Mind AssociationThe Poetics of AristotleAuthor(s): R. P. HardieReviewed work(s):Source: Mind, New Series, Vol. 4, No. 15 (Jul., 1895), pp. 350-364Published by: Oxford University Press on behalf of the Mind AssociationStable URL: http://www.jstor.org/stable/2247536 .Accessed: 23/09/2012 12:16Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact email@example.com.. Oxford University Press and Mind Association are collaborating with JSTOR to digitize, preserve and extend access to Mind.http://www.jstor.org
V.-THE POETICS OF ARISTOTLE. BY R. P. HARDIE. WITHIN the last few years three important contributions have been made by English writers to the study of Aristotles theory of fine art, Mr Prickards Lecture, the chapters on Greek Art in Mr Bosanquets History of ?sthetic, and Professor Butchers Essays, accompanied by a critical text and a translation of the Poetics. It may not therefore be unfitting at the present time to call attention to a department of Aristotles thought that has received from specialists in philosophy less attentionperhaps than it deserves. The book last inentioned is theimmediate occasion of the following exposition, which is anattempt to bring together and state clearly certain parts ofAristotles doctrine, as found chiefly in the Poetics. I willmake no apology for following the text somewhat closely, sincea precise knowledge of what Aristotle actually says is anessential preliminary to understanding his theory of aestheticin its more general aspects, The first five chapters of the Poetics may be describedas introductory to the main subject of the treatise, the dis-cussion of tragedy, epic, and comedy. A definition anddivision of poetic art (7roctruc2) are given in chapters 1, 2, 3.In the fourth and fifth chapters 7rotqrt1c is considered teleo-logically and with respect to its historical development. The de$nition given is informal and defective, for it conta.insonly the genus to which 7rov?)rt/q belongs. Aristotle enumeratessix arts (or, more accurately, the whole of each of four arts andmost of each of other two) which he apparently assumes tobe the chief arts to which the word is applied. All these,he says, are kinds of imitation (1uti,uso). How then, accord-ing to Aristotle, does 7rotyruT differ from other kinds of qouts? We naturally turn first to his division of WOtqTUCin order to find data for the solution of this problem. Aristotles first principle of division is that of medium(ots or eV os pt,"toi3vVat),and he explains that all the species
THE POETICS OF ARISTOTLE. .351,of 7rot?rrt,ic use as their media one or more of the threepv9p,o", Xo6yo, A,ppovia. If then a mediumcommonand peculiar to the species of 7rovWrt1c) could be found, using that medium would be the required differentia. The obvious suggestion is that in general vp6vOpu, Xoeyog and 4p,tovta are modifications of the ultimate medium, what is audible (ro alovc-rov). But, according to Aristotle, dancing, which uses AvO,uos apart from aippovia and XO6yo, is included under Vrotlyruc, and it appeals to sight, as also do acting and the other spectacular parts of drama. Hence this suggestion falls to the ground, and we must look elsewhere -for the differentia of 7rot,it. One may mention in passing that the other arts referred to by Aristotle, which use colours and forms, are evidently painting and sculpture, especially the former (for sculpture is not explicitly mentioned in the Poetics, except perhaps incidentally in ch. 25). Both of these use vo oparvor, but the use of that medium is not peculiar to them. The second principle of division does not seem to helpus, as it is applied in both 7ypaotlcKand 7rov,qrt1. In Aristotlesview both these arts imitate agents (Qnpa`drrorra4)and maytherefore be divided according as the agents imitated are goodor bad in respect of character (9o4). Such at least isthe usual interpretation, which assumes that as nrpactg implies Mos, so conversely 9o9 implies 7rpa&st. But Aristotle doesnot actually say that o 7ypacfer imitate 7rpcaTz0ova: he onlysays that they imitate Vq. Lastly, Aristotle divides 7rotqr/c according as it imitatesall the agents c04 wpaJTTovTa9 or not, so distinguishing dramafrom the other species. In another passage he says thatin tragedy the imitators 7rotolvat T2)v F ad v by acting-7rpaTTVoTev. Hence plainly in drama not only is 7rpa2t thesubject imitated, but the medium too is 7rpaitS. This at onceenables us to solve our problem. For 7rpa2tv is a kind of CtV?70-L(Orpavtg = KIvflo.t + rcXog). And we may reasonablyconjectur6 that the use of 1c1vi7o-tq a medium is peculiar asto wovcip-TW differentiates it from other kinds of dpo-s andKCg7-t&Sq being taken to mean sensations of sight or hearing, thatare successive in time. This conjecture is confirmedby passagesfrom other writings of Aristotle, for instance Probl. 920 a 3, 8a71 0l pV8O/ot Ica6t ra teXl cfov?) ovic-a ?Oe70v EOtICE6v; v oTKtV20Et ei-a 6vc&r7rep /cat at rpa4ee"; Another passage quotedby Prof. Butcher (p. 125) is as follows:-8a TriTo acXOvcrorv t~ ~~~~~~~?/.ovoVO 09os 6Xet Tmw alTflTC05V ; v OT KLfl& X,XEL F4OV.,V,X,OUX oT cIM70rtV?eX 0p0^, OV%a av4 o oo ,Le as oecoePrb.. T919 eb7roV26 THr ToTdov oca0ab8avo, 0 70ebP}eeZ; (Probl. 919b 26). HIere povz can
