0 Aristotle Poetics TransMalcolmHeath PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 70

ARISTOTLE

Poetics

Translated with an introduction and notes by


MALCOLM HEATH

P E N G UI N B O O K S
PENGUIN BOOKS

Published by the Penguin Group


Penguin Books Ltd, 80 Strand, London WC2R ORL. England
Penguin Putnam Inc., 375 Hudson Street, New York, New York 10014, USA C O N T E N T S
Penguin Books Australia Ltd, 250 Camberwell Road, Camberwcll, Victoria 3124. Australia
Penguin Books Canada Ltd, IO Alcorn Avenue, Toronto, Ontario. Canada M4V 3B2
Penguin Books India (P) Ltd, 11 Community Centre, Panchsheel Park. New Delhi- 110 017, India
Penguin Books (NZ) Ltd, Cnr Rosedale and Airborne Roads, Albany, Auckland, New Zealand
Penguin Books (South Africa) (Pty) Ltd. 24 Sturdee Avenue, Rosebank 2196. South Africa INTRODUCTION vii
Penguin Books Ltd, Registered Offices: 80 Strand, London WC2R OR L , England 1 . Human culture, poetry and the Poetics IX

www.pcnguin.com
2. Imitation Xll

This edition fine published 1996 3. Aristotle's history of poetry xv


035
4. The analysis of tragedy XVlll

Copyright 0 M>lcolm Heath, 1996 XX!!


5. Plot: the basics
All rights reterved
6. Reversal and recognition xxvm
The monl right of the editor bas been asterted
7. The best kinds of tragic plot XXX1

Filmtet in 10/tipt Monotype Dembo


8. The pleasures of tragedy xxxv
Typctet by Datix International Limited, Bungay, Suffolk
Printed in England by Clays Ltd, St Ives pie 9. The other parts of tragedy .xliii
10. Tragedy: miscellaneous aspects .xlviii
Except in the United Sates of America, this book is sold subject
to the condition that it shall not, by way of trade or otherwise, be lent, II. Epic liv
re-sold, hired ou� or otherwise circuhted without the publisher's
prior consent in any form of binding or cover other than that in 12. Comedy !xii
which it is published and without a similar condition including this
13. Further reading lxiv
condition being impoted on the subtequent purchaser
14. Reference conventions lxvi
ISBN-13: 978-0-14�636-4
NOTES TO THE INTRODUCTION lxviii
www.greenpenguin.co.uk
SYNOPSIS OF THE POETICS l.xxii

Penguin Books is co mmitted to a suscainab1e


MIX
future for our business, our readers and our planet. POETICS
P•perfrom
_.... ....... This book is made from Foresc Scewardship
!'.,� FSC" C018179 Council™ cercified paper.

NOTES TO THE TRANSLATION 49


I N T R O D U C T I O N

Aristotle was much admired in the ancient world for the elegance
and clarity of his style. Unfortunately, the writings which earned
him that esteem have not survived. What we read today are not the
books which Aristotle prepared and polis�ed for publication, but
notes (perhaps in many cases lecture-notes) compiled for his own
use or the use of his students. This has one great advantage: the Aris­
totelian works available to us, making no concessions to a lay reader­
ship, are the ones which intellectually sophisticated commentators
in late antiquity found philosophically most rewarding. But there are
also disadvantages. These texts were not designed for public con­
sumption, and are consequently often very difficult to understand.
The process by which they took their present form is unclear; in
some cases there are signs of editorial activity (either by Aristotle
himself or by a later hand ) ; so different versions may have been
spliced together, and what is presented as a single continuous text
may in fact juxtapose different stages in the development of
Aristotle's thinking. In general the style is cryptic, condensed and
allusive; the Poetics, in particular, contains many passages which are
more than usually obscure, as the notes to this translation will testify.
This situation has a paradoxical consequence. The works which
did most to disseminate Aristotle's ideas on poetry in the ancient
world were the three books On Poets (written, like Plato's works, in
the form of a dialogue and presumably more lucid than the extant
Poetics) and the six books of Homeric Problems, which discussed pas­
sages in Homer faulted by critics as implausible, inconsistent or
morally improper. Except for scattered fragments quoted by other
ancient authors these two works have been lost. The 1 Poetics itself
does not seem to have been widely known in antiquity. By contrast,

vii
I N T R O DU C T I O N I N T R O DUC TIO N

since the Renaissance its influence on literary theorists and critics subtle intellect. Concepts and arguments which at first seem im­
has been massive; but the obscurities of the text have left it open to a penetrable often prove to make illuminating sense when further
wide range of conflicting interpretations. There have been, and still reflection brings to light their underlying rationale. For me, the
are, fundamental disagreements about the meaning even of key challenge of trying to understand Aristotle's thought is the main
concepts, like hamartia and katharsis. reason why the Poetics continues to be such a rewarding text to
The historic influence of the Poetics is one reason why it merits study. This desire to understand is something with which Aristotle
continued attention. Much of Western thinking about poetry and himself would have been in sympathy; indeed, it provides us with an
drama from the sixteenth century onwards will be obscure to those excellent starting-point in trying to place the Poetics in the broader
who are unfamiliar with the text which lies behind it. It is of course context of his philosophical work.
possible to take an interest in the variety of meanings which this text
has had for successive generations of later readers without concern­
ing oneself with the meanings which Aristotle himself might have I. Human culture, poetry and the Poetics
been seeking to convey. There are, however, various reasons why it
may also be worth taking an interest in what Aristotle meant. 'All human beings by nature desire knowledge.' This, the opening
One reason is that it may help us to a better understanding of sentence of the Metaphysics (98oa1),3 states a fundamental premise of
_
Greek tragedy. We have in the Poetics an analysis of tragedy by an Aristotle's understanding of what it is to be human. He points out as
intelligent and well-informed observer who was much closer, evidence for his claim the pleasure we take in looking at things and
chronologically and culturally, to the plays than we are; it would be assimilating information through our senses. Sensation is of course
irresponsible for the student of Greek tragedy to ignore his testi­ shared with many other animals, and the accretion of sensation
mony. To be sure, Aristotle was not a direct contemporary of the through memory into experience is shared with some. But humans
great fifth-century tragedians whose plays have survived;' and al­ are unique in their capacity to derive universal judgements from
though he numbered later tragedians (such as Theodectes) among their experiences. Animals act by instinct or acquired habit, but
his acquaintances, tragedy in the fourth century was not the same as humans are capable of acting from understanding: they know (as a
tragedy in the fifth - as Aristotle himself was aware. There is no dog might know) that this is the thing to do in a certain situation,
reason to assume that Aristotle's understanding of tragedy was either but they may also understand (as a dog cannot) why it is the thing to
faultless or uncontroversial. It is widely accepted that he failed to do. This is what Aristotle calls in Greek tekhne; the word is con­
appreciate fully the significance of the gods in fifth-century tragedy; ventionally translated as 'craft', 'skill' or 'art', but Aristotle defines
and it is clear from polemical remarks in the Poetics itself that his tekhne as a productive capacity informed by an understanding of its
views on a number of issues diverged from those of some at least of intrinsic rationale (cf. Nicomachean Ethics, 114oa20f.). For Aristotle,
his contemporaries. We must show due caution, then, in using the the evolution of human culture is in large part the evolution of
Poetics as an aid to understanding the nature of Greek tragedy. But tekhne. The first arts which human beings developed were those
that is true of any body of evidence, and is no reason to neglect it. concerned with producing the necessities of their existence. Then
Another reason why the Poetics is worth studying closely is the came recreational arts - those which, while not necessary, enhance
quality of its thought. Aristotle had an exceptionally penetrating and the quality of human life. In due course, activities arose which

Vlll ix
I N T R O DUC T I O N I N T R O DU C T I O N

simply satisfy the desire to know (Aristotle's example is mathemat­ perceive similarities; this, he says, is a natural gift and cannot be
ics). Ultimately, philosophy emerged. Philosophy is rooted in the taught (59a4-8). Aristotle is unlikely to have assumed, therefore, that
basic human instinct to seek knowledge: the world puzzles us and reading the Poetics would make someone good at composing poetry,
arouses our sense of wonder, and so prompts us to look for explana­ and it is unrealistic to think of the Poetics as a do-it-yourself manual
tions. Philosophy is therefore the sophisticated descendant of for would-be poets. Aristotle's interest is philosophical; that is, it is
primitive myth-making responses to an astonishing world. driven by his desire to understand. The production of good poems is
Aristotle's sketch of human culture is relevant in a number of an activity that can be understood, and the Poetics is an attempt to lay
ways to the Poetics. Human beings produce, among other things, that intelligibility open to inspection. It is no surprise that Aristotle
poems, and the production of poems too can be a tekhne; it is an thought this attempt worth making. His own appetite for under­
activity with its own intrinsic rationale, and it can be rendered intel­ standing was omnivorous (he did pioneering work in - among other·
ligible. This does not mean that poets themselves necessarily under­ things - logic, physics and metaphysics, biology, psychology, ethics
stand what they are doing. In the Poetics Aristotle does not treat it as and politics), and poetry was an important feature in the public cul­
a matter of any consequence whether a given poet has a reflective ture of ancient Greek communities. Indeed, poetry had always ex­
understanding of his craft. In chapter 8 he leaves open the question cited the wonder which Aristotle sees as the root of philosophical
of whether Homer's grasp of correct plot-structure was due to enquiry: the Greeks habitually talked of the intense pleasure to be
tekhne or to instinct (5 ra24). In chapter 14 he suggests that early derived from poetry, and of the bewitching enchantment it could
dramatists discovered the best stories to use in tragedy by chance work.
rather than by tekhne (54a9-12); trial and error established a reper­ Aristotle may also have believed that understanding of the kind
toire of first-rate tragic stories, but the dramatists would not have he pursues in the Poetics can enhance the pleasure which poetry
been able to explain why those stories were best (as Aristotle thinks gives us. A pointer to this can be found in one of his biological
he can). In his discussion of tekhne at the beginning of the Metaphys­ works, where he pre-empts objections to the trivial and unpleasant
ics Aristotle notes that unreflective experience may produce the nature of some of the organisms he will be studying (Parts ofAnimals,
same result as tekhne (981a12-15). In general, the ability to do some­ 645a8-15):
thing well does not depend on understanding, nor does understand­
There are animals which are unattractive to the senses when one studies
ing necessarily imply an ability to do it well. A joiner taught to make
them; but even in these, nature's craftsmanship provides innumerable p leas­
a piece of furniture in a particular way may d? it perfectly, even if
ures for those who can discern the causes and have an aptitude for p hilo­
he does not understand the reasons why that is the best way to do it;
sophy. It would be unreasonable - in fact, absurd - if we got pleasure from
he may even do it better than a colleague who has more understand­
studying pictures of these things, because then we are at the same ti me
ing but less manual dexterity.
studying the art [tekhne] whi ch crafted them (e.g. the art of painting or
There are reasons why this principle might apply to poetry espe­
sculpture), but did not get even more pleasure from studying the actual
cially. Poets must be able to project themselves into the emotions of
products of nature - at least when we can make out their causes.
others; natural talent, or even a touch of insanity, are necessary for
this (55a30-4). Moreover, metaphor (which Aristotle regards as the Understanding why an insect or a worm has the form it does is a
most important feature of poetic language) depends on the ability to source of pleasure; this is analogous to the pleasure which can be

x xi
IN T R OpU C T I O N IN T R O D U C T I O N

derived from the study of a painting if one has an understanding believes both painting and poetry to be forms of mimesis, a word
of its tekhne that is, of the reasons why this was the right way to
- which I shall translate as 'imitation'.4 Many scholars would object to
depict that subject. The analogy implies a sophisticated observer. We this rendering, and prefer 'representation'. All translations are, of
have seen that Aristotle leaves it an open question whether poets course, to some extent inadequate, and 'imitation' is by no means
proceed by tekhne, by instinct or by trial and error, and the same can perfect; but there are two reasons why 'representation' may be par­
be assumed for painters. We cannot suppose that any casual observer ticularly unhelpful in this context. First, it fails to capture an essential
will have a more explicit grasp of the requirements of tekhne than element in Aristotle's concept of mimesis - that of a similarity which
the artist; so Aristotle must be comparing the expert naturalist's does not rest wholly on convention. For example, an arbitrary
pleasure in the understanding of natural phenomena to the pleasure symbol on a map may 'represent' an airport, but the representation is
which an expert critic derives from his or her appreciation of a purely conventional; the symbol is not a mimesis of the airport. A
painter's skill. As we shall see, Aristotle regards poetry and painting scaled outline of its runways would be a mimesis. Aristotle is quite
as fundamentally similar activities; so the Poetics may enhance the explicit in linking mimesis and similarity even in cases where we
pleasure we derive from a well-constructed play by helping us to would find it odd to speak of 'imitation'; he says, for example, that
understand why it is good. melody and rhythm can be 'likenesses' and 'imitations' of character
and emotion (Politics, 134oa18-23, 38f.), effectively equating the two
terms. Secondly, 'representation' fails to capture the full range of
2. Imitation Aristotle's concept. The use of a quasi-technical term of modern
aesthetics may tend to obscure the continuity which Aristotle per­
The passage from Parts of Animals explains a pleasure which expert ceives between mimesis in painting, poetry and music and in other,
critics can derive from the skilful depiction even of unprepossessing non-artistic forms of activity, such as the mimicry of animal noises
objects. The expertise on which this pleasure depends is not to be and other sounds (Poetics, 47a20, cf. Plato, Republic, 397a, Laws, 669c­
expected of the majority of people who view such paintings; but d ) and children's play-acting (Poetics, 48b7f., cf. Politics, 1336a32-4).
everyone can gain pleasure from them. A more general explanation This continuity is essential to Aristotle's argument in chapter 4 of
for this phenomenon is needed, therefore, and Aristotle provides it the Poetics; his point there is precisely that poetry is an expression of
in chapter 4 of the Poetics. Here he roots the visual arts in the human a human instinct for mimesis that is also displayed in more element­
desire for knowledge {48b10-19, cf. Rhetoric, 1371b4-10). Someone ary forms of behaviour.
looking at a painting needs to recognize it as a depiction of a given Aristotle's contention, then, is that human beings are by nature
object (as a portrait of, for example, Socrates). This act of recognition prone to engage in the creation of likenesses, and to respond to
involves an exercise of our capacity for cognition; and the exercise of likenesses with pleasure, and he explains this instinct by reference
any capacity is, for Aristotle, in itself pleasurable (Nicomachean Ethics, to their innate desire for knowledge. A likeness is (by definition) a
I 174b14-5a21). likeness of something; to take part in the activity of making and re­
This passage in chapter 4 of the Poetics is one of many in which sponding to likenesses we must recognize the relationship between
Aristotle refers to painting and the visual arts in order to make a the likeness and its object. This engages and satisfies the desire to
point about poetry. He regards these analogies as valid because he exercise our distinctively human power of understanding, and is

Xll Xlll
I N T R O D U C TIO N I N T R O DUC T I O N

therefore pleasurable. This exercise of our capacity for understand­ had no objection to poetic plots based on them (6ob35-61a1). A
ing is, to be sure, a rudimentary one. But Aristotle's purpose here is poem which recounts the actions of a vengeful deity is not an imita­
to explain how poetry and painting are rooted in basic instincts tion of something which (in Aristotle's view) does or could exist in
shared by even the least intellectually sophisticated people. W hen the real world; but it is an imitation of the kind of thing which
Aristotle discerns the roots of philosophy in a primitive sense of would necessarily or probably happen if the traditional beliefs about
wonder there is no implication that philosophy has been explained the gods were true.
away or cheapened. Likewise, the contention that poetry is ulti­
mately explicable as an expression of the elementary human desire
for knowledge does not, and is not meant to, provide an exhaustive 3. Aristotle's history ofpoetry
account of poetry. Indeed, Aristotle goes on in the rest of chapter 4
to trace the development of poetry from primitive improvisation to Having established poetry's ultin1ate anthropological roots at the
increasingly complex and sophisticated forms. We shall see in due beginning of chapter 4, Aristotle goes on to sketch the process by
course that the Poetics implies a correspondingly complex analysis of which more sophisticated forms of poetry evolved. The theoretical
what poetry has to offer its audiences (§8). framework for this outline history of poetry is provided by the first
The notion of poetry as imitation raises many complex problems three chapters, which construct a matrix of three different ways in
which cannot be explored in depth here.5 But some points must be which poems can be distinguished from other kinds of imitation
stressed. An imitation need not be a straightforward copy of the and from each other - in terms of their medium, object and mode.
object imitated; the similarity between the object and its likeness First, why poetry? The pleasure which human beings by nature
may reside in a more oblique and abstract correspondence (as derive from imitations in general does not explain why they took up
the doctrine that music contains 'likenesses' of states of character this form of imitation in particular. According to chapter r, poetry
proves). Nor need an imitation be a likeness of an object which is differentiated from other kinds of imitation by the medium in
actually exists. It is clear from chapter 9 that Aristotle is indifferent to which it produces its likenesses: poetry is imitation in rhythmical
whether the events recounted in a poem did or did not happen in language, with or without melodic accompaniment. It is important
reality; historiographical texts are judged by the criterion of fidelity to note that the weight of emphasis in Aristotle's analysis falls on the
to real events, but poems are not. Aristotle's concept of poetry as element of imitation.6 His contemporaries would have been content
imitation is therefore consistent with (although not identical to) that to define poetry as composition in verse; by contrast, Aristotle
of fiction. Indeed, the events in a poem do not even have to conform denies that non-imitative forms of verse are poetry. In saying this he
to the basic structure of reality. Aristotle's dictum that poetry is is not making an aesthetic judgement. As an example of someone
concerned, not with what has happened, but with 'the kind of thing who composes verse but is not a poet Aristotle cites the philosopher
that would happen, i.e. what is possible in accordance with prob­ Empedocles; but we know that he had a high regard for the artistic
ability or necessity' (51a36-8) must be read in the light of the appli­ qualities ofEmpedocles' verse. Aristotle is not distinguishing poetry
cation of his critical principles in chapter 25. For example, Aristotle from other forms of verse in terms of its linguistic artistry; he is con­
did not believe that the theology built into traditional Greek myths cerned solely with the use to which the verse is put - is it imitation
was true; but (unlike some earlier philosophers, including Plato) he or not? Nevertheless, verse is a traditional medium for imitation, and

xiv xv
I N T R O DUC T I O N I N T R O D U C TIO N

this fact needs to be explained. Aristotle believes that human beings from two premises. The first is that poetry is better if it has a struc­
have an instinct for rhythm and melody, as they do for imitation. tured plot. He infers that the earliest poems would have recounted
This claim, made in passing in chapter 4 (48b20f.), is echoed later the glorious deeds of some god or hero admiringly, or the wicked or
when rhythm and song are identified as features which make the inept deeds of some inferior character in a satirical vein, without
language of tragedy pleasurable (49b28f.). The pleasure which there being any structured sequence to the events described. In due
human beings take in rhythm and melody makes it natural that their course, these disjointed strings of admirable or contemptible actions
instinct for imitation should be expressed in the medium of verse gave way to connected narratives, as in epic. The importance of a
and song. coherently structured plot is a crucial element in the Poetics, and we
Aristotle conjectures, not unreasonably, that imitation in verse shall examine its implications in detail later (§5-7).
and song would have begun with simple improvisations, out of Aristotle's second premise is that poetry is better if (to use the
which the highly sophisticated forms of poetry known in his own terminology of chapter 3) its mode is dramatic rather than narrative.
day would gradually have developed. Aristotle's account of this de­ Poetry is imitation; it seeks to create likenesses, and the likeness is
velopment is sometimes criticized for imposing models of growth greater if the words of those involved in the action are presented
derived from his biological research. This is misleading. In one sense, directly rather than being mediated by a narrator. Hence Aristotle's
the evolution of poetic forms is a natural process. If, as Aristotle be­ admiring remarks on Homer in chapter 24 (6oas-9):
lieves, poetry expresses a number of human instincts (such as those
Homer deserves praise for many reasons, but above all because he alone
for imitation, rhythm and melody), it is an activity in which human
among poets is not ignorant of what he should do in his own person. The
beings will by nature tend to engage. Moreover, poets are likely over
poet in person should say as little as possible; that is not what makes him an
time to discover better and better ways of doing it, if only by ex­
imitator. Other poets perform in person throughout, and imitate little and
periment; thus tragic poets (in Aristotle's view) found the best kind
seldom.
of tragic plot by trial and error (54a9-I2). Once the optimum form
of anything has been achieved, further development of it is by defin­ The implication that epic narrative as such is not imitation seems at
ition impossible; thereafter, there can only be (at best) a proliferation first sight to contradict what is said of it elsewhere in the Poetics, but
of different instances of that optimum form. It is in this sense that the paradox is only superficial. The purest form of poetic imitation
Aristotle can say that 'after undergoing many transformations tra­ is in the dramatic mode; other modes are imitative, but not in the
gedy came to rest, because it had attained its natural state' (49ar4f.). same degree.8 The Homeric poems, with their high proportion of
But in chapters 4 and 5 he displays considerable interest in indi­ direct speech, therefore represent the highest possible development
vidual contributions to the development of poetry, and he is aware of epic, but also disclose a potential which cannot be fully realized
that social and institutional factors, as well as individual incompet­ within the constraints of narrative form. So Homer points the way
ence, may inhibit the continued realization of the optimum form towards drama; but drama proper was (in Aristotle's view) a separate
(51b35-52a1, 53a33-5). Clearly, then, Aristotle saw the history of and subsequent development. His hypothesis is that it originated as
poetry as a social, and not simply as a natural, phenomenon. 7 an adaptation of improvisatory poetic forms in which a soloist led
The assumed trajectory of development will depend on what and responded to a chorus (49a9-14); the distinction between
kind of poetry one regards as most sophisticated. Aristotle works chorus and chorus-leader opens up the possibility of dialogue, and

XVI XV!l
I N T R O DUC T I O N I N T R O DUC T I O N

hence of drama. The leader thus became the first actor; as poets also be constituent parts of tragedy. To see what Aristotle means by
perceived and increasingly exploited the dramatic possibilities of these two terms, imagine that you have left me alone with your
spoken dialogue the number of actors was increased to three, silver spoons. Broadly speaking, there are two factors that will de­
and the chorus declined in importance (49ar5-19) - indeed, by termine whether or not I steal them. One is whether I am honest;
Aristotle's day its importance had declined more than he thought this is the kind of thing which Aristotle means by character - an
appropriate (s6a25-32). agent's settled moral disposition. The other relevant factor is how I
So Aristotle sees the history of poetry as a development towards interpret the situation: do I think that I am likely to avoid suspicion
greater coherence in plot-structure, and towards the more truly imi­ if I take the spoons? This is what Aristotle means by reasoning. If I
tative dramatic mode. But the dichotomy between the imitation am dishonest and reason that I can get away with it, I am likely to
of admirable and inferior agents and activities which he assumes steal the spoons; to use a phrase that recurs persistently throughout
was present in the earliest poetry remains constant. Chapter 2 the Poetics, it is 'necessary or probable' that I will steal the spoons if I
explained its theoretical rationale, and it can be observed to persist am dishonest and think I can get away with it. Thus character sets
throughout the history of poetry: praise-poems are balanced by my agenda (what would I like to do?), and reasoning relates that
lampoons, heroic epic by narrative burlesques, tragedy by comedy. agenda to a given situation (what is it feasible to do in these
The Poetics concentrates on tragedy, the most highly developed circumstances?).
form of poetry concerned with superior persons.Epic is given rel­ Plot, character and reasoning relate to the object of tragic imita­
atively brief treatment as a pendant to tragedy. A full discussion of tion. The medium of tragedy is rhythmical language, sometimes on
comedy is promised (49b2rf.), but the promise is not fulfilled in the its own and sometimes combined with melody. This gives us two
extant Poetics; this is one of several indications that the text we have further constituents of tragedy: diction and lyric poetry, respectively
is incomplete. the spoken and the sung parts of the play's verbal text.
Tragedy is poetic imitation in the dramatic mode. It is designed to
be acted out on-stage, where the action (unlike the action of an
4. The analysis of tragedy epic) can be seen. So tragedy also includes spectacle (the translation is
conventional and unsatisfactory: it refers to everything that is visible
The framework for the analysis of tragedy is set out in chapter 6. A on stage, and is not limited simply to striking effects). We must be
famous definition states what tragedy is; from this Aristotle deduces cautious here. A tragedy is a poem, not a performance. A tragedy
the constituent parts of tragedy; he then ranks these constituents in which, for whatever reason, is never performed is no less a tragedy;
order of importance, giving primacy to plot. and a tragedy may be good, even if its performance is botched. So
Tragedy, like all poetry, is an imitation. Specifically, it is an imita­ what is actually seen by a given audience on a given occasion is in­
tion of a certain kind of action. So one constituent part of tragedy is cidental to the play as such. Spectacle is a part of tragedy in the sense
plot, the ordered sequence of events which make up the action being that tragedy (unlike epic) is potentially performable; so the poet has
imitated. An action is performed by agents, and agents necessarily a duty to ensure that his text can be performed without visual
have moral and intellectual characteristics, expressed in what they absurdity (at 55a22--9 Aristotle cryptically mentions one poet's
do and say. From this we can deduce that character and reasoning will failure to achieve this). However, the actual realization of his text in

xvm XIX
INTRODUCTION I N T R O DUC T I O N

visible stage-action is not the poet's responsibility, but that of the success or failure. It should be noted that the Greek word praxis has a
stage-manager and director. It is therefore not surprising that (at the wider range of meanings than its conventional English translation,
end of chapter 6) Aristotle rates spectacle as the least important of 'action'. If, using the corresponding verb, I ask someone 'Ti pratteis?'
tragedy's constituents. I might be saying either 'What are you doing?' or 'How are you?' So
More striking, perhaps, is the relatively low priority he gives to praxis means 'action' not just in the sense of what someone does but
the verbal text of tragedy.9 If poetry were defined as composition in also in the sense of how theyfare. Aristotle will say in chapter 11 that
verse, this down-rating of the linguistic text would be impossible to suffering ( pathos) is 'an action [praxis] that involves destruction or
defend; but Aristotle has firmly rejected that definition in chapter 1, pain' (s2b I 1f.); the apparent paradox in describing suffering as an ac­
and from his perspective the low priority attached to language tion disappears when one takes account of the broad sense of praxis.
makes sense. Rhythmical language is tragedy's medium; it is a means Success and failure depend on action, therefore. But why does
to tragedy's end, that end being the imitation of an action. From this Aristotle refer to success or failure? According to his definition,
it is a reasonable inference that the choice of the action to be imit­ tragedy 'effect[s] through pity and fear the purification of such emo­
ated is more crucial to achieving tragedy's effect than the way in tions' (49b27f.). We shall return in §8 to the difficult problem posed
which the imitation is realized in words. This is not to deny that an by 'purification' (katharsis) ; here it is sufficient to note that tragedy
incompetent composer of verse could ruin a well-chosen tragic aims to excite a response of pity and fear. Tragedy is 'an imitation . . .
plot, nor that the deficiencies of a second-rate plot could be hidden of events that evoke fear and pity' (52a2f.). These emotions (which
by inspired verse. But the language is there to help realize the plot's Aristotle analyses in detail in Book 2 of the Rhetoric, chapters 5 and
potential, and in that sense is subordinate and secondary. 8) are responses to success and failure; for example, we pity the
Aristotle's arguments for the primacy of plot are therefore prim­ talented person who is prevented by adverse circumstance from
arily arguments for the primacy of plot over character. He begins achieving the success he or she deserves. (Note that in Aristotle's
by claiming that 'tragedy is not an imitation of persons, but of ac­ view we do not pity someone for a lack of talent; it is the lack of
tions and of life' (soa16f.). The reason he gives is that good and bad success which we pity, and our pity is greater when it is a talented
fortune ('well-being and ill-being', as Aristotle puts it, highlighting person who fails than when the failure is in keeping with a person's
the contrast in the Greek by an unusual choice of word for mis­ ability.) So Aristotle's first argument for the primacy of plot is as
fortune) depend on action. An outstandingly talented person is not follows: tragedy aims to excite fear and pity; these emotions are
necessarily outstandingly successful; talents have to be exercised. As responses to success and failure; success and failure depend on
Aristotle observes, in an athletic competition the prize is not award­ action; hence action is the most essential thing in tragedy; therefore
ed to the athlete in best condition, but to the one who actually plot is the most important element.
comes first (Eudemian Ethics, 1219b9f., cf. Nicomachean Ethics, The second argument approaches from the opposite direction:
1099a3-5). We can speak of success and failure, therefore, only in re­ tragedy is impossible without plot, but it is possible without char­
lation to the exercise of someone's abilities; and the outcome of this acter; if character is dispensable, it cannot be as important as plot.
exercise will not be determined by the person's abilities alone, but is The idea of drama without the depiction of character may seem
also influenced by the opportunities they have, and so forth. In this surprising. In the context of the dichotomy of character and reason­
sense, therefore, it is action and not only character that determines ing, the implication is a play in which action arises out of people's

