Aristotle on Metaphor

Aristotle on Metaphor Author(s): John T. Kirby Source: The American Journal of Philology, Vol. 118, No. 4 (Winter, 1997), pp. 517-554 Published by: The Johns Hopkins University Press Stable URL: http://www.jstor.org/stable/1562051 Accessed: 17/09/2008 02:08 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=jhup. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected]. The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and extend access to The American Journal of Philology. http://www.jstor.org ARISTOTLE ON METAPHOR JOHN T. KIRBY MuchMadnessis divinestSenseTo a discerningEye-Emily Dickinson OURS IS AN AGE of metaphor. Wayne Booth, in his inimitable fashion, remarks, There were no conferenceson metaphor,ever, in any culture,until our own centurywas alreadymiddle-aged. As late as 1927,John Middleton Murry,complainingabout the superficialityof most discussionsof metaphor,could say, "Thereare not many of them."... Explicitdiscussionsof something called metaphor have multiplied astronomicallyin the past fifty years.... studentsof metaphorhave positivelypullulated.1 In the postmodern era philosophers of language, particularly those outside the analytic tradition, tend to think more and more in terms of all language as being metaphorical. This trend, however, is not a new one; it seems to have its roots back as far as Heraclitus, and, in the modern world, was certainly espoused by Giambattista Vico in his Nuova Scienza (1725). What we may call the Viconian tradition was embraced on the continent by Nietzsche, and, in the Anglo-American world, by Ivor Richards.2 Nietzsche's thought had far-reaching consequences in terms of its influence on the work of Paul de Man and Jacques Derrida, who in the late 1960s and the 1970s found themselves in the center of a vortex of controversy over the notion of "deconstruction," and its interrogation not only of Western metaphysics but also of human language and thought generally. Now, it seemed, such simple formulae as "This X is Y" called into question the whole process of naming and predication. But even among those for whom this is an unconvincing position, the problem of metaphor continues to be a fascinating one. How to define 'Booth 1978a, 49 (a footnote in the text has been omitted). 2See Vico [1725] 1968, Nietzsche [1873] 1989, Richards 1936. American Journal of Philology 118 (1997) 517-554 ? 1997 by The Johns Hopkins University Press 518 JOHN T. KIRBY it? What are its uses? What is its relation to literal, or nonfigured,language? And what its relationto humancognition?3 Booth goes on to say, "therehave been many more discussionsof what people from the Greek philosopherson called metaphorthan any bibliographycould show."4That being so, a sustained study of Aristotle's concept of metaphorneeds no furtherjustification,since it is to him that we owe the terms in which the debate was framed for many hundredsof years. Indeed, here as often, even those who wish to propose new or differentparametersfor the analysisof metaphormust do so againstthe grainof the Aristoteliantradition.5This, if nothingelse, is a measureof the tremendousinfluenceAristotle has had on the history of Westernrhetoricand poetics. As on many other topics, it is now fashionableto condescend to Aristotle for the limitationsof his study of metaphor,or-more aggressively still-to find fault with its parameters.Certainlyhe did not preempt any furtherdiscussionon the issue;nor, I imagine,would he have wanted to. But I surmise that there is more to be learned from an appreciativestudy of his methodshere than one mightinitiallysuppose.In this study, then, after a few remarkson recent studies of metaphor,I propose to examine the state of the questionbefore Aristotle, and then 3The bibliography on metaphor is staggering, but with pluck and perseverance one can enter what Kenneth Burke would call the "scholarly conversation" on the topic. For treatment of the Aristotelian theory in particular, see Stanford 1936, McCall 1969, Derrida [1971a] 1982, [1971b] 1982, Ricoeur [1975] 1977, Tamba-Mecz and Veyne 1979, Swiggers 1985, Lloyd 1987, and Laks 1994. For a perceptive and innovative analysis of some ancient uses of metaphor see duBois 1988. Important reading for an understanding of twentieth-century scholarly thinking about metaphor in general are Nietzsche [1873] 1989, Barfield 1928, Richards 1936, Foss 1949, Black 1955, Jakobson 1956 (with Lodge 1977), Lacan [1957] 1977, Brooke-Rose 1958, Knights and Cottle 1960, Turbayne 1962, Wheelwright 1962, Greimas [1966] 1983, Hester 1967, Goodman 1968, Dubois et al. [1970] 1981, Genette [1970] 1982, Greimas [1970-83] 1987, Shibles 1971b and 1972, Mooij 1976, Levin 1977, Booth 1978a and 1978b, de Man 1978, Harries 1978a and 1978b, Ricoeur 1978, Ortony 1979 and 1993, Ruegg 1979, Sacks 1979, Lakoff and Johnson 1980, Johnson 1981, Miall 1982, Eco 1983, Haverkamp 1983, Gumpel 1984, Martinich 1984, Mac Cormac 1985, Paprott6 1985, Cooper 1986; Eco, Niklas, and Edeline [1986] 1994; Skulsky 1986, Kittay 1987, Lakoff 1987, Turner 1987, Haley 1988, Lakoff and Turner 1989, Quinn 1991, Lakoff 1993, Gibbs 1994-to list only a very few. One might begin to look for further bibliography in Shibles 1971a, Haverkamp 1983, and van Noppen 1985, and by keeping an eye on the journal Metaphor and Symbolic Activity. 4Booth 1978a, 50 (emphasis in text). 5Most explicitly, but as only one among others, see Gumpel 1984. ARISTOTLE ON METAPHOR 519 to look at what he has to say about metaphor in the Poetics and Rhetoric. In their attempts to cope with the notion of language generally, and specifically with the questions of literal versus figurative language, scholars over the last hundred years or so have themselves appropriated a number of metaphors purporting to describe the phenomenon. These include (above all) models based on comparison and interaction. The comparison model, it is typically said, conforms more to the classical approach to metaphor, whereas those based on interaction stem primarily from the work of Richards and Black.6 Indeed it is from this notion of interaction that the terms "tenor" and "vehicle" were born, the former referring to the underlying idea that is illustrated or illuminated by the latter, which is "applied" to it; and Richards is right both to point out [1] that the word "metaphor" is sometimes used to signify what he means by "vehicle," and sometimes to mean the symbiosis resulting from the conjunction of tenor and vehicle, and [2] that "metaphor," in the sense of this symbiosis, is impossible without just such an interaction of tenor and vehicle.7 Other approaches have been taken, such as that of speech-act theory, which posits an "incongruence between the meaning of a sentence and of a speaker's utterance";8 the imagination model, which highlights the complementary roles of the cognitive, imaginative, and emotional components of metaphor;9 and the emotive model, which (unlike the imagination model) "denies metaphor any cognitive import," asserting instead that while the cognitive function of language controls the statement of facts, the emotive function expresses feelings and attitudes.10 An ambitious linguistic/rhetorical approach was undertaken by the Groupe It of Liege in their Rhetorique generale. This work attempts, in the French style, to isolate a (nonfigurative) "degree zero" of discourse, from which all manifestations of "figurative" language-including such 6Richards1936,Black 1955.See also Perelman[1958]1969(esp. 399 n. 177).Black speaksof the comparisonmodel as a subsetof the substitutionmodel,in whichwhatRichardscalls the vehicle is substitutedfor the tenor. 7So too for Aristotle, metaphoracould mean both "a [particular]metaphor"and "metaphor"in general;see Stanford1936,6-14. 8Eco, Niklas, and Edeline [1986]1994,544. For more on speech-act theory and metaphorsee Searle 1979and 1980,and Mac Cormac1985,ch. 6. 9Ricoeur[1975]1977. 10Eco,Niklas,and Edeline [1986]1994,547. 520 JOHN T. KIRBY seemingly innocent phenomena as repetitions or unexpected suppressions-are viewed as deviations, among which metaphor would of course be counted.'1 One of the most important and influential recent treatments of metaphor is the vast survey by Paul Ricoeur, which (among other things) explores the notion of "metaphorical truth" and the possibility that metaphoric language need not be tied strictly to referential meaning.12 Some very stimulating and useful models-perhaps the most promising of those mentioned so far here-are based on various aspects of cognitive theory, such as that of Earl Mac Cormac and (especially) of George Lakoff and his collaborators. Mac Cormac's model incorporates the notion of words as defined by fuzzy sets,13and envisions metaphor "(1) as a cognitive process by which new concepts are expressed and suggested, and (2) as a cultural process by which language itself changes."14 The model proposed by the Lakoff school conceives of metaphor as a process of "mapping" from a source domain to some target domain.15 While Lakoff is quite emphatic about distinguishing his cognitive model from what he calls the "classical theory," I think it is possible not only to show some important adumbrations of his model in Aristotle, but also to reconcile the two to a significant extent-of which, more anon. With all these (and myriad other) studies of metaphor, it is a great irony that Umberto Eco can say, "of the thousands and thousands of pages written about the metaphor, few add anything of substance to the first two or three fundamental concepts stated by Aristotle."16If that is so-and I am inclined to say that it is-perhaps the journey to the fons "1Dubois et al. [1970] 1981. Indeed the et stands for "metaphor," as we are told in a "liminary note" (p. xix of the English version): "Roman Jakobson was one of the first to draw attention to the operative value of concepts already discussed in Aristotle. In homage to these two witnesses, it is quite natural that we have chosen as our symbol the first letter of the Greek word designating the most prestigious of metaboles." 12Ricoeur [1975] 1977. 13"Fuzzylogic" is the topic of much current research and debate; for a cursory introduction see Zadeh 1965 and 1975, Kosko 1991 and 1993, McNeill and Freiberger 1993. 14MacCormac 1985 (this citation is from pp. 5-6). 15Lakoff and Johnson 1980, Lakoff 1987, Turner 1987, Lakoff and Turner 1989. Lakoff 1993 offers what is essentially a state-of-the-art overview of the cognitive approach to metaphor. 16Eco [1983] 1984, 88. ARISTOTLE ON METAPHOR 521 et origo is worth the effort after all. Let us begin, then, by reachingback furtherstill:before the beginning,as it were. GREEK METAPHORBEFORE ARISTOTLE There is a huge body of Greek literature before Aristotle, of course, and space does not permit an exhaustiveanalysisof it here. For their particularsignificance,however-and what that is, in each case, will become clear-I have singled out three authorsfrom that earlier period for individualconsideration:Homer, Isocrates,and Plato. Homer Greek metaphoris as old as Greek literature-which is to say, as old as Homer. The astute reader will instantlysay, '"Ah,but Homer is most famous for his similes."Just so: but as Aristotle classifiedthe simile as a species of the genus metaphor,17 he himself would have considered the Homeric similes to be metaphors.18In fact there is no doubt that, by this definition,the similes in the Iliad and Odysseyconstitute the best-known group of metaphors in Greek literature up to Aristotle's time. These range from a few words to several lines in length;of the latter G. S. Kirk says, "Theexpandedsimile, in which the details of the image are developed far beyond the point of comparison,and for their own sake, is one of the chief glories of the Iliad. The simile is a deliberate and highly wroughtstylisticdevice, as carefulin its languagewhich is often untraditionalin appearance,because the subject matter is often untraditionaltoo-as in its variety and its placing in the narrative."19 Sometimes in Homer we find what we would call metaphorside '7Rhetoric 3.4.1, 1406b20. (My "Bekker" citations of the Poetics and Rhetoric in this essay are keyed to the lineation of the superb editions by Rudolf Kassel-Kassel 1965 and 1976.) 