Let us discuss the art of poetry in general and its species—the effect which each species of poetry has and the correct way to construct plots if the composition is to be of high quality, as well as the number and nature of its component parts, and any other questions that arise within the same field of enquiry. We should begin, as it natural, by taking first principles first.
Epic poetry and the composition of tragedy, as well as comedy and the arts of dithyrambic poetry and (for the most part) of music for pipe or lyre, are all (taken together) imitations. They can be differentiated from each other in three respects: in respect of their different media of imitation, or different objects, or a different mode (i.e. a different manner).
Those who imitate, imitate objects; and these must be either admirable or inferior. (Character almost always corresponds to just these two categories, since everyone is differentiated in character by defect or excellence.). Alternatively they must be better people than we are, or worse, or of the same sort (compare painters: Polygnotus portrayed better people, Pauson worse people, Dionysius people similar to us).
In general, two causes seem likely to have given rise to the art of poetry, both of them natural.
Imitation comes naturally to human beings from childhood (and in this they differ from other animals, i.e. in having a strong propensity to imitation and in learning their earliest lessons through imitation); so does the universal pleasure in imitations. Wat happens in practice is evidence of this: we take delight in viewing the most accurate possible images of objects which in themselves cause distress when we see them (e.g. the shapes of the lowest species of animal, and corpses). The reason for this is that understanding is extremely pleasant, not just for philosophers but for others too in the same way, despite their limited capacity for it.
Tragedy is an imitation of an action that is admirable, complete and possesses magnitude; in language made pleasurable, each of its species separated in different parts; performed by actors, not through narration; effecting through pity and fear the purification of such emotions.
(By “language made pleasurable” I mean that which possesses rhythm and melody, i.e. song. By the separation of its species I mean that some parts are composed in verse alone; others by contrast make use of song.)
So tragedy as a whole necessarily has six component parts, which determine the tragedy’s quality. The medium of imitation comprises two parts, the mode one, and object three; and there is nothing apart from these.
Tragedy is not an imitation of persons, but of actions and of life. Well-being and ill-being reside in actions, and the goal of life is an activity, not a quality; people possess certain qualities in accordance with their character, but they achieve well-being or its opposite on the basis of how they fare. So the imitation of a character is not the purpose of what the agents do; character is included along with and on account of the actions. So the events, i.e. the plot, are what tragedy is there for, and that is the most important thing of all.
Any beautiful object, whether a living organism or any other entity composed of parts, must not only possess those parts in proper order, but its magnitude also should not be arbitrary; beauty consists in magnitude as well as order. For this reason no organism could be beautiful if it is excessively small (since observation becomes confused as it comes close to having no perceptible duration in time) or excessively large (since the observation is then not simultaneous, and the observers find that the sense of unity and wholeness is lost from the observation, e.g. if there were an animal a thousand miles long). So just as in the case of physical objects and living organisms, they should possess a certain magnitude, and this should be such as can readily be taken in at one view, so in the case of plots: they should have a certain length, and this should be such as can readily be held in memory.
It is also clear from what has been said that the function of the poet is not say what has happened, but to say the kind of thing that would happen, i.e. what is possible in accordance with probability or necessity. The historian and the poet are not distinguished by their use of verse or prose; it would be possible to turn the works of Herodotus into verse, and it would be a history in verse just as much as in prose. The distinction is this: the one says what has happened, the other the kind of thing that would happen.
So there are these two parts of the plot—reversal and recognition; a third is suffering. Of these, reversal and recognition have already been discussed; suffering is an action that involves destruction or pain (e.g. deaths in full view, extreme agony, woundings and so on).
The construction of the best tragedy should be complex rather than simple; and it should also be an imitation of events that evoke fear and pity, since that is the distinctive feature of this kind of imitation. So it is clear first of all that decent men should not be seen undergoing a change from good fortune to bad fortune—this does not evoke fear or pity, but disgust. Nor should depraved people be seen undergoing a change from bad fortune to good fortune—this is the least tragic of all: it has none of the right effects, since it neither agreeable, nor does it evoke pity or fear. Nor again should a very wicked person fall from good fortune to bad fortune—that kind of structure would be agreeable, but would not excite pity or fear, since the one has to do with someone who is suffering undeservedly, the other with someone who is like ourselves (I mean, pity has to do with the underserving sufferer, fear with the person like us); so what happens will evoke neither pity nor fear.
We are left, therefore, with the person intermediate between these. This is the sort of person who is not outstanding in moral excellence or justice; on the other hand, the change of bad fortune which he undergoes is not due to any moral defect or depravity, but to an error of some kind. He is one of those people who are held in great esteem and enjoy great good fortune, like Oedipus, Thyestes, and distinguished men from that kind of family.
