Beyond the individual in the evolution of language
Hawkey, David J. C.
This thesis concerns the evolution of language. A proliferation of theoretical models have been presented in recent years purporting to offer evolutionary accounts for various aspects of modern languages. These models rely heavily on abstract mechanistic models of the production and reception of language by modern humans, drawn from various approaches in linguistics which aim at such models. A very basic and ubiquitous assumption is that expressions have meaning in virtue of being associated with internal representations, and that therefore the evolution of language can be modelled on the basis of individuals trying to produce external manifestations of these internal “meanings”. I examine the role of this assumption in language evolution theorising, and review evidence from neuroscience and first language acquisition relevant to the validity of this assumption. The chaotic nature of the relationship between “meaning” and the brain undermines the supposition that the evolution of language was driven by spontaneous association between internal structures and external forms. I then turn to the philosophical basis of language evolution theorising, adopting a Wittgensteinian perspective on the cognitive interpretation of linguistic theories. I argue that the theoretical apparatus of such approaches is embedded in language games whose complicated rules relate to linguistic behaviour (and idealisations of that behaviour) but not to neural organisation. The reinterpretation of such descriptions of language as descriptions of the internal structures of language users is rejected as a grammatical confusion: if the rules for constructing linguistic theory descriptions do not mention neural structures, then theoretical descriptions of the linguistic abilities of an individual say nothing non-trivial about their internal brain structure. I do not deny that it would, in principle, be possible to reduce linguistic theories (reinterpreted as mechanistic descriptions) to neural structures, but claim that this possibility is guaranteed only by leaving the practice of re-describing physical brain descriptions entirely unconstrained. Thus the idea that we can reasonably infer the behaviour of humans and prehumans in more primitive communicative environments by manipulation of the models of linguistic theories is unfounded: we have no idea how such a manipulation would translate into statements about neural organisation, and so no idea how plausible such statements about earlier neural organisation (and the resultant behaviours) are. As such, cognitive interpretations of linguistic theories provide no better ground for statements about behaviour during earlier stages in the evolution of language than guessing. Rejecting internal-mechanism based accounts as unfounded leaves the evolution of language unexplained. In the latter parts of this thesis, I offer a more neutral approach which is sensitive to the limited possibilities available for making predictions about human (and pre-human) behaviour at earlier stages in the evolution of language. Rather than focusing on the individual and imputed internal language machinery, the account considers the communicative affordances available to individuals. The shifts in what individuals can learn to do in interaction with others, that result in turn from the learning of interactive practices by others, form the basis of this account. General trends in the development of communicative affordances are used to account for generalisations over attested semantic change, and to suggest how certain aspects of modern language use developed without simply assuming that it is “natural” for humans to (spontaneously) behave in these ways. The model is used in an account of the evolution and common structure of colour terms across different languages.