In Poetic Effects Adrian Pilkington proposes that Sperber and Wilson’s (1986;

In Poetic Effects Adrian Pilkington proposes that Sperber and Wilson's (1986; 1995) relevance theory can help fill the explanatory gap between textual patterns and their potential aesthetic effects. Before applying relevance theory to poetic texts, Pilkington first presents his position with regard to literary theory. This position is not unproblematic, however, and therefore requires some discussion. Pilkington contrasts two well-known, though somewhat old-fashioned, approaches to literariness: text-based and semiotic/structuralist approaches. Pilkington's allegiance is clearly to the former. In fact, his attack on structuralism is surprisingly harsh, even unfair. For instance, he fleetingly ridicules the structuralist custom of identifying binary oppositions in a text by criticizing, in a hypothetical example, the opposition between sun and moon because 'it can mean almost anything. It has been used to signal the following distinctions: male/female, strength/weakness, reason/emotion, constancy/ fickle-ness' (p.24). Unfortunately, Pilkington here misses the significance of the fact that traditionally the items of these pairs have been systematically correlated, so that investigating this opposition may not at all be the gratuitous exercise he makes it out to be. Pilkington's fault-finding with Culler's (1975) 'reading conventions' also strikes me as prejudiced. He calls the seven conventions paraphrased from Culler's book 'vague and ad hoc', while the 'fiction convention' is even 'completely vacuous', and he concludes that the conventions 'clearly can form no part of a serious pragmatic account of literary communication' (p.25). Similarly, Pilkington rejects the work of reader-response theorist Siegfried Schmidt, who discusses literary reading strategies in terms of the aesthetic and polyvalence conventions (as opposed to the 'fact' and 'monovalence' conventions that characterize the reading of non-poetic texts; see e.g. Schmidt, 1991). Pilkington claims that neither Culler nor Schmidt presents 'the kind of proposals that can generate serious theoretical and empirical claims' (p.32). This is simply not true. Empirical work rooted in this tradition has been done for instance by Zwaan (1993), Steen (1994) and Forceville (1999), yielding non-trivial conclusions about the influence of genre on the interpretation of texts. Pilkington appears to suspect the 'reading strategy' advocates of seeing as much poetic merit in an alliterating advertising catchphrase as in a Shake-spearean sonnet, and of reducing 'aesthetic experience' to a routine application of a set of rules. But Culler's demonstration that a 'poetic' reading of a news