Poetics

Aristotle · a new plain-English translation from the original language

1| [1447a] Let us discuss the art of poetry itself and its species, the capacity each has, and how plots must be constructed if the poetry is to turn out well; further, of how many and what kind of parts it consists; and likewise about the other matters that belong to the same inquiry — beginning, as is natural, from first things first. Epic poetry, then, and the poetry of tragedy, and further comedy and dithyrambic poetry, and most flute-playing and lyre-playing, all happen to be, taken as a whole, imitations. But they differ from one another in three respects: either in imitating by different means, or in imitating different objects, or in imitating in a different manner and not in the same way. For just as some people, imitating many things, represent them by colors and shapes (some through craft, others through habit), while others do so through the voice, so too in the arts mentioned all of them make their imitation in rhythm and speech and harmony, but using these either separately or mixed together. For example, flute-playing and lyre-playing use only harmony and rhythm, and so do any other arts of that kind of capacity, such as that of the pipes; while dancers imitate by rhythm itself without harmony — for they too, through rhythms given shape, imitate characters, emotions, and actions.

2| But [epic poetry], which uses bare speech alone, or the art that uses [1447b] meters — whether mixing them with one another or using a single kind of meter — has, up to now, no name. Indeed we could give no single name shared by Sophron's and Xenarchus's mimes and by the Socratic dialogues, nor one covering an imitation carried out in trimeters, elegiacs, or any other such meter. Yet people, by attaching "making" to the meter, call some "elegy-makers" and others "epic-makers," naming them not by their imitation but indiscriminately by their meter — for even if they produce something medical or scientific in meter, people are accustomed to call them so. But there is nothing common to Homer and Empedocles except the meter; hence it is right to call the one a poet, but the other a natural scientist rather than a poet. Likewise, even if someone were to make an imitation mixing all the meters together, as Chaeremon did in composing the Centaur, a rhapsody mixed from all the meters, he too must be called a poet. On these matters, then, let this distinction stand. There are, further, some arts that use all the means mentioned — I mean rhythm, melody, and meter — as does the poetry of dithyrambs and of nomes, and both tragedy and comedy.

3| These differ in that some use all the means at once, others part by part. These, then, are the differences of the arts I speak of, in the means by which they make their imitation. [1448a] Since those who imitate imitate people acting, and these must be either good or base — for characters nearly always answer to these two alone, since all people differ in character by vice and virtue — they must imitate people either better than we are, or worse, or of the same sort, as painters do: Polygnotus painted people better than they are, Pauson worse, and Dionysius painted them as they are. It is clear that each of the imitations mentioned will admit these differences, and that each will be distinct from the others by imitating different things in this way. For in dancing and flute-playing and lyre-playing these dissimilarities can arise, and likewise in prose speech and in bare-meter poetry: Homer makes people better, Cleophon makes them as they are, and Hegemon of Thasos, who first composed parodies, and Nicochares, who composed the Deiliad, make them worse. Similarly too with dithyrambs and with nomes, just as one might imitate the Cyclopes as Timotheus and Philoxenus did [corrupt text]. In this very difference tragedy stands apart from comedy, for comedy wishes to imitate people worse, and tragedy better, than those of the present day. And there is yet a third difference among these: the manner in which one might imitate each of these things.

4| For it is possible, using the same means and imitating the same objects, to imitate at one time by narrating — either becoming someone else, as Homer does, or remaining the same person and not changing — or by presenting all the persons imitated as acting and doing something themselves [corrupt text]. It is in these three differences, then, that imitation consists, as we said at the outset: in the means, the objects, and the manner. So that in one respect Sophocles would be the same kind of imitator as Homer, since both imitate good people, while in another respect he would be the same as Aristophanes, since both imitate people acting and doing things. This, some say, is why such works are called dramas, because they imitate people doing things. For this reason too the Dorians lay claim to both tragedy and comedy — the Megarians, both those here in Attica, on the ground that comedy arose among them when they had democracy, and those from Sicily, since the poet Epicharmus was from there and lived much earlier than Chionides and Magnes; and some of those in the Peloponnese lay claim to tragedy — making the names their evidence. They say that they themselves call the outlying villages komai, while the Athenians call them demoi, on the ground that comedians (komoidoi) are named not from komazein, "to revel," but from their wandering through the villages (kata komas) when they were driven out in dishonor from the city.

5| [1448b] And they say that they themselves call "to make" [a thing happen] dran, while the Athenians call it prattein. Concerning the differences of imitation, then — how many they are and what they are — let this account stand. As for what generally gave birth to the art of poetry, there seem to have been two causes, and these natural ones. For imitating is innate in human beings from childhood, and in this they differ from the other animals, that they are the most imitative of creatures and make their first steps in learning through imitation; and also innate is the delight everyone takes in works of imitation. A sign of this is what happens in fact: things we look upon with pain when we see them in reality, we take delight in contemplating when they are rendered in the most exact likenesses — for example, the forms of the most despised animals, and of corpses. The reason for this too is that learning is most pleasant, not only to philosophers but likewise to everyone else, though the others share in it only briefly. For this is why people delight in seeing images — because it turns out, as they contemplate them, that they are learning and inferring what each thing is, as for instance that "this is that"; since if one happens not to have seen the thing before, the image will not produce pleasure as an imitation but on account of its execution, or its color, or some other such cause. Since imitation is natural to us, and so too are harmony and rhythm — for meters are evidently parts of rhythms — those who from the beginning were most naturally suited to these, advancing little by little, gave birth to poetry out of improvisations.

6| Poetry then split apart according to the individual characters of its makers: the more dignified poets imitated fine actions and the actions of people of that sort, while the more common poets imitated the actions of base people, at first composing invectives, just as the others composed hymns and encomia. Now of the poets before Homer we cannot name any such poem, though it is likely there were many; but starting from Homer there are examples, such as his own Margites and works of that kind. In these works the iambic meter came in, as fitting — hence it is even now called iambic, because in this meter people lampooned (iambized) one another. And among the ancients some became poets of heroic verse, others poets of iambic verse. Just as Homer was above all the poet of serious things — for he alone did not merely compose well, but also made dramatic imitations — so too he was the first to display the shape of comedy, by dramatizing the ridiculous rather than invective. For the Margites bears the same relation to comedies as the Iliad and the Odyssey bear to tragedies,

7| [1449a] so too this poem stands in relation to comedies. Once tragedy and comedy had come to light, those who turned to each kind of poetry according to their own natural bent became, some of them, writers of comedy instead of iambic verse, and others, instead of epic, became producers of tragedies, because these forms were grander and held in higher esteem than those. Now to examine whether tragedy is by now adequately developed in its formal elements or not, and to judge this both in itself and in relation to the theaters, is another discussion. At all events, arising from an improvisatory beginning — both tragedy itself and comedy, the one from those who led off the dithyramb, the other from those who led the phallic songs which even now still remain customary in many of the cities — it grew little by little as the poets advanced whatever part of it became evident; and tragedy, after passing through many changes, came to a stop, since it had attained its own nature. And it was Aeschylus who first brought the number of actors from one to two, and who diminished the role of the chorus and prepared for speech to take the leading part; three actors and scene-painting were Sophocles' doing. Further, there is its scale: it was late in acquiring dignity out of small plots and comic diction, through its change from the satyric, and its meter changed from the tetrameter to the iambic.

