Prototype effect on subject change in translation A study in a French⇄Finnish bidirectional literary corpus This doctoral dissertation investigates the link between translation and prototype effect through the study of syntactic subject change in a French-Finnish literary bidirectional parallel corpus built specifically for the study. It examines the controversial topic of so-called translation universals and more widely the cognitive approach in the explanation of translation tendencies. More specifically, the study investigates the subject changes found in the corpus through six explanatory hypotheses proposed in translation studies to explain typical features of translations (translation universals, translational laws, translation figures, the gravitational pull hypothesis, orthonymy/prototipicality and anthropocentrism). The subject change is a particularly good candidate to test the explanatory power of the proposed hypotheses because the grammars of both Finnish and French describe the prototypical subject as being animate and cognitively more salient than an inanimate subject. The study found 320 relevant subject changes in the Finnish-->French and 408 in the French-->Finnish parallel corpora. These are categorized into four different types: 1. animation, 2. inanimation, 3. neutral animated change and 4. neutral inanimated change. The most frequent ones corresponded to animations of the subject, i.e. in most cases there occurred a humanisation of the subject in translation (60% in the French-->Finnish corpus and 49,5% in the Finnish-->French corpus). A thorough analysis of the relevant changes brought up five contextual factors that seemed to promote subject change in the corpus. Three factors are semantic (action, experience and possession) and two are pragmatic (particularization and homogenisation). The main contributions of the study are a methodology developed to analyse subject change between two languages from different language families, making use of the notion of shadow translation and a theoretical framework in the form of hypotheses built upon the five factors found. The hypotheses proposed make predictions about subject change that can be easily tested and further developed in other studies. Finally, the analysis shows that prototypicality is the most powerful explanation for the subject changes found in the corpus. This suggests that the most common way to say something and the anthropocentric conception of the world has an influence on the reading of the source text and thus also on the translation of the syntactic subject.