Different views of language
Goethe once wrote:
Everyone thinks because he can talk, that he can therefore talk about language.
(Ein jeder, weil er spricht, glaubt auch ĂŒber die Sprache sprechen zu können.)
and it is certainly true that most people hold decided opinions about language in general and about their native language in particular. This is understandable enough, in so far as we have all learnt to speak our own native language fluently, and this alone has provided us with words and expressions to talk about language, such as speak, word, language, accent, put it another way, ambiguous. Furthermore, most of us have learnt to read and write, linguistic skills which carry with them terminology like letter, spell, prefix, sentence. Many of us even learnt some form of traditional grammar at school and got quite used to looking at language and taking it apart, using a variety of technical labels, e.g. verb, clause, infinitive, parse, and we may have learnt to apply these labels to foreign languages as well as our own. So altogether we apparently have some grounds for regarding ourselves as entitled to talk about language.
However, about the same time that we learnt to talk, most of us also learnt to walk and have since become competent walkers. We have also acquired a terminology for talking about walking; we distinguish walking, strolling and marching, for instance, and we talk about someoneâs gait, and whether he walked fast, straight, nervously, etc. But how many of us can give a concise, step-by-step account of what the action of walking involves, of what bones, muscles, ligaments, etc. are involved and what they do, of how human decisions and energy are transformed into physical movement? In our defence, we would say that we feel this is the task not of the man-in-the-street but of the professional physiologist, anatomist, physicist or whatever the relevant speciality might be. So we are naturally led to ask why there should not be a specialist in language â in linguistics, in fact â to give us a technical account of his field and explain many aspects that the layman is incapable of explaining.
Language has, of course, already been studied for many centuries from a number of points of view, in particular by philosophers, by traditional grammarians and by literary critics. Ancient philosophers (the Stoics, the Alexandrians, Plato, Aristotle) all discussed the nature and origin of language. Although philosophy has at times taken language for granted, the twentieth century has seen philosophers concern themselves fundamentally with such problems as those of meaning, reference and truth (cf. the work of Russell, Ayer), and linguistic philosophy has used language as a key for clearing up misconceptions about theories of knowledge, existence, good and evil, and so on (cf. the work of Wittgenstein, Ryle and Austin).
Traditional grammar grew out of work by ancient writers on philosophy and language but more particularly out of works devoted to the study of Greek (e.g. Dionysius Thrax) and Latin (e.g. Varro, Priscian). As these became dead languages, so Latin and Greek grammars were able to become codified systems and gain extra respect because of the learned status of the works written in those languages. Thus traditional grammar was prescriptive, laying down rules for the âcorrectâ use of the language (see below pp. 53â4); grammar had become part of the social etiquette, first of the learned world, and then later, when it was applied to âvernacularâ languages like English, French and Russian, of the polite world.
The literary critic has always had to contend with language. Since the writer has language as his medium of expression, his work must be judged partly on the basis of his use of language. The ancient art of rhetoric similarly depended on an analysis of language texts and has provided some of the notions and techniques of the literary critic, e.g. metaphor, paradox.
In more recent times other specialists have devoted special attention to language. Psychologists, social anthropologists and sociologists, speech pathologists and computer scientists, to pick out just a few, have all studied linguistic problems associated with their own fields. But each specialist has been concerned with the particular aspects of language that touch on his or her own studies; the psychologist seeing language in part as a manifestation of mental activity or behaviour, the speech pathologist being interested in the normal process of language acquisition and retention and how various abnormal patterns deviate from this, and so on.
So most studies of language outside linguistics, whether ancient or modern, have a particular axe to grind and therefore, quite naturally, slant their account of language in a particular way. It is left to the (general) linguist to study language in a neutral unslanted way: to study language for its own sake. He is interested in its inherent nature, rather than in its importance for something else; he has no ulterior motive. While a philosopher may see language as an imperfect and misleading code for expressing logical relations, a psychologist may see language as a key to the understanding of the mind, or a literary critic may see language as a a medium for literature, the linguist just wants to know what language itself is like and how it works. It is, moreover, vital to have a neutral, unslanted account of language, because, although each outside specialist may see very deeply into his own problems, he will lack an overview and as a consequence may overlook many important points and issues. There is of course nothing to prevent a specialist taking the linguistâs more general account and adapting it to his own particular needs.
Different linguistic theories
So, accepting the need for a purely linguistic account of language, where precisely do we find it? The point of asking this question is that, although some writers would have it otherwise, there is no single, generally accepted body of linguistic theory, but rather a range of competing schools. Now it is the aim of this book to bring together the common aims, principles and methods of these various theories and to show that despite their differences there are more things they agree about than disagree about. It will nevertheless be necessary to begin with a very brief indication of the identity of these schools and of the principal points of disagreement between them. In this account reference will be made not only to schools as such but to a number of influential individual linguists who have had independent views of their own.
The most well-known, and probably the most influential, school during the 1960s and 1970s has been that of transformational-generative grammar. This theory was originally propounded by N. Chomsky in his Syntactic Structures (1957) and subsequently modified in Aspects of the Theory of Syntax (1965). The twin keynotes of Chomskyâs approach have been the insistence on âgenerationâ, i.e. explicit specification of sentences and their structures through rewrite rules (see chapter 4), and the use of âtransformationsâ, rules for relating sentences with different structures or for relating the âsurface structureâ and âdeep structureâ of a single sentence (see also chapter 8). Transformational-generative grammar can, however, scarcely be regarded as a single school any more: a division of views opened up in the later 1960s between those, like J. D. McCawley, J. R. Ross, C. J. Fillmore and many others, who believe all syntax should be semantically based, thus merging semantics with âdeep syntaxâ, and those, like Chomsky himself and R. S. Jackendoff, who believe that a grammar should have independent âdeep syntaxâ and semantic components, and that in the semantic interpretation of a sentence both âdeep structureâ and âsurface structureâ should play a part (see discussion in chapter 8).
