in quite another way, is hardly convincing. There are too many ‘ordinary language’ counterexamples. We still say that the sun rises in the east every morning, while believing that it remains where it was (relative to the Earth). We still say that night falls, while believing that nothing nocturnal descends from above. We still say that winter comes, while believing it had no journey to make. We still say that time passes quickly or slowly, while believing neither that it goes any­ where else nor that it can be slowed down or accelerated. Examples could be multiplied endlessly. All Berlin seems to have in mind when claiming that determinism conflicts with ordinary language is that ordinary language allows us to talk about ‘free will’. So it may. But perhaps ‘free will’ is just a metaphor too. Here Berlin’s previous evasiveness regarding the criteria for metaphor returns to thwart him and undermine the thesis that he wishes to champion. He digs the pit deeper for himself by conceding that a belief in free will might be a ‘necessary illusion’, but that if it is ‘it is so deep and so pervasive that it is not felt as such’. We can, he supposes, try to convince ourselves that we are systematically deluded. ‘But unless we attempt to think out the implications of this possibility, and alter our modes of thoughts and speech to allow for it accordingly, this hypothesis remains hollow’ (Berlin 1954: 34). Thus in the end Berlin puts up a linguistic barrier that it is impossible

HI S T OR Y A ND O RDI NA RY L A N G U A G E

151

for the determinist to jump over. He demands that ordinary language be radically altered before he will concede any substance to the hypothesis. It is rather like a theologian demanding that all words relating to God be eliminated from everyday speech before admitting the possibility of atheism as a viable position. Not only is it silly linguistics, but silly philosophy to boot. In Section IV of H istorical Inevitability, Berlin returns once again to the topic of ordinary language. History, he tells us, employs few, if any, concepts or categories peculiar to itself, but broadly speaking, only those of common sense, or of ordinary speech. The central concepts of history - the ways in which events or situations are ‘explained’, are shown to be connected or unconnected with one another - the use of such crucial terms as ‘because’ and ‘therefore’, ‘inevitable’ and ‘possible’ and ‘probable’, ‘surprising’ and ‘unexpected’, ‘influential’ and ‘trivial’, ‘central’ and ‘accidental’, and so forth, is much the same as that which it has in ordinary, non-technical thought and speech. As history becomes specialized, e.g. in such disciplines as the history of science or of commerce or of art, technical terms do begin to make their appearance, and to that degree something approaching, but still somewhat remotely, the natural sciences, begins to occur, and the elimination of a good many of the normal moral and psychological concepts of daily speech becomes possible and, according to some, desirable and perhaps even indispensable. But in the realm of general history, social, political, and cultural - what goes by the name of history without specific qualification - this is not so. There we explain and elucidate as we explain and elucidate in ordinary life. [. . .] All attempts to construct special sets of concepts and special techniques for history (e.g. by Marxists) have broken down because they proved sterile, for they either misdescribed - overschematized - our exper­ ience, or they were not felt to provide answers to our questions. (Berlin 1954: 51-2)

These are feeble forms of demonstration. The failure of particular attempts to provide history with its own concepts and terminology does not show that such a project is impossible of achievement. Nor does the reliance of traditional historians on traditional forms of discourse prove anything more than their reluctance to attempt something more adven­ turous. Berlin’s insistence on ‘the ordinary’ is by now becoming monotonous. It also becomes obfuscating in the absence of any attempt to explicate what is ordinary and what is not, and accompanied by an obdurate refusal to concede that ordinary language speaks no more ‘for’ or ‘against’ determinism than it speaks ‘for’ or ‘against’ socialism, satanism or any other ‘-ism’. The reader is cast in the role of‘ordinary’ reader, who should know what is ordinary from personal experience. The buck stops there, and conveniently excuses the philosopher from further historical duties.

152

THE L I N G U I S T I C S OF HI S TORY

We are made to feel stupid if we cannot work out for ourselves what because means, or therefore, and so on through as much of the vocabulary as the philosopher cares to cite. This is argument by bullying. It is also argument that presents the Western language myth in its least attractive light. There is an additional dimension to Berlin’s argument that goes beyond anything Gardiner advances. Berlin carries the ‘ordinary-language’ campaign into the enemy’s camp and complains about their use of language. Relativism, he claims, is a position sustained by ‘misuse of words’ (Berlin 1954: 68). Both relativism and determinism have adopted modes of speech that ‘have peopled the air with supernatural entities of great power’ (by talking about the ‘Collectivist Spirit’, ‘The Myth of the Twentieth Century’, ‘The Contemporary Collapse of Values’, ‘Modern Man’, ‘The Last Stage of Capitalism’, etc.). He inveighs against ‘the barbarous vocabulary of the new mythologies’, the ‘theoretical construc­ tions obscured by picturesque metaphors’ and those who ‘speak in immense, unsubstantiated images and similes’. He laments the correlative decline of the traditional ‘terminology of praise and condemnation’ and its fashionable rejection as ‘uncivilized and obscurantist’. If we take determinism seriously, then ‘the changes in our language, our moral notions, our attitudes towards one another, our views of history, of society and everything else will be too profound to be even adumbrated’ (Berlin 1954: 75). Our words - our modes of speech and thought - would be transformed in literally unimaginable ways; the notions of choice, of voluntary action, of responsibility, freedom, are so deeply embedded in our outlook, that our new life, as creatures in a world genuinely lacking these concepts, can, I should maintain, literally not be conceived by us. (Berlin 1954: 75)

As an afterthought he adds in a footnote: As for the attempt to ‘reinterpret’ these notions so as to bring them into conformity with determinism, this can be achieved only at the cost of altering their meanings beyond applicability to our normal experience. (Berlin 1954: 75)

Here we begin to sense that something much deeper lies behind Berlin’s advocacy of ordinary language than a traditionalist’s respect for the past practice of historians or even a philosopher’s insistence that history is not the right forum for rigid definitions. It is more like a profound conviction that ordinary language embodies universal truths

HI S T OR Y A ND OR D I NA R Y L A N G U A G E

153

derived from human experience; so that whatever ‘goes against’ ordinary language is automatically suspect or even marked out as a perverted doctrine. And indeed in the last pages of H istorical Inevitability this view of language surfaces quite plainly. There Berlin speaks of using words according to the standards not merely of the twentieth century, which is soon over, or of our declining capitalist society, but of the human race at all the times and in all the places in which we have known it. (Berlin 1954: 77)

So in the end it emerges that the historian is committed to ordinary language, not fa u te de mieux but because ordinary language itself enshrines and preserves the assumptions and values that mortals have to accept in all ages and in all societies if they are to make any sense at all of the world in which they live. It sounds suspiciously like Vico’s ‘mental language common to all nations’. This was supposedly a language ‘which uniformly grasps the substance of things feasible in human social life and expresses it with as many diverse modifications as these same things may have diverse aspects’ (Vico 1744: 161). For Vico, this was the historian’s way out of the problem posed by the existence of so many diverse tongues and customs: the possibility that ‘linguistic scholars will be enabled to construct a mental vocabulary common to all the various articulate languages living and dead’ (Vico 1744: 162). Vico envisaged the Idea of a Mental Dictionary for assigning meanings to all the different articulate languages, reducing them all to certain unities of ideas in substance, which, considered from various points of view, have come to be expressed by different words in each. (Vico 1744: 445)

That noble linguistic work, had scholars been able to produce it, would clearly have proved an immense benefit to the historian, an advance far more worthy of being called a ‘linguistic turn’ than the roundabout thus designated in the twentieth century. But perhaps there was no need for it after all, if, as Berlin hints, a proper understanding of ‘ordinary language’ is all that is necessary. Vox populi, vox dei. Twenty years later, Hayden White takes up the well-worn theme of ordinary language, but gives it a new/old twist. ‘The historian’s character­ istic instrument of encodation, communication, and exchange is ordinary educated speech’ (White 1978: 94). Here the term encodation reveals straight away White’s acceptance of the language myth in its traditional form: to speak is to ‘encode’ ideas, the code being provided by the

154

THE L I N G U I S T I C S OF HI S TORY

linguistic community. What is novel, and somewhat disconcerting, is the notion of ordinary ‘educated’ speech. Some might even regard this as an oxymoron, given that education, throughout most of Western history, was the privilege of a few, and that for so many centuries history was ‘encoded’ in Greek and Latin texts which were incomprehensible to the great majority o f‘ordinary’ people, who were in any case illiterate. So just what ‘ordinary’ implies in White’s formulation is by no means perspicuous. What follows is even less so. He writes: This [sc. the historian’s reliance on ordinary educated speech] implies that the only instruments that he has for endowing his data with meaning, of rendering the strange familiar, and of rendering the mysterious past comprehensible, are the techniques of figurative language. (White 1978: 94)

Here, if we are to believe our eyes, the ‘ordinary-language’ train has gone off the rails altogether. Telling your readers that Napoleon’s army had to withdraw from Moscow or that the original cubist painters all lived in Paris immediately before 1914 does not require the resources of ‘figurative’ language. There is nothing ‘strange’ here that requires to be ‘rendered familiar’, no opaque past that needs to be made ‘compre­ hensible’ by such devices as White proceeds to recommend, viz. meta­ phor, metonymy, synecdoche and irony. These are antiquated rhetorical terms dug up from the historical debris of Western education. They obscure more than they clarify, because they were originally part of a system of education that no longer exists and will never be revived. One can go through Polybius or Livy and tick off the textual deployment of such rhetorical devices. One can even extend the exercise to the texts of Gibbon, Macaulay or Winston Churchill. But all that shows in the end is that certain rhetorical categories can be applied, if you insist, to almost any text. (Just as one can apply a ‘parts of speech’ system in almost any grammatical analysis.) About the specificity of historians’ texts - what makes them ‘historical’ - it says nothing. Sadly, the ‘ordinary-language’ defence of history ends up, by a curious twist of fate, relying on apologists who reinvoke an outmoded educational system and its canon of figurative distinctions. But that may be because it was a mistake in the first place (pace Herodotus et al.) to place history under the linguistic aegis of the rhetorician. The ‘ordinary-language’ defence of history thus comes to be entrusted to the hands of theorists latterly known a ‘narrativists’. Their constant theme is that all historians in the end have to tell ‘a story’; and how a story should be told is not for historians to decide, because ‘story’ is a concept

HI S T OR Y AND OR D I NA R Y L A N G U A G E

155

far broader than ‘history’. Your account of a happy holiday by the sea-side is a story. So is Aesop’s fable of the fox and the crow. On this view, historians have ultimately to submit to the same criteria as other story­ tellers. Whether their story is ‘true’ is another question: it does not even arise until they have produced a story in the first place. According to Rayment-Pickard, the narrativists divide into two camps. For the ‘high narrativists’ the basic assumption is that ‘the world always presents itself to us as text, so that the relationship between an historical text and any other presentation of the world only concerns a distinction of genre’ (Rayment-Pickard 2000: 276). By contrast, the ‘low narrativists’ (e.g. Paul Ricceur) are theorists who admit that ‘there is some congruence between the narrative of history and the world’s happening’ (RaymentPickard 2000: 276). On that showing, I suggest, they might as well be called ‘non-narrativists’ or at least ‘mimimal narrativists’, since the notion of ‘some congruence’ is so vague as to make the position vacuous. (One might as well say that a totally unreliable clock exhibits ‘some congruence’ with Greenwich mean time merely by managing to keep going.) The narrativists have clearly taken on board the historical fall-out from structuralism. But they have collected it in the form of scattered intellectual debris. They have been quite unable to piece it together. What they make of it depends entirely on their own various political and academic programmes. Thus we have feminists claiming that history has been written by men, blacks claiming that history has been written by whites, proletarians claiming that history has been written by capitalists, fascists claiming that history has been written by anti-fascists, and so on ad infinitum. The question then becomes not what ‘the facts’ show, but whom you trust (given that ‘the facts’ are a historical fabrication anyway, licensed by the independence of language from reality). The trouble is that none of these re-tellings of history has any new philosophy of language on which to base its acccount. They are stuck in the post-structuralist trough, where all are free to wallow. And, sensing this lack, they frequently have recourse to pre-structuralist linguistics fa u te de mieux. A case study in these consciously slanted re-tellings of history is provided by Rosalind Miles’s The Women s H istory o f the World. Miles disavows any ‘pretence to the traditional historical fiction of impartiality’ (Miles 1988: 15). This sounds like a bold adventure into lands where no reocentric historian had ever set foot. For the notion of ‘impartiality’ presupposes that there is a world of facts that the historian can choose to

156

THE L I N G U I S T I C S OF HI S TORY

report objectively or otherwise. If historical impartiality is a fiction, then presumably there is no such world, either for feminist historians or for their opponents, the male chauvinist historians. For Miles does not claim that, by some obscure divine dispensation, only women can see history ‘as it really was’. But in the very same chapter Miles announces that the aim of a women’s history of the world is to ‘fill the gaps left by conventional history’s preoccupation with male doings’ (Miles 1988: 12-13). Thus she takes up Virginia Woolfs complaint that ‘women have been all but absent from history’. So it seems that reocentric history was on the right lines after all: it just left some ‘gaps’ in the account. It is not that feminist history has a different date for the battle of Hastings, or even ignores the battle in favour of telling us what Harold’s queen and her ladies were doing on the same day. When it suits her purpose, Miles is willing to admit men to the ranks of what she calls ‘reliable historians’. She wheels on Herodotus, Pliny, Strabo and Plutarch in order to lend credence to the existence of the Amazons (Miles 1988: 51). The irony pervading this strategy of justification hinges on Miles’s evident failure to realize that the whole idea of radically different histories of the world written from a male viewpoint, a female viewpoint (and presumably indefinitely many other possible viewpoints) would never have been taken seriously before the structuralist revolution. For it was structuralism which made it intellectually respectable to treat words (including words favoured by traditional male historians, such as king, queen, governm ent and revolution) as leading a life of their own, not answerable to any world of ‘facts’, and hence available to construct as many different versions of reality as different language-users might wish. But, symptomatically, Miles has little to say about language at all. She does not, for instance, claim that feminist history needs a new vocabulary of its own. Although she challenges male history, she does not challenge the language in which it is written. Why not? Well, the snag is that if we go the whole hog with this divorce between words and facts, then the potential post-structuralist proliferation of rival histories, mutually in­ compatible, raises the awkward question of their validity and the criteria by which validity is to be assessed. So Miles, wishing to present a different but ‘true’ historical narrative, is forced in self-defence to revert to the pre-structuralist view that, after all, words are answerable to facts, and enter the plea that her feminist version states facts that male historians omitted or were blind to (thus tacitly conceding that some male historians did after all give factual accounts of the past). The problem here is that feminists have never presented any coherent case for supposing that e.g. English is an unsuitable language for the

HI S T OR Y AND O RDI NA RY L A N G U A G E

157

feminist historian. On the contrary, whether intentionally or not, the feminist case for rewriting history has relied essentially both on ‘ordinary language’ and on acceptance of a very traditional (surrogational) approach to the interpretation of ‘ordinary language’. This becomes apparent as soon as one inspects any clearly articulated statement of the feminist ‘challenge’ to traditional history. Beverley Southgate (Southgate 2001), for example, sets out the case as follows. (1) There were very few women historians in the Western tradition, and those few (such as Christine de Pisan and Catherine Macaulay) tend to be discounted. But, as Southgate admits, their readmission or even their promotion to the status of privileged historians, does not convincingly underwrite the feminist claim that this would ‘give us for the first time a true history o f humanity’ (Southgate 2001, citing the feminist historians Anderson and Zinsser). (2) In some cases, what male historians view as ‘progress’ was not so for women. Examples: the so-called ‘age of reason’, in which women were still being burned as witches; the ‘Renaissance’, a male-oriented expansion of horizons, in which women did not participate (Kelly J. 1984). Arguably, their position had deteriorated since the Middle Ages. (3) The language of history is ‘a language that derives from and in turn underpins an essentially “patriarchal” structure’ (Southgate 2001: 103). In these ways, Southgate concludes, ‘feminism has challenged the chronological and conceptual frameworks within which history has conventionally been done’ (Southgate 2001: 104). It is only the third of these claims that overtly engages linguistic issues as such. But if taken at face value, the claim seems to be self-defeating. For nothing about ‘the French language’ of her day prevented Christine de Pisan from expressing views that would have been unwelcome to her male contemporaries. Nor could it reasonably be supposed that any woman writer living in the Renaissance or during the Enlightenment was inhibited from denouncing the deplorable conditions for women of the time by reason of being caught in a linguistic trap that occluded any expression of opinions contrary to prevailing male orthodoxies. After all, many other counter-orthodox opinions were expressed in these times (which often had dramatic results, including very unpleasant ones for those who ventured to express them). But to suggest that the ‘silence’ of women has a linguistic explanation is as implausible as suggesting that the ‘silence’ of Protestantism before Luther was due to some deficiency of the German language. Feminist historians (although, for lack of historical perspective, they do not see it this way) are caught in a post-structuralist dilemma of their own making. Most of them dare not break radically enough with the hallowed

158

THE L I N G U I S T I C S OF HI S TORY

language myth of the Western tradition. This becomes evident when one considers the attempts that began in the 1970s to stake out a distinctively feminist position on language. The main charge against ordinary language (usually assumed to be English) was that it was - or had been in the past - ‘biassed’ against women. Few of those who made this charge were themselves linguists; but a notable exception was Robin Lakoff, who in 1975 published Language and Woman s Place. Lakoff s book had an eye-catching cover depicting the face of a woman with a length of tape plastered across her mouth. The suggestion, obviously, was that there is something forcibly preventing women from speaking, i.e. from expressing themselves orally. Lakoffs basic claim is that this ‘something’ is linguistic: women are victims of ‘linguistic discrimination’. Ordinary language, according to Lakoff, sub­ divides into two: there is ‘women’s language’, taught to little girls, and ‘men’s language’, taught to little boys but forbidden to little girls. Exactly what the source of this linguistic discrimination is never becomes very clear in her book; for although Lakoff refers very generally to a source in ‘language’, and occasionally mentions other languages in passing, her key illustrative examples turn out to be English. The reader is never quite sure whether English is a particularly nasty language, and women do better in, say, Hausa or Telugu; or whether Lakoff thinks that there is something in the very workings of language itself that is permanently unfair to women. ‘Language uses us as much as we use language’, she declares in Chapter 1. The metaphysics of this statement will strike many as leaving room for clarification, to say the least. What language is using us for, and why, also seem to call for explanation; but none is forth­ coming. Although Lakoff never makes her own theoretical position as a linguist very clear, it seems obvious from her terminology that she is heavily committed to acceptance of some form of the traditional language myth. She accepts such mythical constructs as ‘standard American English’ with its ‘native speakers’, its ‘syntactic rules’ and its ‘literal meanings’. She assumes without argument that linguistic communication is a telementational process. Our ‘choice of forms of expression is guided by the thoughts we want to express’ and ‘the way we feel about things in the real world governs the way we express ourselves about those things’. Here we see classic surrogational semantics (of the psychocentric variety) already installed and ready to go. Hence anything that hinders or distorts the verbal expression of one’s true thoughts or feelings is, for Lakoff, objectionable by definition (i.e. in some way runs counter to the proper psychocentric function of words).

HI S T OR Y A ND O R DI NA R Y L A N G U A G E

159

Lakoffs evidence for linguistic discrimination is likewise psychocentric: she says her ‘data’ have ‘been gathered mainly by introspection’. In defence of this suspect procedure, she asserts that any procedure ‘is at some point introspective’. Thus for Lakoff, it seems, linguistics (and presumably history too) are essentially subjective investigations, and all claims to objectivity are either naive or specious. Yet she does not hesitate to maintain that ‘the majority of claims I make will hold for the majority of speakers of English’. How she knows this, or how many of the millions of speakers of English she has tested in support of this claim, are not revealed. Lakoff makes out her argument by presenting a selection of examples that reflect either remembered or imagined conversation. She extrapolates from these examples conclusions about what can and cannot be said. One of them is ‘Rhonda is a mistress’ (Lakoff 1975: 29), which Lakoff claims ‘one cannot say’. One linguistic objection to this line of argument is that not only is it perfectly possible to say Rhonda is a mistress but there are contexts in which that would be the appropriate thing to say, and the truth to boot. Furthermore, no traditional grammarian would have any objection at all to this sentence. One has to infer that what Lakoff intended to claim (despite her clumsy formulation of the point) was that ‘Rhonda is a mistress’ is an observation that requires some preparation or frame­ setting in the exchanges of everyday conversation. But what she actually claims (mistakenly) is that mistress ‘requires a masculine noun in the possessive to precede it’. This requirement Lakoff describes metalinguistically as if it were a grammatical rule of the English language (NB masculine noun, possessive, precede), rather than a matter of contextualized assumptions built up in the course of conversation. That leaves her open to the obvious riposte from traditional grammarians that she is confusing grammatical restraints, pertaining to what Saussure called la langue, with violations of social expectations, pertaining to what Saussure called la parole. (For those who cannot see the problem, compare Rhonda is a mistress with Jen n ifer is an assistant.) Lakoff summarized her conception of ‘women’s language’ under nine heads. (1) Vocabulary relating specifically to ‘women’s work’. (2) Use of ‘empty’ adjectives (e.g. divine, charm ing, cute). (3) Question intonation replacing declarative intonation. (4) Use of ‘hedges’ (w ell, kinda, sorta, etc.). (5) Use of ‘intensive’ so. (6) ‘Hypercorrect’ grammar. (7) ‘Superpolite’ forms {please, thank you). (8) Women ‘don’t tell jokes’ {sic). (9) Women ‘speak in italics’ {sic). That American women share an over­ whelming proportion of their vocabulary with American men apparently

160

THE L I N G U I S T I C S OF HI S TORY

counted for nothing in Lakoff s view, and even she was forced to admit that ‘there is no syntactic rule in English that only women may use’. So what remains of the claim that a separate ‘women’s language’ actually exists? As Deborah Cameron points out, in a review of feminist writing about language, ‘Lakoff s picture is a stereotype for which she presents no real evidence, and other women who have tested her hypothesis against a quantity of speech data have found it wanting on several counts’ (Cameron 1985: 34). Lakoff concluded her book with the express hope that it ‘will be one small first step in the direction of a wider option of life styles, for men and women’ (Lakoff 1975: 83). For linguists, this avowal could hardly be read other than as a veiled allusion to the notion propagated by more extreme forms of the Whorfian hypothesis; namely, that we are all ‘prisoners’ of whatever language we speak, and cannot see the world otherwise than through the distorting lens of its lexical and grammatical structure. So inviting readers to embrace ‘a wider option of life styles’ contained the implicit assumption that there was currently something inhibiting them from doing so. Lakoff, however, was quite careful to avoid any explicit commitment to Whorfianism, doubtless being fully aware of its very dubious theoretical standing among academic linguists. Although Lakoff had ventured out on a shaky linguistic limb in favour of the feminist cause, her position did not gain immediate approval from other feminists. According to Dale Spender, author of a book challengingly entitled Man M ade Language (1980), Lakoff herself was a victim of ‘the sexist assumptions of the linguistic paradigm in which she worked’ (Spender 1980: 8). Lakoff, according to Spender, ‘accepts that men’s language is superior’. She also ‘takes male language as the norm and measures women against it’. A feminist position on language, in Spender’s view, has to sweep away all these assumptions and establish something much more radical. Spender’s book ‘received enough media coverage to put the subject of women and language on the map for many people outside the women’s movement’ (Cameron 1985: 3). But it did not extricate anyone from the confusion between langue and parole that Lakoff had got into. Rather, it compounded that confusion by adding to it a separate confusion between metalinguistic analysis and linguistic ‘facts’. Cameron (rightly) took Spend­ er to task for conflating the two. Spender had claimed that women are relegated to a ‘negative semantic space’ in the English language. In support of this, she referred to the way linguists have classified words like girl and w ife not as being semantically ‘ + female’ but as ‘ — male’. Spender, according to Cameron, ‘seems to feel that the bias is not only in the analyst

HI S T OR Y A ND O RDI NA RY L A N G U A G E

161

but actually in the language’ (Cameron 1985: 60). Cameron puts her point forcefully: ‘To say that the language embodies value judgements based on these plus and minus categories is a theoretical vulgarity even the most chauvinist linguist would repudiate’ (Cameron 1985: 60). But, forceful as it may be, it is not always grasped by those feminists who continue to rage against ‘the English language’ as ‘sexist’. Cameron remains one of the few feminist linguists of note who have managed to sort the wheat from the chaff where linguistic theory is concerned. Feminism, it might be claimed, has nevertheless made an impact on ‘ordinary language’ at least insofar as writers are nowadays often reluctant, for example, to perpetuate the old usage of English he as the ‘genderneutral’ pronoun covering both sexes (lest they be accused of ‘male chauvinism’). Some conspicuously affect the use of she as gender-neutral (‘The historian must always respect her sources’). The same stylists will go to extraordinary lengths to avoid writing e.g. ‘A medieval knight had specific obligations to his lord’. And since ‘A medieval knight had specific obligations to his or her lord’ sounds absurd, given that medieval knights were men, they will concoct grammatical contortions like ‘Specific obli­ gations to their lord were incumbent upon a medieval knight’. Here we see ‘ordinary language’ being passed through the prescriptive sieve of political correctness. One suspects that history will show such usages to have been a temporary fad confined to the (written) language of professional academics. But the point is that insisting on them at all as correctives to traditional forms of expression has theoretical implications. The whole notion that one can put the world to rights by making the necessary adjustments to languages is itself based on structuralist assumptions (or, at least, some would say, a naive or misguided application of structuralism). A quite different but no less revealing example of the post-structuralist impasse is to be seen in the writings of L. O. Mink, who claims that ‘we cannot refer to events as such, but only to events under a description’ (Mink 1987: 199). This, again, is a position which rests in the end on structuralist foundations; in this case, the denial that there is any direct correlation between words and the world. (Hence one cannot refer - as traditional historians thought they were doing - to ‘events as such’.) In order to give this denial its full weight, I will cite the immediately preceding paragraph in Mink’s essay. There we are told that: narrative form in history, as in fiction, is an artifice, the product of individual imagination. Yet at the same time it is accepted as claiming truth - that is, as representing a real ensemble of interrelationships in past actuality. Nor can we say

162

THE L I N G U I S T I C S OF HI S TORY

that narrative form is like a hypothesis in science, which is the product of individual imagination but once suggested leads to research that can confirm or disconflrm it. The crucial difference is that the narrative combination of relations is simply not subject to confirmation or disconfirmation, as any one of them taken separately might be. So we have a second dilemma about historical narrative: as historical it claims to represent, through its form, part of the real complexity of the past, but as a narrative it is a product of imaginative construction, which cannot defend its claim to truth by any accepted procedure of argument or authentication. (Mink 1987: 199)