352 R. P. HARDIE:hardly be used in its strict sense of alone, for according to thePoetics ro oparovhas 9o9, and if vo adcovo-rov can imitate Sosgbecause it has xcivs7r-1s,plainly dancing which has 1cxviwqa-t cando so too. Perhaps we might say that all the species ofvrovq-rt1c involve /vpos,, using that word in the wide sensewhich it seems to bear in 1Rhet.1408b 30 (tlO p,vOpv 8&6e/ev vra Xo7yov, pet7po 8 &u7 ?robt7ca tyap eo-rat). In thePoetics 3v9,to seems to be limited to rhythm of a moreregular kind. aippovia, it may be added, is not supposed to exist apart from pv6poq. Such then is the meaning that Aristotle at the beginning ofthe Poetics seems to attach to the word ?rotqTrc. It maybe expressed as the art of composition, 7roteLV in this sensebeing applied alike to composing a poem, a piece of music,and the steps of a dance (,rott Xopot avtro rTa o-XpaT 6VOLOVV, Aeschylus is made to say in a fr. of Aristophanes). Itis narrower of course than the widest sense of vrotetV whichis coextensive with all production or art, whether fine oruseful. But there is an unmistakable tendency in the Poeticsto limit still further the use of 71ot?qrT1c. A principle ofselection is at work which tends to separate the important casesof composition and to apply wotqptlc to these alone. Wesee this perhaps for the first time in the division of 7r0tfl8T1K??.The division according to medium, that is, to the use of one ormore of the three pvOuosg,Xo6ryo9,app0ov[a, yields seven possiblecases, of which two are (tacitly) rejected, since ap,lovta isnot independent of pv64tok. Of the remaining five Aristotleseems to group together two, Xo&0os, #tXos and Xo6yo9 e/JepEo,(of which the latter includes any pto-t,s that uses TerpoVexcepttragedy, comedy, and the dithyramb,) and to regret the want ofa word to denote this class. Again, in distinguishing, by meansof the third principle of division, drama from the other kindsof irotq7tic, he makes no use of the fact that drama usesmusic and 6tls in addition to X7yo;: he merely shows howdrama can be regarded as the limiting case of epic, when theelement of narrative is at a minimum. Putting together thesefacts we see easily that Aristotle might very well remove fromtragedy, comedy and the dithyramb their musical and spectacu-lar elements and class the remainder along with Xoryo9 ItXoand Xoryosl et1epos. -The class so formed would evidentlycorrespond precisely to the meaning of our poetry. Otherpassages, in the sequel of the Poetics, show, I think, quiteclearly that Aristotles thought tends in the direction of solimiting votq07tk. Of the three principles of division used by Aristotle the
THE POETICS OF. ARISTOTLE. 353 last two had already been distinguished by Plato (Republic, bk. III. a XeTc7eoI,cd,< XeIcTeov). The great advance made by Aristotle is thus the introduction of the conception of medium (vlx, in his metaphysical termninology). This conception neces- sarily modifies in an important way the meaning of ,^t4t4tlnt. If the special function of vtX is not recognised, the imitation of a thing will be regarded as an imitation in pari materid. Hence from this point of view, which is Platos, the copy of a thing must be either a mere repetition of the thing or it must differ from it as the unreal differs from the real, as appearance or illusion differs from reality. There is no other way of stating how they differ. But when it is recognised that two things having the same d8os may differ in respect of v5Xq, there is no longer any reason why the copy should be regarded as an attempt to rival reality. The imitation is simply the solution of an artistic problem:-Given xy where x is elBO9 and y vXq, to express x in terms of a new medium, y. The relation of xy to xy is naturally expressed by imitation or tt,pio-vt in its ordinary meaning. We may call the other relation, that of xy to x (or of xy to x), expression. Now both Plato and Aristotle use P lrpo-ts of the latter relation as well as of the former. In the case of Plato this is due to the fact that in his theory x the t86a is merely another concrete reality, over and above, and somehow external to, xy. But the case of Aristotle is different. He must have been aware, to some extent at least, of the perpendicular relation, so to speak, of xy to x as distinct in kind from the horizontal relation of xy to xy. We must therefore suppose that to him ,ttlbvs~ had a meaning much wider than that ordinarily attached to the term. In his attack on fine art, Plato insists that the artist has no real knowledge (7rto-r?7,uq) of the object he imitates. In other words the artist gives a merely factual representation of the object; he never sees behind the appearance, and copies the appearance merely, precisely as a photographic camera re- produces an object. But Aristotles doctrine enables us to putthe matter in a very different way. We may say that thereis first the concrete object, then the elBO9 in the mind of theartist, then its expression by him. And we may go a stepfurther. We may dismiss as unnecessary the given concretereality, and start with the E48o in the mind of the artist.Plainly, for a true theory of art, expression is the essentialoperation, not imitation. It might be objected to this interpretation of Aristotle thatit hardly explains the prominent position given to dramaamong the poetic arts. In drama it might be said we havethe art of illusion at its height; speech imitated by speech,