The most direct influence on Chomsky in his work was Z. S. Harris, his teacher, with whom he developed the notion of transformation. Harris, one of the most original and systematic thinkers in linguistics, propounded a kind of transformation that partially agreed with Chomskyâs earlier view, and one that we shall find fruitful; Harris also always stressed explicitness in a grammar.
We may trace Harrisâs and Chomskyâs work back to the twin father-figures of American linguistics, L. Bloomfield and E. Sapir. They each produced an important book called Language in 1933 and 1921 respectively (although Bloomfieldâs is a revision of an earlier book). The two books illustrate the contrasting qualities of the two men: while Bloomfieldâs is attentive to language detail, careful to scrutinize any theoretical proposal he makes and sceptical about the utility of studying meaning, Sapir is more imaginative, more intuitive and more enterprising. While Chomskyâs work can be seen as owing something to both, Harris is more directly a Bloomfieldian.
In fact the 1930s, 1940s and early 1950s were undoubtedly the era of Bloomfieldianism in the United States and the main alternative linguistic approaches available in North America today are provided either by individual post-Bloomfieldians like C. F. Hockett and the more transformationally inclined W. L. Chafe or by two schools arising out of Bloomfieldianism, tagmemics and stratificational grammar.
K. L. Pike first formulated the notion of a âtagmemeâ, a minimal grammatical pattern, in terms of which all grammatical structures have to be described. The idea was further developed by R. E. Longacre, who identified the âfour fundamental insights of tagmemicsâ as the functional value of tagmemes (as subjects, objects, etc.), their grouping into sequences (syntagmemes), their occurrence at different âlevelsâ (word, phrase, etc.) and the possibility for embedding and related phenomena (âlevel skipsâ, âlayeringâ and âloopbacksâ; see chapter 9).
S. M. Lamb emphasizes levels of a different kind in his âstrataâ, which are levels of abstraction or realization as well as of size unit, ranging from more abstract semantic and lexical units through morphemic units to phonemic units. The different units are linked through different kinds of realization, but also through rules of grouping called âtacticsâ, which capture the structurings at grammatical and other levels.
So far we have spoken only of American linguistics, and it is to some extent true that American and European linguistics followed different lines in the pre-Chomsky era.
One unifying factor was their common heritage from Ferdinand de Saussure, the father of modern linguistics (if anyone was), whose planned Cours de linguistique gĂ©nĂ©rale was realized after his death by his pupils, in 1915. De Saussure was the first to distinguish clearly synchronic studies of a language â those that consider the state of a language at a particular point in time â from diachronic studies â which have the history of language change as their focus of attention. His other, perhaps more important, insights concerned the nature of the âlanguageâ that we study: that we should concentrate on the abstract linguistic system (la langue) rather than the actual speech (la parole) and the essence of a language is not in its external aspects â phonetic expression or semantic reference â but in its internal system.
De Saussureâs most faithful followers form the Geneva school (C. Bally, A. SĂ©chehaye, H. Frei, R. Godel), but the most logical development of his views is to be seen in the Copenhagen school, and particularly in the work of L. Hjelmslev. Linguistics becomes for Hjelmslev an autonomous discipline and is therefore given the new name of âglossematicsâ, which describes language as an abstract system defined by its own internal relations. The theories of the Soviet linguist, S. K. Shaumyan, sometimes termed âappli-cational grammarâ, may too be regarded as embodying de Saussureâs formal principles, but Shaumyanâs grammar also claims to be generative in Chomskyâs sense.
Equally close to de Saussure was the work of the pre-war Prague school. Inspired by the expatriate Russians, N. S. Trubetzkoy and R. Jakobson, but with important contributions from Czechoslovak linguists (V. Mathesius, B. Trnka, J. Vachek), the Prague school linguists made striking progress in the field of phonological theory. Since 1945 a new generation of Prague linguists (F. Danes, J. Firbas) have made notable contributions to aspects of grammatical theory concerned with relations between sentences in a text (or âdiscourseâ) (see chapter 12).
In the United Kingdom it was J. R. Firth of the University of London who set the tone. Very few Europeans shared the extreme scepticism or pessimism about semantics felt by Bloomfield and his pupils in America, and Firth was even positive about meaning. He believed in studying language in the context of situation, and that meaning could be discerned at different linguistic levels. Firth also differentiated himself from Bloomfield in not giving undue weight to âchainâ relations â or âbracketingâ (see chapter 6) â compared with âchoiceâ relations-or âlabellingâ (see chapter 7). Firthâs most original pupil has been M. A. K. Halliday, whose work has been described as neo-Firthian but is mostly known now as âsystemic grammarâ. Halliday further emphasizes the plane of âchoiceâ relations, viewing language in general, and grammar in particular, as a whole system of choices or options with complex relations between them; he has also made a special study of textual relations along the same lines at the Prague school linguists.
To conclude our ultra-brief survey of approaches to linguistics we must mention two earlier European individual linguists, O. Jespersen and L. TesniĂšre. Jespersen, who was active throughout the first half of this century, besides his earlier phonetic work and his lengthy and erudite Modern English Grammar, wrote in a stimulating and insightful way in his general works, The Philosophy of Grammar (1924) and Analytic Syntax (1969). His notions of ârankâ in the sense of a scale of modifiers, of ânexusâ and the many transformational relations he expose...