xx XXl
I N T R O DUC T I O N I N T R O DUC T I O N

perception of and reasoning about the possibilities of a situation cept of completeness or wholeness (the terms are effectively equivalent)
without any impression being conveyed of their underlying moral introduces a famous dictum: 'A whole is that which has a beginning,
disposition. Knowledge of an individual's character is not essential a middle and an end ' (5ob26f.). This is not (despite appearances) a
to an understanding of their actions; we can hear reports of things trivial observation. To call something a beginning, a middle or an
done by complete strangers and recognize that their actions make end in Aristotle's sense is not simply to comment on the position in
sense in human terms, or be perplexed because they apparently do a series which it happens to have; the positions described are not
not. So in a tragedy without character motivation would be handled random, but necessary. Aristotle is talking about an ordered structure.
impersonally (this is what someone would do in this situation) rather His definitions of beginning, middle and end show that there are
than concretely (this is what a person with this particular set of char­ two aspects of the structure of a plot which he wants to bring out
acteristics would do). Aristotle does not suggest that such a tragedy when he uses these terms. First, the plot consists of a connected series
would be as good as a tragedy with character; in fact, we know from of events: one thing follows on another as a necessary consequence.
the discussion of Homer's treatment of character that he would not Secondly, the plot consists of a self-contained series of events: the first
(6oarnf.). His point is only that if such a tragedy is possible in prin­ thing in the series is in some sense self-explanatory - it is not a
ciple, then character cannot be essential to tragedy in the way that necessary consequence of something else; equally, the last event in
plot is. the series brings it to a definite end - there is no further necessary
consequence in the series. Another term for this self-containment is
closure. The series of events which constitutes a well-formed plot is
5 . Plot: the basics therefore closed at both ends, and connected in between.
Consider by way of illustration a simple story: 'Bill strangled a cat.
Aristotle's emphasis on the primacy of plot is reflected in the Ben strangled a cat.' This is not a 'complete' plot in Aristotelian
amount of space he allocates to it: chapters 7-14 are devoted almost terms. The two events have no necessary connection. So let us try
entirely to an analysis of plot. There are three stages in this analysis. again: 'Bill stra�gled Ben's cat. So Ben strangled Bill's cat in retali­
Chapters 7-9 specify the minimum conditions which any plot must ation.' This is better: we can now see how the two events hang to­
satisfy if it is to be well-formed. The end of chapter 9 and the next gether; the series of events is connected. But is it self-contained?
two chapters distinguish two kinds of plot, simple and complex; since W hy did the cat-strangling start in the first place? And was that the
Aristotle will claim that complex plots (in the technical sense he has end of it? Let us try once more: 'Bill thought that his cat was going
defined ) are superior to simple ones, this part of the discussion to lose to Ben's in the cat-show. So he strangled Ben's cat. Ben
moves beyond the minimum conditions required of any plot and strangled Bill's cat in retaliation. They never spoke to each other
begins to consider what makes one well-formed plot better than again.' Now the story, connected and self-contained as it is, does sat­
another. This question is tackled more systematically in chapters isfy Aristotle's criteria for being whole or complete. It has few other
1 3-14, which address the question of the best kind of tragic plot. '0 virtues, but at this stage we are talking about the minimum criteria for
The definition at the beginning of chapter 6 stated that tragedy is being a well-formed plot; we are not concerned with the qualities
an imitation of an action that is complete and has magnitude. These which make one plot better than another.
two concepts are taken up in chapter 7. The exposition of the con- Aristotle is often quoted as if he had said that a play has a beginning,

XXll XXlll
INTRODUCTION I N T R O D U C T I ON

a middle and an end. This is wrong. It is the plot, the underlying change from bad fortune to good fortune, or from good to bad. The
sequence of actions, that has this structure. To illustrate the distinc­ plot must have sufficient scope for such a change to take place.
tion, consider Harold Pinter's Betrayal. This play traces the breakdown Moreover, it must take place in the way prescribed in the discussion
.
of a relationship in reverse chronological order: the opening scene of completeness. It is not enough to juxtapose prosperity and
shows the end of the process, and the scenes progress backward in misery; the change from one to another must be the result of a
time to its beginning. So the plot (the events which in Aristotle's sequence of necessarily connected events. If a successful cat-breeder
terms are the object being imitated ) runs in the opposite direction wins a glittering prize and promptly commits suicide, the plot must
to the play which imitates it; the beginning of the one is the end of show why he did so - out of remorse (as it may be) for having
the other, and vice versa. To take a classical example, in the Odyssey strangled his rival's cat.
the wanderings of Odysseus are included within the story of his The change of fortune is a new element, but fits in with Aris- .
home-corning by way of a retrospective narrative placed in the totle's first argument for the primacy of plot. The change from
mouth of Odysseus himself. If the wanderings are counted as part of good fortune to bad eorresponds to the failure in action that evokes
the plot, then the order of events in text and plot is different; if they pity. Unless the plot of tragedy has sufficient scope to allow for
are not counted as part of the plot (as the synopsis at 55br6-23 may such a change, the emotional effect at which tragedy aims cannot
imply), then part of the text is devoted to something other than an be achieved. We must note, however, that Aristotle envisages the
exposition of the plot. In either case, text and plot are distinct. Plot is changes of fortune going in either direction; a change from bad for­
therefore not co-extensive with the play; this is why Aristotle can tune to good is also mentioned as a possibility. This may seem sur­
refer to parts of the plot which fall outside the play (53b3 rf., 54b2-8, prising: what has a change to good fortune to do with fear and pity?
55b24f., 6oa27-32). The reader should be careful not to forget the This is a question to which we shall return when we come to the
level of abstraction at which Aristotle is working throughout the discussion of the best kind (or kinds) of tragic plot in chapters I 3
chapters on plot: he is not concerned here with the construction of and 14.
the verbal artefacts which are tragedies, but with the design of the W hen Aristotle discussed wholeness at the beginning of chapter
patterns of events which underlie them. 7 he talked about things following one on another in a necessary se­
After defining completeness, Aristotle moves on to magnitude. In quence. By the end of chapter 7 this requirement has been modified;
one sense it is trivial to say that a tragic plot must have magnitude: a here we find 'a series of events occurring sequentially in accordance
plot of zero extent would not be a plot at all, since it would contain with probability or necessity' (5 ra12f.); the pairing of necessity with
no events. The real question Aristotle wants to raise is more interest­ probability will recur throughout the Poetics. In other works Aris­
ing: what is the correct magnitude of a tragic plot? In practice, the totle explains these terms as referring to what happens always or
time available for a performance is a key determining factor, but this usually; even in the early part of chapter 7 he was willing to speak
is a contingent fact about the organization of a particular theatrical about things following 'necessarily or in general' (5ob30). It is neces­
event, and throws no light on the art of poetry as such. In principle, sary that the sun will rise tomorrow morning; the sun always rises in
the upper limit is determined by what an audience can grasp at one the morning. It is probable that I will get out of bed tomorrow
time: it has to be possible for them to remember what is in the plot. morning; in general I do get up in the morning, but I might stay in
The lower limit is determined by the need for the plot to include a bed all day if I am ill or have died or am feeling exceptionally idle.

xxiv xxv
I N T R O DUC T I ON I N T R O DUC T I O N

Events in the human sphere are generally no more than probable; so is not the kind of thing that would happen. If I tell you that Bill got
this qualification of the initial formula makes it a little more realistic. another bank statement today, you will not jump to the conclusion
Chapter 8 introduces the concept of unity. Any imitation is uni­ that he is going to strangle another of Ben's cats tomorrow. In other
fied if it imitates a single thing; so an imitation of an action will be words, when Aristotle speaks of 'the kind of thing that would
unified if it imitates a single action (51a30-32). 1 1 By what criterion, happen', he is not talking about individual events but about con­
then, are we to judge the singleness of an action? It cannot mean a nected sequences of events. If a poet wants to construct a plot out of
single event; if a tragic plot involves a change of fortune, the single a given sequence of events, it is not enough that those events actually
action will inevitably include a series of events. The analysis of com­ happened; what is essential is that they are connected with each
pleteness tells us that it must be a self-contained series of connected other in the way defined in chapter 7; and if they are so connected, it
events, and Aristotle's criterion of unity does not add anything in does not matter whether they actually happened or not.
substance to this analysis (nor should we expect it to: something Another way of putting this, which Aristotle discusses in chapter
complete and whole is a unity). But it allows him to formulate a 9, is to say that poetry 'tends to express universals' (51b6f.). We must
negative point more sharply: the fact that a plot is concerned with be careful here. Poetic plots do not deal in generalizations ('people
the actions and experiences of a single person is not enough to make usually get up in the morning'); they make statements about what a
it unified, since there may be no necessary or probable connection particular individual does at a particular time ('Bill got up this morn­
between them. Consider another simple story: 'Bill got a statement ing'). Indeed, the actions with which tragic plots are concerned are
from the bank; the next day he strangled Ben's cat.' This is not a typically so exceptional that it would be absurd to talk of generaliza­
unified plot by Aristotelian criteria; even though Bill is involved in tion. Orestes killed his mother; but it is not true that people gener­
both the incidents reported, there does not seem to be any necessary ally kill their mothers, nor even that people like Orestes generally
or probable connection between them. 12 kill their mothers in such circumstances; such circumstances do not
At the beginning of chapter 9 Aristotle takes a further step: 'It is arise in general - that is one reason why Orestes' situation is such a
also clear from what has been said that the function of the poet is not potent basis for tragedy. But if the plot of Aeschylus' Oresteia is well­
to say what has happened, but to say the kind of thing that would formed, then it is true that a person like Orestes would necessarily
happen, i.e. what is possible in accordance with probability or neces­ or probably kill his mother in such circumstances. So behind the
sity' (5ra36-8) . The distinction between what did happen and what particular statement about what Orestes did lies a premise about
would happen is not as sharp as it might seem at first glance; later in what such a person would necessarily or probably do in such cir­
the chapter Aristotle observes that 'there is nothing to prevent some cumstances; and this premise is a universal truth, however ex­
of the things which have happened from being the kind of thing ceptional such persons and such circumstances may be in actuality.
which probably would happen' (51b3of.). The two classes therefore Poetry is concerned with particular sequences of events; but the
overlap; the poet can say what did happen, but only if it is also the connection between those events means that they instantiate uni­
kind of thing that would happen. In what cases, then, can something versal structures.
happen without being the kind of thing that would happen? Con­ To talk about universal structures is to talk about the things in
sider our last example. Even if this sequence of events actually hap­ which philosophy is interested. The universality of poetry therefore
pened, the lack of a necessary or probable connection means that it gives it something in common with philosophy. This is not to say

XXVI XXVll
I N T R O DUC T I O N I N T R O D U C TIO N

that poetry is philosophy carried on by other means. Philosophy is particularly effective: 'these effects occur above all when things
directly concerned with universal truths, but poetry's .concern is come about contrary to expectation but because of one another'
indirect: the universality of poetry is a by-product of its aim to con­ (52a3f.). 'Contrary to expectation' introduces the notion of aston­
struct effective plots. Consequently, while philosophy is concerned ishment, while 'because of one another' provides an anchor to the
with universal truths, what lies behind an effective poetic plot discussion of necessary or probable connection that has gone before.
may be the universalization of a conventional falsehood; hence, as Astonishment and connection are both desirable if the emotional
we have seen (§2), Aristotle has no objections to plots based on impact of the plot is to be maximized.
traditional beliefs about the gods, even though he would dismiss If connection is one of the things which increase emotional
those beliefS on philosophical grounds. impact, we can see why Aristotle's theory of poetic plots places so
Historiography, by contrast, although bound to the truth of what much emphasis on it. But his slightly convoluted illustration of the .
happened, has no commitment to universality; history records what importance of connection also suggests a degree of flexibility in
happened within a given period of time irrespective of whether the practice. A man called Mitys is murdered; later, a statue of Mitys
events form a sequence linked by necessity or probability. '3 Aristotle topples and kills his murderer. There is in fact no causal connection
explicitly rejects plots constructed like works of historiography in here, but it looks as if there was or ought to be such a connection;
chapter 23,just as he had rejected plots constructed like biographies the two events seem to hang together, as if fate were bringing the
in chapter 8 - a conclusion anticipated in his argument for the pri­ murderer to justice or Mitys' vengeance were reaching out beyond
macy of plot on the grounds that the emotions which tragedy aims the grave. Aristotle's argument is that the illusion of a connection
to evoke are responses to success and failure: hence 'tragedy is not an increases our sense of astonishment at this series of events; if an
imitation of persons, but of actions' (5oa r 6f.). illusion has that effect, then surely an actual connection must do so
as well. There is ample evidence later in the Poetics of Aristotle's
flexibility in this regard. In chapters 24 and 25 he approaches 'ir­
6. Reversal and recognition rationalities' in the plot with caution, but does not absolutely rule
them out. Aristo�e sets out his requirements for a well-formed plot
The end of chapter 9 introduces astonishment (to thaumaston in as things which in principle poetry ought to aim for, but his applica­
Greek covers a range of related ideas: surprise, amazement, wonder) tion ofthis principle is by no means rigid or doctrinaire; as we shall see
into the discussion. Astonishment is a good thing in a tragic plot, but (§ r r) , he recognizes that departures from the norm he has defined
it is not a necessary thing in the way that wholeness and unity are. So may be advantageous in some circumstances.
here Aristotle moves on from the minimum conditions for a well­ Astonishment is not explicitly mentioned in chapters IO and I I,

formed plot, and begins to consider what makes one such plot but it is a crucial concept underlying the distinctions and definitions
better than another. He refers at once to pity and fear: tragedy is an which they contain. Chapter IO defines two classes of plot. A simple
'imitation not just of a complete action, but also ofevents that evoke plot satisfies three conditions: the events are 'in the sense defined
fear and pity' (52ar-3). In judging the quality of one tragic plot over continuous and unified ' (that is, connected in accordance with
against another, it is their emotional impact to which Aristotle ap­ necessity or probability) ; there is a change of fortune (as specified
peals. He identifies two things which make a sequence of events in chapter 7); and there is no reversal or recognition. A complex plot

XXVlll xxix
I N T R O DUC T I O N I N T R O DUC T I O N

also satisfies the first two conditions, but unlike the simple plot it reveals that, because things are not what they seemed, what a person
does have reversal or recognition. These two terms are themselves has done or is about to do is not what he thought it was - for ex­
defined in chapter l l . ample, that he has not killed a hostile stranger, as he supposed, but his
Recognition (anagnorisis) is 'a change from ignorance to know­ own father. This parallel between reversal and recognition points the
ledge' (52a29-3 l ) . The point is that this change affects the good or way to tragic error in chapter 1 3 , and to the role of ignorance in
bad fortune of the person involved: Oedipus learns that he has killed chapter 1 4.
his father and married his mother, and this recognition is the final
blow that shatters his world. Note that recognition is associated
above all with 'close relationship and enmity', on the grounds that 7. The best kinds of tragic plot
such relationships have the closest bearing on an individual's good
and bad fortune: Oedipus' world would not have been shattered if Chapters l 3 and 14 address the question of the best kind of tragic
the man he had killed had turned out to be a complete stranger. This plot. Both chapters assume that this is the plot that is most effective
premise will be examined more closely in chapter 1 4 (see §7) . in arousing pity and fear, but they take different lines of approach
Reversal (peripeteia) is less straightforward. It is emphatically not to and reach seemingly incompatible conclusions.
be equated with the tragic change of fortune: a change of fortune is The approach in chapter l 3 is to analyse in detail the tragic
a characteristic of all tragic plots, simple as well as complex, while change of fortune. Aristotle identifies two variables in this change,
reversal is distinctive to complex plots. But Aristotle's defmition is which between them determine our emotional response to it: one is
vague: 'there is a change to the opposite in the actions being per­ the direction of change (from good fortune to bad, or from bad to
formed, as stated ' (52a22f.). 'As stated ' must refer to the discussion good ); the other is the moral status of the person or persons in­
of astonishment; reversal involves an astonishing inversion of the ex­ volved in the change (virtuous or wicked ) . Since pity and fear are
pected outcome of some action, but that astonishment should not responses to bad fortune, the change from good fortune to bad is
be achieved at the cost of necessary or probable connection. So, for rated more highly than the reverse. But the fall of an outstandingly
example, Oedipus' discovery of the terrible truth is the paradoxical virtuous character into misfortune is morally repellent and disgusts
but necessary consequence of the arrival of a messenger who aims to us, while the fall of a bad character into misfortune is morally satisfy­
bring good news and does everything he can to put an end to Oedi­ ing and pleases us; in both cases the response of pity is blocked by a
pus' worries; here reversal and recognition reinforce each other contrary reaction. So the ideal tragic plot cannot be constructed
(52a3 2f.). around an exceptionally virtuous person or a wicked person; it must
The astonishment produced by reversal involves an overturn of therefore be based on someone between these two - broadly speak­
expectation. There is therefore a close parallel between reversal and ing virtuous, but not outstandingly so. Because their virtue is not
recognition: both reveal that the situation in which a character has outstanding, we do not find their downfall morally repellent;
been acting was misinterpreted. Reversal reveals that, because things because their downfall is undeserved, we can pity them.
are not what they seemed, the outcome of a person's actions will be If a tragic character does not fall into misfortune because he or
other than what had been expected - for example, that the recipient she deserves to, what is the reason for the change in their fortune?
of a message will be devastated rather than reassured. Recognition Aristotle's answer introduces one of the most famous, and most

xxx XXXl
I N T R O DUC T I O N INTRODUCTION

often misunderstood, concepts of the Poetics: error. The Greek word Hamartia, then, includes errors made in ignorance or through mis­
hamartia covers making a mistake or getting something wrong in the judgement; but it will also include moral errors of a kind which do
most general sense; so the word itself gives little help in interpreting not imply wickedness. Aristotle's attempt to prescribe the best kind
Aristotle's precise meaning, and we must be guided by the context. of tragic plot is therefore not as narrowly prescriptive as it may seem
This at once excludes the interpretation of hamartia as a moral flaw: at first sight. His procedure is negative. He excludes various kinds of
the second time Aristotle uses the word he speaks of a 'serious' ha­ plot which he thinks demonstrably less than ideal; but that leaves
martia ( 5 3 a 1 6) ; but a serious moral flaw would be precisely the wick­ considerable scope for diversity. The change to bad fortune must
edness that Aristotle has ruled out. To avoid this inconsistency some come about because of a hamartia (that is, not deservedly) ; but since
interpreters have concluded that hamartia has no moral content at all . hamartia can take a variety of forms, the best kind of tragic plot is not
On this view, Aristotle is referring exclusively to intellectual errors - narrowly prescribed.
to ignorance and mistakes of fact. This dovetails well with his Aristotle's starting-point in chapter 1 4 is the relationship between
account of reversal and recognition; both arise out of a misinter­ characters. The tragic effect is enhanced when people inflict harm
pretation of the circumstances in which a person is acting. So error on those 'closely connected with them'. This rather clumsy expres­
in this sense is central to the complex plots which Aristotle favours. sion (which we met also in chapter r 1 , in connection with recogni­
But this interpretation of hamartia is too restricted. First, the as­ tion) is an attempt to render the Greek word philos, conventionally
sumption that moral and intellectual errors can be sharply dis­ translated as 'friend ' but in reality of much wider application. One's
tinguished is not correct. I might get an arithmetical calculation philoi include all those to whom one is bound by ties ofmutual obli­
wrong because I am not paying attention; ifl am an engineer design­ gation - above all, the members of one's own family. The obligation
ing a bridge, that negligence clearly has a moral as well as an intel­ to help friends and family, to protect them and to promote their
lectual dimension. If a bridge collapses with great loss of life because interests, is one of the most fundamental principles of ancient Greek
of an engineer's negligent calculation, the moral aspect of this error ethics; indeed, a common summary of virtue (that is, of the kind of
is serious; but this does not necessarily mean that the engineer is a behaviour most to be admired ) was 'help your friends, harm your
wicked person. We can imagine circumstances under which the enemies'. Aristotle's point, then, is that a tragic plot is more likely to
negligence, though blameworthy, is understandable. Suppose that he evoke fear and pity if a person inflicts harm on a philos, someone
had just been told that he had terminal cancer: he ought to have taken close to them. His next step is to correlate this point with two other
more care, but we can understand his failure to do so. Secondly, it is variables: first, whether the agent actually goes through with the
possible to construct situations in which someone who knows what harmful act or not; secondly, whether the agent knew what he or
he or she is doing, and so is not subj ect to any intellectual error, does she was doing or not.
something they ought not to do, but understandably so. Perhaps they The reappearance of the idea that acting in ignorance of the true
give way to intolerable pressure or provocation. The circumstances situation may be the basis for tragedy gives chapter 1 4 something in
under which they act therefore mitigate the moral error in some common with chapter 1 1 , where reversal and recognition both in­
way. The error may still be a serious one, in the sense that it has volved this kind of misapprehension, and with chapter r 3 , where
disastrous consequences; but the mitigating circumstances mean that action in ignorance is one possible form of error. Furthermore,
it does not express a serious flaw in their moral character. Aristotle argues here that plots based on ignorance are superior to

XXXl1 XXXlll
IN T R O D U C T I O N I N T R O DUC T I O N

those in which harmful action is either planned or carried out with Aristotle's two lines of argument therefore reach different, and
full knowledge of the circumstances. He uses the same word (miaros) apparently inconsistent, conclusions. Since there is an explicit cross­
here as he did in chapter l 3 to describe the suffering of an outstand­ reference from chapter 1 4 (s4a9) back to chapter 1 3 (s3a1 8-22),
ingly virtuous person. When someone knowingly plans or inflicts there is little doubt that he meant the two chapters to stand together,
injury on one of the people with whom he or she is most closely despite the seeming inconsistency. This paradox has not been satis­
connected we feel disgust, and our sense of revulsion interferes with factorily explained. If we insist that there is one kind of tragic plot
the emotions of fear and pity. It is better, therefore, to have the that is best, then the two chapters are contradictory. But we have
character act in ignorance; there is then no sense of outrage to already seen that the concept of error in chapter l 3 is designedly
interfere with our sense of pity. Indeed, someone who unwittingly open-ended, and, on the assumption that there may be a variety of
harms a person close to them is to be pitied; so in these situations we excellent tragic plots, the two chapters could be allowed to reach
can pity the agent as well as the victim. different conclusions without contradiction. It must be conceded,
The next step in Aristotle's argument is something of a surprise. however, that Aristotle has not presented his arguments in this light.
It is best of all, he says, if the injury is planned in ignorance but the
plan is not carried out. That is, it is better if the identity of the
intended victim is discovered before the injury is inflicted, so that 8. The pleasures of tragedy
disaster can be averted, rather than after the injury is inflicted, when
the disaster can only. be mourned. In chapter 1 3 the best plot is one In the introduction to chapter 1 4 Aristotle says the tragic poet
in which a moderately virtuous person moved from good fortune to 'should not seek every pleasure from tragedy, but the one that is
bad fortune. The second-best plot is the kind which has a double characteristic of it' ( s 3 b r nf.); that is, he continues, 'the poet should
line of development, with the good characters ultimately enjoying produce the pleasure which comes from pity and fear, and should do
good fortune, while the bad characters end up in misfortune. A play so by means of imitation'. This stress on the characteristic pleasure
of this kind might dramatize a conflict between a good character (oikeia hedone) of tragedy has two functions. First, it serves to dis­
and a bad character. For example, the good character may start off tinguish what is appropriate to tragedy from what is appropriate to
in bad fortune because he or she is · oppressed by the bad character; other forms of poetry; thus the second-best kind of plot in chapter
this kind of play will evoke fear and pity because of the good char­ l 3, with a happy ending for the good characters, gives a pleasure
acter's initial misery. Or one could imagine plots in which fear more akin to the characteristic pleasure of comedy (s3a3 5f.) . Sec­
and pity are evoked by the apparent imminence of a fall into ondly, it distinguishes the characteristically tragic pleasure from
misery, which is averted at the last moment. Plots like these will other pleasures which tragedy arouses, but which are not distinctive
excite the tragic emotions in spite of their happy ending, since to it. We return here, therefore, to a point made earlier (§ 1 ) : that
the characters we sympathize with pass through or anticipate Aristotle's analysis of what tragedy offers to its audiences is complex
misfortune; but they are not, according to the analysis in chapter and multi-layered.
1 3 , the best kind of tragic plot. In chapter 14, by contrast, the best We start on relatively firm ground. In the definition of tragedy at
kind is precisely one in which a change to bad fortune is imminent the beginning of chapter 6 Aristotle says that tragedy is composed in
but does not occur. 'language made pleasurable' (49b2 5), which is explained as speech

