The Revolutions of Wisdom "d0e10410"

Aristotle's Critique of Metaphor

In Aristotle, the shift in emphasis towards a more negative evaluation—at least in certain contexts—is marked. First, he frequently censures the metaphors and images used by his predecessors. Thus Empedocles' notion of the salt sea as the sweat of the earth is "adequate, perhaps, for poetic purposes" but "inadequate for understand-

[35] See especially Goldschmidt 1947a.

[36] When the use of paradigms is itself illustrated and explained by a paradigm in the Politicus , 277d ff., the Eleatic Stranger takes the case of children learning to read. Once they have learnt to recognise letters in short and easy syllables, they can be taught to recognise them also in more complex combinations, by juxtaposing the known and the unknown and pointing to the same likeness and nature in both cases. In both the Sophist and the Politicus , paradigms serve to provide practice in method, when the method of division, to be used on the sophist and on the kingly art, is first exemplified with the easier cases of angling (Sph. 218e ff.; cf. 218b–d) and of weaving (Plt. 279a ff.; cf. 286a–b).

[37] In the illustration of children being taught to read, the instructor himself already knows the letters. But in the problems investigated in the Sophist and Politicus , the hunt for the sophist and for the definition of the kingly art are represented as searches , where neither the leader, the Eleatic Stranger, nor his interlocutors have the answers when they set out. Moreover, in both dialogues the paradigm that is chosen as an illustration is particularly relevant to the substantive subject under investigation. In the Sophist , when the activity of dividing angling illustrates the method, angling turns out to be like sophistry (Sph. 221d8ff.), and in the Politicus weaving is chosen at the outset (Plt. 279a7ff.) for its similarity to politics (cf. Plt. 308d ff.). In these examples the activity in the case of the paradigm is an instance, not merely a likeness, of the activity also exemplified in the larger case.

― 184 ―

ing the nature of the thing."^[38] Other images of Empedocles and other pre-Socratic philosophers are criticised on the grounds that they are based on superficial similarities—or on none, that the illustrations are obscure, or crude, or in need of qualification.^[39] Thus milk, he insists at one point, is formed by a process of concoction, not putrefaction, so Empedocles was wrong, or he used a bad metaphor, when he spoke of it as "whitish pus."^[40] Similarly, Plato's own theory of Forms as a whole is dismissed on the grounds that to say that the Forms are "models and that other things share in them is to speak nonsense and to use poetic metaphors"^[41] —where again we may remark that poetic is used as a term of censure.

Aristotle is especially uncompromising in his criticisms of the use of

in the context of his formal logic and theory of demonstration,

, for him, being defined as the transfer of a term appropriate to one domain to another.^[42] In the Posterior Analytics he condemns them as a whole, especially their use in definitions. "If one

[39] See, for example, Top . 127a17ff., GA 747a34ff., 752b25ff., and cf. De sensu (Sens.) 437b9ff., PA 652b7ff. Cf. Bremer 1980.

[41] Metaph. 991a20ff., 1079b24ff. Other comparisons used by Plato are criticised at Pol. 1264b4ff., 1265b18ff., for example.

― 185 ―

should not argue in metaphors, it is clear that one should not use metaphors or metaphorical expressions in giving definitions."^[43] In the Topics , too, he repeats the criticism of definitions that contain metaphors on the grounds that "every metaphorical expression is obscure."^[44]

There is, to be sure, another side to the picture. Elsewhere when he discusses style, especially,^[45] he approves of certain types of metaphor, particularly those that express a proportion, for these, he says, are vivid, witty, and clear^[46] (by which he does not mean to deny that from another point of view they are still "obscure"). He praises in the poet the ability to deploy metaphor and to discern resemblances; the latter is a skill that the philosopher too will need to exhibit.^[47] In the Topics , moreover, the "investigation of likeness" is said to be a useful means by which to become well-supplied with arguments and even also, in certain contexts, for rendering definitions, that is, in securing the genera for them.^[48] In the Sophistici Elenchi he is not above recommending

[45] See Rh. 1405a8ff., 1407a14ff., 1410b36–1411b23.

[46] For example at Rh. 1405a8ff., 1410b13ff.

[48] Top. 105a21ff., 108a7ff., b7ff.

