Grammaticalization changes the meaning of the item affected to a grammatical meaning or function. The question of what it is, if anything, that characterizes grammatical meaning and distinguishes it from lexical meaning has vexed generations of linguists. Are there inherently lexical notions distinct from grammatical notions? We can anticipate here that no satisfactory positive answer is possible, for a very basic reason: Highly grammaticalized forms have primarily a structural function rather than a meaning. The endpoint of the grammaticalization of a definite article is a noun-phrase marker. The endpoint of the grammaticalization of an abstract preposition – viz. Engl. to – may be an infinitive marker. In Neo-Melanesian (an English-based pidgin), the preverbal subject pronoun he and the postverbal object pronoun him have both agglutinated to the verb and now function as a verb marker and as a transitivity marker, resp. Such markers only facilitate the processing of linguistic messages by articulating their structure and increasing their redudancy. Thus, the most categorical answer to the question of what the typical grammatical meaning is, would be: it is zero. The primary function of grammar is not to convey certain meanings – this is the task of the lexicon – but to facilitate cognition and communication by structuring it, thus blowing it up with redundancy.

If one wants to subsume the structural function of such grammatical formatives under an extended concept of meaning, one calls it ‘metalinguistic meaning’. To recall earlier examples of this: The metalinguistic meaning of a gender feature of a noun is ‘this noun belongs to subcategory X of nouns’; and by this feature it may be referred to in endophora. The metalinguistic meaning of Engl. that as a universal subordinator is ‘what follows has the status of a noun phrase and may thus be referential’.

There are, however, stages of desemanticization between the meaning of a lexical source and its reduction to such a metalinguistic meaning and finally to zero. Even if there is no common denominator on which the meanings of grammatical formatives may be put, maybe there is at least a closed set of grammatical meanings that may be transported in the grammars of languages. A possible approach to this question is the compilation of a lexicon of such words which provide the sources of documented grammaticalization paths. There are two such dictionaries: Lessau 1994 and Heine & Kuteva 2002. Meritorious as these works are, they are not meant to offer a systematic theory of grammatical meaning. However, one could go about it by inducing a system of grammatical meanings from the complete set of the entries of these dictionaries.

It may be mentioned in passing that some have postulated a set of universal semantic constants. The ‘semantic primes’ (Wierzbicka 1996) are such a set. Some of these are commonly represented by grammatical items in languages, including be, this, not, if and maybe some more. The question is whether these notions are necessarily “grammaticalized”, i.e. are the significata of grammatical items in all languages. The answer is a clear ‘no’. Even assuming that such semantic primes are universal, i.e. part of the semantics of all language systems, none of them must be the meaning of a grammatical item. Even the operator not may be a verb. In English, this is only a possibility (); in Finnish, there is no other negator.

Research into functional and cognitive linguistics of the last half century has achieved important advances towards a theory of grammatical meaning by comparing grammatical categories and constructions across languages. The underlying idea is the following: Human language is an activity that combines cognition and communication. Certain aspects of these areas are central to linguistic activity, so it is worthwhile for language users to take them into account in forming grammmatical structure. This induces in postulating a set of functional domains of language.

Reference

Wierzbicka, Anna 1996, Semantics. Primes and universals. Oxford & New York: Oxford University Press.