18Onthe Homeric simile see esp. Frankel 1921, Shipp 1953, Lee 1964, Hogan 1966, Scott 1974, Moulton 1977, Nimis 1987, Janko 1992 passim. The summary in Edwards 1991, 24-41, is dense with information. On the Homeric metaphor (as distinguished from the simile) see Keith 1914, Parry 1933, and Stanford 1936, ch. 7. 19Kirk1962, 346. 522 JOHN T KIRBY by side with simile. In this example the metaphor is characteristically brief, while the simile (describing the same phenomenon) is elaborated at leisure: These [the Aiantes] were armed,and about them went a cloud of foot-soldiers. As from his watchingplace a goatherdwatchesa cloud move on its way over the sea before the drive of the west wind; far away thoughhe be he watchesit, blackerthan pitch is, movingacrossthe sea and piling the stormbefore it, and as he sees it he shiversand driveshis flocksto a cavern; so about the two Aiantes moved the battalions, close-compactedof strongand god-supportedyoung fighters, black,andjagged with spear and shield,to the terrorof battle.20 (Il. 4.274-82) Here the metaphor, "a cloud of foot-soldiers," occupies one line, while the elaborating simile takes up the following eight. This disparity of length between metaphor and simile is what Aristotle has in mind at Rhetoric 3.10.3, 1410b18.In other places, such as Iliad 2.455-83, Homer proliferates his similes with prodigal richness, in order to signal the significance of the event that is about to occur.21Moreover, in direct discourse the simile may be an important key to characterization: it is Achilles who utters the greatest number of similes in the Iliad.22 Ancient as well as modern scholars have pronounced on the function of the simile in Homer: The scholia consider the similes contributeauxesis (fullness), enargeia (vividness), sapheneia (clarity),poikilia (variety) and kosmos (decoration).... M. Coffey,AJP 78 (1957) 118,has categorizedtheir functionsas illustratingthe movement of an individual,a group, or a thing;the appearance of a hero, group, or thing;noise; measurementof time, space, and numbers;a situation; and psychological characteristics,including 20Ihave cited the translationof Lattimore1951.Richardson(1993, 100) cites II. 14.16-20as anotherexampleof the neat juxtapositionof a simile and a metaphor. 21So too Richardson 1993, 123. 22Moulton1977,100,cited also in Janko1992,316;Richardson1993,133.Jankohere points specificallyto Aristotle:"Homergives him [Achilles]the acute perceptionsof a bard-imagery is the proof of poetic genius (Aristotle,Poet. [22,] 1459a7)."For Achilles as a makerof similessee also Martin1989,193,204-5. ARISTOTLE ON METAPHOR 523 decision-making.W.A. Camps,An Introductionto Homer (Oxford1980) 56, sums up the uses:to suggestinwardfeelings and states of mind;to illustrate the distinctivequalities of things, actions, or processes;and to render effects of multitudeand mass. More specifically,M. Mueller,The Iliad (London1984)108-24, notes that a similemarksa passageas worthy of specialattention,slowingdown the narrativeas expansionsand digressions do.... it adds colour and a new dimensionto whateveris the focus of attention.Besides this, because of its characteristicallyeverydaycontent the Homeric simile for a moment unites narratorand audience in theirworld,not that of the heroes, as together they marvelat the mighty deeds of the past.23 Isocrates So much for Homer. Subsequent writers, both of poetry and of prose, make equally abundant use of metaphor and simile: one could produce innumerable examples from Pindar, the Presocratics, Attic tragedy and comedy, and the historians and orators.24 But the simple use of metaphor, of course, is different from the naming-or the scholarly consideration-of it. Aristotle did not invent the word metaphora, though (as a noun) there is no extant use of it earlier than Isocrates, who (writing ca. 374 B.C.E.) himself frames the discourse in terms of the prose/verse distinction. For him, there is an everyday Greek that is used in ordinary discursive situations, and that is typically barren of ornament; and, in distinction to this, a more ornamental diction that is available to poets. (A semiotician would say that ordinary Athenian discourse was semiotically unmarked, while verse was marked, not only by the use of meter, but by the use of this ornamental diction). Prominent among such kosmoi, in Isocrates' view, was something he called metaphora: To the poets are granted numerousornaments(kosmoi) [of language], for... they can express themselves (delosai) not only in ordinarylanguage (tois tetagmenoisonomasin),but also by the use of foreign words (xenois), neologisms (kainois), and metaphors (metaphorais)... but to writersof prose (tois de peri tous logous) none of such [resources]are permitted; they must strictly (apotomos) use both words (onomaton) and 23Edwards 1991, 38. 24Fora book-length treatment of metaphor in Pindar see Steiner 1986. 524 JOHNT.KIRBY ideas (enthumematon)[of a certaincategory:][1] of words,only those that are in the [ordinary]languageof the polis (ton onomatontois politikois); [2] of ideas, only those that are closely relevantto the matterat hand (ton enthumematontois peri autastaspraxeis).25 (Euag. 8-10) This passage raises a number of issues that are worth considering. First of all, its list of word categories shows some similarity to (and some differences from) the lists in Aristotle's Rhetoric and Poetics-as well as that in the Rhetorica ad Alexandrum (23.1, 1434b33-34), which has simple (haplous), compound (sunthetos), metaphorical (metapheron). In the Poetics (21, 1457bl-3)-evidently the earlier of the two Aristotelian lists-we find (in addition to the simple/double/triple/compound distinction, i.e., haploun/diploun/triploun/pollaploun) a list of eight: [1] [2] [3] [4] [5] [6] [7] [8] the ordinary word (kurion) the "strange" or "foreign" word (glotta) metaphor (metaphora) the ornamental word (kosmos) the made-up word or neologism (pepoiemenon) the lengthened word (epektetamenon) the shortened form: (hupheiremenon) the altered form (exellagmenon). In the Rhetoric (3.2.5-6, 1404b28-32), which (so to speak) footnotes this passage from the Poetics, Aristotle refers especially to glottai, dipla, pepoiemena, kuria, and metaphors. Significantly, he also singles out these last two as the most suitable to prose:26 "The ordinary and native [term] (kurion kai oikeion) and metaphor are alone useful in the style (lexis) of prose; for everyone converses (pantes dialegontai) using metaphors and native and ordinary 25Thetranslationof this importantpassage(as elsewherein this essay,unless otherwise specified) is my own. I would apologize for its complexity,were Isocrates'own prose not equallyconvoluted.(N.B.:tois de peri tous logous may mean, not prose writers in general,but specificallyoratorsand/orlogographers.) 26Histerm is "barewords,"psiloi logoi. Psilos is the word one would use, e.g., to describea tree branchstrippedof its bark,or deforestedland;appliedto language,it implies that stylisticornamentis somethingaddedto a plain substrateof discourse-an idea centralto the projectof Dubois et al. [1970]1981.In the Poetics(6, 1449b28-31)Aristotle speaksof hedusmenoslogos, language"sweetened"by the additionof metricalregularity and (sometimes)melody-again, two notableformsof semioticmarking. ARISTOTLEON METAPHOR 525 terms" (Rhet. 3.2.6, 1404b34-36).27 In this regard he is so diametrically opposed to the Isocratean doctrine that it is difficult not to believe that he is specifically and consciously addressing it.28 Some of his other kuria, glottai, and pepoiemena-also correspond to items-namely items in the Isocratean list (politika, xena, and kaina respectively).29 Why did Isocrates want to ban metaphor from nonpoetic discourse? The question becomes the more intriguing when we observe that his own language here contains words that are, in our terms at least, demonstrably metaphoric: not only the term metaphora itself,30 but also the adverb apotomos, which I have translated "strictly" but which most exactly means "in a cut-off fashion" (from apotemnein, "to cut off"). It is possible that these were "dead metaphors" by the time Isocrates was writing, but even if that is the case, one need not look far in his own prose to find clear examples of what we would call metaphor: On accountof those who lightlyadvocatewar,we have alreadyfallen into (periepesomen[the word may also mean "be dashed against"])many (On the Peace 12) great misfortunes. Many treatmentsof all sorts have been found by the physiciansfor the diseases of the body, but for minds that are ignorantand full of wicked desires there is no other medication(pharmakon)than discourse(logos) that daresto strikeout at (epiplettein)their transgressions... (On the Peace39) 27Translations from the Rhetoricare taken, or adapted,from Kennedy1991;translationsfrom the Poetics,fromJanko1987. 28SimilarlyO'Sullivan (1992, 51), who however also points out that Aristotle praises Isocrates'prose style with the word hedeia, "sweet"(Rhet.3.9.8, 1410a19-20).Is this partlymotivatedby Isocrates'use of metaphor,despite his own stricturesagainstit? After all, Aristotlesays sweetness(to hedu) is one of the greatvirtuesof metaphor(Rhet. 3.2.8, 1405a8-9). 291take oikeion to be synonymouswith,if not exactlyidenticalto, kurion.If there is a distinctionbetweenthem,it may be that the oikeionis the one "native"to a regionor dialect (from oikos, "home"),while the kurionis the one in commonestuse (from kuros, Typicallythese oughtto be congruentsets, althoughsometimesthe kurionis "authority"). not (or was not originally)the oikeion:for example,"cheese-and-tomatopie"is a designationmade up entirelyof oikeia,but the kuriononoma (even in English)is now surely "pizza";I imaginethat, in ordinaryconversation,English-speakersdo not think of the latteras a glotta. 30Kennedy(1991,222 n. 25) pointsout that metaphorais itself a metaphor(the idea of carrying). 526 JOHN T KIRBY [Ourforebearsthoughtthat]thosewhomakecopiouslawsdo so asbarricrimes. ers [orbulwarks, emphragmata] against[rampant] (Areopagiticus40) ThusI am led to believe ratherthat metaphora(like enthumema)meant something different to Aristotle than to Isocrates.Even setting aside the famous dictum ascribed to Aristotle by Diogenes Laertius, "It is shamefulto be silent and let Isocratesspeak,"31there is the possibility that Aristotle was for a time the pupil of Isocrates.32In any case Isocrates'school was at that time the principalrivalof Plato'sAcademy in Athens,33so it would have been easy for Aristotle to hear of his teachings on language, and to read whichever of his works had been published.Is it possible that by metaphorIsocratesmeant elaborateformal analogiesin the style of the Homericsimile, and only that?We shall perhapsnever know the sure answer,but if that were so, it would explain why he restrictsmetaphorato verse while at the same time using, in his own prose, what modern thinkerswould call metaphor.34These differ only in degree, and not in kind, from the (sometimes arresting) metaphorsone finds in the lyric and dramaticpoets of the great age of Athens, such as PindarPythian 10.47-48, where Perseus, carryingthe head of Medusa, is said to be "bringingstony death"to the people of Seriphos;Aeschylus Agamemnon 36-37, '"Agreat ox stands upon my tongue"(i.e., I will say nothing);Sophocles Oedipusthe King 153,"I am stretchedout [on the torture-rack]"(i.e., in mental/emotionaldistress). It may also be worthwhileto call attention to Isocrates'mention of enthumematahere, because of the massive semantic shift this word underwentat the handsof Aristotle.I have translatedenthumematanot as "enthymemes"but as "ideas,"that is, something the rhetor has in mind (en thumoi).Isocratesuses the word also in an importantpassage in the Panathenaicusthat may shed light on its meaning: 31'Aiskhronsiopan,Xenokratend'ean legein."Lives of the Philosophers5.3;cited also by Philodemus On Rhetoric (II 50 Sudhaus). Virtually all editors emend Xenokraten to Isokraten; the latter was the version known to Quintilian (3.1.14). 