We will speak later about imitation in epic poetry and about comedy.¹ Let us now resume our discussion of tragedy by deducing the definition of its essential nature which follows from what we have said.

Tragedy, then, is an imitation, through action rather than narration, of a serious, complete, and ample action, by means of language rendered pleasant at different places in the constituent parts by each of the aids [used to make language more delightful], in which imitation there is also effected through pity and fear its catharsis ² of these and similar emotions.

By “language rendered pleasant” I mean language which has rhythm, [instrumental] melody, and song. By “aids at different places” I mean that some parts use metre alone while in other parts the language is aided by song. Since the imitation in tragedy is achieved through action, the orderly arrangement of what appeals to the eye must first of all necessarily be an essential part of tragedy. After this would come music and diction, since it is through these that the imitation is effected. By “diction” here I mean the metrical arrangement. What is meant by “music” is entirely obvious.

Now tragedy is an imitation of an action being carried out by certain individuals who must be certain kinds of persons in character and in thinking ³—the two criteria by which we determine the quality of an action; for character and one’s thinking are two natural causes of action, and it is because of these that all men fail or succeed. The plot is the imitation of the action to be presented. I here mean by “plot” the arrangement of the incidents. By “character” I mean that by which we determine what kinds of men are being presented; and by “one’s thinking” I mean that which manifests itself in all the characters say when they present an argument or even make evident an opinion. Therefore, every tragedy must necessarily have six elements according to which the quality of the tragedy is determined: ⁴ (1) plot, (2) character [indicants], (3) thought, (4) spectacle, (5) diction, and (6) music. We thus have three parts (1, 2, 3) which tell us what the tragedy proposes to achieve an imitation of, one (4) which tells us how it proposes to do so, and two (5 & 6) which tell us the means to be used. These constitute all the parts. Therefore, just about all the dramatists use these constituent forms—which is very natural since every drama equally admits of spectacle, character [indicants], plot, diction, music, and thought.

But the most important of these is the arrangement of the incidents of the plot; for tragedy is not the portrayal of men [as such], but of action, of life.⁵ Happiness and misery are the result of action, and the end [of life(?)] is a certain kind of action and not a quality. Men are the certain kinds of individuals they are as a result of their character; but they become happy or miserable as a result of their actions. Consequently, dramatists do not employ action in order to achieve character portrayal, but they include character because of its relation to action. Therefore, the incidents and plot constitute the end of tragedy, and the end is the greatest thing of all. Moreover, without action there could be no tragedy, but there could be tragedy without character. In fact, most of the new tragedies are without character. Speaking generally, there are many portrayers of imitations of which the same is true, as is seen in the case of the painters Zeuxis and Polygnotus. Polygnotus’ works contain excellent portrayals of character, but there is no character portrayal in a painting of Zeuxis. Furthermore, if anyone should arrange in proper order a number of speeches which showed moral character and were well written from the point of view of diction and the thought [of those represented], he would not achieve the end of tragedy. This end would be much better achieved by a tragedy which was inferior in these elements, yet had a plot and arrangement of incidents. Moreover, the two most compelling elements in tragedy—reversals and recognitions—are parts of the plot. A further proof is the fact that even those who undertake to write tragedy can become proficient in matters of diction and character [indicants] sooner than they can in arranging the incidents of the plot, as is proved by almost all the earlier poets. Therefore, the first principle and, as it were, the soul of tragedy is the plot.

The second is the character [indicants]. It is the same also in painting; for if anyone should make a painting by smearing the most beautiful colors at random on a surface his painting would not give as much pleasure as a [mere] figure done in outline. Tragedy is an imitation of action, and is for that reason principally concerned with characters in action.

The third part is the thinking ability of the characters, which is the ability to [think out and] say (1) what is possible [within the limits of the situation] and (2) what is fitting—a function which is the same as that of language when used in statesmanship and in oratory. Thus it happens that the early poets had their characters speak like statesmen, but contemporary ones have theirs speak like rhetoricians.

Character [indicants] are those elements which make clear, in matters where it is not otherwise evident, a person’s choice—what sorts of things he chooses or avoids. Therefore, speeches in which a character is in no way confronted with something he must choose or reject, do not have character [indicants].

Thought manifests itself in what the characters say as they prove or disprove something or make evident something universal.