354 R. P. HARDIE: action by action. The reply to this is obvious. The caseseems one of imitation in pari materid, but then it is not imitation of a given object. So that just because it is imitation in pari materta, we -can be sure that the artisticinterest lies in expression alone. After all, speech and actionare themselves the chief cases of conscious expression in thereal world. Aristotles real view that fine art is the expression of theuniversal, can be illustrated from many passages in his works. When for instance he says that music is the most imitativeof the arts, he can only mean that music most perfectlyexpresses Oos,which is itself of course imperceptible to sense.And in fact, as we shall see presently, he in a way regardsmusic as the typical art, for he finds in the effects of certainkinds of music a guide to the special effect of tragedy. Anotherexample is the well-known saying that poetry is OtXoo-o&fx6repovthan history. For plainly in the former the universal findsfree expression, in the latter it is thwarted by matter, bymerely particular facts. It is hardly necessary to point outthat we must at the same time take care not to emphasizeunduly the universal element in poetry. We must keep inmind that for poetry it is essential that this element should beexpressed in matter of some sort. It is in this respect thatscience differs from poetry. The whole aim of the former is tokeep the E8oq abstract, and therefore science uses not EbKOvE9but o-qAEFa aov-p,u3oXa, or which never really express the Et&9 atall, but are of use merely to suggest the abstract E8o9 qudabstract. I may add for the sake of completeness that there is noexplicit reference in Aristotle to the artists sense of medium,that is, of the kind of expression appropriate to a given medium,for instance the way in which locks of hair should be representedin inarble. In fact there seems to be only one reference to thisidea in the whole of Greek literature (Athenaeus, XIII. p.603 E-604 D). For the abstract theory of vesthetic, there are two im-portant passages in the fourth and fifth chapters, which contain,as I have already mentioned, an account of the historicaldevelopment of poetry. The first is at the beginning of ch. iv.where the origin of poetry is referred to two aiz-iat cvo-tlcat(= psychological conditions, we might say): the second is at theend of ch. v. where the length (b,/coq) of tragedy is distinguishedfrom that of epic. Aristotle begins his explanation of the two airTat cfvo-caiby remarking that the tendencies to imitate and to takepleasure in imitations are innate in man. He illustrates the
THE PRkTICS OF.ARISTOTLE. 355 former by the fact that man learns his first lessons (saOioet9s) by means of imitation, the latter by the fact that we take pleasure in seeing imitations of objects which are themselves unpleasing, the reason being that when we see the imitations we take pleasure in recognizing them, in learning and inferring Icaz (ptav0avEwev o-vXXoryi2eo-at) what each of them is. Thus the two tendencies mentioned are made to depend on the natural love of knowledge. The ordinary interpretation of the passage identifies the acbdat Ovo-cal with these two tendencies. The objection to this interpretation is that these tendencies are not independent, but both are referred to the same acrka. Mr Bosanquet, following the reading of AC, takes the passages otherwise. He finds the two aitica in the tendency to imitate and to rejoice in imitations, and the tendency towards knowledge. A similar objection can be made to this view. Its two abrtat are not independent, and we should have expected Aristotle to refer the origin of poetry, not to two a-tbat, but to one, namely, the instinct for knowledge. If we are right then in looking for two independent acrua, we may be sure that up to the point we have reached inreading the passage only one actdta has been mentioned, the tendency to Fd,o-t9, regarded as a primitive way of knowing, in which the essence or EZtoqof a thing is apprehended, not abstractly, but concretely in the form of, say, a mimetic move-ment. And there is really no difficulty in finding the second-a-tla, in one of the following sentences, which begins: KaTa vowVv 0OV709 y/UlZJ TOv ,/A4E&c7,Oat ICalt rTq acpFLOv1acgab rovpv9iAovJ...Prof. Butcher translates as follows: Imitation, then,is one instinct of our nature. Next there is the instinct forharmony and rhythm... This interpretation of the wholepassage seems much the best. The objection to it lies, I fancy,in the abrupt way in which the instinct for harmony andrhythm is introduced. But I would suggest that the transitionto the second aibria is to be found in the preceding sentence,which is to the effect that when an object iinitated has notbeen seen before, so that the pleasure of recognition cannot bepresent, there may still be pleasure, which will be due, not tothe imitation as such, but to the execution (aqrepryao-ia), thecolouring (Xpoa) or some such cause. Here plainly two kindsof pleasure which are necessarily independent, are referred to,and there is no difficulty in supposing d7repyao-la and Xpoa tobe intended by Aristotle to correspond, roughly, in 7ypaq5t1)to apluovba and pv,4ts, in 7oty-rty. It is hardly necessary topoint out that Aristotles two atriat (to assume the correctnessof our interpretation) are the two instincts, the imitative andthe decorative, to which anthr"opologists assign the origin of art.
