If the features we have looked at so far in this section – gaps in the text, weak implicatures, the characteristic of fitting many different contexts, conveying ideologies, attitudes and feelings – are all typical of literary texts, then translations will need to make sure that none of these characteristics gets lost in the translation. All of them have the effect of demanding a greater involvement on the part of the reader than one would usually expect in a non-literary text. Unlike von Törne's poem, an actual left-wing newspaper report of the Nazi murderers would probably have supplied the subjunctive auxiliary, thus indicating a refusal on the newspaper's part to believe their ignorance. An extreme right-wing paper might have used an indicative to endorse it. And so on. One prerequisite for the survival of such possibilities in the translation is the recognition of subtle stylistic features such as ambiguity, contrasts, change of perspective, by the translator.

Cognitive approaches to style and translation rely on the interplay of stylistic universals with stylistic characteristics peculiar to an individual language, culture or view. Fowler (1977a) plays down the universal aspects because his purpose is to emphasize the importance of social context; however, he acknowledges that the two interact. And it has been recognized by countless critics (e.g. Vendler 1995) that literature appears to work by combining the universal with the particular; see also Goldsworthy (1998:43). Another way to see this is that different readers construct different mental pictures of the text (different text worlds, in Semino's vicw: 1997:192) because of their different backgrounds.

For translation, a cognitive view might suggest that what is universal will be more easily translated than what is culturally or linguistically diverse. Thus Gutt (2005) suggests that a translation of the New Testament story of the man ill with with paralysis, which includes his friends lowering him through the roof, will not make sense in a target culture which has steep thatched roofs. One possibility is to generalize so much as to reduce the cultural conflict to what is mutually comprehensible, even leaving out the mention of the roof if necessary.

Studies like that by Gutt (2000), which use Relevance Theory as a basis, differ from earlier views of the translation of mind in the text in that Relevance Theory does not speak of re-experiencing the author's thoughts, as for example Schleiermacher did, but of "enlarging mutual cognitive environments (MacKenzie 2002:4; Sperber & Wilson 1995:38ff.). This suggests that the readers Gutt mentions in the roof example just given cannot ever share the exact same connotations of roofs, whatever those were, with the original authors and readers, whoever they were. A common denominator, as Gutt (2005) suggests, is one way of creating the greatest overlap or greatest mutuality in cognitive environments of source-text author, source-text reader and target-text reader. Providing content in footnotes and illustrations is another, very different, way, but their aim is the same. Such strategies relate to the view of translation as a third language: a space where cognitive environments, because shared, are changed. Translation in this sense is attempting to bring about a meeting of minds. What this account ignores, though, is that the search for contexts which allow the source text to make sense could be part of the aim of the text.This will certainly be true if it is a literary text, where shared environments will not be provided but will be constructed by the reader.

But not only general knowledge of the world can be universal; so can what Kiparsky (1987) calls "categories of literariness" , such as metaphor or ambiguity. If these are not arbitrary but are psychological givens, it is necessary to consider, in looking at their translation, both the universal basis and what is specific to a language,culture, author, or text.

责任编辑：admin