Here we see a post-structuralist panic, with ensuing rush for the lifeboats. It is not just that the narrative construct is unveriflable, but that, even when broken down into individual segments, these segments cannot be related to ‘events as such’. It is unclear whether Mink is claiming that events do not have names, but only descriptions. If so, many linguists would wish to take issue with him. Or is he claiming, like Bertrand Russell, that names are disguised descriptions anyway? If so, many linguists would take issue with that too. I do not, for the moment, intend to take sides on these controversial matters, but simply to point out that these are issues of linguistic theory. The historian cannot evade them by claiming immunity under some ‘protection of ordinary language’ act. How Mink and others might resolve their problem I shall come back to subsequently. Here it suffices to make the point that the problem is generated by its location in a post-structuralist desert. There the languageuser is found wandering in a no-man’s-land, where no one (male or female, historian or histopathologist) is quite sure about how words relate to anything at all. Marx saw no great problem about language in relation to history. But that is a measure of his unreflecting commitment to the surrogational seman­ tics of his academic upbringing, and in particular to the psychocentric notion that in language sound is the ‘material’ vehicle for ideas. From the outset, ‘spirit’ is cursed with the ‘burden’ of matter, which appears in this case in the form of agitated layers of air, sounds, in short, of language. Language is as old as consciousness, language is practical consciousness, as it exists for other men, and thus as it first really exists for myself as well. Language, like conscious­ ness, arises only from the need, the necessity of intercourse with other men. (Marx 1963: 85-6 [from German Ideology, 1845-6])

Later Marxist historians discovered all kinds of problems with the doctrine that ‘language is practical consciousness’, particularly in the

HI S T OR Y A ND OR D I NA R Y L A N G U A G E

163

aftermath of structuralism. As a consequence, they became increasingly aware of the contradictions inherent in their own linguistic practice. John Warren cites one prominent Marxist historian as reassuring an audience that ‘his vocabulary had changed since his earliest published work, but that his ideas were really the same’ (Warren 1999: 90). This takes the psychocentric biscuit as an apologia for shifting your position. It also adds a historical dimension to linguistic Humpty-Dumptyism: not only can words mean what one wants them to mean, but they can retrospectively be given meanings that one now wants them to have had. Even the great Stalin found himself in difficulties when forced to grapple with the theoretical question of whether, for the true Marxist, the Russian language belonged to the base or the superstructure (Stalin 1954). Stalin pointed out that Marx held language to be ‘the immediate reality of thought’. According to Stalin, this meant that the very idea of ‘thinking without language’ was an ‘idealist’ mistake. But getting from this position to the claim that Russian was the common language of the entire Russian people required a considerable feat of casuistry. What is remarkable about Stalin’s nationalistic ‘solution’ is not so much his denunciation of those comrades ‘who have come the conclusion that national language is a fiction, and that only class languages exist in reality’, or even the historicism that allows him to claim that a language ‘is the product of a whole number of epochs’ and ‘lives immeasurably longer than any basis or any superstructure’, but his tacit acceptance of a framework of discussion which owes so much to that most bourgeois of linguistic thinkers, Saussure. (For further comments, see Harris R. 2001: 203-4.) This is not the place to attempt to write an updated Marxism and Problems o f Linguistics, even in outline. I shall here focus on one particular example, that of Herbert Marcuse. Marcuse not only appeals overtly to philosophy of language but relates it directly to the question of history. In One-Dimensional Man (subtitle: Studies in the Ideology o f A dvanced Industrial S ociety) Marcuse adopts a distinctly Whorfian version of structuralism, while managing to avoid any discussion of Whorf, or even of Saussure (Marcuse 1964). Marcuse clearly disagrees with Stalin’s historical ‘solution’ to the linguistic problem, although that too is referred to only obliquely. We are told that the authoritarian ritualization of discourse is more striking where it affects the dialectical language itself. The requirements of competitive industrialization, and the total subjection of man to the productive apparatus appears in the authoritarian transformation of the Marxist into the Stalinist and post-Stalinist language. These

164

THE L I N G U I S T I C S OF HI S TORY

requirements, as interpreted by the leadership which controls the apparatus, defines what is right and wrong, true and false. (Marcuse 1964: 101)

We should note the term ritualization\ for that by implication impugns the format of Stalin’s pronouncements in Marxism and Problems o f Linguistics. There we had a pedagogic question-and-answer procedure (Q. ‘Is it true that. . . ? A. No, comrade, it is not true th at. . .’) which is reminiscent of some kind of catechism, where the neophyte’s naive queries are given authoritative replies from on high. Marcuse takes seriously Orwellian worries about the unscrupulous manipulation of language by those in power. At the same time, he does not wish to surrender ordinary language to their control. As a result, like some feminists, he is driven to distinguish between two ‘languages’. These are not, however, ‘men’s language’ versus ‘women’s language’, but a ‘closed language’ versus an ‘open language’. Like a good Marxist, he cites the Communist M anifesto as an example of the latter. (This takes some swallowing, in view of the manifestly propagandist tenor of that docu­ ment.) An open language, says Marcuse, allows us to think: a closed language does not. His authorities on language are, interestingly, Hegel, Humboldt and Roland Barthes. J. L. Austin and Wittgenstein are paraded in the rogues’ gallery as examples o f‘academic sado-masochism’ (Marcuse 1964: 173). Marcuse reserves his most barbed sallies for the concept of ‘ordinary language’ as presented by ‘ordinary language philosophers’. Throughout the work of the linguistic analysts, there is this familiarity with the chap on the street whose talk plays such a leading role in linguistic philosophy. The chumminess of speech is essential inasmuch as it excludes from the beginning the high-brow vocabulary of ‘metaphysics’; it militates against intelligent non­ conformity; it ridicules the egghead. The language of John Doe and Richard Roe is the language which the man on the street actually speaks; it is the language which expresses his behavior; it is therefore the token of concreteness. However, it is also the token of a false concreteness. The language which provides most of the material for the analysis is a purged language, purged not only of its ‘unorthodox’ vocabulary, but also of the means for expressing any other contents than those furnished to the individuals by their society. (Marcuse 1964: 174)

Where does history come into the picture? For Marcuse, straight away. A ‘closed language’ has as its very objective the suppression of history. In its own development, dialectical thought came to comprehend the historical character of the contradictions and the process of their mediation as historical process. Thus the ‘other’ dimension of thought appeared to be historical dimension - the potentiality as historical possibility, its realization as historical event.

HI S T O R Y A ND O R D I NA R Y L A N G U A G E

165

The suppression of this dimension in the societal universe of operational rationality is a suppression o f history, and this is not an academic but a political affair. It is suppression of the society’s own past - and of its future, inasmuch as this future invokes the qualitative change, the negation of the present. A universe of discourse in which the categories of freedom have become interchangeable and even identical with their opposites is not only practicing Orwellian or Aesopian language but is repulsing and forgetting the historical reality - the horror of fascism; the idea of socialism; the preconditions of democracy; the content of freedom. (Marcuse 1964: 97-8)

For Marcuse, this closed language (which he also refers to as the ‘functional language’) is ‘a radically anti-historical language’ (Marcuse 1964: 98). But it not only suppresses history: it deliberately obscures the relationship of words to reality. It is ‘the language of one-dimensional thought’ (Marcuse 1964: 95). Marcuse contrasts its implications with those o f‘classical philosophy of grammar which transcends the behavioral universe and relates linguistic to ontological categories’. Here the under­ lying surrogationism of Marcuse’s view of language surfaces from the sea of jargon at last. He needed the flirtation with (Whorfian) structuralism in order to condemn the iniquities of a closed language. But then it drops out of sight. We are left in the end with a recognition of ‘ontological categories’ to which words are - or should be - answerable. Even today, we still hear historians calling for old-fashioned clarity of expression. According to Arthur Marwick, the language of the historian ‘should be as precise as the mathematical symbols of the scientist’ (Marwick 2002: 10). How that could be, unless history were reduced to statistics, will be unclear to many. But not to Marwick. It is just a question of the historian being ‘absolutely clear in his/her own mind what it is he/she is endeavouring to convey to the reader’. Given that, then presumably the resources of ‘ordinary language’ are sufficient to do the rest. Just a matter of selecting the right words to match those absolutely clear ideas. If this is intended as a rehabilitation rather than a satire of Lockean psychocentrism, it seems to be in the hands of someone who has sat deaf throughout the entire post-Lockean debate on language. Even Locke did not suppose that being clear about the ideas in one’s own mind guarantees being able to find words to convey those ideas to others: on the contrary, for Locke that is just where linguistic problems begin. Marwick has been given a rough ride even by fellow historians. Richard Evans takes him to task for decrying ‘the techniques of deconstruction or discourse analysis’ as compared with the ‘sophisticated methods histor­

166

THE L I N G U I S T I C S OF HI S TORY

ians have been developing over the years’ and cites examples from Marwick’s own publications to show his reliance not on any ‘sophisticated methods’ but, as Evans puts it, ‘personal opinions and gems of homespun wisdom’ (Evans 2000: 72). Nor does Marwick’s own handling of language escape: ‘The main characteristic of the Marwickian metaphor seems to be its complete absence of identifiable meaning’ (Evans 2000: 73). Quite a drubbing for someone who champions a historian’s language which is as precise as the mathematical language of the scientist. Unless, that is to say, Marwick is more subtle than anyone takes him for, and holds that the supposed ‘precision’ of scientific language is an illusion too. And there are many who might agree with that. Where ordinary language is concerned, Evans is an interesting example of the gamekeeper turned poacher and back again. He dismisses in short order the notion that history can dispense with technical terminology. He points out that in econometric history and demographic history there is no way of avoiding technical terms, not to mention such topics as the medieval manorial economy, the structure of landholding and serfdom in eighteenth-century central Europe, the legal basis of the slaveholding system in the old American South, or the debate on stamp duty and other aspects of taxation policy in pre-revolutionary America [. . .].(Evans 2000: 67)

But Evans is sceptical about the equation between ‘technical’ and ‘scientific’. He is a severe judge in general of his colleagues’ ability to operate in their own chosen medium. ‘Few historians write compet­ ently; fewer still display any real mastery of the language in which they publish their work. Most history books are hopelessly unreadable’ (Evans 2000: 70). Nevertheless, he upholds ‘the historian’s general responsibility to write as clearly and unpretentiously as is possible under the circumstances’ and endorses Lawrence Stone’s belief that ‘one should always try to write plain English’. ‘Clarity of presentation, after all, is a necessary part of intellectual precision: there is nothing necessarily “unscientific” or sloppy about it, quite the reverse’ (Evans 2000: 67). So we end up back in the school classroom after all. The traditional commitment to surrogational semantics is reaffirmed in the guise of a critique of the historian’s ‘style’. Naughty Marwick is readmitted, but made to stand in the corner wearing a dunce’s cap. His fate could have been worse in France, where the rhetoric of Rivarol’s linguistic jingoism still echoes down lycee corridors. Here surrogational semantics masquer­ ades as patriotism. Ce qui n est pas clair n est pas franpais. Lack of

HI S T OR Y AND OR D I NA R Y L A N G U A G E

167

linguistic clarity is a sin against the nation; and no one is allowed to be in favour of sin. Although much of the huffing-and-puffing about the merits and deficiencies of ordinary language seems to me inspired by a misguided Whorfianism, there is one important point on which the ordinarylanguage thesis is right. It relates to the argument introduced in my opening chapter to the effect that if the Martian language has no past tense or other linguistic devices for time reference, it will be impossible for any history to be written in Martian. I do not propose to rehearse here any of the conundrums about time and language that modern philo­ sophers have unearthed. (For a survey in relation to history see Meiland 1965.) It seems to me not worth worrying about whether the use of a past tense can be taught by ostensive definition, or what it might mean, if anything, to say ‘It is now five o’clock on the sun.’ What matters is that, as regards what traditional grammarians used to call the ‘grammar’ of tenses and related expressions of time measurement, historians are committed to exactly the same usages as the rest of us. The system that made it possible for me to make (and keep) an appointment with my dentist last week, or to say that last year I went to Oslo, is none other than the system that makes it possible for the historian to date the major battles of the World War I, or discuss the succession of events in the French Revolution. There is no tense and there are no temporal adverbs that only historians use. In this respect at least, the language of the historian is ordinary language, and there is no other language that will do the job. Where some philosophers and scientists seem to me to have got it wrong (not because their thinking about time was pre-Einsteinian, but because their linguistics was preSaussurean) was in supposing or imagining that there could be some independent, language-neutral mechanism of time measurement, a kind of God’s clock that has been keeping time for all eternity, and by reference to which all earthly clocks and calendars could, in principle, be ‘checked’. Time measurement and time reference are always linguistic matters, and there are no eternal languages. To grasp that is the first step towards constructing a viable linguistics of history.

CHAPTER SIX

The autonomy of history

When Collingwood championed what he called (unfortunately) ‘the autonomy of history’ (Collingwood 1946) he was referring to an achieve­ ment of historians, and specifically what he took the historians of recent generations to have achieved. This achievement, as described by Joseph Levine in The Autonomy o f H istory, was the establishment of the independence of their discipline from all other disciplines: ‘It was neither literature, nor theology, nor natural science: its logic was neither deduc­ tive nor inductive; its procedures were evidently its own’ (Levine 1999: viii). Collingwood could hardly have foreseen the day when this vaunted autonomy would become the Achilles’ heel of the embattled historian, under attack from critics who equated that kind of autonomy with lack of accountability, made less forgivable by a spurious claim to be accounted a ‘science’ into the bargain. And indeed for Collingwood the two went hand in hand. ‘Like every science,’ he says, ‘history is autonomous’ (Colling­ wood 1946: 256), as if it were obvious that autonomy is a necessary condition for being a science. Since Collingwood’s day, such claims have increasingly come to be regarded as intellectual arrogance. In the modern climate of opinion, establishing ‘what really happened’ in the past is not seen as something that can be safely left to professional academics, and certainly not to the closed shop of one particular academic discipline, whose members often cannot agree among themselves. Recognizing the non-autonomy of history seems to me a prequisite for developing what I shall call an ‘integrationist’ linguistics of history. Is there any way of rescuing history from the post-structuralist slough of despond? In this chapter and the next I shall try to outline a linguistics of history that bypasses those tedious debates about ‘the facts’ between autonomist historians and their postmodernist critics. ‘Pomophobia’, or fear of postmodernism, has recently been identified as the psychosis that many historians now suffer from (Southgate 2001: 148-53). I call these 168

THE A U T O N O M Y OF HI S TOR Y

169

debates ‘tedious’ because they are so often based on question-begging linguistic assumptions on both sides (surrogational assumptions on one side and structuralist or pseudo-structuralist assumptions on the other). Against both parties, I shall argue that history is, always and everywhere, a reflection of its own integrational mechanisms, because that is the way historical meaning is made. History is not a special case: all meaning­ making, at all levels of communication in all societies, operates in the same way. History does not have a semantics of its own. Let me now try to supply some content for that cryptic proposal. In previous chapters I have been working towards trying to establish four main propositions. (1) History-making is a collective communicational process, and history is nothing other than its product, by whatever means this can be achieved in particular cases. (2) Mainstream historians in the Western tradition did not first adopt a philosophy of history and then select a philosophy of language to go with it. It was the other way round. The positions the historian was free to take vis-d-vis describing the past were determined, either consciously or unconsciously, by the assumptions about language that the historian felt obliged to adopt. (3) The assump­ tions about language that are relevant for historians depend on the communicational organization of their society. Thus it makes a great deal of difference whether or not the historian is a member of a literate society. Literate societies have different historical criteria from pre­ literate societies. In literate societies, the view of the past is answerable to writings that survive from the earlier times. In other words, in order to construct or conceive of a history of X, one has to have a way of isolating some thread of events connected across time, and ignoring concomitant events that ex hypothesi do not concern X. In a literate society concerned with its own history, the possibility of identifying such a thread depends primarily upon surviving texts. (4) Even within a literate society, different philosophies of language may be dominant at different periods. Historians are inevitably affected by this. Before we can understand the concept of history that we are presented with in any given historical account, we need to understand the view of language and communication on which it was based. And there is no point trying to rewrite history ourselves, as Wilde urged us to do, unless we know on what linguistic basis we are doing it. The current slough of despond, if my analysis is right, results directly from the conflict between irreconcilable models of communication that Western historians and their critics have adopted, and the mud churned up by this conflict. If there is a way out of this slough, then - I would argue - it has to be based not on further efforts to resolve the conflict (because there is no resolution) but on a different model of communica­

170

THE L I N G U I S T I C S OF HI S TORY

tion altogether. This would require a philosophy of language that super­ seded both twentieth-century structuralism and the dominant surrogationalism of the preceding centuries. The obvious candidate for such a role at the present time is integrationism (Wolf and Love 1992; Toolan 1996; Harris R. 1998; Harris and Wolf 1998; Davis H. G. 2001; Davis and Taylor 2003), since integrationism makes a more radical break with the old ‘sender-receiver’ models of human communication than any other contemporary school of linguistic thought. Integrationism rejects both the psychological assumption that words enable ideas to be trans­ mitted from one person’s mind to another’s, as well as the sociological assumption that words are controlled by social consensus in order to make this transmission possible. In an integrationist perspective, words are neither labels for things nor labels for ideas, but integrated components of creative communicational endeavours. Any attempt to treat words as decontextualized or decontextualizable items inevitably leads to fundamental misunderstandings concerning their communicational function. Integrationist semantics treats meanings as being created ambulando to meet the requirements of particular communication situations. Thus, according to circum­ stances, one and the same set of statements may, and commonly does, mean different things for different people. A form of words does not carry a permanent meaning or meanings around with it. This integrationist tenet has obvious implications for the historian. Whatever a historian may write about some event in the past (whether in ‘ordinary language’ or any other kind of language) cannot be written in such a way as to express one and the same meaning for all potential readers. For that would demand of language more than language can give. For the integrationist, writing is not a visually encoded form of speech, any more than speech is an orally encoded form of thinking. The alleged ‘codes’ are cultural fictions to be explained; not ‘given’ as the basis for explaining communication. The integrationist approach thus reverses traditional explanatory priorities, and I shall accept the validity of that reversal from now on. There is a good reason for doing so where history, in particular, is concerned. Integrationism has been described more than once as taking a ‘Heraclitan’ view of communication. That should make it particularly appropriate to capture what A. J. P. Taylor, in one of the epigraphs quoted at the front of this book, referred to as ‘the one truth we seek to recapture when we write history’ (Taylor 1976: 17): its ceaseless flux. At the same time, in keeping with another of the epigraphs cited (Kolakowski 1972: 138), the integrationist sees no Cratyline reason to ‘renounce traditional modes of speech’, but nevertheless seeks to free

THE A U T O N O M Y OF HI S TORY

171

words - and particularly words belonging to our metalinguistic termi­ nology - from the dubious metaphysics so frequently foisted upon them. H. G. Wells once published a story called ‘The truth about Pyecraft’. Needless to say, the ‘truth’ about Pyecraft was fictional. But, within the framework of the story, that did not either preclude or invalidate a distinction between what was true about Pyecraft and what was not. Likewise, I need to make it clear that my framework for discussing truth and meaning is an integrationist framework. It will not impress those already committed to other frameworks (other philosophies of language), although it might possibly bring them to see exactly what their own commitments are. I favour this framework because, although implying certain criticisms of the traditional history I was reared on, it does not commit me to rejecting the idea of truth altogether, or denying the existence of the past (which, as a lay member of the community, I am already committed to talking about anyway), or emulating any of the more acrobatic feats, such as proclaiming the end of history, that in recent years have made the intellectual headlines. To present the results of research and reflection about the past, whether in books, lectures or any other medium, is to engage deliberately in some form of communication. On this probably few historians or philosophers of history would disagree with integrationists. But that engagement can be realized in a variety of ways. From an integrationist perspective, there is no such thing as ‘just a form of communication’. There is no question of separating the thing-to-be-communicated from the communication-ofthe-thing-in-question. Communication is not an activity or a facility added on to life as an optional extra. Communication is intrinsic to life, at least as the majority of human beings live it. Communication, in whatever form it may take, integrates that existence. All forms of communication have integrational functions in the life of the society that sponsors them: that is their raison d'etre. They articulate relations between individuals and groups of individuals that would otherwise be impossible. From this point of view, history-making is an essentially integrated and integrating enterprise. An integrationist linguistics of history (as of any other field) requires, first, an identification of the activities to be integrated. I propose that the integration requires contributions from at least three classes of history-maker. The first type of human contribu­ tion to history (quite apart from the non-human contributions made by floods, earthquakes and other natural disasters) consists in actually doing something, or taking part in events, then or later deemed worthy of public remembrance, such as fighting in World War I, or inventing the internal

172

THE L I N G U I S T I C S OF HI S TORY

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 an integrationist would say that res gestae do not actually become history (or, more precisely, become ‘historified’, as I shall say from now on) until integrated with two other essential contributions. One of the other two is supplied by those involved in constructing accounts of res gesta e, including those people commonly known in the Western tradition as ‘historians’, who have made it their business to produce such accounts. Historians make up the smallest of the three classes of history-maker. Historians do not necessarily witness personally or participate in the events they give an account of. So if they claim the status of ‘historian’ it must be in general on some other ground (than personal experience) of having knowledge of what happened. Occasionally exceptional individuals do write accounts of their own res gestae. 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. It would be grotesque to ask Caesar what basis he had for writing such an account. But it is in no way grotesque to ask a modern historian the same question, i.e. how the historian comes to be in a position to write about what happened in Gaul in the first century b c . Caesar did indeed ‘write himself into history’. But he would not have been able to do that without a third contribution that he was in no position to make. The third contribution required for the integrational process that - in my view —is the making of history comes neither from res gestae nor from the historian. If we stop short at what Julius Caesar did and what Julius Caesar wrote, we have failed to come to terms with the phenomenon of history-making in all its complexity. What is missing is the reception of Caesar’s deeds and Caesar’s writings not only by the public of Caesar’s day but by later ages. The onlookers too make history, as do their descendants. They constitute by far the largest body of history-makers. They make it in various ways, sometimes in obvious and sometimes in less obvious forms. A statue of Caesar is not a historical account of anything, but its being set up at all is a significant contribution to history-making. The same applies when statues of former leaders are defaced or toppled from their plinths. An anniversary parade is not an account of anything either; but it is a living contribution to history-making nonetheless. Nor is having one’s likeness struck on the face of a coin an account of a past event, but it bears witness all the same. And much of what Hobsbawm defines as ‘invented tradition’ (Hobsbawm 1983: 1) is part of this contribution too. More generally, however, the onlookers and their

THE A U T O N O M Y OF HI S TORY

173

descendants contribute through an untold number of actions, reactions and interactions that are never put on record and never could be. The factors that are decisive in shaping this third contribution are as varied and imponderable as the circumstances of the contributions. At one end of the spectrum we cannot discount the influence of individual historians or even philosophers of history. At the other end we have to take into account how well or ill informed people may be about the world in which they live (to say nothing of the world of their ancestors), through what particular channels such information reaches them, and against what background of hopes and fears they evaluate it. At this point the making of history merges with the formation of a Weltanschauung. It is the omission of the first and third classes of history-maker that would make an integrationist reluctant to accept Oakeshott’s account of history-making, where it seems that the burden falls entirely on the shoulders of the historian: for history, says Oakeshott, ‘is “made” by nobody save the historian; to write history is the only way of making it’ (Oakeshott 1933: 99). Oakeshott is presenting what, from an integrationist point of view, is a psychocentric semantics of history. But historians who have recoiled from this as opening the door wide to subjectivity of the most undesirable kind often seem to think that the balance can be put right simply by insisting that the historian face ‘the facts’. E. H. Carr, for example, not only concedes that ‘the facts of history never come to us “pure”, since they do not and cannot exist in a pure form: they are always refracted through the mind of the recorder’ (Carr E. H. 1987: 22) but is sufficiently psychocentric himself to claim that ‘history cannot be written unless the historian can achieve some kind of contact with the mind of those about whom he is writing’ (Carr E. H. 1987: 24). How anyone alive today could be sure they had made contact with the mind of Julius Caesar I find it hard to see. Nevertheless Carr, who wishes to salvage objectivity for the historian, attempts to do so by confronting the latter squarely with ‘the facts’ (’impure’ and ‘refracted’ though they may be). The historian and the facts of history are necessary to one another. The historian without his facts is rootless and futile; the facts without their historian are dead and meaningless. M y first answer therefore to the question ‘What is history?’ is that it is a continuous process of interaction between the historian and his facts, an unending dialogue between the present and the past. (Carr E. H. 1987: 30)

This might almost sound like integrationism avant la lettre (particularly when Carr speaks of ‘a continuous process’) were it not for (1) its

174

THE L I N G U I S T I C S OF HI S TOR Y

commitment to a simple binary relation between writing on the one hand and what is written about on the other, and (2) the bizarre notion of a dialogue between present and past. Dialogue is a two-way business. It presupposes the possibility of reactions by both parties. I cannot have any dialogue with Aristotle, or with his contemporaries either, because they all died long before I was born. And those with whom I can hold a dialogue eo ipso belong to my present. All I can do where Aristotle is concerned is not hold a dialogue of any kind but try to take account of what he said, or what I now think he might have said. And that is a communicational endeavour, but not one that involves a transmission of thoughts from Aristotle’s mind to mine (much less a transmission in the opposite direction). Three important caveats must be added to my preliminary sketch if it is not to create a misleading impression of the integrationist approach I am taking. The first is that what is here being proposed is a linguistics of human history. The integrational parameters would be different in the case of, say, a history of the solar system or a history of birds of the corvidae family. The second caveat is that the three contributions I have distinguished and which, for convenience, I will call (1) res gestae, (2) historiae and (3) opinio - must not be thought of as three independently given components that are assembled at will as one might assemble individual bricks to build a brick wall. For that would be to conceive of each as already having its own historical status and would render the entire analysis regressive. It is the integration itself that confers any such status and defines res gesta e, historiae and opinio in relation to one another. My main complaint against historians is that, with the exception of some who recognize the importance of ‘oral history’ (see Chapter 7), they make little attempt to research the formation of opinio, and thus leave a gaping hole in the fabric of historification. Opinio is the work of that elusive figure whom Carl Becker, in a famous presidential address to the American Historical Association in 1931, called ‘Mr Everyman’. Berate him as we will for not reading our books, Mr. Everyman is stronger than we [academic historians] are, and sooner or later we must adapt our knowledge to his necessities. Otherwise he will leave us to our own devices, leave us it may be to cultivate a species of dry professional arrogance growing out of the thin soil of antiquarian research. (Becker 1966: 252)

