|
Discourse LIKE has received much attention in the pragmatic literature (e.g., Schourup 1983; Andersen 1997 et seq.), but it has never been investigated from a variationist perspective. Consequently, this dissertation presents an accountable analysis of LIKE is a large corpus of contemporary English. The hypothesis developed in this work is that LIKE is not random, but interacts with syntactic structure in regular and predictable ways. To address this issue, the variable context is circumscribed according to structural criteria and the analyses are embedded within current Minimalist Theory (e.g., Chomsky 1995 et seq.). Over 20,000 structurally defined contexts are examined, comprising data from 97 speakers between the ages of 10 and 87. This method reveals that LIKE is (1) highly constrained by the syntax and (2) occurs in specific positions among speakers of all ages. Indeed, examination of language-internal constraints reveals that the community shares a single variable grammar for LIKE (Poplack & Tagliamonte 2001). This feature is shown to have developed gradually and systematically, arriving at its current state through regular processes of language change. Using the grammaticalization models proposed by Traugott (1997 [1995]) and Brinton (forthcoming), it is argued that after initially developing as a discourse marker, where it occurs clause-initially and links sequences of dialogue (Fraser 1988, 1990), LIKE then begins to enter syntactic structure, spreading to one maximal projection at a time.
|
E-mail this message to a friend
