Shifting frontiers in ancient theories of metaphor

BARKER À . OFFICERS OF THE SOCIETY 3 1999-2000 AR President: VARA , \À M. D. REEVE \ % Council: (Ye C. F. L. AUSTIN (Treasurer), C. M. KELLY (Meetings Secretary), a J.S. WARING (Programme Secretary), T. MEISSNER, S. VOUTSAKI, R. B. B. WARDY Editors of the Proceedings: P. R. HARDIE, C. M. KELLY . | N er The objects of the Society are the furtherance of classical studies in general, and in particular the discussion and publication of critical researches into the languages, literature and civilisation of Greece and Rome. Any person interested in the ancient world is eligible for membership. The subscription of a member resident in Cambridge is £12.00 annually, except that the subscription of a resident research student or senior member (aged 65 and over) is £7.50. The subscription of a member resident elsewhere is £7.50 annually. Members receive notices of all meetings of the Society and copies of the Proceedings. Any library may subscribe to the Society and receive copies of the Proceedings. The subscription for libraries is £15.00 per annum. The Society is responsible for two publications. / Loi Proceedings of the Cambridge Philological Society, containing papers read to the Society and other articles, appears annually. The Editors welcome unsolicited contributions. Contributions intended for the Proceedings should be addressed to either Dr P. R. Hardie, New Hall, Cambridge CB3 ODF, or Dr C. M. Kelly, Corpus Christi College, Cambridge CB2 1RH. Supplements 10 the Proceedings is an occasional series designed to accommodate works of intermediate size, i.e. of about 150 pages. Proposals for monographs to be published in this series should be addressed to either of the Editors. Applications for membership, and all other correspondence relating to the Society, should be This paper is concerned with one little-known but intriguing and conceptually promising episode in the history of Greek thought about metaphor. Remarks made by two distinguished scholars will help us to get some preliminary bearings. In ancient discussions of rhetoric, says D.A. Russell, there was ‘a sharp distinction between content (fo legomenon) and verbal form (lexis). With some hazy and uncertain exceptions, ancient writers on poetry also adhered firmly to this distinction’!. Qualifications are added later in the book; but Russell leaves us with the clear impression that no Greek or Roman theorist made significant concessions to any nonsense about the medium being the message; and that whatever may be true of isolated examples of critical practice, all general theories about the elements of poetry assumed that discussions of what is said can be conducted quite independently of discussions of how it is said. In so far as connections were envisaged at all, Russell maintains, it was in terms of a rather vague notion of ‘suitability’: many writers cite with approval the Gorgian slogan, ‘great words suit great things’? Nearly fifty years earlier, W.B. Stanford had made a similar point about metaphor in particular. The starting-point for almost all substantial Greek discussions of the subject is provided by the accounts given by Aristotle in his Poetics and Rhetoric, and especially, on Stanford's view, by the brief definition at Poetics 1457b, petapood ÒÉ EOTLV OVOLLATOS GAAOTOLOU ETLPOEG, ‘Metaphor consists in applying to a thing a word that belongs to something-else’ (Stanford’s translation). After some pages of discussion, he comments: ‘Aristotle’s definition never suggests that the GAAOTELOV Svottc adds its share of to Eévov to the meaning as well as the diction of its context’. Metaphor is regarded, almost everywhere in the theoretical tradition, as a subject falling wholly within the domain of A€Etc. Stanford identified just one exception to this rule, in the writings of Hermogenes of Tarsus, a rhetorician of the second century A.D. What Stanford calls Hermogenes’ ‘vision of the true faith’ is expressed as follows. Toom) ott TO wij EE Úroxeytévou TMEGYLAtOS GALOTELOU dE OLAVTLAOV Övona Oetvar, HOLVOV eivat SUVEULEVOY Kal TOU UMOXEWLEVOU AO TOD ÉEWDEV Énpaivoné vou, Ö XUhETTOL AOL HETAPODA TUK Tog ypaunortıxoic.* Stanford argues that this constitutes a ‘significant challenge’ to addressed to Dr C. F. L. Austin, Trinity Hall, Cambridge CB2 1TJ. ! D.A. Russell, Criticism in antiquity (London 1981) 4. > Ibid. 6. 130; compare DK 82Al. 2 W.B. Stanford, Greek metaphor: studies in theory and practice (Oxford 1936) 18. + See L. Spengel, Rhetores graeci (Leipzig 1853-6), H 254.

London 1941) 4 1 See L. Spengel, Rhetores graced (Leipzig 1084-0), 1284. ? Ibid. 6, 130; compare DK B2A ‘ WB. Stanford, Greek metaphor: studies in theory and practice (Oxford 1946) 18 I DA, Russell, Crigteism in antiquity tod btoxeutevon xal tod EBwOev bupoivapévou, 6 nahelraı Kal weropood mage rolg yooppatimots.” Stanford argues that thin constitutes a ‘significant challenge’ to nodyparos dAhorgiou dt onravemoy Övona Oetven, nowov elvan duvanevov zal Tarsus, a rhetorician of the second century A.D, What Stanford calls Hermogenes’ ‘vision of the true faith’ is expressed an follows, Toom) dor vd pi OG broxmpévou just one exception to this rule, In the writings of Hermogenes of Stanford identified is regarded, almost everywhere in the theoretical tradition, as a subject falling wholly within the domain of ACs, its share of TO Bévov to the meaning as well as the diction of its context ', Metaphor he comments: ‘Aristotle's definition never suggests that the GAAOtQLOV Övopa adds of discussion, that belongs to something else’ (Stanford's translation), After some pages bot òvóparos Gd kotolov Énupoott, ‘metaphor consists in applying to a thing a word especially, on Stanford's view, by the brief definition at Poetics 1457b, wetapoge dé subject is provided by the accounts given by Aristotle in his Poerios and Rhetoric, and in particular, The starting-point for almost all substantial Creek discussions of the maintains, It was in terms of a rather vague potion of ‘suitability’: many writers cite with approval the Gorgian slogan, ‘great words suit great things”, Nearly fifty years earlier, W‚B, Stanford had made à similar point about metaphor discussions of how it is said, In so far as connections Were envisaged at all, Russell Qualifications are added later in the book; but Russell leaves us with the clear impression that no Greek or Roman theorist made significant concessions to any nonsense about the medium being the message; and that whatever may be true of isolated examples of critical practice, all general theories about the elements of poetry assumed that discussions of what is said ean be eondueted quite independently of content (to legomenon) and verbal form (evi), With some hazy and uncertain exceptions, ancient writers on poetry also fdhered firmly to this distinction‘! two distinguished scholars will help us to get some preliminary bearings, In ancient discussions of rhetoric, says DLA, Russell, there wan ‘a sharp distinction between promising episode in the history of Greek thought about metaphor, Remarks made by This paper is concerned with one litlesknowA but intriguing and conceptually