356 R. P. HARDIE: According. to some anthropologists, for instance Mr Andrew Lang (Custom and Myth, last chapter), the latter instinct is really the more important. My conclusion the.n is that Aristotle supposes poetry (,ulwqct,s ta TreXvni) to have been gradually developed out of instinctive mimicry (,A qo-tgs ata 4voEw9), having throughout as its aim the expression of order or beauty. Although in the preceding divisions tragedy has been clearly distinguished from epic, at the end of the fifth chapter a fresh differentia is given, the ground of division being the length of the dramatic jo-ts^, which Aristotle probablysupposes to correspond on the whole to the length of the 7rpac4t imitated. Epic is said to be aopwto9 TooXpovc, whiletragedy (the rpatv imitated, presumably) is limited to acircuit of the sun or a little more. The difficulties of preciselyinterpreting this passage need not detain us, and its realmeaning will be discussed more appropriately at a later stage.The passage is mentioned now, partly because it is the firstexplicit appearance of a consideration of value, partly because itis used in the definition of tragedy, to which I may now proceed. As Aristotle himself remarks, the definition of tragedyfollows from what has been already said. It is in fact adefinition resulting from a division, after the familiar mannerof Plato in his later dialogues. The definition as actuallystated, however, contains an important addition, a statement ofthe effect (egpyov or reXo) of tragedy. To understand theappearance of this new conception we must keep in mind thatthroughout the Poetics there is, as it were, a background ofteleological considerations. From time to time Aristotle com-pletes his thought by bringing forward from this backgroundconceptions that refer to the end of all fine art. There is, forinstance an obvious reference of the kind in the third line ofthe Poetics where he declares his intention of discussing how aplot is to be composed Et bEtXXet KaXw9 )etev v ohotL19fs, thatis, if it is to be a good poem, capable of attaining the r4Xo9of poetry. I have already pointed out that the length ofa tragedy is determninedreally by consideration of the good or,what is the same thing for fine art, ro caXov. Dismissing for the present all references of the kind, let usconsider the part of the definition which results from whatprecedes. If we neglect, for the sake of simplicity, the musical elementin tragedy, the divisions which lead to the definition maybe summarized thus: 7rovrTtiJis divided into that which uses,erT,O9 + the rest, the formerclass is again divided accord- XOMr/yosing as it representspeople who are 7Jrov8atot or 4aiv0Xot KaTa
THE POETICS OF, ARISTOTLE. 307To Oog,the former class again being divided according as therepresentation is dramatic or not. Hence tragedy (1) usesE1114ETpo0 Xoyo9, (2) represents ol T-7rov&asot, is dramatic. (3)Corresponding to these marks we find tragedy defined as the. 4t,ilwats? (2) 7rpateco TwvovSalta... (1) 8vp Eva Xo"yc... (3) wpGairrwvKat ovi &7 a7raryryEtas. The 7rpait9 represented isfurther described as reXela Ica&t ,.eLEyedo%ovoa, a mark derivedfrom the discussion of the 0I/cog of tragedy. The 7rpats isTeXEta becauseit is Wptoypevq and Aristotle adds that it musthave (some) magnitude because, as he afterwards explains,To rEX0eVov may have none. TEXEIa 7rpa4,&9 may be said tobe to 7rpaS as 7rpac2v? is to ict&Vyst 7yEVeO-&9).It means (ora whole life or a briefer set of events that determine the weal orwoe of a lifetime. If I am right in supposing that Aristotles method ofdefining tragedy is intended to be formal and scientific, itobviously follows that the word arov&ata must have the samemeaning in the definition that o-rov&aJio bore in the precedingdivisions. It would be fatal to scientific procedure, if, while wewere proceeding to the final o-vvayr),y,?, one of the terms usedin the division should in the meantime have changed its mean-ing. If technical words are to be used at all, their meaningsare to be determined, not from the ordinary use of language,but as it were from a set of precise equations. But we findthat the traditional translation of the word o-wovSaia in thedefinition of tragedy is serious not noble (or something ofthe kind). This seems to me to break the rules of scientificprocedure. In fact one might say that it introduces a circleinto the definition, for the meaning attached to avoov&ada ispractically derived from tragedy itself, the word that is beingdefined. The chief argument for the traditional translation of theword is, I suppose, that it brings out adequately the impliedantithesis to comedy (Butcher, p. 224). The reference is to adefinition of comedy, prefixed to the history of comedy inchapter 5. - This definition fits in very ill with the context, andlooks like a later insertion. It is hardly credible that Aristotleshould have given at this point a definition of comedy, when hehad not prefixed a corresponding definition of tragedy to thehistory of the latter. The definition of comedy, or rather of thekind of character or action represented, is to the effect that itrepresents a species of To 4aiXov, namely, T6 7Eyotov. But notonly is there no corresponding limitation of To -rov3aZov in thepart of the Poetics preceding the definition of tragedy but, aswe shall see presently, there is no distinction like that ofserious from noble in what follows, where the 1Oog and