THE A U T O N O M Y OF HI S TOR Y

175

As a footnote to this second caveat, it should be stressed that the relationship between all three components is dynamic. As time goes by, earlier historiae or their residue are progressively absorbed into opinio, while the rosters of res gestae are continually being confirmed, modified or reorganized accordingly. Thus an integrationist (apart from shuddering at ‘true’) would endorse Croce’s celebrated dictum that ‘all true history is contemporary history’; but only on the understanding that this applies not just to individual historiae considered in isolation, but to the whole history-making process. It should likewise be noted that, by reason of this dynamic relationship, there is no question for the integrationist of identifying historiae with texts. No one today can read Caesar’s account of the campaigns in Gaul as Caesar’s contemporaries (including those who fought in Gaul) would have read it. The Latin text may have remained unaltered, but it is the modern reading or readings that yield the modern historiae. A no less important footnote is the following. All this cuts across any traditional (reocentric) distinction between fact and fiction, and by implication expands the class of historians more widely than conservative historians might like. It means, for example, that Shakespeare becomes a historian - arguably the most influential historian of his age, and certainly more influential than his contemporary Raleigh. The same example illustrates the fluid interrelationship between res gesta e, historiae and opinio. Dr Johnson thought that in Julius Caesar Shakespeare stuck too rigidly to what Johnson called ‘the real story’. (So for Johnson there was a ‘real story’, even if it is not quite clear how we are to disentangle this story from others that were and might have been told.) But Shakespeare got the res gestae from North’s translation of Amyot’s translation of Plutarch, who in turn got them from Roman sources. Moreover, as T. J. B. Spencer points out, Shakespeare’s treatment of the res gestae was limited by the consideration that Caesar and Brutus were already ‘historical figures which were firmly established in the popular imagination’ (Spencer 1963: 19). Here at least we have a recognition of opinio and Mr Everyman. Roman history was a favourite source of material for dramatists of the period. Ben Jonson, Beaumont and Fletcher, Thomas Heywood, George Chapman, and others, all put ‘Roman tragedies’ on the stage. Shakespeare’s notably unheroic portrayal of the great Roman statesman would in turn contribute to a modification of Elizabethan opinio concerning his assassination. The third caveat concerns the use of ‘historical examples’ in my explanation of the integration of res gesta e, historiae and opinio. This might appear to presuppose as ‘historical facts’ that Caesar did conquer

176

THE L I N G U I S T I C S OF HI S TORY

Gaul, American astronauts did land on the moon, and so on, when such things are open to dispute. That would make the integrationist analysis blatantly reocentric and immediately destroy it as having any theoretically distinctive position. But it would be perfectly possible to state the relationship between res gesta e, historiae and opinio in entirely abstract terms without any such presuppositions: the ‘historical examples’ I have given are there for demonstration purposes only, to show the kind o f process history-making is. These are shop-window dummies, not being offered for sale. Given these three caveats, one key idea that integrationism brings to philosophy of history is that history-making, like language-making, involves a non-stop process of contextualization and recontextualization, in a multiplicity of ways and a multiplicity of settings for a multiplicity of audiences. I cannot see why this should not be acceptable in principle to many historians, and perhaps it is. In any case, that is how the integra­ tionist views language and other forms of communication. To the extent that language-making involves history-making, and vice versa, the two are inseparable. Furthermore, both are cumulative. The cumulative product of history-making over a period of time is a body of historified material which accrues as a layer of cultural sedimentation. A very similar process involving language-based materials is the way myths and legends build up over time, such as the traditional matiere de Bretagne in medieval France. In this respect history and myth/legend are cousins. This has an important consequence: for the integrationist, ‘Caesar crossed the Rubicon’ ends up not as a ‘truth’ about the past that the historian in person has verified or can verify, but as a reference to one particular item of traditionally historified information, just as ‘Orpheus loved Eurydice’ identifies a particular item of established mythological information. Another key idea that integrationism highlights is the impossibility of characterizing in general terms anything called ‘the language of history’. Whatever language (or other communicational devices, verbal or non­ verbal) historians may use at any particular time will presuppose assump­ tions that do not necessarily make sense for any earlier or later time. That is why, for instance, translation poses such serious problems for his­ torians. I shall return to this below. For convenience, I will call this conclusion about the absence of any permanent linguistic basis for historical statements ‘the provisionality of historiae\ That will doubtless be seen by pomophobic historians as yet another postmodern attack on their discipline. But the fallacy of postmodernism, as far as integrationists are concerned, involves seizing upon an incontestable banality, i.e. that contexts differ, and projecting it back upon the evaluation of judgments,

THE A U T O N O M Y OF HI S TORY

177

as if to produce some higher-order level of multiple truths. (Postmodern attitudes reflect a deep-seated wish to have one’s surrogational cake and eat it.) Another component of an integrationist philosophy of history concerns the communicational format in which historiae are presented. Probably this is less controversial. Different formats are appropriate for different audiences. Although it is tempting to make generalizations about what ‘the historian’ does or should do, an integrationist will insist that behind the anonymous figure of ‘the historian’ we always find particular his­ torians addressing particular groups of individuals. Thus there are different levels and layers of historification. The historia written for schoolchildren - and the language in which it is written - will differ from the historia written for the writer’s own academic colleagues. (Attempts to ‘sanitize’ the history of school textbooks have recently become a focus of controversy. This controversy itself shows how ill-advised it is to general­ ize about ‘the historian’ or about what society expects of such a person.) Different communicational formats (verbal vs. non-verbal, visual vs. non­ visual, static vs. dynamic, etc.) automatically favour or highlight different aspects of history and different types of narrative. To take an obvious example, the limitations of the hand-cranked cine-camera and the problems of setting it up on the front line seriously limited the available documentary footage of military action in World War I. M utatis mutandis, every format in which the past is reported has corresponding strengths and weaknesses which account for certain inclusions, omissions and emphases. These features imprint their own characteristic ‘profile’ on the historiae. I do not think it is possible to interchange formats salva veritate in the presentation of historical accounts. What emerges in each case is a function of the integrational devices involved in the history­ making. Probably most integrationists would agree with Carr that not any fact about the past is a historical fact, but prefer to put that less misleadingly: not every belief about the past becomes historified. All historified belief is open to dispute and revision, because history-making, like language­ making, is open-ended; but whether it is open to falsiflability in the Popperian sense is a trickier question. There can be no question of falsiflability if there was never any unambiguous question of truth versus falsehood in the first place. What misleads people here is a confusion between presumptions of factuality and truth-claims. The difference between history on the one hand and myth, legend and fiction on the other is that in one case there is a presumption of factuality and in the

178

THE L I N G U I S T I C S OF HI S TORY

other case a presumption of non-factuality. But this does not translate neatly into a set of true statements versus a set of false statements. There is a presumption of factuality in Shakespeare’s play Julius Caesar, whereas there is none in The Tempest. But it would be absurd to go through Julius Caesar trying to compile a list of the historical truths it contains. Not even the most ignorant of groundlings at the Globe, I imagine, thought that Caesar actually spoke English. But if that essential feature of Shake­ speare’s play is ‘false’ it is difficult to avoid the conclusion that the whole thing is ‘false’ too. And not just ‘false’. The notion of English-speaking Romans is historically absurd. The absurdity, nevertheless, does not preclude a presumption of factuality. Here I should make clear that what I am counting as a presumption of factuality is not to be equated with the kind of presump­ tion that some historians take to be a sine qua non for all forms of history. For example, in his discussion of ‘history and truth’ Willie Thompson claims: If history is to be distinguished ontologically from myth or fiction certain presumptions are necessary, the main one being that it is possible in principle to reconstruct, via representation, a no-longer-present reality so that it can be understood in terms not essentially different from the understanding of any segment of contemporary reality, physical or human. (Thompson W. 2000: 108)

Thompson adds immediately: ‘If the very possibility of such under­ standing is denied, no further discussion is possible.’ No further dis­ cussion with Thompson, at least. But that is not a great loss, because Thompson has already overstated the ‘presumptions’ he is demanding on behalf of historians. The very last thing our understanding of the past requires is that it should be ‘not essentially different’ from our under­ standing of what is going on in the present. On the contrary, that would mean we were really confused about the difference (as, perhaps, some historians are). There may be some simplistic historical accounts and some very naive readers fulfilling Thompson’s conditions. But to read of Caesar’s assassination as if it were an item in today’s newspaper is already to misunderstand it. Items reported in today’s newspaper have potential implications for those who may read the paper, implications affecting their immediate behaviour, whether in buying gas masks or buying stocks and shares, that have no counterpart when it comes to reading in a history book the details of Caesar’s assassination. (We do not throw down the history book and say to one another ‘Let us hasten to the forum and see how the mob may react.’) It may be that Thompson believes that we have

THE A U T O N O M Y OF HI S TORY

179

no option but to project our assumptions about the present back upon the past. That is not, however, what I construe as being involved in a presumption of factuality. I have no idea what I ‘would have thought’ about Caesar’s assassination as a contem porary event. I cannot cash that particular counterfactuality in terms of my own experience. (If that counts as a confession to having ‘no sense of history’, so be it. I can of course play speculative games of imagination about what I might have thought, if I can somehow suppress most of the things I now believe, and which inevitably colour any interest I might have at all in what happened to Caesar. But that seems to me a quite different and inherently problematic kind of game. It cannot, in any case, explain how I am supposed to understand what historians tell me about Caesar.) The presumption of factuality I am attributing to Shakespeare when he wrote Julius Caesar does not include any Thompsonian delusions about ‘representation’ or equivalences between present and past. Shakespeare represented Caesar as referring to himself in the third person; but whether that was to be construed as ‘not essentially different’ from how referring to oneself thus in Elizabethan England would be construed is dubious, to say the least. Shakespeare represented Caesar as a certain type of person, acting from certain motives understandable to an Eliza­ bethan audience; but whether Shakespeare thought that this was how the ‘real’ Caesar acted is another question altogether. What Shakespeare and his audience presumed - their joint presumption of factuality - was, quite simply, that the main characters and events in the play had identifiable prototypes in the Rome of the first century b c . N o more is necessary. So how, from an integrationist point of view, should we construe whatever ‘representations’ the historian does offer? To adapt an example from J. L. Austin, if I draw a crude sketch-map of France in the form of a hexagon, in order to show you, by marking a cross in the hexagon, more or less where Paris is, there is a presumption of factuality. The sketch serves the integrational function of linking something I think I know with your presumed ignorance of the matter. There is a presumption of factuality because I am not drawing an imaginary country that I have just invented off the top of my head (although exactly the same sketch could serve that purpose too). But it would be ridiculous to object that my sketch was worthless because it did not accurately represent the coastline between Calais and Dunkerque. That just does not enter into the relevant integrational relationship. The representation the sketch gives is neither cartographically ‘true’ nor cartographically ‘false’. It is - I shall borrow Austin’s way of putting it - just ‘rough’ (Austin 1962: 142). Shakespeare’s hist or ia of Julius Caesar’s res gestae is rough in much the same way; as is

180

THE L I N G U I S T I C S OF HI S TORY

likewise a modern historian’s account of Caesar’s res gesta e, whatever care the historian takes in assessing the available evidence or to avoid including details that a rival historian might challenge. The modern historian’s account will be rough in different respects from Shakespeare’s, but it will be rough all the same, because most of what Caesar did during his lifetime went unrecorded and is in any case of no historical interest whatsoever, just as the coastline between Calais and Dunkerque is of no relevance to my sketch-map of France. All historiae are rough. That, I think, is exactly the point Descartes was making (Descartes 1637: 1.114). If so, as an integrationist, I agree with him. Someone will object that the roughness of my sketch-map does not give me licence to mark Paris anyw here I like in the hexagon. No, agreed, it does not. I mislead you if my map shows Paris as located somewhere on the Mediterranean coast: this is not geography but fantasy (much as it might be desirable to have Paris on the Cote d’Azur). But that still does not mean that there is any clear dividing line between correct and incorrect cartographical representations. All maps of France are rough, because a line drawn on paper cannot transcend the limitations of the semiological conventions, if any, under which it was drawn. Within those conventions, questions of accuracy may arise; but not outside them. Questions of accuracy, in any case, are not necessarily questions of truth, even though they are integrational questions. It is important —for my purposes and, I think, for Austin’s too —not to construe roughness as a matter of degrees of approximation to ‘the truth’. For then we should be back to the reocentric starting point. That confusion is endemic in the distinction some literary scholars have tried to draw between ‘historical fiction’ and ‘fictitious history’. According to Mary Lascelles, we have to distinguish between a narrator offering (1) a story which does not contradict the historical record, but concerns imagined events in which historical figures feature, and (2) a story about ‘events which could not possibly have befallen the historical characters to whom they are attributed’ (Lascelles 1980: 112). Lascelles devotes a chapter of her book to ‘The historical event that never happened’ and discusses a variety of examples from English literature. An integrationist objection would be that her argument assumes that there is a single scale of approximation to ‘the truth’, along which the historian’s account always takes priority over the poet’s or the novelist’s, while lower down the same scale ‘historical fiction’ likewise takes priority over ‘fictitious history’. But why for example getting a date ‘wrong’ should demote the account to a lower rank than one in which the question is evaded by mentioning no date is never explained. Furthermore, the ranking fails

THE A U T O N O M Y OF HI S TORY

181

altogether to recognize that the historian’s account is already ‘historical fiction’ to the extent that it attributes possible but unprovable causes, effects, motivations and beliefs in the interests of rendering its narrative coherent and plausible. Those reluctant to see historical truth slip through their fingers in this way will doubtless protest. Can we not at least draw up some minimal requirements for historians? For instance, ought we not to insist that they get the names right? Or who killed whom? But then what would one make of the historicity of the rest of their account? The whole notion that at any point in time we could, if we wished, step back from, or outside, the processes of history-making and discover independently which historiae were ‘true’ and which ‘false’ is the his­ torian’s counterpart to the Saussurean conception of linguistic diachrony. It presupposes the validity of one ‘true’ perspective which turns out itself to be the product of (linguistic) theory. What has been historified as ‘the history of the English language’ is itself a theoretical construct that makes no sense without a methodology for tracing back sound changes, etymol­ ogies, etc. And these in turn have no point of departure in individual instances without a synchronic reification of the items in question. Saussurean diachrony is a failed attempt to find a place within linguistics for the ‘scientific’ study of linguistic change. It fails because it begs the very question of historical connexion. History implies con­ tinuity, not just chronological succession, and linguists constantly confuse the two. The confusion became virtually axiomatic for some linguists in the nineteenth century, as we see in Whitney’s dogmatic assertion that ‘every part and particle of every existing language is a historical product, the final result of a series of changes, working themselves out in time’ (Whitney 1867: 25). This is a splendid motto for gung-ho historical linguists. But it achieves an impressive panoramic generality by leaving the notion of ‘change’ ambiguous as between (1) cases where A is discarded and B (or perhaps nothing at all) comes along to replace it, and (2) cases where Aj is altered into a modified form A2. Changes can be of either kind. But to say that B has replaced A is to make a quite different claim from saying that Aj has been altered into A2. In the former case, it seems very odd to say that B is a ‘historical product’ of A. To say that my room at this moment, its decoration and each item of furniture standing in it, including that vase of flowers on the window-sill, is in some sense the result of a series of changes may well be indisputable. But it is to say very little, and certainly to say nothing explanatory. Perhaps yesterday or last year there were different objects in a somewhat different arrangement, and tomorrow or

182

THE L I N G U I S T I C S OF HI S TORY

next year they may doubtless be different again. Certainly the flowers will not last many days. Perhaps it will no longer be ‘my room’ at all. But to describe the room as a whole, or the things in it, as historical products of what was there previously is not merely odd but misleading. The new armchair I am now sitting in is a substitute for the old but still serviceable armchair that was thrown out a few months ago. There is no historical continuity between them: one has simply, as it happened, replaced the other. The old one could have been kept, and there might now be two armchairs in the room. Pointing to the existence of the old armchair does not in itself provide any historical explanation for the presence of the new one: if anything, it is the other way round. The acquisition of the new chair is more likely to provide part of the explanation for the absence of the old one. But in any case, the new chair is not a ‘descendant’ or evolutionary offspring of the old chair, even though they may fulfil the same function as items of furniture. Neither is the historical ‘product’ of the other. If linguists choose to believe that every item they deal with has a historical antecedent that somehow, qua progenitor, explains its ex­ istence, that is their business. But it makes their curious linguistics of history more, rather than less, esoteric. If the notion that B is the historical ‘product’ of A is problematic as regards items of furniture it becomes even more so in the case of theoretical constructs. From an integrationist point of view, the alleged realia to which statements of linguistic change refer are already abstrac­ tions, which in some cases stretch over generations or even centuries. No bystander ever saw one such abstraction being produced by another (not, at least, in the sense that someone might have seen the battle of Hastings taking place, or witnessed an old armchair being taken out and a new one brought in). Despite what etymologists tell us, the Latin word causa did not somehow evolve into the French word chose. No form or meaning ever evolved into a different form or meaning, any more than the wax candle evolved into the electric light, and it is quite misleading to adopt a terminology that suggests otherwise. On the other hand, it might make perfectly good sense to claim that over time electric lights have, at least to some extent, replaced wax candles. But if we wish to make a similar kind of claim for words or sounds, we need - as in the case of the electric light and the candle - to abandon this muddled evolutionary talk and set about trying to establish precise criteria for recognizing when and how one item may replace another. This has never been a priority in historical linguistics, because appeal to the vague notion of gradual development seemed to make it unnecessary. And if we are serious about trying, we have to be prepared to contemplate reaching the conclusion that there

THE A U T O N O M Y OF HI S TOR Y

183

may be after all no clear way of construing the proposition that French chose replaced Latin causa, just as it seems doubtful whether there is much sartorial sense to be made of the claim that the French two-piece suit replaced or is descended from the Roman toga. Diachrony is a notion that brushes the real problem of replacem ent (as opposed to mere succession) under the carpet. Diachrony makes theoret­ ical sense only when counterposed to synchrony, i.e. to the existence of an etat de langue. This in turn assumes that at any given point in the ongoing diachronic flow we can in principle stop and draw up an inventory of the current linguistic facts. Both are historical illusions, generated by a deeply flawed philosophy of language. The notion that historians ought to use a language dedicated to expressing the truth (or at least the ‘historical’ truth) is no less flawed. Whatever language the historians of one generation may have used to describe certain past events, there is no guarantee that it will meet the demands of the next generation. Or, to put the same point in a slightly different way, there is an ongoing compromise or re-integration between the opinio of one generation and the opinio of its successor. The difference is not always obvious, but in cases where it is there can be little doubt about ‘the provisionality of historiae\ For instance, it seems clear that in the nineteenth century there arose a potentially disastrous clash between various ways of talking about - and viewing - the past. Historians had to accommodate a radical expansion of the domain of history. One factor in this expansion was the Darwinian notion of evolution, which overnight rendered obsolete any attempt to restrict history to ‘recorded’ history in the tradition of Herodotus. It is interesting to note that the term prehistory does not appear until the middle of the nineteenth century (Daniel 1962), and its very appearance highlights the tacit assumption that, heretofore, history and written records had been more or less co-terminous. Developments in archeology were now making it clear that there was more to be discovered about the past than could be vouchsafed by written or oral testimony. The discovery of ancient skeletons, ancient tools and ancient paintings, all with a story to tell, exposed the limitations of Herodotean methods of investigating the past. Henceforward, the testi­ mony would no longer be that of the scribe or the teller of tales. It would include testimony revealed by the spade and the techniques of dating what the spade had uncovered. This, for an integrationist, is a good example of how the process of historification may change by reason of being required to integrate new

184

THE L I N G U I S T I C S OF HI S TORY

activities unknown to or ignored by earlier historians. Digging, as far as we know, was not an activity Livy was any good at. As the activities integrated change, philosophy of history has to be revised accordingly. So it was with the nineteenth-century recognition of ‘prehistory’. Unless historians were deemed to have become overnight competent archeologists, this meant that the good ship Autonomy o f H istory was already holed well below the water line. (Collingwood, himself an archeologist, and perhaps for that reason, never saw this.) That sudden expansion of the historical domain in the nineteenth century meant that (some of) the past was still present, if one knew where to dig below ground. Above ground, at about the same time, came the no less eye-opening discovery that many primitive communities were still living ‘in the past’ and thus offered present witness to the history of humanity. One defining moment in this revaluation of the past, as Godfrey Lienhardt points out, was Darwin’s reaction to seeing for himself the inhabitants of Tierra del Fuego. The astonishment which I felt on first seeing a party of Fuegians on a wild and broken shore will never be forgotten by me, for the reflection at once rushed into my mind - such were our ancestors. These men were absolutely naked and bedaubed with paint, their long hair was tangled, their mouths frothed with excitement, and their expression wild, startled and distrustful. They possessed hardly any arts, and like wild animals lived on what they could catch; they had no government and were merciless to everyone not of their own small tribe. (Lienhardt 1966: 10)

Darwin, in short, had the ‘time-machine’ experience: the impression and conviction - of being able to see the past in the present. And that was also the conviction of a whole generation who laid the academic foun­ dations of European anthropology. Inevitably, ‘history’ would never be the same again. Further and no less important integrational factors in this expansion of the historical domain were political and religious. In the nineteenth century, European governments presiding over a process of imperial conquest found themselves confronted with serious questions about the treatment of colonized peoples. Should administrators interfere with the customs of their reluctant subjects, or leave them alone? Should the natives be allowed, for instance, to carry on head-hunting or burning alive the wives of deceased husbands? Such questions could hardly be answered without taking a view on the duty to spread the message of Christianity. For the doubts that had formerly been raised about whether ‘the natives’ were human beings at all now seemed to be resolved. Given

THE A U T O N O M Y OF HI S TORY

185

that they were human, in spite of their nakedness, long hair, and frothing mouths, it became incumbent on historians to include them in the historical record, and on missionaries and civil servants to ‘civilize’ them. The more such questions were debated, the more apparent it became that just as the languages of the world were not simply alternative sets of labels for the same reality, so the customs and beliefs of different peoples were not just alternative ways of expressing the same set of values. But historians were slow to realize the implications of this for their own discipline and its preoccupation with the past. It is comic that, belatedly, British historians should nowadays regard themselves as under attack from fashionable French theorists (identified as the disseminators of pernicious anti-historical doctrines). They should look back, rather, to Victorian texts such as Tylor’s Prim itive Culture (1871). There they will find, in the notion o f‘culture’ itself, all the ammunition needed to debunk claims about the autonomy of history. For ‘culture’, as it was to be conceived by nineteenth- and twentieth-century anthropologists, already subsumes ‘language’ as one of its components (Marett 1929). And once that move is made, it is only a matter of time before the relativist conclusion is reached; namely, that each culture is a law unto itself, a structure sui generis, and cannot be judged or understood in terms other than its own. What other conclusion could be reached? Structuralism supplied the coup de grace. But perhaps it was scarcely needed. The pre-relativist conception of history was already on its last legs. Other flaws the integrationist detects in the traditional Western view of communication concern the very notion of ‘a language’. Perhaps the simplest way of introducing these is to consider what modifications an integrationist might like to make to the following statement by Humboldt: Only in the individual does language receive its ultimate determinacy. Nobody means by a word precisely and exactly what his neighbour does, and the difference, be it ever so small, vibrates, like a ripple in water, throughout the entire language. Thus all understanding is always at the same time a not-understanding, all concurrence in thought and feeling at the same time a divergence. (Humboldt 1836: 63)

Here we see one of the most penetrating of post-Renaissance thinkers about language struggling against the traditional assumptions of the Western language myth. The passage I have quoted deserves in itself a chapter of explication and probably a whole book, to say nothing of the

186

THE L I N G U I S T I C S OF HI S TOR Y

work from which it comes, which is the Introduction to Humboldt’s posthumously published Uber die K aw i-S prache a u f der Insel Jaw a. Two brief comments will have to serve my purpose here. 1. Once we accept, with Humboldt, that no two individuals use words in exactly the same way, we are already questioning the fixed-code determinacy traditionally assumed to reside in ‘the language’ (English, French, Swahili, etc.). For this determinacy has no basis in the day-to-day praxis of verbal communication. Humboldt accordingly relocates this determinacy at the level of the individual. (So the question already arises for Humboldt as to whether history is the same for any two individuals whose understanding of the past is based on their reading of texts.) An integrationist would go further. The determinacy, i f there is any, has to be sought at the level of the particular communication situation. 2. An integrationist would add the following rider. Such determinacy as can be found is always pro hac v ice, and always more than linguistic determinacy in the orthodox view. For it involves the integration of more factors than most linguists would be willing to call ‘linguistic’. The two comments made above amount to saying that integrationism is looking to dispense altogether with the notion of ‘a language’ as part of its theoretical apparatus. Integrational linguistics is in this sense aiming at being a ‘linguistics without languages’. ‘To say that there are no languages is not to say that anyone who talks about entities called English or French or Swahili must eo ipso be talking nonsense’ (Joseph, Love and Taylor 2001: 204). But it is not to say that they are talking sense either, or what kind of sense they might be talking. For the integrationist, those language-names (English, French, Swahili, etc.) are based on variously motivated and questionable extrapolations (social, political, educational, etc.) from first-order communicational experience. Furthermore, these extrapolations do not provide any empirical or theoretical justification for singling out determinate sets of units or activities corresponding to any one such name. It is not that there are no ‘facts of English’, but rather that English is not some comprehensive set of linguistic ‘facts’ (whether known to linguists or not, and whether known as English or by some other name). In neither space nor time do we find any clear criteria for identifying what ‘a language’ consists in. Accompanying this denial is a more general integrationist claim. Even if, for purposes of research, the linguist were deliberately to restrict attention to a community whose members all agreed that they ‘spoke the same language’, there would be no way of determining, within the totality of ongoing communicational and other activities taking place in that community, what subset of those activities constituted ‘speaking that

THE A U T O N O M Y OF HI S TOR Y

187

language’. In other words, for the integrationist it is not just the frontier between one language and another language that is indeterminate, but the entire boundary between the linguistic and the non-linguistic. The underlying reason for this double indeterminacy, according to the integrationist, is, precisely, that communication is not a separate, isolable human activity but, under all circumstances and conditions, an inte­ gration of activities. The integration is time-bound: it involves an interlocking of past, present and (anticipated) future experience. We do not speak or write, listen or read, in a temporal vacuum. As participants in a communication process, we cannot somehow ignore earlier events or other events that are going on simultaneously, nor the likelihood that still other events will follow in their train. And if per impossibile we could and did, then we should be regarded by others as suffering from severe amnesia and other perceptual or cognitive deficits. Integrational linguistics thus offers a theoretical framework in which the linguist can happily dispense with languages, with e'tats de langue, and with determinacy of form and meaning altogether, in return for recog­ nition of communication as the non-stop, locally focussed and adaptable creative process that lay experience shows it to be. That is why (as I argued in Chapter 1) it is very unwise - from an integrational viewpoint for historians to try to defend their subject on the basis of definitions of ‘the word history’. That strategy already begs too many linguistic questions. All definitions are context-bound. Presented with these radical integrationist theses, doubtless the first question a historian will be prompted to ask is: ‘Does this leave any foothold for my kind of account of the past at all?’. The answer is, perhaps unexpectedly, that it not only affords a foothold but provides an eminently defensible basis for historical accounts. With certain provisos, however. (It will not accommodate everything the more ambitious his­ torians tried to include, such as revealing the workings of divine pro­ vidence, or the ‘real motives’ of our ancestors; but then it is doubtful whether any theory could. And it will have to sacrifice overambitious claims about the autonomy of the discipline.) But at least philosophy of history will be no worse off than under the pre-integrationist dispen­ sation, where at every turn it faced an awkward choice between surrogational and structuralist apologetics. In recent years some historians have made matters more difficult by constructing a historical role for themselves as champions of common sense and old-fashioned academic values. Attempts to protect the discipline from corrosion by subversive ideas (often lumped together

