De la théorie de la description à la description chez Julien Gracq

Sammanfattning: This thesis deals with some aspects of description in Julien Gracq’s fiction. These aspects are selected as a result of a double investigation. Firstly, the theory of description as it has taken form with the contributions of Barthes, Genette, Ricardou, Bal, Riffaterre, Hamon, Weinrich, Adam and Petitjean, Reuter, Molino, Brassart is presented. Secondly, the main critical analyses of description in Gracq’s fiction are exposed. The results of this double investigation show the gap between, on the one hand a formal, non-referential theory and on the other an interpretative and partially referential criticism. Two main domains appear where a confrontation between theory and criticism seems to be necessary. The second part of the thesis deals with this confrontation. The first domain concerns the relations between description and narration. Hamon’s notion of ”demarcating system” is questioned. The way Gracq’s critics extract the descriptions they analyse suggests that the ”demarcating system” is a problematic notion, especially when confronted with the notion of ”descriptive system”. Iteration according to Genette’s definition is also studied. Iterative description and iterative narration in Gracq’s fiction have some important implications for the theoretical model of description. On the one hand, Hamon’s notion of ”semantic equivalence” appears inappropriate in iterative descriptions; on the other hand, iterative narration seems to be a hybrid, disturbing some established theoretical notions. The results show that iterative narration should be considered as a descriptive mode and that it should be taken into account by the theorists of description. The second domain concerns the relations between description and representation. The principle causes of the weak representational value of descriptions in Gracq’s fiction are exposed: metaphors, evaluation, modalisation. There is a noticeable distance between the theoretical study of legibility and the analysis of description’s representational value by Gracq’s critics. In order to test the theoretical hypothesis of the cohesive function assumed by the ”plans de texte” (Adam) or the ”grilles descriptives” (Hamon), an empirical experiment is carried out. The results show the necessity to revise the theoretical assumption of semantic coherence as a structuring principle for description. The conclusions of this thesis validate the hypothesis of the relevance of the confrontation between the theoretical and the critical field. The use of a new corpus turns out to be rewarding. In order to improve the theory of description some modifications are proposed (study of iteration and of the representational and referential dimensions of description); some new aspects of Gracq’s description are also discovered, in the light of this theory.