8| For at first they used the tetrameter because the poetry was satyric and more suited to dance, but once speech had come in, nature herself found the fitting meter; for the iambic is the most speech-like of meters. A sign of this: we speak very many iambics in conversation with one another, but hexameters rarely, and only when we depart from the rhythm of speech. Further, there is the number of episodes. Let what has been said stand for us regarding these and the other elements, as each is said to have been given its ornament; for to go through each of them one by one would perhaps be a great task. Comedy is, as we said, an imitation of people who are worse, yet not in the sense of the whole of vice, but the laughable is a part of the shameful. For the laughable is a certain error and ugliness that is painless and not destructive, just as, to take an obvious case, the comic mask is something ugly and distorted but without pain. Now the changes tragedy went through, and those responsible for them, have not gone unnoticed, but comedy escaped notice because it was not treated seriously from the start.

9| [1449b] For it was only late that the archon granted a comic chorus; before that the performers were volunteers. And it was only once comedy already had certain formal features that those said to be its poets are remembered. Who supplied it with masks, or prologues, or a plurality of actors, and such things, is unknown. The making of plots [Epicharmus and Phormis] came originally from Sicily, but of the Athenians, Crates was the first to begin, abandoning the iambic form, to compose speeches and plots of a general kind. Now epic poetry followed tragedy up to the point of being an imitation, in verse, of people of high standing; but they differ in that epic has a single meter and is narrative; and further, in length: tragedy tries so far as possible to keep within a single revolution of the sun, or to vary from it a little, while epic is unlimited in time — and in this too they differ, although at first they did this equally in tragedies as in epics. The parts are in some cases the same, in others peculiar to tragedy; hence whoever knows about tragedy, good and bad, knows about epic too; for whatever epic has, tragedy possesses, but not everything tragedy has is found in epic. We shall speak later about the art that imitates in hexameters, and about comedy.

10| Let us now speak about tragedy, taking up the definition of its substance that arises from what has been said. Taking this as our starting point, then: tragedy is the representation of a grave and finished action possessing magnitude, cast in speech that has been rendered pleasurable, each element used separately according to its part of the work, carried out by agents performing rather than by report, and bringing about, by means of pity and fear, the purging of such affections. By language made pleasurable I mean language that has rhythm and harmony [and melody]; and by "separately in the various kinds" I mean that some parts are carried out through meter alone, and again others through melody. Since it is by people acting that they make the imitation, in the first place the arrangement of the spectacle would necessarily be some part of tragedy; then the making of melody and diction, for it is in these that they make the imitation. By diction I mean the actual composition of the meters, and by melody-making I mean that whose power is entirely evident. Since it is an imitation of an action, and the action is carried out by certain agents who act, and these must necessarily be of a certain quality both in character and in thought (for it is on account of these that we also say the actions are of a certain quality,

11| [1450a] [there are by nature two causes of actions, thought and character], and it is in accordance with these that everyone either succeeds or fails), the action itself is represented through the plot — and by "plot" here I mean the arrangement of the incidents — whereas the characters are what makes us call the agents such-and-such in quality, and thought appears whenever, through speech, they show something or state a judgment. Necessarily, then, there are six parts of every tragedy, in accordance with which tragedy is of a certain quality: these are plot, characters, diction, thought, spectacle, and melody-making. For two of the parts are that by which they imitate, one is the manner in which they imitate, and three are what they imitate, and beyond these there is nothing. Not a few of them, so to speak, have used these elements; for indeed every play has spectacle and character and plot and diction and melody and thought likewise. Chief among these is how the incidents are put together. For tragedy represents actions and a life, not persons as such, [and happiness and unhappiness lie in action, and the end is a certain kind of action, not a quality; and people are of a certain quality in accordance with their characters, but happy or the opposite in accordance with their actions].

12| So it is not to represent character that agents act, but they acquire character for the sake of the actions; hence the events and their arrangement form tragedy's goal, and that goal outweighs everything else. Further, without action there could be no tragedy, but without characters there could be; for the tragedies of most of the moderns are lacking in character, and in general there are many poets of this kind, just as, among painters, Zeuxis stands in relation to Polygnotus; for Polygnotus is a good painter of character, while Zeuxis' painting has no character in it. Again, if someone sets in a row speeches expressive of character, and well made in diction and thought, he will not achieve what was said to be the function of tragedy; a tragedy that uses these things less well but has a plot and a structure of incidents will do so far more. Besides this, the greatest things by which tragedy sways the soul are parts of the plot — reversals and recognitions. A further sign: those who attempt to compose are able to be exact in diction and characters before they can construct the incidents, as is true of practically all the first poets. The plot, then, is the origin and, as it were, the soul of tragedy, and second come the characters (for it is much the same as in painting too;

13| [1450b] for if someone were to smear on the most beautiful colors at random, he would not give as much pleasure as if he had drawn an outline in white). And tragedy is an imitation of an action, and on that account chiefly of the people who act. Third is thought: this is the ability to say what is in a situation and what is fitting, which in the case of speeches is the function of political and rhetorical skill; for the poets of old made their characters speak in a political manner, but those of today speak in a rhetorical manner. Character is that sort of thing which reveals the choice [in cases where it is not clear what someone chooses or avoids]; hence speeches in which there is nothing at all that the speaker chooses or avoids have no character. Thought is present in those speeches where they demonstrate that something is or is not the case, or declare some general opinion. Fourth of the elements of speech is diction; by this, as was said before, I mean the expression of meaning through words, which has the same force whether in verse or in prose. Of the remaining elements, melody-making is the greatest of the pleasurable additions, while spectacle is a thing that sways the soul but is the least artistic and least proper to the poetic craft; for the power of tragedy exists even without performance and actors, and moreover the craft of the property-maker has more authority over the production of spectacle than does that of the poets.

14| Now that these matters have been distinguished, let us next discuss what kind of arrangement the events ought to have, since this stands as the first and greatest matter of tragedy. We have established that tragedy represents an action that is finished and entire and carries a certain size — for a thing may be entire while having no size at all. An entire thing is one possessing a start, a mid-point, and a close. A start is that which does not necessarily come after something else, though something else naturally exists or arises after it. A close, conversely, is that which by nature comes after something else, either necessarily or in most cases, while nothing further comes after it. A mid-point is that which comes after something else and is itself followed by yet another thing. Well-built plots, then, must neither start nor close at some arbitrary point, but must follow the patterns just given. Further, since anything beautiful — whether an animal or anything else composed of parts — must not only have these parts arranged in order but must also have a magnitude that is not random (for beauty consists in magnitude and order), it follows that a very tiny animal could not become beautiful (for the perception of it grows blurred as it approaches the span of an imperceptible instant of time), nor could a very huge one

15| [1451a] (for then the perception does not occur all at once, but the unity and wholeness of the object escapes those who are perceiving it), as would happen if an animal were ten thousand stades long. So just as with bodies and with animals a certain magnitude is needed, and this magnitude must be one that can be taken in at a single view, so too with plots: a certain length is needed, and this length must be one that can be easily held in memory. As for the limit of length that concerns dramatic contests and the audience's perception, that is not a matter belonging to the art itself — for if a hundred tragedies had to compete, they would be timed by water-clocks, as they say was once done on other occasions too. But the limit set by the nature of the subject matter itself is this: the greater the magnitude, up to the point where the whole remains clearly graspable, the finer the plot is in respect of magnitude. To state the matter simply: whatever magnitude, in a sequence of events occurring in accordance with probability or necessity, permits a change from bad fortune to good fortune or from good fortune to bad, is an adequate limit of magnitude. A plot is not one, as some suppose, simply because it is about one person — for many things, indeed countless things, happen to one person, and some of these do not combine to form any unity; and likewise there are many actions of one person from which no single action results.