188

THE L I N G U I S T I C S OF HI S TORY

indiscriminately as either ‘relativist’ or ‘postmodern’) commonly take the form of gritty or even bellicose reassertions of the old reocentric faith, accompanied by misrepresentations of the views of its critics. R. J. Evans’s robust In D efence o f H istory, although an admirable tour de fo r ce, is an example of this. It is a publication that attracted widespread praise and blame in roughly equal proportions. In Evans’s book there is no conception of any general linguistic problem about justification of historical accounts. According to Evans’s own statement of aims, he was concerned with ‘how we define and achieve such things as truth and objectivity, whatever kind of history we are practising’ (Evans 2000: 259). But the notion that ‘how we define and achieve such things’ does indeed vary according to what ‘kind of history’ one practices, and that that has anything to do with the historian’s philosophy of language, is conspicuous by its absence. Saussure, to be sure, is identified as one of the sources for the scepticism about objectivity that Evans wishes to rebut. But Evans’s (presumably ‘objective’) account of Saussure’s views would hardly pass muster at undergraduate level. We are told that Saussure argued that ‘words, or what he called signifiers, were defined not by their relation to the things they denoted (the signified) but by their relation to each other (e.g. “dog” as opposed to “cat”)’ (Evans 2000: 95). When a reviewer questioned whether the Saussurean signifie is the ‘thing denoted’, Evans replied blandly that this was indeed ‘a howler’ he had made, but dismissed this, along with other howlers that the same reviewer had drawn attention to, by claiming that they made no difference at all to ‘the central arguments of my book’ (Evans 2000: 294). Here the champion of objectivity was forced to fall back on a relativist (rhetorical) manoeuvre. He allows himself any number of historical howlers, provided the historical message is preserved. It sounds like a version of papal in­ fallibility. The reviewer, said Evans, had employed the old tactic of ‘distracting the reader’s attention’ from those arguments by focussing on the book’s ‘inaccuracies of detail on peripheral points (its brief and marginal allusions to Saussure, Derrida and Lyotard).’ It is the phrase ‘peripheral points’ in this cavalier papal reply which shows how far Evans was from understanding what was at issue here. Even less did he understand that his own defence of history involved taking a particular - and very debatable - position in philosophy of language. One might be forgiven for thinking that, as far as Evans could see, the Saussurean revolution in linguistic thought had no more direct impact on the historian’s concerns than Einstein’s disquieting claims about the concept of time had on governments using the Gregorian calendar. Here we see how the disciplinary tunnel vision of the academic historian shuts out (in

THE A U T O N O M Y OF HI S TORY

189

spite of Evans’s disclaimers) everything going on in neighbouring disciplines, or takes note of it only to marginalize its relevance. If this is the autonomy of history, it is purchased at the price of wilful ignorance. Evans’s one serious attempt to deal with a linguistic issue addresses a question the traditional historian can scarcely avoid: the vexed question of translation. This is interesting, because translation is perhaps the most obvious linguistic chink in the autonomous historian’s armour. Levine in The Autonomy o f H istory goes so far as to admit that ‘unless one can recover the original meaning of the documents (things as well as words) that provide us with access to the events of the past, historical knowledge is impossible’ (Levine 1999: xii). Taken at face value, this makes the credentials of the autonomous historian depend on expertise in trans­ lation. (The stress is particularly important for the autonomy thesis: the historian cannot afford to rely blindly on other people s translations.) Evans seems committed to a closely related position. As Evans sees it, historical documents are ‘traces’ of the past. Accordingly, there arises the question of how we translate these traces into a language that we ourselves can understand. Often they are written not only literally in a different tongue (Latin, Greek, AngloSaxon or whatever it might be) but also metaphorically, so that the same word used in the seventeenth century as in the twentieth or twenty-first may have a subtly different meaning. (Evans 2000: 89)

The way the question is put is revealing. According to Evans, it is only ‘metaphorically’ that seventeenth-century English is a different language from the English of today. He does not seem to realize - or realizes but ignores - how controversial that linguistic claim is. In his view, the ‘same word’ may be used at times differing by three hundred years or more, but remains the ‘same word’ even though its meaning has changed over time. If Evans had deliberately been trying to give an illustration of the substantialist fallacy he could hardly have done better. The example he chooses is the word fa m ily, which, he says, could include your living-in servants in the seventeenth century, but today would not (even if you had living-in servants). In this single example Evans in effect dismisses structuralist semantics, but without even realizing what he is doing. (For a structuralist, what would count is not who your family consists of, but whether the word fa m ily is now in a different linguistic relationship with servant; and that would depend on the presence or absence of other co-existing terms.) Evans solemnly assures us that semantic change

190

THE L I N G U I S T I C S OF HI S TORY

(although he does not call it that) is so serious a problem in Germany that a multi-volume encyclopedia has been produced to enable German historians to tackle it. I am sure we can all hope that the encyclopedia provides the answers to their prayers and doubts, but an integrationist would regard this as an attempt at self-levitation by tugging hard at one’s own shoelaces. Although what Evans says about translation is far from convincing, it is to his credit that he does at least say something, whereas so many accounts by historians concerning their professional activities say nothing about it at all. Jordanova devotes a whole chapter to ‘skills that historians can reasonably be expected to have’ (Jordanova 2000: 172-98), in which translation skills are not mentioned once. Similarly Marwick, in discuss­ ing the historian’s sources, refers cryptically to ‘problems arising from archaic or obscure languages’ (Marwick 2001: 182) but never confronts the issue of translation as such. Even Beverley Southgate, who discusses ‘history and language’ at some length (Southgate 2001: 74—81), manages to avoid saying anything about translation. One gains the impression that for historians translation is the skeleton in the family closet: it is there, but not to be mentioned in front of outsiders. This conspiracy of silence is essential if the autonomy of history is to be maintained. What if different but well-qualified historians produce different translations? How can we decide which is ‘correct’? Here at last Evans seems on the verge of admitting the existence of a theoretical difficulty. But no: he informs us that it is possible to reconstruct the meanings which past language had for those who used it because the individual words and concepts we come across in it were part of a system of meaning, so their meaning can be pinned down in terms of the other words and concepts used in the system. (Evans 2000: 90)

For a moment this sounds as if the Saussurean penny had dropped at long last: Evans seems to be offering a holistic concept of the signifie. But the penny evidently got stuck half-way down, for Evans shies away from drawing the Saussurean conclusion that translation equivalences between holistic systems may be impossible. On the contrary, Evans’s historian, having studied the structure of the system thoroughly, going through ‘hundreds, even thousands of documents’ in the process, is eventually in a position to ‘isolate’ the meaning of any desired word. In brief: if you want to know what an Anglo-Saxon word meant, you can find out by studying Anglo-Saxon, and this somehow makes you an Anglo-Saxon (at least in your historical imagination). Evans does not seem to realize that this old-

THE A U T O N O M Y OF HI S TORY

191

fashioned pedagogic solution brings him exactly back to starting point again. How do we learn a dead language. Not by translation? Evans commends ‘the method pioneered by Quentin Skinner, of fixing the meaning of a past text by comparing the language it uses with the language used in other contemporaneous texts’ (Evans 2000: 276). ‘Pioneered’? It sounds to a linguist as if at least one historian had belatedly tried to apply the principles of structuralist lexicology already familiar to linguists for several decades. But the principles do not seem to be adopted very systematically in Skinner’s own practice. For instance, in M achiavelli (Skinner 1981) the author simply refuses to translate Machiavelli’s key term virtu on the ground that ‘it cannot be translated into modern English by any single word or manageable series of para­ phrases’ (Skinner 1981: vi). But there is no attempt by Skinner to identify the meaning of this term by a semantic analysis of ‘the language used in other contemporaneous texts’. Refusing to translate the term is Skinner’s way of avoiding the very pertinent and difficult question of whether Machiavelli himself always used virtu in the same sense: Skinner simply declares ex cathedra that Machiavelli used the term ‘with complete consistency’. But how Skinner knows this remains undisclosed. The assumption throughout Skinner’s book is that Skinner knows what virtu meant for Machiavelli. The historian’s claim to cross-cultural linguistic omniscience could hardly be more blatant. He has somehow managed to crack the structural ‘code’ of Florentine Italian in the early sixteenth century, but is keeping the code-book to himself. The question of translation happens to be as good a peg as any on which to hang a comparison between an integrational approach and the more usual theoretical alternatives. 1. R eocentric translation. Roughly: the translator must find a word that ‘stands for’ exactly the same thing as the word to be translated. Thus, when referring to features of the landscape, what makes French arbre the correct translation for English tree is a fact of nature: a tree is a tree on either side of the English Channel (or anywhere else in the world, for that matter). 2. P sychocentric translation. Roughly: the translator must find a word that expresses exactly the same idea as was in the mind of the original writer. There is no fact of nature to appeal to. Thus arbre correctly translates tree if and only if arbre expresses what was originally expressed by tree in the text to be translated. The complication lies in finding out what was in the original writer’s mind when using the word. 3. Contractualist translation. Roughly: the translator must find a word

192

THE L I N G U I S T I C S OF HI S TOR Y

with a role in the vocabulary of the target language which maps as closely as possible on to the role occupied in the source language by the word to be translated. Thus the general justification for proposing arbre as a translation of tree will include such considerations as the following: both words are common nouns that function grammatically in comparable ways in sentence structure in their respective languages, both are similarly related to words for ‘bush’, ‘branch’, ‘leaf, ‘wood’, ‘forest’, both have comparable extended uses (fam ily tree : arbre genealogique), etc. 3a. Structuralist translation. Roughly: when each word is defined holistically in terms of the total vocabulary, it turns out that there are strictly no interlingual equivalences. Any impression that there are is a linguistic illusion. As semantic systems, English and French are anisomorphic. Thus however closely arbre appears to match tree, there will always be irreducible differences not apparent on first inspection. These differences will show up in cases where an English expression including the word tree has no corresponding French expression including arbre (e.g. apple tree, but not * arbre pomme\ shoe tree, but not * arbre chaussure). 4. Integrational translation. Roughly: translation is always ad hoc and indeterminate. There are no interlingual translation equivalences (although for rather different reasons than those proposed by the structuralist). Arbre may sometimes - or often - work as a translation of tree, but that depends entirely on the communicational context. There is no guarantee that what works in one case will work in another. There are no ‘rules’ for translation. What sets the integrationist position apart from the others is that 1, 2, 3 and 3a are all code-based. In general, code-based approaches to trans­ lation offer two types of theory: ‘transference’ theories and ‘mimetic’ (or ‘replicational’) theories. When Dr Johnson defined translate as ‘to change into another language, retaining the sense’ he was offering a transference theory. Transference theories characteristically treat translation as a set of procedures for extracting a ‘message’ from a text in one language and re­ encoding it in another language. Thus the ‘object’ of translation is the message, not the text itself. On the other hand, mimetic or replicational theories treat translation as a set of procedures for ‘copying’ a text, or certain selected features of a text, from one language into another. Here the ‘object’ of translation is the text itself, and the aim is to produce (as faithfully as possible) a matching text. In philosophy, a powerful case for the indeterminacy of translation has been made out by Quine (Quine 1960; Quine 1970; for criticism, Kirk 1986; Harris R. 1986). The integrationist case is more pragmatic and layoriented. It rests on the observation that the practice of translation rarely

THE A U T O N O M Y OF HI S TOR Y

193

requires the translator to address the general question of the semantic relationship between expression a in language A and expression b in language B. That issue is in most cases a red herring, even supposing that the question can be formulated with sufficient precision to make a clear answer possible. What the translator needs to do is to make sure that, in a particular context and for particular purposes, expression b will not mislead the reader of the translation; at least, no more than expression a might have misled a reader of the original text. To describe the task in this negative fashion is not a strategem of perversity, but reflects the psychological reality of interpreting texts, in whatever language: the reader is engaged in a constant struggle against their opacity. Writing never shows what it says. If it did, it would not be writing. For the integrationist, however, indeterminacy of translation is no more than a particular case of expectable indeterminacy of meaning in all forms of communication. As a translator myself, and frequent user of other people’s translations, I have never thought that I was giving or getting a ‘true’ version of the original, but merely - I hoped - a useful substitute (useful, that is, for certain limited purposes) but a potentially hazardous one. So I would regard it as a mistake in the first place to identify any source narrative with the source text: for ‘no tale is told with absolute completeness’ (Morris 1998: 315). The reader must always ‘supply’ (as it is often described) more than the text gives; but no single reader knows everything that might be supplied, or whether supplying any specific item or items is too little or too much. For texts can be overinterpreted as easily as underinter preted. It is likewise a mistake to identify the source narrative —as a transference theory will do —with the meaning of the text, and for the same kinds of reason. There are interesting mimetic theories of translation (e.g. Catford 1965) but I shall say no more about them here because, as far as I can see, in practice Western translators from antiquity onwards, insofar as they have ventured to make their views explicit, usually seem to be following some kind of transference theory (Nida 1964; Kelly L. 1979). Transference theories have been analysed by Haas as having the following structure. They present translation as an operation with three terms: two expressions and a meaning they share. When we translate, we seem to establish a relation of three distinct entities, each separately apprehended: the two expressions seen on paper or heard in the air, and the meaning in the translator’s mind. The meaning, presumably, we ‘retain’ and translate; we ‘transfer’ it from one expression to the other. Strictly, then, when a sentence or speech or novel is translated, say, from French into English, what is supposed to be translated or transferred is not the French sentence or speech or

194

THE L I N G U I S T I C S OF HI S TOR Y

novel at all; it is something utterly different, something inaudible and invisible the ‘meaning’ itself [. . .]. (Haas 1962 : 208)

As this analysis makes clear, transference theories depend on the adoption of a certain metaphysics of meaning; and to anyone who does not accept the metaphysics the theoretical account proposed will seem entirely unconvincing, and probably obfuscatory. The first point to make here is that it will not do for historians who value the autonomy of history to pooh-pooh the metaphysics, or to say that it lies outside their province qua historians: they do, after all, claim to be reliable translators of texts from the past. (Or at least Evans make this claim on their behalf.) So if they are not espousing transference theories at all, we need to know what different criteria of competence underwrite their claim. The metaphysics in any case cannot be ignored. For i f ‘meaning’ is to be admitted at all as relevant to translation, it becomes important to know how it is to be identified. And Haas proceeds to make a claim that many would reject utterly; namely, that ‘the meaning itself which is allegedly ‘transferred’ in translation is ‘not in French nor in English nor in any language whatever’ (Haas 1962: 208). In short, the meaning of the text is not only inaudible and invisible but something divorced from linguistic expression altogether. Saussure, for one, would have totally disagreed, and so would integrationists. Less controversially, Haas also points out that a transference theory presupposes a particular theory of the linguistic sign; namely, that it is a bipartite unit, uniting a form and a meaning. Languages, we might add, are tacitly envisaged as sets of such units, plus rules for their combination into sentences. In short, languages are construed as codes, and the problems of translation are seen as arising not from the ‘difficulty’ of the text but from the incompatibility of the codes. The perceived mismatching is often taken by structuralists, and particularly in those versions of structuralism that embrace the ‘Sapir-Whorf hypothesis’ (Whorf 1939; Mathiot 1979; Werner 1994: for criticism, Lounsbury 1969; Black 1969), as powerful evidence against reocentric semantics. It does not take much reflection on transference theories of translation to realize that they are no more than extensions of certain ideas that derive directly from the traditional language myth of Western culture. The notion o f‘transferring’ a meaning from an old text to a new text is calqued on the more general assumption that words function telementationally to transmit thoughts from one mind to another. In translation, all that happens is a simple ‘code-switching’ in the middle of the process. Exactly

THE A U T O N O M Y OF HI S TORY

195

how it happens may be mysterious, but that that is what happens is not in doubt for a moment. An integrational approach to translation rejects all these mysterious processes as phantoms conjured up in the wake of the basic theoretical mistake: this consists in reifying words as invariants that have been given a collective authorization in order to make linguistic communication possible. Instead, integrational semiology proposes to see communication in terms of the continual creation of new signs, rather than the use of old signs already supplied from some hypothetical community store. That creative process is identified in integrationist theory with the integration of different activities at a variety of biomechanical and social levels, and in an endless variety of different circumstances. Thus,/#r from being invariants, signs have to have the malleability that fits them to serve all kinds of unforeseen purposes, as well as the familar routines of daily interaction. It is the latter type of deployment that, uncritically scrutinized and taken as ‘typical’, creates the impression that somewhere and some­ how there must exist a fixed communal inventory of signs from which we can all draw tokens whenever we need to, rather like drawing cash from a perpetual cash machine. Such an impression is reinforced, at least in literate societies, by educational systems based on the preservation and circulation of authoritative texts, and, in recent centuries, on deliberate attempts to impose linguistic uniformity and repress variation, usually for political reasons. The compilation of dictionaries and grammar books serves to reinforce this repression. But although language develops in the individual as a ceaseless accretion of communicational experience, its application to a current communication situation —the integrationist will point out - is always a new application. Nor is this surprising. Our past experience as pedestrians is never sufficient to guarantee that we shall never get knocked down when crossing the road. And regardless of how many runs I have scored in the past, there is no guarantee that I shall score any at all in next Saturday’s match against Ship ton-under-Wych wood. Interlingual translation, for the integrationist, is an inescapably risky enterprise in which the translator is forever proposing verbal substitutions that cannot be shown to hold as general interlingual equivalences. They are proposed nevertheless because - in that particular instance - the substitution is better than nothing. At least, in the translator’s estimation. Others may take a different view and declare that it would have been better to admit untranslatability than mislead the reader. Translation can never be ‘verified’: it attempts to bridge an unbridgeable gap by creating ad hoc connexions between texts and communities. This is not to say that gross errors cannot be detected; but it does not follow from this that there

196

THE L I N G U I S T I C S OF HI S TORY

must have been all the time a ‘correct’ translation that the translator failed to give. Translation too is ‘rough’. How ‘good’ a translation is will depend on the integrational parameters set by the communication situation in question. These will include the aims - and claims - of the translator as well as the expectations of the reader. The contention that the meaning is invisible, inaudible and not ‘in’ any language whatever is a form of linguistic mysticism. The transferencetheory alternative is to suppose that the ‘meaning’ is somehow bound up in the form of words in which it originally appeared. But then we are back with the old problem of explaining how, in that case, it can be transferred to a quite different form of words in a different language. These equally unwelcome options are the cleft stick of translation theory. But if the historian claims to be a reliable translator, we are owed at least an indication of which theoretical option has been selected. Otherwise, the credentials of this honest toiler sifting through mountains of documents and dozen of dictionaries are automatically called in question. Hard work does not validate translation. The integrationist can offer the translator a way out of that particular dilemma; but the way out comes at a price. The price to be paid is dropping all the theoretical options discussed above. For some translators, that would come cheap. Translators of poetry might even welcome it as a liberation. But can the historian afford it? For part of the historian’s case — or some historians’ case - seems to be that, by means of correct translation, (some of) ‘the truth’ about the past has been preserved. And this might appear to sit awkwardly with the integrationist view of translation as a creative act of communication, which both requires and produces new signs that come with no semantic guarantee. If the historian is to be an integrationist, that linguistic responsibility cannot be shirked. The ‘truth’ in the translation, if there is any, will have to be the historian’s truth. History-making becomes truth-making. Translation thus provides an important test case for the autonomy of history. For translation is an integrational mechanism that the historian can hardly dispense with. But where is the evidence that the historian understands this mechanism? Where are the treatises on translation by these allegedly autonomous historians? Where do they reveal the solutions that non-historians have spent much time discussing, but failed to produce? The historian has no professional credibility in this domain: for determining how many troops the Greeks had at the battle of Thermopylae is clearly not on a par with determining how to translate Plato’s various uses of the term logos.

THE A U T O N O M Y OF HI S TORY

197

As Collingwood pointed out, it is only by insisting on the dubious translation equivalence between an ancient Greek term (techne) and a modern English term {art) that the art historian can conjure up something called ‘Greek art’ and provide a chronological narrative to go with it. Similarly, as Kenneth Appiah observes, it is possible for the Guggenheim to mount a major exhibition o f‘African art’ when neither the term Africa nor the term art corresponds to any concepts that were in the minds of those who, in the past, made the objects on display (Appiah 1997). The integrationist point is not just that patently misleading and inadequate interlingual extrapolations are possible, but that, however dubious, they are often necessary in order to set up the categories which allow discussion of the categorized subject to take place at all. It is by means of such extrapolations that historians divide up the incalculable vastness of ‘the past’ into the history of X, the history of Y, and so on ad infinitum. The history of X in turn subdivides into the history of X/, the history of x2, etc. An endless process of reflexive parturition, justified at every turn by the generation of its own discursive products. This is indeed autonomy of a certain kind, but hardly the autonomy of history that Collingwood championed. Not least of the reasons for dismissing any notion o f‘the end of history’ as premature is that the process will presumably continue for as long as it can draw on the communicational resources of self-renewal. But that is a conclusion about communication, not a guarantee from the past.

CHAPTER SEVEN

The truth about history

‘What is truth?’ asked jesting Pilate, and would not stay for an answer. On the question of truth, historians and philosophers of history fall into four main categories (with some sliding from one to another). There are those who maintain or assume the following positions. 1. Establishing the truth is the historian’s main aim. This, although often difficult, is in principle possible and in some cases attainable. 2. All that historians can do is establish probabilities. The truth about what actually happened can never be ‘proved’ conclusively. 3. There is no such thing as ‘the truth’, but there are multiple truths, depending on the viewpoint adopted. These truths are not necessarily consistent one with another. 4. Truth is illusory, a rhetorical construct or a strong conviction disguised as an independent reality.

Accepting or rejecting any of these positions is meaningless unless underpinned by some explication of the concept ‘truth’. Many historians have at this point followed Pilate’s lead and made for the exit. But some philosophers of history at least stay long enough to mention the corres­ pondence theory of truth. Whether this will suffice to put historians’ truth-claims on a firm philosophical footing is a different matter. Aristotle’s way of stating the correspondence theory (M etaphysics 101 lb26) leaves an awkward problem, for it does not tell us how we are to tell, when something is said to be the case, whether it is the case, i.e. what ‘being the case’ consists in. Except, circularly, conforming to what the case is said to be. The problem here for the historian is that in most statements about the past there is no question of inviting the inquirer to ‘look and see’ whether or not it is the case, as one might respond to someone who questioned the veracity of the statement ‘It’s raining.’ (‘Look out of the window.’) No one can now look and see whether Caesar crossed the Rubicon. So the present statement ‘Caesar crossed the 198

THE TRUTH A B OU T HI S TORY

199

Rubicon’ cannot be linked to anything currently observable that might be a plausible candidate for its ‘being the case’ that he did. Eric Hobsbawm, acclaimed as ‘one of the outstanding historians of our age’, takes a manifestly Aristotelian stand when he states that ‘without the distinction between what is and what is not so, there can be no history.’ Rome defeated and destroyed Carthage in the Punic Wars, not the other way round. [. . .] relativism will not do in history any more than in law courts. Whether the accused in a murder trial is or is not guilty depends on the assessment of oldfashioned positivist evidence, if such evidence is available. Any innocent readers who find themselves in the dock will do well to appeal to it. It is the lawyers for the guilty ones who fall back on postmodern lines of defence. (Hobsbawm 1997: ix)

Were it not for Hobsbawm’s reputation, one might suspect this of being a parody of the reocentric position on historical truth. Like the court, the historian sits in judgment on the past, and that judgment determines guilt or innocence. But this attitude conflicts with that of those historians who protest that ‘it is for others to judge’ (Evans 2003: 20). In a shrewd discussion of the relevance of truth to historians’ accounts of the past, Leon Goldstein criticizes Mandelbaum’s view that the correspon­ dence theory is presupposed in every work of history (Mandelbaum 1938). Mandelbaum had insisted on distinguishing between, on the one hand, what makes an error an error and, on the other hand, the grounds for error. The former is the failure of correspondence, not the latter. Which assumes the possibility of confronting the erroneous statement with what actually happened and comparing the two. But does this comparison ever take place? Goldstein points out that ‘when we examine the practice of historical research itself we never discover a point at which we test the claims of a historian against the actuality of the past’ (Goldstein 1976:41). Instead, what we find are juxtapositions of one statement with another statement, some being accepted as reliable and others rejected; but never a point at which a statement is matched against an independently accessible state of affairs in order to determine whether the two correspond. Presumably the nearest one could get to this would be on-the-spot checking of an eye-witness account against the events as they were taking place (as in ‘It’s raining’). If one is listening to a live radio commentary on an event at which one is actually present (as is nowadays a mundane occurrence at international sporting events), it is indeed possible to come to a judgment about whether or not, or how well, the commentator is describing what happens. In the case of past events, where one was not present, this is not possible.