141-357), Nelson Goodman, Metaphor as moonlighting’ (75-80) See particularly Paul Ricoeur, The metaphorical process as cognition, imagination , twenty years after the Chicago yinposiim) that Stanford's simple formulations are ade Metaphors mean’ (Sacks 2945), Davidson's arguments are vigorously eritieized by but their essays will also help to undermine the Illusion (if any still remains, sixty y positions of views such as Sunford's are most explicitly ehallenged in Donald symposium in 1978, and collected in $, Sacks ted), On metaphor (Chicago and London| * This will be obvious to anyone who browses, for instance, through the papers delive RC ma Chia actions may quite properly be represented, on his view, in sharply CORErasing yays on different occasions and for different purposes, Secondly, we find in this author, have offered Stanford and Russell much more substantial traces of a Chillen Aristotelian tradition; but ihey seem to have overlooked him, At the simple st level, this theorist is unimpressed by the notion, implicit in Russell's ‘Gorgian | an, (hat a given kind of action or event has some straightforwardly identifi greatness or whatever, for which the poet must seek ‘suitable’ EXPLAiHy 6 been removed, I can find nothing in the text to encourage Stanford’s re, ‘significant challenge’ ta taditional views, insisting’ that the replacem description by a metaphor involves changes in meaning as well as in die I can see it says nothing about these issues at all, and provides no oyiq vce | existence in Hermogenes lime of any account of metaphor which departs uhsanlıve Ways from Aristotle's, There is, however, another theorist, not usually noted for his originale , a phrase that corresponds to nothing in the Greek whatever, Once these dog ion. have them jointly, He certainly says nothing about its uniting them ‘in a compoy concept, matters, Hermogenes does not that a term used metaphorically nos in it significance’ the two objects, merely (and much more vaguely) that it nbelong | marlans’ (p‚ 14), Some of the looseness in this translation is harmless; hy some of it sentence so as to unite in its significance both the subject at Issue and thedraneous object of reference in à composite concept; this is also called Metaphor, he gram his clair that itis implicit in Hermogenes’ remarks certainly needs to be iwstioned The text's one direct reference to ‘meaning’ or ‘significance’, contained je single word onpavtınöv, Is evidently too slight to carry the weight of such an Ipretution The interpretation seems todepend, in fact, on elements of Stanford'str{Shion which (I suggest) are wholly unjustified, His version runs as follows, “It is Oblique Language when a term not re vt to the subject matter but signifying some extraneous object of reference Is intro red intoa Of metaphor in its own right, though many more recent writers Would div igs pur meaning and not merely on diction, expresses its unique force in language b, 18), Now Stanford's own view of metaphor is that its ‘unique power jy Int that of Meaning two things at onee’; and he finds in Hermogenes an important à ation of his own position, This Is not the place to debate the credentials of Stanlon analysis received views, and that Hermogenes, ‘by insisting on the effect of Mphor on co } (1450b12 ff) is the exprennion, bounveia, of u thought, through ® Particularly Rep, 102c-9Nh is treated unambiguously as a kind of Övope, Hence discussions of it fall squarely under the heading of Atkıs; and so they do too, as Stanford complained, in most of the other Greek literary theorists, The consequence is that a metaphorical allusion to a thin differs only in its verbal form from a teferenee that unen the normal, non-metaphorical name for it, its xúorov Övope, A metaphorieal expression changes the sense of what is said no more than does the replacement of the wuo1ov Övope by an equivalent foreign or dialect expression, a yA@rra, It is, quite precisely, a figure of speech, a special way of transmitting to an audience the same thought that the wúgrov Övona would convey The subject is pursued at some length; and both here and in the Rhetoric a perapood are classified under various headings, An Ove may be wiv — that is, the normal designation proper, in the relevant language, to the thing referred to = or a yh@rra — a foreign or dialect borrowing = or a petapogd OF a HOO LOE, or it may be rerouévov or Öxteraqevov or bepgonpivov or ölnkkaybvov. Phe point to which I want to draw attention ie the presence in this list of petapood, (1456619 ff); but it rapidly becomes clear that Dis main interest is in vópera. These words, Ott tig Övopaotag. The ‘parts’ of MIE are orougelov, aukkafh), aivdeapos, Övopa, Ofte, do0eev, mrnorg and Adyag, each of whieh is briefly characterized Mic, we are told too) to discuss Ete in some detail, been adequately explored in the Rhetoric, but poen on (un Indeed he does in the Rhetoric subject at all, only an appendage made necessary by the poxOnolc of the audience (3.1.5); and he describes It an mavra paveaala nal Mode ty degoe (3.1.6). In the Poetics he is very brief about dicevarer, On the grounds (1486032 11.) that it has strategy of a speech, he explains that the study Of helte In not genuinely intrinsic to the distinct. Thus in the Khetorie, after completing hie neeount of Sidevouws, the sense and presentation of these toplon leaves us In ne doubt that he conceived them as quite rather repetitious theoretical enunelations of the Inter Wadition, Aristotle's own suggest, it is clearly the Artstotelian seeounts of dtanola and levis which underlie the classic formulation in Aristotle's Poertes and Rhetoric, und us Stanford and Russell lhe distinction between thought or meaning, ununlly Bidou, on the one hand, and its verbal ‘dress’, diction, Weave on the other, goes bek at least to Pluto,® It receives its main business of this paper In to explore the very Interesting account he presents, But I also have à historical hypothesis to offer, 1 shall argue that the original formulation of Aristides’ account must be due to the Stolen, and that we are justified in perceiving here the footprints of a full-sente Stole theory of metaphor Let us first articulate a little more fully the outlines of (he ‘mainstream’ position disrupts the old division between content and dletion, thought and Its expression, The Aristides Quintilianus, a way of dividing up the territory of literary analysis which relocates familiar topics In quite untamiliar parts of the terrain, and in so doing radically