32A theory advanced by Chroust 1973, 97. Isocrates, in turn, thought of himself as a follower of Socrates; see testimonia in Kennedy 1963, 179-80 and n. 85. 33See, e.g., Dion. Hal. Isoc. 1. 34For more on what Aristotle would call the simile see below, and Tamba-Mecz and Vey 1979. ARISTOTLE ON METAPHOR 527 Leaving aside all these [other types of discourse],I focused instead on those [kinds]that gave advice about what would benefitboth the Athenians and the other Greeks,and those that were full of manyenthumemata, and with not a few instancesof antithesisand parisosis,and of the other devices that shine in orations and induce the audience to marvel and (? 2) approve. George Kennedy, referring to this locus, opines that "in Isocrates it [enthumemata] appears to refer to elaborately developed sentences (Panathenaicus 2)."35 In this passage, however, enthumemata occurs in conjunction with antithesis and parisosis, both specific and even technical terms denoting devices of style; so for Isocrates the enthumema may have been a certain identifiable pattern of structure or syntax. Isocrates is said to have studied with the sophist Gorgias in his youth,36 and both antithesis and parisosis are closely identified with the Gorgianic style; might then enthumema as Isocrates uses it here also refer to some aspect of Gorgias' hallmark style, or of his rhetorical teaching? For Aristotle, by contrast, the enthumema is not a stylistic or (merely) syntactic phenomenon, but rather an inventional vehicle, the very mechanism by which ratiocination is articulated rhetorically. Indeed it was one of the greatest achievements of Aristotle to discover the syllogism, and to identify its valid forms; he was proud of this discovery, as he shows at the end of the Sophistical Refutations, and understandably so. He defined the enthumema as the use of syllogistic demonstration in a rhetorical context,37 and identified the material used for its premises as semeia, signs, and eikota, probabilities.38 This shift from the 35Kennedy 1963, 99. 36Cic.Or. 176; see further testimonia in Too 1995, 235-39. 37Rhet. 1.1.11, 1355a3-14; 1.2.8-9, 1356a34-bl7; Top. 8.14, 164a6. We must constantly remind ourselves of how little we really know about ancient semantics. For Sopho- cles (born perhapssixty years before Isocrates,and a decade or more before Gorgias)it seems that enthumemaalreadymeant "argument"or "reason"(OC 292, 1199,cited in Kennedy 1963,99); thus it is possible that even before Aristotle [1] identifiedthe valid forms of the syllogismand [2] identifiedthe enthymemeas a species of syllogism,enthumemamay have been a technicalterm used in certainquartersto refer to a recognizable syntactic/stylistic structurecomposedof a main-clauseassertionbackedup by an explanatorygarclause.Kamerbeek(1984,61-62, 169)says thatthese Sophocleanloci are the first occurrencesof the word in Greek and comparesenthumion,"scruple"(Hdt. 8.54), with Soph.OT 739 and Eur.Herac.722. 38Rhet.1.2.14-18,1357a32-b25;cf. Pr.Anal. 2.27,70a3-b38. 528 JOHN T. KIRBY Isocratean conception of enthumema may well be all of a piece with Aristotle's similarly momentous transmogrification of Isocrates' notion of metaphora, from the realm of the sheerly ornamental and superficial to that of the substantive-that of the semiotic, as we shall see. Plato Whether or not Aristotle ever studied with Isocrates, no such uncertainty attends his long apprenticeship to Plato.39Much of Aristotle's work on rhetoric and poetics responds to, and subverts, the work of Plato on those subjects; one would like to know what Plato would make of the term "metaphor." He does use the verbal form metapherein, in the sense of "transfer"; at Critias 113a he speaks of transferring words (onomata), in the sense of "translation" from one language to another, and at Timaeus 26c, of transferring ideas from fiction to reality (metenegkontes epi talethes). As W. B. Stanford points out,40the closest Plato comes to an actual discussion of verbal metaphor is in Theaetetus,where he says (of the Heracliteans): "If you should ask one of them something, they draw out (anaspontes) some riddling little words (rhematiska ainigmatode) as if from a quiver, and shoot them off, and if you should seek to make sense of what has been said, you will be stricken anew by some other verbal distortion (heteroi... metonomasmenoi)."41 He is quite right about Heraclitus and his ilk, of course, but what fairly jumps off the page at us is the metaphor depicting language as bows and arrows. I take it that this passage is a reference to Aristophanes' Clouds 942-44: "And then, out of the things he says, I shall shoot at (katatoxeuso) him with new [little] words (rhematioisin) and thoughts"-but there it is the Hetton Logos (the personified Weaker Argument, or False Logic) that 39Guthrie (1981, 22) points out that this spanned a period of twenty years. Speaking of apprenticeships, it should be remembered that Plato, the faithful student of Socrates, has Socrates give lavish praise to the work of Isocrates at the end of Phaedrus, in what appear to me to be exceedingly sardonic terms. The implication here (it seems) is that whereas Plato purported to carry out the quest for the philosophical rhetoric desiderated by Socrates in this dialogue, Isocrates did not fulfill the great potential he exhibited in his youth. (See similarly Kennedy 1963, 79.) 40Stanford 1936, 4. 41Theaet. 180a. Some philological notes: the verb "shoot" (apotoxeusousi) specifically means "shoot from a bow" (toxon). The suffix of rhematiska is diminutive and has here a pejorative sense. ARISTOTLE ON METAPHOR 529 is speaking, the representative of the new sophistic education. But it could be that Plato is slyly taking aim here, with an Aristophanic arrow, at Heraclitus himself, who famously said, "They do not grasp how being at variance it agrees with itself [lit. how being brought apart it is brought together with itself]: there is a back-stretched connexion, as in the bow and the lyre."42 In our quest for the use of metaphor in Plato's prose, however, we have even less distance to go than in the case of Isocrates. The truth of Aristotle's remark about the ubiquitous use of metaphor in conversation is abundantly borne out in the Platonic dialogues: SOCRATES. Whoops! Have we gotten here too late, and at the end of a feast, as the saying goes? CALLICLES. Yes indeed, and a very elegant feast too, for just a little while ago Gorgias gave us an abundant and beautiful display. (Gorg. 447a) ... you are exactly like the flat stingrayof the sea. Wheneveranyone approachesand touches it, it numbshim, and you seem to have done some(Meno 80a) thing like that to me just now. ... I was afraidthatwhen Agathonfinishedup, he wouldsend the head of Gorgias, fearsome in discourse (Gorgiou kephalen deinou legein en toi logoi), againstmy own speech,and make me stone in muteness(me lithon tei aphoniai poieseien). (Symp. 198c)43 I say that he [Socrates] is exactly like (homoiotaton) those statues of Silenus that sit in the statue-carvers' shops, which the craftsmen carve play- ing the syrinxor the aulos, but which being opened up in the middle,appear to have imagesof the gods inside.And moreoverI say that he is like (eoikenai) the satyr Marsyas.44 (Symp. 215a-b) 42Heraclitus fr.22 B51 DK; the translationis from Kirk[1957]1983,192.Curiously, we have yet anotherfragmentof Heraclituson the bow (22 B48 DK): "Thename of the bow (scil.bi6s) is life (bios), but its functionis death."In this paronomasticaphorism,the connectionbetween the bow and languageis made, by virtue of the pun, as obvious as possible. 43Thementionof Gorgiashere is not only a referenceto the Gorgianiceffusionof Agathon'sspeech but also a pun on Gorgon:it was said that if one looked at the head of Medusa,one wouldturnto stone (againsee, e.g., Pind.Pyth. 10.46-48). 44Silenuswas a satyr-like creature,half-equine and physicallyunattractive;like satyrs,he was associatedwith debaucheryand lust. Marsyaswas a satyr whose musical ability had legendarypowers;he even presumedto compete against Apollo (see, e.g., Apollod.Bibl. 1.4.2,Hyg.Fab.165,Plin.HN 16.89,Ov.Met.6.382-400).The point of these comparisonsis to portraySocratesas outwardlyugly,even grotesque,but inwardlybeautiful. 530 JOHN T. KIRBY One could go on at tremendous length; suffice it to say that metaphoric language completely pervades Plato's writings.45 On occasion he moves into the use of what we might call extended metaphor, for example when he compares the psyche to a charioteer and two horses, one white, one black (Phaedr. 246a-57a). This is not vastly different from his use of parables or myths, such as that of Er in the Republic, or of Theuth in Phaedrus itself. Although it is true that Plato does not use the term metaphora, he does use the term eikon, "likeness," not only of physical/visual resemblances but also of verbal comparisons that we would call similes.46 Marsh McCall estimates that, of the more than sixty occurrences of the word eikon in Plato, about a third are in a rhetorical context.47 Besides the one already cited from Symposium 215a-b, which uses the verbal form eoikenai, one might adduce the following: Well then, I will tell you anotherimage (eikona) from the same school as this one just now... (Gorg.493d) I see why you made a simile (eikasas) about me. SOCRATES. MENO.Why do you supposethat was? SOCRATES. To get me to make one about you in return (hina se anteikaso). And I know this about all beautifulpeople:they enjoy being compared to something.It works out well for them, I suppose, since the images (eikones) of the beautifulare also beautiful.But I will not make a simile about you in return(ouk anteikasomaise). (Meno 80c)48 Consider this, whether there is something to what I am saying:for it seems that,just like Simmias,I have need of a simile (eikonos). (Phaed.87b) In addition to these loci, we find that Plato uses eikon in the more general sense of imagery or illustration. 45Indeed Plato was actually criticized for his use of metaphor, by such ancient literary critics as Dionysius of Halicarnassus (Dem. 5) and Caecilius of Calacte (see Ps.-Lon- ginus On the Sublime32.8). 46Thoroughly documented by McCall (1969, 11-18), though he strictly opposes the translation of eikon as "metaphor." 47McCall 1969, 12. 48Socrates seems to be referring to an Athenian party game here: "In conversation at a party or elsewhere, to provide amusement, one man would compare another to something funny, and his victim would then compare him to something in return" (MacDowell 1971, 304). ARISTOTLEON METAPHOR 531 What is importantabout all these instances is that in each case eikon signifies the representationof something-material or conceptual-by something else; and that fact will have importantreverberations later on, as we shall see. Plato, then, makes plentiful use of metaphor,both brief and extended, in his own writing;he has a name for such verbal analogies;and on at least one occasion, he has a characterrefer to the phenomenon itself. He does not, however, provide a theory of metaphor as such, not even cloaked in the narrativeguise of elenchus.This is perhapsthe more curious in view of the representationalnature of metaphor.It would not overstatethe case to say that Plato's whole ontology,at least as laid out by Socratesin the Republic,is rooted in semiotics:the relationship to the Formsof objects in the phenomenalworld is spoken of, not only in termsof "participation" (methexis,e.g., Phaed. 