Language and History

Integrationist Perspectives

Nigel Love

• Publisher: Routledge
• Year: 2006
• ISBN:
  9781134370191
• Edition: 1
• Topic: History
• Subtopic: World History
• Index: History

When linguistics was first established as an academic discipline in the nineteenth century, it was envisaged as an essentially historical study. Languages were to be treated as historical objects, evolving through gradual but constant processes of change over long periods of time. In recent years, however, there has been much discussion by historians of a 'linguistic turn' in their own discipline, and, in linguistics, integrationist theory has mounted a radical challenge to the traditional notion of 'languages' as possible objects of inquiry. Language and History develops the integrationist critique of orthodox linguistics, while at the same time extending its implications to the field of history. By doing so, it throws light on what is now recognized by many historians to be a 'crisis' in their own discipline. Underlying the post-modernist scepticism about traditional forms of historiography, the integrationist approach reveals a more deep-seated problem concerning the interface between philosophy of history and philosophy of language. With chapters from a range of leading international contributors, Language and History represents a significant contribution to the developing work of the integrationists.

1 Language, history and Language and History

Nigel Love

Defining a topic for discussion by conjoining two large, vague nouns with ‘and’ is not calculated to elicit narrowly focused debate. On the contrary, it is a procedure that apparently invites contributors to say anything that in some way relates one designated thing or phenomenon to the other; and when the things or phenomena are in themselves complex and multifaceted, and the connections and relationships between them numerous and varied, the likeliest outcome is no more than an enhanced and perhaps rueful appreciation of how many disparate issues can be made to fall under the heading in question. This short-coming of the ‘X and Y’ formula may appear especially salient in the case of ‘language and history’. To take only the most obvious point, it seems to embrace questions pertaining to the language of history and questions pertaining to the history of language. Aren’t those two (at least two) different subjects? Is there anything to be gained by bringing them together in a single volume? Does this collection go beyond merely bringing them together and perform the altogether more useful and significant operation of yoking them together?

That is for its readers to say. At any rate, without claiming that everything in this book is directed towards sustaining the thesis that the two kinds of linguo-historical question are related, an attempt can be made to spell out that thesis and to draw attention to those aspects of the following chapters that support or illustrate it.

A commonsensical view of the relation between language and history might run as follows. In the primary sense of the word, history is what happened in the past. In a secondary sense, sometimes covered by the word ‘historiography’, history is what historians produce: in essence, collections of statements, factual or interpretative, as to (some of ) what happened. Language serves the historian, as it serves any other species of statement-maker, by providing the linguistic equipment for making statements. That equipment consists of words whose meanings convey the substance of the statements the historian wishes to make. A language is a device for encoding systematically recoverable semantic content. What could be simpler than that? A famous epistemology – Popper’s – has been founded on treating language and languages as thus unproblematically capable of enshrining ‘objective knowledge’ (see e.g. his 1972 book of that title).

That matters are not quite so simple may be initially insinuated with the observation that languages themselves (allegedly) have histories; and this, if true, must be expected to complicate their role as repositories of stable meanings. But once again common sense offers a bluff response. Languages do indeed change over time, and the reader of non-contemporary texts must be alert to the interpretative difficulties this may cause. But can we not rely on linguistic historians (philologists, historical linguists) to reconstruct the history of languages and thereby alleviate such difficulties? The temporal instability of languages may in certain ways be problematic, but the problems can in principle be overcome.

Christopher Hutton’s chapter in this volume raises by implication the question of how far common sense of this order is the product of a linguistic culture whose writing happens to involve the deployment of characters that have long since come to be treated as visually arbitrary. This is not so with Chinese, for instance. Chinese characters offer a ‘fundamental semiotic challenge’, being read ‘in complex ways as having both a pictorial and a phonetic element’. That is to say, the meaning and history of the word taken to be represented by a character is bound up with questions about what the character depicts or portrays and how that portrayal has evolved. The whole enterprise of reconstructing linguistic history is thus from the outset liable to be conceived in ways that differ radically from those taken for granted in other cultures, with a range of consequences not just for how in detail ancient texts are interpreted but for what it is to interpret them.

Even if we confine ourselves to the world of Western languages and linguistics, one would have more confidence in the deliverances of common sense on these questions if linguistic historiography were theoretically better founded than it is. The purpose of that enterprise is to investigate and describe the history of languages, i.e., how they change through time. But the fact is that there is no agreement among historical linguists as to either (a) what a language is, or (b) what it is that changes when a linguistic change occurs or (c) what kind of event, if any, a linguistic change might be. This is an unpromising state of affairs.