meanings and associations will amount to an examination of the meanings and assoclations of the words in whieh the metaphors are expressed, Aristides’ treatment is no exception, Direct references to words, and to the spewlal effects created by their indirect least, which are said to have the payehologioal influence which metuphors and other enables the thought to be conveyed more vividly und effectively, The thought itself h tradition, His account comes in Book 2 of the De musica, written perhaps in the thi be used in ethics century À, D, In setting off to discuss the details of how novou) can under the heading of Évvoux, and intellectual milieu Introduction can usefully be connulted, however, on issues surrounding the questionof the works Mathiewn, Aristides Quintitianus: On music in three books (New Haven and London 983), Mathie MOANING are both thoroughly Obscure and une huracteristie of Aristotje's overall apphach, and i seen elenr thy they failed to make any Impact on the dominant strand of the later tradition For the oxi, see Winnington-ingram’s Teubner edition (Leipzig 1963), There is a Iran ban nıy Grm: offered Vth MUNA Writings, vol, 2 (Cambridge 1989), which | believe to be preferable to thi Interesting view, That passage certainty deserves close serutiny For prevent purposes however, we nt PUL it asidde, on the grounds that its implications (if indeed it has any) about the power ofmetaphor wall SEEN thar what a metaphor conveys differs in meaning from what is conveyed By A ctrresponding Statement. But there are certain passages, notubly Aher, 3.2.13, which may indicate A stronger and n Aristotle: but I am not whol ‘tn king these statements I follow what I think is the usual interpretatioof CONVied that it does him justice, It is true that his discussions of the effects of metiphors focus| on their capacity to make Ideas vivid, or to give pleasure to an audience, and Hat they do| CIALLY of the mutter, Metaphors and related figures, like direct descriptions, are inevitabl formulations, Hence un expressed in words, and cannot be detached from their verbal attempt io discuss them will refer, inter alla, to words; any attempt to exemplify not be liken to imply that words are irrelevant to metaphor, in Aristides’ understandin This (perhaps alarmingly epigrammatic) way of putting the point should of © and huve become, instead, figures of thought, boundary between this discussion and that of Alig is, an I have said, quite emphat of spe indicated, For Aristides, it would seem, these figures have ceased to be figures apparatus on whieh, as I argue, he actually drew. ® Tdo not mean to argue that Aristides In er | here with a fully developed distinction between ‘sense’ and ‘reference’, as articulated by Frege) but I do mean that any philosophieally adequate analysis af his strategy would find it useful to deploy a distinetien of just that sort, The thearetionl paraphernatiaon which (I believe) he consciously builds, and whieh Lexamine below, ean luirly be said to have affinities with Frege's. But it in certainly not identical with it: and renders who persevere to the end of thix paper may be inclined to judge that the Fregean analysts would five served Aristides’ purposes better than did the Stoic themselves. His concern at 69,1 ff is strietly with the semantic aspect of language, and with words in their guise as vehicles of meaning, What the educator is looking for are passage here; we shall return to It later.) He denoriben It au naldevng bx tv evvoudy, education ‘from’, or ‘on the basis of, or even ‘eonstructed out of conceptions, The resources used for ‘influencing the soul’ Gpuyayayla) in the procedures he is examining are evidently quite different from those involved in manipulations of Aökıc (75,14 ff.), which exploit only the effeeta of the shapes and sounds of the words they present different conceptions of it; they have different meanings.” Perhaps the clearest indication of Aristides’ position comes at 69, 1-B, where he introduces the mode of education with whieh he is currently concerned, and lists the various figures that are useful for providing it, (shall not consider every nuance of the of referring to the same thing (70.201) have different effeets on their hearers because what is signified by the words (tdi onpauvópever, an 0. at 70,11, 71,2). Different ways shorthand, of the psychological effects of hari wards’, ‘sotenwn words’, and so forth, he also describes them, more elaborately, as the effects, not of words as such, but of Big, being due to the author's inattention or an accidental jumbling of the text, His handlin th of the figures is tied carefully and closely into his general account ol Évvoicu, and expliciily under the heading, not of Thus although Aristides sometimes writes, In what I take to be u perfectly intelligible which as I have said is deliberate and systematic, pointy at the very least to a modi fication of that position, It suggests that the difference between a literal and a metaphorical expression Is, or at any rate Inoluden, a difference between the !vvorau, the ‘conceptions’ of the thing described whieh the different verbal expressions convey, but of #vvora, There is no question of thi on the quite different topie of Mig (Gopovte and Oude, beginning of ch. 11 as a start get their own discussions later). He includes a lengthy disquisition on figures of spect all sound Metiphors, similes, synecdoches, «AAnyoglar, and various others, That COMMonplace enough. The surprise is that this disquisition falls wholly and qu it alters only the verbal form, Aristides’ treatment of (ch one long chapter clearecho of the same Platonic passage. He subsequently devotes 9), and parts of the chapters before and after it, to Evvoran, and emphatically marks t when he tells us that dvvoux is in all respects the leader, ngoxadıyyelraı — anothe ENPFemed is prior to the choice of means for its expression, is immediately reinforcel be characterized through its contrast with that of Aristotle, By subsuming metaphor under the category of Allie, Aristotle and hin followers implieitly deny that the substitution of a metaphor for a direet description affeets the meaning of what is said: like conception’) will stand on a different level trom the others, as that which is toh None of this is in dispute, The position Lam attelbuting to Aristides can mont simply direct our feelings about the death of Kuphorbun (71,20) and so on, the elements 10 which the poetic educator must attend are Evvout, ALS, Gpeovic and read as somethin OVO,» (65,234), Our expectation that !vvora (which we naturally education, he begins with what looks like u routine Hist, plainly Platome In its ancestry figures are designed, according to Aristides, 0 exereine, It in ‘harsh words’ that affect our attitude to Ares and Aphrodite (70,214), ‘words that trouble the mind’ whieh convey a particular impression OF Achilles” laments (71 14-155, ‘bright words’ that CONcelved ay being prior to and essentially independent of the chosen form d XPression, whether the latter is figurative or KUQIOV,’ U is here that Aristides Quintilianus departs most strikingly from the fawuiliu uses are abundant in the relevant pages of His text, Min words, in certain passages at will present the examples as sequencen of words; and any attempt to examine their The Only point of using a metaphor is that sometimes, for a particular audience,| 5 SHIFTING FRONTIERS IN ANCIENT THEORIES OF METAPHOR