XXXlV xxxv
I N T R O DUC T I O N I N T R O DUC T I O N

with rhythm and (in some parts) melody. The fact that tragedy is distinctive to tragedy; the same process of recognition and under­
written in verse, and that part of it is sung, gives us pleasure. This is standing the plot as a connected sequence of events is involved in
consistent with the observation in chapter 4 that rhythm and watching comedy. Nor does the cognitive pleasure come from pity
melody are natural to human beings (48b20f.); for Aristotle 'natural' and fear; if the cognitive pleasure can be derived from comedy just
implies naturally pleasurable. Likewise, at the end of chapter 6 song as much as from tragedy, it must be neutral as to the emotions
is described as 'the most important of the sources of pleasure' which accompany it. The cognitive pleasure is therefore, like the
(5ob 1 6); and the pleasure to be derived from music is identified in pleasures of text and performance, a pleasure derived from tragedy,
chapter 26 as one of the points in which tragedy is superior to epic but not its distinctive and 'characteristic' pleasure.
(62a 1 5- 1 7). The same passage probably (it is not quite certain what The next step brings us to a much-discussed, and probably insol­
Aristotle wrote) identifies spectacle as an additional source of pleas­ uble problem: katharsis. This concept appears just once and fleetingly
ure. So there is pleasure to be got both from the verbal text of tra­ in the Poetics, at the end of the definition of tragedy in chapter 6:
gedy, and from its visual and aural realization - from what is seen on 'effecting through pity and fear the purification [katharsis] of such
stage and heard when a play is performed. But these pleasures of the emotions' (49b27f.). It reappears at somewhat greater length in the
verbal text and performance are not distinctive to tragedy; they are last book of the Politics; but there Aristotle says that he need give
present in comedy as well. And we have already seen that Aristotle only a brief account, since he has discussed it in more detail in his
rates lyric poetry and spectacle as the least important parts of tragedy. Poetics ( 1 3 4 1b3 8-40) . The text of the Poetics as we now have it is
Another pleasure is implied by the discussion of poetry as imita­ probably incomplete; internal and external evidence suggests that
tion at the beginning of chapter 4 (§2). Aristotle observes that imita­ there was originally a second book, including (at least) the promised
tion is naturally pleasurable to human beings, and explains this with discussion of comedy whose absence we have already noted (§3) . A
reference to the process ofrecognition which it involves (48b 1 2- 1 7) . fuller discussion of katharsis may originally have appeared in the
When we look a t a picture o f Socrates we have t o identify i t as a missing book. But that is of little comfort; we must do our best with
picture of Socrates; this exercise of our capacity to learn and under­ the little we have.
stand is (in Aristotle's view) pleasurable. If tragedy is an imitation, The context of Aristotle's reference to katharsis in the Politics is a
then in watching and responding to a tragedy we must engage in a discussion of various uses of music. For children, music has an educa­
similar process of recognition and understanding. In the light of tive function; for adults it has a role in relaxation and leisure; but it
Aristotle's analysis of plot we can see that the process involved in can also be used to bring about katharsis. Aristotle's example refers
watching a tragedy is somewhat more complex than that involved to people prone to 'enthusiasm', by which he means hysterical or
in recognizing a picture as a picture of Socrates. When we watch ecstatic frenzy such as that associated with certain religious cults,
a tragedy we have to follow the plot, which means recognizing that like the cult of Dionysus. Aristotle observes that music which
the events are a connected sequence, and recognizing that this stimulates their frenzy can have a calming effect on such people
sequence corresponds to some universal pattern (i.e. that the events ( 1 3 42a4-1 5):
are connected in accordance with necessity or probability) . But this
The emotion which affects some minds violently exists in all , b u t in differ­
is not unduly demanding, and the process will still therefore be
ent degrees, e.g. pity and fear, and also enthusiasm; for some people are
inherently pleasurable. 14 This cognitive pleasure is, however, not

xxxvi XXXV!l
I N T R O D U C TI O N I NTRODUCTION

prone to this disturbance, and we can observe the effect of sacred music on Deficiency as well as excess of emotion is a deviation from the eth­
such people: whenever they make use of songs which arouse the mind to ical ideal. If I am paralysed with fear at the sight of a mouse, my fear
frenzy, they are calmed and attain as it were healing and katharsis. Necessar­ is inappropriate and excessive; that is a sign of cowardice. But if l sit
ily, precisely the same effect applies to those prone to pity or fear or, in gen­ nonchalantly in the path of an oncoming steam-roller, then my lack
eral, any other emotion, and to others to the extent that each is susceptible of fear is equally inappropriate and excessive; that is a sign of reck­
to such things: for all there occurs katharsis and pleasurable relief . lessness. Courage, located somewhere between cowardice and reck­
lessness, recognizes real dangers and responds appropriately to them.
So the relief that katharsis brings is pleasurable. A pleasure that comes
It is appropriate to feel fear in battle; if you did not, you would be
from the katharsis of pity and fear is, at any rate, not shared with
prone to act rashly, endangering your own and your comrades' lives.
comedy. So the question arises whether this kathartic pleasure is the
But this fear should not be overwhelming; otherwise, you might
characteristic pleasure of tragedy.
desert your post. So for Aristotle the crucial point is not, as it is with
In trying to make sense of the notion of a katharsis of pity and fear,
Plato, to suppress your emotions; it is rather to feel the right degree
one thing must be stressed: Aristotle does not think that emotions
of emotion in the right circumstances.
are bad things in themselves. In this respect his outlook differs from
So the katharsis of fear and pity cannot be understood as getting
that of Plato, whose critique of poetry in Book ro of the Republic is
rid of those feelings. Nevertheless, Aristotle does talk about katharsis
based in part on a profound suspicion of the emotions. One of
in quasi-medical terms in the passage from the Politics quoted above.
Plato's complaints is that poetry arouses emotion, and in so doing
In those subject to enthusiasm kathartic music brings about 'as it
increases our tendency to be emotional; but in his view we should
were healing and katharsis'; and all those prone to pity, fear or any
be bringing our emotions under control, not strengthening them
other emotion enjoy 'katharsis and pleasurable relief'. References to
in this way (Republic, 605e-6d ) . Aristotle has a more sophisticated
'healing' and 'relief' imply that katharsis does in some sense put right
and reasonable view of emotions. They are not irrational impulses.
something that is wrong with us. What, then, might it be that is 'as it
They are grounded in our understanding, since an emotional re­
were heal[ed]' through the katharsis of pity and fear, if not the emo­
sponse to a situation presupposes an interpretation of it (as fear im­
tions themselves? The obvious answer is: an excess of those emotions.
plies an assessment of the situation as one that threatens pain or
The katharsis of pity and fear would (if this is correct) work on
injury); and since such responses can be more or less appropriate to
people whose disordered emotional susceptibilities make them
the situation, they are open to ethical evaluation. So there is an
prone to feel these emotions at the wrong time, in response to the
intimate link between emotion and virtue (Nicomachean Ethics,
wrong things, with regard to the wrong people, for the wrong
1 1 06b 1 8-23):
reason or in the wrong way. By stimulating the emotion to which
For example, fear, confidence, desire, anger, pity and i n general pleasure and they are excessively prone, tragedy discharges the tendency to
distress can be experienced in greater or lesser degree, and in both cases excess; it thus relieves the pressure which their disordered emotional
wrongly. To feel them at the right time, in response to the right things, with make-up exerts on them, so that in ordinary life they will not be so
regard to the right people, for the right reason and in the right way - that is prone to indulge the emotion in question. On this interpretation,
the mean and the optimum, which is the characteristic of virtue. · katharsis does not purge the emotion, in the sense of getting rid of it;
it gets rid of an emotional excess and thus leaves the emotion in a

XXXVlll XXXI X
I N T R O DU C T I O N I N T R O DU C T I O N

more balanced state, mitigating the tendency to feel it inappropri­ tragic plot (53a33f.). I n chapter 26, when he is argumg for the
ately. Why should this be pleasurable? From an Aristotelian point superiority of tragedy over epic, Aristotle has to counter the objec­
of view any process that restores one to a natural or healthy state tions of those who think that tragedy appeals to vulgar and inferior
is pleasurable (Nicomachean Ethics, 1 1 52b r r-20, I I 54b r 7- r 9) . If audiences; but if the characteristic pleasure of tragedy was one that
you are prone to feel some emotion to excess, then yo"! are in had most appeal to spectators whose emotions (and, therefore, moral
an unnatural state; the disorder of your emotional condition is character) were significantly disordered, Aristotle's high regard for
analogous to having a disorder in your physical condition. When tragedy would be hard to sustain. Aristotle seems therefore to be
you are thirsty, satisfying your thirst is pleasurable; when someone committed .to the view that the characteristic pleasure of tragedy
has trodden on your toe, it is nice to feel the throbbing die away; is one which will appeal at least as much to the better and more
something similar will apply to the restoration of your emotions to virtuous members of the audience; the pleasure of katharsis does
a natural or healthy state. not fit this description.
If this approach to the problem of katharsis is even broadly cor­ This conclusion runs counter to the widespread assumption that
rect, ' 5 certain consequences seem to follow. The kathartic effect ap­ the reference to katharsis in the definition of tragedy in chapter 6 is
plies to someone watching a tragedy only to the extent that his or meant to state the 'final cause' of tragedy - that is, the end or
her emotional state is disordered; the more prone someone is to feel purpose for the sake of which tragedy exists. But there is no reason
excessive or inappropriate emotion, the more benefit he or she to expect an Aristotelian definition to state the final cause of a
stands to derive from katharsis, and (presumably) the more pleasure. phenomenon (important though that may be in a complete analysis
This is obvious: if a process is like healing, it applies in particular to of it); a definition states what a thing is - its essence (49b23f.) or
those who have most wrong with them. But Aristotle would cer­ form. 16 Thus the definition of tragedy (49b24-8) tells us what tra­
tainly not have accepted that everyone's emotional dispositions are gedy is in general terms ('an imitation . . . ') . It then uses the matrix
seriously disordered; even the quotation from the Politics shows that constructed in chapters 1-3 to differentiate tragedy from other kinds
he saw various degrees of proneness to emotions, and only a minor­ of imitation in terms of its object (' . . . of an action that is admirable,
ity are violently (that is, excessively) prone to them. Given the close complete and possesses magnitude . . .'), medium ( . . . in language
'

connection between emotion and virtue, to say that everyone's made pleasurable, each of its species separated in different parts . . . ')
emotions were seriously out of order would for Aristotle be to say and mode (' . . . performed by actors, not through narration . . . ). It
'

that there were no virtuous people, which is absurd. So katharsis is fundamental to an understanding of tragedy that the action which
will not apply, or will not apply in the same degree, to all members it imitates is one evocative of fear and pity (52a 1-3 , 52b32); so this
of the audience of a tragedy; the effect will be least on those whose too is specified, but it is also (and crucially, in the light of the Pla­
emotional dispositions are least disordered. tonic critique of tragic emotionalism) explained that the evocation
This in turn implies that the kathartic pleasure is not the charac­ of fear and pity is potentially kathartic. A play which was likely to
teristic pleasure of tragedy. Aristotle believes that the better mem­ encourage rather than assuage a tendency to emotional excess
bers of the audience are more responsive to the best kind of tragedy; would not, in Aristotle's view, be a properly constructed tragedy.
note especially, at the end of chapter r 3 , that it is 'the weakness of To say that tragedy should excite emotions in a way that does
audiences' which leads tragedians to prefer the second-best kind of not do ethical harm to its audiences is to state something about

xl xli
I N T R O DUC T I O N I N T R O DUC T I O N

the essence of tragedy, therefore, but it does not imply that the ology rather than poetics. We can only speculate on what resolution
regulation of emotions is the function or purpose of tragedy. '7 of the paradox he would have favoured had he addressed the prob­
On the interpretation that I have outlined, then, katharsis is not lem explicitly. Since it continues to exercise modern philosophers, it
the function of tragedy, but a beneficial effect which tragedy has on would perhaps be surprising if Aristotle had found a compelling
some members of the audience. Aristotle has constructed a multi­ solution.
layered defence against Plato's critique of the appeal of tragedy to
emotion. The first layer is provided by his more sophisticated ethical
theory, in which emotions are recognized as in themselves good; the 9. The other parts of tragedy
Platonic premise that emotions should be suppressed as far as pos­
sible is denied. The second layer is provided by the premise that the The analysis of plot is Aristotle's main concern in chapters 7 to 1 4;
effect of tragedy and its characteristic pleasure is most available to in chapter I 5 he turns to the second most important of the six parts
the best members of the audience; if that is so, then it cannot be the identified in chapter 6, character. This was defined in chapter 6 in
case that tragedy merely panders to the distorted tastes of the vulgar terms of choice (prohairesis) . Character is imitated when what is said
and emotionally disordered. But the audience of a tragedy is likely or (presumably) done reveals the nature of the choice that is made,
nevertheless to include people who are vulgar and emotionally dis­ and hence the underlying moral disposition of the person who is
ordered, and a Platonist critic might claim that there is a danger that speaking or acting. So when Aristotle talks about character he is not
tragedy will have a bad effect on them; even if tragedy does not de­ talking about the quirks and details of someone's individuality, but
liberately pander to their distorted tastes, it may still be true that about the structure of their moral dispositions in so far as it becomes
stimulating their emotions will increase their tendency to feel those clear through what they say and do.
emotions to excess in ordinary life. The final layer of the defence, It is not surprising, therefore, that the first thing he specifies is that
therefore, is the contention that, even for these people, the emo­ the characters should be morally good - or (a qualification added a
tional stimulus provided by tragedy is beneficial, because it can re­ few lines later) that they should not be morally bad unnecessarily. The
lieve and reduce the pressure towards emotional excess by which example given is Menelaus in Euripides' Orestes. Menelaus squirms
they are afflicted. out ofhis obligation to offer protection and assistance to his nephew
The characteristic pleasure of tragedy is therefore not to be iden­ Orestes, a philos (see §7) towards whom his obligations are particu­
tified with the pleasures of text and performance, with the cognitive larly strong. This is 'unnecessary' presumably in the sense that the
pleasure, or with the 'pleasurable relief' of katharsis. The natural in­ plot would have got along perfectly well if Menelaus had not been
ference is that the experience of tragic emotion is pleasurable in such a despicable character. If Menelaus fails to support his nephew,
itself. This is a paradox, since pity and fear are forms of distress (Rhet­ this may be because he is a coward or because he is self-seeking; but
oric, 1 3 82a2 1 , 1 3 85b3 1 ) . But the paradox is one with which the there was no need to make him a self-seeking coward. So Aristotle is
Greeks were familiar; the sophist Gorgias had described the effects not saying that tragedy must be populated only by virtuous people,
of poetry as 'fearful shuddering, tearful pity and a yearning that is but that the characters should be as virtuous as is possible given the
fond of grief' (fragment I r .9) . Aristotle could well take the paradox demands of the plot.
as a given in this context, since its resolution is a question for psych- This requirement must be seen in relation to comments made

xlii xliii
IN T R O D U C T I ON I N T R O DUC T I O N

earlier in the Poetics on the nature of the people involved in the woman in tragedy should be a good woman; she should have the
action of a tragedy. Two passages are particularly relevant: chapter 2 virtues appropriate to a woman, and only those virtues. The fourth
(on the objects of imitation) , and chapter 1 3 (on the moral qualities requirement, too, is straightforward: characters should be consistent.
of the person who undergoes a change of fortune). Chapter 2 con­ This obviously follows . from the requirement of necessary or
trasted 'admirable' people with 'inferior' people (48a2). 'Admirable' probable connection. If someone in a tragedy acts inconsistently
(spoudaios) reappears in the definition of tragedy in chapter 6: 'tra­ and unpredictably, then one cannot say that what they do follows
gedy is an imitation of an action that is admirable' (49b24) ; spoudaios necessarily or probably on what has gone before. The one exception
.
is often translated 'serious' here, but Aristotle means an action in­ is where we have been led to expect someone to behave unpredict­
volving the admirable persons (persons 'better than we are') speci­ ably (perhaps because they are mentally or emotionally unstable, or
fied as the object of tragic imitation in chapter 2. Like many terms of are facing an irresolvable dilemma) ; then, as Aristotle neatly puts it,
commendation and disparagement in Greek, spoudaios and its oppos­ they are consistently inconsistent.
ite embrace social status as well as moral qualities. Thus in chapter I 3 The third requirement is more diffi cult. Literally, Aristotle says
it is stated in passing that the person whose fortunes change should that the character should be 'like' (in Greek homoios) : like what? Aris­
be 'one of those people who are held in great esteem and enjoy totle has used this word twice before in connection with character.
great good fortune' (s3aro), and 'inferior' recurs in chapter I 5 in a In chapter 1 3 he says that fear is felt for someone 'like us' (s3 a5f.) ;
clearly status-oriented context: 'there is such a thing as a good and in chapter 2 he uses the phrase for people 'of the same sort' as
woman and a good slave, even though one of these is perhaps ourselves, as distinct from those better or worse than we are (48a4-
deficient and the other generally speaking inferior' (s4a20-22) . One 1 4). But this creates a problem: in chapter 2 Aristotle says that tra­
cannot expect aristocratic nobility of character from persons of low gedy is concerned with people better than we are, while chapter I 3
status, but they can still be morally good in terms of status-related indicates that tragedy's distinctive emotional response is concerned
norms of behaviour; a slave can be a good slave - loyal, hard­ with people like ourselves. A passage near the end of chapter I 5 may
working and so on. In tragedy he should be; by contrast the slaves in point to a resolution of the apparent inconsistency. There Aristotle
comedy (whatever their virtues) are likely to be disobedient, lazy, compares poets to portrait-painters. Portrait-painters, he says, 'paint
dishonest and self-seeking. Conversely, high-status characters such people as they are, but make them better-looking' (s4b r of.). There
as Menelaus can be morally bad. Tragedy, therefore, is essentially is therefore a combination oflikeness and idealization in portraiture;
concerned with people who are of high status and of good moral a painter might keep Cromwell's warts, but make them seem less
character; there will be peripheral figures (slaves and so forth) of ugly than they really are. In the same way characters can be made
lower status, but they cannot be at the centre of tragedy's interest better than we are while still retaining some imperfections of
and should at least be good of their kind; high-status characters in character; in this respect they will be like us, despite the element of
tragedy can be morally bad, but not if they are meant to be a focus idealization. This would agree with the requirement in chapter 1 3
for our pity, and only if and to the extent that the plot requires this. that tragic characters should be virtuous, but not outstandingly so.
The second of Aristotle's four requirements for character, ap­ They are like us, in that they fall short of the moral perfection
propriateness, is very close to the idea of character being 'good of its whose downfall we would find outrageous; but they still tend to the
kind ' which Aristotle has just applied to persons of low status. A better rather than the worse (cf. 5 3 a 1 6f.).

xliv xiv
IN T R O DUC T I O N IN T R O DUC T I O N

Aristotle's discussion of character is superficial by comparison not seem so apposite. Chapter 20 is an essay on linguistic theory
with the analysis of plot. Even less attention is given to the third part with no particular bearing on poetry. Chapter 2 1 is more directly
of tragedy, reasoning, treated almost parenthetically in chapter 1 9. relevant; it classifies various kinds of departure from standard dis­
Here we find that the concept has undergone a degree of expansion. course, and we learn in the next chapter that such departures from
In chapter 6 reasoning and character were paired together as the two the linguistic norm are characteristic of poetic language. But even
bases of action: ifl am a coward and perceive the current situation as in chapter 22 it is striking that Aristotle is not concerned with the
dangerous, then I am likely to run away. Both my moral disposition diction of tragedy in particular. It is tempting to infer that Aristotle,
(my cowardice) and my understanding of the situation (my belief feeling that something had to be said about poetic diction, in­
that I am in danger) enter into the motivation of action. More for­ corporated material which he had conveniently to hand even
mally, reasoning was defined in chapter 6 in terms of demonstrating though it was not ideally adapted to its context. 19
that something is or is not the case, or making a general statement The section on poetic language concludes Aristotle's discussion
(5oa6f. , b1 i f.). The agent's understanding of the current situation is of tragedy; in chapter 23 he turns his attention to epic. But there is
revealed by what he or she says: an assertion or denial about the par­ clearly some unfinished business. Only four of the six parts identi­
ticular circumstances ('this situation is dangerous', or 'this situation fied in chapter 6 have been discussed; lyric poetry and spectacle have
is not safe') , or a generalization ('it is dangerous to sit in front of not received separate attention. Lyric poetry is a combination of
moving steam-rollers'). When Aristotle talks about reasoning in words and melody. Its verbal dimension could conceivably be seen
chapter 1 9, the elements of assertion, denial and generalization are as treated implicitly in the chapters on poetic language, since their
still present, but others have been added. Reasoning here includes focus is not restricted to spoken dramatic verse; the musical dimen­
speech in which the characters arouse emotion, or make things look sion is perhaps passed over on the grounds that it is part of the real­
important or unimportant (55b37-56a2). In other words, it has ex­ ization of the tragic text in performance. A similar consideration
tended to cover all the ways in which in a tragedy one person can applies to spectacle. As we have already seen (§4), Aristotle says at
use language to influence another. It includes now, not just the ob­ the end of chapter 6 that spectacle is not really integral to the poet's
servation that it is dangerous to sit in front of a steam-roller, but also art; the tragic poet's job is to produce a text that can be performed,
the persuasive devices which one character might use to frighten but the performance itself is a separate matter. Aristotle was certainly
another with the prospect ofsitting in front ofa steam-roller or to con­ aware of the potency of performance; it is precisely that potency
vince them thatit is not something to be entered into lightly. The tech­ which tempts tragedians to rely on it to the neglect of plot, a ten­
niques for this, as Aristotle observes, can be got out of the Rhetoric. 1 8 dency which Aristotle deplores in chapter 1 4 (5 3b7f.). That passage
From reasoning it is a short step to the language in which the does not imply that the use ofspectacle to complement and enhance
reasoning is conveyed, and the next section (from the latter part the effect of a well-constructed plot is improper, and the comments
of chapter 1 9 through to chapter 22) is concerned with diction. at the beginning of chapter 1 7 on the need to visualize the action
This may seem a generous allowance of space, in view of the com­ suggest that Aristotle recognized the importance of a poet's being
paratively low importance which Aristotle attaches to the poet's aware of, and taking steps to enhance, the effect of his text in per­
function as a maker of verses, as distinct from a maker of plots formance. One factor which may help to explain his reluctance to
(5 1 b27--9) . On closer inspection, the material in these chapters does be more positive emerges from chapter 26, which shows that hostile

xlvi xivii
INTRODUCTION INTRODUCTION

critics of tragedy saw performance as one of the objectionable It is no surprise that Aristotle's underlying concern with the
things about it; so Aristotle as a defender of tragedy had a reason not connectedness of the events which make up a plot is repeatedly in
to emphasize performance. In his reply to these critics he argues evidence. For example, chapter 1 6 turns to recognition (defined in
that, since tragedy can have its effect without being performed, it is chapter l l) and classifies a variety of different techniques by which a
not intrinsically more vulgar than epic; but he apparently goes on to recognition can be engineered. Since recognition is one component
make the point that music and spectacle give pleasure (62a 1 6f.). of the complex plot, which in Aristotle's view is more effective than
Thus he concedes the objection to performance for the sake of the simple plot, this classification is not of purely abstract interest;
the argument, and shows that it does not entail the inferiority of the question of how best to handle one of the devices which do
tragedy; but then he reverses the argument and claims that per­ most to enhance tragic effect is an important and eminently prac­
formance is a positive advantage. However, an oddity in the syntax tical one. Aristotle's classification works up from recognitions that
of the Greek at this point has led some editors to conclude that the are in essence accidental to those that are integral to the structure of
words 'and spectacle' have been inserted in error by a later copyist, the plot. Thus the least artistic kind of recognition is that prompted
and that Aristotle mentioned only music; so we cannot in the end by a visible sign. Odysseus is recognized because he has a distinctive
be sure how positively Aristotle expresses himself about spectacle. scar, but, as Aristotle implies in chapter 8 (5 la25-8), his acquisition
of the scar has no causal link with the plot of the Odyssey. By con­
trast, the best class arises by necessity or probability from the plot
r o. Tragedy: miscellaneous aspects itself. In commending recognitions of this kind Aristotle picks up
the idea stated at the end of chapter 9, that when something happens
We have still not finished with tragedy. We jumped from character, both unexpectedly and nevertheless as a necessary or probable con­
in chapter 1 5, to reasoning, in chapter 1 9; but there is a lot of ma­ sequence of what has gone before, this combination increases the
terial in between. This material is not bound to the framework of audience's astonishment and thus enhances the emotional impact
the six constituent parts of tragedy established in chapter 6; its pre­ of events. This is important: Aristotle's preoccupation with neces­
sentation is loosely ordered (at times, disorderly) and sometimes ob­ sary and probable connection is not the product of an abstract
scure to the point of unintelligibility. But it should not be neglected. formalism; he believes that there is an intimate connection between
The material returns repeatedly to Aristotle's central preoccupation the cohesion of the plot and the emotional impact at which tragedy
with plot; but, by contrast with the broad issues of principle treated aims.
in chapters 7-1 4, the emphasis in these later chapters tends to be The latter part of chapter 1 7 likewise explores some procedural
more practical. These chapters are, broadly speaking, concerned implications of the theory of proper plot-structure set out in chap­
with matters of technique, and so they sometimes look beyond the ters 7- 14. A tragedy aims to excite pity and fear, so a story is needed
design of the abstracted plot towards its concrete realization in a that will have a powerful emotional effect; we know from the discus­
play. Thus chapter l 7 starts off from the perspective of a poet 'when sions in chapters 1 3 and 1 4 which kind or kinds of story are most
constructing plots and working them out complete with their lin­ appropriate. Greek mythology is a rich repertoire of such stories, so
guistic expression' (55a22f.); chapter 1 8 includes comments on the it makes sense to turn there to find a suitable subject. We might look,
handling of the chorus (56a25-3 2) . for example, for a story involving acts of violence between close kin

xiviii xlix
INTRODUCTION INTRODUCTION

(53b1 5-22). The most fundamental obligations of all are those of likely to be the climax of our plot: if Oedipus acts in ignorance, he
a child to its father and mother; so a man who kills his father or must eventually find out what he has done, and the devastating
sleeps with his mother is doing two of the most terrible things impact of this discovery on him will be a perfect tragic climax; here
imaginable. This would be the most potent material for tragedy one the analyses of recognition in chapters I I and 1 6 will be relevant.
could hope to find. Oedipus, of course, did both of these things. So But the premise that Oedipus must be unaware of the identity of
let us assume that we have chosen Oedipus as the subject of our his father and mother also tells us something about how the early
tragedy.2° Where do we go from there? What we do not do, accord­ stages of the plot must be developed. His ignorance must surely
ing to Aristotle, is ask ourselves what was the first thing that hap­ imply that he was separated from his parents in infancy. Since the
pened to Oedipus, what happened next and so on. A historian Greeks practised the exposure of unwanted infants, that is not
might work like that; as we have seen (§5), historians in Aristotle's a problem in principle; we just have to supply some motive for
view are concerned with what happened, and so their narratives Oedipus' parents to expose him, and construct a mechanism to
simply reproduce a given sequence of events. But Aristotle's poet is ensure that despite being exposed he survives to adulthood.
not concerned with what did happen, but with the kind of thing We have now embarked on the process which Aristotle describes
that would happen in accordance with necessity �r probability as 'turn[ing] the story into episodes' (55b 1 f.) . The abstract outline to
(5 1 a3 7f.). So we should not just take over a ready-made series of which we have reduced the story needs to be converted into a series
events as our plot. Having selected Oedipus as our subject, we ought of concrete events, or episodes; and that series must be linked by
(as it were) to forget about Oedipus and rethink the story in more necessary or probable connections if we are to satisfy Aristotle's
abstract terms. Someone kills his father _and sleeps with his mother; requirements and avoid the defective plot-structure which he de­
how can we represent this as something that would happen, neces­ scribed earlier as 'episodic' - 'one in which the sequence of episodes
sarily or probably? In other words, how can we imagine this as part is neither necessary nor probable' (5 r b34f.). Since the central events
of a series of connected events? The poet is thus advised to work in our plot are to be Oedipus' killing of his father and sleeping with
down from the most abstract possible formulation of the story of his mother, it would make sense to have those events drive the
Oedipus through more and more concrete specifications of it until others. In both cases we can draw on supernatural assistance. If a
a final version is reached in which Oedipus' killing of his father baby is going to grow up to do such terrible things, a prophecy to
and sleeping with his mother are embodied as a connected series that effect would provide the parents with a very strong motive
of events. indeed to dispose of their baby before he grows up; so we will
How might we proceed? Since tragedy aims to evoke pity and postulate such a prophecy, and Oedipus will be exposed precisely
fear we cannot have Oedipus do these things knowingly. Someone because it is foretold that he will grow up to kill his father and sleep
who knowingly killed his father and slept with his mother would be with his mother. But when he has done these things, he will be
monstrously wicked; his actions would disgust us, and his downfall terribly polluted. In Greek religious thought, it was very dangerous
would evoke satisfaction rather than pity. So Oedipus must act in for a city to harbour polluted persons; they could bring down
ignorance of the identity of his father and mother; recall here the disaster on the whole community. If this happened (if, let us say,
significance of ignorance in chapter I 4, and of error in chapter I 3 . a plague broke out) then the city would naturally try to identify the
Given Oedipus' lack of this knowledge, we can begin to see what is supposed source of the pollution in order to get rid of it; and this