100c3-6), but This fact also in terms of mimesis (cf., e.g., Rep. 510b4,mimetheisin).49 problematizesfor us not only Plato's deep absorptionin the topic of mimesisbut also the fact that he does not reallyaddressthe relationship between mimesisand metaphor.50 As usual,he has less to say about the of compositionalaspects language and rhetoricthan about their metaphysical and ethical ramifications;thus a fully systematicinvestigation of the phenomenonhad to await the attention of Plato's most brilliant pupil. METAPHORIN ARISTOTLE In what follows I discussthis investigation,as it is conductedfirst in the Poeticsand then in the Rhetoric,and above all drawingattention to the guidingprincipleof the Aristotelianmodel. The Poetics: A Semiotic Model for Metaphor What is a metaphor,in Aristotle'sview? He offers a more or less formaldefinitionat Poetics21, 1457b6-7:"Metaphorais the epiphoraof the name (onoma) of something [to something else]." Because of the 49For more on the Forms see (among many others) Ross 1951, Vlastos 1954 and 1965, and Fine 1993. 5?Formore on mimesis see, e.g., Kirby 1991b; to the references there (esp. in n. 2) add Asmis 1992 and Golden 1992, 41-62. 532 JOHN T. KIRBY limitations of the Greek language at this point, Aristotle risks tautology, as both metaphora and epiphora stem from the same root pher-/phor-, "carry."A similar strain occurs in his famous definition of the syllogism at Prior Analytics 1.1, 24b18-20, "a sullogismos [/V leg-/log-] is a logos [V/ leg-/log-] in which, certain things having been asserted, something other than what has been asserted necessarily follows from their being so" (cf. Top. 1, 100a25-27). In both instances Aristotle's best recourse is to depend on the addition, or alteration, of a prefix. Meta- as a prefix often indicates a change of some sort; it may mean "across," so that metaphora is literally a "carrying across" or transference from one point to another. Indeed the Latin transferre (supine translatum), from which our English transfer and translation are derived, seems an exact rendering of metapherein. In the rhetorical context, the transfer seems to be that of a name from one item-the ordinary item for which that name is the "literal" term-to a new (and unaccustomed) one, where its applicability may be highly figurative. Conversely, though this is not the Aristotelian schema, we might envision a transfer of meaning or significance from one term to another. With epiphora we are on slightly different ground.51Epi- as a prefix may designate movement over or beyond boundaries. Too, it may have a sense of addition, or (as per LSJ s.v. epi G.I.4) "accumulation of one thing over or besides another." Thus epipherein may mean to put, or pile, something on top of something else (so Ar. Peace 167; Xen. Anab. 3.5.10; Hdt. passim). The noun epiphora may mean an additional payment (so Thuc. 6.31, IG I2 205) or (in later Greek) the second clause in a sentence (so Dion. Hal. Dem. 20) or even the conclusion of a syllogism (so, e.g., Chrysippus in SVF II 80). In Aristotle's usage it probably has the older sense of "piling up": the new or additional designation of a(n) (unusual or unaccustomed) name to something that already has a(n) (ordinary) name. Hence with cumbrous exactitude we might translate it "additional assignment"; of the various renditions offered in published translations of the Poetics-"giving" (Bywater), "movement" (Kennedy), "transfer(ence)" (Golden, Telford), "application" (Butcher, 51It is worth noting that Wheelwright (1962) uses the term "epiphor" to designate "metaphor in the conventional sense" (73)-i.e., metaphor that expresses something relatively well-known-and the term "diaphor" (from diaphora, "difference") to designate a metaphor that suggests a possible new meaning by emphasizing differences rather than similarities. Cf. Mac Cormac 1985, 38-42. ARISTOTLE ON METAPHOR 533 Else, Grube, Hutton, Janko), probably the latter is the most successful. "Transference" would be a fine translation of metaphora itself. But I suggest that if we limit ourselves to looking at this brief definition in our quest for a category to accommodate Aristotle's concept of metaphor, we risk eclipsing other useful information. With respect to the word onoma, I should point out that while Aristotle seems primarily to be thinking of nouns, this word does not completely restrict metaphor to nouns; onomata may include adjectives, which in practice are of course often metaphorical in value, and even verbs.52 After this proposition of a genus, Aristotle (as is his wont) inventories its various species. In so doing, he himself uses the language of scientific taxonomy: [1] a noun designating a genus may be transferred to a species, or [2] that of a species transferred to a genus; [3] one species-name may be transferred to another; or [4] we may find metaphor by proportional analogy, where, if (to use modern notation) A: B:: C: D, B may be used to represent D or vice versa.53Examples of each of these species of metaphor are given: [1] E.g., "Here stands my ship": for [the species] lying at anchor is [a part of the genus] standing. [2] E.g., "Truly has Odysseus done ten thousand deeds of worth": for [the species] "ten thousand" is part of the genus "many," and [Homer] uses it here instead of "a lot." [3] E.g., [Killing a man by] "draining out his life with bronze" [i.e., a weapon], and [drawing water by] "cutting it with long-edged bronze" [i.e., a bowl]: for here [the poet] calls cutting "draining"and draining "cutting." Both are [species of the genus] "taking away." [4] E.g., The wine bowl stands to Dionysus as the shield does to Ares: so [the poet] will call a wine bowl "shield of Dionysus" and a shield "wine-bowl of Ares." Again, as old age stands to life, so the evening 52On the onoma cf. Poet. 20, 1457a10-14; De Interp. 2, 16al9-b5. 53On the cognitive mechanism at work in proportionality cf. Post. An. 1.5, 74al7-25 (with the note ad loc. in Ross 1949, 525). The concept of analogy itself is redolent of the semiosis of metaphor, although Aristotle does not pursue it as such; see his remarks on analogy in Post. An. 1.10, 76a37-76b2; 2.14, 98a20-23; 2.17, 99a8-16. 534 JOHN T KIRBY standsto the day:so [the poet] will call evening "old age of the day," as Empedocles does, and old age "the evening of life" or "the sunset of life."54 In chapter 22 of the Poetics (1458a21-23)we are told that metaphor, like foreign terms (glottai) and unusualword forms, is a kind of "alien" term-used to achieve "impressiveness"(semnotes) and the avoidanceof familiarlanguage.This ensures that the languagewill not be commonplaceor low. If an entire compositionis made of metaphors, it will be a riddle-the latter being defined as talking of real things while making impossible combinationsof them (e.g., "I saw one man glue bronze on another with fire"to mean "I saw a physicianapply a heated bronze cupping glass to another [in order to draw blood]"). A lesser degree or concentrationof metaphorwill be more cognitivelyaccessible to the audience,but will still importthe desired impressiveness to the style. At the end of Poetics22, where Aristotle concludeshis discussion of metaphor, we learn two extremely important things on the topic. I take them up in reverse order:[1] Metaphor,like the "regular"word (to kurion) and the ornamentalterm (kosmos), is especially suited to iambicverse, because the iambus (the meter used, e.g., for spoken dialogue in Attic tragedyand comedy) is metricallythe closest to the prosodic patterns of ordinary speech, and expressions like metaphor are those best suited to ordinary speech.55 [2] Metaphor is the most impor- tant device to be skilledin, and one cannotlearnthat skill from another: one must have the innate gift.56This gift, moreover,depends on seeing likenesses in things (to to homoion theoresai, 1459a7-8). And here, I sub- mit, is where we begin to find a new avenue of approachto the question of how to categorizethe Aristotelianapproachto metaphor.I now draw the reader's attention to the opening portion of De Interpretatione, which is one of the first systematicformulationsof semiotic theory in Westernliterature: 54Janko(1987,129) maintainsthat only type [4] is equivalentto the modernnotion of "metaphor." Manywould,I think,disagreewithhim;but certainlyit receivesthe fullest treatmentfrom Aristotle and is said to be the best-liked (Rhet.3.10.7,1411al).The four types are well treatedin Bywater1909,282-84, and Eco, Niklas,and Edeline [1986]1994. 55Thatmetaphormay be used both to impartsemnotes(Poet. 22, 1458a21)and as an aspect of ordinarydiscourse(22, 1459a10-14;cf. Rhet.3.2.6, 1404b31-37)is a measure of its broadapplicability. 560n euphuiasee Poet. 17,1455a31-34;22, 1459a6-7. 535 ONMETAPHOR ARISTOTLE Spoken sounds (ta en tei phonei) are tokens (sumbola)of experiencesin the psyche (ton en tei psukheipathemata),and writtenmarks(ta graphomena) [are tokens] of spoken sounds.And just as writtenmarks (grammata)are not the same for everyone,neitherare spoken sounds(phonai). But what these are signs (semeia)of in the firstplace-pathemata in the psyche-are the same for all; and what these pathemataare likenesses (homoiomata) of-actual things (pragmata)-are also the same.57 (De Interp.1, 16a3-8) Charles Sanders Peirce, the first great American semiotician, spoke of "semiosis" as the operation of signs of all kinds-the interaction of a triad of semiotic elements. These elements are [1] the object, the thing being represented; [2] the "representamen" or sign-whatever thing it is that stands for or signifies the object to someone-and [3] the interpretant, the all-important vehicle whereby the human brain makes the connection between sign and object-the concept whereby one grasps this connection, so that (for example) a red octagonal traffic sign is accepted as signifying that an automobile must come to a halt at that point. In the passage from De Interpretatione cited here, we have all the elements of the Peircean triad of semiosis: the sign (sumbolon or semeion), the object it represents (pragma), and the interpretant (en tei psukhei pathemata).58 Indeed this text must have influenced Peirce's formulation profoundly. Of particular interest in this passage is the word homoiomata, "likenesses" or "representations," which puts us in mind of representamen, Peirce's technical designation of the sign. Now a sumbolon and a semeion are also representations, but these need not have any visual similarity to, or organic connection with, the thing represented; indeed in writing-systems the latter will only be so in the case of pictograms, such as hieroglyphics. (In sound-systems, perhaps onomatopoeic words come the closest to such constitutional similarity.) Other forms of representation, however-those that we may term homoiomata-depend 57On this passage see Eco 1984, 27-29. 