There is an important participant missing in existing models of translation: the translating institutions (corporations, churches, governments, newspapers) which directly or indirectly use the services of translators. In my view, the goals of a translating institution are what determines the general approach taken in the translations it produces: whether they are relatively literal or free, whether the language is conventional or innovative, whether metaphors are eliminated or retained, and so forth.
Here is an example. In Freud and Man's Soul, Bruno Bettelheim (1984, p. 53 ff.) claims that the English translators of Freud were wrong to substitute Greco-Latin words for Freud's everyday German. Why, he asks, is "das Es" rendered as "the Id" rather than "the It"? Why not the same approach as in the French translation ("le Ça")? According to Bettelheim, the reason is that in the English-speaking world, the psychoanalytic associations under whose aegis the translations were done wanted to create a science of psychoanalysis, so they needed to eliminate the humanist style and substitute a scientific one.
Assuming for the sake of argument that Bettelheim's understanding of Freud as a humanist is correct, the important thing to see here is that the non-equivalence was deliberate: no change in level of language was necessary in order to adapt the text to the target culture. The change was determined by the goal of the translating institution. Indeed there is a problem with Bettelheim's calling this a mistranslation, when what he is really objecting to is the goal of the institution. Given that goal, the approach the translators used was the right one. To speak of mistranslation in such cases suggests that there exists a "correct" way of translating which can be determined just
by looking at the wording and purpose of the source-text, and constraints imposed by the target culture. (Translation criticism might benefit from a limitation on the term "mistranslation", restricting it to mechanical reading and writing errors, and errors arising from the translator's lack of language knowledge, referential knowledge and cultural knowledge.)
The institutional factor in translation is to be distinguished from two factors commonly mentioned in translation studies: the "customer" factor and the "cultural" factor. Mistranslation (as just defined) can arise from a customer's action, for example when the customer sets too short a deadline or fails to provide documentation. But the customer does not determine the approach to translation. Customers may well ask for literal renderings, but whether or not literal renderings get produced will depend on whether the institution's doctrine of translation allows for this approach.
The cultural factor in translation is a set of background conditions whereas as the institution is an actor in the translation process. Furthermore, the institution does not act on behalf of an entire culture; radier it serves specific groups by producing translations that address specific readerships. The translation method dictated by its goals may actually hinder understanding for other readers, as in the Freud case.
Note, incidentally, that the translating institution is not necessarily located within the target culture. It can be a bi- or multi-lingual institution, or it can be associated with a third culture (e.g. the American Bible Society, which translates Ancient Greek into the languages of Third World countries), or with the source culture (e.g. French-to-English translation by the Québec Government as opposed to the federal or Ontario government).
Returning to the main argument, many readers will perhaps agree that translation of Freud or the Bible is institutionally determined. But what about the translation of a memo on acid rain? What happens when a public ...