From the Tree to the Labyrinth

2

Metaphor as Knowledge

Aristotle’s Medieval (Mis)Fortunes

In Chapter 1 we observed that Aristotle’s major contribution to the theory of metaphor lay in the emphasis he placed on its cognitive value. Since we are accustomed to seeing the Middle Ages as the age of the rediscovery of Aristotle and indeed of his near-canonization, it should prove interesting to inquire whether the Middle Ages somehow picked up on and profited from this suggestion of his. Let us say from the outset that our investigation was sparked by the conviction that the answer is in the negative. What we must try to understand, then, is why there exists no medieval theory of metaphor as an instrument of knowledge, at least in the aforementioned Aristotelian sense. The answer, which we will attempt to document in what follows, is that not only did medieval authors gain access to the Poetics and the Rhetoric at a very late date, but they also became acquainted with these texts in translations that were, to say the least, somewhat misleading. We will see later (in Chapter 3) what the other sources of medieval reflection on metaphor were, and what other tools (such as, for example, the concept of analogia entis or “analogy of being”) they did attribute a cognitive function to.

2.1. The Latin Aristotle

It is no secret how protracted and tormented were the fortunes of the Aristoteles Latinus. In the sixth century Boethius had translated the entire Organon, but for centuries only one section of it, the so-called Logica Vetus—translations, in other words, of the Categories and the De interpretatione, accompanied by a version of the Isagoge by Porphyry and a number of treatises by Boethius on categorical and hypothetical syllogisms, on division and on topics—was in circulation, and that for the most part in a corrupt form.¹ Boethius had also translated the Prior Analytics, the Topics, and the Sophistical Refutations, but these works did not circulate at all until they were revised or retranslated, from Greek or Arabic,² along with the Posterior Analytics, in the twelfth century. True, this last-named work had also been translated by Boethius, but his version had been lost and remained practically unknown.³ With the twelfth century the Libri Naturales also make their appearance: the Physics, the De coelo et mundo, the De generatione et corruptione, the Meteorologica, the De anima, the Parva Naturalia are translated, first from Arabic then from Greek. The Metaphysics too appeared, first in partial form in a translatio vetustissima by James of Venice, while another extended portion appeared—translated from the Greek—in the same century (the so-called translatio media). Thomas Aquinas will own a complete version only when William of Moerbeke, completing his rendering, will make Book K available to him. Partial versions of the Greek text of the Libri Morales also go back to the twelfth century. In the mid-thirteenth century Robert Grosseteste translated the Nicomachean Ethics, later revised by Moerbeke, and it will be the 1260s before the latter will provide a complete version of the Politics. It is likewise in the thirteenth century that Michael Scotus made versions of the books on animals from the Arabic, while at a slightly later date Moerbeke will also translate them from Greek. A rendering of De motu animalium by yet another translator was known to Albertus Magnus.

Coming to the two texts that most concern us, we note that Moerbeke did not translate the Poetics until 1278—in other words, after Thomas’s death in 1274⁴—while Averroes’s Middle Commentary—composed in 1175—appears, translated by Hermann the German (Hermannus Alemannus), around 1256.

In the same year Hermann translated the Rhetoric from the Arabic. This translation is accompanied by the anonymous Translatio Vetus, from the Greek. And finally, around 1269 or 1270, there appears a version from the Greek by Moerbeke.

Thus, the Rhetoric and Poetics, when they finally appear in Latin, do so at an advanced date (and at a moment when a Logica Modernorum is on the rise—more interested in the Organon than in the remainder of Aristotle’s works). Thomas is the typical example of a thinker who was not influenced by any suggestion of Aristotle’s on this subject, and his theory of metaphor that “non supergreditur modum litteralis sensus” (“does not exceed the literal sense”) offers sufficient proof of this fact.⁵

2.2. The

Poetics:

Averroes’s Commentary and Hermann’s Translation

Averroes did not know Greek, he scarcely knew Syriac, and he was reading Aristotle in a tenth-century Arabic translation, derived in turn from a Syriac version.⁶ Both he and his sources have trouble rendering the various aspects of Greek poetry and dramaturgy to which Aristotle refers, and consequently try to adapt their examples to the Arabic literary tradition. Imagine what the Latin reader was able to make of Aristotle with the aid of Hermann the German’s Latin translation of an Arabic text, based in turn on an attempt to fathom the Syriac version of an unknown Greek original!

Furthermore, Hermann decided to translate only Averroes’s commentary, because, on account of the different metrical systems and the obscurity of the lexicon, he was not able to make complete sense of Aristotle’s work from the Arabic version, as he remarks in his “Proem.”⁷

Today we possess an English translation of Averroes’s Arabic text (Butterworth 1986) and, when we compare the two, we must admit that Hermann did not go wrong on the fundamental points. But he certainly adds to the confusion when he attempts to translate the poetic examples from Arabic; and he occasionally decides to replace them with Latin examples taken from the rhetorical tradition. When, for instance, Averroes proposes as an example of metaphor a fine line of Arabic poetry “the horses of youth and its trappings have been removed” (Butterworth 1986: 61), meaning that, in old age, love and war, activities associated with youth, are no longer practicable, Hermann substitutes the tired old chestnuts, pratum ridet (“the meadow smiles”) and litus aratur (“the strand is plowed”).⁸ In addition he gets badly tangled up in the rhetorical terminology. He translates what was intended as the term for metonymy as translatio and the term for metaphor as transumptio, but when Averroes cites both as species of the genus “substitution,” he proceeds to use the term concambium (p. 42). When Averroes says that poetic discourse is imitative, Hermann translates with the adjective imaginative, with quite drastic results for the comprehensibility of the text (ibid.).

