This study has two purposes. As a social-psychological contribution to the theory of translation, it points to some of the advantages and drawbacks of the researcher's participation in the translation of scientific texts. As a contribution to social-psychological theory, it wishes to demonstrate that forms of cooperation cannot be planned in abstracto, without taking the overall social activity of the actors into account, of which participation in cooperation is only one part.
One of the most original and ingenious inventions in the social sciences dates back to the early 1950's: the game known as the Prisoner's Dilemma (see Rapoport 1982). With perplexing accuracy, it puts it quite plainly that, first, action taken by individuals upon perfectly rational deliberation does not necessarily lead to collective rationality. Also, showing how a social structure may produce forces motivating the individual, the Prisoner's Dilemma has something to give to social psychologists. Even in the event that the prisoners have had the opportunity to discuss different strategies and jointly decide on adopting one, each is tempted to betray the other - and both are afraid that they will be betrayed. Psychological motives, the temptation and the fear, arise out of the logic of the social situation.
Our intention has been to show that translation as a social activity involves motivating forces, assumptions to do with competence, and restrictive factors that all shape the scientist-translator cooperation irrespective of their deliberate pursuits. Therefore, rather than planning it oh an abstract basis, the working method has to be deduced from the logic of action. In doing so, we will see that cooperation cannot be symmetric; the weight is necessarily on the translator's role. The scientist, then, comes into the picture when the translator needs help; he does not have to be prepared for regular and face-to-face interaction with the translator, but only to make sure that the message of his text is conveyed (provided that he has the competence in the target language). Cooperation between translator and editor, in turn, is much more dependent on face-to-face interaction.