200

THE L I N G U I S T I C S OF HI S TORY

If historians abandon the correspondence theory, as it seems they must, can they find solace in the rival philosophical theory, which is the socalled ‘cohesion’ or ‘coherence’ theory of truth? This theory casts a true statement as being one that is consistent with a large and systematic body of other such statements. (Thus, to determine the truth or otherwise of the mathematical statement ‘2 + 2 = 4’, we look not to any correspond­ ing state of affairs in the external world but to other mathematical statements within the same system.) This seems to be more congenial to Goldstein’s way of thinking about historical truths, for he writes: We have no access to the real past in any way that would make the correspondence theory of truth - no matter how sophisticated a version of it one might devise applicable to historical claims to knowledge. We do have access to the historical past, but our only approach to it is historical research itself. [. . .] Our problem is the fairly concrete one of determining whether or not the actual discipline of history as it is practiced by historians is adequate to the task of establishing warranted claims to knowledge. (Goldstein 1976: 195)

Thus it would seem that the best the historian can hope for is to give an account consisting of statements which are consistent with other state­ ments, including those of other authorities, witnesses of past events, statements about surviving artifacts or archeological finds, and so on, the whole body of these forming an internally self-supporting edifice of claims. So the entitlement of ‘Caesar crossed the Rubicon’ to be con­ sidered ‘true’ would depend on how securely it was lodged in some such edifice, and not on an ascertainable relationship with a past event involving Caesar and the river in question. This looks, on the face of it, like adopting a much more modest stance than that of the correspondence theorist. For it comes down to admitting that, where the past is concerned, truth is consensual, and the historian is the accredited professional in charge of establishing the relevant con­ sensus. But, although more modest, is it not also a leap from the frying pan into the fire? For it affords an obvious re-entry for the unresolved problem of translation and for more dire semantic problems as well. What methodology is available to enable the historian to determine whether one statement is consistent or inconsistent with another, particularly if the two are made at different times by different people? Worse still, however, is the intrinsic failure to give a convincing explanation of what is meant by making such statements as ‘Caesar crossed the Rubicon’ or saying that they are ‘true’. Here the correspondence theorist can fight back and argue that taking something to be a historical truth is not just a matter of taking it to be something agreed among themselves by historians. The consensus

THE TRUTH A B O U T HI S TORY

201

is hollow if it cannot bridge the gap between the statement and what Caesar did. Nor will it do for the cohesion theorist to respond that the consensus is of course a consensus about what Caesar did; for once the correspondence theory has been abandoned there is no longer any ground for explaining what it is for a statement to be ‘about’ something in the ‘real world’. Where the integrationist differs from the cohesion theorist hinges principally on this latter point. Once a society has accepted the historification of a certain corpus of beliefs or assumptions about its past, history has already been made. For many generations of Greeks it just was a fact that Paris abducted Helen and thus precipitated the Trojan war. All that dissenters can do - if they think it worth the candle - is campaign for a few small adjustments in the internal structure of the corpus. Such minor ‘re-historifications’ are likely to be accepted only when they remove existing anomalies in the internal structure, or fill gaps in it, and not because they supposedly reflect more accurately something outside it altogether. To cite a present-day geographical parallel, it is possible to argue that Paris is on the river Yonne, not the Seine (because the Yonne has the major water flow at the point of junction). But the chances of integrating that minor re-geographication into maps and atlases are slim, because ‘Paris is on the Seine’ has long since ceased to be an empirical proposition. It has become a traditional part of the framework within which French people (and foreigners) conceptualize the geography of France. For similar reasons, historians nowadays are not likely to waste much time over when and where Paris ‘really did’ abduct Helen, or Caesar ‘really did’ cross the Rubicon, and, if so, how you could ‘prove’ it. This, it should be noted, is not to dismiss such questions as ‘meaningless’. They are meaningful, but have simply been by-passed in the ongoing process of historification. You can always recoup such details and demand a reassessment. The mistake consists in thinking that the reassessment will reveal God’s truth about it. Or, if not exactly God’s truth, a better approximation to God’s truth. The reocentric temptation for the historian to play God’s emissary on Earth is strong. This is not to deny the possibility of replacing one whole corpus of historified beliefs by another. For the integrationist, this is always possible. Some such replacements, arguably, are taking place at the present time. There may even be rival contenders for these replacements. And that is probably one reason why history is a discipline in crisis. But such replacements rarely take place piecemeal, by arguing about indiv­ idual ‘facts’, such as Caesar’s crossing or not crossing the Rubicon. They reflect more general shifts in attitude to whole sectors of the past and to

202

THE L I N G U I S T I C S OF HI S TORY

the political and moral values seen as being reflected in what happened. Thus, for example, I assume that no British historian would nowadays write a history of the British Empire up to the end of the nineteenth century in the way it would have been written in 1910. This it not because individual Tacts’ have been re-dated or re-located but because the imperial corpus as a whole has been reviewed in the light of changed attitudes to imperialism. I would be very surprised to find in a history book published in 1910 an account of British rule in India which included a passage like the following: The alienated native rulers were not fools. They knew that the annexations were more often than not driven by British strategic and financial interests, rather than by any high-minded commitment to ‘improved’ government. The massive ex­ ercises in data-gathering, like the Great Trigonometric Survey of India, were intended in the first instance to supply military intelligence. The railways and trunk-road extensions would make not only the penetration of Indian markets but also the deployment of troops easier and faster. (Schama 2002a: 318)

In my terms, the key to understanding the change between 1910 and 2002 would not be that new Tacts’ about the British in India had come to light, nor a change in any direct relationship between res gestae and historiaey but between both and opinio. At the opposite end of the scale in the process of historification come those cases, popular with certain historians nowadays, where the historia concerns some forgotten episode that the historian has unearthed and endowed with a new ‘significance’ which allegedly throws light on some aspect of the past. Robert Darnton’s essay on ‘the great cat massacre’ (Darnton 1984) is as good an example as any. The setting is a printer’s workshop in the rue St-Severin in Paris in the late 1730s. Two rebellious apprentices, realizing that the pampered cat of their master’s wife leads a better life than they do, decide to take their revenge. One of them climbs on the roof at night above the master’s bedroom and keeps the couple awake by piercing caterwauling. After a succession of sleepless nights, the printer declares war on the neighbourhood cats, and his employees round up and kill or maim all they can lay hands on. In the general melee of the cat hunt, they do not fail to dispatch the detested cat of the master’s household, but hide the corpse. The captive survivors of the local cat population are then solemnly put on trial, found guilty and hanged. The printer’s wife bemoans the disappearance of her favourite, but cannot pin the blame on anyone. The printer complains about the lost working hours, but cannot deny that he ordered the massacre. Subsequently the

THE TRUTH A B O U T HI S TORY

203

whole episode is mimed repeatedly by the apprentices, to the great amusement of themselves and their workmates. This story, recorded in an obscure memoir published by one of the participants some twenty years later, is retold and reinterpreted by the historian after an interval of two centuries. With the benefit of hindsight, and drawing heavily on later work in European folklore, the historian discovers arcane meanings in all the details. The cat is an animal traditionally associated with women, sex and witchcraft. The apprentices took revenge not only by killing the particular cat they loathed, but by symbolically raping its mistress and cuckolding the printer. It was a ‘metonymic insult’. The whole episode is seen as prefiguring the develop­ ment of ‘working-class militancy’ and the September Massacres of the French Revolution. Whether this is a plausible interpretation is not the point. From an integrationist perspective, it represents a different mode of historification. The historian is careful not to claim that the participants themselves interpreted it thus. But what his treatment does is give the tale a ‘historical meaning’ that it would otherwise lack. It becomes a semiological link in a chain that goes back to the superstititions of the Middle Ages and forward to the revolutionary dream of liberte, egalite,fratern ite. It will be evident that the way I have presented the above arguments and cases was deliberately designed to draw attention to the parallels between certain positions in philosophy of history and certain positions in philosophy of language. These parallels, I would maintain, are difficult to deny or discount. The historian’s dilemma over admitting translation as one way of getting historical ‘evidence’ mirrors the conflict between certain theories of meaning. The awkward choice between hitching historiae to a ‘correspondence’ or a ‘cohesion’ interpretation of truth reflects the clash between reocentric and contractual theories. The semantics of the great cat massacre rests on an appeal to cultural etymologies. Nor are these parallels fortuitous. On the contrary, they arise because in the Western tradition the historian always operates within the space afforded by certain assumptions about language, and in that space the possibilities are strictly limited by acceptance of a long-standing myth about language. That overarching language myth holds the various possibilities together and, at the same time, generates the various tensions between them, as I have tried to show. But more remains to be said about communication, truth and factuality. Presumptions of wow-factuality are sometimes made quite explicitly.

204

THE L I N G U I S T I C S OF HI S TORY

The characters and situations in this book do not portray, nor are they intended to portray, any real person, persons or institution, living or dead. The characters, events, corporations and firms depicted in this film are fictitious. Any similarity to actual persons, living or dead, or to actual corporations or firms, is purely coincidental.

These are not commitments to telling all and only lies. The reasons why such declarations are made are to be found in a social context in which litigious people can invoke laws in order to sue authors, publishers and film-makers for misrepresentation, and the guilty parties may have to pay out large sums of money. That is the integrational function of the declaration. We might think of historians as traditionally operating under an undeclared presumption which, if declared, might read: The characters, situations and events in this book portray, and are intended to portray, real persons, real situations and real events, unless indicated to the contrary.

However, it is not difficult to cite works of high quality that make a mockery of this rigid distinction. One example would be the short stories of Saadat Hasan Manto. In Manto’s collection M ottled Dawn (1997) only one ‘historical figure’ appears, the Muslim leader Mohammad Ali Jinnah, and even then in a minor role; but Manto’s snapshots of life at certain times and places portray the political, social and psychological horrors of the partition of India more vividly than any conventional historian’s account. Although ‘fiction’ (just how ‘fictional’ is debatable), they are stories that have already - I would maintain - become part of, or are becoming part of, the public opinio of this controversial episode in the history of the subcontinent. Because history-making is ongoing, that is just what one might expect. I find it interesting to compare Manto’s M ottled Dawn with Natalie Zemon Davis’s absorbing essay ‘The rites of violence’ (Davis N. Z. 1975), which is concerned with religious persecution in sixteenth-century France. The comparison suggests itself because both Manto and Davis - in my view - are fascinated with the mystery of how supposedly ‘civilized’ human beings can quite easily bring themselves to behave in quite uncivilized - even barbarous - ways towards their fellows and neighbours, given certain historical conditions. (Here there is no ‘relativist’ dissent about what counts as barbarous. All those involved - narrators and dramatis personae included - recognize the barbarity.) But the differences in treatment are striking. Davis is concerned with such

THE TRUTH A B OU T HI S TORY

205

matters as the definition of the term religious riot and the ‘social meaning’ of such events. Whereas I imagine that for Manto these concerns would bear witness to an academic fa ilu re to confront the reality in question. Historification as practised by the professional historian, in other words, typically holds reality at arm’s length by interposing an apparatus of classification and explanation, complete with meticulously documented notes, references and statistics. By reason of this interposition there is indeed a paradox in the very notion of ‘objectivity’, if that is supposed to be what the professional historian is thereby achieving. For it is achieved by what an integrationist would call decontextualization, a deliberate distancing of the ‘object’ from any serious engagement with it in human terms. (And within ‘human terms’ I include, it hardly need be said, the emotions.) Under the aegis of this decontextualized concept of ‘objectivity’, I seem to see the ominous reappearance centre-stage of Hobsbawm’s reocentric judge-historian, appointed to officiate over sorting out the follies of humanity and pronouncing sentence. In Manto’s world, no one is sentenced. No one is excused either. What happened is never given an ‘interpretation’ (as in the manner suggested by Davis’s title). That might look like the refusal of a challenge. But reading Manto (even at one remove, i.e. in translation) makes one realize that this is no such abdication. On the contrary, it is perhaps the measure of a superior historical consciousness. How are presumptions of factuality realized in textual practice? This is a crucial question in the linguistics of history. One of the important integrational mechanisms is that of the proper name, that I have already drawn attention to in earlier chapters and must come back to now. For an integrationist, neither a descriptive theory of proper names (Russell 1918) nor a causal theory (Donnellan 1972) is satisfactory as it stands, although both incorporate acute observations about proper names in ‘ordinary language’. What is missing in both cases, however, is a failure to recognize that proper names are no less context-dependent than any other type of linguistic expression. In an integrationist discussion of the linguistic category ‘proper name’, Paul Hopper argues that proper names are always ‘emergent’ (Hopper 1990). They emerge ‘in direct response to the referential needs of the discourse’. If this is so, then I suggest that in historical writing there are broadly two types of case to consider. The first may be illustrated by the following passage from Powicke’s classic The R efor­ mation in England:

206

THE L I N G U I S T I C S OF HI S TORY

Most good men were in agreement upon the moral issue. They regretted the loss of the opportunity to convert ecclesiastical property to religious uses, which in their eyes included social and educational uses, although most monastic property was, of course, essentially secular in character and always subject in England to the common law. Latimer, as is well known, held this view, and Cranmer in the end of Edward VTs reign resisted the wanton spoliation of the Church. But by this time the greed of Northumberland and his satellites was but part of a policy which would have destroyed the traditional status of the Church of England and might well have involved its ultimate reconstruction on a non-episcopal basis, as abhorrent to Ridley and Cranmer as it was to Gardiner and Bonner. (Powicke 1941: 35)

Here the reader is just expected to know who the dramatis personae were: they are simply identified by name - names such as Latimer, Cranmer, Northumberland, Ridley, et al. Nor have they been previously introduced in the narrative, except in the same ‘you-of-course-knowwho-I-am-talking-about’ manner. Contrast this with the following extract from Christopher Hill’s Intellectual Origins o f the English R evolution: Take Robert Recorde, for instance, ‘the founder of the English school of mathe­ matics’, who died in 1558. He was a Cambridge graduate, but failing to establish himself in either university he moved to London, where he earned his living as a physician, and gave popular lectures in mathematics. He also depended on the patronage of the Muscovy Company, to which many of his works were dedicated. His writings were all in English, and long remained the standard popular textbooks: his arithmetic ( The Ground o f Artes) had at least twenty-six editions between 1540 and 1662. (Hill 1997: 17-18)

Here, evidently, the reader is not expected to know who Robert Recorde was, and not, perhaps, even to have heard of him; but, once introduced, he reappears frequently and is thereafter referred to simply as ‘Recorde’. No such introduction is given to anyone in Powicke’s Reformation. Doubtless it will be said that there is nothing very remarkable here. We do, after all, when introducing strangers on social occasions, often feel it appropriate to explain briefly who they are, where they come from, or how they came to be present. (‘This is Bob Flybynight. He’s a visiting fellow of All Souls.’) But the point about historical accounts such as those cited above is that in such accounts, unlike on social occasions, the ‘real person’ in question is always missing. What we are left with, as Russell would have insisted, is just a description; and where there is no descrip­ tion all we have is just the name itself. In order to understand who is

THE TRUTH A B OUT HI S TOR Y

207

referred to we have no clue but the occurrence of the name in a particular context in a particular work. At the opposite extreme from this stands the historical biography, where the identificatory information is expanded to become the work itself. Thus whereas Hill tells us only as much about Robert Recorde as is required for understanding his role in Hill’s story, a biography of Recorde would try to include as much about the man and his career as could be discovered (his place and date of birth, his parentage, his upbringing, etc.). That still does not blunt Russell’s point: for what is a biography in the end but a long description? An integrationist would add that however long the description, it is still context-bound. In verbal description itself there is no linguistic bridge to the external world; for a fictional life can be just as minutely described. Whatever presumption of factuality there may be lies outside the words, and not inside. So, as readers, we are still left with the integrational work to do on our own, inferring from the contextual clues what connexions to make. Furthermore, as any reader of Hill’s book will discover, although people like Robert Recorde are duly introduced with an appropriate identificatory description, including dates, there is no similar description attached to more familiar names ( William the Conqueror, H enry VIII, Francis B acon, etc.). In short, there is a twotier structure of names. The ‘you-of-course-know-who-I-am-talkingabout’ names provide a framework for the ‘perhaps-you-have-neverheard-of names, the latter nested inside the former. So this still does not get us outside the historia: it actually serves to close off the supply of information available. Nor do we get outside the historia when the name Jinnah occurs in Manto’s narrative. What the name Jinnah does is simply establish a presumption of factuality. But what is remarkable is the scrupulous care, apparent to the attentive reader, that Manto takes to avoid any pres­ umption that Jinnah was a ‘good guy’ or a ‘bad guy’ (except to the principal character in the story). The integrationist point is that Manto had to do that, because otherwise his story might have been construed as making a ‘historical’ statement. (For which he might have been prosecuted. He was prosecuted anyway, but on the bizarre charge of ‘obscenity’. The historical establishment can always sniff out its enemies.) Of recent years, as I mentioned much earlier, historians have been exercised by the so-called ‘linguistic turn’ in their own discipline. And one of the questions that the ‘linguistic turn’ seems to raise in a particularly acute form is whether historical truth is not a linguistic

208

THE L I N G U I S T I C S OF HI S TORY

illusion. From an integrational point of view, however, the linguistic turn itself appears to be based on a pot-pourri of linguistic confusions. I have already referred to Mink’s philosophical diagnosis of the historian’s linguistic difficulty as turning on the impossibility of ever referring to ‘events as such’ (Mink 1987: 199). Here integrationists could at least help Mink to see what the problem actually is. What is worrying Mink is a puzzle which arises out of his own semantic assumptions. In his essay ‘Narrative form as a cognitive instrument’ he is bothered by the availability of 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’ pre­ supposes 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 linguistic conundrum and Mink’s analysis of it are generated by tacitly adopting a reocentric semantics. What Mink evidently fails to see is just that. He wants to have events structured in such a way that individual 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 is in supposing that there must be some unique and independently specifiable portion of the past that will act as the basis 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, for the integrationist, on the specifics of that particular episode of communication. In no less desperate linguistic straits is Hayden White, who has been described as ‘the most influential historian of historical ideas and practices writing today’ (Evans 2000: 276) and ‘the most articulate and persuasive proponent of [. . .] the “high” narrativist position’ (Rayment-Pickard 2000: 280). White’s approach to history (White 1973; White 1978) comes in direct line of descent from the literary critic Northrop Frye (Frye 1957).

THE TRUTH A B OUT HI S TOR Y

209

Its only ‘linguistic’ basis is the work of Roman Jakobson, and this because of Jakobson’s recognition of a ‘poetic’ function in language. For White, the poetic function reveals a neglected dimension in every historical account. In Jakobson’s view, stylistics must seek to analyze the poetic dimension of every merely putatively prose discourse [. . .]. This conflation of the prosaic and the poetic within a general theory of discourse has important implications for our understanding of what is involved in those fields of study which, like historio­ graphy, seek to be objective and realistic in their representations of the world but which, by virtue of the unacknowledged poetic element in their discourse, hide their own subjectivity and culture-boundedness from themselves. (White 1978: 104)

Whether White ever understood Jakobson’s account of the ‘poetic function’ is doubtful, but that is not a matter worth pursuing here. Jakobson’s theory of linguistic communication is in any case a code-based theory, with the familiar ‘sender-receiver’ model at its core (Jakobson and Halle 1956; Jakobson 1960). White appears to accept all this as uncontroversial. He then proceeds to impose his own misinterpretation on what Jakobson has written. If Jakobson is right, then historical writing must be analyzed primarily as a kind of prose discourse before its claims to objectivity and truthfulness can be tested. This means subjecting any historical discourse to a rhetorical analysis, so as to disclose the poetical understructure of what is meant to pass for a modest prose representation of reality. (White 1978: 105)

Jakobson never suggested any such thing. In the paper to which White refers, Jakobson says nothing about historical writing at all (Jakobson 1960), much less that before ‘claims to objectivity and truthfulness can be tested’, there must be a prior rhetorical analysis. How that priority could be achieved methodologically White never explains, nor why in any case it is not possible to test for truth without any rhetorical analysis at all. We may, doubtless, find it difficult nowadays to test ‘Caesar crossed the Rubicon’ for truth (at least, under a correspondence theory); but that difficulty will not be resolved by what any rhetorical analyses might conceivably bring to light about the poetic function of the words in which the statement is made. What Jakobson does say - and White ignores - is: ‘The verbal structure of a message depends primarily on the predominant function’ (Jakobson 1960: 353). Which, in the case of historical accounts, would be what Jakobson calls the ‘referential’ function. So it would be extremely odd, if not perverse, for historians - and their readers - to give priority instead to the poetic function.

210

THE L I N G U I S T I C S OF HI S TOR Y

What is going on here? We are looking at a desperate but ill-informed attempt to find a reputable basis in linguistic theory for the postmodern ‘narrativist’ attack on the traditional historian. The irony is that the attackers do not understand the linguistic theory themselves, let alone that if they did understand it, their attack would be undermined rather than supported. The troops are sent into battle, protected only by the postmodern theoretical ointment which they have been indoctrinated into believing can deflect all wounds. What has tended to happen in recent years is that the lunatic fringe of postmodernists has joined forces with the lunatic fringe of Marxists in their eagerness to debunk traditional (or ‘bourgeois’) history, with its ‘grand narratives’ of the past, allegedly constructed to promote the interests of the ruling classes. Having accomplished this demolition job to their own satisfaction (but no one else’s), the radicals then encounter a dilemma. Should they renounce history totally as a bourgeois invention, denying the existence of historical truth altogether, and claiming that ‘the historian’s work reduces to its ideological positions’ (Wordsworth 1987: 116)? But then they would have slain a paper tiger. Furthermore, such a renunciation would leave no credible locus standi for a postmodern history at all. The alternative is to opt for installing their own version(s) of history in place of the discredited product. But then they find themselves adopting much the same ‘historical methods’ as their debunked predecessors, re-assessing the same or similar evidence, and so on. So what emerges eventually is a compromise: The removal of ‘objective truth’ as a meaningful goal is counterbalanced by a perceived need for many different accounts of the past - none claiming any special privilege, but each providing some illumination from its own perspective. (Southgate 2001: 9)

The problem with this pluralist solution is that not all of these different accounts of the past, although each may be internally coherent, can be accepted simultaneously. It is the old difficulty of reconciling dis­ crepancies, already familiar long before Derrida appeared on the scene. Without quite realizing how they have reached this position, the post­ modernists find themselves facing a snag endemic in the coherence theory of truth: an impasse is reached when conflicting ‘coherences’ confront one another. Under cover of this discordant confrontation, the notion of ‘objective truth’ that was so noisily ejected waits for surreptitious re­ admission by the back door. For without it there is no basis for rejecting any account of the past (other than one’s own personal preferences and

THE TRUTH A B OUT HI S TOR Y

211

ideological sympathies). Here Becker’s Mr Everyman emerges trium­ phant as ‘his own historian’. Some postmodern historians evidently think that the solution is to fragment the notion of historical truth into a plurality of different truths. Thus Richard Price in A labis World (Price 1990) presents a ‘multivocal narrative’ in which the various ‘voices’ of the Saramaka people, the Dutch officials and the German Moravian missionaries are separated out (and presented visually in different type faces) in an attempt to describe a situation in the Dutch colony of Suriname in the eighteenth century. The rhetorical technique is original. But an integrationist would point out that it solves no historical problem. The twentieth-century historian is still in charge of integrating those ‘voices’ on the printed page (and deciding how many to recognize). It clearly matters how an integration between res gesta e, historiae and opinio is achieved. One key to this lies in the format in which the historiae are presented. We cannot abstract from this and suppose that it makes no difference whether the historian opts for one format or another. It makes a great deal of difference, granted that those options o f integration are available. For an integrationist, it goes without saying that a written historical account is a different communicational enterprise altogether from an oral account. In pre-literate societies, a written account is not an option at all. But in early literate societies it often happens that features of oral accounts are carried over into the written accounts that supersede them. Trad­ itional narrative formats as they emerged in Greek and Latin literature are almost certainly, in various respects, hangovers from an earlier, pre­ literate tradition, which survived in Europe for centuries. Much of what went back to Homer (where the ‘historical’ debate started) was still going strong in the European Middle Ages in such formats as the chanson de geste. Although eventually ‘written down’ and thus surviving as texts, the chansons de geste were originally oral narrative compositions. They were the stock-in-trade of professionals who thereby earned a living. That demanded skills of extemporization and a memory for formulaic ex­ pressions that no writer required (Lord 1960; Parry 1971). The Bodleian manuscript of the Chanson de Roland is manifestly a fossilized, static version of a dynamic oral original (although at how many removes from the original it is difficult to say). The narrativists have a point. At least, they have a point if their critique of the traditional historians boils down to saying: ‘The conventional narrative, in the Western tradition, with its Aristotelian beginning,

212

THE L I N G U I S T I C S OF HI S TORY

middle and end, and its cast of dramatis personae, is a form quite unsuited to capturing all the realities of the past. It is a Procrustean literary bed into which historians have forced their material, using all the rhetorical devices that are traditionally associated with narrative form.’ If this is so, then the question of whether historians have ‘got their facts right’ is overshadowed by a much more fundamental question: what kinds of ‘facts’ does the traditional narrative form allow? And what kinds of ‘facts’ are excluded? Clearly, the ‘facts’ that are favoured are ‘facts’ about human agents, their doings, their achievements and their conflicts, whereas the ‘facts’ that tend to be disfavoured are those better presented in maps, tables, graphs and columns of statistics rather than in story form. But it would be odd to conclude from this that historiae in a literate society have to exclude all ‘facts’ that are not readily narratable. On the contrary, a book, as opposed to an oral account, offers ample opportunity to include relevant figures, calculations, diagrams and illustrations of all kinds. So the narrativists’ song-and-dance about narrative form seems to be something of a red herring. Some of the ‘new historians’ - for instance, Emmanuel Le Roy Ladurie in Le Carnaval de Romans (Le Roy Ladurie 1979) - have already broken with the traditional narrative format without any assis­ tance from narrativists. My discussion of narrative in the context of an integrationist linguistics of history seems to call for a more explicit account of what I am assuming a narrative to be. Here I take my lead from Michael Toolan, whose ‘first attempt at a minimalist definition’ reads: ‘a perceived sequence of nonrandomly connected events’ (Toolan 2001: 6). The attraction of this, for an integrationist, lies in the role attributed to perceptual involvement. In other words, my perception - or somebody’s perception - is necessary for integrating events into a narrative. (The events could be connected in all kinds of other ways, even non-random ways. But they do not become items in a narrative until perceived as sequentially ordered and inter­ connected.) It seems to me to follow from this that how they are perceived as narrative will depend in part on the material form in which they are presented. It also follows, from an integrationist perspective, that historical narrative is not confined to verbal narrative. When the past is presented pictorially, different factors come into play. One of the outstanding examples from the ancient world was Trajan’s column (Rossi 1971). This told the (official Roman) story of the Dacian wars in the format of a marble scroll of reliefs, arranged in spiral around a vertical axis. Such a

THE TRUTH A B OUT HI S TORY

213

presentation already constrains the narrative in ways that demand a totally different appreciation from that which might be appropriate if reading, say, Dio Cassius’s comments on the Dacian wars. From an integrationist perspective, there is no way we can give Dio Cassius priority qua historian over the master-sculptor of Trajan’s column; or vice versa. That Dio Cassius was a second-rate bumbler, whereas the designer of Trajan’s column was a genius, makes no difference either. Some may doubt whether the term narrative is properly applicable to such a work as Trajan’s column, given that, whatever inferences the spectator may perhaps make, the work itself contains no actual statement of any connected sequence of events. But that would be to adopt a narrower definition of narrative than the definition Toolan is proposing. In the composition of Trajan’s column there can hardly be any doubt that the artist constructed the whole set of images as a ‘perceived sequence’ (although one which in practice it is difficult for the spectator to follow closely without having either a telescope or a specially constructed viewing platform). The principle of ‘perceived sequence’ is evident, in the first place, from the adoption of the spiral arrangement, which allows an unbroken ribbon of scenes to be superimposed on the seventeen circular marble drums of which the column is physically constructed. The spiral, in other words, itself assigns a continous visual sequencing, without which that particular format would lack any rationale. But it is an open question whether this visual sequencing is more important than other visual patterns that the spectator may notice; or whether the composition as a whole conforms to Richard Brilliant’s definition of ‘truly visual narrative’: one which ‘has no visible text to which it must defer’ (Brilliant 1984: 17). Here we come back to basic questions of communication in a literate society. The only ‘visible text’ in the case of Trajan’s column is the dedication at the base. But that itself, although giving no detailed commentary on the visual narrative, leaves us in no doubt as to the historical purpose of the monument. Which in turn presupposes a familiarity with current opinio and with texts now lost, including Trajan’s own commentarii on the Dacian campaigns, which the artist commissioned could hardly have failed to consult and respect. An integrationist might add that it is in any case obvious that this mounument is the product of a literate society. For what the designer has done - with considerable ingenuity - is adapt to the demands of sculpture the physical form of a written document, the Roman volumen. In the volumen, the text is normally inscribed on the inside of the roll; here the artist has displayed his visual narrative on the ‘outside’ of the roll.