Things go from bad to worse when Averroes, for “peripeteia” and “anagnorisis,” uses terms equivalent to “reversal” and “discovery”; the best Hermann can come up with is circulatio and directio, choices that are of little help in making the concepts clear (p. 53).

But the blame is not all Hermann’s. Butterworth is convinced that the Middle Commentary has been unjustly condemned and is more useful than previously thought, and he may be right as far as the comprehension of Averroes goes, but he is overindulgent with Averroes when it comes to a proper understanding of Aristotle.

Many readers will recall Borges’s 1947 short story entitled “Averroës’ Search” (in Borges 1998) in which the Argentinian writer imagines Abū al-Walīd Muḥammad ibn-Aḥmad ibn-Muḥammad ibn-Rushd (aka Averroes) as he endeavors to write a commentary on Aristotle’s Poetics. What bothers him is that he does not know the meaning of the words “tragedy” and “comedy,” which he had already come across nine years earlier when reading the Rhetoric. The problem is an obvious one, since these artistic forms were unknown in the Arabic tradition. The irony of Borges’s story stems from the fact that, while Averroes is struggling over the meaning of these obscure terms, beneath his windows a group of children is role-playing, impersonating a muezzin, a minaret, and a congregation, in other words, they are performing theater, but neither they nor Averroes are aware of the fact. Later on, somebody tells the philosopher about a strange ceremony he once witnessed in China, and from the description the reader is able to deduce that it was a theatrical performance—but the characters in the story are not so perceptive. At the end of this veritable comedy of errors, Averroes returns to his meditations on Aristotle and concludes: “Aristu [Aristotle] gives the name “tragedy” to panegyrics and the name “comedy” to satires and anathemas. There are many admirable tragedies and comedies in the Qur’àn and the mu’allaqat of the mosque” (Borges 1998: 241).

Readers tend to attribute this paradoxical situation to Borges’ imagination, but what he describes was precisely the quandary that beset Averroes.⁹ In the Middle Commentary, everything Aristotle has to say about tragedy is referred by Averroes to poetry in general, and more particularly to the poetic forms known as laudatio (praise) and vituperatio (blame). This epideictic poetry makes use of representations, but—though Averroes reminds us how men take pleasure in imitating things, not only in words but also through images, song, and dance—he speaks of them as exclusively verbal representations. Such representations are intended to instigate to virtuous actions, and their intent, then, is moralizing. Aristotle’s pragma thus becomes a virtuous and voluntary operation (Hermann: “operatio virtuosa, que habet potentiam universalem in rebus virtuosis, non potentiam particularem in unaquaque rerum virtuosarum” [“a virtuous deed that has a universal power with respect to virtuous matters, not a particular power with respect to one or another virtuous matter”] [p. 47]). Averroes understands that poetry tends toward the universal, and that its end is to arouse pity and fear in order to impress the minds of the audience. But for him these procedures too are calculated to render persuasive certain moral values, and this moralizing notion of poetry prevents Averroes from understanding Aristotle’s notion of the fundamentally cathartic (and not didactic) function of the tragic action.

The situation becomes still more “Borgesian” when Averroes finds himself obliged to comment on Poetica 1450a 7–14, where Aristotle lists the six components of tragedy. For Aristotle, as we know, these are mythos (plot), ethos (character), lexis (diction), dianoia (thought), opsis (spectacle), and melos (song or melody). Averroes interprets the first term to mean “mythic statements” (Hermann translates “sermo fabularis”), the second as “character” (Hermann: “consuetudines”), the third as “meter” (Hermann: “metrum seu pondus”), the fourth as “belief,” that is, as “the ability to represent what exists or does not exist in such and such a way” (Butterworth 1986: 78); for Hermann: “credulitas” or “potentia representandi rem sic esse aut sic non esse” (“the ability to represent the thing as it is or as it is not”). The sixth component is correctly interpreted as “melody” (tonus), but evidently Averroes is thinking of a poetic melody, not of the presence of musicians onstage. Things get more (or less) dramatic when we come to the fifth component, opsis. Averroes cannot envisage a spectacular representation of actions, and in translating opsis as nazar he has in mind something that leads to the “discovery of the correctness of a belief” (Butterworth 1986: 76),¹⁰ in other words a type of argumentation that demonstrates the correctness of the beliefs represented (for moral purposes). And all Hermann can do is to go along, so he translates “consideratio, scilicet argumentatio seu probatio rectitudinis credulitatis aut operationis non per sermonum persuasivum (hoc enim non pertinet huic arti neque est conveniens ei) sed per sermonem representativum” (“an examination or argument or proof of the correctness of a belief or the correctness of a deed, not by means of a persuasive statement [for that is not applicable to this art nor appropriate for it] but by means of a representative statement”).

Having failed in this way to grasp the meaning of spectacle, Averroes goes on to remark (Butterworth 1986: 79) that epideictic poetry “does not use the art of dissimulation and delivery the way rhetoric does,” and Hermann translates “non utitur carmen laudativum arte gesticulationis neque vultuum acceptione sicut utitur hiis retorica” (“laudatory verse does not take advantage of the art of gesticulation nor of putting on facial expressions the way rhetoric does”) (p. 49). On the other hand, Averroes had been led astray by 1450b 15 et seq., where Aristotle says that spectacle, however effective it may be, is not essential to the poetic art, since tragedy can also function without performance and without actors. In this way, Aristotle’s concession (tragedy can also be read) is transformed into the elimination of the opsis. Consequently, at least in the form in which it reaches its medieval readers, Aristotle’s text appears to exclude the only truly theatrical aspect of tragedy.