Ii
I N T R O DU C T I O N I N T R O DUC T I O N

will give us a plausible mechanism for setting in train the enquiries If a plot consists of complication and resolution, then excellence
which will lead to the discovery of the truth. Again, a prophecy can in plot-construction (which, for Aristotle, is the key element of the
help us out here: the Thebans would naturally turn to an oracle for poet's art) must embrace both parts. The comment, later in chapter
help in identifying the cause of their problem, and we can frame the I 8, that 'many poets . . . handle the resolution badly' (56a9f.) harks

response to point the enquiry in the right direction." back to a section in the middle of chapter I 5 , interrupting the discus­
Chapter I 8 looks at the structure of a plot from another perspec­ sion of character. According to the Greek text,22 resolution should
tive. Complication and resolution (in Greek desis and lusis, literally arise from the plot (54a3 7f.) . That is, the poet should put together a
'tying' and 'untying') define the two stages of a plot hinging on the sequence of events in which the change of fortune and its con­
change of fortune which has been central to Aristotle's account of sequences are a necessary or probable consequence of everything
tragedy since chapter 7. The complication is everything up to the that has gone before. It is an instance of the kind of unskilful reso­
beginning of the change of fortune; the resolution everything from lution Aristotle mentions in chapter r 8 when the poet has to resort
there on. So, for example, in Sophocles' Oedipus the change of for­ to an arbitrary device to get the plot to work out the way he wants.
tune begins when news of the death of Oedipus' supposed father Aristotle's examples of arbitrary resolution in chapter r 5 involve
arrives and the truth about his real parentage begins to come to divine intervention. One is the end of Euripides' Medea. Aristotle
light; that point was already identified in chapter I I as the play's feels that the playwright has created a situation in which Medea
reversal (52a24-6) . Note that (to return to a distinction made in §5) cannot be extricated from danger in any necessary or probable way;
complication and resolution are not parts ofthe play, but parts of the he has therefore resorted to the premise that the gods can do any­
action which is embodied in the play's plot. The complication may thing, solving the problem by the arbitrary introduction of a god­
include events which occur before the start of the play and which given flying chariot in which Medea can escape. Aristotle's rejection
influence what happens in the play itself, but which are reported of this kind of device does not imply that divine involvement is in­
rather than enacted; in Oedipus everything involved in Oedipus' appropriate in general. One use of the gods which he explicitly ap­
_
birth, exposure and survival, the circumstances under which he proves is to convey information about those parts of the plot which
killed his father and married his mother, the onset of the plague, and fall outside the play. A god explaining the background to the play in
the appeal to Delphi which that crisis prompts and which sets in a prologue, or a god appearing to foretell the future at the end of the
train the enquiry that leads to the discovery of the truth all this is
- play, would be typical examples. But there is of course much middle
crucial to the plot, but precedes the play's opening scene and is ground between the extremes of arbitrary intervention and the
gradually disclosed to the audience as the play proceeds. One could, mere conveying of information. Aphrodite in Euripides' Hippolytus
in principle, have a play which consisted solely of resolution; in such does not just appear .in the prologue to convey information; it is
a play the change of fortune would already have begun, and the play her resentment which has set events in motion, and her intervention
would simply trace its completion and consequences (Aeschylus' in human affairs is essential to the causal structure of the plot. There
Persians would perhaps be an example) . The fact that Aristotle says is no reason why Aristotle should object to this. Her intervention
that the complication includes what is outside the play and 'often' is well-motivated (she is angered by the way in which Hippolytus
some of what is inside it confirms that he recognized the possibility dishonours her) , and its consequences follow by necessity or
of a play consisting solely of resolution. probability. To be sure, Aristotle did not believe in such deities;

Iii !iii
INTRODUCTION INTRODUCTION

but we know from chapter 25 (6ob3 5-<5 1 a 1 ) that the truth of the contained part of the whole story of the Trojan war (the quarrel
theological premises which poets use is not a matter of great con­ between Achilles and Agamemnon in the tenth year of the war) .
cern to him. Hence he can accept unconcernedly the role which Although Aristotle does not make the point here, the same could
Poseidon plays in the plot of the Odyssey (55b 1 8) . be said of the Odyssey. A plot embracing everything that happened
to Odysseus would not be unified (as we know from chapter 8).
A plot embracing the whole story of his wanderings and home­
1 1 . Epic coming would be unified; it would comprise 'a single action of
many parts'. The plot of the actual Odyssey concentrates on the last
It should come as no surprise that Aristotle's first thought when he stage of Odysseus' homecoming, skilfully using his earlier adventures
turns to epic in chapter 2 3 is about the structure of the plot. In epic, in the way that Aristotle says the catalogue of ships is used in the
as in tragedy, the plot should be unified, with a beginning, middle Iliad, as 'episodes . . . to diversify his composition' (59a3 5-'7) .23
and end suitably connected. The contrast between epic and histori­ The beginning of chapter 24 proposes four overlapping types of
ography is familiar already from chapter 8 (where the contrast was epic plot. This classification is applied to the two Homeric epics: the
with biography) and chapter 9; but a more complex set of contrasts fliad is simple and based on suffering, the Odyssey is complex and
is constructed here. The chapter begins by opposing unified poetic based on character. The comment on the Odyssey is obvious
plots and defective plots with a quasi-historical structure. But it enough: the poem is full of recognition-scenes, and the antithesis be­
would be no great tribute to Homer to say that his plots are not tween good (Odysseus, his family and his loyal servants) and bad (the
defective; Homer is outstanding because he has handled plot­ unruly suitors) makes character important. In chapter r 3 the Odyssey
structure in the best possible way. To make the desired distinction was cited as an example of the 'double' plot which ends in good for­
between plots that are simply well-formed and Homer's unique tune for the good characters; so its happy ending contrasts with the
brilliance in the handling of plot, Aristotle introduces a third ele­ suffering which pervades the plot of the fliad, and which gives that
ment into the comparison. There are epics which have defective plots, poem its powerfully tragic character. On the other hand, Aristotle
constructed like works of history or biography about a single period believes that complex plots are superior in principle to simple ones;
of time or a single person irrespective of the causal connectedness of so one implication of the attempt to classify the poems here must be
the events narrated; most epic poets write like this. Other epics have that the Odyssey is, at least in this one respect, superior in its plot­
plots which are unified in Aristotle's sense, but which are very large structure to the fliad, and by implication emotionally more powerful.
indeed. An epic about the whole Trojan War would have been of Any reader who dissents from this conclusion should consider
that kind; it would have had a beginning and an end (59a3 r f.) , and so whether the fault lies in Aristotle's theory, or in his description of
been unified, but the vast mass of material in the middle would have the poems. In the fliad Achilles' actions lead, contrary to his expect­
made it hard to grasp the plot as a whole. In chapter 7 Aristotle ation, to the very thing he least wanted - the death of his dearest
stated that the upper limit on a plot is determined by what can be comrade. There is thus a strong case that the fliad exploits reversal,
held in the memory; so this kind of plot, embodying a 'single action and that its plot is complex. Indeed, the centrality of this reversal to
of many parts' (59b 1), though in principle well-formed, would be the poem's emotional impact could be seen as strongly supporting
pushing at the upper limit. By contrast, Homer selected a single self- Aristotle's preference for complex plots.

liv Iv
I N T R O DU C T I O N I N T R O DU C T I O N

The last part of chapter 24 returns to the theme of Homer's rationality because the action is narrated but not seen (we may recall
merits. Aristotle has already commented (in chapters 8 and 23) on here the comments at the beginning of chapter 1 7 on the poet's
Homer's skill in the handling of plot. Chapter 4 noted Homer's need to visualize the tragic action) ; since he goes on to say that
unique versatility (he alone worked on both sides of the bifurcation irrationality generates astonishment, and astonishment gives pleas­
in poetry between imitations of admirable and inferior characters, ure, the implication is that it is legitimate for epic to exploit its
composing.both epic and narrative burlesque) and his adumbration greater tolerance of the irrational.
of the dramatic mode: he lets the characters speak for themselves. These two concessions have something in common. Irration­
The latter point rests on the high proportion of direct speech in his alities can be included in the parts of the action which lie outside a
narratives, which Aristotle again singles out as an aspect of Homer's play, and in the action of an epic; in neither case will the irrationality
unique excellence (6oa5-1 1 ) . We may note here too that Aristotle is be seen by the audience. Keeping an irrationality out of sight of the ·

favourably impressed by Homer's powers of characterization: 'none audience makes it less salient, and so helps keep intact the impression
of them are characterless: they have character.' Despite the priority that everything is properly connected. The example of Mitys' mur­
which Aristotle gives to plot over character, he recognizes the con­ derer at the end of chapter 9 points in the same direction: an illusion
tribution which character makes to the quality of a poem; and of connectedness can have the same effect as connectedness itself
although the possibility of plot divorced from the imitation of (see §6). The next part of chapter 24, which says that Homer 'taught
character was recognized in chapter 6 (5oa23-9), this statement other poets the right way to tell falsehoods' (6oa1 8f.), suggests a
about Homer confirms that Aristotle would not rate such a com­ further way to smuggle an irrationality into a poem while giving the
position as highly as one with character. impression that everything makes sense. This is to exploit the human
There follows a section on astonishment and the irrational. Aris­ tendency to make fallacious inferences. (Aristotle's explanation sug­
totle's theory of plot, with its strong emphasis on necessary and gests a fallacy of the kind: 'If Daisy is a cow, then Daisy has four legs;
probable connection, tends to exclude irrationalities. If we apply Daisy has four legs; so Daisy is a cow' .) If one is not alert and critical
Aristotle's criterion of good plot-structure strictly we will have to (the audience of a play or an epic recitation is likely to expect that
say that if all the events follow one another in accordance with events will be connected, and may not be in the frame of mind to
necessity or probability, then the sequence of events is rational; if focus on logical puzzles) then the fallacy may pass unnoticed; the
not, the plot is defective. Even Aristotle's approval of astonishment, poet will then have persuaded the audience by subterfuge that
as in complex plots with reversal, does not alter this; as he says at the something is necessary or probable when it is not. Later in the
end of chapter 9, the proper emotional effect depends on astonish­ chapter we read that one should prefer 'probable impossibilities to
ment and connection together (5 1 a 1-6) . But how strictly should implausible possibilities' (6oa26f.). Again, the emphasis is on main­
Aristotle's criteria be applied in practice? One concession has al­ taining the impression of connectedness: something that is (in real­
ready been made in chapter 1 5 : there should not be any irrationality ity) necessary or probable but which looks as if it is not damages that
in the action, but if there is it should be outside the play (54b6f.). impression, and should be avoided; something which is not neces­
This implies that Aristotle would prefer a strict application of his sary or probable but which looks as if it is maintains the impression
criteria, but is able to tolerate a departure from them in certain and is therefore technically superior. Here too a preference for
circumstances. Chapter 24 notes that epic is more tolerant of ir- having no irrationalities is combined with a willingness to tolerate

!vi !vii
INTRODUCTION INTRODUCTION

them. Near the end of chapter 24 Aristotle seems to suggest yet passages; chapter 25 offers an extremely condensed synopsis of this
another way to conceal irrationalities: you may distract attention material, and is correspondingly difficult to interpret. Rather than
from them by the brilliance of the writing, as Homer conceals the pursue all the details, it may be more helpful to look at the general
absurdity involved in the narrative of Odysseus' arrival in Ithaca principles which Aristotle states and consider how they fit in with
'with other good qualities' (6oa34-b2) . the rest of the theory articulated in the Poetics.
Why, ifAristotle prefers the avoidance of irrationalities, is he will­ Aristotle suggests three general principles. The first is based on
ing to tolerate them? Chapter 25 suggests one answer: ' If impossi­ the obj ect of poetic imitation: poetry is not bound to imitate what is
bilities have been included in a poem, that is an error; but it is the case, but can imitate what ought to be the case or is said to be the
correct if it attains the end of the art itself . . . i.e. if it makes either case. This looks similar to what is said in chapter 9: the poet is not
this or some other part have greater impact' (6ob22-6) . Aristotle tied down to real events, but can use invented plots as well. But this ·

recognizes that an irrationality can enhance the effect of a poem, later passage is more radical in its implications. Suppose a poet con­
provided that it is concealed; in such a case he does not object. A structs a plot in which a god inflicts terrible punishment on a
poet designing a plot should aim (other things being equal) at con­ human being who has given offence; is that necessary or probable?
sistency with what is necessary or probable; but where the poem's In reality, no; philosophers will tell us that this is not a true concep­
emotional impact can be increased by a departure from necessity or tion of divinity. But in terms of traditional Greek religious beliefs -
probability, that departure is legitimate. Necessary or probable con­ that is, what is generally said to be the case about the way gods
nection is desirable in general not for any abstract a priori reason, but behave - the god's action is certainly necessary or probable. So the
because in practice it generally serves to increase the impact of plot of a poem does not have to conform to the underlying universal
events (5 2a r-6). So one should always retain the impression of patterns to which events conform in the real world; it is enough if it
necessary or probable connection by distracting from or concealing conforms to the underlying universal patterns to which events con­
(in one way or another) the irrationality; but sometimes a false form in some imaginary or fictitious world, such as that of traditional
impression will be more effective than the real thing. 24 religion. Aristotle's second general principle is based on the
Chapter 2 5 , from which this point has been taken, is concerned medium ofpoetic imitation, language. We know from the discussion
with the objections people raise to poems, and with the range of of poetic style in chapters 2 r and 22 that poetry is distinguished by
possible responses or solutions to those objections. The principles it departures from standard speech; such departures can be used to
enunciates can be applied to any poetic form (some examples from answer objections ifit can be argued that the objector has misunder­
tragedy are mentioned in 6oa3 0-3 2); the discussion is attached to stood one of the non-standard usages characteristic of poetic dic­
the section on epic because the critique of implausible, inconsistent tion. Thirdly, Aristotle insists on a distinction between the standards
and immoral elements in poetry had been most comprehensively by which poetry and the products of any other tekhne are to be
applied to Homer. Books 2 and 3 of Plato's Republic collect a certain judged. This principle, of course, applies to other arts as well. If I
an1ount of such material; the fourth-century sophist Zoilus earned am illustrating a zoology textbook, I ought to get the details of an
the nickname 'Homer's Scourge' (Homeromastix) for his nine books animal's anatomy right and resist the temptation to draw imaginary
devoted to identifying faults in Homer's poems. The six books of beasts; if I am painting pictures to hang in an art-gallery, I can
Aristotle's Homeric Problems collected discussions of such problem legitimately sacrifice strict zoological accuracy in the interests of

)viii !ix
INTRODUCTION INTRODUCTION

the painting's balance or composition, and dragons and unicorns audiences encourage the degenerate tendency of tragedy to turn
may be more piquant than warthogs. away from the proper tragic pleasure to one more characteristic
The last chapter of the Poetics (or of the part of it that has sur­ of comedy (s3a3 3-5). But he rejects the critics' other premises
vived ) raises the question of whether epic or tragedy is superior. (62a5- 1 8) . Vulgarity of performance is not limited to actors; epic
The way this question is introduced suggests that Aristotle is taking recitations can also be exaggerated and vulgar.25 Conversely, per­
up an existing discussion, rather than proposing the question him­ formance is not essential to tragedy, since a tragedy can have its
self; this is confirmed when he summarizes the position of those effect when it is only read; so if the performance of a tragedy is
who criticize tragedy. But we know already that Aristotle does have vulgar, that is a fault in the performance but not a fault in the tra­
a view on this question. The history of poetry in chapter 4 treats gedy. Moreover, the potential for performance means that tragedy
drama as the fulfilment of the development of poetry: it is some­ has the elements of spectacle and song, giving it a range of additional
thing which the quasi-dramatic mode of Homer's epics adum­ pleasures which epic lacks.
brated, and when drama proper appeared it supplanted epic because Not all of Aristotle's arguments in this chapter are equally com­
of its intrinsically superior qualities (49a2-6) . Aristotle, then, will be pelling. He goes on to claim that tragedy is superior to epic because
firmly aligned with the supporters of tragedy. it is more concentrated; pleasure is better concentrated than spread
We have mentioned some of the criticisms which Plato makes in out thinly (62b 1-3 ) . The flaw in this argument is evident from the
the Republic of poetry in general, and of tragedy in particular. More example he gives: Sophocles' Oedipus would not be effective if ex­
immediately relevant here is the second book of the LAws, where panded to the length of the Iliad. But that is scarcely to the point. No
Plato develops the idea that the quality of a poetic genre is correl­ one would deny that the Oedipus at its present length is superior to
ated with the quality of the audience that appreciates it. Small chil­ the Oedipus expanded to the length of the Iliad; the real question is
dren like puppet-shows; older children like comedy; respectable whether the Oedipus at its present length is superior to the Iliad at its
adults like tragedy; but older men, mature and experienced, prefer present length. Expanding a play to the scale of an epic does not
epic (LAws, 65 7c-9c). The critics of tragedy to whom Aristotle re­ show that tragedy is better than epic; it shows only that tragedies
sponds make a similar assumption (6 1 b27-62�) : acting out a role is should not be expanded to the scale of an epic - which we know
vulgar, as the vulgar performances of actors and such-like show; already, since Aristotle explained in chapter 1 8 that tragedies con­
the only reason tragedy includes this vulgarity is to communicate tain less material (s6a rn- 1 5 ) . Furthermore, two passages elsewhere
effectively to the low-grade members of the audience, who could in the Poetics are difficult to .reconcile with Aristotle's assumption
not follow what is happening or respond to it without this kind of here that greater concentration is superior. In chapter 7 he said that,
exaggerated acting; hence tragedy is designed to appeal to the lowest within the upper limit oflength imposed by memory, the larger the
grade of audience. Without them, it would be possible to eliminate magnitude of the plot the better (5 I al of.); in chapter 24 he said that
the vulgarity of stage-performance; but then, of course, you would the larger scale of epic gave positive advantages over tragedy in
have a kind of epic. terms of its grandeur and variety: 'similarity quickly palls, and may
The premise that the quality of a poetic genre can be judged from cause tragedies to fail' (s9b28-3 1 ) . It is not clear, therefore, that
the quality of the audience that appreciates it is one which Aristotle Aristotle has managed to give convincing form to this argument, or
would have accepted in principle; in chapter 1 3 he said that weak to make it consistent with claims made elsewhere in the Poetics.

Ix !xi
INTRODUCTION INTRODUCTION

sometimes assumed that this expresses a preference for comedy in


1 2 . Comedy the more recent style; in fact it only implies that the more recent
style of comedy is a better guide to the standards of behaviour to be
The loss of the extended analysis ofcomedy which the original Poet­ observed in everyday life. We know from chapter 25 of the Poetics
ics probably contained makes it difficult to be sure what Aristotle's that 'correctness is not the same thing in ethics and poetry' (6ob 1 3 f.)
views on comedy would have been.26 The starting-point must be and that 'in evaluating any utterance or action, one must take into
the brief account in chapter 5 : 'Comedy is . . . an imitation of in­ account not just the moral qualities of what is actually done or said,
ferior people' (49a3 2f.) . As we have seen (§9), 'inferior' has both but also the identity of the agent or speaker' (6 1 a4-7); greater pro­
moral and social implications. The central figures of comedy will in­ priety is not self-evidently an advantage in a poetic genre devoted to
clude the lowly persons (such as peasants and slaves) who are only the imitation of inferior agents.27 Of course, Aristotle would have
peripheral in tragedy; and comic characters, even those of high disapproved of an excess of vituperation and obscenity in comic
status, will tend to behave badly. In tragedy, moral badness is abuse; but we have no way of telling what he would have judged
acceptable in characterization only to the extent that the structure excessive in a comedy.
of the plot demands it; but comedy deliberately incorporates moral A point which needs to be treated with some care in this connec­
badness in its characters, since comedy aims to evoke laughter and tion is the contrast which Aristotle makes between comedy with
'the laughable is a species of what is disgraceful' (49a3 3f.). The 'universalized stories and plots', introduced in Athens by Crates, and
response at which comedy aims is antithetical to that of tragedy. the older style of comedy in 'the form of a lampoon' (49b7-<)) . The
Tragedy aims to evoke pity and fear, which Aristotle defines as contrast of comedy and lampoon reappears in chapter 9, where the
reactions to painful and destructive harm (Rhetoric, 1 3 82a2 1 f. , concept ofuniversality is explained more fully (5 1 b 1 1 - 1 5 ) . It should
1 3 8 5 b 1 3f.). Hence tragedy involves suffering, which is 'an action not be read as a contrast between the abuse of named, real indi­
that involves destruction or pain' (52b 1 rf.); but comedy eschews it: viduals and ludicrous stories about fictitious characters. As we have
'the laughable is an error or disgrace that does not involve pain or seen (§5), universality in Aristotle's sense depends entirely on the
destruction' (49a34f.). If the story of Orestes were burlesqued in a structure of the poetic plot. If events are connected with each other
comedy, it would not end with Orestes killing Aegisthus but with in accordance with necessity or probability, they constitute a univer­
their reconciliation (53a3 5-<)) ; the abandonment of Orestes' obliga­ salized plot (5 1 b4-<)) ; and there is nothing to prevent real events, or
tion to avenge his father's murder would be a laughable disgrace. the actions ofreal individuals, satisfying this criterion in comedy any
One type of moral badness is a tendency to abuse others, and the more than in tragedy (cf. 5 1 b2!)--32). By 'the form of a lampoon',
characters of many extant Greek comedies indulge in liberal and therefore, Aristotle must mean a disjointed series ofjokes or comic
often obscene abuse both of each other and of real contemporaries. routines with no necessary or probable connection between them;
Aristotle took it for granted that comic poets would use slanderous comedy is better if it links its jokes and comic routines into a con­
and indecent language (Rhetoric, 1 3 84b!)-1 1 , Politics, I 3 46b I 3-23). In nected sequence.
the Nicomachean Ethics (I 1 28a22-5) _he notes a transition from the The requirements concerning plot-structure must be applied to
open indecency of older comedy to innuendo in more recent comedy with due regard to the pragmatism which we have already
comedy, and comments on the greater propriety of the latter. It is observed in Aristotle's treatment of tragedy and epic. Indeed, in

!xii !xiii
I NTRODUCTION I NTRODUCTION

comedy the scope for legitimate departure from strict connection in and J. Lallot's Aristote: La Poetique (Editions du Seuil, 1 980) contains
accordance with necessity and probability is greater, since dis­ an introduction, the Greek text with facing French translation, and
continuities in the action may have a comic effect. To allow such a commentary. The edition I have worked from is R. Kassel's
discontinuities in comic plots does not negate the contrast between Oxford Classical Text (Oxford, 1 965; reprinted with a commentary
the connectedness characteristic of comedy with universalized plots by D. W Lucas, Oxford, 1 968); but with such a difficult text there is
and the disconnected lampoon-like comedy which it superseded. often great uncertainty as to what Aristotle wrote, and l have freely
Where there is nothing but disjointed jokes or comic routines, dis­ departed from the readings printed by Kassel where this seemed
continuity has no comic effect; it is only in the context of a generally appropriate.
sustained impression of continuity that a particular violation of Humphrey House in Aristotle's Poetics (Hart-Davis, 1 956) gives
necessary and probable sequence will seem incongruous and laugh­ a short and accessible, though now somewhat dated, overview. The
able. But, again, we have no way of telling what Aristotle would only full-length systematic study in English is Stephen Halliwell's
have regarded as an excessive violation of connectedness in a Aristotle's Poetics (Duckworth, 1 986; this is not the same as his com­
comedy. mentary, mentioned above); chapter IO provides a good starting­
point for exploration of the history of interpretations of the Poetics
and of its influence. Halliwell's book has a very different vision of
I 3. Further reading Aristotle's project from mine; but it tends (as one reviewer put it) to
'float at an Olympian distance from the text', and the style makes it
This introduction began by mentioning some of the problems harder than it should be to get to grips with his arguments. By con­
which the Poetics poses to its interpreters; it scarcely needs to be said, trast, Elizabeth Belfiore's Tragic Pleasures: Aristotle on plot and emotion
therefore, that there are many and very diverse alternatives available (Princeton, 1 992), a book full of fresh ideas to think about and argue
to the account which I have sketched in the body of the with, is lucid and exceptionally stimulating. Anyone who has read
introduction. this book and Leon Golden's A ristotle on Tragic and Comic Mimesis
Among other translations which might be compared and con­ (Scholar Press, 1 992) will appreciate how radically contemporary
trasted with mine, I would recommend in particular those by Mar­ experts can disagree on the interpretation of even the most funda­
garet Hubbard, in Classical Literary Criticism, ed. D. Russell and M. mental concepts of the Poetics. The excellent collection Essays on
Winterbottom (Oxford, 1 972); Richard Janko's Aristotle: Poetics Aristotle's Poetics, edited by A. 0. Rorty (Princeton, 1 992), also illus­
(Hackett, 1 987) ; and Stephen Halliwell's The Poetics of Aristotle trates the diversity of approaches to the text currently on offer.
(Duckworth, 1 987) . Janko's version, which 'attempts to follow the For a short general introduction to Aristotle's philosophy see J. L.
original Greek closely, with minimal alterations for the sake of nat­ Ackrill, Aristotle the Philosopher (Oxford, 1 9 8 1 ) or Jonathan Barnes,
ural English', is especially suitable for close study, and has extensive Aristotle (Oxford, 1 982); at greater length, W K. C. Guthrie, Aristotle:
notes on points of detail.28 Halliwell's translation is equipped with a an Encounter, the sixth volume of Guthrie's History of Greek Philo­
more discursive commentary. sophy (Cambridge, 1 9 8 1 ) . Jonathan Lear's Aristotle: the desire to
Halliwell has also published an English translation with facing understand (Cambridge, 1 98 8) offers a philosophically challenging
Greek text in the Loeb Classical Library ( 1 995). R. Dupont-Roe's approach to Aristotle's thought; the introductory chapter, which

!xiv !xv
INTRODUCTION INTRODUCTION

explains the book's subtitle, is particularly relevant to this introduc­ column numbers in the margins. There are usually about 3 8 lines in
tion's starting-point. a Bekker column, although the number does vary.
For a broader perspective on ancient literary criticism I would References to Aristotle's other works are likewise given either by
recommend D. A. Russell's Criticism in Antiquity (Duckworth, 1 9 8 1 ) book and chapter, or by page, column and line in Bekker.
and The Cambridge History of Literary Criticism. Volume I : Classical
Criticism, ed. George A. Kennedy (Cambridge, 1989) .