It is possible for the evocation of fear and pity to result from the spectacle, and also from the structure of events itself. The latter is preferable and is the mark of a better poet. The plot should be constructed in such a way that, even without seeing it, anyone who hears the events which occur shudders and feels pity at what happens; this how someone would react on hearing the plot of the Oedipus. Producing this effect through spectacle is less artistic, and is dependent on the production. Those who use spectacle to produce an effect which is not evocative of fear, but simply monstrous, have nothing to do with tragedy; one should not seek every pleasure from tragedy, but the one that is characteristic of it. And since the poet should produce the pleasure which comes from pity and fear, and should do so by means of imitation, clearly this must be brought about in the events.
(Clearly, therefore, the resolutions of plots should also come about from the plot itself, and not by means of a theatrical device, as in the Medea, or the events concerned with the launching of the ships in the Iliad. A theatrical device may be used for things outside the play—whether prior events which are beyond human knowledge, or subsequent events which need prediction and narration since we grant that the gods can see everything. But there should be nothing irrational in the events themselves; or, failing that, it should be outside the play, as for example in Sophocles’s Oedipus.)
Since tragedy is an imitation of people better than we are, one should imitate good portrait-painters. In rendering the individual form, they paint people as they are, but make them better-looking. In the same way the poet who is imitating people who are irascible or lazy or who have other traits of character of that sort should portray them as having these characteristics, but also as decent people. For example, Homer portrayed Achilles as both a good man and a paradigm of obstinacy.
The best recognition of all is that which arises out of the actual course of events, where the emotional impact is achieved through events that are probable, as in Sophocles’ Oedipus and the Iphigeneia (her wish to send a letter is probable). Only this kind does without contrived tokens and necklaces. Second-best are those which arise from inference.
The most important quality in diction is clarity, provided there is no loss of dignity. The clearest diction is that based on current words; but that lacks dignity (as can be seen from the poetry of Cleophon, and that of Sthenelus). By contrast, diction that is distinguished and out of the ordinary when it makes use of exotic expressions—by which I mean non-standard words, metaphor, lengthening, and anything contrary to current usage. […] So what is needed is some kind of mixture of these two things: one of them will make the diction of the ordinary and avoid a loss of dignity (i.e. non-standard words, metaphor, ornament and other categories I mentioned earlier), while current usage will contribute clarity.
Homer deserves praise for many reasons, but above all because he alone among poets is not ignorant of what he should do in his own person. The poet in person should say as little as possible; that is not what makes him an imitator. Other poets perform in person throughout, and imitate little and seldom; but after a brief preamble Homer introduces a man or a woman or some other character—and none of them are characterless: they have character.
While it is true that astonishment is an effect which should be sought in tragedy, the irrational (which is the most important source of astonishment) is more feasible in epic, because one is not looking at the agent. The pursuit of Hector would seem preposterous on stage, with the others standing by and taking no part in the pursuit while Achilles shakes his head to restrain them; but in epic it escapes notices.
Homer, in particular, taught other poets the right way to tell falsehoods. This the false inference In cases where the existence or occurrence of A implies the existence or occurrence of B, people imagine that if B is the case than A also exists or occurs—which is fallacious. So if A is false, but its existence would entail the existences or occurrence of B, one should add B; then, on the basis of its knowledge that B is true, our mind falsely infers the reality of A as well. An example of this can be found in the bath-scene.
Probable impossibilities are preferable to implausible possibilities. Stories should not be constructed from irrational parts; so far as possible they should contain nothing irrational—or, failing that, it should be outside the narration (like Oedipus’ ignorance of the manner of Laius’ death) and not in the play itself (like the report of the Pythian Games in Electra, or the man who comes from Tegea to Mysia without speaking in the Mysians). Saying that the plot would have been ruined otherwise is absurd; plots should not be constructed like that in the first place. But is one does posit an irrationality and it seems more or less rational, even an oddity is possible; the irrationalities involved in Odysseus’ being put ashore in the Odyssey would be manifestly intolerable if a second-rate poet had composed them, but as it is the poet conceals the absurdity with other good qualities, and makes it a source of pleasure.
Furthermore, if the objection is that something is not true, perhaps it is as it ought to be; e.g. Sophocles said that he portrayed people as they should be, Euripides as they are. That is the solution to use.
Tragedy has everything epic does (and it can even make use of its verse-form), and additionally it has a major component part music and spectacle; this is a source of intense pleasure. […] Also, the end of imitation is attained in shorter length; what is more concentrated is more pleasant than what is watered down by being more extended in time ( I mean, for example, if one were to turn Sophocles’ Oedipus into as many lines as the Iliad has).
So tragedy surpasses epic in all these respects, and also in artistic effect (since they should not produce any arbitrary pleasure but the one specified); clearly, then, because it achieves its purpose more effectively than epic, tragedy must be superior.