Going back to Toury's map of Translation Studies, whereas Descriptive Translation Studies and Translation Theory are mutually dependent, the relationship between the two "pure" branches of the discipline and its applied extensions into the world is unidirectional. Each applied extension of Translation Studies (e.g. translator training, translation aids, translation criticism or translation planning) formulates and applies a different set of bridging rules by drawing on theoretical assumptions and predictions as well as empirical evidence concerning translational behaviour (Toury 1995:

17–18). Moreover, each applied activity sets its own bridging rules by also drawing on areas of scholarly enquiry that lie outside Translation Studies. The role of the applied extensions of the discipline is to establish rules of translational behaviour that are expressed in terms of prescriptive statements which "tell others what they should have done and/or should be doing" (1995: 19) if they conform to the norms set by translation practitioners, i.e. translator trainers, translation critics or translation planners.

Chesterman (2000) advocates a causal model for Translation Studies, considered to be the only model that incorporates all four categories of hypotheses formulated in the philosophy of science: interpretive, descriptive, explanatory and predictive. Chesterman also asserts the importance of putting forward and systematically testing, with appropriate empirical tools, explicit explanatory and predictive hypotheses so as to be able "to develop corroborated hypotheses into probabilistic laws, as envisaged by Toury" (2000:26). He believes an explicit causal model of translation would make a significant contribution to the development of a translation theory or theories that would explain the link between causal conditions, translation profile features and observed effects. It would also be applicable in translator training and quality assurance (Chesterman 2000, 2005). Unlike the comparative and process models, a causal model can in fact accommodate prescriptive statements, since stating what translators or translator trainees should or should not do in a given context is the same as formulating implicit predictive hypotheses of effect. These prescriptive hypotheses (e.g. if a given norm is accepted/broken under given conditions, certain desirable/undesirable effects will be produced) can be explicitly formulated and tested in the translation classroom. This can enhance our understanding of how causal conditions, translation profile features and observed effects are related to each other. It could also provide the type of information that a translator, a translator trainee, or a teacher of translation wants to know about translation, for example how to produce high-quality translations or what the likely effects are of a specific translation choice, such as the tendency to avoid the use of lexical anglicisms when translating English economics articles into Italian (Musacchio 2005; Laviosa 2006).

From Toury's perspective, applying a causal model in the translation classroom can provide practitioners with an analytical framework they can draw on for carrying out their own systematic research into the product and process of translation, and this application would not be with a view to accounting "either for possibilities and likelihoods or for facts of actual behaviour" (Toury 1995:19). The latter aims are the tasks of theory and description respectively. The pedagogical application would be with a view to setting norms in a more conscious way, since the applied extensions "cannot be anything but prescriptive" (1995:19), just as, by the same token, "theory should certainly not be concerned with bringing about changes in the world of our experience" (1995:273).

责任编辑：admin