1 4 . Reference conventions

Like many translators of the Poetics I have divided the text into
sections and sub-sections to help the reader follow the structure
of Aristotle's exposition. These divisions, and the accompanying
section-headings, are not part' of the transmitted text and have no
authoritative status. Nor do they correspond to any traditional or
commonly accepted way of dividing the text, so they are of no use in
giving references. For this purpose the reader should note the two
series of numbers in the margin of this translation, relating to the
two conventional ways of giving references in Aristotle.
One series (in roman type) runs from 1 to 26: these are the chap­
ters. The chapter divisions do not go back to Aristotle and are not
always very sensible, but they do provide one convenient and com­
monly accepted set of 'broad-brush' references.
The other series (in italics) runs from 47a to 62b: these refer to
pages and columns in the nineteenth-century edition ofAristotle by
Immanuel Bekker. Bekker's pages were wide enough to have two
columns; so each page is divided into columns a and b. Strictly
speaking, the Bekker pages for the Poetics run from 1 447a to 1 462b,
but the abbreviated form is often used when it is clear from the
context that the reference is to the Poetics rather than to one of
Aristotle's other works. It is possible to give references with great
precision by adding line-numbers to the Bekker pages and columns
(for example, Aristophanes is named in chapter 3 at 48a27); for this
reason modern editions conventionally give Bekker page and

!xvi
NOTES TO THE I NTRODUCTION

IO. Chapter I 2 interrupts the analysis of plot with a brief summary of the
standard parts of the text of a tragedy. It is awkwardly expressed, awkwardly
placed (the reference to 'what has just been said ' at the beginning of chapter
1 3 ignores it) and not very illuminating; some scholars have suggested that
N O T E S T O T H E
Aristotle did not write it, but that is probably wishful thinking.
I N T R O D U C T I O N
1 1 . Note that unity of action in the sense defined here is the only unity Aris­
totle is concerned with in the Poetics; the so-called unities of time and place
1 . In recent times what are probably additional fragments of On Poets have are later inventions.
been discovered on papyrus. 12. Our instinct when confronted with such a report would perhaps be to
2. Aristotle was born in 3 84 BC in Stagira in northern Greece; his father try to establish an implicit connection between events in order to make a
was doctor at the Macedonian court. He went to Athens to study in proper story out of it (was Ben perhaps Bill's bank-manager?). Aristotle's
Plato's Academy at the age of seventeen, and subsequently taught there. criteria for a well-formed plot correspond to our desire to find coherent, in­
He left Athens in 3 47, and in 343/2 was appointed tutor to Alexander telligible sequences in events; the effort that has been invested in blocking
Oater 'the Great') of Macedon. He returned to Athens in 3 3 5 to establish such inferences in some modern literature is evidence of the strength of that
his own philosophical school, the Lyceum. Anti-Macedonian feeling forced desire.
him to leave Athens again in 3 2 3 ; he died the following year. There is no 1 3 . It would be wrong to infer that Aristotle had no regard for history and
consensus as to when in Aristotle's career the Poetics was written. the recording of particular facts. A huge effort in collecting historical and
3. A briefexplanation of the conventions for referring to Aristotle's works is empirical data informs his works on (for example) politics and natural his­
given in § 1 4. tory. The study of particulars is a necessary precursor to the philosopher's
4. This was already a key concept in the discussion of poetry for Plato; see attempt to discern the universal patterns which lie behind them; but it
especially Books 3 and ro of the Republic (3 92d-8b, 595a--608b). is only a precursor, and the universalizing cast of mind characteristic of
5. For example, epic narrative is counted as imitation and historiographical philosophy has more in common with the way a poet must think in con­
narrative is not (if it were, versified history would be poetry, contrary to structing a well-made plot than with the procedures of a historian.
what is said at 5 1 b2-4); since Aristotle assumes the concept of imitation 14. In Rhetoric, 1 4 1 ob r of. Aristotle significantly qualifies the thesis that
without explanation or analysis, it is unclear what criterion underpins this learning is naturally pleasant: 'learning easily is naturally pleasant to all ' .
distinction and what the rationale of that criterion would be. 15. Katharsis i s such a controversial concept that this cannot b e taken for
6. Hence the references to prose in 47a28-b 1 3 ; for Aristotle, imitations in granted; experts on the Poetics continue to produce radically diverse inter­
verse and prose have more in common at the most fundamental level than pretations (see, for example, the books by Belfiore and Golden cited in § 1 3 ) ,
do imitative and non-imitative verse. and there i s n o consensus view.
7. The antithesis is, in one sense, false: human beings are by nature social 16. The 'final' and 'formal' causes are two elements in Aristotle's doctrine
(or, as Aristotle's dictum is most often cited, 'political') animals (Politics, of four causes (or, better, four kinds of explanation); see Physics, 2 . 3 .
l 2 5 3 a 1 - 1 8). 1 7 . I t might b e felt that tragedy must have a role i n its audience's moral
8. A similar point arises in chapter 9. Poetry expresses the universal, as a con­ development if it is to be taken with ultimate seriousness. But from an
sequence of its plot-structure (see §s) ; this does not mean that poetic forms Aristotelian point of view the goal of human existence is not moral
lacking plot-structure (such as lampoon) cannot be poetry, but they fall formation, but the exercise of formed moral character. To be ultimately
short of the ideal to which poetry aspires. serious, therefore, tragedy should have a place in the life of morally mature
9. That spoken language is rated more highly than song reflects the reduced individuals over and above any contribution it may make to their moral
importance of the chorus remarked in chapter 4 (49a 1 6- 1 8 ) . development.

!xviii lxix
NOTES TO THE INTRODUCTION NOTES TO THE INTRODUCTION

18. Even here, Aristotle's thoughts return to plot. The poet who is con­ can be seen in Politics, 1 3 3 6b3-2 3 , where he exempts comedy from a general
structing a plot in order to excite certain emotions in the audience is doing prohibition on slander but recommends that the audience be restricted to
something similar to what a speaker does to excite those emotions in his adults, who will be immune to its potentially harmful moral effects; younger
audience. Unlike the speaker, the poet does this without explicit statement; people, whose moral character is still in the process of formation, should be
he places a series of events in front of us, and leaves us to register for excluded.
ourselves the fact that it merits pity. But the kind of thing the poet puts 28.Janko includes the fragments of On Poets, and a hypothetical reconstruc­
before us is exactly the same as the kind of thing the speaker would put tion of the second book of the Poetics; the latter should be treated with
before us explicitly to excite pity; the same things are pitiful in each case. So caution (see n. 26).
even in constructing the plot, the poet can learn from rhetoric.
19. These chapters can usefully be read in conjunction with the discussion
of style in Book 3 of the Rhetoric (chapters r - r 2) .

20. Aristotle's example in chapter 17 is in fact based on Euripides' Iphigeneia


in Tauris, which he cites as often as Sophocles' Oedipus, although the
two plays are very different. It is a fundamental mistake to suppose that
Aristotle's theory of tragedy as a whole is modelled exclusively on the
Oedipus, or applicable primarily to plays of that kind.
2 1 . The completion of this plot is left as an exercise for the reader.
22. The Arabic translation of the Poetics (itself derived from a ninth-century
translation into Syriac) says that resolution should arise from 'character'; but
this is almost certainly. not what Aristotle wrote.
23. 'Episode' is an elusive term, used in a variety of ways in the Poetics. The
usage here, in which episodes may be derived from events outside the plot,
seems to be different from the one noted in chapter r 7 , in which episodes are
parts of the tragic plot in their most concrete realization ( 5 5 b r f. ; see § 1 0) .
24. This pragmatic accommodation o f general principle t o the needs of
particular cases is consistent with Aristotle's understanding of the relation
between tekhne and practice (see Metaphysics, 9 8 1 a 1 2-24) .
2 5 . The intensely emotional atmosphere o fa n epic recitation can b e gauged
from Plato's Ion (53 5b-e) .
26. Some have held that Aristotle's views on comedy can be recovered from
the Tractatus Coislinianus, a much later text derived (it has been argued ) from
an abridged version of the second book of the Poetics. This view has been
defended most forcefully in recent years by Richard Janko (see § 1 3) . But
few scholars currently accept this view of the tractate' s origins; fewer still
believe that in its present, severely mangled form much reliance can be
placed in it as a source for what Aristotle wrote.
27. The imitation of such agents and their behaviour does raise a concern
about its ethical effect on impressionable young people. Aristotle's response

lxx
SYNOPSIS OF THE P O E TICS

6.4 Recognition
6.5 Suffering
Chapter I 2 6.6 Quantitative parts of tragedy
S Y N O P S I S O F T H E P O E TI C S 7. T H E B E S T K I N D S OF TRAGIC PLOT

Chapter I 3 7 . I First introduction


7.2 First deduction
Chapter I I . INTRODUCTION Chapter I 4 7.3 Second introduction
2. P O E T RY A S A S P E C I E S O F I M I T AT I O N 7.4 Second deduction
2. 1 Medium 8. O T H E R A S P E C T S O F TRAG EDY

Chapter 2 2.2 Object Chapter I 5 8 . I Character


Chapter 3 2 . 3 Mode Chapter I 6 8.2 Kinds of recognition
3. T H E A N T H R O P O L O G Y A N D H I S T O RY Chapter 1 7 8 . 3 Visualizing the action
O F P O ET RY 8 . 4 Outlines and episodization
Chapter 4 3 . I Origins Chapter I 8 8 . 5 Complication and resolution
3 .2 Early history 8 .6 Kinds of tragedy
3 . 3 Tragedy 8 . 7 Tragedy and epic
Chapter 5 3 .4 Comedy 8 . 8 Astonishment
3 . 5 Epic 8.9 The chorus
4. T R A G E D Y : D E F I N I T I O N A N D A N A LY S I S 9. DICTION

Chapter 6 4. I Definition Chapter I 9 9 . I Introduction


4.2 Component parts Chapter 20 9.2 Basic concepts
4.3 The primacy of plot Chapter 2 I 9.3 Classification of nouns
4.4 The ranking completed Chapter 22 9.4 Qualities ofpoetic style
5. PLOT: BASIC CONCEPTS I O. E P I C

Chapter 7 5 . I Completeness Chapter 23 IO. I Plot


5 . 2 Magnitude Chapter 24 I 0.2 Kinds and parts of epic
Chapter 8 5 . 3 Unity 10.3 Differences between tragedy and epic
5 .4 Determinate structure 10.4 Quasi-dramatic epic
Chapter 9 5 . 5 Universality 1 0 . 5 Astonishment and irrationalities
5 .6 Defective plots 1 0.6 Diction
6. PLOT: SPECIES AND COMPONENTS I I . P R O B L E M S A N D S O L UT I O N S

6 . I Astonishment Chapter 2 5 I I . I . Principles


Chapter IO 6.2 Simple and complex plots I I .2 Applications
Chapter I I 6.3 Reversal I 1 .3 Conclusion

boll lxxi ii
SYNOPSIS OF THE P O E TI C S

12. C O M PA R A T I V E E VA L U A T I O N O F E P I C

A N D T R A G E DY

Chapter 26 1 2 . 1 The case against tragedy


1 2 . 2 Reply
13. CONCLUSION

Poetics
I . INTRODUCTION

Let us discuss the art of poetry in general and its species - the effect r

which each species of poetry has and the correct way to construct 47a

plots if the composition is to be of high quality, as well as the


number and nature of its component parts, and any other questions .
that arise within the same field of enquiry. We should begin, as is
natural, by taking first principles first.

2 . P O E T RY A S A S P E C I E S O F I M I TAT I O N

Epic poetry and the composition of tragedy, as well as comedy and


the arts of dithyrambic poetry and (for the most part) of music for
pipe or lyre, are all (taken together) imitations. 1 They can be differen­
tiated from each other in three respects: in respect of their different
media of imitation, or different objects, or a different mode (i.e. a differ­
ent manner) .

2 . 1 Medium

Some people use the medium of colour and shape to produce imita­
tions of various objects by making visual images (some through art,
some through practice) ; others do this by means of the voice.2 Sim­
ilarly in the case of the arts I have mentioned: in all of them the
medium of imitation is rhythm, language and melody, but these may
POETICS 2.J M 0DE

be employed either separately or in combination. For example,


music for pipe or lyre (and any other arts which have a similar effect, 2.2 Object
e.g. music for pan-pipes) uses melody and rhythm only, while dance
uses rhythm by itself and without melody (since dancers too imitate Those who imitate, imitate agents; and these must be either admir- 2

character, emotion and action by means of rhythm expressed in able or inferior. (Character almost always corresponds to j ust these 4sa
movement) . two categories, since everyone is differentiated in character by
The art which uses language unaccompanied, either in prose or in defect or excellence.) Alternatively they must be better people than
47b verse (either combining verse-forms with each other or using a we are, or worse, or of the same sort (compare painters: Polygnotus
single kind of verse), remains without a name to the present day. We portrayed better people, Pauson worse people, Dionysius people
have no general term referring to the mimes ofSophron and Xenar­ similar to us) .7 So it is clear that each of the kinds of imitation men­
chus and Socratic dialogues, 3 nor to any imitation that one might tioned above will exhibit these differences, and will be distinguished
produce using iambic trimeters, elegiac couplets or any other such by the imitation of distinct objects in this way. These dissimilarities
verse-form. Admittedly people attach 'poetry' to the name of the are possible in dance and in music for pipe or lyre, and also in con­
verse-form, and thus refer to 'elegiac poets' and 'hexameter poets'; nection with language and unaccompanied verse (for example,
i.e. they do not call people 'poets' because they produce imitations, Homer imitates better people; Cleophon people similar to us; He­
but indiscriminately on the basis of their use of verse. In fact, even if geman ofThasos, who invented parodies, or Nicochares, the author
someone publishes a medical or scientific text in verse, people are in of the Deiliad, worse people) ;8 similarly in connection with dithy­
the habit of applying the same term. But Homer and Empedocles rambs and nomes (one could imitate as Timotheus and Philoxenus
have nothing in common except the form of verse they use; so it did the Cyclopes) .9 The very same difference distinguishes tragedy
would be fair to call the former a poet, but the latter a natural scien­ and comedy from each other; the latter aims to imitate people worse
tist rather than a poet. 4 On the same principle, even if someone than our contemporaries, the former better.
should produce an imitation by combining all the verse-forms (as
Chaeremon composed his Centaur, which is a rhapsody combining
all the verse-forrns)5 he should still be termed a poet. So this is the 2 . 3 Mode
way distinctions are to be drawn in this area.
There are also some arts which use all the media mentioned A third difference between them is the mode in which one may imi­
above (i.e. rhythm, melody and verse) , e.g. dithyrambic and nomic tate each of these objects. It is possible to imitate the same objects in
poetry,6 tragedy and comedy; these differ in that the former use the same medium sometimes by narrating (either using a different
them all simultaneously, the latter in distinct parts. persona, as in Homer's poetry, '0 or as the same person without vari­
These, then, are what I mean by differences between the arts in ation) , or else with all the imitators as agents and engaged in activity.
the medium of imitation. So imitation can be differentiated in these three respects, as we
said at the outset: medium, object and mode. So in one respect
Sophocles would be the same kind of imitator as Homer, since both
imitate admirable people, but in another the same as Aristophanes,

4 5
POETICS J .2 E A R L Y H I S T 0 R Y

smce both imitate agents and people doing things. 1 1 This is the animal, and corpses) . The reason for this is that understanding is ex­
reason - some say - for the term 'drama': i.e. that the poets imitate tremely pleasant, not just for philosophers but for others too in the
people doing things. It is in consequence of this too that the Dorians same way, despite their limited capacity for it. This is the reason why
lay claim to tragedy and comedy. The Megarians lay claim to people take delight in seeing images; what happens is that as they
comedy - both those on the mainland (who allege that it arose in view them they come to understand and work out what each thing
the period of their democracy) , and those in Sicily (that being the is (e.g. 'This is so-and-so') . If one happens not to have seen the thing
birthplace of the poet Epicharmus, who was much earlier than before, it will not give pleasure as an imitation, but because of its
Chionides and Magnes) ; and some of the Peloponnesians lay claim execution or colour, or for some other reason.
to tragedy. They use the names as evidence. They say that they call Given, then, that imitation is natural to us, and also melody and
outlying villages komai, while Athenians call them demoi, the as­ rhythm (it being obvious that verse-forms are segments of ·

sumption being that comedians were so-called not from the revel or rhythm) , ' 3 from the beginning those who had the strongest natural
komos, but because they toured the villages when expelled from the inclination towards these things generated poetry out of improvised
4ab town in disgrace. And they say that they use the term dran for 'do', activities by a process of gradual innovation.
the Athenians prattein.
So much, then, for the number of ways in which imitation is
differentiated, and what they are. 3 .2 Early history

Poetry bifurcated in accordance with the corresponding kinds of


character: more serious-minded people imitated fine actions, i.e.
those of fine persons; more trivial people imitated those of inferior
3. THE ANTHROPOLOGY AND persons (the latter at first composing invectives, while the others
H I S T O RY O F P O E T RY composed hymns and encomia) . We are not in a position to identify a
poem of the latter kind'4 by any of the poets who preceded Homer,
3.1 Origins although they are likely to have been numerous; but beginning with
Homer we can do so (e.g. his Margites and similar poems). Because
4 In general, two causes seem likely to have given rise to the art of of its suitability, the iambic verse-form developed in these poems;
12
poetry, both of them natural. indeed, the reason it is now called 'iambic' is that they wrote lam­
Imitation comes naturally to human beings from childhood (and poons or iamboi against each other in that verse-form. And so some
in this they differ from other animals, i.e. in having a strong pro­ of the ancients became composers of heroic poetry, others of lam­
pensity to imitation and in learning their earliest lessons through poons. But just as Homer was the outstanding poet of the serious
imitation); so does the universal pleasure in imitations. What hap­ kind, since he did not just compose well but also made his imitations
pens in practice is evidence of this: we take delight in viewing the dramatic, 1 5 so too he was the first to adumbrate the form of comedy;
most accurate possible images of objects which in themselves cause what he composed was not an invective, but a dramatization of the
distress when we see them (e.g. the shapes of the lowest species of laughable. His Margites stands in the same relation to comedy as the

6 7
POETICS 3.5 EPIC

49a fliad and Odyssey do to tragedy. When tragedy and comedy made orated may be taken as read; it would probably be a major under­
their appearance those who inclined towards either kind of poetry taking to go through them all individually.
became, in accordance with their nature, poets of comedy (instead
of lampoons) or of tragedy (instead of epic), because these forms
were greater and more highly esteemed than the others. 3 . 4 Comedy

Comedy is (as we have said) an imitation of inferior people - not,


3 . 3 Tragedy however, with respect to every kind of defect: the laughable is a spe­
cies of what is disgraceful. The laughable is an error or disgrace that
This is not the place for a detailed investigation of whether or not does not involve pain or destruction; for example, a comic mask is .
tragedy is now sufficiently developed with respect to its formal con­ ugly and distorted, but does not involve pain. '9
stituents (judged both in its own right and in relation to theatrical The transformations which tragedy has undergone, and those re­
performances) . But originally it developed from improvisations. sponsible for them, have not been forgotten; but, because it was not
(This is true of tragedy, and also of comedy: the former arose from taken seriously, little attention was paid to comedy at first. Indeed, it 49b
the leaders of the dithyramb, the latter from the leaders of the phallic was relatively late that the archon first granted a comic chorus;20
songs which are still customary even now in many cities.) 1 6 Then before that the performers were volunteers. So comedy already had
tragedy was gradually enhanced as people developed each new some of its features before there is any mention of those identified as
aspect of it that came to light. After undergoing many transforma­ comic poets, and it is not known who introduced masks, prologues,
tions tragedy came to rest, because it had attained its natural state. the number of actors and so forth. But plot-construction came ori­
The number of actors was increased from one to two by Aeschy­ ginally from Sicily; among Athenian poets it was Crates who first
lus, who also reduced the choral parts and made the spoken word abandoned the form of a lampoon and began to construct universal­
21
play the leading role; the third actor and scene-painting were intro­ ized stories and plots.
duced by Sophocles. In addition, the magnitude increased from
short plots; and in place of comic diction, as a consequence of a
change from the satyric style, '7 tragedy acquired dignity at a late 3 . 5 Epic
stage, and the iambic verse-form was adopted instead of the trochaic
tetrameter. (They used tetrameters at first because the composition Epic poetry corresponds to tragedy in so far as it is an imitation in
was satyric in manner, and more akin to dance. But when speech verse of admirable people. But they differ in that epic uses one
was introduced nature itself found the appropriate form of verse, verse-form alone, and is narrative. They also differ in length, since
iambic being the verse-form closest to speech. There is evidence of tragedy tries so far as possible to keep within a single day, or not to
this: we speak iambics in conversation with each other very often, exceed it by much, whereas epic is unrestricted in time, and differs
but rarely dactylic hexameters - and only when we depart from the in this respect. (At first, however, people used to make no distinction
normal conversational tone.) 1 8 As for the number of episodes and between tragedy and epic in this respect.)
other such features, the way each of them is said to have been elab- Some of the component parts are common to both, others are

8 9
POETICS 4.3 THE PRIMACY 0 F PL0T

peculiar to tragedy. Consequently anyone who understands what is actual composition of the verse; what is meant by 'lyric poetry' is
good and bad in tragedy also understands about epic, since anything self-evident.)
that epic poetry has is also present in tragedy, but what is present in Now, tragedy is an imitation of an action, and the action is per­
tragedy is not all in epic poetry. formed by certain agents. These must be people of a certain kind
with respect to their character and reasoning. (It is on the basis of
people's character and reasoning that we say that their actions are of 5oa

a certain kind, and in respect of their actions that people enjoy suc­
cess or failure.) So plot is the imitation of the action (by 'plot' here I
4. TRAGEDY: DEFINITION mean the organization of events) ; character is that in respect of which
A N D A N A LY S I S we say that the agent is of a certain kind; and reasoning is the speech
which the agents use to argue a case or put forward an opinion.
4. I Definition So tragedy as a whole necessarily has six component parts, which
determine the tragedy's quality: i.e. plot, character, diction, reason­
6 We shall discuss the art of imitation in hexameter verse and com­ ing, spectacle and lyric poetry. The medium of imitation comprises
edy later;22 as for tragedy, let us resume the discussion by stating two parts, the mode one, and the object three; and there is nothing
the definition of its essence on the basis of what has already been apart from these.
said.
Tragedy is an imitation of an action that is admirable, complete
and possesses magnitude; in language made pleasurable, each of its 4.3 The primacy ofplot
species separated in different parts; performed by actors, not through
narration; effecting through pity and fear the purification23 of such Virtually all tragedians, one might say, use these formal elements; for
emotions. in fact every drama alike has spectacle, character, plot, diction, song
(By 'language made pleasurable' I mean that which possesses and reasoning.24 But the most important of them is the structure of
rhythm and melody, i.e. song. By the separation of its species I mean the events:
that some parts are composed in verse alone; others by contrast
(i) Tragedy is not an imitation of persons, but of actions and of life.
make use of song.)
Well-being and ill-being reside in action, and the goal oflife is an
activity, not a quality; people possess certain qualities in accord­
ance with their character, but they achieve well-being or its op­
4.2 Component parts
posite on the basis of how they fare. So the imitation of character
is not the purpose of what the agents do; character is included
Since the imitation is performed by actors, it follows first of all that
along with and on account of the actions. So the events, i.e. the
the management of the spectacle must be a component part of tra­
plot, are what tragedy is there for, and that is the most important
gedy. Then there is lyric poetry and diction, since these are the medium
thing of all.
in which the actors perform the imitation. (By 'diction' I mean the

IO II
POETICS 5.1 COMPLETENESS

(ii) Furthermore, there could not be a tragedy without action, but something is or is not the case, or put forward some universal
there could be one without character. The tragedies of most proposition.
modern poets lack character, and in general there are many such Fourth is diction. By 'diction' I mean, as was said before, verbal
poets. Compare, among painters, the relation between Zeuxis and expression; this has the same effect both in verse and in prose
Polygnotus: the latter is good at depicting character, but Zeuxis' speeches.
painting has no character.25 Of the remaining parts, song is the most important of the sources
(iii) Also, if one were to compose a series of speeches expressive of of pleasure. Spectacle is attractive, but is very inartistic and is least
character, however successful they are in terms of diction and rea­ germane to the art of poetry. For the effect of tragedy is not de­
soning, it will not achieve the stated function of tragedy; a tragedy pendent on performance and actors; also, the art of the property­
which, though it uses these elements less adequately, has a plot manager has more relevance to the production of visual effects than
and a structure of events will do so much more effectively. does that of the poets.
(iv) Additionally, the most important devices by which tragedy sways
emotion are parts of the plot, i . e. reversals and recognitions. 26
(v) A further indication is that those who are trying to write poetry
are capable of accuracy in diction and character before they can
construct the events; compare too almost all the early poets. 5. P LO T : BA S I C C O N C E P T S

4.4 The ranking completed Given these definitions, let us discuss next what qualities the struc- 7
ture of the events should have, since this is the first and most import­
So the plot is the source and (as it were) the soul of tragedy; char- ant part of tragedy.
5ob acter is second. (It is much the same in the case of painting: if some­
one were to apply exquisitely beautiful colours at random he would
give less pleasure than if he had outlined an image in black and 5 . I Completeness
white.) Tragedy is an imitation of an action, and on account above
all of the action it is an imitation of agents. We have laid down that tragedy is an imitation of a complete, i.e.
Third is reasoning. This is the ability to say what is implicit in a whole, action, possessing a certain magnitude. (There is such a thing
situation and appropriate to it, which in prose is the function of the as a whole which possesses no magnitude.) A whole is that which has
arts of statesmanship and of rhetoric. Older poets used to make a beginning, a middle and an end. A beginning is that which itself
people speak like statesmen; contemporary poets make them speak does not follow necessarily from anything else, but some second
rhetorically.27 Character is the kind of thing which discloses the thing naturally exists or occurs after it. Conversely, an end is that
nature of a choice; for this reason speeches in which there is which does itself naturally follow from something else, either neces­
nothing at all which the speaker chooses or avoids do not possess sarily or in general, but there is nothing else after it. A middle is that
character. Reasoning refers to the means by which people argue that which itself comes after something else, and some other thing comes

12 13
POETICS 5.4 DETERMI N ATE STRUCTURE

after it. Well-constructed plots should therefore not begin or end at


any arbitrary point, but should employ the stated forms. 5 . 3 Unity

A plot is not (as some think) unified because it is concerned with a


5 .2 Magnitude single person. An indeterminately large number of things happen to
any one person, not all of which constitute a unity; likewise a single

Any beautiful object, whether a living organism or any other entity individual performs many actions, and they do not make up a single

composed of parts, must not only possess those parts in proper order, action. So it is clear that a mistake has been made by all those poets

but its magnitude also should not be arbitrary; beauty consists in who have composed a Heracleid or Theseid, or poems of that kind, on
magnitude as well as order. For this reason no organism could be the assumption that,just because Heracles was one person, the plot .

beautiful if it is excessively small (since observation becomes con­ too is bound to be unified.Just as Homer excels in other respects, he

fused as it comes close to having no perceptible duration in time) or seems to have seen this point clearly as well, whether through art or

5 1 a excessively large (since the observation is then not simultaneous, and instinct. When he composed the Odyssey he did not include every­

the observers find that the sense of unity and wholeness is lost from thing which happened to Odysseus (e.g. the wounding on Parnassus

their observation, e.g. if there were an animal a thousand miles long) . and the pretence of madness during the mobilization: the occur­

So just as in the case of physical objects and living organisms, they rence of either of these events did not make the occurrence of the

should possess a certain magnitude, and this should be such as can other necessary or probable); 29 instead, he constructed the Odyssey
readily be taken in at one view, so in the case of plots: they should about a single action of the kind we are discussing. The same is true

have a certain length, and this should be such as can readily be held of the fliad.
in memory.
The definition of length which is determined by theatrical per­
formances and perception is not relevant to the art of poetry; if it 5 .4 Determinate structure
were necessary to perform a hundred tragedies they would time the
performances by the clock, as they say used to be done on other oc­ Just as in other imitative arts the imitation is unified if it imitates a

casions. 2 8 But the definition which agrees with the actual nature of single object, so too the plot, as the imitation of an action, should

the matter is that invariably the greater the plot is (up to the limits of imitate a single, unified action - and one that is also a whole. So the

simultaneous perspicuity) the more beautiful it is with respect to structure of the various sections of the events must be such that the

magnitude; or, to state a straightforward definition, 'the magnitude transposition or removal of any one section dislocates and changes

in which a series of events occurring sequentially in accordance the whole. If the presence or absence of something has no discern­

with probability or necessity gives rise to a change from good for­ ible effect, it is not a part of the whole.

tune to bad fortune, or from bad fortune to good fortune', is an


adequate definition of magnitude.

