Abstract

In this article I posit a pragmatic model of how sound-patterning in poetry contributes to interpretation, with particular attention to the cognitive processes involved. The model is based in cognitive pragmatics (relevance theory) and psycholinguistics (spreading activation models), and the focus here is largely on segmental patterning such as rhyme and alliteration, though there is also some reference to metrical features. The article draws on relevance theory notions of literary language conveying a range of weak implicatures and thus resulting in diffuse and non-propositional impressions (Sperber and Wilson 1995 [1986]), notions used so far in relevance theory approaches to literary features such as repetition and metaphor. The article develops Pilkington's (2000) argument that poetic form, in slowing down lines and anticipating rhyming words, allows for longer access and deeper exploration of information associated with lexical concepts. Such exploration can lead to a potentially wide range of interpretive material. Here I focus on associative, iconic, and mimetic effects, which, I argue, are enabled through the spreading of semantic properties along sequences of relatively contiguous similar sounding words, and through the association of formal properties with semantic and encyclopaedic information. Through the analysis, my aim is also to account for the instability of poetic interpretation. The argument is presented with reference to poetry by Pope and Keats.