58The Peircean semiotic system is dense and at first forbidding, but the reader may be promised that any effort invested in its understanding will be recompensed a hundredfold. Some preliminary bibliography may be suggested: Peirce [1893-1910] 1940, Parmentier 1987, Merrell 1992, 1995a, 1995b. 536 JOHN T. KIRBY precisely upon such similarities or connections.59 In painting, for example, a portrait depends for its success first and foremost upon its effectiveness as a homoioma or likeness of the subject. Thus, when in response to the opening of T. S. Eliot's "Love Song of J. Alfred Prufrock," Let us go then, you and I, When the eveningis spreadout againstthe sky Like a patient etherisedupon a table, C. S. Lewis wrote, I am so coarse,the thingspoets see Are obstinatelyinvisibleto me. For twentyyears I've staredmy level best To see if evening-any evening-would suggest A patient etherisedupon a table; In vain.I simplywasn'table,60 he was objecting to the metaphor on the grounds that it was an insufficient homoi6ma of the thing described. Aristotle's assertion that to craft good metaphors (to... eu metapherein) depends upon being able to perceive likenesses-presumably likenesses in things that seem dissimilar, or at least likenesses that might not initially suggest themselves-is a key point. Of this latter ability he says, in a different context, The observationof likeness (he... tou homoiou theoria)is useful with a view both to inductivearguments(epaktikouslogous) and to hypothetical deductions (ex hupotheseossullogismous),and also with a view to the productionof definitions(ten apodosinton horismon).It is useful for inductivearguments,becauseit is by meansof an inductionof particularsin cases that are similar(kath'hekastaepi ton homoion) that we claimto induce the universal(to katholou);for it is not easy to do this if we do not know the points of likeness (ta homoia). It is useful for hypotheticalde- 590n representation and similarity see Belfiore 1992, 48-53, 63-65. She also (63 n. 46) draws attention to Sorabji 1972, 2-7. 60Excerpt from 'A Confession" in Poems by C. S. Lewis ? 1964 by the Executors of the Estate of C. S. Lewis; renewed 1992 by C. S. Lewis Pte. Ltd. Reprinted by permission of Harcourt Brace & Company and Curtis Brown Ltd. ARISTOTLEON METAPHOR 537 ductions because it is a reputable opinion (endoxon) that among similars (ton homoion) what is true of one is true also of the rest.61 (Top. 1.18, 108b7-14) In other words, the observation of likeness is a crucial cognitive step in the process of reasoning about the world-and also in the practice of articulating one's perceptions. Aristotle is talking here primarily about dialectical discourse, but the principles would seem to hold true for discourse of any kind, including the interior monologue. Now the articulation of a likeness is a kind of representation, and representation is itself quintessentially semiotic insofar as one thing signifies another. But the world is full of semiosis, and we should not forget that the entire Poetics itself is built on another semiotic infrastructure: that of mimesis.62 Like verbal metaphor, the whole process of artistic mimesis-be it in painting, music, or the composition of a tragedy-depends upon the artist's ability to perceive likenesses and to represent them (mimeisthai). Thus, for example, it is desiderated that the characters of a tragedy be good likenesses (to homoion, Poet. 15, 1454a24).63 Thus it is too that the pleasure we take in artistic mimesis depends upon recognition of likenesses: ... everyonedelightsin representations.An indicationof this is whathappens in practice:we delight in looking at the most proficientimages of thingsthat in themselveswe see with pain, e.g., the shapesof the most despised wild animalsand of corpses.The cause of this is that learningis very pleasant,not only for philosophers,but for others likewise,although the latter share in it to a small extent.For this reason they delightin seeing images,because it comes about that they learn as they observe (sumbaineitheorountasmanthanein),and inferwhateach thingis, e.g., that this person [in the picture] [represents] that one (houtos ekeinos). For if one has not seen the thing [that is represented] before, [its image] will not produce pleasure qua representation (hei mimema), but because of its [artistic] accomplishment, color, or some other such cause. (Poet. 4, 1448b8-19) 61Thistranslation is adapted from that of Pickard-Cambridge in Barnes 1984, I 180. For more on similarity and difference cf. Post. An. 2.13, 97b7-8. 62On this topic see Kirby 1991b. 63Admittedly this term is not entirely clear: does it mean "like [the mythic exemplar]"? "[life]like"? "like [us, the spectators]"? See Kirby 1991a, 201 and n. 12. 538 JOHN T. KIRBY In view of this, I propose that in our assessmentof metaphorin Aristotle we focus on a semiotic model.64Such a model, based on the triadicrelationshipof sign, object, and interpretant,will in fact make it possible to incorporatethe featuresand strengthsof other models, such as those based on comparison,substitution,or interaction,and will be suitable for the assessmentof verbal texts in either prose or verse, as well as of visual and other nonverbaltexts. I should like to think that a Peircean semiotic model would appeal to Tamba-Mecz and Veyne (1979),who, despite their rejectionof the Saussureanmodel,65nonetheless seem to insist on an interpretationthat is in fact compatiblewith Peircean semiotics.66Such a system ought to resonate favorablywith the work of the semioticianA. J. Greimas,of MichaelHaley, and possibly even, when some allowanceshave been made, with that of Liselotte Gumpel.67Certainlyit would accord with the work of Umberto Eco, who has written that "when closely studied in connection with verbal language, the metaphor becomes a source of scandal in a merely linguistic framework,because it is in fact a semiotic phenomenonpermitted by almost all semiotic systems."68 And, with some carefulattention to semanticdistinctions,I thinkthat a semioticpresentationof the Aristotelian model can be made congenial to Lakoff's cognitive approach. One of his majorobjectionsto the "classicaltheory"is that it sets off a 64See, e.g., Haley 1988 (esp. chs. 1-2) for a Peircean approach to poetic metaphor. I want it to be clear that the distinction between semiotics and semantics made by Benveniste (1967) and adopted by Ricoeur-where "the sign is the unit of semiotics while the sentence is the unit of semantics" ([1975] 1977, 69)-has no impact on my formulation here: as far as I am concerned, an entire sentence, or a paragraph, or indeed a whole book, may be a sign, or set of signs. 65E.g., "Ce ne sont pas des signes linguistiques dont le signifiant renvoie a un signifie" (89); "sa [Aristotle's] linguistique ignore la notion stoicienne ou saussurienne de signifie" (93). 66E.g., "La metaphora n'est donc pas une figure de rh6torique, au sens oiu nous l'entendons aujourd'hui, qui ressortit a l'analyse formelle et s6mantique du seul langage. Elle met en oeuvre toute une th6orie sur les rapports entre les choses, la pens6e, les mots. A travers la parole, Aristote voit les m6canismes d'intellection; derriere les mots, il vise les choses. Ou plutot il essaie de faire les deux a la fois, se livrant a un constant va-et-vient entre les 'intellections' et les noms, la pens6e et le langage" (81); "le mot ne signifie rien: il renvoie a une intellection, a un noema, et, par la, a une chose" (89). 67Greimas [1966] 1983, [1970-83] 1987; Haley 1988, Gumpel 1984. 68Eco [1983] 1984, 88. See too the treatment in Eco, Niklas, and Edeline [1986] 1994. ARISTOTLE ON METAPHOR 539 putative "figurative"or "poetic"languagein contradistinctionto everyday usage; and, while it is true that Aristotle has a notion of kuria or ordinaryterms, he by no means restrictedmetaphor to poetic or extraordinarycontexts. On the contrary,as we have seen, he explicitly contradictsIsocrateson this count;and his remarkthat "everyoneconverses using metaphors"may be seen as actuallyprescientof the cognitive approach.69 Another of Lakoff's objectionsto the so-called classicaltheory is that it locates the phenomenon of metaphorin the realm of language, Here again,the not (like the cognitivetheory) in the realmof thought.70 Aristotelian semiotic approach actually anticipatesthe new cognitive model-which is itself manifestly semiotic in nature: the notion of "cross-domainmapping in the conceptual system"71is not a bad description at all, in fact, of the phenomenon of semiosis, in which information from a source-domain (the object) is conceptually(i.e., by interpretants)mapped onto a target-domain, which mappingis represented by a sign;and, because of what Peirce calls "unlimitedsemiosis," the pertinent interpretantsmay themselves be represented by new signs, being connected to the latter by new interpretants,and so on ad infinitum. 69Lakoff would probably point (disapprovingly) to Aristotle's distinction between the kurion kai oikeion and metaphora, as evidence of a literal/figurative dichotomy; but we must be wary of mapping Lakoff's "classical theory" precipitously onto what Aristotle has actually said. "kurion kai oikeion" is by no means the same as "literal"; rather, the Aristotelian phrase refers to familiarity and typicality of usage. Aristotle valorizes kuria, not because they are more "literal" or come closer to unmediated truth, but because-unlike, e.g., glottai-they are semiotically the most immediately productive of interpretants (Rhet. 3.10.2, 1410bl12-13)and thus conduce especially to learning. But so, he asseverates, does metaphor (ibid.). 70For an egregious example of the shift, in non-Aristotelian paradigms, from the conceptual to the purely linguistic, consider the case of homonymy. For the modern reader the words "isle" and "aisle" are homonyms; for Aristotle the homonyms in this situation would be the actual things themselves-the piece of land surrounded by water, and the walkway. (See Categ. 1, lal-5, and the note ad loc. in Ackrill 1963, 71: "it is important to recognize from the start that the Categories is not primarily or explicitly about names, but about the things that names signify.") The instances in the Rhetoric of the term hom6numia(i) are typically translated "homonym(s)," but most accurately they should be "situation(s) where there is homonymy"; see Rhet. 2.24.2, 1401a13;3.11.8, 1412bll, 12; and even, I would say, 3.2.7,1404b38. 71Lakoff 1993, 203. 540 JOHN T. KIRBY A semiotic model of metaphor, in which signs are connected to objects by interpretants, would have the following virtues. [1] It would steer clear of arguments over comparison (or substitution) versus interaction, because it incorporates the crucial assertions and the strengths of all of these models. [2] It would highlight the quintessentially cognitive nature of the Aristotelian formulation. And [3] with very little adjustment for the context in question, it would make Aristotle's theory of metaphor applicable even to nonverbal situations-for, as students of Peirce know, the linguistic sign is but one of myriad kinds of sign, and semiosis is all around us, in what Merrell calls "our perfusive, pervasive universe."72Aristotle was well aware of this, as the opening chapters of the Poetics show.73 Metaphor in the Rhetoric With this background from the Poetics we are now equipped to look carefully at Aristotle's treatment of metaphor in the Rhetoric. By the time he comes to write the third book of the Rhetoric, he finds it convenient to refer us to the earlier discussion: "Now what each kind of word is and how many species of metaphor there are and that metaphor has very great effect both in poetry and speeches has been said, as noted above, in the Poetics" (Rhet. 3.2.7, 1405a3-6).74 The text that concerns us begins at 3.2.1 (1404bl), where the virtue of style (lexeos arete) is defined as "being clear" (saphe einai), on the grounds that logos75 is a kind of sign (semeion), and that, to do its work, it must make (things) clear (ean me deloi ou poiesei to heautou ergon). The word "work" here, ergon, is closely tied up with Aristotle's notion of a final cause or pur- 72Merrell 1991. Here too one ought to have a look at Goodman 1968 and Hausman 1989. 730n this see Kirby 1991b, 119-22. 74The topic of the relative dates of the Rhetoric and Poetics is hopelessly problematic; he may well have revised both after the composition of both. See Kirby 1991a, 198 and n. 4. 