In this indispensable essay, the French literary theorist Gerard Genette sets out to disentangle the 'knot of confusions' that has, for more than two centuries, surrounded one of the most familiar of generic classifications, the division of literature into three kinds: lyric, epic and dramatic. Genette shows how this 'all-tooseductive' triad, which came to dominate Romantic genre theory (especially in Germany) and has also been the foundation of many modem genre-systems, conflates two logically dissimilar categories: 'mode' (an essentially linguistic category, specifying means of enunciation: e.g. narration, dramatic imitation) and 'genre' (a literary category, defined by thematic as well as formal criteria, and indicating a historically existent type: e.g. the epic); and how this tripartite division, habitually but erroneously attributed to Aristotle, has impeded the development of a coherent classification of literature and adequate theory of genre. The critique is brilliantly sustained and meticulously documented, and has profound implications not only for our understanding of the history of genre theory, but also for our reckoning of the tasks that still confront the genre theorist. Yet the tone of the essay is anything but censorious: rarely has so venerable a tradition of intellectual confusion been overturned with such modesty and good humour - qualities which animate all of Genette's work, even his most technical. Moreover, as he would be the first to acknowledge, Genette himself is by no means free of the systematising impulse whose potential dangers he so lucidly analyses, nor of the 'terminological rapture' which Derrida (below, p. 224) sees as the mark of the true genre theorist ('architext' and 'archigenre' are

but two of his many neologisms). The short extract reproduced here formed the original conclusion of the essay; the book version, which runs to 85 pages, carries a postscript in which Genette responds to criticisms of the essay, reviews yet another flawed genre-system, and - very tentatively - sketches his own (threedimensional) model of the relation between modal, thematic, and formal categories.