358 R. P. HARDIE:7rpatl represented by tragedy are discussed with considerabledetail. One cannot suppose for a moment that Aristotle used adistinction of such importance without explicitly referring to it. Another argument in favour of the usual translation of o-rrov&adais that the transference of the epithet from theperson to the action is a matter of no small import (Butcher,p. 218), effecting, that is, a change in the meaning of the word.Now it is hard to see how this can be maintained in the face ofAristotles reiterated statement that the 7rpFt9S has the same0tolrTp as the 0o which it displays. The lepryetattoo thatproduce the gt9 are the same in quality as the eltl. No doubt -7rov8a-o0 does mean serious. But when once the word wasadopted as the recognized technical or semi-technical adjectivecorresponding to dpETr, it could not at the same time retain allthe meanings that it had in ordinary use. Much more evidenceseems to me to be required to establish the customary renderingof o7rov8ata. According to Aristotle there are two airtat of 7rpa-tq, 460Sand 8davoua,of which the former is inore important, as indicatedby the explicit reference to it by means of the adjective a7rov- Sata in the definition of tragedy. The subject of tragedy maythus be said to consist of three parts, 7rpa~tqS, Oos, and Sta&vota.I will consider briefly the last two, before proceeding to discussAristotles account of the first. Aristotles treatment of these two parts can, I think, bebrought under a general principle, which may be explained asfollows. The word "npa`tS is most frequently used in thesense of a visible act or line of conduct, and such a 7rpaitS maybe said to have an obvious meaning. On the one hand, it isnot a mere event, but is the act of an intelligent agent, who hasan end in view; on the other, it has a certain quality, or, inother words, the fact that a particular -re`xois chosen implies acertain character on the part of the agent. (In a less strictsense, the act may be said to have intellectual quality too.)Now, besides the visible vrpatq, in the case of ordinary conductinvolving two or more people a certain amount of Xoyosq, orexplanation to one another, of their thoughts, feelings, etc., isnecessary. Still further in the case of Ft4lot4vkst, 6oyosq ex- orplanation of a new kind is required to make clear to theaudience what is going on in the minds of the agents; for thatis not expressed sufficiently in either 7rpa5tq or the naturalkXoyoq between the agents. If then we subtract the visible7rpa5tS which explains itself, the rest of the ,xno-tqswill consistof parts which will be expressive or imitative of what wouldnot -otherwise be manifest (Oavepov or SAXov)to the otheragents in the 7rpa&tqor to the audience. This residuum of
THE POETICS OF. ARISTOTLE. 359what would otherwise be unexpressed taken as a whole isequivalent to x7os + tatvota. The expression or imitation of the 7rpa&tq is called thepuilo9. But there are no special words for the ,4.s, of r0Ooand of Stavota. Hence both ?Oo9 and tahvota are used in ahighly ambiguous way, and an essential preliminary to under-standing the passages in which these words are used, is todistinguish clearly between the two senses of each. Both Oos and Stacvota are implied in a way, as I have said,in the visible rpailtS. We have only to ask then in what partsof the Xo6yo9is each of them to be found? Aristotle naturallyseparates 7rpoaipeot9, which is the immediate mental conditionand equivalent of ?rpa4tq, from other mental states. Its expres-sion or revelation is 76oq in the second sense of the word, in thesense of p,wo7G-tq i6ovm. The most important passage bear- ToViing on this is translated by Prof Butcher (following the ParisianMS. AC)thus:- Character is that which reveals moral purpose;it shows what kind of things, in cases of doubt, a man choosesor avoids. A dialogue, therefore, which in no way indicateswhat the speaker chooses or avoids, is not expressive of cha-racter. In the first sentence 4Oo0Sseemnsto be taken in thesense, not of gtU?Jo-tq ToV i7Oovs, but of iOo& itself. (See pp.117, 315.) The objection to this rendering is that it hardlyagrees sufficiently with Aristotles actual ethical theory. Weshould rather expect him to say that 7rpoaipeocr9reveals Oo9.Besides, in cases of doubt seems too vague for ev ot0 ov)c 60S-Xov which would naturally mean in cases where it is notplain (i.e. to the spectators) what the agent is really choosingor avoiding. For these reasons the reading given in Christstext seems preferable1: it might be paraphrased as follows.0o9 (i.e. ,ptiApqO-t ToV i77ov9) is what reveals wrrpoatpeotq,sothat those Xo&yot have no i6oso where it is not made manifestwhat the man chooses or avoids, or where the speaker is notchoosing or avoiding anything at all. In a word, in the firstkind of case the 7rpoaipeot,l is not expressed, in the secondthere is no irpoatpeo-tq to express. The third and final account of Ooqs occurs in ch. 15, but thedetails there given do not concern us. 3t,avota is explained, like i3Oo9, three passages, of which inthe last is most explicit, while the second is hardly differentfrom the first. Part of the second Prof. Butcher translatesthus:- Thought, on the other hand, is that whereby we provethat,something is or is not, or state a genieral maxim. Again 1 EOT & o /L ETV ro toWTrov 0 ?Xo0 rT7v lrpoapEcrv &OVITp OVK EXOVuaLVMOov r6W Xoyiv, EV OlS OVK (TT aiXOV, oirotartv 7 7poapeTat I (EvyaE, 7 EvOtt ,178 OcAwC tTUv o0 rT 7rpoatpELTrat 1 (IEvyEt o (A) V.