751hesitate to translate this infinitely slippery term here. It seems to be susceptible of at least the same polysemy as English "speech" and French "discours"-i.e., to admit both the notion of "verbal communication" and of "oration." I am inclined to the former of these translations here, as one would more naturally think of a complete oration as a collection of semeia; but we should keep in mind that Aristotle is first and foremost addressing the topic of stylistic excellence in the specific context of oratory. ARISTOTLE ON METAPHOR 541 pose.76 By the same token, Aristotle is aware of the rhetorical value of deviating from prevailing usage (exallaxai), especially by the use of "strange language" (poiein xenen ten dialekton).77 The particular goal here, besides the achievement of semnotes or elevated style, seems to be the sheer pleasure of the audience; as we learn from a pair of linked syllogisms (one partially implicit): "One should make one's language unfamiliar (xenen). For listeners marvel at (thaumastai... eisin) things that are distant (ton aponton [here a synonym of xenen]), and what is marvelous is pleasant (hedu [lit. 'sweet'])." This is acknowledged as a common goal in both prose and verse, but (we are told) prose has fewer resources in this regard than verse-namely, kuria and metaphor (3.2.6, 1404b31-33; cf. also above on 1404b34-36). In 3.2.8 (1405a6-10) Aristotle begins to focus carefully on the nature of metaphor: it excels particularly in the aspects he has been discussing-clarity (to saphes), sweetness (to hedu), and unfamiliarity (to xenikon). Moreover, in an unusual aside reminiscent of his remark in the Poetics (22, 1459a6-7), he informs us that one cannot learn its use from another. He has already (Rhet. 3.2.3-6, 1404b) touched on the need to make language appropriate, in order to seem natural; here (3.2.9, 1405a10-14) he advises that metaphors too should be appropriate. This natural quality is seen, metaphorically enough, in terms of "theft"; by this he seems to mean, generally, the use of unfamiliar language in a way that does not seem unnatural (3.2.5, 1404b24) and, specifically, the use of metaphor in a way that, because appropriate to its context, does not call attention to itself (3.2.10, 1405a31).78 76See Phys. 2.3, 194b32-195a3; Metaphys. A 2, 1013a32-b3; Kirby 1991a, 210-13. 77Xenous ("strangers,"perhaps "foreigners") is here explicitly contrasted with politas ("citizens," i.e., probably "Athenians"), and one might reasonably think that Aristotle is speaking of the use of glottai; but that term appears a few lines later (3.2.5, 1404b28), in a list that is (as we have said) not dissimilar to the one in Poetics 21. It looks rather as if poiein xenen may itself be a metaphor for "strangeness"-hence Kennedy's translation "make the language unfamiliar" (1991, 221) and footnote ad loc. on the Russian Formalist doctrine of defamiliarization (ostraneniye). Yet again I would draw attention to the fact that Aristotle is talking about what might be called semiotic marking. 78Kleptein carries a specific connotation of "stealing" but, more generally, of "stealth." For forms of kleptein meaning "conceal" see, e.g., Pind. 01. 6.36; Soph. Aj. 1137, El. 37; Plato Laws 10, 910B1;and Rhet. ad Alex. 35.4, 1440b21.It is possible that Aristotle's concept of artistic concealment here ought to be connected with Gorgias' doctrine of apate, the "deception" wrought by Attic tragedy (fr. 82 B23 DK). On apate see, e.g., Rosenmeyer 1955 and Segal 1962. 542 JOHN T. KIRBY In 3.2.12(1405a35-b6)a curiousand importantuse of metaphoris delineated:the namingof thingsthat do not have propernames of their own (metapherein (epi) ta anonuma onomasmenos). As Kennedy points out,79this is related to the figureknown in Greek as katakhresisand in Latin as abusio;of key importancefor our study,however,is how Aristotle explicitlyrecognizesthat there will be cases when our only semiotic recoursewill be to metaphor-another way in which he foreshadows the modern cognitive approach.In the same context (1405b5;cf. 3.11.6,1412a23-24),metaphor is said to be like the ainigma or riddle (metaphorai... ainittontai,lit. "metaphorsspeak in riddles").In keeping with our semiotic approachto metaphor,we are well equipped to understandthis, because in either case, the cognitive connection that enables the reader/listenerto understandthe meaningis an interpretant. Take the most famous ainigma of antiquity,the riddle posed by the Sphinx:"Therewalks upon the earth a creaturetwo-footed and fourfooted and three-footed, that has but one voice; and alone of all creatures that move on land and in the air and in the sea, it is able to change its nature.Indeed when it goes supportedon the most feet, then is the speed of its limbs the feeblest."80The riddle perpendsa series of signs: the four-footed, two-footed, and three-footed creaturesthat are one creature.On solving the riddle, the reader/listenercomes to see that these representthree related objects:the baby,the grownman, and the elderly man, respectively.By the same token, it is easy to see that these signs representtheir objectsby virtue of metaphors;moreover,that the that arisesin processof semiosisis not completewithoutthe interpretant the reader'sor listener'smind,to make sense of the connectionbetween sign and object. When that connection is made, several things happen: firstof all, the reader/listenersays houtos ekeinos-"This is a representation of that"-and takes, presumably,some pleasure, both in the recognitionand in the poet's craft (Poet. 4, 1448b8-19).Also as a result of this process, the reader/listenerlearns something new about what, followingShakespeare,we may call the Ages of Man;and Aristotle says that learning is a source of pleasure (Rhet. 3.10.2, 1410b10-12).More79Kennedy 1991, 224 n. 32. 80I have cited the version attributed to Asclepiades, as given in Athen. Deipn. 456B. It is found also in two manuscripts of Sophocles and in the scholia on Eur. Phoen. 50. The original Greek consists of five verses of dactylic hexameter. It is possible that this version dates from Aristotle's own lifetime. A later version, apparently abbreviated from this one, is preserved by Apollodorus (Bibl. 3.5.8). ARISTOTLE ON METAPHOR 543 over, because the interpretantsconnecting the signs in the riddle and their objects are not immediately obvious, there will be something quasi-philosophicalabout pursuingthe interpretationof these metaphors (Rhet.3.11.5,1412a9-12). The next thesis is that the source of the metaphorshouldbe something beautiful (3.2.13, 1405b6-21).This value, like all our most basic ideologicalunderpinnings,is one that has come undersevere scrutinyin the postmodern era: why should the beautiful be valorized over the ugly?But Aristotle seems to have an explicitlyutilitariananswerfor the aspiringrhetor:beauty is here connectedboth with appropriatenessand with the vivid evocation of interpretants in the audience (3.2.13, 1405bll-13).81What is certain,in view of his heavily semiotic approach to metaphor overall, is that he is not here espousing a merely ornamental (Isocratean) model for metaphor.The rest of Rhetoric 3.2 is devoted to exploring appropriatenessand beauty in epithets and diminutives. Chapter3.3 deals with psukhra,vices (lit. "chills")of style. These arise from four sources:"doublewords"(dipla onomata);foreign terms (glottai);epithets, when these are too large, untimely,or too numerous; and metaphor.Again it is importantto note that Aristotle's guidelines are context-sensitive, for (assumingthat the epithet is what he means by kosmos at Poetics 21, 1457b2)82each of these four items appearsin the list in Poetics 21. The point here in Rhetoric3.3 is that what is appropriateto verse may not be so in prose oratory.Hence he criticizes certaindipla because they seem "poetic"(3.3.1,1406a6);the injudicious use of epithets will make a prose discourseseem (too much) like poetry (3.3.3, 1406al4). So it is with metaphor:what might be appropriatein comic or tragicdramais not, as a rule, fittingfor oratory (3.3.4, 1406b6-8).Too, if the orator has to "reach too far," the metaphors will be unclear 81In this connection I think back to what Aristotle said (Poet. 4, 1448b8-12) about the pleasure that can be taken even from the images of disgusting things such as corpses. Perhaps there he was referring specifically to our ability to admire technical prowess even when the subject matter is repulsive. (I surmise, for example, that this was part of the project of Robert Mapplethorpe.) That would be a rather different matter, then, than the verbal (oral) evocation of interpretants in oratory. On the passage in Poetics 4 see Belfiore 1992, 61-63. 82So (rightly, I think) Dupont-Roc and Lallot 1980, 342, and Janko 1987, 130-31; but see Bywater 1909, 280-81, and Stanford 1936, 7. 544 JOHN T KIRBY (asapheis, 3.3.4, 1406b8)83 and too poetic (poietikos... agan, 3.3.4, 1406b10-11). Chapter4 launchesdirectlyinto a discussionof the eikon or simile. As we saw earlier,Plato not only uses similes himself in his writing but also speaks of eikon in several contexts, sometimes to mean what we would ourselves term similes. Thus it is not surprisingto find that Aristotle coopts the term here (or that three of his examples in 3.4.3, 1406b32-1407al,are from the Republic).For him, a simile is a species of the genus metaphor,84and the one differs only slightlyfrom the other (3.4.1, 1406b20).The nature of that difference has been minutely and perceptivelystudied by Tamba-Meczand Veyne,85and has to do with the presence (in the simile) of an explicit comparativephrase,the prothesisor "setting-forth"of the comparison:"thatis to say, the poet creates an expectationin the listener or readerthat demandsfulfilment."86 Like the virtuallyuntranslatableparticlesmen and de, like the protasis and apodosisof a condition,the prothesisof a simile and the phrasethat answersit drawtwo (or more) ideas into parallel.In the case of a simile the parallelism,broughtto completionin what we may call sunthesis,87 allows the transferencethat makes metaphor.88 In 3.10Aristotle returnsto the topic of metaphor,underthe heading of ta asteia ("urbanities"is Kennedy's felicitous rendering).89In keeping with the theoretical/investigativegoal of the Rhetoric,90Aristotle seeks to offer an analysisof the essentialnatureof ta asteia,and as in all such endeavors,he recognizeshere the need for what he calls an 83Thisagain harks back to the virtue of clarity;cf. Rhet. 3.2.1, 1404b2-3;3.2.8, 1405a8.(N.B.:Kennedy[1991,228] translatesnot "unclear"but "inappropriate" as if reading aprepeis here. This does not seem to be an attested variant reading but is a plausible conjectural emendation.) 84Students of later classical rhetoric will be interested to note that for Quintilian (8.6.8) the converse is true: the metaphor is a kind of simile. 85Tamba-Mecz and Veyne 1979, esp. 95-97. 86Tamba-Mecz and Veyne 1979, 95 (my translation). 87Cf.Tamba-Mecz and Veyne 1979, 97. 88Though they might well have done, Tamba-Mecz and Veyne do not draw attention to Aristotle's analysis of the period (Rhet. 3.9, 1409a24-1410b5) in connection with this. 89The word, derived from astu, "city,"means "cityish." It is typically understood to connote elegance and wit-perhaps "sophistication" would be another useful translation. 