stated, While the soul inhabits the distant upper regions, it consorts only with reason and is free from desire (mou), But when iin vunlonliy leads it to become attracted to things in this world, it needs and seeks a body, But again, all bodies — indeed ait natural things — are characterized by one or the other of à pair of contrary qualities, mateness and femaleness, or by some mixture of the two, In coming to oceupy a body, the soul acquires an affinity with this body's gendered eharneter; and this makes its and talk about modıynarcı ne ” whose ‘underlyil if material” they onpanvöneva question’, Ihe to which the could straightforwardiy be four id, Thus tl je ‘things INa descriptions refer, must be the same whether the description sire teal or indirect, In that case, plainly, the relevant difference between ad rect and a fiptitive description is not in the thing teferred to, Ww int he worde employed, thagh there will of course be differences there, It is in he ‘things signifie I" by the work, the meanings its ne, its meaning and here U ‘prychuggic® effects are is the idea of something, he dea representing te pets choice olwbject-matter When the notion of + as that lom which the introduced II 27, VOLO poet—composer begin: VS prc CESS O f construction, ont might NIPPOE that an Évvouct Passage is not at the leve of eduenttontentioned & in this = l verbal expression but al IC level of Whitis signified, and of the way in whieh that signifie ‘atio is related u sheobject Which Weing described. We might guess (though ris goes beyond (he txt! (the undenulle difference in verbal form, which is iney pio be CONNIE us a secondary invo tab consequence ol the differ ji maphor esentuly involves ence in meat ing transfer of meaning fro one omain to another | he words are Mlected only as convenient devices through i he achieve rich rat transferee! If Aristides’ approach in va es, a rather rada shift from the ut yparenily | familiar Aristotelian treatmer al t UAF Naast ve ways in ver zt of met yhe yr, there un! which the development might be justified but! hall not try 10 à late them yet, Aristides himself gives no justifieatiions, or not dir order toves what he is up to, we shall need to work ure detail, {ite ind ugh parts ¢ of his tey description remains (he to Évvoqpe, of which I shall say more below, of the word in other parts of Aristides’ text (46.25, 76.20, 44.31) no elear decision in possible. Evaluative implications are quite plainly atmehed also te the twa gsenrmrenves (68.19, 74.7) of the reluted word ‘conceptions’ without any implication tat they are evaluatively lowded, None of the others can plausibly be interpreted in that way; neither can an earlier oeeurrenge at 87.4. About the remaining three appearances 70,18; 72.7), The last two, taken in Iolmlan, might not unreasonably be construed un referring In the chapters with whieh we are concerned, the ward ovc urs even timen (65.241, 24, 67,16: 6N, 14; 69.1: 10 The clain that an Évvout ia ‘never’ valweeneumat may He a Marginal = butonly à marginat —exapgeration leaves, a plant that multiplies itself by stemerooting, a plant with a bitter flavour, The 4 buttercup might be conceived alternatively ae u golden flower, à tussock of green ‘conceptions’, Evvoran ICs not that in our conceptions of the same thing we privilege different observable parts or properties of it, The objects to whieh the #vvorau mentioned here relate are not complex, eonorete things like frogs or butlercups - where Why do these differences in people's tastes count um examples of differing explain, of the male-female polarity in different qualitative domains, with these natural differences, there are differences also in their !vvoncı. Aristides offers some simple examples: some people like bright colours, others dark; some enjoy sponding to those of its own body; and it thereby aequires charnoteristically male, generalized desires specific, It comes to love, value and admire things of either the male or the female kind, or things whose gendered qualities are mixed in proportions corre sweet flavours, others sharp or bitter onen = these being instances, as he goes on to doctrines on this topic, For present purposes the essential ideas of ch. 8 can be simply transformed, Thus the central dist metion between the t woli Nply ange the subject female or mixed dispositions and emotional tendencies, Ihe subject we are interested in now resurfaces (67,15), Since people are constituted eva he has in mind, he could « attached to them: and since the (quasi-Frege an) rte’ | each eseription is the same, this must mean that hed fferent form of s deser on offer eomptually different presentations of the same « rbj e et. In the case of melt jorsi and other ipurative forms, the object is presen ted through the filter, ut& it wen, à Evvaut (01 Bene vóge vov) Which is directly deseri (ht bference of the ' only of something elt ptive be ‘about’ those things, If that order to convey the ONMavöH evaluative attitudes, and an explanation of why they differ from one person to another. It sets out from an elaborate ttle essay on the woul, the second of three substantial passages (chs, 2,8, 17) in which Aristides unravels his engaging and sometimes bizarre ech that there I Whether he speaks literally or metaphorical Iv, then, his utterinees TH] in some sense not so, he would never Need Maphors at all; in ently ass about which the educator has views to convey, the mod natOF gs in question and so forth (69,58), 11 is evid Ihe bulk of the next chapter, oh, 8, ie devated to an account of the origins of says. is in all respects the leader, because without it there will be no choice or rejection, alprorg or puyy, of anything at all (65, 24-6), 10 sentation of its subject in accordance with some evaluative attitude. The !vvou, he una first key point, An Évvoux, in the relevant parts of Avintiden’ text, is never a purely descriptive, value-neutral conception of some subject, but always involves the repre and specifically whether to present it in a favourable or unfavourable light. This gives 7 love, for instance, or the story of Oedipus, Hut thin ie not Aristides’ focus, The original question, for him, is not what subject to present, but, given a subject, how to present it; SHIFTING FRONTIERS IN ANCIENT THEORIEEN OF METAPHOR some defile set of subjects ways of representing a subject It “influence the soul APPFOPHely; and he does that so by his choice of Evora suitable for the purpo The technique of education bx vvÀ vw ich, Aristides BOEK OFS two forms, ‘If 0 vre Well influencing the one can find the things algal jac À Cet on voue ve) (hat soul among (or ‘trom') the underlying naterial of the things inuestion (x ik Oroxegy vis u tay NOU HANS), we shall us hem: b IIE Ware at a loss, we shall track down the heee var by direct me aans (Hebbore)” (69,24); Nsary OF ne and these péOodor turn out to be such things as me hors, MHC Clreumbocutions