214

THE L I N G U I S T I C S OF HI S TORY

Furthermore, as in the volumen, it is designed to be ‘read’ from left to right. In short, the sculptural document is a gigantic volumen, 3 feet wide and 670 feet long, unrolled to its full length and wound round a column 100 feet high. The next question is how adopting this non-verbal format affects telling the historical narrative of the Dacian campaigns. One requirement, patently, is to introduce visual conventions that will express certain information that the verbal historian could have stated explicitly, but which the visual historian cannot. The direction of the gaze, gestures and movements of the figures indicates the actual orientation of the action concerned. Facing forward stands for going east or north-east (in the direction of Dacia, or downstream on the Danube), while facing backwards means going west or south-west (in the direction o f Roman territory or upstream on the Danube). (Rossi 1971: 18)

(Similar ‘visual conventions’ are sometimes found in modern docu­ mentary films. Viewers of the BBC’s The Great War may have been surprised to observe how many left-handed British soldiers there were in 1914—18, and failed to realize that the film-makers had adopted the convention of having the British always attacking from the left and the Germans from the right.) The continuous ‘ribbon’ sequence on Trajan’s column makes it possible to achieve, in principle, a strict correspondence between order of scenes and chronological succession. To what extent the artist attempted to achieve this is difficult to determine, because of the problem at this distance in time in identifying particular episodes or even individual persons. However, it seems clear that the narrative is in part structured by the recurrence of certain types of scene: these include battles, sacrifices, building operations, captives and marches (LehmannHartleben 1926). Trajan, whom Brilliant describes as ‘the hero of this dramatized m ythos of victory’, appears prominently in person at regular intervals, as ‘the awesome protagonist whose effect on events is decisive, almost as if it were he alone against the Dacians’ (Brilliant 1984: 97-9). More interestingly, Brilliant suggests that the artist’s treatment of the subject, the recurrence of motifs, certain vertical alignments and the strategic positioning of certain images reveals the conception of history underlying it. ‘Given the conflicting requirements of continuous narrative and symbolic correspondence, the master of Trajan’s column invented the helical composition as an integrating medium’ (Brilliant 1984: 104). This ‘historian in stone’ was operating in accordance with an ‘obli­

THE TRUTH A B OUT HI S TORY

215

gation to reveal the truth’ rather than to record a mere sequence of events. In this respect, Brilliant sees the artist-historian’s conception of the task as conforming already to Humboldt’s account: An historical presentation, like an artistic presentation, is an imitation of nature. The basis of both is the recognition of the true form, the discovery of the necessary, the elimination of the accidental. [. . .] For it is the greatest virtue of a work of art to reveal the inner truth of forms which is hidden in their actual appearance. (Humboldt 1821; cited in Brilliant 1984: 100)

Ah, truth again; and Humboldt again. But ‘inner truth’ this time. Perhaps we are being invited to see the outer truth as told in the narrative helix of the column, and the innner truth as indicated by the non­ narrative devices, including the artful juxtaposition of images across the helical division of sculptural space. Then we would have to look for truth in quite different places than it might be found in a text by Livy or Tacitus. Truth is format-dependent. What a format can deliver depends on its integrational structure. Take the case of the H istory o f P hotography by Helmut and Alison Gernsheim (since replaced by Helmut Gernsheim’s two large volumes The Origins o f Photography and The Rise o f P hotography 1850-1880). The plates re­ produce images that have already become landmarks in the public opinio of this art form. They include, for example, Nicephore Niepce’s view from his window, and Carjat’s portrait of Baudelaire. (In many cases the images are now more solidly historified than the names of the photo­ graphers.) If we imagine a book which contained just the images, but no accompanying text at all, that could hardly count as a history of photo­ graphy. On the other hand, Gernsheim’s text without the images, although still a history, would be significantly impoverished. Why? Because, in the first place, the detailed discussion of those images in the text would be difficult to follow, and in some places impossible. In other words, it is misleading to suppose that the pictures are included as illustrations of the text: the relationship is quite different. The pictures show what the text presupposes. Their juxtaposition to Gernsheim’s comments constitutes a form of deixis. It would be no less misleading to say that this is a case where the historian has been able actually to present the historical evidence, instead of merely citing it. The pictures themselves prove nothing: they might have been altered, or Gernsheim might have attributed them to the wrong photographers. They are neither illustrations nor evidence but an integral part of the format the historian

216

THE L I N G U I S T I C S OF HI S TORY

has chosen, which combines showing what cannot be said with saying what cannot be shown. Whenever we have a format that integrates verbal material with non­ verbal material on a large scale, that itself affects narrative structure and narrative content. A much earlier - and quite different - example is the Bayeux tapestry. Here the Latin text plays second fiddle to the pictures. The subordination is evident in the positioning of the Latin sentences as ‘captions’ to what is shown in the tapestry. There are even instances where the chronological succession seems to have been reversed for no better reason than a more attractive pictorial composition (Gibbs-Smith 1973: 9). While it cannot be gainsaid that without the text it would be hard to make very much sense of the pictorial narrative, the ‘events’ in the story are selected for their pictorial value. We become aware that in visual historiae there is a different concept of ‘detail’ altogether: what is included answers to different demands from the details of verbal narrative. Thus, for instance, Harold {dux Anglorum) is shown at the beginning setting out for Bosham, apparently taking with him all the equipment for medieval hunting, including a hawk and a pack of hounds. Well, perhaps he did. But whatever criteria you apply, the dogs and the bird hardly merit this bid for historification. They play no further role in the story. Nor does Harold’s journey to Bosham deserve a mention, except that it provides a splendid pictorial panel. Similar considerations apply to the contemporary phenomenon of ‘telehistory’, where, like the medieval jon gleu r, the historian is earning an income based on attracting as large an audience as possible. Simon Schama, a highly-paid jon gleu r of the small screen, has himself drawn attention to the existence in the Middle Ages of what he calls ‘a strong, unofficial tradition of performative history’ and credited A. J. P. Taylor (‘grand-daddy of all television historians’) with reviving that tradition (Schama 2002b: 44). It is a tradition that allegedly operates in a quite different way from that of ‘print historians’. With this an integrationist would wholeheartedly agree. And perhaps an integrationist might go further and say that between the two there is no common measure. This is not to say that one never has the opportunity to see a television programme and read a book on the same historical subject by the same jongleur. Such opportunities are nowadays commonplace, and observers sufficiently sophisticated constantly point out the differences. For instance, Dorothy Thompson, reviewing the final volume of Schama’s A H istory o f Britain, says that ‘a splendid TV history loses its lustre in print’ (Thompson D. 2003). That is understating the case. What happened was that a very successful television presentation could not

THE TRUTH A B OUT HI S TORY

217

be satisfactorily translated into an illustrated book (as the publishers evidently wanted for commercial reasons), because the narrative structure imposed by a sequence of scenes does not automatically correspond to a sequence of verbal/paginated statements, even when supported at inter­ vals by impressive, high-quality photographs (including a double-page spread of Queen Victoria’s funeral cortege). In the TV programme, the backbone of narrative is the sequence of scenes, not any sequence of verbal statements, even those made by the presenter. (They serve, rather, to link the scenes.) There is no question here of which version tells (or tells better) the historical ‘truth’. We have two different formats, one more successful in this case than the other: but there are as many different types of format as there are possibilities of integrating various modes of communication. Schama won his spurs as a print historian before anyone decided to find out whether he was any good as a telehistorian. But that is typical of a communicational transition in history-making, and Schama seems to grasp more clearly than most just what is involved in such a transition. In his BBC H istory Lecture (Schama 2002b), he lists four ‘mistakes’ about history that underlie the criticisms often directed against telehistory. They are: (1) ‘that real history is essentially coterminous with the printed book’, (2) ‘that only printed text is capable of carrying serious argument’, (3) that only ‘hewers at the rock-face of the archives’ have ‘the esoteric knowledge’ that is historical knowledge, and (4) that telehistory is only viable to the extent that the conclusions of print historians ‘are faithfully translated and reproduced on television’. His analysis is one that any integrationist would endorse. But Schama does not go the extra mile and note that in certain non-verbal formats of historification, presumptions of factuality can actually be translated into ‘seeing’ the past (where ‘seeing = believing’), as opposed to seeing what actors suitably dressed up perform for an audience. This is no longer a form of propositional history (where ‘facts’ are reduced to their statement in words), nor a form of dramatized history (where ‘facts’ are supposedly re-enacted). Nonpropositional, non-dramatized history can show us the relevant inform­ ation. The makers of documentaries have no need of verbal descriptions whenever they can incorporate into the format of their historiae film footage of historical events themselves. (All they need to resist is the occasional temptation to make right-handed soldiers left-handed in order to present a visual narrative that is easier for the viewer to follow.) What distinguishes telehistory, from an integrational perspective, and puts it in a history-making class of its own, is the technological possibility its format affords of integrating all previously known formats.

218

THE L I N G U I S T I C S OF HI S TORY

For all that, telehistory a la Schama remains a form of minority entertainment, its significance eclipsed by a related but more profound revolution in modern communication. With the advent of television news reportage, history entered a new era. b c and a d were replaced by b t and a t (’Before Television’ and ‘After Television’). The chronological rules of the game changed once more when the daily or even hourly telecast became the community’s main way of keeping track of national and global events, and the diffusion of that information could, in certain circum­ stances, affect the course of events itself. The biggest mistake it would be possible to make in philosophy of history would be to think of television news bulletins as mere ‘chronicle’, or to regard them as no more than a form of incomplete visual documentation, analogous to the b t historian’s written records. The integrational structure of the format is completely different in the two cases. What the seemingly ubiquitous television camera can do is capture the flow of history in vivo. Television coverage has the potential for instant historification. It does not need to wait for retrospective endorsement by the print historian. At its most impressive, television’s account of an event or sequence of events makes the b t historian’s valiant attempts to ‘reconstruct’ what happened or ‘bring it alive’ seem pathetically inadequate. Whatever verbal postscripts histor­ ians may write to 9/11, nothing will either add to or shake the conviction of those who watched the unforeseen events unfold on their television screens that they were already seeing history in the making. That sense of unimpeachable witness is something that many b t historians strove to create, but failed for want of a format through which it might be achieved. Here at last is a format in which you are invited to ‘see for yourself: the ultimate autoptic test. No historian can make it more real. Doubts about whether what you are seeing ‘truly’ happened can only be doubts of the Orwellian variety or else profound mistrust of the senses: and for neither of these is there any known cure. A much more subtle problem about truth is posed by the emergence of ‘oral history’ in the second half of the twentieth century. Subtler because, in spite of the designation ‘oral’, oral history is actually a construct of print historians. An aura of uncertainty surrounds the term oral history. It is an unhappy designation which suggests (1) that normal history is written, and (2) that, in spite of this, ‘oral’ history achieves what written history achieves, only in oral form. From an integrationist perspective, this would be self­ contradictory. Clarification is not helped by perplexing claims such as ‘oral history is as old as history itself (Thompson P. 2000: 25), which

THE TRUTH A B OUT HI S TORY

219

seems to presuppose that there is something identifiable as ‘history itself, with which oral history is coeval, or perhaps that oral history has always existed alongside non-oral history. Thompson cites a passage from Michelet, in which the author of Histoire de la revolution fra n fa ise speaks of what one can hear of an evening in the village tavern, or in conversation with a passer-by on the road. For Thompson, it seems, Michelet is talking about oral history; whereas in the terms I have used above, it seems clear that what Michelet is talking about is just one kind of contribution to opinio. Later in his book, however, having discussed the collection and sorting of pieces of oral evidence, Thompson addresses a final question: ‘How do we make history from them?’. So, after all, the role of historymaker is reserved for the professional print historian. This seems confirmed by the definition of oral history as ‘the inter­ viewing of eye-witness participants in the events of the past for the purposes of historical reconstruction’ (Perks and Thomson 1998: ix). We are not dealing here with the spontaneous formation of opinio but with solicited testimony. It is not spontaneously solicited either. There are handbooks instructing historians in the techniques of conducting recorded interviews. And then there is the editing of interviews before they make their appearance in quoted form. Here at last we - sometimes find recognition of genuine ‘linguistic’ difficulties: ‘The same statement may have quite contradictory meanings, according to the speaker’s intonation, which cannot be represented objectively in the transcript, but only approximately described in the transcriber’s own words’ (Portelli 1998: 65). However these difficulties might be resolved, if they can be, it remains clear that the objective set by the print historian is a transformation of oral communication into written communication, with the assumption that nothing need be ‘lost’ in the process. But this, for an integrationist, is an extremely problematic assumption, and at worst an attempt to bully us into believing that there is a format-neutral historical truth. It looks suspiciously like the re-emergence of a code-based semantics, plus the naive belief that equivalences can be established across codes. Some oral historians are at least aware of the issue. Peter Read points out that this is ‘a wider problem than how to turn voice into print’ (Read 1998: 414). It is, essentially, a problem about transforming one kind of communicational activity into another and somehow conjuring ‘history’ out of the transformation. What began as ‘a personal, intimate and one-toone communication’ becomes something which is ‘impersonal, public and one-to-many’. The role of the interviewer in this process ‘raises a variety of ethical questions and responsibilities’ (Read 1998: 415).

220

THE L I N G U I S T I C S OF HI S TORY

Indeed it does. But an integrationist would press the point further. Moving the discussion into the area of ethics risks diverting attention from the basic communicational question. It is not just a matter of whether your informant’s naive account has been extracted under false pretences and used for purposes the informant perhaps did not fully appreciate or might not approve. The more fundamental issue is whether the integrational process initiated by the interviewer does not inevitably mould the account itself. Is there any disembodied ‘account’ per se, which existed independently of its elicitation, its possible contextualization and recontextualization, and of any use that might or might not be made of it? Or is that a convenient theoretical assumption made by the interviewer and based on the tacit adoption of a surrogational semantics? An integrationist will suspect the latter. One reason that prompts this suspicion is the readiness with which some oral historians resort to a glib psychocentric interpretation of what their informants say. Responding to doubts about the ‘accuracy’ of what informants allegedly ‘remember’, Alistair Thomson stands the objection on its head and claims that ‘the so-called “unreliability” of memory can be a resource rather than a problem for historical research’ (Thomson 1998: 23). In order to prove his paradoxical case, Thomson cites two examples of questionable accounts given by former Australian servicemen concerning their enlistment in World War I. In one case, the informant according to Thomson —gave an account biassed by the subsequent formation of his own political views, and wanted to ‘forget’ the difficult choice it actually presented at the time. In the other case, the informant likewise ‘forgot’ his own previous antipathy to life in rural Australia, and the fact that enlistment provided an opportunity to escape. In both cases, we have the historian playing God and giving us a commentary on what truly happened, as opposed to what the informants ‘said’ happened. In both cases, God’s role is predicated on a psychocentric semantics: the informant’s statement is treated as meaningful only as an expression of his current beliefs about the past, as distinct from a veridical account of previous experience. This linguistic manoeuvre certainly enhances the historian’s status as an authority on what happened; but what linguistic justification there is for it remains obscure. How, then, does an integrationist deal with the seemingly intractable problem of historical truth? By recognizing the fallaciousness of suppos­ ing that true and fa lse have fixed meanings, and by acknowledging that it is foolish to try to place upon lay metalinguistic terms greater theoretical weight than they can possibly be made to bear (a burden that Western

THE TRUTH A B OUT HI S TORY

221

philosophers have consistently tried to thrust upon them over the centuries). From an integrationist perspective, there is no way that the alleged meanings of true and fa lse can be projected as eternal values into such varied domains of communication as history, politics, religion, aesthetics and mathematics (to mention but a few). In the stages most of us go through in the course of our communicational apprenticeship to the adult world, honesty stands opposed to lies, deceit, or evasion; and these are matters of personal responsibility. (‘Are you telling me the truth?’) It may be abundantly clear that some (revisionist) historians are telling lies about the Holocaust in World War II, and some Serbian politicians are telling lies about the battle of Kosovo in 1389. And it may be fairly easy to see what purpose the lies serve in both cases. But there is a considerable difference between being able to detect a lie and believing that there must be a unique ‘truth’ out there waiting to be told. Truth is context-sensitive. That is one epigrammatic way of stating the integrationist position; but it may be misleading. For it seems to suggest a contrast with a possible alternative in which truth is context-free. The beginning of wisdom where Pilate’s question is concerned is not to set about some desperate search for a definition to provide an all-time answer that would shut Pilate up, but to note how the opposition true fa ls e can be used in different ways in different communication situations. Like all metalinguistic terms, true and fa lse cannot be decontextualized without dire consequences for their interpretation. The notion of ‘absolute truth’ to which some philosophers have appealed is the product of just such a decontextualization. As regards history, there are no statements about the past which must be regarded as true come hell or high water: that brand of linguistic absolutism is incoherent. But it does not follow from this that the truth is whatever we declare it to be. Playing off rigid absolutism against please-yourself relativism is the last thing that will help us make sense of history. Telling historical lies, deplorable though it may be, is also a manifes­ tation of historical consciousness. Historical consciousness takes many forms, some sick and some healthy. When it serves such causes as jingoism and racism, we deplore it. When it prompts ecological concerns for future generations, we applaud it. And perhaps we are not quite sure what to make of it when it sponsors the curious notion that we ought to give ‘historical apologies’ for the alleged wrongdoings of our ancestors to the historical descendants of the wronged parties. The difficulty which pervades the discussion of such questions is the readiness with which people foist a reocentric interpretation on its terms. Hence my liberal use of scare quotes around ‘true’, ‘false’, etc. Not that I

222

THE L I N G U I S T I C S OF HI S TORY

can claim to have used this metalinguistic device consistently throughout my discussion, for I have no idea how to go about setting up a ‘rule’ for scare quotes (even one that I could follow to my own private satisfaction). Nor would I know what wise persons I might consult about setting up such a rule or rules. The bottom line here is that {pace Aristotle and all subsequent reocentric theorists, and regardless of whether you care to call it ‘truth’ or anything else) there is no universal dimension of rightness that applies to all statements alike, depending on whether what is asserted is so or not so, and irrespective of the circumstances of assertion. The point I am making is not new. It was already made by Collingwood, whose objection to decontextualized syllogistic logic was that we cannot tell what a statement asserts unless we know the question which the statement supposedly answers (Collingwood 1944: 27). Unless this is known, two apparently incompatible statements may turn out to be perfectly com­ patible after all. The objection, however, goes deeper than questions of compatibility. To see this one need only consider how unjust it would be if a schoolboy who got his sums wrong were punished for not telling the truth. Or how ridiculous to upbraid the weather forecaster who wrongly predicted snow as a liar. There is, to be sure, a question of whether it did or did not snow. But it is just as confusing to treat what the forecaster said as a matter of ‘truth’ or ‘untruth’ as it would be to treat it as a question of being ‘polite’ or ‘impolite’. These terms are all metalinguistic labels. They are not vacuous and they are not unimportant; but nor do they escape, by reason of being metalinguistic, the conditions of indeterminacy to which all nonmetalinguistic vocabulary is subject. You can, if you like, call the unfortunate weather forecasters liars. That will not prevent me (or anyone else) from saying: ‘No, they are not liars: they just made a mistake’; or perhaps ‘They made no mistake: the weather, as it happens, changed unpredictably.’ Or even, if I am sufficiently casuistic: ‘What they said was absolutely true, because all weather forecasts have an unstated ceteris paribus clause which protects them against the eventuality that it might not turn out that way.’ It takes a philosopher (or a historian?) to decontextualize ‘truth’ and treat it as an entirely person-neutral relationship between a sentence and a ‘state of affairs’. For an integrationist, that already removes any possibility of understanding the complexity of an important network of beliefs (about what is ‘true’) that enter into human communication in all kinds of ways, some of which have very little in common. #

#

*

THE TRUTH A B OUT HI S TORY

223

By this time methinks I hear a clamour arising from enraged objectors stamping on the floor and complaining that the integrationist offers no definition at all of the term context, no explanation of where a context begins or ends, no formal identification of its ‘parameters’, no account of how we know whether something belongs inside the context or outside. Quite right. Well objected. Alas, the reply will not satisfy the objectors, and may even fuel their fury. The reply is that contextualization is what you are doing right now as you fulminate against such an unsatisfactory response, and the context is the framework in which you do it. Whatever that context may be, you know more about it than I do. But I do know that you have to be engaging in contextualization of some kind if you are to make any sense at all of what you are reading. Otherwise your rage is inexplicable; or at least dare I say - irrational? Stop there, for a moment. Has the integrationist already (selfdefeatingly, some may claim in triumph) reified ‘context’, ‘truth’, ‘idea’, ‘sentence’, etc.? If he appears to have done so, it is for purposes of engaging in discussion with those who deny they are guilty of any such misdemeanour. Guilty or not, it does not stop me from going on to say that whether a government spokesman tells the truth about the current economic situation is not on a par with whether you tell the truth about who you were with last night, or whether a geometer tells the truth when claiming that the area of the square on the hypotenuse of a right-angled triangle is equal to the sum of the areas of the squares on the other two sides. And anyone who cannot see the differences deserves all the theoretical confusions that ensue. This may sound at first like giving historians ample opportunity to retreat once more behind an ‘ordinary-language’ defence of history and an ‘ordinary-language’ conception of truth. But not quite. What the inte­ grationist is saying is not that ‘we all know’ what the words true and fa lse mean, so there is no call for the historian to bother about explaining them. The historian is in the same leaky linguistic boat as the rest of us. There is no question of directing the operation from some safe place on shore. (Narrativists please note the ‘metaphor’.) On the contrary, the historian who presents an account as ‘true’ incurs the obligation to tell us, or try to tell us, just what that implies in that particular case. We could all drown anyway; but that is no reason for the historian to shirk the responsibilities of volunteering to act as navigator. It is up to the historian to make clear at some point what kind of communicational authenticity the historia is presented as having in that particular case, which means explaining how its various components and

224

THE L I N G U I S T I C S OF HI S TORY

materials are selected and integrated. That is a considerable demand. Historians are not accustomed to meeting it, because for generations they have coasted along on the built-up prestige that accrues to their discipline. History is now prisoner to its own processes of historiflcation. The demand I am voicing is not met by listing ‘sources’ in footnotes: that merely defers the accountability. It requires the historian to come clean about, for example, what has simply been taken over unquestioned from earlier historians, what has been subjected to fresh research, what relationships have been reinterpreted through the perspective adopted, what has been ignored because deemed to be irrelevant, and what has been highlighted because it happens to have a bearing on contemporary concerns. A historian may well protest that such a demand would require extensive expansions or annotations not just to every paragraph, but to every sentence, and perhaps to every other word. So it might. But if ‘truth’ is what is being claimed, a proliferation of caveats cluttering the text seems a small price to pay in the service of such a noble cause. An integrationist typology of historiflcation, which I do not intend to attempt here, would have to recognize all the differences between orality and literacy, in their various guises, and the differences of context appropriate to each. It would have to recognize the compromises that occur when verbal and non-verbal communication interlock. It would have to avoid begging questions about the relations of words and images to ‘the past’. It would have to face up to the problems of translation instead of avoiding them. It would have to discount all the self-serving linguistic assumptions that historians make to bolster their own credibility. It would have to do many other things. But it would be a step forward if historians stopped taking communication for granted and started doing this bit of theory for themselves. Perhaps some would be reluctant to embark on it, because it might involve recognizing that, just as journalists during the course of the twentieth century had to realize that newspapers were no longer the privileged vehicles for news, so historians must realize that history books are no longer the privileged vehicles for history. The historian’s communicational role has already changed and is still changing. To some this will sound like just another postmodern attack. Admit­ tedly, what the integrationist is saying is that there is no hard-and-fast or universal distinction between historical and non-historical accounts, because the criteria to which accounts of the past may be expected to conform are open-ended. This is a much more revolutionary idea than that endorsed by the average postmodernist. (Relativists like Foucault

THE TRUTH A B OUT HI S TORY

225

with his ‘epistemes’, whether they realize it or not, are actually per­ petuating a structuralist semantics of history.) For the integrationist, Schopenhauer’s denial that history is a science, Marx’s belief in the class struggle, and Oakeshott’s thesis that the past is unknowable are all spin-offs from the Western surrogational semantics dominant in their day. Baudrillard’s contention that the Gulf War never happened reflects an equally dubious counter-surrogational (structuralist) semantics. None of these advances our understanding of how history is made, because they were from the start based on misconceptions about language and communication. Different philosophies of communication project different philosophies of history. Historians, qua communicators, cannot be history-makers without at the same time making and remaking the forms of commu­ nication in which the historiae are presented. The sooner that communicational role is acknowledged, and faced up to, the better. You cannot or should not - be a telehistorian without realizing that what you are doing is merging the presentation of the past with the presentation of popular entertainment. And that, automatically, affects the facets of history that can be presented. (Shakespeare already understood this four centuries ago.) The primary task of the integrational historian, unlike that of the surrogational or the structuralist historian, would be to identify and analyse the mechanisms of historification, in order to construct a defens­ ible general theory of the process itself. It can hardly be said that such a philosophy of history leaves no room for historical research. On the contrary, it opens up a programme that would keep researchers busy for years, if they were bold enough to embark on it. In the first place, we do not know, until the necessary research has been carried out, what the current state of historification is. That is to say, we do not know exactly what people believe about the past of their own society, or clan, or village, to say nothing of what they believe about others’. But we can reasonably expect to find that not everybody believes the same things. Then, for any given item (X) of historified belief (where X might be anything from ‘the battle of Hastings’ to ‘Custer’s last stand’ or ‘the first landing on the moon’), research is needed to show at least the following: (1) how long X has been accepted as ‘historical’, (2) by whom, (3) the current standing of X relative to other items in the individual’s or community’s store of such beliefs, (4) how the current standing of X is maintained and/or chal­ lenged, (5) the perceived relevance/irrelevance of X to current concerns, (6) the communicational formats through which X has been and is still perpetuated, and (7) the perceived credibility of these formats. I see no

226

THE L I N G U I S T I C S OF HI S TORY

dearth of team projects or PhD theses here. They might even tell us something worth knowing about how history is made. Let me conclude by restating briefly what I was trying to do in this and the preceding chapters. I have been trying to make sense of one of the great supercategories of modern civilization: history (or History, if you must). I conclude as follows. History is a body of beliefs, not a body of facts. (The latter view is one interpretation of the former.) History is made, not given. History is constantly changing because the making of it never stops. History is a changing body of beliefs about the past, by reference to which societies and individuals are able to construct an interpretation of the present and a view of their own future. The constitution of this body of beliefs is the product of a complex process I call ‘historification’, about which we know far less than we might like to suppose. But at least it seems clear that historification is very dependent on the modes of communication practised by a society at any given place or time, since these determine what kinds of account of the past are available. What people are prepared to believe about the past depends on how they think the available historical accounts relate to what actually happened. This involves taking a view not only of how any particular account was produced but of what it means or meant. And this, in turn, in the case of verbal accounts, involves taking a view of language and the limits within which, in general, it is possible for words to represent that which no longer exists. My contention is that a literate society will develop different forms of historification from any non-literate society. This is not only because writing enables detailed records to be kept from times past (and, in principle, kept indefinitely), but also because wherever writing becomes an essential cog in social organization and programmes of education it sponsors a reassessment of the role of language in human affairs. Different systems of writing throughout the world vary only within certain limits, and possible models of the relationship between speech and writing are restricted accordingly. Once language is conceived as having both a spoken form and a written form, the latter representing the former, models of this relationship tend to be carried over to interpreting the more comprehensive relationship between words and the world. In the West, where alphabetic writing triumphed, the result was an atomistic view of both speech and writing and the idealization of a one-to-one correlation between sounds and letters. The analogue to this at the verbal level was an atomistic view of words as semantic units and the search for one-to-one

THE TRUTH A B OUT HI S TORY

227

correlations between words and non-verbal items that could be construed as their meanings. This view of how language worked obviously favoured certain forms of historification (particularly those based on the preservation of written texts). It made literacy a professional requirement for historians and sponsored a definition of history that made history co-terminous with the survival of written records. (Everything else was prehistory.) It also, however, generated serious difficulties, since historical accounts could not claim exemption from the problems that were inherent in this view of language. Whereas a surrogational interpretation of words may seem plausible when dealing with correlates in praesentia, its plausibility diminishes when the alleged persons and events are invariably absent. What is required seems to be tantamount to an act of faith in the reliability of earlier testimony, which in turn may rely on still earlier testimony, and so on in an uncomfortable regress. The monitoring of this reliability required a class of experts, suitably trained in the appropriate forms of scholarship. Thus for centuries in the West the social institution of literacy and the pre-eminence of textual forms of historification dictated virtually the whole of the historian’s agenda. When eventually languages themselves came to be treated as historical objects and analysed accordingly, the historian’s problems only increased. For this undermined the plausibility of traditional semantics at a funda­ mental level. The supposed correlations between words and the world were shown to be either unstable or illusory: this left past accounts of the past in an even more parlous position than before. The way in which structuralism cut the Gordian knot of language made matters even worse. For now it was seen that words offered not even a rough guide to contemporary reality, much less a map to explore the past. The brash verbal confidence of the traditional historian suffered a blow from which it has yet to recover. The importance of non-verbal access to the past has augmented proportionately. A stage has now been reached at which any conflict between what the ancient texts say and what the latest dating and identification technologies reveal is almost certain to be resolved in favour of the latter. And if the technologies are challenged, the challenge is not likely to come from the historian, because the historian does not have the requisite scientific training. At the same time, the time-honoured role of the book as an instrument of historification is being challenged by the media. For the media historify as they go, with a minimum of research and a maximum of persuasion. The television report achieves the un­ rivalled power of authentication that comes from transforming the distant viewer into an on-the-spot eye-witness. Writing never managed that.