16| For this reason all the poets who have written a Heracleid, a Theseid, and poems of that kind seem to be in error: they suppose that because Heracles was one person, the plot too ought properly to be one. But Homer, just as he surpasses others in every other respect, seems to have seen this point well too, whether by art or by nature: when shaping the Odyssey he left out the greater part of what actually befell Odysseus — his wound on Parnassus, say, or his pretended madness at the muster of the fleet — since neither event, once the one had happened, made the other either necessary or likely; rather he built the Odyssey, and the Iliad no less, around a single action of the sort we have described. Just as in the other imitative arts a single imitation is an imitation of one thing, so too the plot, since it is an imitation of an action, must be an imitation of one action and of that action as a whole, and the parts of the events must be so structured that if any part is shifted or removed, the whole is disturbed and thrown out of joint; for that whose presence or absence makes no perceptible difference is not a part of the whole at all. What we have said also makes plain that a poet's task is not reporting events that actually occurred, but rather the sort of events that could occur — what probability or necessity makes possible.

17| For the historian and the poet do not differ by speaking in verse or without verse [1451b] (Herodotus's work could be put into verse, and it would be no less a kind of history with meter than it is without meter); rather they differ in this: the one relates what has happened, the other the kind of thing that might happen. For this reason poetry is a more philosophical and more serious thing than history; for poetry speaks more of universals, history of particulars. A universal statement is one about the kind of thing a certain kind of person will, in accordance with probability or necessity, say or do — this is what poetry aims at, even while attaching names to the characters; a particular statement is one about what Alcibiades did or what happened to him. In the case of comedy this has by now become clear: comic poets construct the plot on the basis of what is probable, and only then assign chance names to the characters, rather than proceeding, as the writers of iambic verse do, about a particular individual. But in tragedy poets hold fast to the names of people who actually existed. The reason is that what is possible is persuasive: we do not yet believe that things which have not happened are possible, whereas it is clear that things which have happened are possible — for they would not have happened if they were impossible. Nevertheless, even in tragedies, in some plays only one or two of the names are well known and the rest are invented, while in others none is well known — as, for example, in Agathon's Antheus,

18| for in this play both the events and the names alike have been invented, and it delights no less for that. So one need not at all costs seek to hold fast to the traditional stories about which tragedies are composed. Indeed it is absurd to seek this, since even the well-known stories are well known only to a few, yet they delight everyone all the same. It is clear from this, then, that the poet must be a maker of plots rather than of verses, inasmuch as he is a poet in virtue of imitation, and what he imitates is actions. And even if it turns out that he composes about things that actually happened, he is no less a poet for that; for nothing prevents some of the things that have happened from being the sort of thing that would probably [and possibly] happen, and it is in respect of this that he is their poet. Among simple plots and actions, episodic ones are the worst. I call a plot episodic in which the episodes follow one another without either probability or necessity. Plots of this kind are made by bad poets because of their own deficiency, and by good poets for the sake of the actors: composing show-pieces for competition and stretching the plot beyond its natural capacity, they are often forced to distort the sequence of events.

19| [1452a] Since the imitation is not only of a complete action but also of events that arouse fear and pity, and these come about, and come about most of all [and more], when they occur contrary to expectation yet on account of one another — for in that way the marvelous element will be present more than if the events occurred spontaneously or by chance, since even among chance events those seem most marvelous which appear to have happened as if on purpose, as, for instance, when the statue of Mitys at Argos killed the man responsible for Mitys's death by falling on him as he stood watching a spectacle; for such things seem not to happen at random — so that plots of this sort must necessarily be finer. Among plots, some are simple and some are complex, since the actions of which plots are imitations are themselves already of these two kinds from the start. I call an action simple when, occurring as it has been defined, continuous and unified, the change of fortune comes about without reversal or recognition; and I call it complex when the change of fortune comes about with recognition or reversal or both. These must arise out of the very structure of the plot itself, so that they follow from what has gone before either by necessity or in accordance with probability; for it makes a great difference whether these events happen because of those or merely after them.

20| Reversal is a change of the actions to their opposite, as has been said, and this too, as we say, in accordance with probability or necessity. Take the Oedipus: a messenger arrives meaning to lift Oedipus's spirits and clear away his dread concerning his mother, yet in disclosing his identity he brought about the reverse; and in the Lynceus, one man is being led off to die, and Danaus is following to kill him, but as a result of what has happened one of them dies while the other is saved. Recognition, as the very name signifies, is a change from ignorance to knowledge, disclosing either a bond of friendship or one of enmity, on the part of people marked out for good or bad fortune. Recognition is finest when it occurs at the same time as reversal, as with the one in the Oedipus. There are also other kinds of recognition, for it can occur even in relation to lifeless things and to chance occurrences as has been said it happens, and it is possible to recognize whether someone has or has not done something. But the recognition most proper to the plot and most proper to the action is the kind just described; for a recognition of this sort, joined to reversal, will yield either pity or fear

21| [1452b] (and it is actions of this sort that tragedy has been posited as an imitation of), since both misfortune and good fortune will follow upon such things. Now since recognition is a recognition of certain people by certain people, in some cases it is recognition of one person by the other only, when it is already clear who the other one is, but at other times both parties must recognize each other, as when Iphigenia was recognized by Orestes through her sending of the letter, while a separate recognition was needed to make him known to Iphigenia. These, then, are two parts of the plot, reversal and recognition; and a third is suffering. Of these, reversal and recognition have been discussed; suffering, by contrast, is an act that destroys or causes pain — visible deaths, agonies, woundings, and the like. The parts of tragedy that must be treated as forms we stated earlier, but the parts into which it is divided by quantity, taken separately, are these: prologue, episode, exodos, and choral song, this last divided into parodos and stasimon. These are common to all tragedies, while songs from the stage and laments are proper to some. The prologue is the whole part of a tragedy before the parodos of the chorus; the episode is the whole part of a tragedy between complete choral songs; the exodos is the whole part of a tragedy after which there is no song of the chorus;

22| and of the choral part, the parodos is the whole first utterance of the chorus, the stasimon is a song of the chorus without anapaest or trochee, and the kommos is a lament shared between the chorus and the actors on stage. The parts of tragedy that must be treated as forms we stated earlier, and the parts into which it is divided by quantity, taken separately, are these. Next in order after what has now been said, it would be proper to state what one ought to aim at and what one ought to guard against in constructing plots, and what it is that will produce the effect proper to tragedy. Since, then, the structure of the finest tragedy must be not simple but complex, and must be an imitation of things that evoke fear and pity (for this is distinctive of this kind of imitation), it is clear, first, that decent men should not be shown passing from good fortune to bad fortune, for this is not fearful or pitiable but repugnant; nor should base men be shown passing from bad fortune to good fortune, for this is the most untragic thing of all, since it has none of the qualities it should have — it is neither humane, nor pitiable, nor fearful;

23| [1453a] nor again should an utterly wicked man fall from good fortune into bad fortune. Such a structure might have the humane quality, but it would produce neither pity nor fear, since pity is felt about a man who suffers undeservedly, and fear about a man like ourselves — pity about the undeserving, fear about the man like ourselves — so that what happens will excite neither pity nor fear. What is left, then, is the man in between these. Such is the man who is not outstanding in virtue and justice, and who does not pass into misfortune through vice and wickedness, but through some error, being one of those who are held in great esteem and enjoy good fortune, such as Oedipus and Thyestes and the notable men of families of that kind. It follows, then, that the well-constructed plot must be single rather than double, as some say, and must involve a change not from bad fortune to good fortune but, on the contrary, from good fortune to bad fortune, and this not on account of wickedness but on account of some great error on the part of a man such as has been described, or a better man rather than a worse. And what actually happens confirms this: at first the poets recounted whatever plots came their way, but nowadays the finest tragedies are composed about a few houses — about Alcmaeon, for instance, and Oedipus and Orestes and Meleager and Thyestes and Telephus, and all the others to whom it has befallen to suffer or to do terrible things.