not innocent and direct; it is filtered through the Üvvorar we already have ol = bd THV Moyo xivoupévotg bvvorpaow bxdorot vox native can, Je dx tijs rovaútns ovvndelas te Kal peters !Radev fj eldalova xard rg Euv À Évavriav Pepaudoaca, negbxaon advrov yao À pugij roue ev bavri) wal elnovag | Kor en MOMOYV Hafboroers 100g rouwt <oby Tyetov> À] al rhv miro nn noden ev obv i dt évvowWy dv flo ovyxard0eors À ma Sncene abropnadelag xaredypotwy À Voregov perareodevroy’ al yao, 2 , t we commonly respond to things notjust by noticing them, but with actsof appre ssent, here called ouyxara0toers, to elucidate its relations with the others, and their = lainly have to do something s in Aristides’ account of literary imagery, We must start by looking rather > , passage at the beginning of 2.9, which apparently sets out to coop = { = a th a number of other rather tricky terms, If we are to understand the former, we shall In Aristides’ discussions of these matters, the word Ëvvora is used in conjunction red to, but to some conceptual tertium quid interposed between the two. ferent ways is due neither to the words as such, nor to any change in the object ience to conceive the thing referred to through the ‘filter’ of the evaluative eeptions aroused by the metaphorical image, we manipulate their eae notes the thing itself, The difference between the effects of referring to a buttercup = oon the buttercup, the metaphorical phrases simultaneously awaken. By snising rie fference merely in the audible, visible or grammatical forms of the words used to fer to it. It lies rather in the Üvvorau, the evaluative conceptions, which, in referring id the xÜQtov Övona ‘buttercup’, is not in the object referred to; nor is the re zen gardener's enemy. The difference between these designations, and me . n |. > Metaphors to the category of conceptions, !vvorau, rather than that of M Su, = on e may refer metaphorically to a buttercup either as a golden bowl, for a s begin to see here, in an embryonie way, a reason for assimilating — ings, and it is the quality of these Fvvoran that erucially affects au re = nvisaging brightness, yellowness or whatever, we include in out zo. an valuative loading contributed by ourselves; so that one person, conceiving the = our llow, has in his conception the notions of cheerfulness and light, while anothe ‘ 7 nceiving the same colour, presents to himself the Image of something bilious anc Of familiarity and praetiee, | IMperceptibly and gradually establishes a fixed ot everything, which arouse, Then if oc is effects they are capable of As to Övvonjure, we have subsisting ‘concepts’ whieh the soul envisages OF grunps, than independently or abstractly are psvehologieal states or entities, the soul's ‘coneeivings', as it Were, father and with images and impressions internal to the soul, It rather looks, then, as if in these two related terms, vvout and bvvona, we have references to items of two different sorts, performing distinet roles in the linguistic and conceptual transactions that Aristides js discussing, The evidence we have looked at so Far would suggest that !vvoran, being elosely ansoelated or even identical with On, producing, with which Aristides is concerned) and the Payehologicnl named by the new term, voip, The é wvorjeera are not the same as the Impressions and images embedded in the soul, but are instrumental in the remoulding of them; they seem to be Intermediaries of some sort between people's utterances (which will include, of course, the deliberately educati ve poetic utterances represented in or even constituted by the vor, are these perhaps identical also with the TÚTOL and elxdvee contained in the soul? Il no, they are apparently nor identical with the items it constantly remodels through the bwvorigere that Koyo talk, because the soul contains Brot and elves nor even Constituted by our Evvoutt, and ahange in one ensures OF amounts to change in the other, And now Aristides explains this elalm in ite turn, Oo can be affected by other people's least very closely linked, if not identical, Our fon, our evaluative dispositions, are reflected and Évvorc are at Oog: and since what he is talking about here are the origins of !vvoren, this suggests that doe dafonoerg, the chatterings of the many, can implant nv moAAdy by external suasion, In explaining this assertion Aristides makes the Ineldentul point thateven al conceptions that have been altered from their original form between things we meet and our original, natural constitution: or they may be affinities un obseure expression, but indicating, I think, the way in which we independently leam to recognize meet them, These !vvorar may arise bi’ CO POPOL l or rejection when we use af our vulue-louded conceptions of things Chat we respond to the things themselven with Approva , This now seems straightforward) it ie heea his {crucial paragraph, but it leaven the status of some of the central concepts employed, and the relations between them, In gertain respects unelear, The first claim is that our Ouyxara0toers, our psychologlenl gew of Approval or ansent, arise through Evvorat disposition (Hee), which may be a blensing or a eure, (2.9, 68, 14-22) sort CE OL) and imagen (elndvee) ol everything, and it constantly remodels them in the Light of À vvorjpere (hat are stirred up by apeeeh; and then, ax a result of this O5) as have the doctrines of the selenven, For the soul has in itself impressions unpleasant; and since the objects remain the same while the response is sms e response must be determined not by the things themselves, but by our ie et em, the manner in which we conceive them, Our perceptual experience of the things as y particular evaluative response; in encountering them we encounter nn - “= n themselves neither good nor bad, Nevertheless we do respond to them as de ig - : complexity, Aristides’ point must be that these qualities do not by ao 8 = a ~ through our own inherent capueity to leam, or have later been persuaded to chara ge, Here the chatterings of the many have mul power to instil character 9 Our initial approval CovyncerdOemig) of things in life arises, then, through eon eptions (Évvotat), whieh have either grasped individual things successfully ibjects are simple qualities like brightness or sweetness, devoid —