228

THE L I N G U I S T I C S OF HI S TORY

We are still in the transitional phase when our conceptions of history remain dominated by traditional learning and reliance on the written word. We have not yet taken on board - let alone seen how to deal with the problems posed by the new communication technologies. New criteria of historification were ushered in, whether we like it or not, by the advent of the televisual society, with its insatiable demand to participate as collective witness in the making of history. The traditional linguistics of history that sustained the historian for more than two millennia is past its sell-by date. It will need to be replaced by a more thorough and more comprehensive analysis of the mechanisms of modern communication. Our successors will need to understand just how their beliefs about the past-present continuum are being constructed and maintained. If writing-based forms of historification are to achieve any long-term credibility with the general public of the future, they must abandon a traditional philosophy of language that is no longer viable. They must be seen to be based on communicational assumptions that are coherent and defensible. This requires the historian to come clean about the processes of historification that underlie the account presented. And that applies just as much to new forms of historification as to old. Why is this important? Because otherwise we risk encouraging a kind of rejection of the past that might reasonably be called ‘anti-historical’. It is less a thesis than a state of mind. It is the state of mind reached when the communicational constraints on historification are sensed but not adequately analysed, leaving only scepticism about all forms of historical belief. Such a state of mind is one aspect of a loss of confidence in civilization. Like loss of faith in traditional religions, it provokes a fundamentalist backlash which only widens the gulf between believers and unbelievers. The current stand-off between pomophobes and their opponents is symptomatic of how far we have come in this direction. It is important also for a more fundamental reason. What people call ‘historical consciousness’ is not a matter of standing outside the temporal flow of events in order to comprehend it, but of engaging here and now in certain ongoing communicational processes that are part of the temporal flow itself. We do not have an option about this, even though we may try to convince ourselves otherwise. There is no superior God’s-eye view of either language or history. To try to understand what makes history history, and how it is made, is a step towards answering that most famous of all philosophical challenges: ‘Know thyself.’

References

Abbott, E. A. (1912), H ow to T ell the P a rts o f Speech , A n Introduction to English G ra m m a r , rev. ed., London: Seeley Service. Ankersmit, F. R. (1981), N a rra tiv e Logic. A Sem an tic A n a ly sis o f the H istorian s L anguage , Groningen: Krips ReproMeppel. Ankersmit, F. R. (1994), H istory a n d Tropology. The Rise an d F a ll o f M e ta p h o r, Berkeley: University of California Press. Appiah, K. A. (1997), ‘The arts of Africa’, N ew Y ork R eview o f Books XLIV: 7, 46-51. Aristotle, Com plete W orks. The R evised O xfo rd T ran slation , ed. J. Barnes, Princeton: Princeton University Press, 1984. Arnauld, A. and Lancelot, C. (1660), G ram m aire generale et raisonnee , Paris: Le Petit. Austin, J. L. (1962), H ow to do things with W ords , Oxford: Clarendon. Bailyn, B. (1994), On the Teaching a n d W ritin g o f H istory, Hanover: University Press of New England. Barthes, R. (1966), ‘Introduction a l’analyse structurale des recits’. Repr. in R. Barthes, U a v e n tu re semiologique, Paris: Seuil, 1985, pp. 167—206. Barthes, R. (1967), ‘Le discours de l’histoire’. Repr. in R. Barthes, Essais critiques I V . L e bruissement de la langue, Paris: Seuil, 1984, pp. 153-66. Becker, C. L. (1966), E verym an H is Own H isto ria n , Chicago: Quadrangle. Bede, H istoria Ecclesiastica G entis A nglorum , trans. J. E. King, Loeb Classical Library, London: Heineman, 1930. Bergmann, G. (1964), Logic an d R e a lity , Madison: University of Wisconsin Press. Berlin, I. (1954), H isto rica l In e v ita b ility , Oxford: Oxford University Press. Black, M. (1969), ‘Some troubles with Whorfianism’, in S. Hook (ed.), Language an d P h ilosop h y , New York: New York University Press, pp.30-5. Bloch, M. (1954), The H istorian s C ra ft, trans. P. Putnam, Manchester: Man­ chester University Press. Bopp, F. (1816), Uber das Conjugationssystem der S an sk ritsp ra ch e in Vergleichung w it jenem der griechischen, lateinischen , persischen und germanischen S p ra c h e , Frankfurt-am-Main. Repr. in R. Harris (ed.) (1999), vol. 1. 229

230 Bopp, F. (1833—52),

THE L I N G U I S T I C S OF HI S TORY Vergleichende G ram m a tik des S a n sk rit, Zend, Griechischen,

Berlin. Repr. in R. Harris (ed.) (1999), vols 10-11. Breal, M. J. A. (1897), Essai de sem antique , Paris: Hachette. Brilliant, R. (1984), V isu al N a rra tiv e s. S to ry te llin g in E truscan a n d R om an A r t , Ithaca: Cornell University Press. Briscoe, J. (1996), ‘Livy’, in S. Hornblower and A. Spawforth (eds), The O xfo rd C lassical D ic tio n a ry , 3rd ed., Oxford: Oxford University Press, pp.877-9. Burns, R. M. (2000), ‘On philosophizing about history’, in R. M. Burns and H. Rayment-Pickard (eds), Philosophies o f H isto ry , Oxford: Blackwell, pp.l— 29. Calame-Griaule, G. (1987), Ethnologie et langage: la p a ro le chez les Dogon , 2nd ed., Paris: Institut d’Ethnologie. Cameron, D. (1985), Feminism an d Linguistic T heory , London: Macmillan. Cannon, J. (ed.) (1980), The H istorian a t W o rk , London: Allen & Unwin. Carr, D. (1995), ‘Philosophy of history’, in R. Audi (ed.), The Cam bridge D ictio n ary o f P h ilosop h y , Cambridge: Cambridge University Press, pp.584— 6. Carr, E. H. (1987), W h a t is H istory?, 2nd ed., London: Penguin. Catford, J. C. (1965), A Linguistic Theory o f T ran slation , London: Oxford University Press. Clanchy, M. T. (1993), F rom M em ory to W ritten Record. E ngland 1 0 6 6 - 1 3 0 7 , 2nd ed., Oxford: Blackwell. Collingwood, R. G. (1938), The Prin ciples o f A r t , Oxford: Clarendon. Collingwood, R. G. (1944), A n A u to b iog ra p h y , Harmondsworth: Penguin. Collingwood, R. G. (1946), The Id ea o f H isto ry , Oxford: Clarendon. Cornell, T. J. (1996), ‘Annals, annalists’, in S. Hornblower and A. Spawforth (eds), The O x fo rd C lassical D ic tio n a ry , 3rd ed., Oxford: Oxford University Press, pp.98-9. Cornford, F. M. (1935), P la to 's Theory o f K now ledge , London: Routledge & Kegan Paul. Crystal, D. (1997), A D ictio n ary o f Linguistics an d Phonetics , 4th ed., Oxford: Blackwell. Daniel, G. (1962), The Idea o f P re h isto ry , London: Watts. Danto, A. C. (1985), N a rra tio n a n d K n ow ledge , New York: Columbia University Press. Darnton, R. (1984), The G re a t C a t M assacre and O ther Episodes in French C u ltu ra l H isto ry , New York: Basic. Dauzat, A. (1944), Les etapes de la langue fra n fa is e , Paris: Presses Universitaires de France. Davis, H. G. (2001), W ords. A n In teg ra tio n a l A p p ro a c h , London: Curzon. Davis, H. G. and Taylor, T. J. (eds) (2003), R ethinking Linguistics , London: RoutledgeCurzon. Davis, N. Z. (1975), S o ciety an d C u ltu re in E a rly M odern F ran ce, Stanford: Stanford University Press. Lateinischen, Litthauischen, Gothischen und D eutschen,

REFERENCES

Descartes, R. [1637] (1985),

231

trans. R. Stoothoff, in The trans. J. Cottingham, R. Stoothoff and D. Murdoch, vol. 1, Cambridge: Cambridge University Press. Donnellan, K. S. (1972), ‘Proper names and identifying descriptions’, in D. Davidson and G. Harman (eds), Sem antics o f N a tu ra l L anguage , Dordrecht: Reidel, pp.356-79. Downes, W. (1994), ‘History and linguistics’, in R. E. Asher (ed.), The Encyclo­ pedia o f Language and Linguistics , Oxford: Pergamon, vol. 3, pp. 1578-81. Edmonds, J. M. (1906), A n Introduction to C o m p arative P h ilology f o r C lassical S tu d en ts , Cambridge: Cambridge University Press. Elton, G. R. (1967), The P ra ctice o f H isto ry , Sydney: Sydney University Press. Evans, R. H. (2000), In Defence o f H istory, new ed., London: Granta. Evans, R. H. (2003), ‘Our job is to explain; it is for others to judge’, Times H igher Education S upp lem en t , 13 June, pp.20-1. Ewert, A. (1933), The French L anguage , London: Faber & Faber. Farrar, F. W. (1873), F am ilies o f Speech , 2nd ed., London: Longmans, Green. Fay, B. (1998), ‘The linguistic turn and beyond in contemporary theory of history’, in B. Fay, P. Pomper and R. T. Vann (eds), H istory an d Theory. C ontem porary R eadings , Oxford: Blackwell, pp. 1— 12. Finley, M. I. (1972), ‘Introduction’ to Thucydides , H istory o f the Peloponnesian W a r, trans. R. Warner, rev. ed., London: Penguin. Foster, B. O. (ed.) (1919), L iv y , H istory o f Rome, vol. 1, Loeb Classical Library, London: Heinemann. Frankel, C. (1957), ‘Explanation and interpretation in history’, Philosophy o f Science. Repr. in P. Gardiner (ed.), Theories o f H istory, Glencoe, 111.: Free Press, 1959, pp.408-27. Frye, N. (1957), A n a to m y o f C riticism , Princeton: Princeton University Press. Gadamer, H-G. (1985), T ruth an d M ethod, 2nd ed., New York: Crossroad. Gardiner, P. (1952), The N a tu re o f H istorical E xp lan ation , Oxford: Clarendon. Gernsheim, H. and Gernsheim, A. (1955), The H istory o f P h otog rap h y, Oxford: Oxford University Press. Gibbs-Smith, C. H. (1973), The B a y e u x T apestry, London: Phaidon. Gobineau, J-A. de (1853-5), Essai sur I'inegalite des races humaines, 4 vols, Paris. Goldstein, L. J. (1976), H isto ric a l K n ow in g, Austin: University of Texas Press. Goody, J. (1977), The Dom estication o f the S a v a g e M in d , Cambridge: Cambridge University Press. Goody, J. and Watt, I. P. (1963), ‘The consequences of literacy’, C om p arative Stu d ies in H istory and S o ciety 5: 304— 45. Gordon, W. T. (1997), M a rs h a ll M cL u h an . Escape into Understanding. A B io grap h y, New York: Basic. Grimm, J. (1819), Deutsche G ram m atik , Gottingen. This and subsequent edi­ tions are reprinted in R. Harris (ed.) (1999), vols. 3-9. Haas, W. (1962), ‘The theory of translation’, Philosophy 37: 208—28. Discours de la methode,

Ph ilosophical W ritin gs o f D escartes ,

232

THE L I N G U I S T I C S OF HI S TORY

Hacker, P. M. S. (1996), W ittgenstein s P lace in T w en tieth -C en tu ry A n a ly tic P h ilosop h y , Oxford: Blackwell. Hacker, P. M. S. (1997), ‘The rise of twentieth-century analytic philosophy’, in H. J. Glock (ed.), The Rise o f A n a ly tic P h ilosophy, Oxford: Blackwell, pp.51-76. Harris, R. (1980), The L an g u ag e-M a k ers , London: Duckworth. Harris, R. (1981), The Language M y th , London: Duckworth. Harris, R. (1986), ‘Language difficulties’, Times L ite ra ry Su p p lem en t , 28 No­ vember, p. 1338. Harris, R. (1990), ‘On “folk” and “scientific” linguistic beliefs’, in S. L. Tsohatzidis (ed.), M eanings and P rototypes. Stu d ies in Linguistic C a teg o riza ­ tion, London: Routledge, pp.449-64. Harris, R. (1998), Introduction to In te g ra tio n a l Linguistics, Oxford: Pergamon. Harris, R. (ed.) (1999), Foundations o f Indo-E uropean C o m p arative P h ilo lo gy , 1 8 0 0 - 1 8 5 0 , London: Routledge. Harris, R. (2001), Sau ssu re a n d his In terp reters, Edinburgh: Edinburgh Uni­ versity Press. Harris, R. (ed.) (2002), The Language M y th in W estern C u ltu re, London: Curzon. Harris, R. and Wolf, G. (eds) (1998), In teg ra tio n a l Linguistics: a F irs t R ead er, Oxford: Pergamon. Harris, W. V. (1989), A n cien t L ite ra c y , Cambridge, Mass.: Harvard University Press. Havelock, E. A. (1963), P reface to P la to , Cambridge, Mass.: Harvard University Press. Havelock, E. A. (1976), O rigins o f W estern L ite ra c y , Toronto: Ontario Institute for Studies in Education. Havelock, E. A. (1982), The L ite ra te R evolution in Greece a n d its C u ltu ra l Consequences, Princeton: Princeton University Press. Herodotus, The H istories, trans. A. de Selincourt, rev. ed., London: Penguin, 1972. Hexter, J. H. (1961), R eap p raisals in H isto ry, London: Longman. Hexter, J. H. (1967), ‘The rhetoric of history’, H istory an d Theory 6: 1- 14. Hill, C. (1997), In tellectu a l O rigins o f the English R evolution R evisited, Oxford: Clarendon. Hobbes, T. [1651] (1968), L evia th a n , ed. C. B. Macpherson, Harmondsworth: Penguin. Hobsbawm, E. (1983), ‘Introduction: inventing traditions’, in E. Hobsbawm and T. Ranger (eds), The In vention o f T radition , Cambridge: Cambridge Uni­ versity Press, Canto ed., pp. 1—14. Hobsbawm, E. (1997), On H istory, London: Weidenfeld & Nicolson. Hopper, P. (1990), ‘The emergence of the category “proper name” in discourse’, in H. G. Davis and T. J. Taylor (eds), R edefining Linguistics, London: Routledge, pp. 149-62. Hubbell, H. M. (1962), C icero , O ra to r, rev. ed., Loeb Classical Library, Cam­ bridge, Mass.: Harvard University Press.

REFERENCES

233

Humboldt, W. von [1821] (1967), ‘On the historian’s task’, H istory an d Theory 6: 57-71. Humboldt, W. von [1836] (1988), On Language , trans. P. Heath, Cambridge: Cambridge University Press. Hume, D. [1777] (1975), Enquiries concerning H um an U nderstanding an d con­ cerning the P rin ciples o f M o ra ls , ed. L. A. Selby-Bigge, 3rd ed., rev. P. H. Nidditch, Oxford: Clarendon. References are to numbered sections of the text printed in this edition. Iggers, G. G. (1997), H istoriograph y in the Twentieth C en tu ry. From S cien tific O bjectivity to the Postm odern C hallenge , Middletown, Conn.: Wesleyan Uni­ versity Press. Jakobson, R. (1960), ‘Closing statement: linguistics and poetics’, in T. A. Sebeok (ed.), S ty le in L anguage , Cambridge, Mass.: Massachusetts Institute of Technology Press/New York: Wiley. Jakobson, R. and Halle, M. (1956), Fundam entals o f L anguage , ’s-Gravenhage: Mouton. James, W. (1904), ‘Does “consciousness” exist?’, J o u r n a l o f Philosophy, P sychol­ ogy a n d S cien tific M ethods 1, no. 18. Repr. in Essays in R a d ic a l Em piricism , New York: Longmans, Green, 1912. Jenkins, K. (1991), R e-T h in k in g H isto ry , London: Routledge. Jones, W. [1786] (1807), Discourses d elivered a t the A sia tic k S o c ie ty , in The W orks o f S i r W illiam Jo n e s , ed. Lady A. Jones, London: Robinson, vol. 3. Jordanova, L. (2000), H isto ry in P ra c tic e , London: Arnold. Joseph, J. E., Love, N. and Taylor, T. J. (2001), L an d m arks in Linguistic Thought I I . The W estern T radition in the Twentieth C e n tu ry , London: Routledge. Kelley, D. R. (1998), Faces o f H istory. H istorical In q u iry fro m H erodotus to H erd er , New Haven: Yale University Press. Kelly, J. (1984), H istory an d T heory: The Essays o f J o a n K e lly , Chicago: University of Chicago Press. Kelly, L. (1979), The True In terp reter. A H istory o f T ranslation Theory and P ractice in the W est , Oxford: Blackwell. Kennedy, G. (1963), The A r t o f Persuasion in G reece , Princeton: Princeton University Press. Kirk, G. S., Raven, J. E. and Schofield, M. (1983), The Presocratic Philosophers , 2nd ed., Cambridge: Cambridge University Press. Kirk, R. (1986), T ranslation D eterm ined , Oxford: Clarendon. Kolakowski, L. (1972), P o sitiv ist Philosophy. From Hume to the V ienna C ircle , trans. N. Guterman, Harmondsworth: Penguin. Lakoff, R. (1975), Language and W om an s P la c e , New York: Harper. Lallot, J. (1989), L a gram m aire de D enys le T hrace , Paris: Centre National de la Recherche Scientifique. Lascelles, M. (1980), The S to ry -T e lle r R etrieves the P a s t , Oxford: Clarendon. Lehmann-Hartleben, K. (1926), D ie Traianssdule, ein romisches K u n stw erk zu Beginn der S p d ta n tik e , Berlin.

234

THE L I N G U I S T I C S OF HI S TORY

Le Quesne, A. L. (1982), C a rly le , Oxford: Oxford University Press. Le Roy Ladurie, E. (1979), L e C a r n a v a l de R om ans , Paris: Gallimard. Levine, J. M. (1999), The A u ton om y o f H istory. T ruth a n d M eth od fro m Erasm us to Gibbon, Chicago: University of Chicago Press. Levi-Strauss, C. (1958), A nthropologie stru ctu rale, Paris: Plon. Translations are from and page references to S tr u c tu r a l A n th rop o lo g y , trans. C. Jacobson and B. Grundfest Schoepf, Harmondsworth: Penguin, 1968. Levi-Strauss, C. (1962), L a pensee sauvage, Paris: Plon. Lienhardt, G. (1966), S o c ia l A n th ro p o lo g y , 2nd ed., London: Oxford University Press. Livy, A b Urbe C o n d ita , trans. B. O. Foster, Loeb Classical Library, London: Heinemann, 1919. Locke, J. [1706] (1959), A n E ssay Concerning H um an U nderstanding , 5th ed., ed. A. C. Fraser, 1894. Repr. New York, Dover, 1959. Logan, R. K. (1986), The A lp h a b et E ffect. The Im p act o f the Phonetic A lp h a b et on the D evelopm ent o f W estern C iv iliz a tio n , New York: Morrow. Lord, A. B. (1960), The S in g er o f Tales, Cambridge, Mass.: Harvard University Press. Lounsbury, F. G. (1969), ‘Language and culture’, in S. Hook (ed.), Language an d P h ilosophy, New York: New York University Press, pp.3-29. McLuhan, M. (1951), The M ech a n ica l B rid e : F o lk lore o f In d u s tria l M a n , New York: Vanguard. McLuhan, M. (1962), The Gutenberg G a la x y : The M ak in g o f T ypographic M a n , Toronto: University of Toronto Press. McLuhan, M. (1964), U nderstanding M e d ia : The Extensions o f M a n , New York: McGraw-Hill. McLuhan, M. (1967), The M edium is the M assag e: A n In ve n to ry o f E ffects, New York: Bantam. McWhorter, J. H. (2002), The P o w er o f B abel, London: Heinemann. Mandelbaum, M. (1938), The Problem o f H isto rica l Know ledge, New York: Liveright. Manto, S. H. (1997), M o ttle d D aw n . F ifty Sketches and S to rie s o f P a rtitio n , trans. K. Hasan, New Delhi: Penguin. Marcuse, H. (1964), O ne-D im ensional M a n . S tu d ies in the Ideology o f A d v an ced In d u stria l S o ciety, London: Routledge & Kegan Paul. Marett, R. R. (1929), ‘Anthropology’, E ncyclopaedia B rita n n ic a , 14th ed., vol. 2. pp.41-6. Marwick, A. (1970), The N a tu re o f H istory, London: Macmillan. Quotations are from and page references to the corrected 1976 reprint. Marwick, A. (2001), The N ew N a tu re o f H istory. K now ledge , Evidence , Language, Houndmills: Palgrave. Marwick, A. (2002), ‘Knowledge and language: history, the humanities, the sciences’, H istory 87: 33-17. Marx, K. (1963), Selected W ritin gs in Sociology an d S o c ia l Philosophy, eds. T. B. Bottomore and M. Rubel, Harmondsworth: Penguin.

REFERENCES

235

Mathiot, M. (ed.) (1979), E thnolinguistics: B oas, S a p ir a n d W h o r f R evisited , The Hague: Mouton. Meiland, J. W. (1965), Scepticism a n d H isto rica l K now ledge , New York: Random House. Meillet, A. (1903), Introduction a retu d e com p arative des langues indo- europeennes, Paris: Hachette. Meillet, A. (1925), L a methode com parative en linguistique historique, Oslo: Aschehoug. Michael, I. (1970), English G ra m m a tic a l Categories a n d the T radition to 1 8 0 0 , Cambridge: Cambridge University Press. Miles, R. (1988), The Women s H istory o f the W o rld , London: Michael Joseph. Repr. London: HarperCollins, 1993. Page references are to the reprint. Mill, J. S. (1884), A System o f Logic , 8th ed., London: Longmans, Green. Repr. London: Longman, 1970. Miller, J. (1971), M c L u h a n , London: Fontana/Collins. Mink, L. O. (1987), H isto rica l U nderstanding , Ithaca: Cornell University Press. Morris, M. (1998), ‘What problems? On learning to translate’, in R. Harris and G. Wolf (eds), In te g ra tio n a l Linguistics: A F irs t R ead er , Oxford: Pergamon, pp.313-23. Muller, F. M. [1851] (1995), ‘Comparative philology’, Edinburgh R eview 94. Repr. in Language a n d Linguistics , vol. 3 (The Wellesley Series), London: Routledge/Thoemmes. Page references are to the reprint. Muller, F. M. (1861), Lectures on the Science o f Language. Page references are to the 4th ed., London: Longman, Green, Longman, Roberts & Green, 1864. Muller, F. M. (1864), Lectures on the Science o f Language, Second Series. Page references are to the 4th ed., London: Longman, Green, Longman, Roberts & Green, 1864. Murdoch, I. (1977), The F ire an d the S u n . W h y P la to B anished the A rtis ts , Oxford: Oxford University Press. Murray, G. (1920), ‘Preface’ to A ris to tle on the A r t o f P o e try , trans. I. Bywater, Oxford: Clarendon. Nida, E. A. (1964), T ow ard a Science o f T ran slatin g , Leiden: Brill. Oakeshott, M. (1933), Experience an d its M odes, Cambridge: Cambridge Uni­ versity Press. Ong, W. J. (1982), O ra lity an d L ite ra c y . The Technologizing o f the W o rd , London: Methuen. Orwell, G. (1946), ‘Politics and the English language’. Originally published in H o rizo n , April, 1946. Repr. in The Collected Essays, Jo u rn a lism a n d L etters o f George O rw ell , eds. S. Orwell and I. Angus, vol. 4, Harmondsworth: Penguin, 1970, pp. 156-70. Orwell, G. (1949), N ineteen E ig h ty -F o u r , London: Seeker & Warburg. Repr. Harmondsworth: Penguin, 1954. Parry, M. (1971), The M ak in g o f H om eric Verse: The Collected P ap ers o f M ilm a n P a r r y , Oxford: Clarendon.

