ABSTRACT
Everyone agrees that the question of literary genres is fairly tangled. We can distinguish three factors that conspire to this result. The first is due to the recent history, romantic and post-romantic, of generic theories and of the status these theories have granted generic categories. The second results from the difficulty of defining the status of the textual classes that are supposed to form genres. The third results from a specificity of generic names, the fact that they are partly self-referential and consequently contribute to the institution of the historical reality they claim to describe. I propose here to take the measure of these difficulties and to indicate a few ways out. The theory of genres that will be outlined implicitly will be much less ambitious than the ones that are generally recognized. I hope what is lost in theoretical force will be at least partly counterbalanced by a greater descriptive exactness.