Finally, we have a total misunderstanding apropos of 1451b 1–14, where Aristotle opposes poetry to history, in the sense that poetry narrates possible actions, either probable or necessary, but always general, while the historian expounds real but particular events. Here Averroes radically misinterprets: he says that the poet speaks of existing and possible matters and that he often speaks about general things, while “the one who invents parables and stories” (in other words, those who for Aristotle were the historians) feign false things, inventing individuals who do not exist and finding names for them (Butterworth 1986: 83–84). Hermann translates “poete vere ponunt nomina rebus existentibus, et fortassis loquuntur in universalibus” (“the poets on the other hand use names for existing [viz. individual] things, and sometimes they also speak in general terms”) (p. 52) and, transforming the historian into a fictor (in other words, a narrator of fables), he says that he “fingit individua quae penitus non habent existentiam in re, et ponitur eis nomina” (“he invents individuals who do not exist at all in reality, and gives them names”).

Averroes seems sensitive to the thematics of metaphor, as he brings it up right away at the start of his commentary (Butterworth 1986: 60–61), whereas Aristotle himself has nothing to say about it and confines himself to discussing imitation. For Averroes, poetic compositions are imitative when they compare one thing to another, and he gives the example of cases in which one thing is described “as if” it were another (speaking of these “particles of comparison,” Hermann will use the term “sinkategoremata similitudinis” (“the syncategorematic terms of the comparison”) (p. 42);¹¹ but he also cites cases of “substitution,” a generic procedure of which metaphor and metonymy are subspecies. Apropos of metaphor Averroes speaks immediately of analogy, that is, of a four-term relationship. In this same context he makes an affirmation typical of Arabic philosophy, which will come to have a notable influence on Latin thought, namely, that poetics belongs to the art of logic.¹²

In another context, not found in Aristotle, Averroes, discussing sense-perceptible things represented by means of other equally sense-perceptible things, seems to be alluding to metaphors, since he speaks of the knowledge produced by the names of constellations like Cancer (in the sense of “crab”). He appears to be saying that these juxtapositions generate uncertainty (at least they are introduced by expressions of uncertainty) and therefore some kind of cognitive effort, while comparisons that do not generate uncertainty are less interesting (p. 97). Hermann translates: “Quedem earum sunt ut fiat representatio rerum sensibilium per res sensibiles quarum natura sit ut quasi in dubio ponant aspectorem, et estimare faciant eum presentes esse res ipsas” (“Among them is for the representation of sense-perceptible things to be made by means of sense-perceptible things, such that anyone who looks at them becomes uncertain and fancies that they are indeed those things”) (p. 59). Here we could be getting close to a cognitive notion of tropes. But a little earlier Averroes has declared that these imitative pictures must conform to commonly used formulas in a clear fashion, so as not to create difficulties. The doubt is resolved when we realize that he is commenting on 1454b 19–21, where ways of making the recognition or agnition more interesting are analyzed, and therefore the uncertainty is due to the recognizability of characteristic signs or tokens (Aristotle is talking about scars, necklaces, etc.). Perhaps it is because Averroes is not thinking of coups de théâtre that he treats the matter with some hesitancy (otherwise he would never have introduced the Cancer example), and Hermann follows him with the same hesitant confusion.

Metaphor also seems to crop up apropos of 1455a 4–6. Aristotle is concerned with agnition through syllogism, as when, in Aeschylus’s Libation Bearers (Choephoroi), Electra argues that someone identical to herself has arrived, but nobody is identical to her except Orestes. Averroes interprets this to mean that what is being spoken of is an individual who is like another individual because of a similar constitution or temperament (p. 104). Hermann is drawn by this discourse on similarity to speak of metaphorica assimilatio (“metaphorical comparison”) (p. 60), which is quite evidently a misreading.

We do come to metaphor apropos of 1457b et seq. A noun, as Aristotle says, can be “ordinary” or “rare” or “metaphorical” or “ornamental” (along with other categories less interesting from our point of view). Averroes (Butterworth 1986: 121–122) accepts this distinction, as does Hermann (p. 67), who defines metaphor as primarium, intromissum aliunde, transumptum, or facticium (“original, introduced from elsewhere, taken over from an extrinsic usage, or artificial”). Similarly observed is the Aristotelian distinction between metaphors from genus to species and vice versa, from species to species, or by analogy, and even the example of old age as the evening of life is preserved.

Averroes however (as well as his translator) follows the letter of Aristotle: it is certainly useful to use unusual words if one wishes to strike the reader’s imagination, but one must not exaggerate, so as not to fall into riddles. As for the passage in 1459a 8, in which Aristotle introduces the knowledge of the related concept (with the verb theorein), Averroes does not seem to grasp the suggestion and confines himself to saying that “when the similarity in the substitution is very strong, it makes both the imitation and the understanding more excellent” (Butterworth 1986: 134). Hermann translates “quando enim commutatio vehementis fuerit assimilationis, inducet bonitatem imaginationis et comprehensionem complectiorem rei representatae simul” (“when indeed the reciprocal opposition is that of a very strong similarity, it leads to both the good quality of the imagery and a more comprehensive understanding of the thing it represents”) (p. 71). All this is certainly much weaker than it was in Aristotle’s text.

Overall, it is difficult to say what effect Averroes’s commentary might have had on the imagination of the Latins, since what they were confronted with were metaphors taken from Arabic poetry inadequately translated by Hermann. They certainly must have sounded odd to the ears of the Latin reader, and they might therefore have suggested an invitation to be daring. And what can we say of the effect that might have been produced by metaphors such as “Iam sol inclinatur et nondum perfecisti, et subdivisus in horizonte est quasi oculus strabi vel lusci” (“Now the sun is declining but has not completely set, it appears split on the horizon like the eye of a squinter or a person with one eye”) or “Non est denigratus oculos antimonio pulvere, ut nigros habent oculos a natura” (“Someone who has blackened their eyes with kohl is not like someone who has black eyes by nature”) (pp. 59–60)?

