Patterns of cohesion in spoken text

Editors’ introduction

The article by Susan & Geoff Thompson, like several others in this volume, relates to the issue of text chunks and linkages, but in their case the text is a corpus of television advertisements where viewers (and analysts) have to deal with information at various levels simultaneously: written text, voice-over, moving images. In the case of television advertisements, there is great coherence between these three.

However, the strength of the paper lies essentially in the combination of two modes of analysis, of “soundings” and “wordings” (we must look elsewhere for an analysis of “imagings”), an analysis which finds that the two modes operate not only in complement with each other but also with points of contrast.

The chief problem they address is that of cohesion: how do chunks relate to each other? What makes the advert a single audiovisual unit, a text, as opposed to a set of elements? To answer this, they draw upon the notions of repetition and replacement (Winter 1974, 1979). In this view, repetition operates within a text “frame” with “slots”, typically a fairly narrow phrasal context, and serves to signal continuity of aboutness. Darnton’s contribution in this volume presents a fairly similar view, while Scott’s takes repetition at a text level. Replacement puts new words in familiar slots. Conjunction, on the other hand, deals with the signalling of connectedness and can be achieved basically by claiming a Matching relation or a Logical sequence relation. A further pattern enabling linkage is Theme choice, which serves to glue the text together by claiming a relatedness of topic focus. That is, by preserving consistency in what is chosen to start each clause, the television advertiser manages to maintain a continuity of the relevant aspect of “what we’re on about”: namely promoting batteries or deodorants.

Note that parallelism of frames and slots, and conjunction and Theme in this view operate at a fairly local, narrow level in connecting two or more consecutive chunks. That is, they are likely on average to be connecting points of detail rather than major changes of gist. As the Thompsons point out, however, there are also larger-scale operators, such as those in the Problem-Solution discourse pattern, which serve to connect and relate and mark off larger chunks where a major change of emphasis and content may be introduced. Examples of larger chunk linkage operators would be “Meanwhile, back at the ranch” in a novel, or “And finally” in the television news, or “But first” in television announcements of upcoming programmes.

It is particularly interesting that evidence is found that sound linkages may well operate to hold together larger rather than smaller text chunks. The patterns of text are thus complex: some linkages may cluster around detail, and others may relate these to larger text divisions.