236

THE L I N G U I S T I C S OF HI S TORY

Pears, D. F. (ed.) (1956), The R evolution in P h ilosophy , London: Macmillan. Perks, R. and Thomson, A. (eds) (1998), The O ra l H isto ry R ead er , London: Routledge. Plato, Com plete W orks , ed. J. M. Cooper, Indianapolis: Hackett, 1977. Pope, M. K. (1934), From L a tin to M odern F ren ch , Manchester: Manchester University Press. Popper, K. R. (1962), The Open S o ciety an d its Enemies, 4th ed., 2 vols, London: Routledge & Kegan Paul. Portelli, A. (1998), ‘What makes oral history different’, in Perks and Thomson (1998), pp.63-74. Pott, A.F. (1856), D ie Ungleichheit menschlicher Rassen , Meyer: Lemgo & Detmold. Powicke, M. (1941), The R eform ation in E ngland , London: Oxford University Press. Price, R. (1990), A la b is W orld , Baltimore: Johns Hopkins University Press. Quine, W. V. O. (1960), W o rd a n d O bject, Cambridge, Mass.: Massachusetts Institute of Technology Press/New York: Wiley. Quine, W. V. O. (1970), ‘On the reasons for indeterminacy of translation’, J o u r n a l o f Philosophy 67: 178-83. Quintilian, In stitu tio O ra to ria , trans. H. E. Butler, Loeb Classical Library (4 vols), London: Heinemann, 1920-2. Rask, R. K. (1818), Undersogelse om det gam le nordiske eller islandske sprogs oprindelse, Copenhagen. Repr. in R. Harris (ed.) (1999), vol. 2. Rayment-Pickard, H. (2000), ‘Narrativism’, in R. M. Burns and H. RaymentPickard (eds), Philosophies o f H istory. From Enlightenm ent to P ostm od ern ity, Oxford: Blackwell, pp.274—84. Read, P. (1998), ‘Presenting voices in different media’, in Perks and Thomson (1998), pp.414—20. Robins, R. H. (1997), A S h o rt H istory o f Linguistics , 4th ed., London: Longman. Rorty, R. M. (ed.) (1967), The Linguistic Turn. Essays in Ph ilosophical M eth o d , Chicago: University of Chicago Press. Rossi, L. (1971), T ra ja n s Colum n a n d the D acian W a rs, London: Thames & Hudson. Rowse, A. L. (1946), The Use o f H istory, London: Hodder and Stoughton. Russell, B. (1918), The Philosophy o f L og ical Atom ism . Repr. in B. Russell, Logic an d Know ledge, ed. R. C. Marsh, London: Routledge, 1992. Page references are to the reprint. Ryle, G. (1966), P la to 's Progress, Cambridge: Cambridge University Press. Saussure, F. de (1922), Cours de linguistique generale, 2nd ed., Paris: Payot. Translations quoted are from R. Harris, F . de S au ssu re , Course in G en era l Linguistics, London: Duckworth, 1983. Schama, S. (2002a), A H istory o f B rita in , vol. 3. The F a te o f E m pire 1 7 7 6 - 2 0 0 0 , London: BBC Worldwide. Schama, S. (2002b), ‘Television and the trouble with history’, B B C H isto ry M a g a zin e 3: 7, 42-4, 3: 8, 40-3.

REFERENCES

237

Schmidt, J. (1872), D ie V erw andtschaftsverhaltnisse der indogermanischen S p r a chen , Weimar: Bohlau. Shotwell, J. T. (1910), ‘History’, Encyclopcedia B rita n n ic a , 11th ed., vol. 13, pp.527-33. Skinner, Q. (1981), M a c h ia v e lli, Oxford: Oxford University Press. Southgate, B. (2001), H isto ry : W h a t an d W h y?, London: Routledge. Spencer, T. J. B. (1963), Shakesp eare: The Roman P la y s, London: Longmans, Green. Spender, D. (1980), M a n M a d e Language, London: Routledge & Kegan Paul. Stalin, J. (1954), M a rx ism a n d Problem s o f Linguistics, Moscow: Foreign Lan­ guages Publishing House. Stone, L. (ed.) (1965), S o c ia l Change a n d R evolution in E n gland 1 5 4 0 - 1 6 4 0 , London: Longman. Strettell, A. [1851] (1993), ‘Comparative philology’, N orth B ritish R eview 16. Repr. in B ritish Linguistics in the Eighteenth C en tu ry, London: Routledge/ Thoemmes. Page references are to the reprint. Taylor, A. J. P. (1976), Essays in English H istory, Harmondworth: Penguin. Thapar, R. (2000), C u ltu ra l P asts. Essays in E a rly In d ian H istory, New Delhi: Oxford University Press. Thomas, R. (1989), O ra l T radition an d W ritten R ecord in C lassical A then s, Cambridge: Cambridge University Press. Thompson, D. (2003), ‘Many hands make a light work’, Times H igher Education Supplem ent, 10 January, p. 25. Thompson, P. (2000), The Voice o f the P ast. O ra l H istory, 3rd ed., Oxford: Oxford University Press. Thompson, W. (2000), W h a t H appened to H isto ry?, London: Pluto. Thomson, A. (1998), ‘Unreliable memories? The use and abuse of oral history’, in W. Lamont (ed.), H isto rica l C ontroversies a n d H istorians, London: Uni­ versity College London Press, pp.23—34. Thucydides, H istory o f the Peloponnesian W a r, trans. R. Warner, rev. ed., London: Penguin, 1972. Toews, J. E. (1987), ‘Intellectual history after the linguistic turn: the autonomy of meaning and the irreducibility of experience’, A m erican H isto rica l R eview 92: 879-907. Toolan, M. (1996), T o ta l Speech. A n In teg ra tio n al Linguistic A p p roach to Language, Durham: Duke University Press. Toolan, M. (2001), N a rra tiv e . A C ritic a l Linguistic Introduction, 2nd ed., London: Routledge. Tosh, J. (1984), The P u rsu it o f H istory. A im s , M ethods and N ew D irections in the S tu d y o f M od ern H istory, London: Longman. Tylor, E. B. (1871), P rim itiv e C u ltu re, London: Murray. Varro, D e L ingua L a tin a , trans. R. G. Kent, Loeb Classical Library, London: Heinemann, 1938. Vico, G. [1744] (1984), The N ew Science o f G iam b attista Vico, 3rd ed., trans. T. G. Bergin and M. H. Fish., rev. ed., Ithaca: Cornell University Press.

238

THE L I N G U I S T I C S OF HI S TORY

Voloshinov, V. N. (1929), M arx ism a n d the Philosophy o f L anguage, trans. L. Matejka and I. R. Titunik, Cambridge, Mass.: Harvard University Press, 1986. Walsh, W. H. (1967), A n Introduction to Philosophy o f H isto ry , 3rd ed., London: Hutchinson. Warnock, G. J. (1958), English P h ilosophy since 1 9 0 0 , London: Oxford University Press. Warren, J. (1999), H istory a n d the H istorian s , London: Hodder & Stoughton. Wartburg, W. von (1934), E volution et structure de la langue fr a n fa is e , Leipzig: Teubner. Wells, H. G. (1946), A S h o rt H istory o f the W o rld , rev. ed., Harmondsworth: Penguin. Werner, O. (1994), ‘Sapir-Whorf hypothesis’, in R. E. Asher (ed.), The E ncyclopedia o f Language an d Linguistics , Oxford: Pergamon, vol. 7, pp.3656-62. Wettstein, H. (1995), ‘Causal theory of proper names’, in R. Audi (ed.), The Cam bridge D ictio n ary o f P h ilosop h y , Cambridge: Cambridge University Press, pp. 109-10. White, H. (1973), M etah isto ry. The H isto rica l Im agination in N in eteen th -C en tu ry E urope , Baltimore: Johns Hopkins University Press. White, H. (1978), Tropics o f Discourse. E ssays in C u ltu ra l C riticism , Baltimore: Johns Hopkins University Press. White, H. (1987), The C ontent o f the Form . N a rra tiv e Discourse a n d H isto rica l R epresentation , Baltimore: Johns Hopkins University Press. Whitney, W. D. [1867] (1971), L anguage an d the S tu d y o f L anguage , New York: Scribner. Repr. in W h itney on Language. S elected W ritin gs o f W illiam D w ight W h itn ey, ed. M. Silverstein, Cambridge, Mass.: Massachusetts Institute of Technology Press. Page references are to the reprint. Whorf, B. L. [1939] (1956), ‘The relation of habitual thought and behavior to language’. Repr. in Language, Thought an d R e a lity . S elected W ritin gs o f B enjam in Lee W h orf, ed. J. B. Carroll, Cambridge, Mass.: Massachusetts Institute of Technology Press. Williams, B. (2002), T ruth a n d Truthfulness. A n E ssay in G enealogy, Princeton: Princeton University Press. Wolf, G. (1991), M ich el Bre'al. The Beginnings o f Sem an tics, London: Duckworth. Wolf, G. and Love, N. (eds) (1992), ‘Integrational linguistics: an introductory survey’, Proceedings o f the X V t h In te rn a tio n a l Congress o f Linguists, Quebec, Presses de l’Universite Laval, vol. 1, pp.313-20. Wordsworth, A. (1987), ‘Derrida and Foucault: writing the history of histori­ city’, in D. Attridge, G. Bennington and R. Young (eds), P ost-stru ctu ralism a n d the Q uestion o f H istory, Cambridge: Cambridge University Press, pp. 1lb25.

Index

Abbott, E. A., 56-7 Aeschylus, 43 alphabetic mind, 50-1, 55, 60-3, 656 ambiguity, 15-17, 23, 25, 29-30 Amyot, J., 175 Anderson, B. S., 157 Ankersmit, F. R., viii, 9, 12 annalists, 81 Appiah, K. A., 197 Aquinas, T., 84 Aristotle, viii, 8, 34, 36, 38—43, 4650, 59-60, 68-70, 72, 78-80, 86, 96, 124, 139, 145, 147-8, 174, 198-9, 211, 222 Arnauld, A., 56 Austin, J. L., 164, 179-80 Bacon, F., ix, 149 Bailyn, B., 16 Barthes, R., 121-5, 128, 132, 164 Baudelaire, C. P., 215 Baudrillard, J., 225 Bayeux tapestry, 216 Beaumont, F., 175 Becker, C. L., 174, 211 Bede, 82, 84 Bergmann, G., 11, 12, 17 Berlin, I., 26, 146^53 Bible, 57, 84, 93, 97, 121 Black, M., 194 Blanc, L., 122 Bloch, M., viii Bopp, F., 103-5, 109-11, 117, 134 Bossuet, J. B., 122

Breal, M., 110-11, 118-19 Brilliant, R., 213-15 Briscoe, J., 89 Buckle, H. T , 99 Burns, R. M., 12, 84 Bury, J. B., 24 Caesar, 72, 77, 122, 172-3, 175, 178-80, 198, 200-1 Calame-Griaule, G., 14 Cameron, D., 160-1 Cannon, J., ix, 16 Carjat, E., 215 Carlyle, T., 8, 22 Carr, D., 16 Carr, E. H , vii, 173, 177 Catford, J. C., 193 Chapman, G., 175 Christianity, 82—4, 112 Christine de Pisan, 157 Churchill, W. S., 154 Cicero, 77, 80-1 Clanchy, M. T., 83 Clio, 80 codes, 19-22, 48, 70, 98-9, 119, 130, 139, 153, 170, 186, 191-2, 194, 219 Collingwood, R. G., 30-1, 35, 49, 67, 91-2, 119, 132-3, 168, 184, 197, 222 communication, 1-2, 10, 13, 19-22, 32-3, 43, 47-8, 50-3, 58-9, 66, 778, 86, 94-9, 135, 139, 143-4, 153, 158, 162, 169-71, 174, 176-7, 1867, 193-7, 208-9, 211, 217-27; see also codes, telementation, writing 239

240

THE L I N G U I S T I C S OF HI S TORY

Comte, A., 98, 120 Condorcet, M. J. A. N. C. de, 98-9 contractualism, 3, 5-8, 14, 17, 19, 23, 25, 27, 191, 203 Cornell, T. J., 81 Cornford, F. M., 47-8 Croce, B., 175 Crystal, D., 64 Daniel, G., 183 Danto, A. C , 13-14, 27-8, 31 Darnton, R., 202 Darwin, C. R., 115, 183—4 Dauzat, A., 131 Davis, H. G., 170 Davis, N. Z., 204—5 definition, 47, 74, 139, 152, 167, 187, 205, 223 Democritus, 85 Derrida, J., 188, 210 Descartes, R., 76-7, 180 determinism, 146-7, 149-52 diachronic linguistics, 119-20, 126, 130, 132, 134, 181, 183 Dio Cassius, 213 Dionysius Thrax, 56 Donnellan, K. S., 96-8, 205 Downes, W., vii Edmonds, J. M., 116 Einstein, A., 188 Encyclopedic , 75, 84 Epicurus, 85 etymology, 18, 84, 97, 118-19, 181— 2, 203 Evans, R. H , 9, 165-6, 188-90, 194, 199, 208 Ewert, A., 130-1 Farrar, F. W., 117 Fay, B., 9, 13 feminists, 155-61 fiction, 27, 35, 45, 74-5, 89, 93, 124, 171, 175, 177, 204 Finley, M. I., 41 Fletcher, J., 175 Foster, B. O., 90 Foucault, M., 224

Frankel, C., 26 Frege, G., 11 Frye, N, 208 Fustel de Coulanges, N. D., 122 Gadamer, H.-G., 22-3, 71-2 Gardiner, P., 26, 28—31, 137-48, 150, 152 Gernsheim, A., 215 Gernsheim, H., 215 Gibbon, E., 15, 68, 84, 154 Gibbs-Smith, C. H., 216 Gobineau, J.-A. de, 112 Goldstein, L. J., 199-200 Goody, J., 51, 53, 63 Gordon, W. T., 53 grammar, 56-9, 64, 83, 103, 109-10, 113, 120, 122, 154, 159-60, 165, 167, 195 Grimm, J., 103, 107-9, 111-14, 116, 134 Gulya, J., 117 Gyarmathi, S., 117 Haas, W., 193-4 Hacker, P. M. S., 11 Halle, M., 209 Harris, R., x, 15, 20, 163, 170, 192 Harris, W. V., 51 Havelock, E. A., 50-2, 60, 63-4 Hegel, G. W. F., 92, 98, 121, 164 Heraclitus, 36 Herodotus, 19, 34, 40-3, 49, 55, 63, 65, 67-8, 79-81, 83, 91, 96, 122, 135, 154, 156, 183 Hesiod, 43—4, 80 Hexter, J. H., 9 Heywood, T., 175 Hill, C., 206-7 h is to ric 43, 80-1, 174— 7, 179-81, 183, 202-3, 207, 211, 216-17, 223, 225 historical apologies, 221 historical consciousness, 32, 34, 63, 66, 90, 205, 221, 228 historical explanation, 28, 40-2, 77, 111, 118, 182, 205 historical fiction, 180-1

I NDEX

historical hindsight, 43, 120 historical method, 24—5, 165-5, 210 historicism, 35-7, 49, 101-136, 163 historification, 42-3, 134—6, 172, 176-7, 181, 183, 201-3, 205, 21618, 224-8 historiography, 24, 83 history-making, 34, 42, 65, 169, 171— 3, 175-7, 196, 201, 204, 217-19, 225-6, 228 history of history, 25 history of ideas, 36 Hobbes, T , 69, 132 Hobsbawm, E., 172, 199, 205 Homer, 35, 43-4, 80, 211 Hopper, P., 205 Horne Tooke, J., 118 Hubbell, H. M., 80 Humboldt, W. von, 13, 106, 110-11, 131, 164, 185-6, 215 Hume, D., 73-6, 98-9 Iggers, G. G., 10 integration(ism), vii-viii, 32, 42, 49, 168-73, 175-7, 179-80, 182-7, 191-2, 194-7, 201, 203-5, 207-8, 212-13, 216-25 Jakobson, R. O., 122, 125, 209 James, W., 85 Jenkins, K., 24, 29 Jespersen, O., 122 Johnson, S., 175, 192 Joinville, J. de, 122 Jones, W., 102-6, 110, 117 Jonson, B., 175 Jordanova, L., 9, 11, 190 Joseph, J. E., 186 Kant, I., 98 Kelley, D. R., 74, 82-3 Kelly, J , 157 Kelly, L., 193 Kennedy, G., 79 Kirk, G. S., 44 Kirk, R , 192 Kolakowski, L., 170

241

Lakoff, R., 158-60 Lallot, J., 56 Lancelot, C., 56 language myth, 20-2, 25, 28, 48, 523, 57, 70, 86, 119, 129, 152-3, 158, 185, 194, 203 Lascelles, M., 180 Lehmann-Hartleben, K., 214 Le Quesne, A. L., 8 Le Roy Ladurie, E., 212 Levine, J. M., 168, 189 Levi-Strauss, C., 121-2, 125-8, 132 Lienhardt, G., 184 linguistic discrimination, 158-9 linguistic limits, 13-15 linguistic turn, 9-12, 17, 22, 27, 153, 207-8 literacy, 45, 50-5, 60-8, 79, 90, 154, 169, 195, 211, 213, 2 2 4 - 1 ; see also writing Livy, 82, 84, 88-96, 135, 154, 184, 215 Locke, J., 8, 21, 23, 86-8, 94, 129, 132, 139, 165 Logan, R. K., 51-2 Lord, A. B., 211 Lounsbury, F. G., 194 Love, N. L., 170, 186 Lucian, 74—5 Luther, M., 157 Lyotard, J.-F., 188 Macaulay, C. S., 157 Macaulay, T. B., 24, 68, 154 Machiavelli, N., 122, 191 McLuhan, M., 52-3 McWhorter, J. H., 131 Mandelbaum, M., 199 Manto, S. H., 204-5, 207 Marcuse, H., 163-5 Marett, R. R., 185 Marwick, A., 9-10, 90, 135, 165-6, 190 Marx, K., 92, 98, 162-3, 225 Mathiot, M., 194 Matthew Paris, 83 meaning, 2-8, 10, 14—18, 23-32, 48, 57, 74, 78, 83, 85-8, 94, 118-20,

242

THE L I N G U I S T I C S OF HI S TOR Y

125, 129, 131-2, 144-5, 148-9, 152-4, 158, 163, 166, 169-71, 182, 187-90, 193-4, 196, 200-1, 203, 205, 219-22, 224, 227; see also translation Meiland, J. W., 167 Meillet, A., 110-11, 113, 119 memory, 37—8, 40-1, 45-6, 66, 83, 97, 171, 187 messages, 19-22, 129, 192 metaphor, 144-5, 148-50, 152, 154, 166, 189, 223 Michael, I., 58-9 Michelet, J., 24, 122, 219 Miles, R., 155-6 Mill, J. S., 86, 99, 120 Miller, J., 53, 98 Mink, L. O., 161-2, 208 Morris, M., 193 Muller, F. M., 104-8, 111-12 Murdoch, I., 35 Murray, G., 39 myth(ology), 39, 43-6, 57, 74, 93, 123, 125, 147, 152, 176-7, 203, 214 narrative, 13, 26, 39, 42, 88, 98-9, 122-4, 154—5, 161-2, 177, 180, 204, 207-8, 210-14, 216-17, 223 natural history, 75 Needham, J., 52 Neogrammarians, 101, 108, 118, 120, 134 Newton, I., 64 Nida, E. A., 193 Niebuhr, B. G., 90 Niepce, N., 215 Nietzsche, F., 27 nomenclaturism, 4, 57, 128, 143 North, T., 175 Oakeshott, M., 173, 225 Ong, W. J , 51 opinio, 174-6, 183, 202, 204, 211, 213, 215, 219 oral history, 174, 218-20 ordinary language, 26-7, 93, 134, 137-67, 170, 205, 223

Orwell, G., 74, 78-9, 81, 90-1, 1645, 218 Parry, M., 211 Pears, D. F., 12 performative history, 216 Perks, R., 219 philology, 9, 103-19 Plato, 8, 22, 35-10, 42-9, 53-6, 589, 61-2, 64-7, 69-72, 82, 104, 118, 133, 196 Pliny, 156 Plutarch, 156, 175 poetry, 34—5, 38-9, 41-2, 45-9, 55, 63, 78, 80, 84, 98, 124, 180, 196, 209 Polybius, 90, 154 Pope, M. K., 130 Popper, K. R., 35-6, 38, 177 Portelli, A., 219 postmodernism, 10, 18, 22, 25-6, 79, 134-6, 168, 176-7, 188, 210, 224, 228 Pott, A. F., 112 Powicke, M., 19, 205-6 prehistory, 183— 4, 227 presumptions of factuality, 177-9, 205, 207, 217 presumptions of non-factuality, 178, 203-1 Price, R., 211 print historians, 216-18 Priscian, 59 probability, 75, 99 proper names, 38, 47, 58, 69-71, 76, 92-9, 162, 181, 186, 205-7 psychocentrism, 3-8, 14, 17, 19, 23, 30, 84-8, 91, 94, 128-9, 158-9, 162-3, 165, 173, 191, 220 Quine, W. V. O., 192 Quintilian, 60 Raleigh, W., 175 Ranke, L. von, 75, 105, 113, 135 Rask, R. K., 103, 108-9, 111-12, 117, 134 Raven, J. E., 44

N D EX

Rayment-Pickard, H., 121, 124—5, 155, 208 Read, P., 219 reocentrism, 3-8, 14, 17, 19, 26, 29, 57, 68-79, 82, 8^8, 91-2, 94, 98, 124, 128-9, 133, 155-6, 175-6, 180, 188, 191, 194, 199, 201, 203, 205, 208, 221-2 res gestae , 80, 172, 174— 6, 179-80, 202, 211

rhetoric, 9, 12, 56, 79-80, 83, 124, 154, 166, 188, 198, 209, 211-12 Rivarol, A. de, 166 Robins, R. H., 56, 64 Rorty, R. M., 10 Rossi, L., 212 Rowse, A. L., 24 Russell, B. A. W., 93-7, 162, 205-7 Ryle, G., 36, 55

243

Strabo, 156 Strettell, A., 105-6, 111-12 structuralism, 11, 84, 101, 120-34, 137, 155-7, 161-3, 165, 168-70, 185, 187, 189, 192, 194, 225, 227 substantialism, 91—2, 119, 132, 189 supercategories, viii-ix, 226 surrogation(al)ism, 4—6, 60-2, 65-6, 69-70, 72-3, 77, 79, 83, 86, 88, 90-2, 94, 96, 100-1, 124, 128-9, 132-3, 138, 143, 149, 157-8, 162— 3, 165-6, 169-70, 177, 187, 220, 225, 227 synchronic linguistics, 119, 126-7, 130, 181, 183

Tacitus, 122, 215 Taylor, A. J. P., 170, 216 Taylor, T. J., 170, 186 telehistory, 216—18, 225, 227 Saussure, F. de, 11-12, 21, 53, 101— telementation, 20-2, 53, 59, 86, 94, 2, 109, 113, 119-21, 123-33, 135, 98, 158, 174, 194 teleology, 146, 149 159, 163, 167, 181, 188, 190, 194 Thapar, R., 32 Schama, S., 202, 216—18 Schleicher, A., 115 Thierry, A., 122 Schmidt, J., 115 Thiers, L. A., 122 Schofield, ML, 44 Thomas, R., 81 Schopenhauer, A., 225 Thompson, D., 216 science, 13, 21, 24, 26, 51-2, 67, 75, Thompson, P., 218-19 Thompson, W., 178 86, 90, 101, 104-9, 114, 118, 120, 125-6, 136-44, 147-8, 150-1, 162, Thomson, A., 219-20 165-8, 181, 227 Thucydides, 36, 41-3, 55, 65, 68, 79, semantics see meaning; also 81, 83, 86-8, 91, 122 contractualism, reocentrism, Toews, J. E., 9-10 psychocentrism, structuralism, Toolan, M., 170, 212-13 surrogation(al)ism Tosh, J., 16 Shakespeare, W., 175, 178-80, 225 Toynbee, A. J., 23, 92, 135, 146 Shotwell, J. T., 15, 24 Trajan’s column, 212-14 Skinner, Q., 191 translation, 13, 142, 176, 189-97, Socrates, 35, 43-7, 54—6, 58, 61-2, 200, 203, 205 69-71 Trubetzkoy, N. S., 125 Southgate, B., 11, 157, 168, 190, truth, 7, 13, 20, 24, 35, 43-5, 49-50, 210 56, 59, 65, 67-72, 74-7, 79, 82-4, Spencer, T. J. B., 175 88, 90, 92, 94, 97-100, 122, 152, Spender, D., 160 155-7, 159, 161, 164, 170-1, 176Stalin, J., 163-4 81, 183, 188, 196, 198-228 statistics, 99, 120, 140, 165 Tylor, E. B., 185 Stone, L., 166

244

THE L I N G U I S T I C S OF HI S TORY

Usher, J., 84 Varro, 84-5 Vico, G., 153 Voloshinov, V. N., 134 Voltaire, F. M. A. de, 75 Walsh, W. H., 15-16, 26 Warnock, G. J., 12 Warren, J., 163 Wartburg, W. von, 131 Watt, I. P., 51 Wells, H. G., 63, 66, 79, 171 Werner, O., 194 Wettstein, H., 96 White, H., 9, 12, 26, 124, 153^1, 208-9 Whitney, W. D., 107, 181

Whorf, B. L., 27, 150, 160, 163, 165, 167, 194 Wilde, O, 1, 169 Williams, B., x Wittgenstein, L., 11, 13, 128, 164 Wolf, S. G., 170 women’s language, 158-60, 164 Woolf, V , 156 Wordsworth, A., 210 writing, 33, 50—5, 60-9, 80, 84, 90, 96, 109, 161, 166, 170, 173^, 183, 193, 205, 211, 213, 216-19, 226-8 Xenophanes, 44 Xenophon, 122 Zinsser, J. P., 157