[185]
§1 Definitions. Traditional divisions
Static linguistics, or the description of a linguistic state, may be termed grammar in that very precise sense, by no means uncommon, found in expressions like ‘the grammar of chess’, ‘the grammar of the stock market’, etc. These are all cases of complex systems, involving coexisting values.
Grammar studies the language as a system of means of expression. ‘Grammatical’ implies ‘synchronic’ and ‘meaningful’. Since no linguistic system can straddle several eras simultaneously, as far as we are concerned there is no such thing as ‘historical grammar’. What is called ‘historical grammar’ is in reality simply diachronic linguistics.
Our definition is not in accord with the narrower one usually given. Grammar is conventionally restricted to morphology and syntax combined; and lexicology or the science of words is excluded.
But in the first place, do these traditional divisions correspond to linguistic reality? Are they in harmony with the principles we have already indicated?
Morphology deals with the various classes of words (verbs, nouns, adjectives, pronouns, etc.) and with the different forms of flexion (conjugations, declensions). To distinguish this from syntax, it is claimed that syntax examines the functions associated with linguistic units, while morphology is merely concerned with their forms. Thus morphology simply tells us that the genitive of the Greek word phúlax (‘guardian’) is phúlakos. Syntax informs us about the use of these two forms.
[186]
But this distinction is illusory. The series of forms of the noun phúlax do not constitute a flexional paradigm unless one compares the functions these different forms have. Equally, their functions have nothing to do with morphology unless each answers to a particular phonetic signal. A declension is neither just a list of forms nor just a series of logical abstractions, but a combination of both (cf. p. [144]). Forms and functions are interdependent. It is difficult if not impossible to separate them. Linguistically, morphology has no real, independent object of study: it cannot constitute a discipline distinct from syntax.
On the other hand, is it reasonable to exclude lexicology from grammar? At first sight, words as listed in the dictionary do not seem to lend themselves to grammatical analysis. For grammar is usually limited to studying relations between units. But it soon becomes apparent that many of these relations may be expressed by words just as well as by grammatical devices. In Latin, for example, fīō (‘I become’) and faciō (‘I make’) contrast in the same way as the passive form dīcor (‘I am said’) with its active dı¯cō (‘I say’). In Russian the distinction between the perfective and imperfective is expressed grammatically in the case of ‘to ask’ (sprosít’ vs. sprášivat’), but lexicologically in the case of ‘to say’ (skazát’ vs. govorít’). Prepositions are usually treated under grammar; nonetheless the prepositional phrase en considération de (‘in consideration of’) is essentially lexicological, since it contains the word considération (‘consideration’) used in its proper sense. If we compare the Greek peíthō vs. peíthomai with the French je persuade vs. j’obéis, we see that the same contrast (‘I persuade’ vs. ‘I obey’) is expressed grammatically in the first case and lexicologically in the second. Many relations expressed in certain languages by cases or prepositions are expressed in other languages by compounds, which are closer to words proper, e.g. French royaume des cieux (‘kingdom of heaven’) vs. German Himmelreich (heaven-kingdom’), or by derivational forms, e.g. French moulin à vent (‘windmill’) vs. Polish wiatr-ak, or by single words, e.g. French bois de chauffage (‘firewood’, lit. ‘wood for heating’) vs. Russian drová, or French bois de construction (‘wood for building’) vs. Russian lês. In the same language, we often find single words and compound expressions as alternatives, e.g. French considérer (‘consider’) and prendre en considération (‘take into consideration’), se venger de (‘revenge oneself on’) and tirer vengeance de (‘wreak vengeance upon’).
[187]
Thus from a functional point of view, lexicological and syntactic devices overlap. Moreover, any word which is not a single, unanalysable unit is essentially no different from a phrase, syntactically speaking. For the arrangement of smaller constituent units obeys the same basic principles as govern the formation of groups of words.
To summarise, the traditional divisions of grammar may have some practical utility, but they do not correspond to natural distinctions and are not unified by any logical principle. Grammar needs a different basis, and a better one.
The interpenetration of morphology, syntax and lexicology is explained by the fact that all synchronic features are ultimately of the same kind. No boundary between them can be laid down in advance. Only the distinction earlier drawn between syntagmatic relations and associative relations suggests a classification which is indispensable, and which fulfils the requirements for any grammatical systematisation.
Everything in a given linguistic state should be explicable by reference to a theory of syntagmas and a theory of associations. Certain parts of traditional grammar immediately appear to fall without difficulty under one or the other. Flexion is clearly a typical case of association between forms in speakers’ minds. Syntax, on the other hand, as generally understood nowadays – the theory of word sequences – falls under syntagmatics, since word sequences invariably involve at least two units spatially ordered in some way. All syntagmatic facts are not to be classed as syntax, but all syntactic facts belong to syntagmatics.
[188]
Any point of grammar will illustrate how important it is to examine each question from both points of view. The concept of a word raises two separate problems, depending on whether one considers the word associatively or syntagmatically. The French adjective grand (‘big’) takes two forms in the syntagma, one before a consonant (grã garsõ ‘big boy’) and another before a vowel (grãt ãfã ‘big child’). It also takes two forms associatively, a masculine grã (written grand) and a feminine grãd (written grande).
The aim should be to assign every fact to its proper domain, either syntagmatic or associative, and thus organize the whole subject-matter of grammar on its two natural axes. Only such a systematisation will show to what extent the traditional framework of synchronic linguistics needs to be altered. This is a task which cannot of course be undertaken here. We are simply concerned to lay down principles of a very general kind.