the thing, In the other it in apparently what in state, leading us into that of concelving a situation of chaotic turmoil, d the right ones not ‘in’, but bx, ‘out of", one among’, THE DrOKEYLeVIS Ding. In ier cases We must track down the ones thatare necessary, Ceveryxata., by indire ites. It is obvious that the onpawvönevov cannot sign or word as such; nething signified, not the signifier, But equally, in neither case is the onpauvóje vov The really interesting word here is onnaavöpeva, ‘things signified’. Sometimes we at is, through metaphors, similes, synecdoehes and the rest, Here the word ÖAr can working precisely in its technical, Aristotelian sense; what Aristides presumably ins is that in some cases We can describe the thing, with appropriate psychological fect, by referring to it plainly and directly, in terms that identify it simply as what it while in other cases an indirect form of reference is required, Direct reference is to made, | suppose, through what Aristotle calls the wor Övönara, While the indirect proach, as we have seen, involves ‘figures of speech’ = though that is evidently no iger the appropriate expression = of one sort or another. fluencing the soul, the poet discovers them dx tig drones TOY mOcypdtov we. In the other, we are to track down what is necessary by indirect routes, eOodot midevorg dx TOV Pvvoudy, and says that there are two species of it, It will be recalled at in one of them, when searching for the onpeuvópeve that are suitable for use in ready briefly examined, Aristides charaeterizes the poet/musician's art as that But there are other elements in play here too, At 09,12, in u passage that we have ical at the line expresses a concept well, The idea must be that it changes our psycho jean ‘to a concept of...” But the pi sage as a whole is concerned with techniquesof uyayuyla, ways of altering psychic conditions, It would not be to the point to say vvorav of the confusion he wishes to convey, By itself the phrase eis Ëvvorav night iyone, is not an aspect of me, No more In 1e the thing I refer to or an aspect of it sources for Stale theory, Sextus Empiricus uses it oe ification between onpctuvópevor and rente ould he made, In a Stole context, at this date, and In na reason why it cannot have achaient POHL. there in Sextus” cettique hang sen the ton we nothing —1 made also, or even originated, within the Stale camp 811-12, One might argue with some justice that Mextun In svurvely to be trusted on the Adv. math, of terminology, His tentimen in worth reuanding none (he leus. It shows, IF no more, that the delicacies agony of war, But it does so Oude 76 poflegdv, This ‘fearfuiness’ In the middle term inking the word and the effect: and it is neither a quality of the word in question, an articulate sound, nor the condition directly ouxutar, conceived mere obviously. In the present case, we ue told, the ‘word’ (dvopa) goin displays the what he could not have done on the basis of the ‘natural’ terms, bx tidy ward quot Ovopdrov, he achieven vividly = TOY one pop onavoplvav, It is not it through ompanvopueven, here things signified through words alone andas sue pre en effeets ure produced, Words are important, through metaphor, that paye xucveer odeolv TE wal Eye polen (Il 4,282) ‘wo other passages in Aristides also suggest (hat hie ompeuvójpever have some such sas this, The first is 70,101 1, Here we are told that in Homer's line onnanvönevov are explicitly said to be the same, 1) the more usual bit of jargon, Auwedv, lntredueed) and here Aextov and consistently throughout his passage on Stole semantien, Only at (he end of his treatment least one of our important onnanvönevov, is used of these non-corporeal, nonepsyehologieal somethings in at of reasons, but most obviously beenuse there may be no thing corresponding a number to the sense of my utteranee, It is worth pointing out that Aristides’ word, for while In that tate, His the meaning of the utterance vhich I might express my thought; and this meaning, whieh could be thought of by in chat which vr Lapooal, Nothing has been said to suggest that different souls envisage differen oneepts, or to explain why they should; and It in again more plausible to treat different voran as different psychological conditions or orlentations, which have indeed been counted for in the preceding discussion, Finally, in 70.19 If, he anatomizes the way | which a certain Homeric expression (4, 20,5960) is well designed to lead us els MOT Hard how dragóows brroxelvav nal TOV vvordy yivovel hextov is notabody, noreven a condition af one, Specifically, iin not the body referred the utterance, af the person making nor is Ita payehologioal candition to by an utterance, but the Acro of mine; state or disposition dy a peychologieal My thought or impression uxis Evvorett, and it is plainly ase is at 67,1516, where Aristides claims that differences in people's fvvoru: olv xplained by the fact that the soul's nature is different in different people, tadırav actions are the (hut is, about what utterancen mean, In Staleımetapliynion these items have a curious status. Unlike both the utterance itself and that of whieh it in true (if it is true), the On and naon involved ore natural locate these episodes of imitation in the soul's acts or states ol onceiving than in separately existing concepts which it contemplates, An other cleat ‘Nignitied’ by a metaphor or the likes and the kinde of Hgure that Aristides goes on to to be something ‘extractec discuss need not literally piek out any coal entities In the world at all, Hut it seems clear in any case, as I argued earlier, that the very same thing might be referred to in either way, directly or by pé0obo, and that in auch elreumstaneen the anpenvoneva will not be the same, even though the thing referred to In Aristides seems to be talking, at least roughly, alvaut what the Staten called Arte, mentioned frequently in our passage, and occasionally elsewhere in li identical with the thing to whieh reference Ie made, In one form of reference it is sald HPTING FRONTIER SIN ANCIENT THROIIEN OF METAPHOR ook 2, Some of the occurrences give no real help with the present problem, Others, owever, seem at least on balance to confirm the diagnosis I have just suggested, that vvora are psychological states, entities or conditions, actually existing in the soul, wodifientions or aspects of it, Thus at 57,4, we are told that in povouer what imitates !vvorat are to look at Aristide “handling of both ANDREW BARKER s yet insufficient information to decide, We need rms in other parts of his text,

lights, He points to words alte ejuretig woerd vógov achieves the effect of this prodrng, owû vc ev eness of the things signified’, ouura of By now Aristides has completed his account 4e > _ m naor: } Tete en reen Hyon plain The passage we joked at earlier, which included and elnöves, are explicitly or dispositions of the soul, or else = in the case of rror iy said is that they are What ra. Evvonjpa said to be contained in it, This is not said of ntal in reshaping Instrume are they that and aroused, yuvodpeva, by utterances, Aöyor, have a place between to then, appear, They soul, the in ned the trot and elwóves contai the psychological entities in a person the utterances, the spoken or written signs, andspeech, It seems a very fair guess that to listens or reads which are affected when he in Aristides, specifically meanings loaded with these intermediaries are Meanings, und n of words can provide the right sort of link betwee affective weight. ( Inly the meanings vópwva and bvvarwara are the same the utterance and its effect, Perhaps, then, opec ns, acts, states candidate: dvvorjard, All the others are either quite evidently conditio On), trot, coedveg, vworpvarro and ÜEers references 10 ovyxaraltoers, !vvora, only one plausible a whole barrage of semi-technical notions = seems to contain anything else that he mentions? a ee ’ they cite on the subject of beurd, and their