A speech has two parts. You must state your case, and you must prove it. You cannot either state your case and omit to prove it, or prove it without having first stated it; since any proof must be a proof of something, and the only use of a preliminary statement is the proof that follows it. Of these two parts the first part is called the Statement of the case, the second part the Argument, just as we distinguish between Enunciation and Demonstration. The current division is absurd. For 'narration' surely is part of a forensic speech only: how in a political speech or a speech of display can there be 'narration' in the technical sense? or a reply to a forensic opponent? or an epilogue in closely-reasoned speeches? Again, introduction, comparison of conflicting arguments, and recapitulation are only found in political speeches when there is a struggle between two policies. They may occur then; so may even accusation and defence, often enough; but they form no essential part of a political speech. Even forensic speeches do not always need epilogues; not, for instance, a short speech, nor one in which the facts are easy to remember, the effect of an epilogue being always a reduction in the apparent length. It follows, then, that the only necessary parts of a speech are the Statement and the Argument.
These are the essential features of a speech; and it cannot in any case have more than Introduction, Statement, Argument, and Epilogue.
'Refutation of the Opponent' is part of the arguments: so is 'Comparison' of the opponent's case with your own, for that process is a magnifying of your own case and therefore a part of the arguments, since one who does this proves something. The Introduction does nothing like this; nor does the Epilogue-it merely reminds us of what has been said already. If we make such distinctions we shall end, like Theodorus and his followers, by distinguishing 'narration' proper from 'post-narration' and 'pre-narration', and 'refutation' from 'final refutation'. But we ought only to bring in a new name if it indicates a real species with distinct specific qualities; otherwise the practice is pointless and silly, like the way Licymnius invented names in his Art of Rhetoric-'Secundation', 'Divagation', 'Ramification'.