Language in the world: properties and powers
Linguistic autonomy
Interaction between languages and their speakers
However, all these lines of thinking share a common recognition: the interaction between language as a system and the human beings who make use of it is of much greater significance than the internal properties of languages themselves.
When writers provide illustrations of a phonological sequence or a syntactic construction that is prohibited in any given language, as is frequently done in linguistics textbooks, they simultaneously demonstrate that speakers can, for particular purposes, disregard or overcome these constraints, as well as stray across linguistic boundaries. The constraint may still be operative within that language, but the speaker can work out, for example, (despite not being a speaker of Twi) how to pronounce /mba/, can create the noun 'friendment' as an alternative to 'friendship', can, as we saw, generate '*got I have money n't no' (Finch 1998: 3) (if only to exemplify the importance of word order in English) or '?they sowed their isn't they reaped their same' (for poetic effect, as produced by e e cummings), and so on.
Speaker agency is particularly evident in the area of 'corpus semantics' (Stubbs 2001a), where many of the recurrent patterns are not fully determined, as with frozen metaphors and fixed idioms, but are better conceptualized as formulaic frames, amenable to varying degrees of creativity by speakers. These findings serve as a reminder that, although 'collocational facts are linguistic' (ibid,: 120), and are indicative of the constraints of the language system, meanings are social, the affordances of agency.
From another angle, in accounting for the origins of linguistic variation, Nettle gives considerable weight to the well-documented sociolinguistic phenomenon of speakers' desire to demonstrate their identification with certain groups in preference to others (Le Page and Tabouret-Keller 1985): 'The social-marking theory ... predicts that linguistic boundaries will form around the core networks of cooperation and exchange in which people are involved in their daily lives' (Nettle 1999: 59).
Recognition of the social character of language is also found in a recent collection (Knight et al 2000a) which presents various approaches to research linking the emergence of language to evolutionary development. The editors retain a modified concept of language as internal to the human individual (I-language) but see a diminished role for its innate inheritance, maintaining that: '[i]t is through others' performance - in other words, through language as embodied in social life - that speakers internalize (and, in turn, con tribute to) the language in which they are immersed' (2002b: 11). Language is essential not only for cognition and for the communication of its products, but also, crucially, for the kinds of social contracts which have no material referent. Social facts (Durkheim), institutional facts (Searle), intersubjective knowledge (Popper), are all, as we have repeatedly said, defining features of a social world which is not reducible to individuals. Knight et al.'s interpretation of the evolutionary evidence suggests that 'linguistic communication emerges and varies as an expression of distinctively human coalitionary strategies' (ibid.: 12). Similar lines of argument are to be found at the ontogenetic level in the emergentist literature (see for example Snow 1999).