In fact we have only to consult the few medieval commentaries devoted to Averroes’s treatise, prior at least to the use made of it by Giles of Rome (Egidius Romanus). These texts are reproduced by Dahan (1980: 193–239) and represent a series of glosses on the Translatio Hermanni, a Quaestio in Poetriam and the Expositio super Poetriam of Bartholomew of Bruges. They consist of fairly pedestrian summaries of Averroes’s text, which add nothing useful either to the comprehension of Aristotle or that of Averroes. At most, in the first glosses, where Hermann speaks of translatio and transumptio as two species of concambium, two examples (taken perhaps from Boethius’s De consolatione) are adduced, “sicut enim se habet liberalis ad pecuniam, sic mare ad aquas” (“just as a liberal person handles money, so does the sea its waters”) and “sicut mare arenis siccis aquas ministrat, sic liberalis egentibus pecuniam” (“just as the sea pours its waters on the dry sands, so liberal persons hand out money to those in need”), which appear to be two instances of transumptio.

2.3. The

Poetics:

William of Moerbeke’s Translation

Compared with the commentary of Averroes/Hermann, Moerbeke’s translation strikes us as considerably more faithful to Aristotle, though at times it too falls victim to misleading Greek manuscripts.¹³ When in 1457b 32 Aristotle says that a shield could be called a “a wineless wine bowl” (aoinon), Moerbeke reads oinou and translates “puta si scutum dicat ‘fyalam’ non Martis sed vini,” “as if you were to call the shield, not the cup of Mars but a wine cup”) (p. 27).

Faced with the riddle of the dry suction cup, and following the reading he found in his manuscript, he translates “virilem rubicundum ut est ignitum super virum adherentem” (“a manly red like something fiery sticking to a man”) (p. 28). But he translates “seminans deo conditam flammam” (“sowing the god-created flame”), as well as other citations, correctly.

Tragodia and komodia are translated correctly (albeit with a simple calque). But let us not forget how obscure these terms could appear to a man of the Middle Ages: according to William of Saint-Thierry (Commentarius in Canticum canticorum, PL 180), a comedy is a story that, though it may contain elegiac passages that speak of the pains of love, ends happily; for Honorius of Autun (De animae exilio et patria, PL 172), tragedies are poems that deal with war, such as Lucan’s Pharsalia, while comedies, like the works of Terence, sing of weddings. Dante too refers to his work as a Commedia not because it is a theatrical work but because it has a very happy ending. Hugh of Saint Victor (Didascalicon II, 27, PL 176) says that the art of performance gets the name of “theatrical” art from the word “theater,” a place where the ancient peoples gathered for amusement, and in theaters dramatic events were recited aloud, with readings of poems or representations involving actors and masks. In the Poetria of John of Garland (Johannes de Garlandia), we find a classification of literary genres in which tragedy is defined as “carmen quod incipit a gaudio et terminat in luctu” (“a poem that begins in rejoicing and ends in lamentation”), while comedy is “carmen jocosum incipiens a tristitia et terminans in gaudium” (“a light-hearted poem beginning in sadness and ending in rejoicing”) (cf. De Bruyne, Études II, iii, 3). One of the few texts in which an idea of classical tragedy can be identified (based, however, on hearsay) is the Ars versificatoria of Matthew of Vendôme (II, 5, in Faral 1924), where among the arts tragedy is cited “inter ceteras clamitans boatu” (“shouting various loud cries in the midst of the group”), which (citing Horace, Ars Poet. 97) “projicit ampullas et sexquipedalia verba” (“spews forth bombast and sesquipedalian words”) and, continues Matthew, “pedibus innitens coturnatis, rigida superficie, minaci supercilio, assuetae ferocitatis multifarium intonat conjecturam” (“relying on buskin feet, an inflexible appearance, and a menacing brow, thunders forth a multitude of warnings, all with her customary ferocity”). Not much to go on as a clue to Aristotle’s concept of tragedy, but enough to recognize what ancient theatrical actions were like, seeing that the theater the Middle Ages had in mind evoked the antics of minstrels and histriones, along with the sacred mystery play.

Accordingly, in Moerbeke’s translation (pp. 9–10), mimesis is rendered as imitatio, pity and fear with misericordia and timor, pathos with passio; the six parts of tragedy become fabula, mores, locutio, ratiocinatio, visus, and melodie; it is understood that opsis has to do with the mimic action of the hypocrita or actor; peripetie and anagnorisees (idest recognitiones) are mentioned; and the distinction between the poet and the historian is clear. The oppositions between a clear and a pedestrian style are faithfully presented, though glotta is translated as lingua, making the nature of the barbarism somewhat less than transparent. Moerbeke translates 1457b 6 et seq., where metaphor is defined, in an acceptable manner.

In the crucial 1459a, 8, where Aristotle says that “to use metaphor well implies an ability to see the likenesses in things,” and he uses in this context the verb theorein, Moerbeke translates “nam bene metaphorizare est simile considerare” (“for to coin good metaphors is to consider likeness”) (p. 29). Perhaps the verb considerare has a weaker connotation than the Greek word, but it points in any case to the universe of knowledge.

To sum up, the Latin reader could have acquired a reasonable idea of Aristotle’s text from Moerbeke, but with nothing that underscored with particular energy the cognitive aspect of metaphor’s implications.