15
POET I C S 6. I A S T O N I S H M E N T

plots rather than of verses, insofar as he is a poet with respect to imi­


5. 5 Universality tation, and the object of his imitation is action. Even if in fact he
writes about what has happened, he is none the less a poet; there is
9 It is also clear from what has been said that the function of the poet nothing to prevent some of the things which have happened from
is not to say what has happened, but to say the kind of thing that being the kind of thing which probably would happen, and it is in
would happen, i.e. what is possible in accordance with probability or that respect that he is concerned with them as a poet.
51b necessity. The historian and the poet are not distinguished by their
use of verse or prose; it would be possible to turn the works of
Herodotus into verse, and it would be a history in verse just as much 5.6 Defective plots
as in prose. The distinction is this: the one says what has happened,
the other the kind of thing that would happen. JO Of simple plotsll and actions, the episodic ones are the worst. By an
For this reason poetry is more philosophical and more serious episodic plot I mean one in which the sequence of episodes is neither
than history. Poetry tends to express universals, and history particu­ necessary nor probable. Second-rate poets compose plots of this
lars. The universal is the kind of speech or action which is consonant kind of their own accord; good poets do so on account of the actors
with a person of a given kind in accordance with probability or - in writing pieces for competitive displayJ4 they draw out the plot
necessity; this is what poetry aims at, even though it applies indi­ beyond its potential, and are often forced to distort the sequence. 52a
vidual names. The particular is the actions or experiences of (e.g.)
Alcibiades.
In the case of comedy this is in fact clear. The poets construct the
plot on the basis of probabilities, and then supply names of their own
choosing; they do not write about a particular individual, as the 6. P LO T : S P E C I E S A N D C O M P O N E N T S
lampoonists do. In the case of tragedy they do keep to actual names.
The reason for this is that what is possible is plausible; we are dis­ 6. I Astonishment
inclined to believe that what has not happened is possible, but it is
obvious that what has happened is possible - because it would not The imitation is not just of a complete action, but also of events that
have happened if it were not. To be sure, even in tragedy in some evoke fear and pity. These effects occur above all when things come
cases only one or two of the names are familiar, while the rest are in­ about contrary to expectation but because of one another. This will
vented, and in some none at all , e.g. in Agathon's Antheus;J' in this be more astonishing than if they come about spontaneously or by
play both the events and the names are invented, but it gives no less chance, since even chance events are found most astonishing when
pleasure. So one need not try at all costs to keep to the traditional they appear to have happened as if for a purpose - as, for example,
stories which are the subjects of tragedy; in fact, it would be absurd the statue of Mitys in Argos killed the man who was responsible for
to do so, since even what is familiar is familiar only to a few, and yet Mitys' death by falling on top of him as he was looking at it.JS
gives pleasure to everyone. Things like that are not thought to occur at random. So inevitably
So it is clear from these points that the poet must be a makerJ2 of plots of this kind will be better.

16 17
POETICS 6.6 Q U A N T I T A T I V E P A R T S 0 F T R A G E D Y

is best when it occurs simultaneously with a reversal, like the one in


6.2 Simple and complex plots the Oedipus.
There are indeed other kinds of recognition. Recognition can
ro Some plots are simple, others complex, since the actions of which come about in the manner stated with respect to inanimate and
the plots are imitations are themselves also of these two kinds. By a chance objects; and it is also possible to recognize whether someone
simple action I mean one which is, in the sense defined, continuous has or has not performed some action. But the one that has most to
and unified, and in which the change of fortune comes about with­ do with the plot and most to do with the action is the one I have
out reversal or recognition. By complex, I mean one in which the mentioned. For a recognition and reversal of that kind will involve
change of fortune involves reversal or recognition or both. These pity or fear, and it is a basic premise that tragedy is an imitation of ac- 52b
must arise from the actual structure of the plot, so that they come tions of this kind. Moreover, bad fortune or good fortune will be the ·

about as a result of what has happened before, out of necessity or in outcome in such cases.
accordance with probability. There is an important difference be­ Since the recognition is a recognition of some person or persons,
tween a set of events happening because of certain other events and some involve the recognition of one person only on the part of the
efter certain other events. other, when it is clear who the other is; but sometimes there must be
a recognition on both sides (e.g. Iphigeneia is recognized by Orestes
from the sending of the letter, but the recognition of Orestes by
6 . 3 Reversal I phigeneia had to be different) . 39

rr A reversal is a change to the opposite in the actions being performed,


as stated - and this, as we have been saying, in accordance with prob­ 6. 5 Suffering
ability or necessity. For example, in the Oedipus someone came to
give Oedipus good news and free him from his fear with regard to So there are these two parts of the plot - reversal and recogni­
his mother, but by disclosing Oedipus' identity he brought about tion; a third is suffering. Of these, reversal and recognition have
the opposite result;36 and in the Lynceus, Lynceus himself was being already been discussed; suffering is an action that involves destruction
led off to be killed, with Dana us following to kill him, but it came or pain (e.g. deaths in full view, extreme agony, woundings and
about as a consequence of preceding events that the latter was killed so on) .
and Lynceus was saved. 37

6.6 Quantitative parts of tragedy


6.4 Recognition
We have already mentioned the component parts of tragedy which 1 2
Recognition, as in fact the term indicates, is a change from ignorance should b e regarded as its formal elements. In quantitative terms, the
to knowledge, disclosing either a close relationship38 or enmity, on separate parts into which it is divided are as follows: prologue; epi­
the part of people marked out for good or bad fortune. Recognition sode; finale; choral parts, comprising entry-song and ode - these are

18 19
POETICS 7.2 Fl RST DEDUCTl0N

common to all tragedies, while songs from the stage and dirges are undergoing a change from bad fortune to good fortune - this is
found only in some. the least tragic of all: it has none of the right effects, since it is neither
The prologue is the whole part of a tragedy before the entry-song agreeable, nor does it evoke pity or fear. Nor again should a very 53a
of the chorus; an episode is a whole part of a tragedy between whole wicked person fall from good fortune to bad fortune - that kind of
choral songs; the finale is the whole part of a tragedy after which structure would be agreeable, but would not excite pity or fear, since
there is no choral song. Of the choral part, the entry-song is the first the one has to do with someone who is suffering undeservedly,
whole utterance of a chorus; an ode is a choral song without the other with someone who is like ourselves (I mean, pity has to do
anapaests or trochaics; a dirge is a lament shared by the chorus and with the undeserving sufferer, fear with the person like us); so what
from the stage. happens will evoke neither pity nor fear.
We have already mentioned the component parts of tragedy We are left, therefore, with the person intermediate between .
which should be regarded as its formal elements. In quantitative these. This is the sort of person who is not outstanding in moral ex­
terms, the separate parts into which it is divided are these.40 cellence or justice; on the other hand, the change to bad fortune
which he undergoes is not due to any moral defect or depravity, but
to an error4' of some kind. He is one of those people who are held in
great esteem and enjoy great good fortune, like Oedipus, Thyestes,
and distinguished men from that kind offamily.
7 . THE BEST KINDS OF TRAGIC PLOT It follows that a well-formed plot will be simple42 rather than (as
some people say) double, and that it must involve a change not to
7. r First introduction good fortune from bad fortune, but (on the contrary) from good for­
tune to bad fortune - and this must be due not to depravity but to a
1 3 What, then, should one aim at and what should one avoid in con­ serious error on the part of someone of the kind specified (or better
structing plots? What is the source of the effect at which tragedy than that, rather than worse) . There is evidence of this in practice.
aims? These are the topics which would naturally follow on from At first poets used to pick out stories at random; but nowadays the
what has just been said. best tragedies are constructed around a few households, e.g. about
Alcmeon, Oedipus, Orestes, Meleager, Thyestes, Telephus and any
others whose lot it has been to experience something terrible - or
7 .2 First deduction to perform some terrible action.43
So the best tragedy, in artistic terms, is based on this struc­
The construction of the best tragedy should be complex rather than ture. This is why those who criticize Euripides for doing this in his
simple; and it should also be an imitation of events that evoke fear tragedies, most of which end in bad fortune, are making the same
and pity, since that is the distinctive feature of this kind of imitation. mistake;44 for this is, as has been stated, correct. There is very
So it is clear first of all that decent men should not be seen under­ powerful evidence for this. On stage and in performance people
going a change from good fortune to bad fortune - this does not recognize that plays of this kind (provided that they are successfully
evoke fear or pity, but disgust. Nor should depraved people be seen executed ) are the most tragic, and Euripides, even if his technique is

20 2!
POETICS 7.4 SEC0N D DEDUCTI0N

faulty in other respects, is regarded as the most tragic of poets. Necessarily, we are concerned with interactions between people
Second-best is the structure which some say comes first - that who are closely connected47 with each other, or between enemies,
which has a double structure like the Odyssey, and which ends with or between neutrals. If enemy acts on enemy, there is nothing piti­
the opposite outcome for better and worse people.45 It is thought to able either in the action itself or in its imminence, except in respect
come first because of the weakness of audiences; the poets follow of the actual suffering in itself. Likewise with neutrals. What one
the audiences' lead and compose whatever is to their taste. But this is should look for are situations in which sufferings arise within close
not the pleasure which comes from tragedy; it is more characteristic relationships, e.g. brother kills brother, son father, mother son, or son
of comedy. In comedy even people who are the bitterest enemies in mother - or is on the verge of killing them, or does something else
the story, like Orestes and Aegisthus, go off reconciled in the end, of the same kind.
and no one gets killed by anybody.46 Now, one cannot undo traditional stories (I mean, for ex- ·

ample, Clytaernnestra's death at Orestes' hands, or Eriphyle's at


Alcmeon's);48 but one has to discover for oneself how to use even
7. 3 Second introduction the traditional stories well. Let us state more clearly what this in­
volves. I t is possible for the action to come about in the way that the
14 It is possible for the evocation of fear and pity to result from the old poets used to do it, with people acting in full knowledge and
53b spectacle, and also from the structure of the events itself. The latter is awareness; this is in fact how Euripides portrayed Medea killing her
preferable and is the mark of a better poet. The plot should be con­ children.49 It is also possible for the action to be performed, but for
structed in such a way that, even without seeing it, anyone who the agents to do the terrible deed in ignorance and only then to recog­
hears the events which occur shudders and feels pity at what hap­ nize the close connection, as in Sophocles' Oedipus. (This is outside
pens; this is how someone would react on hearing the plot of the the play: examples in the tragedy itself are Astydamas' Alcmeon or
Oedipus. Producing this effect through spectacle is less artistic, � nd is Telegonus in the Odysseus Wounded.)50 A third possibility besides these
dependent on the production. Those who use spectacle to produce is for someone to be on the verge of performing some irreparable
an effect which is not evocative of fear, but simply monstrous, have deed through ignorance, and for the recognition to pre-empt the act.
nothing to do with tragedy; one should not seek every pleasure from Besides these there is no other possibility: necessarily the agents
tragedy, but the one that is characteristic of it. And since the poet must either act or not act, either knowingly or in ignorance.5'
should produce the pleasure which comes from pity and fear, and Of these, being on the verge of acting wittingly and not doing so
should do so by means of imitation, clearly this must be brought is worst; this is disgusting, and is not tragic since there is no suffering.
about in the events. So no one composes in this way, or only rarely (e.g. Haemon and 54a
Creon in the Antigone) .52 Performing the action is second; but it is
better if the action is performed in ignorance and followed by a re­
7 .4 Second deduction cognition - there is nothing disgusting in this, and the recognition
has great emotional impact. But the last case is best; I mean, for ex­
Let us therefore take up the question of what classes of events appear ample, in the Cresphontes Merope is on the verge of killing her son
terrible or pitiable. but does not do it, but instead recognizes him;53 the same happens

22 23
POETICS 8. I CHARACTER

with sister and brother in the Iphigeneia;54 and in the Helle the son lament in the Scylla6' and Melanippe's speech.62 An example of in­

recognizes his mother when on the verge of handing her over.5 5 consistency is the Iphigeneia in Aulis:63 when she pleads for her life to

For this reason, a s I said some time ago, 5 6 tragedies are concerned be spared she is not at all like her later self - but in characterization,

with a limited number of families. Although their search was guided just as much as in the structure of events, one ought always to look

by chance rather than art, poets discovered how to produce this kind for what is necessary or probable: it should be necessary or probable

of effect in plots; so they are forced to turn to just those households that this kind of person says or does this kind of thing, and it should

in which this kind of suffering has come about. be necessary or probable that this happens after that.
(Clearly, therefore, the resolutions of plots should also come
about from the plot itself, and not by means of a theatrical device, as 54b
in the Medea, or the events concerned with the launching of the
ships in the fliad.64 A theatrical device may be used for things outside

8. O T H E R A S P E C T S O F T R A G E DY the play - whether prior events which are beyond human know­
ledge, or subsequent events which need prediction and narration -

8. r Character since we grant that the gods can see everything. But there should be
nothing irrational in the events themselves; or, failing that, it should

Enough has been said about the structure of events and what plots be outside the play, as for example in Sophocles' Oedipus.)65

15 should be like; as for character, there are four things to aim at: Since tragedy is an imitation of people better than we are, one
should imitate good portrait-painters. In rendering the individual
(i) First and foremost, goodness. As was said earlier, speech or action
form, they paint people as they are, but make them better-looking.
will possess character if it discloses the nature of a deliberate In the same way the poet who is imitating people who are irascible
choice; the character is good if the choice is good. This is possible
or lazy or who have other traits of character of that sort should por­
in each class of person: there is such a thing as a good woman and
tray them as having these characteristics, but also as decent people.
a good slave, even though one of these is perhaps deficient and the
For example, Homer portrayed Achilles as both a good man and a
other generally speaking inferior.57
paradigm of obstinacy.66
(ii) Secondly, appropriateness: it is possible for the character to be
One should observe these points closely, and in addition those
courageous, but for this to be an inappropriate way for a woman
corresponding to the perceptions that are necessary concomitants of
to display courage or cleverness. 5 8
the art of poetry. It is possible to make many mistakes with respect
(iii) Thirdly, likeness: this i s not the same a s making character good
to these. But they have been discussed in sufficient detail in my
and appropriate, as has already been stated.59
published works.67
(iv) Fourthly, consistency: even if the subject of the imitation is incon­
sistent, and that is the kind of character that is presupposed, it
should nevertheless be consistently inconsistent.

An example of unnecessary badness of character is Menelaus in


the Orestes;00 of impropriety and inappropriateness, Odysseus'

25
24
POETICS 8.3 V l S UA L l Z I N G T H E! ACTl0N

probable for Orestes to infer that his sister had been sacrificed
;
8 . 2 Kinds of recognition and so it was now his turn to be sacrificed. Also in Theodectes
Tydeus, that he came to find a son, but is perishing himself. And
16 We have already said what recognition is. Its kinds are: the recognition in the Sons of Phineus; when the women saw the
place they inferred that it was their fate to die there, since that was
(i) First of all, the least artistic kind (and the one which people use
_ where they had been exposed. 76
most, because of their lack of ingenuity) is that by means of
(v) There is also a composite kind arising from a false inference
tokens. Some of these are congenital (e.g. 'the spear the earth­
on the part of the audience. For example, in Odysseus the False
born bear',68 or stars such as Carcinus used in his Thyestes),69 and
Messenger: 77 the fact that he can bend the bow and nobody else is
some are acquired; of the latter, some are physical characteristics
contrived by the poet as a premise, as is his claim that he will
(e.g. scars), others are external (e.g. necklaces, or the use of the
recognize the bow which he has not seen; and although he is
boat in the Tyro) .70 It is possible to make better or worse use of
going to make himself known by means of the former, he actually
these. For example, Odysseus is recognized by means of the scar
does so by means of the latter, which involves a false inference.
both by the nurse and by the swineherds, but in different ways.
(vi) The best recognition of all is that which arises out of the actual
Recognitions that are used only for confirmation are less artistic
course of events, where the emotional impact is achieved through
(so too all recognitions of that kind ) ; recognitions which arise out
events that are probable, as in Sophocles' Oedipus and the Iphige­
of a reversal, as in the bath-scene, are better. 7'
neia (her wish to send a letter is probable) . Qnly this kind does
(ii) Second are those which are contrived by the poet; for that reason
without contrived tokens and necklaces. Second-best are those
they are inartistic. For example, Orestes in the Iphigeneia revealed
which arise from inference.
his own identity; Iphigeneia's identity is revealed by the letter, but
Orestes declares in person what the poet (instead of the plot) re­
quires. This brings it close to the error mentioned above: it would
8. 3 Visualizing the action
have been possible actually to bring tokens with him.7' There is
also the 'voice of the shuttle' in Sophocles' Tereus. 73
When constructing plots and working them out complete with r 7
(iii) The third is by means of memory, when someone grasps the sig-
their linguistic expression, one should so far as possible visualize
55a nificance of something that he sees. This is how it is in Dicaeo­
what is happening. By envisaging things very vividly in this way, as if
genes' Cyprians, where he sees the painting and bursts into tears,
one were actually present at the events themselves, one can find out
and in the tale told to Alcinous, where Odysseus listens to the
what is appropriate, and inconsistencies are least likely to be over­
lyre-player, is reminded of his past and weeps; recognition results
looked. The criticism made of Carcinus provides evidence of this:
in both cases. 74
Arnphiaraus was corning back from the temple; this would have
(iv) Fourth is that which arises from inference. For example, in the
escaped the notice of anyone who did not see it, but it failed in
Choephori: 'someone similar has come; no one is similar except
performance because the audience was dissatisfied with it. 78
Orestes; so he has come' .75 There is also the recognition which
One should also, as far as possible, work plots out using gestures.
Polyidus the sophist suggested for Iphigeneia; he said that it was
Given the same natural talent, those who are actually experiencing

27
POETICS 8.6 K I N D S 0F TRAGE0Y

the emotions are the most convincing; someone who is distressed or reaches home, reveals his identity to a number of people and attacks.
angry acts out distress and irritation most authentically. (This is why He survives and destroys his enemies.' That much is integral; the rest
the art of poetry belongs to people who are naturally gifted or mad; is episodes.
of these, the former are adaptable, and the latter are not in their right
mind.)79
8.5 Complication and resolution

8.4 Outlines and episodization Every tragedy consists of a complication and a resolution. What is 1 8
outside the play, and often some o f what i s inside, comprises the
Stories, even ones which have been the subject of a previous poem, complication; the resolution is the rest. By complication I mean every­
55b should first be set out in universal terms when one is making use of thing from the beginning up to and including the section which
them oneself; on that basis, one should then turn the story into epi­ immediately precedes the change to good fortune or bad fortune; by
sodes and elaborate it. resolution I mean everything from the beginning of the change of
As an example of what I mean by considering the universal, take fortune to the end. Thus in Theodectes' Lynceus the complication
the Iphigeneia: 'A girl has been sacrificed and has disappeared with­ consists of events before the play, the seizure of the child and the dis­
out those who performed the sacrifice being aware of it. Set down closure of the parents; the resolution is everything from the capital
in another country, where it was the custom to sacrifice foreigners charge to the end. 8 2
to the goddess, she becomes the priestess of this rite. It subsequently
happens that the priestess's brother arrives (the fact that the god or­
dered him to go there is outside the universal; so too the reason); 80 8.6 Kinds of tragedy
on his arrival he is captured, but when he is on the verge of being
sacrificed he discloses his identity (either as Euripides did it, or as in There are four kinds of tragedy (since that was also the number of
Polyidus, 8 ' by saying - as was quite probable - that it was his lot, as component parts mentioned ) : 83 complex tragedy, depending en­
well as his sister's, to be sacrificed ) . Escape ensues. ' After that, one tirely on reversal and recognition; tragedy of suffering (e.g. plays
should supply the names and turn the story into episodes. The epi­ about Ajax or Ixion); tragedy of character (e.g. J:.tOmen ef Phthia 56a
sodes must be appropriate - for example, in the case of Orestes the and Peleus) ; and, fourth, simple tragedy (e.g. Daughters ef Phorcys,
fit of madness which resultt:d in his capture, and the escape by means Prometheus and plays set in the underworld ) .
of the purification. B y preference one should try t o include all the component parts,
In plays the episodes are concise, but in epic poetry they are used or failing that most of them and the most important, especially given
to increase the length. The story of the Odyssey is not very long: 'A the captious criticisms which people make of poets nowadays. Be­
man has been away from home for many years; he is kept under cause there have been poets good at each part, people expect indi­
close observation by Poseidon, and is alone; at home affairs are in vidual poets to surpass the particular excellence of every one. 84
such a state that his property is being squandered by the suitors, and The proper basis for contrasting and comparing tragedies is
plots are being laid against his son. Despite being shipwrecked he principally in virtue of the plot, i.e. whether the complication and

28 29
POETICS 9.1 INTRODUCTION

resolution are the same. Many poets are good at complication but more to do with the plot than they do with any other play; this is the
handle the resolution badly; 8 5 but both should be treated with equal reason why they sing interludes. This is a practice which Agathon
care. was the first to start; 89 but what is the difference between singing
interludes and transferring a speech or a whole episode from one
play into another?
8. 7 Tragedy and epic

Bearing in mind what I have already said several times, one should
not compose a tragedy out of a body of material which would serve
for an epic - by which I mean one that contains a multiplicity of 9. DICTION

stories (for example, if one were to use the whole plot of the fliad) .
In epic, because of its length, every part is given the appropriate 9. I Introduction
magnitude; but in plays the result is quite contrary to one's expecta­
tion. There is evidence of this in the fact that everyone who has The other formal elements have been discussed; it remains to I 9
composed a Sack of Troy as a whole, and not piecemeal like Euripi­ discuss diction and reasoning. The discussion of reasoning can be
des, or a Niobe and not like Aeschylus, has either failed or done badly reserved for my Rhetoric, since it has more to do with that field of
in the competition; even Agathon failed in this one respect. 86 enquiry. Under reasoning fall those effects which must be produced
by language; these include proof and refutation, the production of
emotions (e.g. pity, fear, anger, etc.), and also establishing import- 56b
8.8 Astonishment ance or unimportance.
(It is clear that in the events too one should apply the same prin­
In reversals and in simple actions poets use astonishment to achieve ciples when it is necessary to make something seem pitiable or ter­
their chosen aims; 87 this is tragic and agreeable. This happens when rible, important or probable. The only difference is that the one set of
someone who is clever but bad (like Sisyphus) is deceived, or some­ effects should be apparent without explicit statement, while the
one who is courageous but unjust is defeated. There is no violation others must be produced in speech by the speaker, and must come
of probability in this; as Agathon said, it is probable for many im­ about through the spoken word. What would the speaker's function
probable things to happen. 88 be if the necessary effect were evident without the use oflanguage?)
As for diction, one kind of enquiry is into the forms of utterance;
knowledge of these belongs to the art of performance and to the
8 .9 The chorus person who has that kind of expert knowledge - e.g. what is a
command, prayer, narrative, threat, question, answer, and anything
One should handle the chorus as one of the actors; it should be part else of that kind. Knowledge or ignorance of these matters does not
of the whole and should contribute to the performance - not as in give rise to any criticism relevant to the art of poetry that is actually
Euripides, but as in Sophocles. In the other poets the songs have no worth taking seriously; no one could suppose that there is an error

30 31
POETICS 9.2 BASIC C0NCEPTS

in the point Protagoras criticized (i.e. that Homer thinks he is utter­ effects the composition of a single significant vocalization from
ing a prayer but is in fact giving an order when he says 'Goddess, two or more vocalizations, and which should not occur at the
sing the wrath' : 90 Protagoras' point is that telling someone to do beginning of an utterance by itself (e.g. men, de, toi, de) . Or:
something or not is an order) . So let us set that aside as an investiga­ (b) A non-signifying vocalization which is capable of creating a
tion belonging to an art other than that of poetry. single significant vocalization from two or more vocalizations
which are themselves significant (e.g. 'around ' , 'about', etc.).
(iv) A conjunction is a non-signifying vocalization which marks the
9.2 Basic concepts beginning, end or division of an utterance, and which may occur
at the extremities as well as in the middle of an utterance.
20 Diction as a whole has the following elements: phoneme, syllable, (v) A noun93 is a composite significant vocalization which does not
connective, noun, verb, conjunction, inflection, utterance. express tense, no part of which is significant in its own right. (In
nouns comprising two parts we do not treat either part as signifi­
(i) A phoneme is an indivisible vocalization - not any kind, however, cant in its own right: e.g. the element -dorus in the name 'Theod­
but one which can be part of a composite vocalization; some am­ orus' does not signify.)94
mal noises are indivisible, but these are not what I mean by phon­ (vi) A verb is a composite significant vocalization which does express
emes.9' Phonemes are classified as vowels, continuants and mutes: tense, no part of which is significant in its own right Gust as with
(a) a vowel does not involve contact between the organs of speech, nouns). 'Person' or 'white' do not signify tense; the signification
and has audible sound; of 'walks' or 'walked ' includes present and past tense respectively.
(b) a continuant does involve contact between the organs of speech,
(vii) An inflection of a noun or verb is that which expresses (a) case
and has audible sound: e.g. s, r;
('ofhim', 'for him', etc.), (b) number (e.g. 'person', 'persons'), or (c)
(c) a mute does involve contact between the organs of speech, but
modes of expression, e.g. interrogative or imperative (thus 'did he
does not have sound in itself; it becomes audible when com­
walk?' and 'walk!' are inflections of the verb according to these
bined with a phoneme which has audible sound: e.g.g, d.
two categories) .
Phonemes differ in the shape of the mouth, in the point of con­
(viii) An utterance is a composite significant vocalization, part or parts
tact, in the presence or absence of aspiration, in length or brevity,
of which are significant in their own right. Not every utterance is
and in acute, grave or intermediate pitch. Detailed discussion of
composed of a verb and a noun (e.g. the definition of 'human
these differences belongs to the study of verse-forms.
being'); it is possible for an utterance to contain no verb. But it
(ii) A syllable is a non-signifying composite vocalization, comprising a
will always contain a part which signifies something (e.g. 'Cleon'
mute and a phoneme which has audible sound (thus gr is a syllable
in 'Cleon walks'). An utterance may be single in two senses: either
without an a, and also with an a, i.e.gra) . Detailed discussion of the
because it signifies a single object, or because it comprises a con­
differences between syllables also belongs to the study of verse­
nected plurality of utterances (e.g. the Iliad is a single utterance by
forms.
connection, the definition of 'human being' is a single utterance
(iii) A connective92 is: by virtue of signifying a single object).
5 7a (a) A non-signifying vocalization which neither prevents nor

