If we were to project the notion of modified interlingual relativism on to the intralingual situation, the result would be to accept that different genres and text-types, while sharing universal features with other types of language, can be distinguished linguistically. This is in fact the common view of writers such as Reiß (1981). But we could also extend such intralingual variation to take in translation itself, seeing translated texts as embodying a separate language, a different kind of writing from untranslated texts, a "third code", in Frawley's words. It is a view that is backed up by the observations of critics: Bayley, for example, says of Cohn's Rilke (1992) that it invents a new language, "the language of 'Rilke-in-English" . The poet Michael Hamburger has developed such a style in translation partly as a way of respecting the original, not using it "as a springboard" for his own work. So his English translations of Hölderlin use a peculiarly Hölderlin-influenced English: "Into water, the holy-and-sober.

It is often said of Pound that he wrote Homage to Sextus Propertius "in a distinctive kind of 'translatorese'" . And indeed the very notion of translatorese or translationese suggests, usually in a pejorative sense, that translations have a language all of their own. There is empirical evidence for this in studies which have shown that translation tends to obey certain norms irrespective of those that might obtain in source language and target language: Ben-Ari (1998), for example, shows that repetition is avoided in the translation into French of Mann's Buddenbrooks, even though it serves a distinct literary function. French textbooks on stylistics do indeed exhort the writer to avoid repetition, but then so do German ones, e.g. Reiners (1990). Ben-Ari concludes (1998:78) that [a]voiding repetition is such a predominant norm that it seems to be found in all translated texts".

Seeing translation as a separate type of writing from non-translation can be dependent upon the translator's voice, which Schiavi (1966:3) argues is a composite of the source-text author's and the translator's own. This intuitive sense of translation as a distinctive text-type has been borne out by corpus studies such as Kenny (1998). Following Benjamin's view, translation can be seen as an experimental literary practice, which enables a translated text to distance itself from the norms of both source language and target language, thus becoming an "independent form of writing.

In fact, Venuti's description of a translated text stylistically defamiliarized not by an author's personal style but by heterogeneity and marginal uses of language is very close to Pilkington's description of the literary text itself as using "strange language, archaisms, 'phonetic roughening"

(2000:18). This suggests that a literary translation is, above all, consciously literary in that it is defamiliarized. Defamiliarization might thus serve not only to circumvent the differences between languages as Hyde (1993) suggested, but also to make literary translation more literary than non-translated literature.

责任编辑：admin