2.4. Aristotle’s

Rhetoric:

Hermann the German’s Translation

In section 2.1 we recalled that three translations of the Rhetoric had appeared: one from the Arabic by Hermann the German, an anonymous Translatio Vetus, and, between 1269 and 1270, the version of William of Moerbeke, from the Greek. For a long time the received wisdom concerning Hermann’s Rhetoric was imprecise. The title, Averroes in Rhetoricam, led some scholars to conclude that it was a translation of Averroes’s Middle Commentary. Then, because of the existence of other manuscripts that bore the title Didascalia in Rhetoricam Aristotelis ex glosa Alpharabi (whereas al-Farabi’s commentary was incomplete from the start), it was believed that Hermann’s text was based solely on Arabic sources. Only quite recently (Bogges 1971) was it determined that Hermann had translated the text of Aristotle from the Arabic, inserting passages from Averroes’s commentary and from Avicenna’s Shifa when the manuscripts at his disposal were lacunary (but invariably making the insertion explicit). In his translation of al-Farabi’s glosses, Hermann explicitly claims to have translated Aristotle’s Rhetoric from Arabic to Latin, and he repeats the claim in the prologue to his translation of the Rhetoric.

We shall consider later the problems that this extremely arduous translation, of whose insufficiencies the translator himself was fully aware, posed for the Latin reader. Furthermore, we know of only two complete manuscripts and one fragmentary one, which leads us us conclude that it had a very limited circulation.¹⁴

An example of the translator’s embarrassment is provided by the notion of ta asteia. At the end of chapter 10 Hermann decides to skip portions of the text of Aristotle that he is unable to translate, and he goes on to comment: “Plura talia exempla ad idem facientia, quia greca sapiebant sententiam non multum usitatam latinis, dimissa sunt, et subsequitur quasi conclusio auctoris” (“Many like examples of the same import have been omitted for they smacked of the Greek idiom not much used by the Latins, and the author’s conclusions as it were follow immediately after”). To say nothing of the fact that in one manuscript (Toledo, cf. Marmo 1992: 32 n. 8) in chapter 11 we find: “Ideoque pulchre dicit Astisius in suis transsumptionibus quasi ante oculos statuende ea que transumendo loquitur” (“And this is why Astisius expresses himself so well in his transumptions that almost place what he is talking about by transference before your very eyes”). Where the form Astisius suggests that the Arabic original had interpreted asteia as a proper name, and that Hermann had gone along with this interpretation.

2.5. The

Rhetoric: Translatio Vetus

(V) and William of Moerbeke’s Translation (M)

With reference to the key points of Aristotle’s text listed in section 2.1, let us now examine the solutions provided by V and M.¹⁵

1404b 3. That what is “foreign” is delectable and thaumaston (i.e., exciting wonder) is clear enough both in V (“mirabiles enim absentium, delectabile autem mirabile est” [“for those who are admirable are different, but what is admirable is delightful”]) and in M (“admiratores enim advenarum sunt, delectabile autem quod mirabile est” [“for those who admire are strangers, but what is admirable is delightful”]).

1405a 9. V says that “manifestum et delectabile et externum habet maxime metaphora, et assumere non est ipsam ad alio” (“metaphor has most especially evidence, delight and strangeness, and it cannot be received from someone else”). M translates “evidentiam et delectationem et extraneitatem habet maxime metaphora, et accipere ipsam non est ab alio” (“metaphor has most especially distinctness, delight and strangeness, and it cannot be taken from someone else”). Both let it be understood that good metaphors are not made by merely imitating those already codified.

1405a 9. Verbs like phainesthai and skopein are rendered in V with videri and intueri and by M with apparire and intendere. They are in other words verba cognoscendi.

1405a 10. V does not get the quip about pirates calling themselves purveyors and translates “et latrones se ipsos depredatores vocant” (“and pirates call themselves predators”). M on the other hand speaks appropriately of acquisitores.

1405b 13. The idea that metaphor puts matters before our eyes is properly understood (“in faciendo rem coram oculis” in V and “in faciendo rem pre oculis” in M). Similarly, all subsequent translations of the same expression are correct.

1406b 4. The translators are embarrassed, and not without reason, by the distinction between metaphora and eikon. V first translates eikon as conveniens, producing the obscure expression “est autem et conveniens metaphora” (“moreover a metaphor is also befitting”), but right afterward he translates the same term with ymagines. M translates it as assimilatio. In both translations, however, the context makes it clear that what is involved is a simile (for both, Achilles “ut leone fremit” or “fremuit” [“roars like a lion”]).

1410b 10 et seq. We come now to the definitions of ta asteia (“witty and popular sayings”). V renders the term with solatiosa and M conserves asteia. Especially in the latter case, we can only suppose that the medieval reader had no idea what they were talking about (see, in Marmo 1992, the misunderstandings that ensue in Giles of Rome’s commentary). One might have expected the concept to be clarified by the plentiful examples supplied by Aristotle, but unfortunately the translation of these pithy sayings is unsatisfactory. Many of Aristotle’s examples are completely skipped. In V the triremes like “parti-colored mills” become “milonas curvas,” and in M “molares varios.” Sisyphus’s stone that rolls ruthlessly down to the plain becomes in V “lapis … inverecundus ad eum qui est inverecundus” (“a stone … shameless to him who is shameless”) and in M “lapis … qui inverecundus ad facile verecundabilem.” (“a stone … that is shameless to someone who is easily contemptible”). The spear-point that speeds eagerly through the warrior’s breast is not translated in V, while in M it appears as the inexplicable “gibbosa falerizantia.” In V the metaphor of stubble for old age becomes the incomprehensible “quando enim dicit senectutem bonam, facit doctrinam et cognitionem propter genus” (“for when he says that old age is good, he teaches and imparts knowledge to us through the genus”), while M translates more appropriately “quando enim dixit senectutem calamum fecit disciplinam et notitiam per genus” (“when he called old age a stalk, he taught and delivered a notion through its genus”). In 1412a 5, Archytas’s metaphor on the similarity between an arbitrator and an altar (both a refuge for someone who has suffered an injustice) in V becomes “sicut Archites dixit idem esse propter hanc et altare” (“just as Archytas said that there was no difference because of this”), perhaps because his manuscript, instead of “diaiteten” [= arbitrator], read “dia tauten”); M on the other hand is not guilty of the same error. We may well wonder how much intellectual stimulation a medieval reader might have felt in the face of such obscure pseudo-inventions that often come across as insipid or meaningless.