32 33
POETICS 9.3 CLASSIFICATI0N 0F N 0UN S

to Dionysus as a shield does to Ares; so one may call a cup the


9 . 3 Classifi,cation of nouns 'shield ofDionysus', or a shield the 'cup ofAres'. '00 Or old age is to
life as evening is to the day; so one may speak of evening as the old
21 Nouns are classed as simple (by which I mean those not com­ age of the day (as Empedocles does) , '0' and of old age as the even­
pounded from significant parts, e.g. 'earth') or double. Double nouns ing of life, or life's twilight. In some cases there is no existing
may be composed of a significant and a non-signifying element noun for one term of the analogy, but it can nevertheless be ex­
(although within the noun itself there is no distinction between pressed. For example scattering seed is 'sowing', but there is no
significant and non-signifying elements), or of two significant ele­ noun for the scattering of fire from the sun; but this stands in a
ments. 95 One may also have triple, quadruple or even multiplex nouns similar relation to the sun as sowing does to seed; hence the ex­
(e.g. most of those from Marseilles, such as 'Hermoca1coxanthus').96 pression 'sowing the god-created fire'. 1 02 There is another way of .
57b Nouns are classed as current, non-standard, metaphorical, orna- using analogical metaphor: one may refer to something using the
mental, coined, lengthened, shortened and adapted. transferred noun, and negate some ofits proper attributes; e.g. one
By a current noun I mean one which is in use among a given might call a shield not 'the cup of Ares' but 'the wineless cup'.
people; by a non-standard noun I mean one which is in use among
An ornamental noun is . . . '03
other people. Obviously the same noun may be both current and
A coined noun is one that is not in use by anyone, but is posited by
non-standard, but not for the same people. (Sig unon is current
the poet himself. There seems to be a few nouns of this kind (e.g.
among the Cypriots, but non-standard to us; 'spear' is current
'sproutages' for horns and 'invocator' for priest) . '04
among us, but non-standard to them.)
As for lengthening and shortening, a noun is lengthened if it has a 58a
A metaphor is the application of a noun which properly applies to
longer vowel than usual or an extra syllable; a noun is shortened if
something else. The transfer may be from genus to species, from
something has been removed. Examples of lengthening are po/eos
species to genus, from species to species, or by analogy:
(for pole8s, 'of a city') and Pe/€iade8 (for Peleidou, 'of Peleus' son'); of
(i) By a transfer from genus to species I mean (e.g.) 'Here stands my shortening, e.g. kri (for krithe, 'barley') , d8 (for d8ma, 'house') and
ship'; lying at anchor is one kind of standing.97 'from two eyes single ops ' (for opsis, 'sight'). 105
(ii) From species to genus: 'Odysseus has in truth performed ten An adapted noun is one in which part of the word is kept un­
thousand noble deeds'; ten thousand is a large number, and is used changed, and part added; e.g. 'by the rightward breast' (for 'right') . '06
in place of 'many'.98 Nouns themselves may be masculine, feminine or neuter. Mascu­
(iii) Species to species: e.g. 'drawing off the life with bronze' and 'cut­ line nouns are those ending in n, r and s (and its compounds, of
ting off water with edged bronze'; here 'drawing off' means cut­ which there are two, ps and ks) . Feminine nouns are those ending in
ting, and 'cutting' means drawing off- each is a kind ofremoval.99 those vowels which are invariably long, i.e. in e and 8, and (among
(iv) By analogy I mean cases where B stands in a similar relation to A the vowels which are capable of being lengthened ) in a. '07 (So the
as D does to C; one can then mention D instead of B, and vice classes of masculine and feminine nouns turn out to be equal in
versa. Sometimes the thing to which the noun replaced stands in number, since ps and ks are simply compound forms of s.) No noun
relation is expressed; I mean (e.g.) a cup stands in a similar relation ends in a mute or in a short vowel; only three end in i (i.e. meli,

34 35
POET I C S 9.4 QUA LlTl ES 0F P0ET l C STYLE

kommi, pepen), and five in u (i.e. doru, p8u, napu, gonu, astu) . Neuters ing to Marathon' and 'not mixing his hellebore'. ' '0 Admittedly, ob­
end in these and in n, r and s. trusive use of this style is absurd; but moderation is equally necessary
in all aspects of diction; using metaphors, non-standard words and
the other categories in an inappropriate and deliberately absurd way
9.4 Qualities ofpoetic style would produce the same effect. The difference that appropriateness
makes in the case of epic poetry can be observed if one inserts the
22 The most important quality in diction is clarity, provided there is no ordinary words into the verse. Equally in the case of non-standard
loss of dignity. The clearest diction is that based on current words; words, metaphors and the other kinds, the truth of what I am saying
but that lacks dignity (as can be seen from the poetry of Cleophon, is obvious if one substitutes current words. For example, Aeschylus
and that of Sthenelus). 1 08 By contrast, diction is distinguished and and Euripides composed identical lines of iambic verse; but the
out of the ordinary when it makes use of exotic expressions - by change of a single word - a non-standard word in place of a current
which I mean non-standard words, metaphor, lengthening, and any­ one - made one line seem excellent, and the other trivial by com­
thing contrary to current usage. However, if one used nothing else parison. Aeschylus wrote, in his Philoctetes, 'the canker that eats up
the result would be a riddle or gibberish - a riddle if it were made my foot's flesh'; Euripides substituted 'feasts on' for 'eats up'. " ' Also
up entirely of metaphors, gibberish if it were made up entirely of in 'a scant and strengthless and unseemly man' one could substitute
non-standard words. (The essence of a riddle is that it states facts by current words: 'a little, weak, ugly man'. And in 'setting down an un­
means of a combination of impossibilities; this cannot be done by comely chair and scant table' : 'setting down a second-rate chair and
putting other kinds of word together, but it is possible using meta­ little table'. And in 'the sounding sea-shore': 'the shouting sea­
phor; e.g. 'I saw a man welding bronze upon a man with fire', '09 and shore' . . . , Ariphrades, too, ridiculed the tragedians for using expres­
such like. And what is composed of non-standard words is gibber­ sions that nobody would use in conversation, e.g. 'the house with­
ish.) So what is needed is some kind of mixture of these two things: out' (for 'outside the house') , 'of thine', 'Achilles round about' (for 59a
one of them will make the diction out of the ordinary and avoid a 'around Achilles'), etc. Things ohhis sort all make diction out of the
loss of dignity (i.e. non-standard words, metaphor, ornament and ordinary because they are not part of current usage. Ariphrades
the other categories I mentioned earlier), while current usage will failed to understand this. "3
contribute clarity. It is important to use all the things I have mentioned appropri­
5sb A major contribution to a style that is both clear and out of the ately,including compound and non-standard words; but the most im­
ordinary is made by lengthenings, abbreviations and alterations. The portant thing is to be good at using metaphor. This is the one thing
variation from current usage makes the diction out of the ordinary, that cannot be learnt from someone else, and is a sign of natural
because we are not used to it; but it has something in common with talent; for the successful use of metaphor is a matter of perceiving
what we are used to, so it will be clear. The people who find fault similarities. Compound words are most appropriate in dithyramb,
with this kind of style and satirize Homer are therefore mistaken in non-standard words in heroic verse, and metaphor in iambics. In
their criticism; e.g. the elder Eucleides argued that writing poetry is heroic verse all the things I have mentioned have their use; but in
easy if one is allowed to use lengthening as much as one likes, and iambic verse, because of its close resemblance to ordinary speech, "4
composed lampoons in the style in question: 'I saw Epichares walk- the most appropriate words are the ones which could also be used in

37
POETICS I 0. 3 D I F F E R E N C E S B E T W E E N T R A G E D Y A N D E P I C

prose speeches 1.e. current words, metaphor and ornamental composition) . The other poets write about a single person, a single 59b
words. period of time, or a single action of many parts - e.g. the poet of the
Cypria and the Little Iliad. "7 This means that only one tragedy can
be made out of the Iliad and Odyssey, or at most two, but many out
of the Cypria and the Little Iliad (more than eight, e.g. Adjudication of
A rms, Philoctetes, Neoptolemus, Eurypylus, Beggary, Spartan rMimen,
IO. EPIC Sack of Troy, Putting to Sea; also Sinon and Trojan TMimen) .

IO. I Plot
10.2 Kinds and parts of epic
Tragedy and imitation in action has been adequately covered in
23 what has already been said. As for the art of imitation in narrative Epic must also have the same kinds as tragedy; it is either simple or 24
verse, it is clear that the plots ought (as in tragedy) to be constructed complex, or based on character or on suffering. 1 1 8 The component
dramatically; that is, they should be concerned with a unified action, parts, except for lyric poetry and spectacle, are also the same; it too
whole and complete, possessing a beginning, middle parts and an needs reversals, recognitions and sufferings, and the reasoning and
end, so that (like a living organism) the unified whole can effect its diction should be of high quality. Homer was the first to use all of
characteristic pleasure. They should not be organized in the same these elements in a completely satisfactory way. Each of his two
way as histories, in which one has to describe not a single action, but poems has a different structure; the Iliad is simple and based on suf­
a single period of time, i.e. all the events that occurred during that fering, the Odyssey is complex (recognition pervades it) and based
period involving one or more people, each of which has an arbitrary on character. In addition, he excels everyone in diction and
relation to the others. The naval engagement at Salamis and the reasoning.
battle against the Carthaginians in Sicily occurred simultaneously
without in any way tending towards the same end; " 5 in exactly the
same way one thing may follow another in succession over a period 10.3 Differences between tragedy and epic
.
of time without their producing a single result. But perhaps the
majority of poets compose in this way. Epic is differentiated in the length of its plot-structure and in its
So (as we have already said ) Homer's brilliance is evident in this verse-form. The stated definition oflength is adequate; one must be
respect as well, in comparison with other poets. He did not even try able to take in the beginning and the end in one view. This would
to treat the war as a whole, although it does have a beginning and an be the case if the structures were shorter than those of the ancient
end. Had he done so, the plot would have been excessively large and epics, and matched the number of tragedies presented at one sit­
difficult to take in at one view - or, if it had been moderate in mag­ ting. 1 19 Epic has an important distinctive resource for extending its
nitude, it would have been over-complicated in its variety. Instead, length. In tragedy it is not possible to imitate many parts of the
he has taken one part and used many others as episodes (e.g. the cata­ action being carried on simultaneously, but only the one on stage
1 16
logue of ships, and other episodes which he uses to diversify his involving the actors. But in epic, because it is narrative, it is possible

39
POETICS I 0. 5 A S T 0 N I S H M E N T A N D I R R A T I 0 N A L I T I E S

to treat many parts being carried on simultaneously; and these (pro­


vided that they are germane) make the poem more impressive. So
IO. 5 Astonishment and irrationalities
epic has this advantage in achieving grandeur, variety of interest for
the hearer and diversity of episodes; similarity quickly palls, and may While it is true that astonishment is an effect which should be
cause tragedies to fail. sought in tragedy, the irrational (which is the most important source

As for the verse-form, experience has proved the appropriateness of astonishment) is more feasible in epic, because one is not looking
of the heroic verse. If one were to compose a narrative imitation in at the agent. The pursuit of Hector would seem preposterous on

some other verse-form, or a combination of them, it would seem stage, with the others standing by and taking no part in the pursuit

unsuitable. Heroic verse is the most stately and grandiose form of while Achilles shakes his head to restrain them; but in epic it escapes
notice. 121 Astonishment gives pleasure; evidence of this is the fact .
verse; this is why it is particularly receptive to non-standard words
and metaphors (for narrative imitation departs further from the that everyone exaggerates when passing on news, on the assumption

norm than other kinds). Iambic verse and the trochaic tetrameter that they are giving pleasure.
Homer, in particular, taught other poets the right way to tell
6oa express movement (the latter having a dance-like quality, and the
falsehoods. This is the false inference. In cases where the existence or
former being suited to action) . It would be still more peculiar if one
occurrence of A implies the existence or occurrence of B, people
mixed them, as Chaeremon did. 12° For this reason no one has com­
imagine that if B is the case then A also exists or occurs - which is
posed a long structure in any verse-form other than the heroic; as we
fallacious. So ifA is false, but its existence would entail the existence
have said, nature itself teaches people to choose what is appropriate
or occurrence of B, one should add B; then, on the basis of its know­
to it.
ledge that B is true, our mind falsely infers the reality of A as well.
An example of this can be found in the bath-scene. 122
Probable impossibilities are preferable to implausible possibilities.
I0.4 Quasi-dramatic epic
Stories should not be constructed from irrational parts; so far as
possible they should contain nothing irrational - or, failing that, it
Homer deserves praise for many reasons, but above all because he
should be outside the narration (like Oedipus' ignorance of the
alone among poets is not ignorant of what he should do in his own
manner of Laius' death) 123 and not in the play itself (like the report
person. The poet in person should say as little as possible; that is not
of the Pythian Games in the Electra, or the man who comes from
what makes him an imitator. Other poets perform in person
Tegea to Mysia without speaking in the Mysians) . 1 24 Saying that the
throughout, and imitate little and seldom; but after a brief preamble
plot would have been ruined otherwise is absurd; plots should not
Homer introduces a man or woman or some other character - and
be constructed like that in the first place. But if one does posit an ir­
none of them are characterless: they have character.
rationality and it seems more or less rational, even an oddity is pos­
sible; 125 the irrationalities involved in Odysseus' being put ashore in
the Odyssey would be manifestly intolerable if a second-rate poet 6ob
had composed them, but as it is the poet conceals the absurdity with
other good qualities, and makes it a source of pleasure. 126

41
POETICS I I .2 A P P L I C A T I 0 N S

1 0. 6 Diction rr .2 Applications

Diction should be handled with particular care in those parts in So one should solve the objections posed in problems by consider­
which little is happening, and which are expressive neither of char­ ing them on the basis of these principles.
acter nor of reasoning; excessively brilliant diction overshadows
(i) First, those with regard to the art ofpoetry itself. If impossibilities
character and reasoning.
have been included in a poem, that is an error; but it is correct ifit
attains the end of the art itself (the end has been stated above) : i.e.
if it makes either this or some other part have greater impact. An
example is the pursuit of Hector. 128 If, however, it is possible for ·
the end to be achieved as well or better without contravening the
I I . PROBLEMS AND S OLUTIONS
art concerned with those matters, then the error is not correct;
there should if possible be no error at all.
II.I Principles
(ii) Also, which class does the error belong to: those in respect of the
art, or those in respect of some other incidental? It is less serious
25 As for problems and their solutions, their number and the classes
if the artist was unaware of the fact that a female deer does not
into which they fall should become clear if considered in this way:
have antlers than ifhe painted a poor imitation.
(i) The poet is engaged in imitation,just like a painter or anyone else (iii) Furthermore, if the objection is that something is not true, per­
who produces visual images, and the object of his imitation must haps it is as it ought to be; e.g. Sophocles said that he portrayed
in every case be one of three things: either the kind of thing that people as they should be, Euripides as they are. That is the solu­
was or is the case; or the kind of thing that is said or thought to be tion to use.
the case; or the kind of thing that ought to be the case. (iv) If it is neither true nor as it ought to be, one might reply that this
(ii) The diction in which these things are expressed includes non­ is what people say; e.g. stories about the gods: it may be that
standard words, metaphors and many modifications of diction; talking like that is neither an idealization nor the truth, and
these licences are allowed to poets. perhaps Xenophanes was right; 129 but at any rate, that is what 61a
(iii) In addition, correctness is not the same thing in ethics and people say.
poetry, nor in any other art and poetry. Error in poetry is of two (v) Other things, though not idealizations, may perhaps reflect the
kinds, one intrinsic, the other incidental. If someone has chosen way things used to be; e.g. the passage about the weapons, 'their
to imitate accurately but failed to do so because of incompetence, spears stood upright on the butt-end ' - that was the norm then
the fault is intrinsic; but if he has chosen not to do so correctly (as it is even now among the Illyrians) . '30
(e.g. to show a horse with both right legs thrown forward ) the (vi) In evaluating any utterance or action, one must take into account
error is in respect to the particular art (e.g. in respect to medicine not just the moral qualities of what is actually done or said,
or some other art), not in respect to the art of poetry itself. 127 but also the identity of the agent or speaker, the addressee, the

42 43
POETICS I I .J C0N CLUSI0N

occasion, the means, and the motive (e.g. whether it is to bring the adverse verdict is one they have reached by themselves,
about a greater good or avert a greater evil) . they make inferences from it and if anything contradicts their
(vii) Other problems should be solved with an eye to diction. For ex­ own ideas they criticize the poet as if he had expressed their opin­
ample a non-standard word may provide the solution to 'first the ion. This is what has occurred in the case of Icarius. 1 42 People
mules' (perhaps he does not mean mules but sentinels), 1 3 1 Dolan assume that he is a Spartan, and that Telemachus' not meeting
being 'ugly in appearance' (not physically deformed but facially him when he went to Sparta is therefore odd. But perhaps the
disfigured, since Cretans call facial beauty 'beauty of appear-: Cephallenians are right when they say that Odysseus married
ance'), '32 and 'mix the wine stronger' (not undiluted, as for from among them, and that his name is Icadius not Icarius. So
drunkards, but faster). 133 Other things are said metaphorically, e.g. probably the problem is based on a misconception.)
'all the gods and men slept through the night', while at the same
time he says 'when he looked out over the Trojan plain . . . the
sound of pipes and pan-pipes'; 'all' is said metaphorically for I I . 3 Conclusion
'many', since all is a. lot. 134 Also 'alone with no share' is meta­
phorical, the best known instance being unique. 1 35 In general:
(viii) With reference to pronunciation, as in Hippias ofThasos' solu­
(i) Impossibilities should be referred to poetic effect, or idealization
tion to 'we grant him achievement of glory' and 'part is rotted by
of the truth, or opinion. With regard to poetic effect, a plausible
rain' . 1 36
impossibility is preferable to what is implausible but possible.
(ix) Punctuation provides the solution to some problems; e.g.
Again, it is impossible for people to be as Zeuxis painted them,
Empedocles: 'at once mortal things were born that before were
but that is an idealization of the truth; one should surpass the
immortal, and things unmixed formerly mixed '. 1 37
model. 1 43
(x) So does ambiguity; e.g. ' more of the night has passed ' - 'more' is
(ii) Irrationalities should be referred to what people say: that is one
ambiguous. 13 8
solution, and also sometimes that it is not irrational, since it is
(xi) Other problems can be solved with reference to linguistic usage.
probable that improbable things will happen. 1 44
We call diluted wine 'wine'; hence the phrase 'greaves of new­
(iii) Contradictory utterances should be subjected to the same scru­
forged tin' . We call people who work iron 'bronze-smiths'; hence
tiny as refutations in arguments (i.e. is the same thing said, with
Ganymede is said to pour wine for Zeus, although the gods do
reference to the same thing, and in the same sense?) , to establish
not drink wine (this could also be metaphorical) . 1 39
whether the poet contradicts either what he says himself or what
(xii) Whenever a word seems to imply a contradiction, one should
a reasonable person would assume. 145
consider the number of meanings it could bear in the context;
(iv) An objection, either to irrationality or to depravity, is correct
e.g. in 'by it was the bronze spear stayed ' - how many different
when there is no necessity and the poet makes no use of the ir­
possible ways are there for it to be stopped there, in one way or
rationality (as Euripides fails to use Aegeus) or of the wickedness
another, however one might best take itt40 (This is the exact
(as that of Menelaus in the Orestes) . 146
61b opposite of what Glaucon describes/41 when he says that some
people make unreasonable prior assumptions and then, although So the objections people make are of five kinds, i.e. that

44 45
POETICS 1 2.2 R E P L Y

something is impossible, irrational, harmful, contradictory, or con­


trary to correctness in the art. Solutions should be sought from 1 2. 2 Reply
those enumerated; there are twelve of them. 147
(i) First of all, this is not a criticism of the art of poetry but of the art
of performance. A rhapsode performing epic poetry can make
exaggerated use ofgestures (like Sosistratus) ; so can a singer (this is
what Mnasitheus of Opus used to do) . 1 50
I 2 . C O M PA R A T I V E E VA L U A T I O N O F (ii) Next, not all movement is to be disparaged (any more than all
E P I C A N D T R A G E DY dance is), but only that of inferior persons. This is the objection
that used to be made against Callippides, and is made now against
others, on the grounds that the women they imitate are not
26 One might pose the question whether epic imitation or tragic is respectable. 1 5 1
superior. (iii) Also, tragedy has its effect without movement,just as epic does:
its quality is clear from reading.
So if tragedy is superior in other respects, this criticism at any
1 2 . l The case against tragedy rate does not necessarily apply to it. Further:
(iv) Tragedy has everything that epic does (it can even make use ofits
If the less vulgar art is superior, and in all cases what is addressed to a verse-form), and additionally it has as a major component part
superior audience is less vulgar, then it is perfectly clear that the art music and spectacle; this is a source of intense pleasure.
which imitates indiscriminately is vulgar. Assuming that the audi­ (v) Also it has vividness in reading as well as in performance.
ence is incapable of grasping what the performer does not supply in (vi) Also, the end of imitation is attained in shorter length; what is 62b
person, they engage in a great deal of movement (as second-rate more concentrated is more pleasant than what is watered down by
pipers spin round if they have to imitate throwing a discus, and drag being extended in time (I mean, for example, if one were to turn
the chorus-leader about if they have to play the Scylla). 148 Tragedy is Sophocles' Oedipus into as many lines as the fliad has).
like that. This is in fact the opinion which older actors held about (vii) Also the epic poets' imitation is less unified (an indication ohhis
those who came after them; Mynniscus used to call Callippides is that more than one tragedy comes from any given imitation) .
62a 'monkey' because of his excesses, and Pindarus was viewed in much So if they treat a unified plot, either the exposition is brief and
the same way. 149 The whole art of tragedy stands in the same rela­ appears curtailed, or else it adheres to the length of that verse­
tion to epic as these do to the others. So it is argued that epic is form and is diluted. 1 52 (I mean, for example, if it comprises a
addressed to decent audiences who do not need gestures, while number of actions. The fliad and Odyssey have many parts of this
tragedy is addressed to second-rate audiences; if, then, tragedy is kind, which possess magnitude in their own right; and yet the
vulgar, clearly it must be inferior. construction of these poems could not be improved upon, and
they are an imitation of a single action to the greatest possible
degree.)

47
POET I C S

So tragedy surpasses epic in all these respects, and also in artistic


effect (since they should not produce any arbitrary pleasure but the
one specified ) ; clearly, then, because it achieves its purpose more
effectively than epic, tragedy must be superior. N O T E S T O T H E
T R A N S L A T I O N

1 . For 'imitation' (mimesis) see Introduction §2. The dithyramb was a kind of
lyric poetry performed by a chorus. Pipe (aulos) and lyre (kithara) were the
two most common forms of Greek wind and string instrument; the ad­
I 3. C O N C LU S I O N
dition of the pan-pipes (syrinx) below implies a more general concept of
instrumental music.
2. The reference is to the mimicry of, for example, animal noises (cf. Plato,
So much for tragedy and epic, the number and variety of their Republic, 397a, LAws, 669c-d) .
forms and component parts, the causes of their success and failure, 3. Sophron and his son Xenarchus worked in Syracuse in the late fifth cen­
and criticisms and solutions. tury; their sketches of everyday life, and the philosophical dialogues of Plato
and Xenophon (and others whose works have not survived), are cited as
examples of imitation in prose.
4. The fifth-century philosopher Empedocles expounded his theories of
nature in hexameter verse. Aristotle greatly admired the artistic quality of
his work: fragment 70 (from On Poets) comments on his 'Homeric' mastery
of poetic language, and especially his use of metaphor; he is cited several
times in the Poetics ( 5 7b 1 3 f. , 24, 5 8a 5 , 6 1 a24f.). The present point is there­
fore not evaluative, but intended to define a restricted technical usage for
the term 'poetry' .
5. A fourth-century tragic poet; little is known about his Centaur (see also
6oa2), apparently a piece for recitation ('rhapsody') in a variety of metres.
6. The nome, like the dithyramb (n. 1 ) , was a kind of choral lyric.
7. All fifth-century painters. Polygnotus (cf. 5oa27f. and n. 25) is famous;
little is known of the other two, but at Politics, 1 3 4oa3 5-8 Pauson's work is
described as less suitable for young people to view than that of Polygnotus.
8. Cleophon may be the fourth-century tragic poet of that name; on his
style cf. 5 8a 1 8-2 1 . Hegemon wrote epic burlesques in the late fifth century;
Nicochares' Deiliad (derived from deilos, 'cowardly', by analogy with fliad)
also suggests epic burlesque.
9. For dithyramb and nome cf. 47b26 (n. 6). Timotheus (cf. 54a3of. and n . 6 1 )
was a lyric poet o f the late fifth and early fourth centuries noted fo r musical
and stylistic innovations; his poem is presumably mentioned as an example

49
N O T E S TO T H E T R A N SLAT I O N N O T E S TO T H E T R A N S LAT I O N

of a serious treatment of the Cyclops Polyphemus, since the portrayal (49a r 6- 1 9) ; but it remained the custom in Greek theatre for a limited
of Polyphemus by Philoxenus of Cythera was said to be a caricature of number of actors to play all the roles.
the tyrant Dionysius I, whose mistress the poet had seduced. But the text 17. Satyr-plays were mythological burlesques with a chorus of satyrs (idle,
here is uncertain, and we cannot be quite sure what point Aristotle is drunken and lascivious followers of Dionysus, with a mixture of human and
making. animal features), written and produced by tragic dramatists as a regular part
10. For Homer's quasi-dramatic style, making extensive use of direct speech, of the tragic competition. Aristotle infers that they preserve characteristics
cf. 48b34-8, 6oa5-1 r. Aristotle's classification of modes is an adaptation of of tragedy's pre-dramatic choral antecedents (cf. 49a1 0-- 1 3 and n. 1 6) which
Plato's (Republic, 3 92d-4c); unlike Aristotle, Plato regarded the dramatic had disappeared in the evolution of tragedy proper. The text of this whole
mode with disfavour. sentence is extremely uncertain.
1 1 . The verb for 'do' here is dran, whence 'drama'. The Dorian claim to 18. On the characteristics of iambic verse cf. 59ar o-- 1 3 , 59b34-8, Rhetoric,
have invented tragedy and comedy assumes (as becomes clear at the end of I 404a2<)--3 3 .

the following digression) that dran implies an origin among speakers of the 19. B y contrast tragedy evokes fear and pity, emotions which Aristotle
Doric dialect; Aristotle is rightly sceptical of this assumption, and of the al­ defines as responses to painful and destructive harm (Rhetoric, 1 3 82az 1 f. ,
leged derivation of 'comedy' from kome ('village'). The democracy of the r 3 85 b 1 3f.; cf. 5 2b n f. below). A mask like that of the blinded Oedipus, dis­
city of Megara in mainland Greece is dated to the sixth century BC; the torted by agonizing wounds, fulfils the function of tragedy, but would be
Megarians in Sicily are the colonists at Megara Hyblaea. Epicharmus (from out of place in comedy.
another Sicilian city, Syracuse) worked in the late sixth and early fifth cen­ 20. The archon chose the poets to compete in an Athenian dramatic festival,
turies, not really 'much earlier' than the Athenian comic poets Chionides and assigned a wealthy citizen to each to finance the production-costs (es­
and Magnes (active from the 48os and 47os). pecially the costume and training of the chorus) . The earliest official comic
12. The 'two causes' are probably the human instincts for (a) imitation and competition in Athens was in 486 BC.
(b) melody and rhythm (48b2of.). Many interpreters try to identify two dis­ 2 1 . B y 'universalized ' Aristotle means constructed in accordance with
tinct causes related to imitation; but this would not explain poetry as imita­ necessity or probability; cf. chapter 9 below. Crates was an Athenian comic
tion in verse. The human instinct for rhythm and melody is also recognized dramatist active from c. 450 BC.
by Plato (Laws, 653d-654a). 22. By imitation in hexameter verse Aristotle means epic; his discussion
13. A rhythm can in principle be continued indefinitely; a verse-form such begins in chapter 2 3 . The promised discussion of comedy is not included in
as the dactylic hexameter is a defined segment of that potentially infinite the extant Poetics; it was probably contained in a lost second book.
rhythmic continuum. Cf. Rhetoric, 1 408b29. 23. Katharsis: see Introduction §8.
14. That is, one imitating inferior agents. The Margites was not an invective, 24. The text and interpretation of this sentence are extremely uncertain.
but a burlesque narrative about a hero of wide-ranging incompetence (he 25. For Polygnotus cf. 48a5f. (n. 7); he is cited as an example ofpainters good
'knew many things, but knew them all badly'). The attribution of the poem at portraying character at Politics, 1 3 4oa3 7f. Zeuxis, a painter active in the
to Homer makes it antedate the earliest extant lampoons (those of the late fifth and early fourth centuries, reappears at 6 1 b 1 2f. to illustrate idealiza­
seventh-century poet Archilochus) ; but Aristotle reasonably infers that tion in portraiture.
there must have been earlier lampoonists whose poems had not been 26. For reversals and recognitions see chapter 1 1 .
preserved. 27. Statesmanship forms judgements on the right conduct of public affairs;
1 5 . For Homer's quasi-dramatic style cf. 48a2 1 f. (n. )
IO , 6oa5-1 I . statesmen may express these j udgements using natural eloquence, or ex­
16. The assumption is that a chorus-leader singing a solo response to the ploiting the systematized persuasive techniques of rhetoric. Speech expres­
choral parts was the first step towards an actor wholly separate from the sive of character and moral choice (prohairesis) is contrasted with speech
chorus. The number of actors was subsequently increased to two and three based on reasoning at Rhetoric, 1 4 1 7a 1 6-28.

