Gutt (2000), first published in 1991, is the book most translation theorists associate with the application of Relevance Theory to translation. Gutt's aims were laudable: to show that we can explain (rather than merely classify and describe) the facts of translation without the need for a special translation theory, focussing on translation as communication. Like all good theories, that put forward in Gutt's book aims not to drive but to explain its subject. Thus Gutt said, quite rightly, in a recent contribution to the Relevance-Theory email list, that "I do not advocate any particular translation approach" (2004). But Gutt's book is both dense and a little disorderly; it is sometimes difficult to make out a clear line of argument. And the notion that translation does not need its own theory, while clearly a strong argument for Relevance Theory, in the sense that it is shown to explain data outside the range of those it was originally called on to clarify, has been viewed with surprise or wariness by translation theorists such as Malmkjar (1992), though others such as Hatim (2001) have acknowledged its usefulness for translation. Gutt made a number of suggestions about how translation could be explained using Relevance Theory. The most important of these are:
i) translation, as communication, works under the assumption of relevance (that what the translator intends to communicate to the audience
is relevant enough to them to make processing it worthwhile);
ii) a translated text is an instance of interpretive, as opposed to descriptive, use (the translator is saying what someone else meant);
iii) texts in which the way of saying – the style – plays an important role require direct translation, as opposed to indirect translation, which, like indirect quotation, just gives the substance.
In English-speaking countries comparatively little work has been done with Relevance Theory and translation. Exceptions are Boase-Beier (2004a, 2004b),Hatim & Mason (1990), Hatim (2001). The latter, in particular, discusses the difference between notions such as direct and indirect translation, but does not consider how this distinction affects the translation of style in very great detail. In Spain, on the other hand, there are more studies that directly apply a Relevance-Theory approach, such as Dahlgren (1998), Dahlgren (2000), Edwards (2001). As I suggested in the previous section, it may be the focus on relevance itself which has rendered the approach less useful than it might be for translation. Taking relevance as given, here are some insights from Relevance Theory and cognitive pragmatics more generally (the first four are adapted from Boase-Beier 2004a) which might be useful for studying what happens to style when we translate:
i) The notion of mind style can be integrated into translation theory as a set of weak implicatures which are communicative clues to a cognitive state.
ii) Relevance Theory allows for the importance of a cognitive state as that which a translator will try to recreate, rather than meaning in a truth-conditional sense.
iii) By allowing a view of style as weak implicatures, Relevance Theory provides a framework and a legitimation for the translator's interpretive freedom and the creativity of the translation act.
iv) By tying poetic effect to the extra work, in terms of maximal relevance, that stylistic features call for, a Relevance Theory view can help explain the common intuition of translators that preserving style helps recreate the effects of the source text on the target reader.
V)An important difference between the way literary and non-literary texts are translated is that the former will tend to require direct translation (which preserves the style) while the latter will tend to require indirect translation (see Gutt 2000:68ff.).