Curiously, in the same passage, both translators accurately render the conceptual aspect. In V good enthymemes “faciunt nobis doctrinam expeditam” (“they teach us expeditiously”), and in this connection mention is made of “cognitio” (which is Aristotle’s gnosis or “knowing”). M says that good enthymemes “faciunt nos addiscere celeriter” (“make us learn quickly”) and that “cum hoc quod dicuntur notitia fit” (“which are understood the moment they are stated”). Similarly, it is clear, though elliptically expressed, that metaphor must make us see the thing in action and that, like philosophy, it must make us “inspicere” (a good translation of theorein, “to contemplate or consider”) a resemblance “a propriis et non manifestis” (“proper to the object, yet not obvious”) (V), while M speaks less forcefully, but with clarity, of a witty saying that makes us “bene considerare similitudinem in multibus distantibus” (“consider carefully similarity in many disparate things”). When V finds himself faced (in 1412a 17) with the term “epiphaneia” he boldly transliterates it as “epyphania” (while M does not grasp the meaning of apparition and revelation and says “in superficie,” [“on the surface”]).

Correctly rendered is the passage in 1412a, in which Aristotle says that, when confronted with a witty juxtaposition, the surprised reader recognizes that he had not seen things as they were and had been mistaken (even though, immediately following, V, after attempting to translate Stesichorus’s apophthegm of the grasshoppers that will sing to themselves from the ground, skips a short passage on riddles and translates the notion of “novel expressions” with “inania”). M on the other hand translates the passage on riddles (which are able to say new things [“nova dicere”]) and gets across the idea of the unexpected word (“inopinatum”) and the paradox it produces.

To sum up, the two versions might have given some inkling of Aristotle’s position, but it is doubtful whether the meaning of the technical terms was immediately evident, and the translations of the examples were certainly of no help in understanding the definitions any better.

2.6. The Medieval Misfortunes of the

Poetics

and the

Rhetoric

The scant attention the Middle Ages paid to these two translations can be explained in a number of ways. In the first place, up until the twelfth century, rhetoric had belonged to the trivium, but poetics was not included. Thus, observes Dahan (1980), poetics is ignored by Alan of Lille (in his Anticlaudianus), Honorius of Autun, Hugh of Saint Victor, Robert Grosseteste (in his De artibus liberalibus), John of Dacia (in his De divisione scientiae), and many others.

Around the twelfth century, another division of the sciences becomes prevalent, one Stoic in origin, according to which philosophy is subdivided into logic, ethics, and physics, and at this point both poetics and rhetoric were considered part of logic. The idea is already present in Augustine, but see Isidore of Seville’s definition in Etymologiae II, 24, 3: “Philosophiae species tripartita est: una naturalis, quae graece physica appellatur …; altera moralis, quae graece ethica dicitur …; tertia rationalis, quae graece vocabulo logica appellatur” (“There are three kinds of philosophy: one natural [naturalis], which in Greek is ‘physics’ [physica] …; a second moral [moralis], which is called ‘ethics’ [ethica] in Greek …; a third rational [rationalis], which is named with the Greek term ‘logic’ [logica].”)¹⁶

Later in the twelfth century, through the agency of Gundisalvo, the Arabic classification, in which poetics and rhetoric are seen as an integral part of Aristotle’s Organon (see, for instance, Avicenna’s Shifa and the De scientiis of al-Farabi), becomes established in the West. It was in fact as an aid to students of logic that Hermann presented his translation: “suscipiant igitur, si placet, et huius editionis Poetriae translationem viri studiosi, et gaudeant se cum hac adeptos logici negotii Aristotilis complementum (“May then learned men, should it be deemed desirable, take up also the translation of this edition of the Poetria and rejoice to achieve with it a completion of the logical works of Aristotle”).

Though he did not know Aristotle’s Rhetoric, Albertus Magnus considers rhetoric a logical discipline (see, for instance, Liber de praedicabilibus I, 4); and in the Liber Primis Posteriorum Analyticorum he includes poetics under logic (cf. Dahan and Rosier-Catach 1998: 77, as well as Marmo 1990: 159–163).¹⁷

As parts of logic, poetics and rhetoric were understood to be persuasive discourses that could be used for political and moral ends, and it is in fact Gundisalvo who defines poetics as forming part of civil science, which is in its turn part of eloquence, whose purpose is to delight and instruct both in science and proper behavior.

The thinker who crystallizes the Arabic position (rhetoric and poetics as part of logic, and their moral and civic orientation) is Roger Bacon (cf. Rosier-Catach 1998). Bacon, inspired by Gerard of Cremona’s translation of al-Farabi’s De scientiis, is intent, in his Moralis Philosophia (the seventh part of his Opus Majus), on establishing a method for convincing the infidel of the superiority of Christianity, and he finds it in rhetorical and poetic discourse. He is seeking a “sermo potens ad inclinandum mentem” (“speech with the power to persuade the mind”); and language (he affirms in Opus Majus III) is more effective than any war. If dialectical and demonstrative arguments could move the speculative mind, poetics and rhetoric can move the practical intellect (Opus Majus III).

Poetic argument has nothing to do with truth or falsehood. Poetics is the study of ways of moving the listener emotionally by means of a magniloquent style, and the greatest example of poetic discourse is provided by the Holy Scriptures. In the Moralis Philosophia imitation (similitudo) is seen as the way of comparing, for instance, virtue to light and sin to things that are hideous.