It is a linguists’ platitude that what counts as a language in ordinary parlance is a matter of indefinitely variable sociocultural definition; that no linguistic (i.e. internal structural) justification can be found for distinguishing Swedish and Danish as separate languages if Mandarin and Cantonese are to be acknowledged as versions of the same language; that what happened to Dutch when taken to South Africa was on the face of it no more cataclysmic than what happened to English when the Normans invaded, and yet South African Dutch became Afrikaans, while English marched on as English. Historical linguists, for their part, have in general never been particularly concerned to impose some special definition of their own, uniquely required for their historians’ purposes,¹ or to displace ordinary parlance by defining the term ‘a language’ so as to give unequivocal answers to questions about the boundary between one language and another (whether in time, in space, in terms of social stratification . . .), or to specify how exactly the concept of a dialect relates to the concept of a language, or to worry over whether the sense of ‘a language’ in which Spanish, say, is a form of Latin, should or should not take priority over the sense in which Spanish and Latin are different languages. Such issues have traditionally either not been thought to matter or else have been set aside sine die, or at any rate for a rainy one.

In his chapter in this collection Roger Lass considers, from the perspective of a historical linguist working (albeit critically) within the orthodox scholarly tradition that starts with nineteenth-century comparative philology, some of the consequences for linguistic historiography of supposing that a language name like ‘English’ uncontroversially identifies a domain susceptible to historical treatment in terms of a linear narrative. English is especially problematic in this respect because of the relative wealth of documentary evidence available from different times and places, and because of the depth and intensity of study to which surviving texts have been subjected. A paradox emerges: the more that is known of a language’s history the less reason there is to suppose that it has one. Published histories of English are replete with statements to the effect that A ‘becomes’ B in cases where there is no reason to suppose, even granting the standard assumption that in certain circumstances the formula A > B has a coherent interpretation,² that A and B stand in any identifiable relationship of temporal succession. Forms attested in eighth-century Northumbrian are treated as directly ancestral to forms attested in tenth-century West Saxon. It is as if items from one period in Italian were to be linked by the A > B formula to items from a later period in Portuguese. The reason that would be unacceptable is that Italian and Portuguese are well known to be different ‘languages’, whereas from a stand-point that starts by identifying ‘English’ as the object whose history is to be mapped, Northumbrian and West Saxon are not. They are merely ‘dialects’. Thus the historiographical enterprise is distorted from the outset by the requirement that a folk language name correspond to a real-world entity whose history can be likened, in Lass’s trope, to a tube or cylinder whereby ‘the historiographical act consists of pouring in forms at one end . . . and looking to see what occurs at the various graduation marks (the points at which the surviving texts allow us to intersect the “stream” of history), and then finally seeing what comes out of the bottom’.

Lass further observes that ‘historians tend to adopt a manner of speaking in which the collection of materials whose ancestry we are reconstructing, no matter how complex and variable, always at some remove appears to go back to a single ancestral object whose “story” we then find ourselves telling’. We know quite well from reflection on our own first-order experience of language that it is hard to find ‘single objects’ at any level of generalising description. Lass takes the example of ‘the’ Modern English vowel ‘short a’, as in bat, cab, sap, etc., which, looking at the English-speaking world as a whole, corresponds at the very least to all of [a], [ ᵄ], [a], [ᵋ], [e]. But as we go back in time such variation dwindles away: ultimately all of these will be held to have a single, unitary origin in the ‘comparativist fiction of the “dialect-free protolanguage” ’. There is no such phonetic object as ‘short a’ (nor for that matter any unitary phonological analysis in terms of which ‘short a’ would emerge as a single pan-English phoneme), but the very act of setting it up for the modern language establishes it as the reflex of some definite ancestral entity. Thus does a language’s history emerge as an artefact of the particular synchronic analysis taken as a starting point.

Why are these conceptual problems that swirl round the concept of ‘a language’ largely ignored? Partly because they are intractable, and partly because the central proposition on which Western historical linguistics is founded is not that it is possible to identify languages across transformations in time but linguistic substance constitutive of linguistic units, i.e., the component microparts of languages. The issue of which language any units thus historically located count as belonging to can indeed be left to politics, or societal Selbstverstehung, or whatever; the historical linguist is in business on the footing that, given any two états de langue, however crudely identified, it is possible to say whether they do or do not contain a core of material, at the meaningful level of articulation, which counts as ‘the same’ in that one can plausibly postulate continuous direct or indirect historical transmission of linguistic units combining a form with a meaning. Non-meaningful sameness, i.e., the subject matter of phonetic or phonological history, has in practice tended to get the lion’s share of attention in historical linguistics, but in fact it is logically subordinate to meaningful sameness in that one has no basis for historically relating two ‘sounds’ unless there is some justification for relating meaningful units in which those ‘sounds’ appear. So what makes English Indo-European is hundreds of etymologist’s facts such as that the word thatch contains the same root as the Latin verb tego, that feather is ultimately the same, or contains some of the same linguistic substance, as Welsh adar ‘birds’, that cow and the stem of bovine are ultimately cognate, and so on and so forth. What makes English Germanic is hundreds of facts such as that although English father, German Vater, Latin pater are all ‘the same word’, as are English foot, German Fuss, Latin pes, the English forms stand closer to the German than they do to the Latin, in that there are forms, attested or reconstructible, taken to be ancestral to both the English and the German that are less historically remote than forms ancestral to the English, the German and the Latin. And what makes Modern English and Old English the same language historically speaking (even if they are different languages according to most other kinds of speaking) is that many Old English forms are themselves identifiable as the recent forebears of Modern English forms.