901am thinking of such statements as "[rhetoric's] function is not to persuade (peisai), but to see (idein) the available means of persuasion in each case" (Rhet. 1.1.14, 1355bl0-11). ARISTOTLE ON METAPHOR 545 arkheor "firstprinciple."In this case, that is as follows:"Tolearn easily (to gar manthanein rhaidios) is naturally pleasant to all people, and words signify something,so whateverwords create learning(mathesin) Thatthis is indeed an in us are the most pleasant"(3.10.2,1410b10-12).91 arkhe for understandingthe nature and function of metaphorin rhetoric is immediatelyclear from the next sentence:92"metaphorespecially brings this [learning]about; for [e.g.] when he [Homer] calls old age 'stubble,'he creates understanding(mathesin)and knowledge (gnosin) through [reference to] the genus, since both old age and stubble are [species of the genus of] things that have lost their bloom" (3.10.2, In other words, Aristotle has in mind at least two things 1410bl13-15). about how metaphorworks:[1] the cognitive exercise of decoding, or (so to speak) "unraveling"the metaphor,is an occasion of enlightenment, and [2] this exercise is also pleasurablefor the audience. In 3.10.3he again designatesthe simile a species of metaphor,but he devalues it for reasons of verbal economy:the simile is less "sweet" because it takes longer to express (hetton hedu, hoti makroteros, 1410b18).Moreover,the simile does not say that X is Y (1410b19).What Aristotle means here, and in the next clause-by "thusthe psyche also does not seek this"-is a bit opaque. Kennedy translates the latter phrase "nordoes [the listener's]mind seek to understandthis"and says that '"Aristotle,unlike later classical rhetoricians, thus implies that In that case, "thepsyche also does metaphoris a form of predication."93 not seek this" would mean, presumably,that with a simile, the mind does not seek to make the predicativeconnections that it does-for syntacticalreasons-in the case of metaphor. In 3.10.5-6 metaphor is yoked with "shaped"or "figured"language-by which Aristotle means such syntacticstrategies as antithesis-as a source of urbanity.94 Only,it is stipulated,the metaphorshould 91Kennedy (1991, 244 n. 112) is quite right to draw attention to the opening of the Metaphysics here: 'All humans by nature seek to know." Lear (1988) emphasizes that "know" in that context means "understand." 92And corroborated in 3.10.4, 1410b20-21, where things that conduce to quick learning are said by necessity (anagke) to be asteia. The word anagke is Aristotle's signal for the force of syllogistic inference. Taking this as the major premise, then, and his assertion in 3.10.2, 1410b10-15, that metaphor especially promotes learning as the minor, we may infer that metaphor is to be classed under the heading of asteia. 93Kennedy 1991, 245 n. 114, also citing Ricoeur [1975] 1977. 94Again I draw the reader's attention to Isocrates' list of enthumema, antithesis, and parisosis in Panathenaicus 2. 546 JOHN T KIRBY be neither "strange"(allotrian)nor "superficial"(epipolaion).The former is difficultto understand(sunidein), and the latter makes the listener experience nothing-by which I take him to mean that there is none of the sort of cognitive decoding that he refers to in 3.10.2 (1410b13-15).95 At the end of 3.10.6 (1410b35-36)metaphor is listed, along with antithesis and energeia,"actualization,"as a major source of urbanity. From the passage of which this is a final summary,it would appearthat energeia is a synonym of what Aristotle calls pro ommaton poiein, "bringing-before-the-eyes,"of which he says, "thingsshould be seen as being done rather than as in the future"(1410b34-35).And indeed this is made explicitin 3.11.1(1411b25-26),wherepro ommatonpoiein is said to signifywhateveris in activity(energounta).In certaincases (e.g., 3.10.7, 1411a25-bl; 3.11.2-3.11.4,1411b28-1412a9)but not in all (e.g., 3.11.2,1411b26-28),metaphorand energeiaoverlap. At 3.11.5we come to an importantelaborationof the semiotictheory of metaphor:here Aristotle comparesthe understandingof metaphor to the practice of philosophy,saying that "metaphorsshould be transferredfrom things that are related but not obviously so (apo oikeion kai me phaneron), as in philosophy too, it is characteristicof a well-directed (eustokhou)mind to observe the likeness even in things very different"(1412a9-12).Both pursuits,then, depend upon the cognitive decoding or unravelingwe have alreadymentioned;and both depend upon the formationof interpretantsin the mind.Connectedto this is the point made in the next section, where Aristotle maintainsthat "in most cases, ta asteia come via metaphorplus an added surprise(ek tou prosexapatan),for it becomes clearer [to the listener] that he learned somethingdifferentfrom what he believed, and his mind seems to say, 'How true, and I was wrong'" (3.11.6,1412al7-21).96 In 3.11.11-13Aristotle once again examines the topic of similes. Here we are told that, as with the metaphorfrom analogy,there will always be two terms involved. In 3.11.14-15he mentions proverbialex95Kennedy (1991, 245 n. 117) is right to point out that in Poetics 21, 1457b7, metaphor is defined as the epiphora of an onoma allotrion. It may be that in the Rhetoric here, allotrian has a different sense, or that Aristotle originally wrote allotri(oter)an, "too alien"; cf. Nic. Eth. 8.12.4, 1162a3, for the opposed forms allotrioteroi and oikeioteroi. 96Forthe narrative efficacy of events that happen contrary to expectation cf. Poet. 9, 1452al-4. ARISTOTLE ON METAPHOR 547 pressions and hyperboles as types of metaphor as well, associating hyperbolewith the exaggerationof youth. While he does not say so explicitlyhere, he probablyintends to set these two-the use of proverbs and the use of hyperbole-in polar opposition, for, just as he remarks here that it is inappropriatefor an old man to speak in hyperbole (3.11.15,1413bl-2), he has also told us earlier(2.21.9,1395a2-5) that the use of proverbialexpressionsis appropriateto old age but not to youth. Over and over, in both the Poetics and the Rhetoric,we find a nexus of ideas in Aristotle'smetaphortheory:that the phenomenonis a semiotic one; that it involves a cognitiveprocessof decodingon the part of the audience;that it should add sophisticationto the discourse,resultingin pleasureon the partof the audience;and that its use shouldbe suited or appropriateto its context.Farfrom being a dispensable,extraneous trope, metaphorfinds itself, in a numberof ways, at the heart of Aristotelianrhetoricand poetics, precisely because it is fundamentally linked to Aristoteliansemiotics.Because of this, moreover,metaphoris also a centralelement in Aristotle'scognitivesystem.It is to metaphor, he says, that we resort when a thing can be named in no other way;to craft a felicitous metaphoris to give evidence of genius, and is a skill that cannot be learnedfrom another. To make metaphorsis, quintessentially,to be able to perceivelikeness-and thus, almost by definition,difference.In this respect, metaphor epitomizes or recapitulatesin itself all of language-that mysterious, miraculousmeans by which we mirrorthe whole world around us. It is perhaps small wonder, then, that he who searched so deeply into other aspects of semiotics should also have been the first to give us a critical vocabularyfor this sign of signs, the sign that signifies signification.97 PURDUE UNIVERSITY e-mail: [email protected] 97I owe heartfelt thanks to Elizabeth Belfiore, Thomas Broden, Patricia Kenig Curd, Keith Dickson, David Engel, Eugene Garver, Richard Janko, George Kennedy, Howard Mancing, Floyd Merrell, Neil O'Sullivan, Carol Poster, and Anthony Tamburri for their encouragement and their assistance in various ways. 548 JOHN T. KIRBY BIBLIOGRAPHY Ackrill, J. L., ed. 1963. Aristotle's Categories and De Interpretatione. Translated with notes. Oxford: Clarendon Press. Asmis, Elizabeth. 1992. "Plato on Poetic Creativity." In The Cambridge Companion to Plato, edited by Richard Kraut, 338-64. Cambridge: Cambridge University Press. Barfield, Owen. [1928] 1973. Poetic Diction: A Study in Meaning. London: Faber & Gwyer. 3d ed. Middletown, Conn.: Wesleyan University Press. Barnes, Jonathan, ed. 1984. The Complete Works of Aristotle: The Revised Oxford Translation. 2 vols. Princeton: Princeton University Press. Belfiore, Elizabeth. 1992. Tragic Pleasures: Aristotle on Plot and Emotion. Princeton: Princeton University Press. Benveniste, Emile. 1967. "La forme et le sens dans le langage." In Le langage: Actes du XIIIe Congres des Societes de Philosophie de Langue Francaise. Neuchatel: La Baconniere. Black, Max. 1955. "Metaphor." Proceedings of the Aristotelian Society 55 (195455) 273-94. .1979. "More about Metaphor." In Ortony 1993, 19-41. Bontekoe, Ron. 1987. "The Function of Metaphor." Philosophy and Rhetoric 20:209-26. Booth, Wayne. 1978a. "Metaphor as Rhetoric: The Problem of Evaluation." Critical Inquiry 5:49-72. .1978b. "Ten Literal 'Theses.'" Critical Inquiry 5:175-76. Brooke-Rose, Christine. 1958. A Grammar of Metaphor. London: Secker & Warburg. Bywater, Ingram. 1909. Aristotle on the Art of Poetry. A revised text with critical introduction, translation, and commentary. Oxford: Clarendon Press. Chroust, Anton-Hermann. 1973. Aristotle: New Light on His Life and Some of His Lost Works. Vol. I (of two), Some Novel Interpretations of the Man and His Life. Notre Dame: University of Notre Dame. Cooper, David E. 1986. Metaphor. Oxford: Blackwell. Derrida, Jacques. [1971a] 1982. "La mythologie blanche: la m6taphore dans le texte philosophique." First published in Poetique 5 (1971), then in his Marges de la philosophie (Paris: Minuit, 1972). Translated by Alan Bass as pp. 207-71 in Margins of Philosophy (Chicago: University of Chicago Press, 1982). . [1971b] 1982. "Le supplement de la copule: la philosophie devant la linguistique." First published in Langages, 24 December 1971, then in his Marges de la philosophie (Paris: Minuit, 1972). Translated by Alan Bass as pp. 175-205 in Margins of Philosophy (Chicago: University of Chicago Press, 1982). ARISTOTLEON METAPHOR 549 Dubois, J., E Edeline, J.-M. Klinkenberg, P. Minguet, E Pire, and H. Trinon (Groupe [t). [1970] 1981. Rhetorique generale. Paris: Larousse. Translated by Paul B. Burrell and Edgar M. Slotkin as A General Rhetoric. Baltimore: The Johns Hopkins University Press. duBois, Page. 1988. Sowing the Body: Psychoanalysis and Ancient Representations of Women. Chicago: University of Chicago Press. Dupont-Roc, R., and J. Lallot. 1980. Aristote: La Poetique. Paris: Seuil. Eco, Umberto. [1983] 1984. "The Scandal of Metaphor." Poetics Today 4:217-57. Reprinted in Eco 1984, 87-129. .1984. Semiotics and the Philosophy of Language. Bloomington: Indiana University Press. Eco, Umberto, Ursula Niklas, and Francis Edeline. [1986] 1994. "Metaphor." In Encyclopedic Dictionary of Semiotics, edited by T. A. Sebeok, 534-49. 2d ed. Berlin: Mouton/de Gruyter. Edwards, Mark W. 1991. The Iliad: A Commentary. Vol. V, Books 17-20. Cambridge: Cambridge University Press. Fine, Gail. 1993. On Ideas. Oxford: Clarendon Press. Foss, Martin. 1949. Symbol and Metaphor in Human Experience. Princeton: Princeton University Press. Frankel, Hermann. 1921. Die homerischen Gleichnisse. Leipzig: Vandenhoeck & Ruprecht. Genette, Gerard. [1970] 1982. "La rh6torique restreinte." First published in Communications 16 (1970); collected in his Figures III (Paris: Seuil, 1972). Translated as pp. 103-26 of Figures of Literary Discourse (New York: Columbia University Press, 1982). Gibbs, Raymond W. 1994. The Poetics of Mind: Figurative Thought, Language, and Understanding. Cambridge: Cambridge University Press. Golden, Leon. 1992. Aristotle on Tragic and Comic Mimesis. Atlanta: Scholars Press. Goodman, Nelson. 1968. Languages of Art: An Approach to a Theory of Symbols. Indianapolis: Bobbs-Merrill. Greimas, A. J. [1966] 1983. Semantique structurale:recherche de methode. Paris: Larousse. Translated by D. McDowell and R. Schleifer as Structural Semantics: An Attempt at a Method. Lincoln: University of Nebraska Press. . [1970-83] 1987. Du sens: essais semiotiques. 2 vols. Paris: Seuil. Translated by Paul J. Perron and Frank H. Collins as On Meaning: Selected Writings in Semiotic Theory. Minneapolis: University of Minnesota Press. Gumpel, Liselotte. 1984. Metaphor Reexamined:A Non-Aristotelian Perspective. Bloomington: Indiana University Press. Guthrie, W. K. C. 1981. Aristotle: An Encounter. Vol. VI of A History of Greek Philosophy. Cambridge: Cambridge University Press. 