360 R. P. HARDIE: it seems preferable to take &tazvota in the second of its two possible senses. The sentence would then run thus:- &avota (=U11770-l Tr-S 8tavoiacw) is where they (= o or wrpaTTOvTESI XBEyovVTes; word in the text is adro8EtKcvV5ovo-t) the prove that something is or etc. The third passage gives the fullest account of 8tacvota. The key to it is to be found I think in the last sentence of it. Literally translated this sentence is as follows:-For what would be the function of the speaker, if a 8eSE 8&aor Wa (or 8etva avra etc.; the readings vary) were at once (or already)manifest (4avoiTo) apart from what he says. I take this to meanthat Aristotle is here applying the general principle which Ihave ascribed to him, the need, namely, of explaining to theaudience certain things which they could not otherwise know,and which they require to know in order to understand fullythe 7rpaitq that is being represented. Interpreted by this lightthe rest of the passage is plain. It is to the effect that thefunction of 83alvota (in the second sense) is ro re acro8&Evv?vatIcat r XveZv Ica r wraOq -7rapao-KevaTEv, i.e. to prove and torefute, and to supply (to the spectators) 7raJ-6 i.e. the ?raw-WofOt XryOVTr, and not as in the ordinary view to excite the feel-ings for instance of the audience. 8Saivota of course doeseffect the excitation of the feelings of the audience, but itachieves this, not directly, but by means of expressing thefeelings of the agents. 7ro vTaO1 vapaa-Keva&etv might verywell mean to convey an impression, but it could hardly beused with precision to express the excitation of the feelings,for which some such word as lctvetv or J1/roteLEVperhaps wouldbe required. This interpretation seems to be confirmed by thephrase in the immediate context, oo-a v7ro rov Xoyov 8fl rapa-TKlcevao-7vat. It is confirmed also perhaps by the mention not only of gXEoq and 06/3oq but also of opyq Icat oaa rotavra as the ra-dO be expressed or excited, for AXEoqand 6,/38o are the to only feelings of the spectators that Aristotle consistently con- templates; but possibly the reference to rhetoric, under whichart 8Stcvota is said to fall, would suggest the inclusion of other besides these two. In a word then I take Aristotle to -rad-O,mean that 83acvota expresses everything that goes on in themind except 7rpoalpeo-ts. It is hardly necessary to add that insome cases nra-Oo9 a visible emotion or directly audible by isinarticulate cries. or Aristotles account of the pvJOo9 plp7no-v? ? ?rpatew9 T9presents to us two points of interest, corresponding to the twoteleological considerations that enter into the definition oftragedy. The other five ,ep of tragedy may almost be saidto exist for the pVVos,which is itself in turn the immediate
THE POETICS OF ARISTOTLE. 361instrument by which the effect or effects of tragedy are pro-duced. These effects are of two kinds. First there is theeffect of To KcaXozv,which is an end common to all fine art, andsecond there is the special effect ( peyov)of tragedy, the gcaO-apo-t9 of the feelings of pity and fear. The former effect is discussed in ch. 7, 8, 9. He thereexplains the conditions laid down in the definition, that tragedyshall be the p,1po-tv of a wpat9, that is (a) TeXeIa Ica (b)/4y0os eXovo-a. These conditions are independent, for a thinghe says may be OXov(= TENetov) without having pe,.eo9. Thefirst condition of To KcaX)ov simply of course that every work isof art must contain unity or order (r7at,s) in some degree. Itis found at its simplest in Av0,ct69, the natural love of which asI have explained is one of the two fvatlcat a1rat that give riseto fine art. This condition Aristotle explains with directreference to poetry, whose AvJ6os he says must have a begin- ming, middle, and an end, in order to be oXo9. a His explanation of the second condition (b) of so KcaXov ismore interesting, as it is founded on a comparison of poetrywith painting. The reference to painting, it is true, has beendoubted, but, I think, on insufficient grounds. It depends onthe interpretation of the word ?6)0ov, which is Aristotles usualexample of a homonym. The two contexts from which themeaning has to be determined are these:- EnrelTo KcaXovca?iov Icat airaav -77pay,Aa o OVvEo-TipcKEv el Tvwv OV vov ravTa ovpTerarypeva Me xet%e*... and (oZre Mt Kca0a7-ep e7rt TWv 0wpaToU v :/rat e7rt roIv G oZ T o Icat er vi-vO ..ov.... The meaningof 7rpaeyua b. -vVO-rT7lKcV EFC ri-tvzo and Ta acoara seems plainfrom other passages in Aristotle, for instance De Animan,412 a11, where he identifies ovio-ia co ovv0eO&q (substantia composita)with ao-4ara, and