340 Aristotle's treatise on poetry. parts, but in the drama the effect of such a plan is far different from what is expected. As a proof of this, those poets who have formed the wliole of the destruction of Troy into a Tragedy, instead of confining themselves [as Euripides, but not jEschylus, has done, in the story of Niohe to a part, have either been condemDed in the representation, or have contended without success. Even Agathon has failed on this account, and on this only ; for in revolutions, and in actions, also, of the simple kind, these poets succeed wonderfully in what they aim at ; and that is, the union of tragic effect with rtioral tendency: as when, for example, a character of great wisdom, but without integrity, is deceived, like Sisyphus; or a brave, but unjust man, conquered. Such events, as Agathon says, are probable, " as it is probable, in general, that many things should happen contrary to probability." The chorus should be considered as one of the persons in the drama ; should be a part of the whole, and a sharer in the action ; not as in Euripides, but as in Sophocles. As for other poets, their choral songs have no more connexion with their subject than with that of any other Tragedy; and hence they are now become detached pieces, inserted at pleasure ; a practice introduced by Agathon}. Yet where is the difference between this arbitraiy insertion of an ode, and the transposition of a speech, or even of a whole episode, from one Tragedy to another % Cap, XV.2 With respect to the Manners, four things are to be attended to by -Of the best the poet. modes of ex- ^ pressing the to.e'actors'^ ^ ^^^ Greek is 5t6 efx^oXi^ia ^dovaiv, irpdoTOV dp^avros 'AydOojvos too toiovtov, and Ritter, like most of the commentators, understands ep.(36ifji.a as cantica ah argumento tragcedicB aliena et 'pro arbitrio poetce inserta. So that Agathon committed the fault deprecated by Horace {A. P. 193): Actoris partes chorus officiumque virile Defendat, neu quid medios intercinat actus Quod non proposito conducat et hcereat apte. Cicero uses i/x^6iov in the sense of a mere episode. 2 I have transposed this chapter to its proper place after the eighteenth chapter, in compliance with the suggestion of Spengel, who writes as follows (Munich Transactions, u. s. p. 246) : "The chapter about the rfdr) is erroneously inserted here, and is the cause of all the confusion. If it is removed from its present place, the duaypibpiais imme- diately follows ; and it is clear that it is here mentioned and that the remark is made : eipTjTac irpbrepov, — for between the first mention (cc. X. XI,) and the present full dis- cussion many other subjects have been introduced. Now it must be remembered that we do not find in the MSS. such divisions and separations of the clauses as we give in our editions : Hepl jxkv ovv TTJs tCjv TrpayfxaTOJV cvardaews kuI TTolovs Tivds etvai Set rot>s /x69ov5 eiprjrai i/cai'tDs, Tlepl 5^ rd yjdr} riTTapd ecrriv Cov del (TTOxd^eadat. So that the former terminates the chapter, and the latter commences a new one. But such clauses are regarded by the old writers, and in a grammatical sense rightly, as an indivisible whole. I am then convinced that the leaf consisting of forty lines, which con-