50 51
N OTES TO THE TRAN SLATION NOTES TO THE TRANSLATION

28. Litigants in the Athenian courts were allocated a limited time for their 42. Referring to the outcome; this does not contradict the statement that the
speeches, measured by a water-clock. But the last part of this sentence has structure of the plot should be complex rather than (in a different sense)
never been satisfactorily explained. simple (52a 1 2- 1 8 , 52b3 1 f.).
29. As a young man Odysseus was wounded by a boar during a hunt; later 43. Alcmeon, like Orestes, avenged his father's death by killing his mother
he tried to avoid joining the Greek expedition against Troy by pretending (cf. 5 3b24f., 3 3) ; Oedipus killed his father and slept with his mother in
to be insane. There is no causal relationship between these two events. The ignorance of their identity; likewise Meleager and Telephus killed their
wounding is, in fact, recounted in Odyssey, 1 9 . 3 93-466, to explain the scar uncles; Thyestes unwittingly ate his children's flesh (served to him at a feast
which establishes Odysseus' identity (see 54b26-30 and n. 7 1 ) ; but it is not by his brother Atreus in revenge for the seduction of the latter's wife) and
part of the poem's plot, which is Aristotle's current concern. committed incest with his own daughter.
30. For Aristotle's view of history cf. 59a21--9. 44. That_ is, Euripides' critics make the same mistake as the advocates of the
31. Agathon was a prominent Athenian tragic poet of the late fifth century double outcome mentioned in the previous paragraph.
(Plato's Symposium is set at a party celebrating his first victory, in 4 1 6 BC); 45. Odysseus triumphs, the wicked suitors are killed (cf. 5 5b22f.).
Antheus is otherwise unknown.
cf. 56a 1 8f., 24f., 29f. His 46. In tragedy Orestes kills Aegisthus to avenge his father's death. The
32. The same word (poietes) means both 'poet' and 'maker'. Aristotle's point fourth-century comic poet Alexis wrote an Orestes, but we do not know
(here as in chapter 1) is that writing in verse is not sufficient to identify a whether Aristotle is alluding to the plot of that (or some similar) play or
poet; the poet is an inutator of action, and therefore a maker of plots. suggesting a hypothetical extreme.
33. 'Simple' is defined in chapter IO. 47. Philoi: cf. n. 3 8 and Introduction §7.
34. Greek dramatic festivals were competitive events, and there had been a 48. Both cases of matricide (cf. n. 43).
competition between the leading actors (separate from that between the 49. I n Euripides' Medea, Medea punishes the infidelity ofher husband Jason
dramatists) since 449 BC. Aristotle comments on the contemporary domin­ by killing their children.
ance of actors over poets at Rhetoric, 1 403b3 3 . 50. Astydarnas was a leading tragic dramatist in the mid fourth century; it is
35. A n Argive named Mitys i s mentioned i n a speech falsely attributed to not known how he contrived to keep Alcmeon (cf. 5 3 a20 and n. 43 , 5 3 b24)
Demosthenes (59. 3 3 ) ; nothing more is known about the incident referred in ignorance of his mother's identity. In Odysseus Wounded, a lost play by
to here. Sophocles, Telegonus was Odysseus' son by Circe; having never seen his
36. At Sophocles, Oedipus, 92.µf. a messenger brings the good (934) news father, he did not recognize him when they fought.
that Oedipus has succeeded to the Corinthian throne on the death of his 5 1 . One possibility (knowing and not acting) has been overlooked in the
supposed father; when he learns that Oedipus is reluctant to return to Cor­ foregoing enumeration, although it is included (somewhat disrnissively) in
inth for fear of committing incest with his mother, he is eager to allay that the following ranking. It does appear in the Arabic translation (after the ref­
fear too (roo2ff.); but in doing so he sets in train the sequence of events erence to Medea at 5 3 b29); but 'third possibility' at 5 3 b34 suggests that this
which leads to the discovery of Oedipus' parricide and incest. may be a later addition by a reader who has noticed the ornission.
37. For the Lynceus, by Aristotle's friend Theodectes, see 5 5b29-"32 (n. 82) . 52. Sophocles, Antigone, 1 2 3 I-T the distraught Haemon tries to stab his
38. Philia: see Introduction, §7. As Aristotle goes on to say, this is not the father Creon, and then in remorse kills himself.
only kind of recognition; but its close bearing on good and bad fortune 53. The Cresphontes is a lost play by Euripides. When Merope's husband was
makes it particularly effective in tragedy. murdered she managed to smuggle their baby son Cresphontes to safety;
39. For the recognitions in Euripides' Iphigeneia in Tauris see 54a7, 54b3 1-6 many years later she tries to kill the stranger who comes to claim the reward
(n. 72), 5 5a 1 8f., 5 5b3- 1 2 . for killing Cresphontes, but discovers before it is too late that the stranger is
40. For this section see I ntroduction, n. I O . Cresphontes himself, returning in disguise to avenge his father's death.
4 1 . Hamartia: see Introduction §7. 54. See the summary of the story at 5 5b3- 1 2 .

52 53
N OTES TO THE TRAN SLATION N OT E S TO T H E T R A N S LAT I O N

55. Nothing is known of this play; and while Helle's family (including means of a supernatural chariot, and fliad, 2. 109"-2 r o, where Agamemnon
Athamas, Phrixus and !no) was fertile in tragic events, none of the attested proposes abandoning the siege of Troy in an oblique attempt to stimulate
stories corresponds to the incident Aristotle describes here. the army's fighting spirit; but the army, taking the proposal at face value, ac­
56. In chapter 1 3 ( 5 3 a 1 8-22) . cepts it with enthusiasm, and the goddess Athene has to intervene to resolve
57. For Aristotle's views on whether and in what sense slaves and women the crisis which ensues.
can be 'good ' cf. especially Politics, 1 . 1 3 ( 1 2 59b1 8-6ob24). 65. Referring to Oedipus' ignorance of the circumstances of Laius' death
58. The character should display the right kind of goodness; a good woman (see 6oa30).
should be courageous, but not in the same way as a man (cf. Politics, 66. The text of this sentence is uncertain.
l 26oa20-24, 1 2 77b20-2 5 ; 'a man would be regarded as a coward if he were 67. No one knows what this paragraph means. By 'published works'
courageous in the same way that a woman is courageous'). 'Cleverness' Aristotle presumably means his On Poets.
looks forward to the example of Melanippe below (54a3 l and n. 62). 68. The 'earth-born' were the men who sprang from the dragon's teeth
59. The reference is obscure; Aristotle probably means 'like us' (the pre­ sown by Cadmus; their descendants had a birth-mark in the shape of a
condition offear at 5 3 �-6, and cf. 48a4- 1 4) . See I ntroduction §9. spearhead. In one version of the story, Creon recognized Maeon, son of
6o. In Euripides' Orestes Menelaus' failure to support his nephew Orestes Haemon and Antigone, by this mark; but we do not know the source of the
violates the obligations owed to a philos (see Introduction §1); the example line which Aristotle quotes.
recurs at 6 1b2 1 . 69. Carcinus was a tragic poet of the early fourth century (cf. 5 5a26-9) . We
6 1 . A dithyramb (also mentioned at 6 1 b32) by Timotheus (48a 1 5 and n . 9) do not know anything about the plot of this play; for Thyestes see 5 3 a2 1
which portrayed Odysseus lamenting the loss of his comrades, eaten by the (n. 43); the star-shaped birthmark was a characteristic of his family (the
monster Scylla (cf. Odyssey, 1 2.234-59). descendants of Pelops) .
62. The reference is to Euripides' lost Melanippe the Wise. Melanippe gave 70. Tyro set her twin sons by Poseidon adrift in a small boat; in Sophocles'
birth to twins by the god Poseidon, and exposed them; when the babies (lost) Tyro the boat served as a recognition-token.
were found being suckled by a cow Melanippe's father, assuming that the 7 1 . In the bath-scene (Odyssey, 1 9. 3 86-475 ; cf. 6oa25 f. and n . 1 22) the Nurse
cow had given birth to them, decided to have them destroyed as unnatural penetrates Odysseus' disguise when she observes his scar (cf. 5 1 az6 and
monsters. Melanippe tried to prevent this by arguing that the cow could not n. 29). This is an unplanned consequence of Odysseus' own request
have given birth to human children; her speech included advanced cosmo­ ( 1 9 . 3 43-8) that his feet be washed by one of the older female servants, and is
logical and theological arguments, thus displaying a cleverness inappropriate thus linked to a reversal. But in Odyssey, 2 1 . 1 88-224 Odysseus simply de­
in a woman (cf. n. 5 8) . clares his identity to the herdsmen Eumaeus and Philoetius, and shows them
63. In Euripides' Iphigeneia in Au/is Iphigeneia's first reaction o n learning the scar by way of confirmation.
that she is to be sacrificed to Artemis to secure the Greek army's passage to 72. In Euripides' Iphigeneia in Tauris (see 52b6-8 and n. 39) Iphigeneia's
Troy is to plead for her life ( 1 2 n-52); but later she patriotically embraces identity is revealed when she asks one of the two strangers to deliver a
her fate ( 1 3 68-1 40 1 ) . message addressed to her brother Orestes (769"-94) ; Orestes then declares
64. 'Resolution' i s defined i n chapter 1 8 (5 5b24-3 2); c f. 5 6a9f. for poor himself, confirming his identity by displaying knowledge of their home
technique in resolutions. 'Theatrical device' renders mekhane; literally, this (808-26) . Aristotle's point is that he could equally well have brought some
was a crane used in the Greek theatre for the appearance of a god who physical recognition-token with him.
might conclude the play by outlining subsequent events or (less appropri­ 73. Tereus was married to Procne, and raped her sister Philomela. To keep
ately, in Aristotle's view) by imposing an arbitrary resolution on the plot. his crime secret he cut out her tongue, but she wove a tapestry showing
The two examples are Euripides' Medea, in which Medea's escape from what had happened; this picture was the 'voice of the shuttle'.
Corinth after the killing of her children (see 5 3b29 and n. 49) is contrived by 74. Dicaeogenes was a late fifth-century tragedian; nothing is known of his

54 55
N OT E S T O T H E T R A N S LAT I O N NOTES TO THE TRANSLATION

Cyprians. In Odyssey, 8.485-586 Odysseus weeps on hearing a song about found and his parentage revealed; Dana us condemned Lynceus to death but

the fall of Troy; this prompts his host Alcinous to enquire about his identity, somehow (cf. 5 2a27-9 and n. 3 7) this led to a reversal, and it was Danaus

and Odysseus's reply (Odyssey, 9- I 2) is the tale told to Alcinous. who died; Lynceus survived. But the text here is uncertain, and our limited

75. ln Aeschylus' Choephori, 1 66-2 1 r Electra finds a lock of hair and a foot­ knowledge of the play's plot makes it impossible to reconstruct Aristotle's

print at her father's tomb, and infers Orestes' presence from their similarity words with complete confidence.

to her own. Despite this example, Aristotle is not thinking primarily of rec­ 83. A perplexing statement: there has been no mention ofjour parts of tra­

ognition through reasoning from signs, but (as the following examples make gedy before now. To add to the confusion, the name of the fourth kind has

clearer) situations in which one character's reasoning discloses his or her been lost in the Greek text; the conjectural text translated here makes the

identity to another. As with recognition through memory, a character's passage consistent with the cross-reference at 59b7-9 (another widely

spontaneous response to the situation provides a clue by means of which accepted conjecture makes the fourth kind the 'tragedy of spectacle'). The

their identity can be inferred. examples do little to cast light on Aristotle's meaning.

76. Nothing more is known of Polyidus; his suggestion is mentioned again 84. The obscurity of the preceding paragraph casts a shadow over this one as

at 5 5b 1 0f. For Theodectes cf. 52a27-9 (n. 3 7), 5 5 b29-J2 (n. 82); nothing well. Presumably poets with a special talent for the depiction of character

more is known about his Tydeus. The Sons of Phineus is also unknown. were criticized for failing to depict suffering as effectively as poets who

77. Unknown. Aristotle's account is very cryptic. Presumably Odysseus on specialized in that kind of tragedy, and vice versa.
his homecoming concealed his identity by bringing a false report of his own 85. For faulty resolutions cf. 54a37-b2.

death; the audience is led to expect him to establish his identity by stringing 86. The Sack of Troy was the title of an epic poem and of several lost traged­
the bow (which no one but Odysseus could do), but instead he is accepted ies; Euripides' Trojan Women and Hecuba both deal with events drawn from
simply because he recognizes the bow (which anyone who had seen or this larger story. The reference to Niobe is perplexing, since there was no

heard a report of it might do). For the exploitation of false inference by epic on that subject; Aristotle perhaps wrote something different, but the

poets see also 6oa r 8-26. text cannot be corrected with any confidence. We do not know what

78. For Carcinus cf. n. 69; we have no further information about the mistake failure of Agathon (cf. n. 3 r) Aristotle is referring to.

referred to here. 87. The text and interpretation here are uncertain.

79. To imitate convincingly the poet must be able to project himself into the 88. For Agathon cf. n. 3 r ; and cf. 6 1 b 1 5 for the principle stated here.

emotions of the subjects. This is made easier by the versatility of a genius or 89. Instead of composing choral lyrics in the traditional way, the poet could

by the madman's weak grasp on his own identity; more generally, acting out simply mark the points at which the chorus should perform and leave it to

the part with gestures may help. This was evidently a well-established view the producer to choose the songs to be sung in these interludes. In comedy

of a poet's method of work; Aristophanes has fun with it (Thesmophoria­ we can observe this being done sporadically by Aristophanes in his later

zusae, r 56-8). work at the beginning of the fourth century, and consistently by Menander

So. The universal here seems to be more abstract than in chapter 9. There at the end of the fourth century.

universality rested on the necessary or probable connection between events; 90. Iliad, I . I Protagoras of Abdera was a leading fifth-century sophist.

here it designates the barest outline of a story, which only becomes a plot 9 1 . The difference is that animal noises cannot be compounded into

with causally connected events when the outline is turned into episodes. syllables.

The text of this parenthesis is uncertain. 92. The text is in a hopeless muddle here, and Aristotle's definitions of con­
nective and conjunction cannot be restored with any confidence. The re­
8 1 . Cf. 5 5a6-8 (n. 76).
82. Danaus ordered his daughters to kill their husbands; Hypermestra alone construction adopted here counts as 'connectives' (a) the particles, much

disobeyed, sparing her husband Lynceus; she bore his son, presumably keep­ used in Greek, which convey a nuance (e.g. adding emphasis, or highlight­

ing him and the child secret. In Theodectes' play the child must have been ing an antithesis) without changing the structure of the utterance, and (b)

57
NOTES TO THE TRANSLATION N O T E S TO T H E T R A N S L AT I O N

prepositions which link significant words together ('stab in the dark') , and century; his style was mocked by Aristophanes (fragment 1 5 8 Kassel­
possibly also co-ordinating conj unctions ('fog and confusion'); Aristotle's Austin).
'conjunctions' are then words which signal the articulation of complex 109. This riddle describes a doctor applying a heated bronze cup to a wound
utterances ('since the text is obscure, we can only guess', 'we must do what to draw blood; the cup would be kept in place by suction as it cooled.
we can'). 1 10. Eucleides is unknown. The two quotations grotesquely exaggerate a
93. Aristotle's term applies to any signifying word which does not express metrical freedom found in epic poetry. Epichares was an Athenian politi­
tense, including adjectives and pronouns. But in the discussion of style in cian at the end of the fifth century; hellebore was used in the treatment of
chapter 22, where verbs are also included, the term has reverted to its broad­ insanity.
er non-technical sense, 'word ' . 1 1 1 . Aeschylus fragment 2 5 3 ; Euripides fragment 792.
94. T h e element -dorus derives from the word for 'gift'. 1 12. These three examples are from the Odyssey (9.5 I 5, 20.259) and Iliad
95. For example, 'outcome' contains a non-signifying element (the connect­ ( 1 7.265).
ive 'out': 57a6- 1 0) , while 'homecoming' comprises two significant ele­ 1 13 . Nothing is known of Ariphrades. I have omitted one more than
ments; but in both cases it is the whole compound word which we treat as normally untranslatable example from the foregoing list of poetic usages.
significant, not its separate components (cf. 57a 1 2- 1 4) . 1 14. Cf. 49a24-8 (n. 1 8) , 59b34-8.
96. The Hermus, Ca"icus and Xanthus are all rivers in the region o fPhocaea, 1 15. The victories over the Persians at Salamis and over the Carthaginians at
the city from which Marseilles was originally founded; the less extravagant Himera in Sicily were said to have happened on the same day (Herodotus,
compound Hermocalcus is attested as a personal name at one of the col­ Histories 7 . 1 66). For the contrast between poetry and history cf. chapter 9
onies of Marseilles. The text here is uncertain. (5 1a3 8-b7).
97. Oydssey, I . I 8 5 . 1 16. Iliad, 2.484--?79; the catalogue relates to the beginning of the war rather
98. Iliad, 2 . 272. than its tenth year, in which the Iliad is set.
99. Empedocles fragments 1 3 8 and 1 4 3 ; the first quotation refers to a man 1 17. The Cypria recounted the antecedents of the Trojan War; the Little Iliad
being killed with a bronze weapon, the second to water being drawn off in a took up the story from the end of the Iliad.
bronze bowl or bucket. 1 18. Cf. 5 5 b 3 2-56a3 and n. 8 3 .
100. Timotheus fragment 21 Page (PMG 797) . 1 19. That i s , three tragedies, amounting t o 4,000-5 ,ooo lines; b y contrast, the
1 0 1 . Empedocles fragment 1 52 ; but the text here is uncertain, and we Iliad is over 1 5 ,000 lines long and the Odyssey over 1 2,000.
cannot be sure which phrase is being attributed to Empedocles. 120. Cf. 47b2 1 f. (n. 5 ) .
102. Source unknown. 1 2 1 . See Iliad, 22. 1 3 1-207, cited again at 6ob26.Cf. 5 5a22-9 for t h e care that
103. The discussion ofornamental nouns has dropped out of the Greek text; dramatists have to take over what is seen on stage.
Aristotle has in mind the poetic use of epithets (as in 'rosy-fingered dawn'). 122. Cf. 54b26-30 (n. 7 1 ) , and compare the discussion of false inference at
104. The source of 'sproutages' is unknown; 'invocator' (areter) is Homeric. 5 5 a r 2- 1 6. If the stranger is Odysseus, he will have a scar; but his having a
105. All these examples are epic forms: Peleiadeo occurs in the first line of the scar does not (as the Nurse assumes) entail that he is Odysseus. Some think
Iliad. The quotation illustrating ops is Empedocles fragment 8 8 . the reference is to the way the disguised Odysseus deceives Penelope just
106. Iliad, 5 . 3 9 3 ; Homer uses a comparative (dexiteros) i n place o fthe standard before the bath-scene (Odyssey, r9.2 r 3-60) : if the stranger saw Odysseus, he
dexios. will be able to describe him; but his ability to describe Odysseus does not
107. In Greek ps and ks are each written with a single letter. Different letters entail that he saw him.
are used for the long and short forms of the vowels e and o; the other vowels 123. Cf. 54b7f.
have a single letter for both forms. 124. Sophocles, Electra, 680-763 is a false report of Orestes' death in a chariot'
108. For Cleophon cf. 48a12 (n. 8) . Sthenelus was a tragic poet of the fifth race at the Pythian games; the irrationality in question is an anachronism

58 59
N OT E S TO T H E T R A N S L AT I O N N OTES TO THE T R A NSLAT I O N

(since the Pythian games were founded much later) . Aeschylus and of Homer the phrase in question occurs at 2 1 .297 but not at 2. I 5 , where the
Sophocles both wrote a play entitled Mysians, concerned with Telephus; reading is 'sorrow is in store for the Trojans'.) In Iliad, 23 .328 a different
because of the blood-guilt incurred by the killing of his uncle (53a2 1 and reading of the letters gives 'not rotted ' instead of 'rotted ' .
n. 43), he could speak to no one in the course of his lengthy journey. 137. Empedocles (see 47b 1 8 and n . 4) fragment 3 5 . 1 4f. Were the things
125. The text and interpretation are uncertain. ' unmixed formerly' or 'formerly mixed ' ?
126. In Odyssey, 1 3 . u 6-25 the Phaeacians put the sleeping Odysseus ashore 138. lliad, 1 0.252; the line may mean the majority o f the night, i . e two­
in Ithaca without his waking up. This implausible eventuality is contrived to thirds, or more than two-thirds - which in context creates a contradiction.
enhance his homecoming (it makes possible the striking scene in which 139. The examples are from Iliad, 2 1 . 592 (the 'tin' armour must be an alloy
Odysseus is at first uncertain where he is, and then learns from Athene that of tin, which can be called 'tin' in the same way that diluted wine is called
he has arrived home) ; Homer distracts us from the implausibility by (for ex­ 'wine') and Iliad, 20.234 (Ganymede poured nectar for the gods).
ample) switching our attention to a discussion between Zeus and Poseidon. 140. lliad, 20.272. A spear penetrates two layers ofbronze, and is stopped by a
127. The text in this paragraph is damaged, and Aristotle's argument cannot layer of gold; but the gold (being for display) would be the outer layer.
be reconstructed with complete certainty. 141 . Glaucon is unknown. The text here is uncertain, and it may be wrong
128. Cf. 6oa 1 4- 1 7 (n. 1 2 1 ) . to connect Glaucon to what follows.
129. Xenophanes, a poet and philosopher active i n the late sixth and early 142. If it is assumed that Penelope's father Icarius was a Spartan, it is odd
fifth centuries, was critical of anthropomorphic theology, and objected that his grandson Telemachus does not meet him when he visits Sparta in
strongly to the immoralities of the gods as portrayed in poetry. Compare Odyssey, 4; but the Odyssey does not say that he was a Spartan.
the arguments in Plato, Republic, 2 (377d-3 83c). 143 . For Zeuxis cf. 5oa26-8 (n. 25). The text of this sentence is uncertain.
130. Iliad, IO. 1 52f. The objection is to an unfamiliar way ofkeeping spears at 144. Compare the remark attributed to Agathon at 56a23-5 (n. 88).
the ready. 145. For example, Aristotle says that the best kind of tragic plot is 'complex
131. lliad, 1 . 50. The objection is to the triviality of the god Apollo paying rather than simple' (52b3 I f.) and that it is 'simple' ( 5 3 a 1 2f.); but 'simple'
attention to animals when he inflicts a plague on the Greek army; Aristotle's refers in one case to the structure of the plot and in the other to its outcome
solution turns on a similarity between Greek words for 'mule' (oureus) and (n.42), so there is no contradiction.
'sentinel' (ouros) . 146. In Euripides' Medea, 663 Aegeus' arrival has no necessary or probable
132. Iliad, 1 0. 3 1 6. The objection arises because Homer also describes Dolon connection with what precedes it; it is a coincidence, contrived to fur­
as 'fleet of foot'. nish Medea with an offer of asylum. For Menelaus' wickedness cf. 54a28f.
133. Iliad, 9.203. Greeks drank wine diluted; Achilles' instruction seems to (n. 60) .
turn a serious meeting into a drunken party. 147. At first sight there seem to be at least thirteen. The approach adopted
134. Aristotle means to quote lliad, 1 0 . 1 f. (the wording of which is slightly above (treating non-standard words and metaphors as variants of a single
different in our manuscripts of Homer). How can everyone have been solution based on the kinds of departure from current usage listed in chap­
asleep if Zeus heard music (lliad, 1 0 . 1 1-1 3)? ter 2 1 ) is perhaps the least arbitrary of the many that have been proposed.
135. Iliad, I 8.489, Odyssey, 5 . 275. Taken literally, Homer states that the Bear is 148. For Timotheus' Scylla cf. 54a3of. (n. 6 1 ) .
the only constellation which never sets ('alone with no share ofthe baths of 149. Mynniscus performed i n Aeschylus' later plays (in the 46os) and was still
Ocean'); this is false, but the Bear is the best-known of those constellations active in 422 BC, when he won the actors' competition; Callippides (men­
which never set. tioned again at 62a9- 1 1 ; cf. Xenophon, Symposium, 3 . I 1 ) won a prize in 4 1 8
136. Hippias is unknown (this is not the famous sophist, Hippias of Elis). In BC. Nothing is known about Pindarus.
lliad, 2. 1 5 Hippias changes the accent to make 'we grant' into the impera­ 150. Sosistratus and Mnasitheus are unknown.
tive 'grant'; this avoids attributing a lie directly to Zeus. (Jn our manuscripts 1 5 1 . For Callippides cf. 6 1 b3 4-6 (n. 1 49) . The objection is that his style of

60 61
NOTES TO THE TRANSLATION

acting robbed female roles of the restraint and self-control to be looked for
in respectable women.
152. See 59b3 1-6oa5 for the association between the dactylic hexameter
and extended narration. A unified plot must subsume a lot of actions to
achieve the length appropriate to heroic verse (cf. 56a 1 0- 1 9) ; even Homer's
poetry, which is excellently constructed (59a30-37), is diluted as a result of
this, so a fortiori other inferior epics will be open to the same criticism.
ISBN 978-0-1 4-044636-4

111 1 1 11 1111 1111 1 111 11


9 780 1 40 4,4636

You might also like