It is Bacon again, in his Communia Matematica, who will state that poetic argument uses fine discourses so that the soul may be overcome by the love of virtue and learn to hate vice. To this end ornaments such as meter and rhythm can be useful, as is the case in the texts of Scripture.¹⁸

Independently of Bacon, the idea that poetics and rhetoric are part of logic and are concerned with moral and civil knowledge made more and more headway among those who approached the first translations of Aristotle. It is understandable, then, that the thinkers who debated such problems were not especially interested in the semiotics of elocutio, and hence in the technical study of metaphors, but focused their attention more on methods of argumentation.

Thomas demonstrates his familiarity with these translations (except, of course, Moerbeke’s translation of the Poetics), but, in his commentary on Posterior Analytics I, he sees logic as judicative (Prior and Posterior Analytics), sophistic (Sophistical Refutations) and inventive (Topics, Rhetoric, and Poetics). Hence, “poetae est inducere ad aliquod virtuosum per aliquam decentem representationem” (“the poet’s task is to lead us to something virtuous by some excellent description”).

Buridan will allude to the fact that poetics, while it doesn’t put things clearly like rhetoric does, still has the same educative intentions in mind, “scientiam delectabiler obscurare nititur, per verborum transumptionem,” “it endeavors to obscure knowledge delightfully by metalepsis” (cf. Dahan 1998: 186).

But Bacon is the one who points to another reason for the scant currency of these translations and interpretations of Aristotle: they were badly translated and hard to fathom. Bacon says that he knew Hermann personally. In Moralis philosophia VI he claims that Hermann confided in him (“dixit mihi”) that he was insufficiently versed in logic to translate the Rhetoric well, and for the same reasons had not dared to translate the Poetics, confining himself to translating Averroes’s commentary. So, Bacon observed, we can never really know what Aristotle thought about poetics, we can only “get a whiff” of it, not savor it, as is the case with wine that has been poured too many times from one container to another.¹⁹

Bacon claims in Opus Majus I that the moderns neglect two books of logic, one of them translated with a commentary by al-Farabi, the other an exposition of Aristotle by Averroes, translated without reproducing the philosopher’s original text.²⁰ In Opus Majus III he points out once again that there are few Latin translations of Aristotle’s logic and Averroes’s commentaries, and that the few versions extant are not read.²¹ Again, in the Moralis Philosophia (V, 255), he cites Averroes’s commentary on the Poetics as the only available source for Aristotle’s text, but he recognizes that it too is known to only a few.²² In Opus Majus III he complains about the translations of Aristotle’s works, executed “cum defectu translationis et squalore,” (“crudely and with defective translation”)—with the result that nothing can be understood–and he remarks what a loss this has been for the culture of his time.

From these texts we may deduce that Bacon was not yet acquainted with Moerbeke’s translation of the Rhetoric, which would not appear in fact until later and which he might perhaps have treated with greater indulgence. But Bacon’s strictures, which are extremely severe on almost all the translators without proposing new criteria for a correct translation (cf. Lemay 1997), indicate to us that, though they may have enjoyed some limited currency, Averroes’s texts were familiar to few, and viewed with suspicion by those who knew them. It appears that the translations were not readily available in university circles, though in any case Bacon operated outside of those circles. Thomas cites a brief excerpt from Hermann’s translation of the Rhetoric in the Contra Gentiles; and later, in Summa Theologiae I–II, 29, 6, he will quote it once more, but this time in Moerbeke’s version. Moerbeke’s translation will in fact enjoy greater popularity, it will circulate in numerous manuscripts, and it will form the basis for the commentary on the Rhetoric composed by Giles of Rome between 1272 and 1274.²³

Precisely because he has available a less improbable translation than those that came earlier, Giles’s theory of metaphor strikes us as unquestionably more mature. In both Marmo (1998) and chapter 7 of Eco and Marmo (2005) (written entirely by Marmo), we see how Giles worked out a strategy of critical collation among the different versions.

But at this point we are nearing the end of the thirteenth century. Giles’s commentary will be followed by those of John of Jandun and Buridan,²⁴ too late, we might say, for Aristotle’s theory of metaphor to have any decisive influence on scholastic thought. As will be seen in Chapter 3, medieval metaphorology will have other founding texts and other outcomes.

What we have attempted to demonstrate here is how the absence of a cognitive theory of metaphor in the golden age of scholasticism was largely due to the inadequacy of the existing translations.

Paper delivered at the Scuola Superiore di Studi Umanistici of the University of Bologna in March 2001 as part of a series of lectures on the fortunes of Aristotle’s theory of metaphor. Revised and expanded, especially as regards the contribution of Giles of Rome, and published as a collaborative effort by myself and Costantino Marmo with the title “La teoria aristotelica della metafora nel Medioevo,” in Lorusso (2005). The present version is a reelaboration of my 2001 paper, but takes into account observations and clarifications made by Marmo.

1. Boethius’s contributions were the Introductio in syllogismos categoricos, De categoricis syllogismis, De hypotheticis syllogismis, De divisione and De differentiis topicis (PL 64). Only in the thirteenth century would William of Moerbeke bring forth new translations of the Aristotelian treatises, in addition to a translation of Ammonius Hermiae’s Greek commentary on the De interpretatione.

2. Boethius’s translation of the Sophistical Refutations, for example, was revised by James of Venice in the twelfth century and retranslated in the thirteenth century by William of Moerbeke. Its circulation, however, was modest.

3. It is only with the translation from the Greek by James of Venice (twelfth century), its revision by William of Moerbeke (thirteenth century), together with the translation from the Arabic by Gerard of Cremona (end of the twelfth century) and the commentary of Robert Grosseteste (ca. 1230), that this text will enter medieval culture, becoming fully integrated into the so-called Logica Nova.