24| The tragedy that is finest according to the art, then, is built on this structure. Hence those who criticize Euripides for doing this in his tragedies, and for having many of them end in misfortune, are making the same mistake. For this is, as has been said, correct; and the strongest sign of it is this: on the stage and in the dramatic competitions, such plays appear the most tragic of all, if they are brought off successfully, and Euripides, even if he does not manage everything else well, still appears the most tragic of the poets. Second in rank is the structure that some call first, the one that has a double structure, like the Odyssey, and ends in opposite ways for the better and the worse characters. It is thought to be first because of the weakness of audiences, for the poets follow along, composing to gratify the spectators' wishes. But this is not the pleasure that comes from tragedy; it belongs rather to comedy, where those who are the bitterest enemies in the story, such as Orestes and Aegisthus, walk off as friends at the end, and no one is killed by anyone.

25| [1453b] Now the fearful and the pitiable can arise from the spectacle, but they can also arise from the very structure of the events, which is the prior and better way, and the mark of a better poet. For the plot should be so constructed that, even without seeing it performed, the person who hears the events that occur shudders and feels pity at what happens, which is what one would feel on hearing the plot of the Oedipus. To produce this effect through spectacle is less a matter of art, and depends on lavish production. Those who use spectacle to produce not the fearful but merely the monstrous have nothing in common with tragedy, for one should not seek every kind of pleasure from tragedy, but only the pleasure proper to it. And since the poet must produce the pleasure that comes from pity and fear by means of imitation, it is clear that this effect must be built into the events themselves. Let us, then, take up what sorts of events seem terrible and what sorts seem pitiable among those that occur. Necessarily, such actions occur either between people who are friends toward one another, or between enemies, or between people who are neither. Now if an enemy does it to an enemy, there is nothing pitiable either in the doing or in the intending, except so far as the suffering itself goes; nor is there when the people involved are neither friends nor enemies; but when the suffering occurs within relationships of friendship — when, for example, a brother kills a brother, or a son a father, or a mother a son, or a son a mother, or is about to, or does something else of that kind — these are the cases to be sought after.

26| Now the plots that have been handed down one must not break up — I mean, for example, the death of Clytemnestra at the hands of Orestes, and that of Eriphyle at the hands of Alcmaeon — but the poet must himself invent, and use the traditional stories well. Let us say more clearly what we mean by "well." It is possible for the action to take place as the poets of old made it happen, with the people involved acting knowingly and with awareness, as Euripides too made Medea kill her children. Or it is possible for them to act, but to act in ignorance of the terrible thing they are doing, and afterward to recognize the bond of friendship, as with the Oedipus of Sophocles; that case falls outside the drama itself, whereas within the tragedy proper we find something like Astydamas's Alcmaeon, or the Telegonus of the Wounded Odysseus. There is, besides these, a third possibility: someone about to do one of the irreparable deeds in ignorance recognizes the truth before doing it. And beyond these there is no other way; for it is necessary either to act or not to act, and either knowingly or unknowingly. Of these, to know and intend but not to act is the worst; for it has the repugnant quality, and it is not tragic, since it involves no suffering. That is why no one composes it in this way, except rarely,

27| [1454a] if not rarely — as with Haemon toward Creon in the Antigone. Acting is second-best. Better than that is to act in ignorance and then, after acting, to recognize: for the shocking element is absent, and the recognition is startling. Best of all is the last kind — I mean, for example, in the Cresphontes, where Merope is about to kill her son but does not kill him, but instead recognizes him; and in the Iphigenia, where the sister recognizes the brother; and in the Helle, where the son, about to hand his mother over, recognized her. This is why, as was said earlier, tragedies are about few families. For in searching, poets found how to produce this sort of thing in their plots not by craft but by chance, and so they are forced to fall back on those households in which such sufferings have occurred. Concerning the construction of the incidents, then, and what qualities the plots should have, enough has been said. As for character, there are four things one must aim at. One, and first, is that it be good. A speech or action will have character, as was said, if it makes clear a choice of some kind, whatever that choice may be, and the character will be good if the choice is good. This holds within each class of person:

28| for even a woman can be good, and a slave too, although one of these is perhaps the inferior kind, the other altogether base. Second is the fitting: for there is such a thing as a manly character, but it is not fitting for a woman to be manly or clever in that way. Third is likeness. This is different from making the character good and fitting, as said before. Fourth is consistency. For even if the person offering the imitation is himself inconsistent, and such a character is assumed for him, he must nevertheless be consistently inconsistent. A model of character that is bad without need is Menelaus in the Orestes; of the improper and unfitting, Odysseus's wailing in the Scylla and Melanippe's speech; of the inconsistent, the Iphigenia at Aulis, since the girl begging for her own life resembles not at all the woman she later becomes. In matters of character, just as in the construction of the incidents, one must always seek either the necessary or the probable, so that for such a person to say or do such things is necessary or probable, and for this to follow after that is necessary or probable. It is clear, then, that the resolutions of plots too must come about from the plot itself,

29| [1454b] and not, as in the Medea, from a contrivance, or as with the matters of the departure in the Iliad. A contrivance should be used for what lies outside the drama — either events that happened earlier, which a human being could not know, or events that come later, which need to be foretold or reported; for we grant the gods the power to see everything. But there should be nothing irrational within the events themselves; if there is, it should be kept outside the tragedy, as with the irrational element in Sophocles' Oedipus. Since tragedy is an imitation of people better than we are, one should imitate good portrait-painters: for they too, in rendering a person's own form, make a likeness and yet paint it handsomer. In the same way the poet, in imitating people who are irritable or easygoing or have other such traits of character, should, while making them such, make them decent as well. (The text is corrupt at this point in the manuscripts; the sense appears to be something like: take a harsh character, and, as Homer does with Achilles, make him good.) These points, then, must be watched, and besides them, the points bound up with the perceptions that necessarily follow on the poetic craft; for one can err often here too — enough has been said about these matters in the published treatises. What recognition is has been stated before. As for the kinds of recognition, the first is the least artful, and the one used most often for lack of resource — recognition by signs.

30| Of these signs, some are innate — for example, "the spear the earthborn bear," or the stars such as Carcinus gives in the Thyestes — and some are acquired; and of these acquired ones, some are on the body, such as scars, and some external, such as necklaces, or, as in the Tyro, the chest. Even these can be used either better or worse: for example, Odysseus was recognized by his scar in one way by the nurse and in another way by the swineherds. For recognitions used only to establish belief are less artful, and all of that kind are so, while those that arise out of a reversal, as in the Washing Scene, are better. Second are those contrived by the poet, and for that reason inartful. For example, Orestes in the Iphigenia recognized that he was Orestes: she made herself known through the letter, but he himself says what the poet wants said, not what the plot requires. This is why it comes close to the fault mentioned before, since he might just as well have been carrying tokens. And in Sophocles' Tereus there is the voice of the shuttle. The third kind is through memory, by perceiving something on seeing it,

31| [1455a] as in the Cyprians of Dicaeogenes — for on seeing the painting he wept — and as in the tale told to Alcinous, where, hearing the lyre-player and being reminded, he shed tears; and from this they were recognized. The fourth kind comes from reasoning — for example in the Libation-Bearers: someone resembling him has come; no one resembles him except Orestes; therefore Orestes has come. There is also the one devised by Polyidus the sophist concerning the Iphigenia: for he said it was likely that Orestes reasoned that, since his sister had been sacrificed, it fell to him too to be sacrificed. And in Theodectes' Tydeus: that he came expecting to find his son, and instead perishes himself. And the one in the Phineidae: for the women, on seeing the place, reasoned out their fate, that it had been destined for them to die there, since they had also been exposed there. There is also a composite kind that arises from a false inference on the audience's part, as in the Odysseus the False Messenger: that he alone could draw the bow and no one else was a device the poet invented as a starting assumption, as was the claim that he would know the bow on sight, though he had never laid eyes on it; but to construct the scene so that he is recognized by means of that same bow is a false inference drawn by the spectators. Best of all recognitions is the one that arises out of the events themselves, the astonishment coming about through probabilities, as in Sophocles' Oedipus and in the Iphigenia:

32| for it is probable that she would want to have a written message delivered. Recognitions of this kind alone dispense with contrived signs and necklaces. Second-best are those from reasoning. Plots must be constructed, and worked out in diction, by setting the action as far as possible before one's eyes: for in this way, seeing it most vividly, as if present among the very events taking place, [the poet] would discover what is fitting and would be least likely to miss [the] contradictions. A sign of this is the fault found with Carcinus: his Amphiaraus was coming up from the temple, a fact that escaped [the spectator] because he did not see it, but on the stage it fell flat, since the spectators were offended by it. So far as possible, one should also work the action out with the very gestures: for those actually caught up in the emotions are the most convincing, by the very nature of the case, and the person truly storm-tossed storms and the person truly angered rages most truly. This is why poetry belongs to a person of natural gift, or to one touched with madness: for of these, the one kind are easily molded, the other beside themselves. As for the stories, both the traditional ones and those he makes up himself, the poet must first set them out in their universal form,

33| [1455b] and only then, in that form, fill them out with episodes and stretch them to length. What I mean by viewing the universal outline is this — take the Iphigenia, for example: a certain girl was sacrificed and vanished without the knowledge of those who sacrificed her, and was established in another land, where the custom was to sacrifice strangers to the goddess, and she held this priesthood; some time later it happened that the priestess's brother came there — that the god ordained his coming there [for some reason outside the universal outline], and for what purpose, lies outside the plot; having come, and having been seized, and being about to be sacrificed, he was recognized, whether the poet made it happen as Euripides did or as Polyidus did, by saying, in accordance with probability, that it was fated for him too, not only for his sister, to be sacrificed — and from this comes his deliverance. After this, once names have been assigned, one adds the episodes; and one must see to it that the episodes belong to the story, as in the Orestes, the madness through which he was seized, and the deliverance through the purification. In dramas, then, the episodes are compressed, but epic poetry is drawn out by means of them. For the plot of the Odyssey is not long: a man is away from home for many years, watched closely by Poseidon, and alone; and meanwhile things at home stand so that his wealth is being consumed by suitors and his son is being plotted against; he himself arrives, storm-battered, and after making himself known to some people, attacks, and is himself saved while destroying his enemies.

34| This, then, is what is proper to the plot; everything else is episode. Every tragedy has a complication and a resolution: the events outside the play, and often some of those within it, form the complication, and the rest is the resolution. By complication I mean the part from the beginning up to the last point from which the change to good fortune or bad fortune occurs; and by resolution, the part from the beginning of that change to the end. Take Theodectes' play Lynceus: its complication is built from what happened beforehand, the taking of the infant, and then, in turn, the abduction of the pair themselves; its resolution stretches from the charge of killing all the way through to the close. Tragedy divides into four types (matching the number of parts already named): the complex kind, whose whole substance is reversal and recognition; the pathetic kind, such as the Ajaxes and the Ixions; the kind of character, such as the Women of Phthia and the Peleus;

35| [1456a] and the fourth kind (the text is corrupt at this point), such as the Phorcides, the Prometheus, and all the plays set in Hades. One should try, above all, to have all of these elements, or failing that, the greatest and the most of them, especially given how people now unfairly attack poets: since good poets have arisen in each particular part, critics demand that a single poet surpass what was excellent in each one taken separately. It is right to call one tragedy different from another, or the same as another, by nothing so much as by its plot — that is, by tragedies that have the same complication and resolution. Many poets tie the knot well but resolve it badly; both need to be brought off skillfully. One must remember, as has often been said, and not turn a tragedy into an epic structure — by epic I mean a structure with many stories — as if someone were to make the whole story of the Iliad into a single tragedy. For there, because of its length, the parts take on their proper scale, but in plays the result falls far short of what one would expect. The proof is this: those who have dramatized the whole sack of Troy, rather than a part of it as Euripides did, or the whole story of Niobe, rather than as Aeschylus did, either fail outright or fare badly in competition — since even Agathon failed on this point alone. But in reversals, and in simple actions, poets hit wonderfully well what they aim at;

36| for this is both tragic and appeals to our human feeling. This happens whenever a man who is clever but wicked is deceived, as Sisyphus is, or when a man who is brave but unjust is defeated. And this is also probable, as Agathon says, for it is probable that many things happen even against probability. One should also regard the chorus as one of the actors, and as a part of the whole, taking part in the action — not as in Euripides, but as in Sophocles. In the rest of the poets, the songs belong no more to the plot than they would to any other tragedy; that is why they now sing interludes, a practice Agathon was the first to begin. Yet what difference is there between singing interludes and fitting a speech, or even a whole episode, from one play into another? That accounts for the remaining elements; what is left to discuss is verbal expression and thought. Let what concerns thought be set down in the works on rhetoric, since that subject belongs more properly to that discipline. Thought covers everything that has to be produced by means of speech; its parts are demonstrating, refuting, and arousing emotions

37| [1456b] (examples include pity, fear, and anger, among similar feelings), together with amplification and diminishment. It is clear that in the actions too one must draw on the same forms whenever it is necessary to produce effects that are pitiable, terrible, great, or probable; except that they differ in this respect — that the effects in the action must appear without any explanation, while those produced through speech must be brought about by the speaker and arise out of the speech itself. For what would be the speaker's function, if the effect appeared just as it should without being due to the speech? Among the matters concerning diction, one line of inquiry concerns the forms speech can take, and knowing them is the business of the actor's craft and of anyone who has command over it — for example, knowing what a command is, what a prayer is, and what narration, threat, question, and answer are, and anything else of this kind. For knowledge or ignorance of these brings against the poetic art no criticism worth taking seriously. What fault could anyone find in what Protagoras criticizes — that Homer, thinking he is uttering a prayer, actually gives a command when he says, "Sing, goddess, of the wrath"? For to bid someone do something or not, Protagoras says, is a command. So let this be set aside as belonging to another discipline, not to poetics.

38| Of diction as a whole, these are the parts: letter, syllable, connective, noun, verb, article, inflection, and phrase. A letter is an indivisible sound — not just any such sound, but one out of which a composite sound is naturally formed; for animals too have indivisible sounds, none of which I call a letter. The parts of this class are the vowel, the semivowel, and the mute. A vowel is that which has an audible sound without any contact of the speech organs; a semivowel is that which has an audible sound with contact, such as S and R; a mute is that which, even with contact, has by itself no sound at all, but becomes audible only in combination with letters that do have some sound, such as G and D. These differ from one another in the shapes and places of contact in the mouth, in roughness and smoothness, in length and shortness, and further in pitch, whether high, low, or intermediate; the detailed study of each of these belongs properly to works on metrics. A syllable is a non-signifying composite sound made up of a mute and a letter that has sound; for GR without an A is not a syllable, while with an A it is, as in GRA — but to examine these differences too belongs to metrics.

39| A connective is a non-signifying sound [1457a] which neither prevents nor helps form a single signifying sound naturally composed of several sounds, and which is not fitting to place by itself at the beginning of a sentence, whether at the extremes or in the middle — for example, men, ētoi, de. Or alternatively, a connective is a non-signifying sound naturally capable of forming, out of several signifying sounds, a single signifying sound. An article is a non-signifying sound that marks the beginning, end, or boundary of a phrase, such as amphi and peri and the like. Or alternatively, an article is a non-signifying sound which neither prevents nor helps form a single signifying sound naturally composed of several sounds, and which can be placed both at the extremes and in the middle. A noun is a composite signifying sound without reference to time, no part of which is by itself signifying; for in compound words we do not treat each part as signifying on its own — in Theodorus, for instance, the -dorus does not signify anything by itself. A verb is a composite signifying sound with reference to time, no part of which signifies by itself, just as with nouns; for man or white does not signify when, whereas walks or has walked additionally signifies, the one present time, the other past time.

