Coherence

Parts I and II covered words and sentences and part III covered interactive features. But most texts—arguments in particular—consist of sentences combined into meaningful, coherent passages, the subject of part IV, beginning with this chapter. The coherence of a text is a perception built in the minds of readers who make sense of texts both from the bottom up, as they construe linear strings of words, and from the top down, as they impose meaning according to probable surmises about the kind of text they are encountering and what it is trying to accomplish. This chapter takes up the consideration of passage structure at the most basic level of connections between clauses in sequence. It first builds on the work of Halliday and Hasan in describing signs of cohesion between clauses by means of reference, substitution, ellipsis and lexical ties (i.e., words from the same lexical field). Then, using insights into the given/new structure of information in a typical sentence, a more useful account is constructed of how a sequence of clauses can maintain the same topic or use the new information (comment) from a preceding clause to open the next clause. However, clauses in sequence can abandon such explicit signs of connection when they can rely on the audience's background knowledge (in schemas, scripts, or story grammars) to fill in the connections. It is also possible to impose a meaning relation between two clauses with an explicit transition word or phrase. The set of transitions available in English can be grouped under fixed meaning relations, and a system of these relations is possible (offered in an appendix), based on the rhetorical insight that unexpected relations have to be signaled, while expected relations may or may not be, depending on what the arguer wants to emphasize. Most actual texts combine the sources for cuing sentence-to-sentence coherence that are described in this chapter.