Functions of Language - Volume 3, Issue 1, 1996

The category VP has since the late 1950s attracted a considerable amount of interest in both formally and functionally oriented linguistics, though in the latter to a lesser extent. There is a substantial amount of disagreement amongst linguists as to the status of this category, and as to its nature and extent amongst those who adopt it. Unfortunately, however, the literature contains few if any convincing arguments either for or against the category, most linguists apparently being content to accept, reject or ignore it as their intuitions or theoretical percepts guide them, and on the basis of quite flimsy evidence. In this paper I attempt to present some evidence in support of the category. It is impossible, of course, to do this in a theory-neutral way; the challenge in developing arguments is to deploy theoretical constructs as little as possible.

Can is traditionally seen as having three general meanings, ability, permission, and possibility (Leech 1987). However, as noted by Coates (1983), in certain cases the function of can is somewhat similar to that of will, e.g. offering to help by saying / can/will help, if you want y asking for help by saying Can/will you pass the salt please?, etc. Using a database of 119 naturally-occurring tokens of can followed by a main verb with future time reference (Trudeau 1984), the present paper posits 17 distinct functions of can, of which 7 show a semantic affinity with will. Those functions having a semantic affinity with will generally have a number of distinctive characteristics: they are associated with possibility rather than ability or permission, they are characterized by interactive or syntactic rather than semantic properties, and analysis of early 19th century texts shows that they have only developed in the language relatively recently. This suggests that, to a large extent, the possibility function of can represents a recent encroachment into the functional territory of future tense marking.

Evidentiality is traditionally discussed in terms of information source and its relation to the degree of commitment a speaker has with respect to a proposition. Furthermore, it is a notion that is usually thought to apply principally to assertions. A consideration of the full range of uses of the Wanka Quechua direct evidential shows that while these concepts are generally sufficient for the description of prototypical usage, understanding other uses requires explanation along different lines. Certain non-prototypical uses exemplify a type of subjectification, a process of semantic change that has been observed cross-linguistically (cf. Traugott 1989, Langacker 1990). The use of the direct evidential as a grammaticalized marker of content (WH) questions is particularly anomalous from a traditional perspective. However, it can be shown to be a well-motivated extension from other uses within the semantic network.

Systemic-functional linguistics (SFL) is used as a framework within which a child's 'cognitive' development can be seen in linguistic terms as the building of a meaning potential which gains realisation in texts. Data are then presented from a diary study of one child's speech between the ages of two and a half and five years, focussing on the child's use of 'mental' and 'verbal' clauses (such as I think or she said) in order to reveal the child's understandings about information exchange, which constitutes the basis of learning. The naturalistic data display various developments in the child's construals of semiotic exchange, including exploration of 'false' information and the status of perceptual evidence. A general pattern emerges whereby the child moves out from representing and exploring the 'I-you-now' of the ongoing interaction, to a later construal of the world beyond this 'deictic centre', suggesting an intersubjective rather than an 'egocentric' starting point to development.