divides these into ovacaa o-4aTa and therest, the former class again being divided according as they are oreIJAIPvXa advXa. Thus animated bodies would seem to becomposite in the fullest sense of the word. 3ov then in thepresent passage in the Poetics must be equivalent to picture,in which sense however it would naturally suggest to a Greekthe picture of a g6ov in the sense of aw3pa ,evXov. Otherreasons could be given for this interpretation of giov, forinstance the phrase in ch. 23, tv 6o-rrep 63ov etv bXov 7rot, rwotKcetav 780ovn?v. would surely be absurd to talk of an animal Itas giving an otKeta owlOZ.Again, the word ryerypappeva inPhaedrus 264 c (quoted by Prof Butcher, p. 177) really putsthe matter beyond doubt. Assuming then that Aristotle has in view a comparisonbetween painting and poetry, we find a remarkable feature inhis treatment of the former. We might expect the fact that M. 24
362 R. P. HARDIE: painting does not involve movement to appear. But on the contrary we find that Aristotle reduces the p ye6o9 of paintingto terms of time, to the length of time occupiedby the K1CJ1v7o-tof which the perception(abo-Ofqo-L or Oewpt[a) of the pictureconsists (cf. De Anin4d, 425a 16 where, discussing the Kot?va?air-O?ra, he says of them, rav-a yap wravTa ?Icivqet alo-Oavo,ubtEa,osov /eLe7OO KctvE rt). In the Metaphysics (1078 a 31) Aristotleexplains that rd icaX)v is wider than ro adya0ov because itapplies to ra aic1virTawhereas the latter does not, and goes onto divide 70 icaXov into three eta17, Tatg, -vpltzerpta, and Toopta,Uevov. Of these ovpLerpta would apply chiefly to mathe- matical figures and relations, and in art to architecture and to the motionless relations involved in sculptuire (see for instance Galen, de plac. Hipp. et Plat. 5) and perhaps in painting, while ra~t, would refer rather to order in time, ro 69pto-eov being perhaps common to both space and time. However that mav be, Aristotle in the present passage of the Poetics reduces the beauty of the object seen to the Tas implied in our perception of it. Since however the time taken in the perception of a gcov is relatively short, a gjiov that is KIaoJvmay be described as ovfvvo7rTov,while an effective tragedy must have a length that is EtV,V9LOYUevrov. I may point out that this way ofregarding the beauty of a picture naturally reminds us of DeAnimra 424 a 27, 426 b 3 where it is stated that all (efficient) at!o-?Octqdepends on a proportion (X6oyo9). If the stimulus (r7rotoiv) is too great, the XoeyoT dissolved and perception isceases. ro KcaXov thus corresponds simply to aC&9,o-tg its instate of highest efficiency, when its Jvepyeta is at a maximum.If we use intensity for the magnitude of a force or activitywe can express Aristotles meaning more definitely by sayingthat the Evepryetawhose intensity is greatest, is the mostbeautiful, so long as its unity (X6yos, radt,s) is preserved. It isto be observed that the doctrine of the Poetics supplementsthat of the De Animd, for while the latter mentionis only thefailure of aius0oft when the stimulus is too great, the formertakes accoint of the case when the stimulus is too small (whenthe 6tov is 7rauputpov). We can now see clearly how tragedy differs from epic. Thelatter is too long, so that the attention becomes tired, and tendsto diminish in intensity. Tragedy on the other hand is exactlythe right length, thatis, the attention can be stimulated to thegreatest possible degree. At the end of ch. 7 Aristotle definesthe due pkyeOo9 of tragedy: it must contain a pera/3oX fromone to the other of the extremes of 7rpaitts, from SvOrTvXta toeTvXta or vice versa. The aim of tragedy is declared in the definition to be the
THE POETICS OF ARISTOTLE. 363XaOapaovt of pity and terror by means of these emotions them-selves. Before Aristotle the aim of poetry had been regardedas either of two things-(I) to educate, chiefly to improve }Oosq,(2) to give pleasure. But on Aristotles principles Oo9 is notimportant in itself, as it exists merely for vpa~tis,and he seemsaccordingly to have intended adOapo-tsto express the viewthat some kinds of poetry may do good, neither in the way ofimproving the o&, nor of giving pleasure merely, but in athird way, by exciting the feelings. For feeling (7raWos) isclassed with vrpatv in opposition to Ooso, and is in a sense akind of a7rWdeta or EvEpxyua. Such -7raOoor expression of thefeelings must do good, since eVeryEta better than mere is vva,ut?. Besides, it will make for dperT7, the alternative is forirregular and violent outbursts of feeling, and it will givepleasure too, for all &vEppetais accompanied by pleasure. ButI