40| An inflection of a noun or a verb is one that indicates of this or to this and the like, or one that indicates singular or plural, such as men or man, or one that indicates the modes of delivery, such as question or command — for did he walk? or walk! is an inflection of the verb according to these forms. A phrase is a composite signifying sound, some of whose parts signify something by themselves (for not every phrase is composed of verbs and nouns — the definition of man, for example, is not — but a phrase can exist without verbs, though it will always have some part that signifies something), as in Cleon walks, where Cleon signifies something by itself. A phrase is one in two ways: either as signifying one single thing, or as composed of several parts joined by a connective — the Iliad, for instance, is one by connection, while the definition of man is one by signifying one thing. Nouns are of two kinds: simple, by which I mean one not composed of signifying parts, such as earth, and double. Of the double kind, one sort is composed of a signifying part and a non-signifying part, except that within the noun itself that part functions neither as signifying nor as non-signifying, while the other sort is composed of two signifying parts. There could also be a triple, quadruple, or manifold noun, such as many of the names used by the people of Massalia, like Hermocaicoxanthus.

41| . . [1457b] Every noun is either a standard term, or a dialect word, or a metaphor, or an ornament, or a coined word, or a lengthened word, or a shortened word, or an altered word. By "standard" I mean the word each group of people uses; by "dialect word" I mean one that other people use; so it is clear that the same word can be both a dialect word and a standard term, though not for the same people: for the word "sigynon" is a standard term for the Cyprians, but a dialect word for us. Metaphor is the carrying-over of a name that properly belongs elsewhere: the shift runs from a general class down into a particular kind, from a particular kind up into its general class, from one particular kind across into another, or along a proportion. By "from genus to species" I mean, for example, "here my ship stands": for lying at anchor is a kind of standing. By "from species to genus": "truly Odysseus has done ten thousand noble deeds": for "ten thousand" is a large number, and here he has used it in place of "many." By "from species to species," for example: "drawing off the life with the bronze" and "cutting with the long-edged bronze": here he has used "draw off" for "cut" and "cut" for "draw off," since both are a kind of taking away. I call it "by analogy" whenever the second term stands to the first as the fourth stands to the third:

42| for then the speaker will say the fourth in place of the second, or the second in place of the fourth. And sometimes people add, in place of the term replaced, the term to which it stands in relation. I mean, for example, that a cup stands to Dionysus as a shield stands to Ares; so one will call the cup "the shield of Dionysus" and the shield "the cup of Ares." Or again, as old age stands to life, so evening stands to day; so one will call evening "the old age of the day," or, as Empedocles does, call old age "the evening of life" or "the sunset of life." In some cases there is no established name for one of the terms in the analogy, but nonetheless the thing will be spoken of in the same way: for instance, releasing seed onto the ground is called "sowing," but releasing flame from the sun has no name; yet this stands to the sun just as sowing stands to the seed, and so it has been said, "sowing the god-created flame." And it is possible to use this manner of metaphor in another way too, by applying the foreign name and then denying of it one of its own proper attributes—for example, if one were to call the shield not "the cup of Ares" but "the wineless cup" . . . A coined word is one that, being called by no one at all, is set up by the poet himself; for some words do seem to be of this kind, such as calling horns "sprouts" and a priest a "prayer-man." A word is lengthened or shortened,

43| [1458a] the one when it makes use of a vowel longer than its own or of an inserted syllable, the other when something has been removed from it; lengthened, for example, "poleos" for "poleos" and "Peleidou" for "Peleiadeo"; shortened, for example, "kri" and "do," and "one voice results from both." A word is altered whenever, of the word as normally spoken, part is left as it is and part is made up, as in "the right one, on the breast" in place of "the right-hand one." Of the nouns themselves, some are masculine, some feminine, and some between the two: masculine are all that end in N and R and S, and those composed from this last (these are two, Psi and Xi); feminine are those ending in vowels that are always long, namely Eta and Omega, and, among the ones that admit lengthening, those ending in Alpha; so it turns out that the masculine and feminine endings are equal in number, since Psi and Xi are compounds. No noun ends in a mute consonant, nor in a short vowel. Only three end in Iota: honey, gum, pepper. Five end in Upsilon. The nouns between the two genders end in these same letters and in N and S. The excellence of diction is to be clear and not low.

44| The clearest diction, then, is the one made of standard words, but it is low; the poetry of Cleophon and that of Sthenelus are an example. Dignified diction, one that rises above the ordinary, is that which makes use of foreign elements; and by "foreign" I mean dialect word, metaphor, lengthening, and everything that departs from the standard. But if someone makes a whole composition of such things, the result will be either a riddle or barbarism: if it is made of metaphors, a riddle; if of dialect words, barbarism. For this is the very nature of a riddle, to state true things while combining them in an impossible way: this cannot be done by the ordinary combining of other kinds of words, but it can be done by means of metaphors, for example, "I saw a man gluing bronze onto another man with fire," and things of that sort. And compositions made of dialect words are barbarism. So these elements must somehow be blended together: for the dialect word, the metaphor, the ornament, and the other kinds mentioned will keep the diction from being ordinary and low, while the standard word will supply its clarity. Not the least part of what contributes

45| [1458b] to the clarity of diction, and to its not being ordinary, is the lengthening, curtailing, and altering of words: for because such a word is other than the standard form, departing from the customary usage, it will make the diction not ordinary, while because it still shares in the customary usage, it will remain clear. So those critics are not right who fault this manner of diction and mock the poet for it, as the elder Eucleides did, on the ground that it is easy to compose poetry if one is to be allowed to lengthen words as far as one pleases—composing, in that very style, iambic lines such as "I saw Epichares walking to Marathon," and "not (the text is corrupt here) his hellebore." Now to make one's use of this device in some way conspicuous is ridiculous; but due measure applies in common to all the parts of diction alike: for indeed someone using metaphors and dialect words and the other forms unsuitably, and on purpose for the sake of raising a laugh, would produce the very same effect. How great a difference fitness makes can be studied in epic verses by inserting the standard words in place of the others, within the same meter. And likewise with the dialect word, the metaphor, and the other forms: if one substitutes the standard words in their place, one can see that what we say is true; for example, the very same iambic line was composed by both Aeschylus and Euripides, and with only a single word changed—a dialect word put in place of the customary standard one—the one line appears beautiful and the other trivial.

