Despite being neglected by theorists for the larger notion of incipit, it is generally accepted that the first sentence of a novel is special. Taking the point of view of the reader, this dissertation aims at characterizing this specificity and identifying its causes and effects, using the tools and concepts of linguistics. 402 first sentences were studied, taken from novels published from 1980 to 2011 and written in English. Beginnings are cognitive non-sense. The opening sentence inherits this specificity and places the reader in front of the arbitrariness inherent to creation. The emerging fictional discourse also raises the question of its logical status and makes the opening sentence appear as a hinge between reality and fiction. For literature theorists and structuralists, beginnings have functions (programmatic, codifying). Also, the first sentence is often seen as a crucial brick in the fictional apparatus. It is the opening fringe of a new speech territory, ie a game one, and the edge where enunciative authority is being transferred from the author onto the narrator. It sets the enunciative and narrative frame-work of the story, as well as its first contents, of which readers elaborate mental representations. A study of the anatomy of the corpus reveals its lack of grammatical specificity, part of the game being for the opening sentence not to appear as such. The result is an enigmatic sentence which, for the purpose of the game, hijacks cognitive functions meant for serious matters. This cognitive hijacking is a seduction, it facilitates fictional immersion, which also is an aesthetic experience. The specificity of the first sentence is cognitive and affective.