The fourth thing in our list is diction, and by “diction” I mean, just as we stated earlier, expression [of thought] by means of language—a power which is the same in both metrical and non-metrical language.

Of the remaining elements, music has the greatest enriching power; and spectacle, while quite appealing, is the most inartistic and has the least affinity with poetry; for the essential power of tragedy does not depend upon the presentation and the actors. Moreover, for achieving the effects of spectacle, the art of the mechanic of stage properties is more competent than the art of poetry.

Since these matters have been defined, let us discuss next the kind of arrangement of incidents a tragedy must have, since this is the most primary and important essential of tragedy. We have laid it down that tragedy is an imitation of an action which is whole, complete, and of a certain magnitude. Now it is possible to have a whole which does not have magnitude. A whole is that which has a beginning, a middle, and an end. A beginning is that which of necessity does not follow anything, while something by nature follows or results from it. On the other hand, an end is that which naturally, of necessity, or most generally follows something else but nothing follows it. A middle is that which follows and is followed by something. Therefore, those who would arrange plots well must not begin just anywhere in the story nor end at just any point, but they must adhere to the criteria here laid down.

Furthermore, everything beautiful, both animate and those inanimate things which have been created from certain parts, must have not only the arrangement just described, but must have also a proper magnitude; for beauty consists in magnitude and arrangement. And so, a very small animal could not be beautiful, because our perception of it, taking place at close range and in an imperceptible amount of time, is confused. Nor could a very large animal, one a thousand miles long, for instance, be beautiful to us, because we cannot see it all at once and perception of it as a unity and a whole eludes us. Therefore, just as there must be a proper magnitude in the case of animate and created inanimate things—a magnitude which can be easily taken in as a whole by the eye—so must plots have a proper length—one which can be easily retained within the memory.

The limit which dramatic contests and the perceptive powers [of an audience] impose upon the length of a tragedy is not a part of the art of poetry; for if it had been the rule to enter a hundred tragedies ⁶ in the “tragic” contests, the time for each would have been limited by the water clock, as it is said was once the case in some places. The length-limit dictated by the very nature of the thing is always this: the greater the limit, provided the whole continues readily comprehensible, the more beautiful it will be because of its magnitude. And so, a general definition of magnitude for an action would be: that amount of magnitude in which events proceeding in succession according to probability and necessity veer around from bad to good fortune or from good to bad. This is a sufficient delimitation for magnitude.

Some people think a plot can be said to be a unified one if it merely centers about one person. But this is not true; for countless things happen to that one person some of which in no way constitute a unit. In just the same way there are many actions of an individual which do not constitute a single action. Therefore, those poets who write a Heracleid, a Thesiad, and poems of that kind appear to be following a wrong principle; for they think that since Heracles was a single individual any plot dealing with him must of necessity have unity. But Homer, just as he excels in other matters, seems, either through natural insight or art, to have been right in this also; for in composing an Odyssey he did not include an account of everything which happened to Odysseus. He omitted his being struck by a boar on Parnassus,⁷ for instance, and his feigned madness when the Greeks were assembling for Troy, because neither of these by having happened made probable or inevitable some other incident. But Homer has made both the Iliad and the Odyssey centre around an action which is unified in the sense we are demanding. Therefore, just as in other mimetic arts a unified imitation is an imitation of a single thing, in the same way the plot in tragedy, since it is an imitation of an action, must deal with that action and with the whole of it; and the different parts of the action must be so related to each other that if any part is changed or taken away the whole will be altered and disturbed. For anything whose presence or absence makes no discernible difference is no essential part of the whole.

It is evident also, from what has been said, that it is not the function of the poet to relate what has happened but what can happen according to the law of probability or necessity. Whether the writer is a historian or a poet is not to be determined by whether he uses metre or prose; for the writings of Herodotus could be put into metre and still be history just as completely with metre as without it. The difference between a poet and a historian is this: the historian relates what has happened, the poet what could happen. Therefore, poetry is something more philosophic and of more serious import than history; for poetry tends to deal with the general, while history is concerned with delimited particular facts. An instance of “the general” [with which poetry undertakes to deal] is this: what are the sorts of things which, according to the law of probability and necessity, various types of individuals tend to do and say? This is what poetry aims to make evident when it attaches names to characters. An instance of “particular facts” [with which history deals] is: what did Alcibiades do, or what was done to him?

In [present-day] comedy ⁸ this [usin...