This chapter first describes the main features of a translation as a creative product, few traditional notions of the creative process are then discussed, and finally a number of cognitive models and notions are presented, in order to show how they could be applied to the explanation of creative translation processes.

If we can show that translating in general involves creativity, sometimes to a similar degree to that involved in the writing of an independent text, we may, among other things, help improve the status of the profession

One might say that the looser the link (to the source text), the greater the chance for a translator to be creative.

In creativity research the creative product is defined as both novel and appropriate for the task (see Preiser 1976: 5).

Accordingly, we may say that a creative translation is a translation which (a) involves changes when compared with the source text, thereby bringing in something that is novel, and which (b) is also appropriate for the task that was set, i.e. the translation assignment (or purpose).

The traditional idea, to which I do not subscribe at all, is that the creative process is a mystery.

Translation scholars often use the word 'creative', but, with a few exceptions (notably Wilss 1988: 108-128), unfortunately also in a rather prescientific way.

Only recently have attempts been made to approach creative thought in translating in a rational and scientific way (see Beylard-Ozeroff et al. 1998).

Cognitive linguistics provides us with a number of models and notions that may serve to analyze the creative process. .

4.1. Fillmore's scenes and frames

4.2.RonaldLangacker'sfigure/ground alignment

4.3. Roger Schank's thematic organization points (TOPs)

The notions of frames, scenes, core and fuzzy edges, focusing and figure/ground alignment seem to be quite illuminating for understanding the role of comprehension and the creation of sense in the translation process.

As to the verbalization of sense in the target text, the techniques of picking out scene elements, changing ( or preserving) the focus, and using TOPs seem to lead to creative translations.

Our models may, of course, also serve to explain what goes on when translating is not successful. They can, for instance, be used to show that preconceived ideas prevent translators from seeing what the text is really about, as when the scenes in their minds do not fit into the frames of the text, and they might help us in evaluating translations by making use of the notion of a core and fuzzy edges.

The models presented are taken from cognitive linguistics and cognitive psychology (Schank), in other words, from theories of normal language processing.

When applied to translation, the cognitive models might have to be modified, or we may find that some of the models are more applicable than others. But then, the models could also be used to show in what respect translation differs from normal language processing.

We may find that turning around the figure/ground alignment, changing the focus and thinking via TOPs are rather rare in normal language processing and are typical only of solving specific translation problems.

Finally, I would like to mention a pedagogical line of research. Does knowledge of these (possibly modified) models produce creative translators?