4. Furthermore, only two manuscripts are known from this period. It is not until Giorgio Valla’s 1495 Latin translation from the Greek that the Poetics will enter the world of the Humanists. Valla was unaware of Moerbeke’s translation.

5. The quote from Thomas is from Quodlibet VII, q. 6 a. 3 ad 2 (http://www.corpusthomisticum.org/q07.html#68283). We will have more to say on this topic in Chapter 3.

6. The Middle Commentary on the Poetics appeared in English in Butterworth (1986). The text of Hermann can be found under the title Averrois expositio seu Poetria Ibn Rosdin in Minio-Paluello (1968). Citations from either work are to the pages of these modern editions.

7. “Postquam, cum non modico labore consummaveram translationem Rhetorice Aristotelis, volens mittere ad eius Poetriam, tantam inveni difficultatem propter disconvenientiam modi metrificandi in greco cum modo metrificandi in arabico, et propter vocabulorum obscuritatem, et plures alias causas, quod non sum confisus me posse sane et integre illius operis translationis studiis tradere latinorum” (“After having completed, with no small labor, my translation of Aristotle’s Rhetoric, and wishing to dedicate myself to his Poetria, I found myself confronted with enormous difficulties, because of the difference between Greek and Arabic metrical scansion, the obscurity of the terminology, and for a number of other reasons, so that I am not sure I can really offer the translation of that work to the schools of the Latins without misrepresentation”) (p. 41).

8. P. 42. On the circulation of these canonical examples, see Chapter 3.

9. Borges probably got his information from Marcelino Menendez y Pelayo’s meticulous summary of the two commentaries in volume 1 of his Historia de las ideas estéticas en España (1883).

10. For Averroes, see Butterworth (1986: 75–79); for Hermann, see Minio-Paluello (1968: 48–49).

11. [Translator’s note: The Shorter Oxford English Dictionary defines “syncategorematic” as follows: “Of a word: having no meaning by itself, but only in conjunction with one or more other words or concepts.”]

12. For the fortunes of this thesis, see Marmo 1990.

13. Moerbeke’s translation of the Poetics also appears in Minio-Paluello (1968). The page references in parenthesis are to the pages of this edition.

14. For all this information and the passages cited, see Bogges (1971).

15. Textual references and page numbers are to Schneider (1978). References to the original are to Aristotle (1926).

16. See Barney, Lewis, Beach, and Berghof (2006, p. 79).

17. To find classifications that include poetics in an autonomous position, we must wait for Giles of Rome, though Dahan (1980: 178) already finds anticipations of this position in William of Conches and Richard of Saint Victor.

18. “Hoc argumentum utitur sermonibus pulchris et in fine decoris, ut rapiatur animus subito in amorem virtutis et felicitatis, et in odium vicii et pene perpetue que ei respondent. Et ideo sermones poetici qui sunt completi et pulchritudine et efficacia movendi animum debent esse ornati omni vetustate loquendi prosaice at astricti omni lege metri et ritmi, sicut Scriptura Sacra … ut decore et suavitate sermonis animus subito et fortiter moveatur” (“This is argument makes use of beautiful and decorous speech so that the soul will be immediately raised to the love of virtue and happiness and to the hatred of vice, and will scarcely ever be attracted to it. And so poetic speeches that are complete, beautiful and efficacious in moving the soul ought to be dressed out in all proper forms of prosaic speech, and abide by all the laws of meter and rhythm, just like Sacred Scripture … so that by the decor and sweetness of the language, the soul may be strongly and immediately moved”), (cf. Hackett 1997: 136, n. 6).

19. “Studiosi homines possunt a longe olfacere eius sentenciam, non gustare: vinum enim, quod de tercio vase transfusum est, virtutem non retinet in vigore” (“Learned men can get a distant whiff of his meaning, but not taste it. For a wine that has been poured into three successive containers does not keep its virtue in all its strength”) (Moralis Philosophia VI, 267, cited in Rosier-Catach (1998: 95), to whom we are also indebted for the references that follow).

20. “Moderni … duos libros logicae meliores negligunt, quorum unus translatus est cum Commentum Alpharabii super librum illum, et alterius expositio per Averroem facta sine textu Aristotelis est traslata” (“The Moderns neglect the two best books on logic, one of which has been translated with the commentary of al-Farabi, while the commentary on the other composed by Averroes has been translated without Aristotle’s text”).

21. “Quoniam autem libri Logica Aristotelis de his modis, et commentarii Avicennae, deficiuntur apud Latinos, et paucae quae translata sunt, in usu non habentur nec leguntur, ideo non est facile esprimere quod oporteat in hac parte” (“On the other hand, since the books of Aristotle’s Logic regarding these methods, together with Avicenna’s commentaries, are not available to the Latins, and the little that has been translated is not used or read, it is no easy matter to express what needs to be expressed in this part”).

22. “Quoniam vero non habemus in latino librum Aristotelis de hoc argomento ideo vulgus ignorat modum conponendi ipsum; sed tamen illi, qui diligentes sunt, possunt multum de hoc argumento sentire per Commentarium Averrois et [forse in] librum Aristotilis, qui habetur in lingua latina, licet non sit in usu multitudinis” (“Since we do not have, in Latin, Aristotle’s book on this subject [Bacon is alluding to the Poetics], most people therefore do not know the way it was composed; those, however, who are studious can learn much on this topic from Averroes’s Commentary and the book by Aristotle that we do possess in the Latin language, though not many people make use of it”).

23. Interest in the two Aristotelian texts apparently reawakens in the fourteenth century, when citations from Hermann’s translation appear in several florilegia; see Bogges (1970).

24. Still unpublished; see Marmo (1992).