46| For Aeschylus, in the Philoctetes, wrote: "the cancer that eats the flesh of my foot," but the other poet substituted "feasts on" for "eats." And: "but now, being small and worthless and unseemly," if someone were to substitute the standard words and say: "but now, being small and weak and ugly." And: "setting down an unseemly stool and a small table," becomes "setting down a wretched stool and a small table." And "the shores are roaring" becomes "the shores are shrieking." Further, Ariphrades used to mock the tragic poets because they use expressions that no one would use in ordinary speech—for instance, "from the halls" rather than "away from the halls," and "of thee," and "and I him," and "Achilles concerning" rather than "concerning Achilles,"

47| [1459a] and all the other things of that kind. For it is precisely because such expressions are not among the standard words that they all make the diction not ordinary; but Ariphrades did not understand this. Now it is a great thing to use each of the kinds mentioned fittingly—both compound words and dialect words—but by far the greatest thing is to have a gift for metaphor. This alone cannot be acquired from another, and it is a sign of natural talent: for to make good metaphors is to perceive what is similar in things. Of nouns, compound words suit dithyrambs best, dialect words suit heroic verse, and metaphors suit iambic verse. And in heroic verse all the kinds mentioned are useful; but in iambic verse, because it imitates speech as closely as possible, the words that are fitting are those that one might also use in ordinary prose speech, namely the standard word, the metaphor, and the ornament. Concerning tragedy, then, and imitation carried out in action, let what has been said be sufficient for us. As for the narrative kind, which imitates in meter, it is clear that its plots must be constructed dramatically, as in tragedies, and must be about a single action that is whole and complete and has a beginning, middle parts, and an end, so that like a single whole living creature it may produce the pleasure proper to it—and that and its compositions must not be like histories, which are bound to give an account not of one action but of one stretch of time -- everything that occurred within it concerning one person or many, each event related to the rest purely as chance would have it.fittingly—both compound words and dialect words—but by far the greatest thing is to have a gift for metaphor. This alone cannot be acquired from another, and it is a sign of natural talent: for to make good metaphors is to perceive what is similar in things. Of nouns, compound words suit dithyrambs best, dialect words suit heroic verse, and metaphors suit iambic verse. And in heroic verse all the kinds mentioned are useful; but in iambic verse, because it imitates speech as closely as possible, the words that are fitting are those that one might also use in ordinary prose speech, namely the standard word, the metaphor, and the ornament. Concerning tragedy, then, and imitation carried out in action, let what has been said be sufficient for us. As for the narrative kind, which imitates in meter, it is clear that its plots must be constructed dramatically, as in tragedies, and must be about a single action that is whole and complete and has a beginning, middle parts, and an end, so that like a single whole living creature it may produce the pleasure proper to it—and that its compositions must not be like histories, which are bound to give an account not of one action but of one stretch of time -- everything that occurred within it concerning one person or many, each event related to the rest purely as chance would have it.

48| For just as the sea battle at Salamis and the battle of the Carthaginians in Sicily happened at the same time without tending toward the same end at all, so too in successive stretches of time one thing sometimes happens after another, and no single end results from them. Most poets do this. That is why, as we have already said, Homer would appear divinely inspired beyond the others in this respect too: that he did not even attempt to make the whole war his subject, even though it has a beginning and an end—for the plot was bound to become too large and hard to take in at a single view, or else, kept moderate in size, entangled by its variety. As it is, he has taken up one part and used many of the rest as episodes—for instance the catalogue of ships—and diversifies the poem with other episodes [twice]. But the other poets compose about one person, and about one time, and one action with many parts,

49| [1459b] as, for instance, the man who wrote the Cypria and the Little Iliad. Hence out of the Iliad and the Odyssey only one tragedy is made from each, or two at most, while out of the Cypria many are made, and from the little Iliad [[more than] eight, for instance the Judgment of Arms, Philoctetes, Neoptolemus, Eurypylus, the Begging, the Spartan Women, the Sack of Troy, and the Sailing Away [and Sinon and the Trojan Women]]. Further, epic must have the same species as tragedy: it must be either simple or complex, or one of character or one of suffering. And its components, too, match those of tragedy save for song-composition and spectacle, since it likewise requires reversal, recognition, and suffering. Further, it must handle thought and diction well. Homer has used all of these, and was the first to do so, and did so adequately. For of his two poems, the Iliad is constructed as simple and one of suffering, while the Odyssey is complex (for there is recognition running throughout it) and one of character; and besides this he has surpassed all others in diction and in thought. Epic differs from tragedy in the length of its structure and in its meter. The limit of length already stated is sufficient: it must be possible to take in the beginning and the end together in a single view. This would be so if the structures were shorter than those of the old epics, and reached up to the length of the number of tragedies set for one hearing.

50| Epic has, for extending its magnitude, a great deal that is peculiar to it, because in tragedy it is not possible to represent many parts occurring at the same time, but only the part that is on the stage and involves the actors; whereas in epic, because it is narrative, it is possible to make many parts being completed at the same time, and these, being proper to the subject, increase the poem's bulk. So epic has this advantage toward grandeur, and toward shifting the listener's attention and diversifying the work with dissimilar episodes; for it is sameness that quickly fills one up and causes tragedies to fail. As for meter, experience has fitted the heroic meter to epic. For if someone were to compose a narrative imitation in some other single meter, or in several meters, it would appear unfitting; for the heroic is the most stately and weighty of meters (which is also why it admits foreign words and metaphors most readily, since narrative imitation too stands apart from the others), while the iambic and the trochaic tetrameter are meters of movement, the one suited to dancing, the other to action.

51| [1460a] Further, it is even more absurd if one mixes these meters together, as Chaeremon did. That is why no one has composed a long structure in any meter other than the heroic; rather, as we have said, nature itself teaches us to choose what is fitting to it. Homer deserves praise for many other things, and especially because he alone among the poets is not ignorant of what he himself ought to do. For the poet himself ought to speak as little as possible, since in speaking he is not being an imitator. The other poets compete in their own persons throughout, and imitate only a little and rarely; but Homer, after a brief prelude, at once brings in a man or a woman or some other character, and none of them without character, but each possessing one. Now the marvelous must be produced in tragedies, but epic allows more scope for the irrational, through which the marvelous is chiefly achieved, because one does not look at the person acting. For the events of the pursuit of Hector, if placed on the stage, would appear ridiculous—the Greeks standing still and not pursuing, and Achilles shaking his head to hold them back—but in the epic verses this passes unnoticed. And the marvelous is pleasant: a sign of this is that everyone, in reporting a story, adds something to it, since this pleases their hearers. It is Homer above all who has taught the rest how one ought to tell lies.

52| This is a false inference. For people suppose that, when this thing being so that thing is, or when this coming to be that comes to be, then if the later thing is true, the earlier thing too is true or has come to be—but this is false. Hence, if the first thing is false, but there is something else that necessarily is or comes to be if the first were true, one must add that: for because we know this second thing to be true, our mind falsely infers, on that basis, that the first thing is true as well. An example of this is the passage from the Washing scene. One ought to prefer probable impossibilities to improbable possibilities; and stories should not be composed of irrational parts—rather, so far as possible they should contain nothing irrational, or if they do, it should fall outside the plot itself, as with Oedipus not knowing how Laius died, and not within the drama, as with the messengers reporting the Pythian games in the Electra, or the mute character in the Mysians who comes from Tegea into Mysia. To say that otherwise the plot would have been destroyed is ridiculous, for one ought not to construct such plots from the beginning. But if one has set such a thing in place and it appears to admit of a more plausible account, it is nevertheless absurd since the irrational elements in the Odyssey concerning the putting ashore would clearly have been intolerable, had an inferior poet composed them;

53| [1460b] but as it is, the poet conceals the absurdity, sweetening it with his other virtues. In diction one must labor especially over the idle passages, those that convey neither character nor thought; for excessively brilliant diction, in turn, obscures the characters and the thoughts. Concerning problems and their solutions, one can see how many and what kinds there are if one considers the matter as follows. Since the poet is an imitator just like a painter or any other maker of images, he must always imitate one of three things in number: either how something was or is, or how people claim and believe it to be, or how it should be. These are conveyed in diction, in which there are foreign words and metaphors and many modifications of diction, for we grant these to poets. Besides this, correctness is not the same thing for political science and for poetics, nor for any other art and poetics. Within poetics itself there are two kinds of error: one belonging to it in its own right, the other incidental. For if the poet set out to represent something and failed through . . . incapacity, the error is his own; but if he chose incorrectly—representing, say, a horse throwing forward both right legs at once—or made an error belonging to some particular art, such as an error in medicine or in whatever other art [or has represented impossibilities], that is not an error in the poet's own right.