550 JOHN T. KIRBY Haley, Michael C. 1988. The Semeiosis of Poetic Metaphor. Bloomington: Indiana University Press. Harries, Karsten. 1978a. "Metaphor and Transcendence." Critical Inquiry 5: 73-90. .1978b. "The Many Uses of Metaphor." Critical Inquiry 5:167-74. Hausman, C. 1989. Metaphor and Art. Cambridge: Cambridge University Press. Haverkamp, Anselm, ed. 1983. Theorie der Metapher. Darmstadt: Wissenschaftliche Buchgesellschaft. Hester, Marcus B. 1967. The Meaning of Poetic Metaphor: An Analysis in the Light of Wittgenstein'sClaim That Meaning Is Use. The Hague: Mouton. Hogan, J. C. 1966. The Oral Nature of the Homeric Simile. Diss. Cornell University. Jakobson, Roman. [1956] 1987. "Two Aspects of Language and Two Types of Aphasic Disorders." In The Fundamentals of Language, edited by R. Jakobson and R. Halle. The Hague: Mouton. Reprinted in his Language in Literature, 95-114. Cambridge: Harvard University Press. . [1960] 1987."Linguistics and Poetics." Closing Statement in Style in Language, edited by T. A. Sebeok. Cambridge: MIT Press. Reprinted in his Language in Literature, 62-94. Cambridge: Harvard University Press. Janko, Richard, ed. 1987.Aristotle: Poetics I. Indianapolis: Hackett. .1992. The Iliad:A Commentary. Vol. IV, Books 13-16. Cambridge: Cambridge University Press. Johnson, Mark, ed. 1981. Philosophical Perspectives on Metaphor. Minneapolis: University of Minnesota Press. Kamerbeek, J. C. 1984. The Plays of Sophocles: Commentaries. Part VII, The Oedipus Coloneus. Leiden: E. J. Brill. Kassel, Rudolf, ed. 1965. Aristotelis De Arte Poetica Liber. Oxford: Clarendon Press. , ed. Aristotelis Ars Rhetorica. Berlin: de Gruyter. Keith, A. L. 1914. Simile and Metaphor in Greek Poetry from Homer to Aeschylus. Diss. University of Chicago. Kennedy, George. 1963. The Art of Persuasion in Greece. Princeton: Princeton University Press. , ed. 1991.Aristotle on Rhetoric:A Theory of Civic Discourse. New York: Oxford University Press. Kirby, John T. 1990. "The 'Great Triangle' in Early Greek Rhetoric and Poetics." Rhetorica 8:213-28. . 1991a. 'Aristotle's Poetics: The Rhetorical Principle." Arethusa 24: 197-218. . 1991b. "Mimesis and Diegesis: Foundations of Aesthetic Theory in Plato and Aristotle." Helios 18:113-28. . 1992. "Rhetoric and Poetics in Hesiod." Ramus 21:34-60. Kirk, G. S. 1962. The Songs of Homer. Cambridge: Cambridge University Press. ARISTOTLEON METAPHOR 551 Kirk, G. S., J. E. Raven, and M. Schofield, eds. [1957] 1983. The Presocratic Philosophers. 2d ed. Cambridge: Cambridge University Press. Kittay, Eva Feder. 1987. Metaphor: Its Cognitive Force and Linguistic Structure. Oxford: Clarendon Press. Knights, L. C., and Basil Cottle. 1960. Metaphor and Symbol. London: Butterworths. Kosko, Bart. 1991. Neural Networks and Fuzzy Systems. Englewood Cliffs, N.J.: Prentice-Hall. . 1993. Fuzzy Thinking: The New Science of Fuzzy Logic. New York: Hyperion. Lacan, Jacques. [1957] 1977. "L'instance de la lettre dans l'inconscient ou la raison depuis Freud." First published in La Psychanalyse 3 (1957); collected in his Ecrits (Paris: Seuil, 1966). Translated by Alan Sheridan as "The Agency of the Letter in the Unconscious, or Reason since Freud," in Ecrits:A Selection, 146-78. New York: Norton. Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. . 1993. "The Contemporary Theory of Metaphor." In Ortony 1993, 202-51. Lakoff, George, and Mark Johnson. 1980. Metaphors We Live By. Chicago: University of Chicago Press. Lakoff, George, and Mark Turner. 1989. More Than Cool Reason:A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press. Laks, Andr6. 1994. "Substitution et connaissance: une interpretation unitaire (ou presque) de la theorie aristotelicienne de la metaphore." In Aristotle's Rhetoric: Philosophical Essays, edited by David J. Furley and Alexander Nehamas, 283-305. Princeton: Princeton University Press. Lattimore, Richmond, trans. 1951. The Iliad of Homer. Chicago: University of Chicago Press. Lear, Jonathan. 1988. Aristotle: The Desire to Understand. Cambridge: Cambridge University Press. Lee, D. J. N. 1964. The Similes of the Iliad and the Odyssey Compared. Parkville, Victoria: Melbourne University Press. Levin, Samuel R. 1977. The Semantics of Metaphor. Baltimore: The Johns Hopkins University Press. Lewis, C. S. 1964. Poems. Edited by Walter Hooper. New York: Harcourt Brace Jovanovich. Lloyd, G. E. R. 1987. The Revolutions of Wisdom: Studies in the Claims and Practice of Ancient Greek Science. Berkeley and Los Angeles: University of California Press. Lodge, David. [1977] 1988. The Modes of Modern Writing:Metaphor, Metonymy, and the Typology of Modern Literature. London: Edward Arnold. Reprinted Chicago: University of Chicago Press. 552 JOHN T. KIRBY McCall, Marsh H., Jr. 1969. Ancient Rhetorical Theories of Simile and Comparison. Cambridge: Harvard University Press. Mac Cormac, E. R. 1985. A Cognitive Theory of Metaphor. Cambridge: MIT Press. MacDowell, D. M., ed. 1971.Aristophanes: Wasps. Oxford: Clarendon Press. McNeill, Daniel, and Paul Freiberger. 1993. Fuzzy Logic: The Revolutionary Computer Technology That Is Changing Our World. New York: Simon & Schuster. Man, Paul de. 1978. "The Epistemology of Metaphor." Critical Inquiry 5: 13-30. Martin, Richard P. 1989. The Language of Heroes: Speech and Performance in the Iliad. Ithaca: Cornell University Press. Martinich, A. P. 1984. '"ATheory for Metaphor." Journal of Literary Semantics 13:35-36. Merrell, Floyd. 1991. Signs Becoming Signs: Our Perfusive, Pervasive Universe. Bloomington: Indiana University Press. .1992. Sign, Textuality,World. Bloomington: Indiana University Press. .1995a. Semiosis in the Postmodern Age. West Lafayette, Ind.: Purdue University Press. . 1995b. Peirce's Semiotics Now: A Primer. Toronto: Canadian Scholars' Press. Miall, David, ed. 1982. Metaphor: Problems and Perspectives. Brighton: Harvester Press. Mooij, J. J. A. 1976. A Study of Metaphor: On the Nature of Metaphorical Expressions, with Special Reference to Their Reference. Amsterdam: NorthHolland. Moulton, Carroll. 1977. Similes in the Homeric Poems. Gottingen: Vandenhoeck & Ruprecht. Nietzsche, Friedrich. [1873] 1989. "Uber Wahrheit und Luge im aussermoralischen Sinne." In Nietzsche: Werke.Kritische Gesamtausgabe, Abt. 3, Bd. 2, edited by G. Colli and M. Montinari, 369-84. Berlin: de Gruyter, 1973. Translated as "On Truth and Lying in an Extramoral Sense." In Friedrich Nietzsche on Rhetoric and Language, edited by S. Gilman et al., 246-57. New York: Oxford University Press. Nimis, S. A. 1987. Narrative Semiotics in the Epic Tradition:The Simile. Bloomington: Indiana University Press. Ortony, Andrew, ed. [1979] 1993. Metaphor and Thought. 2d ed. (much revised and enlarged). Cambridge: Cambridge University Press. O'Sullivan, Neil. 1992. Alcidamas, Aristophanes and the Beginnings of Greek Stylistic Theory. Stuttgart: Franz Steiner. Paprotte, Wolf, and Ren6 Dirven. 1985. The Ubiquity of Metaphor: Metaphor in Language and Thought. Amsterdam: Benjamins. ARISTOTLEON METAPHOR 553 Parmentier, Richard. [1987] 1994. "Peirce Divested for Non-Intimates." Recherches semiotiques/Semiotic Inquiry 7:19-39. Reprinted as chapter 1 of his Signs in Society: Studies in Semiotic Anthropology. Bloomington: Indiana University Press. Parry, Milman. [1933] 1971. "The Traditional Metaphor in Homer." Classical Philology 28:30-43. Collected in his The Making of Homeric Verse, 36575. Oxford: Clarendon Press. Peirce, Charles Sanders. [1893-1910] 1940. "Logic as Semiotic: The Theory of Signs." In The Philosophy of Peirce: Selected Writings, edited by Justus Buchler, 98-119. London: Routledge & Kegan Paul. Reprinted 1955 and thereafter as Philosophical Writings of Peirce. New York: Dover Books. Perelman, Chaim, and Lucie Olbrechts-Tyteca. [1958] 1969. Traite de l'argumentation: la nouvelle rhetorique. Paris: Presses Universitaires de France. Translated by J. Wilkinson and P. Weaver as The New Rhetoric:A Treatise on Argumentation. Notre Dame: University of Notre Dame Press. Quinn, Naomi. 1991. "The Cultural Basis of Metaphor." In Beyond Metaphor: The Theory of Tropes in Anthropology, edited by James W. Fernandez, 56-93. Stanford: Stanford University Press. Richards, I. A. 1936. The Philosophy of Rhetoric. London: Oxford University Press. Richardson, Nicholas. 1993. The Iliad: A Commentary. Vol. VI, Books 21-24. Cambridge: Cambridge University Press. Ricoeur, Paul. [1975] 1977. La metaphore vive. Paris: Seuil. Translated by Robert Czerny et al. as The Rule of Metaphor: Multi-disciplinary Studies of the Creation of Meaning in Language. Toronto: University of Toronto Press. .1978. "The Metaphorical Process as Cognition, Imagination, and Feeling." Critical Inquiry 5:143-59. Reprinted in Johnson 1981, 228-47. Rosenmeyer, Thomas G. 1955. "Gorgias, Aeschylus, and Apate." American Journal of Philology 76:225-60. Ross, W. D. 1949. Aristotle's Prior and Posterior Analytics. Revised text with introduction and commentary. Oxford: Clarendon Press. . [1951] 1953. Plato's Theory of Ideas. 2d ed. Oxford: Clarendon Press. Ruegg, Maria. 1979. "Metaphor and Metonymy: The Logic of Structuralist Rhetoric." Glyph 6:141-57. Sacks, Sheldon, ed. 1979. On Metaphor. Chicago: University of Chicago Press. Scott, W. C. 1974. The Oral Nature of the Homeric Simile. Leiden: E. J. Brill. Searle, J. 1979. "Metaphor." In Ortony 1993, 83-111. . 1980. Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Segal, Charles P. 1962. "Gorgias and the Psychology of the Logos." Harvard Studies in Classical Philology 66:99-155. 554 JOHN T KIRBY Shibles, Warren A. 1971a. Metaphor: An Annotated Bibliography and History. Whitewater, Wis.: Language Press. .1971b. An Analysis of Metaphor. The Hague: Mouton. , ed. 1972. Essays on Metaphor. Whitewater, Wis.: Language Press. Skulsky, Harold. 1986. "Metaphorese." Nous 20:351-69. Sorabji, Richard. 1972. Aristotle on Memory. Providence: Brown University Press. Stanford, W. B. 1936. Greek Metaphor. Oxford: Blackwell. Steiner, Deborah. 1986. The Crown of Song: Metaphor in Pindar. New York: Oxford University Press. Swiggers, Pierre. 1984. "Cognitive Aspects of Aristotle's Theory of Metaphor." Glotta 62:40-45. Tamba-Mecz, I., and P. Veyne. 1979. "Metaphora et comparaison selon Aristote." Revue des Etudes Grecques 92:77-98. Too, Y. L. 1995. The Rhetoric of Identity in Isocrates. Cambridge: Cambridge University Press. Turbayne, Colin Murray. 1962. The Myth of Metaphor. New Haven: Yale University Press. Turner, Mark. 1987. Death Is the Mother of Beauty: Mind, Metaphor, Criticism. Chicago: University of Chicago Press. Van Noppen, Jean Pierre, et al., eds. 1985. Metaphor: A Bibliography of Post1970 Publications. Amsterdam: J. Benjamins. Vico, Giambattista. [1725] 1968. La nuova scienza. 3d ed. 1744. The standard Italian edition (of the 1744 text) is that of Fausto Nicolini (Bari: Laterza, 1928). Translated by Thomas Goddard Bergin and Max Harold Fisch as The New Science of Giambattista Vico. Ithaca: Cornell University Press. Vlastos, Gregory. [1954] 1995. "The Third Man Argument in the Parmenides." Philosophical Review 63:319-49. Reprinted in his Studies in Greek Philosophy, vol. II, Socrates, Plato, and Their Tradition, edited by D. W. Graham, 166-90. Princeton: Princeton University Press. . [1965] 1981. "Degrees of Reality in Plato." In New Essays on Plato and Aristotle, edited by Renford Bambrough, 1-19. New York: Humanities Press. Reprinted in Gregory Vlastos, Platonic Studies, 58-75. 2d ed. [lst ed. 1973.] Princeton: Princeton University Press. Wheelwright, Philip. 1962. Metaphor and Reality. Bloomington: Indiana University Press. Zadeh, Lotfi. 1965. "Fuzzy Sets." Information and Control 8, no. 3 (June): 338-53. . 1975. "Fuzzy Logic and Approximate Reasoning." Synthese 30:407-28.