will not discuss further the question of the meaning oficaiOapc7t, a subject treated at length in all works on thePoetics. Little is known of KaOapo-ts, beyond the fact that itis an idea taken by Aristotle from medical science, in the firstplace to explain the effect of a certain kind of music, and thenapplied to tragedy. The immediate end of tragedy is the excitation of the feel-ings of Xeos and foR/3O. This end Aristotle of course simplyassumes, and then proceeds to investigate its bearing on theplot, and, through the plot, on the 4Oo?S represented. I do notpropose to enter into the details of this investigation, whichbegins in ch. 9, and occupies (with a digression on o? con-sidered in itself apart from plot) the following eight chapters.Its main results may be stated briefly as follows: The p6era/3oX?) represented is now determined as a changefrom eivxia to 8vorTvyta. In general the latter will take theform of an actual visible 7racOos, such as otv Vj T)avep a&avarot. The tragedy will be more effective, if the ,/ekTaGoX)77is a 7rept7reTEta, which Aristotle defines as 7 Et" TO evaVTtOV OWv7rpaTTrTo/EVW0V /6eTa/30X?, a change by which a train of actionproduces the opposite of the result intended. This again willbe most effective, when it is accompanied by an awakening orrecognition (avayvwptats@)on the part of the hero, for instance,when he finds out that a person, previously treated as anenemy, is really a friend or relative. A visible ra-Oo9is notrequired; the purely mental horror of a situation will achievethe required effect. The nature of the V9os? represented isdetermined by the consideration that we feel f06/3o only inconnection with people like ourselves. The hero, therefore,though good on the whole, must not be aoo/pa wtretK?c,,andmust fall into misfortune & a,4kaprltavTva. 24-2
364 R. P. HARDIE: THE POETICS OF ARISTOTLE. We are now in a position to make clear the only remaining -question of importance, namely Aristotles doctrine that Oos,is subordinate to p3v9os0. This doctrine is supposed to be at vari- ance with modem theory, and has been more adversely criticised than any other part of the Poetics. Mr Bosanquet considers. the question at length. He quotes the passage in ch. 6, and the following words may perhaps be taken to represent his, main conclusion as to Aristotles meaning: he may not have been contrasting the plot, as a mere puzzle and solution, with the portrayal of individual human character, but he may rather- have intended to oppose the man as revealed in action, or in speech which contributes to the march of incident, with mono- logue or conversation simply intended to emphasize this or that type of disposition in the interlocutors (p. 73). It is not difficult to see that this is a correct interpretation of Aristotles, view. According to Aristotle 9os, ought not to be separated from 7rpa&t. The connection between them is vital, for Oos, exists for 7rpa5t. Good Oos,is simply good that is, gefrom which good 7rpait4 is likely to result. The 7rpats, hasthe same quality as the Oos,from which it springs. Hence 420ets rparybtat (ch. 6) cannot mean tragedies that have no Oos,in any sense of the word, for the nrpa5~t,s represented mustimply Ooso a determinate kind, but must mean rather- oftragedies that contain few or no X6yot expressive of Oosv.Such tragedies would naturally be contrasted with tragedies,,that have no 7rpdets and consist wholly of pfetsOtlca4 i.e.Xoyot expressive of Oos, as it is in itself entirely apart from -7rpatq. But the most important passage bearing on the ques-tion is that in ch. 13, where Voa is discussed for the last timein its relation to /,tVO`os, where it is explained that the andmisfortunes of the hero should arise St a6papirtvav itaa. It has isbeen suggested that a6iap-rta intenitionally used, not aipaip-T,ua, to convey the idea of an error due to a flaw in thecharacter (cf. the distinction between ad&,caand ad8cq1,.ta inNic. Eth. bk. v. ch. 8, and the defilition of the subject of~comedy, Poet. ch. 5, where 6A,apTq4pr according to this sugges- a,tion, is plainty the right word to use). If this interpretation iscorrect, we have here a decisive instance of the inseparabilityof Pvi3os, and Oos. But in any case the drift of the wholepassage implies that iOo, does not necessarily mean to Aristotlea simple generic type (BIosanquet, 73), but that its complexity p.is precisely on a level with the complexity of the plot. At another time I may be allowed to discuss in these pagessome of the more general questions connected with the Poetics,,especially its relation to other Greek theories of aesthetic.