dlseunaian dd mb | ne F4 : 7vy u Passages = ’ 8-202, subject, wand DN, Nediley, The Hellenistic philosophers (Camb idge "But substantial à Sedley's : note an help is given byv Long Long and 1987) 12 Alexander ; In Ar. ; ; Top, 489, 12-16,| StanStone 12-100; | Diogenes Laertlus 7.60-1, See passages vk A As ken OD. 10 and 30C renpectively.in nat pretend to understand all (hin? and fortunately we do his te 1 fer prreken > Diogenes Laerdus, and are neither HME nor mate, ennn ded: = ” | Wurts or pavELop dre diavalue, wong io Stobnous’ À Ach Du l'hey are neither para nor Codd, AMY ‘Alexander: À jen ged GNA are denied even that zaak sig of dachpard, incorporeal ‚ CvVOT status ee mn ol what is said — in the ease of _stateme Évvotuara are more like concepts, Pi emd eek an Philosophers would call ‘universals’, There lean iocinted mis M oestinetion, rg real bodily ns ent 2 a While Aexrd and onpawvópeve, though nat « si and Te ure standardiy, | think, Ihe "ble oo. sé à om | Now dvvonpa is another teehnieal teem In tole Frs shy: but in Stoic ivvoñpara are apparently not identical with heerd and ov eee j ~ a zum some pardeular mind, Hf then the Ivvoran very important, it is they that can be ‘useful’ Now the onpavópeva are evidently this poyayoryu. They do the work, But they are mentioned under or ‘necessary for with ed identifi be to long discussion, Are they then title only three times In Aristides’ identical with any item essentially located in not identical with the onnmvonmve, are states of mind, they are of figurative expression in Homer: the figures ave th En = En omens examples things as the charm of dawn, the lenrlulnenn of war, the Maas of acute 4 — - er he task af eftoetively ve these figures is to present an objeet or nellan mace f mo vn “according ze + — « Er ing discussion _ nn driv to the affective load they carry, That indeed WabAslotides’ that they are something like ‘meanings’, OHV Ee of metaphors and so on has been about: and we are again objects to which the words refer, They are, heverthelens Nas on à " un rare hin Don 0 sien = certainly obviously, are they the psyehotogieal condition, nd 88 ot = hot words, or groups of words = the term thought and impulse display the Adyag that ij ot een weh ad an often, Aviation’ diseumsten 1e za vs en eve (dyamorndg)', Here, elements lo the Larger cansteuctions that are built up OP HoH and th Mt en smaller these nich erento ee dore ars hyo) vel fall, In this case the larger construettons and cheerfulness give birth to simple und charming hi ” Po ‘ x hed pra köyor and their species) and he goes om dede — ra "vvoruara are Interwoven with one Ant thei u sand effects exemplified From Ihe works af Hamer vo nee und the like, the 100801 Houehed on earlier Pach evaluative light. It does this not through Hie werde simplyres an ye : h Nten = zen. the Hence clauses of ulfective associations of their meanings, referred be oun bvvot en to as classes of meanings, ONMEHVOLMVE, OF Concepts, oriauvöpeva are not rat el 1 à et ze: me, on the other as what is er by a bert bvvorjpara resurface explicitly in a later ang, at the uess en 7 - things viewed from opposite angles; ont SHIETING FRONTIERS IN ANCIENT THRON OF METAPHOR are they psychological states, They are the items referred to by words, and neither be his audience, and cannot therefore and poet the to “significations available equally semanties, In Aristides, then, as in Sextus’ account af Stole attention 1s directed, In its force, with Odyug, which is word puyfivan in line 133, treated earlier as negative ce in reference that gives differen u nat is plainly positive, in the next line, Once again It are evaluatively loaded nwva omjavo The mind, some in dvarmarng, nor anything and the things to which their concepts that mediate between the minds of the audience here of the onpawoptvov bvavuórnrog. He seems to be alluding to the conjunction by ‘an interweaving of the opposit Homer praised, and comments that ily neither blamed nor ofte brawougdvne dla, intercourse that in customar intermediacy, love to Brisels (//, 9,133-4), 10 describe this as a one of \ to bed with Aphrodite MAD) ffoyvve in the accountof Ares going such as pui mocv, 9), contrasting them with the cepver Oleta of the line about Odysseus (Od. 8.208 (Od, 23,296), He goes vn Abnrooro saaiod Oeojoy Wovro and Penelope, domdo that he has never made on, at 71.13, speaking of the verses in which Agamemnon says sort in different evaluative represents (WO events of the same ss’ of the serred ranks of spearsmen, It is referred to by that word, the ‘spikine its referent, an affective meaning or signinor word the something ditter ent, neither an object or state of affairs is conceived and fication through whose special colouring word generate in the soul a represented. Only through such an intermediary can the , referent its to particular emotional attitude the devices through which Homer Secondly, a little later on, Aristides is discussing

subtleties, rather than as a proof that the whole theory must after all be allen to Stoicism edgement is to be construed, | suggest, merely as a symptom of his lack of interest in refined ontological complicated. But the fact that Aristides’ account passes these problems by without the least acknow! (his is a conchision that Stolon who sought to understand the power of language could not embrace without further discussion: the difficulty had to be eireumvented, and the route around It was always likely to be to Aewrd too, though there the metaphysical issues are slightly leas problematic) is that they are incapable of entering into causal relationships, As the PCPS reader pointed out to me, this entails that they cannot net upon anything, and hence = it would seem — cannot produce changes In anyone's soul, Hence no one and nothing can (apparently) be affected or altered by the meaning of anything that is said, Rather obviously, See for example Aetius 4.41.14, and other passages collected in Long and Sedley's Section 39, with their discusston, It is used here, Lihlak, In just dre right manner and contest for a Stoke discussion, See in particular Diogenes Laerthus 10,33 (Long and Sedley 171) One consetuence of the peculiar ontological status assigned by States to bvvoryecera, (and that assigned troc: perhaps also 1Oog and ÉE15, though these are of course in much more general curreney, The only point at which Aristides’ usage, on my interpretation, is plainly at odds with that of the Stoies is over an issue so recondite and philosophically specialized that the slip is eminently understandable,!® Aristides, after all, does not write as a Stoic; Stoic terms and distinetions, The key terms are !vvora, anpawöpevov and Évvémper, As further Stoic borrowings we can certainly add owynarddroıg, and probably also Ëvvorau that are ‘natural’ and those that are culturally or idiosyneratically constructed. 