But where do such ‘facts’ come from? Feather and adar, for instance, have neither a single phonetic segment in common (granted a non-rhotic English speaker) nor anything but the loosest semantic connection. Crucially required, it might seem, is some way of perceiving the ‘sameness’ allegedly lurking behind obvious – in many cases, gross – discrepancies. But this is the wrong way to look at it. It is not a matter of first noticing a (non-existent) sameness and then working back to ancestral forms whose divergent development in different branches of Indo-European has caused the differences. It is a matter of demonstrating that, provided it is not on semantic grounds wholly implausible to associate feather with adar, a story can be told whereby given, among many other things, a certain reconstructed Indo-European root or form, certain established sound changes distinguishing Germanic from Celtic . . ., continuous transmission through many generations might be expected to yield feather or something like it as the contemporary end point of one line of development and adar or something like it as the contemporary end point of the other. The alleged ‘sameness’ is an artefact of the notion of ‘continuous transmission’. What is claimed to have been continuously transmitted? The only possible answer is: a metalinguistic reification located in the mind of the historical linguist.

In trying to rationalise this onto-epistemological curiosity, two kinds of explanation suggest themselves. In ‘History and comparative philology’ Roy Harris discusses issues in the historiography of historical linguistics itself. He suggests that the main achievement of early scholars in the tradition of comparative philology, and their main claim to have founded a science of language, was that, going beyond merely observing many formal-semantic correspondences across Indo-European languages, they devised ways of systematising those observations and reducing them to simple correlational formulae. But this immediately invites the question what the correlations mean and why they should hold. ‘Interlinguistic correspondence formulae (of the kind typified by Grimm’s Law) are no more than that: i.e., formulae. In themselves they contain no historical information and capture no “facts” other than the observations and assumptions (whether accurate or inaccurate) that went into their formulation.’ So an interpretation is required. The preferred interpretation, much influenced by a contemporary revolution in biological science, was to suppose they represented the different development, along separate evolutionary lines, of linguistic entities that could be traced back to a common ancestor. Thus did comparative philology find itself committed to belief in the survival across millennia of sounds and forms which although in many cases transformed beyond recognition nonetheless mystically retained their identity. In the biological case items that did indeed retain their identity across such time spans were eventually identified and called genes. In the linguistic case, needless to say, nothing analogous to genes has ever turned up.

Another kind of explanation would point out that the ‘sameness’ of linguistic forms across long time spans is no more than a logical extension of a commonplace feature of our understanding of how languages work ‘synchronically’. If you and shortly afterwards I each produce an utterance that might be written down as ‘feather’, then irrespective of any phonetic differences between them we have both said ‘the same thing’. That in itself, properly understood, may be uncontentious enough. What is highly contentious is that ‘saying the same thing’ in such a case is to be interpreted as instantiating an abstract invariant, viz. the ‘thing’ (in this case the word feather) that has been said twice.³ Once one makes that reifying move there is nowhere to stop. For if the abstraction in question remains the same across minor phonetic differences (as between your pronunciation and mine) and tiny time gaps (as between your utterance and mine), at what specifiable degree of enlargement do the differences and the gaps begin to matter, and why? A viable concept of the limits of ‘a language’ would help here. But it is not clear that there is any to be had.

What, then, is a linguistic change, granted a historical linguistics crucially founded on identifying linguistic continuities through time? Is there a theoretical perspective in which it makes sense to conceive of linguistic objects as changing and yet mysteriously staying the same?

Let us approach an answer by sketching the kind of account a historical linguist might give of a small lexical change in English that seems to be taking place right now, if indeed it has not already been accomplished. English contains a word font, which belongs primarily to the technical vocabulary of printers, and which according to the second edition (1989) of the Oxford English Dictionary⁴ refers to ‘a complete set or assortment of type of a particular face and size’. It has a number of homonyms, and a variant fount.

As against that, the word-processing program with which I am now writing offers me (i) a choice of ‘fonts’, and (ii) a choice of ‘font sizes’, each choice independent of the other. This suggests that, at any rate for the designers of Microsoft Word, the word font does not mean what the OED says it means, in at least two respects. For the fonts in question are not ‘assortments of type’, nor are they of any particular size.

What does the linguist make of these discrepancies? He is likely to detect a paradigmatic case of semantic change, recent or ‘in pr...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contributors
• Preface
• 1 Language, history and Language and History
• 2 The end of linear narrative?: Reflections on the historiography of English
• 3 History and Comparative Philology
• 4 Word-stories: Etymology as history
• 5 Language: object or event?: The integration of language and life
• 6 Indeterminacy of meaning and semantic change
• 7 On the cusp: Antoine Meillet as a sociologist of language
• 8 ‘The grammatical being called a nation’: History and the construction of political and linguistic nationalism
• 9 How to make history with words
• 10 Talking about what happened
• 11 Part of the meaning/history of euro: Integrational corpus linguistics
• 12 Language and prehistory
• 13 Bridges to history: Biomechanical constraints in language
• References