A corpus-based analysis of extended multiple themes in PresE

Abstract

This corpus-based study reformulates Halliday's (1994: 55) notion of Multiple Theme, i.e., textual and/or interpersonal items occurring before a simple Topical Theme (or clause initial transitivity/mood element) (e.g., Well, but then, Ann, surely, wouldn't the best idea be to join the group?) (cf. Berry 1982, 1995; Lautamatti 1978; Young 1980; Vasconcellos 1992). Firstly, the label Extended Multiple Theme is here proposed as a cover-term for Topical Themes co-occurring with pre-topical and/or post-topical textual and/or interpersonal elements. And secondly, Extended Multiple Themes are suggested to: (i) allow for recursiveness within the three functional slots; (ii) tend to abide by Dik's (1989: 342) Principle of Centripetal Organisation; and (iii) substantiate the layering hypothesis posited for example in Dik 's Functional Grammar or in Role and Reference Grammar (cf. Hengeveld 1989; Van Valin Jr. 1993). These claims were deduced from the application of three multivariate statistical tests, namely, the Logistic Regression Technique, the Fisher's Exact Test, and the x2 Test, to the tokens of Extended Multiple Themes found in real Present-day English texts, that is to say, in the Lancaster Spoken English Corpus.