54| Hence one must resolve the objections raised in these problems by examining them under these headings. First, those bearing on the art itself: if impossibilities have been represented, this is an error—but it is correct if the poem thereby achieves its own end (for the end has been stated), that is, if this makes that very part, or some other part, more astounding. The pursuit of Hector is an example. If, however, the end could have been achieved either better, or no worse, while still observing the correctness proper to that art, then [the error stands] and is not right; for one ought, wherever possible, to make no error at all. Further, one must ask to which class the error belongs: one against the art itself, or an incidental one against some other point? For it is a lesser fault not to know that a female deer has no horns than to draw one unrecognizably. Besides this, if the objection is that something is not true, one may perhaps answer that it is as it ought to be—just as Sophocles said that he himself portrayed people as they ought to be, while Euripides portrayed them as they are; this is the way to resolve such an objection. But if neither answer applies, one may say that this is simply what people say, as with the stories about the gods; for perhaps it is neither better nor true to speak of them so, but it may simply be as Xenophanes held—still, that is what people say.

55| [1461a] Other things again are perhaps not better as put, but that was simply how matters stood — for example the point about the weapons, "their spears stood upright on the butt-spike": that was the custom then, just as it still is among the Illyrians now. As to whether something has been well said or well done by someone, one must not look only at the deed or the saying itself and ask whether it is good or bad, but also at the agent or speaker — the person addressed, the occasion, the instrument used, and the purpose behind it: for example, whether a greater good was the goal, meant to result, or a greater evil, meant to be prevented. Other problems must be resolved by attending to the diction — for example, to a rare word, as in "the mules first": perhaps he does not mean the mules but the sentries. And the line on Dolon, "who indeed was poor in looks," points not to a misshapen body but to an unsightly face — since among the Cretans, a good-looking face is what "handsome" designates. And "mix it zoroteron" does not mean unmixed, as for heavy drinkers, but "more quickly." Other expressions are said by way of metaphor, as when "all the gods and men slept the whole night through," while at the same time he says, "whenever he looked out over the Trojan plain, [he heard] the din of flutes and pipes": for "all" is said instead of "many" by metaphor, since "all" is a kind of "much."

56| And "alone she has no share" is likewise by metaphor, for the best-known thing is called the only one. Other problems are solved by accent, as Hippias of Thasos solved "we grant him to win his prayer" and "the part of it that rots in the rain." Others are solved by punctuation, as with Empedocles: "suddenly things grew mortal that before had learned to be immortal, and things unmixed before became mixed." Others are solved by ambiguity, as "more of the night has gone by": for "more" is ambiguous. Others are solved by the ordinary usage of speech. They call mixed wine simply "wine," and from this usage comes the phrase "a greave of freshly-worked tin"; and those who work iron are called "bronze-workers," and from this usage it is said that Ganymede pours wine for Zeus, though the gods do not drink wine. This last case, however, could also be explained as metaphor. And whenever a word seems to signify some contradiction, one must also consider in how many ways it could bear that meaning in the passage in question — for example, in "there the bronze spear was stopped," one should ask in how many ways it might be stopped "there," this way or that, in whichever sense would most naturally be understood: taking it as the direct opposite, or as Glaucon states it —

57| [1461b] namely that some people assume something without good reason, and then, having themselves passed a verdict on it, reason from that assumption, and censure the poet for having said what merely seems to them to have been said, whenever it conflicts with their own opinion. This is what has happened with the matter concerning Icarius. For they suppose him to be a Spartan, and so find it strange that Telemachus did not meet him when he went to Lacedaemon. But perhaps the matter stands as the Cephallenians say: they claim that Odysseus married a wife from among their own people, and that the man's name was Icadius, not Icarius; the difficulty, it seems likely, arises from an error (the text here is uncertain). Broadly speaking, an impossibility needs defending on one of three grounds: the demands of poetic composition, a standard of excellence, or common belief. For with a view to poetry, a persuasive impossibility is preferable to an unpersuasive possibility; ... that men should be such as Zeuxis painted them, but better — for the model ought to surpass its subject. To the charges that things are irrational one should answer in this way too: that sometimes a thing is not in fact irrational, for it is likely that even improbable things happen. Apparently contradictory statements call for the same scrutiny used in refuting arguments: check that the term applies identically — same object, same respect, same manner — ...to determine whether the poet contradicts himself [corrupt text], or contradicts what he himself says elsewhere, or what a person of sound judgment would suppose.. Censure is justified in the case of irrationality and of depravity, whenever, though there is no necessity for it, the poet makes use of the irrational — as Euripides does with Aegeus — or of baseness of character, as with the character of Menelaus in the Orestes.

58| The criticisms brought against poets, then, fall under five kinds: works are censured as impossible, or as irrational, or as harmful, or as self-contradictory, or as contrary to correctness as defined by the craft. The solutions to these must be examined under the categories just stated; there are twelve of them. One might raise the question whether epic imitation or tragic imitation is the better. For if the less vulgar art is the better, and the art addressed to the better sort of spectators is always of that kind, then it is quite obvious that an art which imitates everything indiscriminately is vulgar: performers, supposing the audience will not understand unless they themselves add something, make a great deal of movement — as inferior flute-players roll about if a discus-throw must be represented, and drag the chorus-leader about if they are playing Scylla. Tragedy, then, is of this sort, just as the earlier actors thought of their successors: for Mynniscus used to call Callippides "the ape," on the ground that he overacted so excessively, and the same opinion was held of Pindarus as well.

59| [1462a] And as these actors stand to one another, so tragic art as a whole stands to epic poetry. Epic, they say, is addressed to respectable spectators who need no bodily gestures at all, while tragedy is addressed to inferior ones; and if it is vulgar, it is clearly the worse art. But in the first place, this charge is not against the poetic art but against the actor's art, since it is possible to overdo the gestures while merely reciting, as Sosistratus did, or while singing, as Mnasitheus of Opus used to do. Next, it is not every instance of movement that deserves censure — dancing as a whole is no more guilty than movement as a whole — rather only the movement performed by inferior performers — which is exactly what Callippides was criticized for, and what others are criticized for now, on the ground that they represent women who are not respectable. Further, tragedy achieves its own proper effect even without movement, just as epic does: for through reading alone it becomes clear what kind of thing it is; so if it is superior in every other respect, this defect need not belong to it of necessity. Then again, tragedy has everything that epic has — it too can make use of meter — and it has besides, as no small part of itself, music [and spectacle], through which its pleasures are established most vividly; further, it possesses vividness both when read and when performed on stage: [1462b] and further, its imitation reaches its end within a shorter length, since what is concentrated is more pleasant than what is diluted over a long stretch of time — I mean, for example, if one were to set Sophocles' Oedipus in as many verses as the Iliad.

60| Further, the imitation carried out by epic poets is less unified — a sign of this is that from any single epic imitation several tragedies can be produced — so that if epic poets treat a single plot, it either appears truncated when shown briefly, or, if it keeps to the length proper to the meter, appears watered down. I mean, for instance, a case where the poem is composed of several actions, as the Iliad has many such parts, and the Odyssey too, parts which even by themselves possess magnitude; and yet these epics are put together as skillfully as can be managed, and come as close as possible to representing one unified action. Given, therefore, that tragedy has the advantage on every one of these points, and additionally in the function proper to the art — since these poems ought to produce not just any pleasure whatever, but the pleasure already specified — it is clear that, since it attains its end more fully, tragedy would be superior to epic poetry. Concerning tragedy and epic poetry, then — both in themselves and in their species and parts, how many these are and in what they differ, what the causes are of their being good or not, ...and concerning objections and their solutions — let this much be said. . . . [text breaks off]

← All of Aristotle: The Complete Works of Aristotle