1 Aristides has taken the topic ‘figures of speech’, which was regularly located in literary theory within the field of diction, Allie, and has transformed it into the topic ‘figures of thought’, whose investigation is concerned with the quite different and explicitly distinguished domain of conceivings and concepts, !vvorau and dvvorjpara, What I am suggesting is that this unusual and interesting tinsference is achieved through a moderately systematic and moderately faithful deployment of a collection of is equipped; and just as for Aristides these !vvoran may arise Ov aùropa0elas or the result of external persuasion and remodelling, so the Stoles distinguish between in the Stoie repertoire, and in a role very close to that in which Aristides uses them that is, as psychological states of ‘conceiving’, Acording to the Stoics, as in Aristides, impressions of external things are received through the !vvora with which the mind Turning back briefly to the subject of &vvown, itis clear that they too belong squarely they are the Meanings of sentences or the concepts that correspond to individual expressions, On the other hand they clearly belong, even in Stoleism, to the same general area of thought; they both exist = LE (hat is the right word (and It clearly is not) in a metaphysical limbo between words and things, and between words and the psychological conditions that their reception and comprehension produces, Both are, in a broad sense, meanings, and dvvorpara, and that this extended both to their semantic roles and to their ontological status, The difference between their roles, at least seems tolerably unprob lematie, Lam right about Aristides, however, he (reats these terms as interchangeable; they indicate only different aspects of the same things, meanings, regardless of whether purposes, What is clear is that the Stoies intended a distinction between oryauvöpeva 14 and other L ‘ by a viorv Bvoa, while ble = ican a ae = a "di “a understood, en by zZ un auılon OF objet to our minds through the that few Greek the - =z 4 Wt his move In potentially a fruitful one, and Anstiden wenns te = ecn dent vane vip ean, HELL (eea in . ed be hie wentment as eanceptually Healy derived from the Staip permire; and tuut within a this Stolelzing material ix emb edde ed Amp ich the subjowt that differ In ver te the B p ann ways (vam all others subjeet | that have surviv and terminals healt sources of many ie na = += tenor, and tov have almarbod Cor partly absorbed) ideas from © compounds Henn ln wa, ALt ie nut always make vonsistent sense, and which certainly ennn 4 , thatge his l Villon te ae fully cae ng the views of ity particular wehool’, I do not claim | je Y CON Met tet alone Tally adequate) or vonsi stentiy Stoic in JON) (hae i m SURES an - a bo Artstatellanten, Amen other Things, we not hard to find, What I rarely u source of genuine intel inspired in the Organization are ve orginal, He of his borrowings, very ix an unanhamed eclectic, fitfully 5 6 hi Wied 10 explore, And this brings me to my final point, It veer = sin : the Ontic Begotter of this radical shiftin the acne able that Aviatiden was himself way ideas developed ots : pr 7 Meruty unalyain, deploying in an entirely novel aanb nein Std Ze different purponen and in different contexts of one is trek … es 7 mal fe 0 Dur donveptions of other things, vakButof this shae > = —_ i HO want to saddle myself wich (he improbable al is cru cin na me ve OF all people, has found the magic key to with affective loadin, ‘ = eren One In whieh aur nation of some (hing is hung about rocha = = d, n MAL CHUNG A Metapher 1e not primarily à kind of expression, Is awoeiat 2 ren hoe Which the literal meaning of the metaphorical expression bu Hier of @ M the meta an that however metaphorteal reference is to be er or modified verbal OApronniens, and no on, all in the met Yı ‘ law A, usE lexicı > * 2 ng rent but otherwise equivalent ways of referring to the same thing. ” ideaal 2 1 hochanc eof even acddronstng this Inne, There seer to be u good deal Neos 10 Heat DOLE &véperrer, metapharienl expreusions, forelpn ash zag which VOW ix used — and hence how Metaphor van be à useful weupan in the puy poetor the psychological mantgalaterPheAristoteltun approach, ate D won (0 in an Hg Nipnifionntly different Tron that In whieh it in seen when pe wr Solved, there will still be the quentin how wl Bee pe Bee seems likely lo remain not only contention but | ‘ ection her that refer, If at all, to something quite different = in of course day, out So long an weandhavermnigewendsvan hand,tl and on the other thething things Picked by them thereby Darupthe in ane thought saroble (| erken. 10 something which might have heen referred io but a racer Promising move, The question how metaphors wark — how th mnd + pr, Irom the domain of Mig to that of Évvoi ix not à in > lt ‚no Abo supgosted that the decision EEE OF Metaphors hin predilections, lor (he most part, are for à sort OF advanced neo-Pythugoreunism 4 bordering on 'ulimentary Neoplatonism, Ney be We trying to expound es ze talyne Stole f im of tin re doctrine, only to make his own modest une

live inter to our body of knowledge about the Stoies, Apart from its intrinsic interest, it is farther evidence, if such were needed, of the thoroughly systematic nature of their thought, ANDREW BARKER expect my revisions to have convinced that rey Fr, but I subset that there is a subset il case to answer of the Cambridge Philological Society In January of this paper were presented at meetings IN Kurier versions Lam gratefulto members of both audiences 1991, and the West Midlands Classios Seminar in March 1997, lor comments and suggestions made on those occasions. Lowe a speetal debt to this journal's anonymous reader for the trenehant and far-reaching comments which he or she made on a previous draft, I do not UNIVERSITY OF BIRMINGHAM then HET resc dill weı imn liter gra Hur us tl This in C The judy ‘trut un prey ori | fore si test ius vorn ure, pers| only it on literary techniques were closely integrated with their suggesting that their reflections semantics, and so also with their metaphysics and psychology, IF we cannot assign these ideas an author or a date, that only puts them in the same situation as very many other '® indisputably Stoic doctrines, to wi voul in so far as we have evidence of its nature and contents, What I believe I have shown, however, is that we can add with some confidence one more small but intriguing Item came from, seem never to have reached the mainstream of Greek literary discussion, direct clue to the identity of the subtle and original Stole theorist whose existence Lam postulating (hough if I had to put my money anywhere, I would be inclined to have of Babylon), Nor can Lexplain why these ideas, wherever they little Hutter on Diogenes lectual novelty, It is overwhelmingly likely that he found literary figures treated in this way and in these terms in some earlier authority, whose analyses he took over and adapted to his own themes; and there is every reason to believe that the authority was not merely conversant with Stoic terminology but thoroughly committed to central tenets of peculiarly Stoie doctrine, As far as I have been able to discover, we have no