― 186 ―

metaphor as a way of making an account difficult to refute without, he hopes, being found out—a backhanded recommendation, to be sure.^[49]

Nevertheless Aristotle's appeal to the contrast between the "proper" and the "alien" or transferred uses of terms runs counter to some modern preoccupations. It leaves little room for any concession to an interaction view of metaphor or for the idea that a metaphor may create a similarity as much as show one.^[50] Indeed, the assumption in the background is that the comparisons implicit in proportional metaphora can be spelt out fully in literal terms without loss: his notion of metaphora already presupposes that there are two distinct, independently identifiable fields between which a transfer has taken place and in only one of which the term transferred is "proper." Finally, his analysis of analogical argument in the form of the paradigm concentrates on its shortcomings judged from the standpoint of the theory of the syllogism.^[51] It proceeds from particular case to particular case, whereas for the argument to be valid it must proceed first by a complete induction to a universal rule, which is then applied deductively to the particular case in question in the conclusion.^[52]

The concessions that Aristotle makes, from time to time, to the usefulness of various modes of reasoning based on likenesses do not do much to mitigate an attitude that is otherwise strongly critical. But these formal condemnations provoke a series of questions. First, as regards Aristotle himself, how far does his actual practice tally with the implications of those formal condemnations, and insofar as it does not,

[49] SE 176b20ff., 24f. In other texts too Aristotle describes how the dialectician may exploit similarities to deceive an opponent, especially at Top. 156b10ff., SE 174a37ff., and he reverts with great frequency throughout the Topics to the topic of the inspection of similarities in general, e.g., 114b25ff., 124a15ff., 136b33ff., 138a30ff.

[51] See especially APr. 2.24.68b38ff. Paradigms are also discussed, from the point of view of their use in rhetoric, at Rh. 2.20.1393a22–1394a18, and cf. 1357b25ff., 1368a29ff.

[52] See APr. 69a13ff.; cf. Rh. 1357b26ff.

― 187 ―

how can we explain the driving force behind his critique? Is it really the case that he entirely purges his speculative thought of metaphor, heeding his own warning that "every metaphorical expression is obscure"? Does he quite fail to recognise the role of models in both philosophy and science? What kind of impoverished science would it be that did without theoretical terms drawing, implicitly or explicitly, on models and even metaphorai? But if and when he in fact uses such terms, how could he square them with his theory of definition and demonstration? The questions may be raised in relation to Aristotle himself in the first place, particularly with regard to his contributions to the inquiry concerning nature. But this in turn will lead us to broach similar issues in relation to the use of theoretical and technical terms, and the role of metaphors and models, in Greek science more generally.

The Art of Nature

We may concentrate first on the inquiry into nature in Aristotle, and the term "nature" itself offers an excellent starting-point. As is well known, nature is defined in terms of an innate capacity for movement, and the power at work in what has that capacity (especially living creatures) is captured in a wealth of images, comparisons, and analogies. When Aristotle describes the growth of the embryo or the structure of the bones or the blood-vessels in the body or the way in which the blood is used as the material for the other parts in the body, he compares nature to a modeller in clay,^[53] to a painter (sketching a figure in outline and then applying the colours),^[54] to a housebuilder (laying out the stones along the foundations of the house),^[55] and to a good housekeeper (not wasting material).^[56] Several more images are borrowed directly from, or at least echo, those that Plato had used when describing

[53] PA 654b29ff.; cf. GA 730b27ff. and other comparisons incorporating the idea of the construction of a framework for a model, e.g., HA 515a34ff., GA 743a1ff.

[54] GA 743b20ff.; cf. 764b30f.

[55] PA 668a16ff.

[56] GA 744b16ff., PA 675b20ff.

― 188 ―

the work of the Craftsman in the Timaeus . Aristotle too compares the blood-vascular system to a system of irrigation channels,^[57] and he too compares the crisscrossing of the blood-vessels to a wickerwork structure.^[58] More simply, nature is repeatedly described as creating,

, devising,

, and adorning,

, living creatures or their parts,^[59] and most frequently of all, of course, her purposeful activity is expressed in the phrase "nature does nothing in vain,"

.^[60]

For a philosopher who condemned all metaphor as obscure, Aristotle is, one might think, extraordinarily free with implicit and explicit comparisons of every kind between the role of

and the

. But the first-stage defence he would offer is not far to seek. It is above all in relation to the workings of the final cause that these comparisons are developed. Both domains, Aristotle would insist, exemplify finality, though its modality in each is different: he points out, for instance, that nature does not deliberate, just as he also recognises that there are exceptions to finality, failures to secure the good, in both artistic and natural productions.^[61] But in many of the comparisons he draws he would claim that there is no question of transferring conclusions from one particular instance to another directly (thereby encountering the difficulty he mentioned in his analysis of analogical argument). Rather, both particulars fall under a general rule for which he believes he has ample grounds. Art can be used to illustrate nature because both domains manifest certain general principles concerning, for example, the adaptation of form to function, the hierarchisation of ends, and the relationship between the end to be attained and the character of the matter necessary to attain it. To quote just one prominent example:

[57] PA 668a13ff.; cf. Plato Ti. 77cff.

[59] E.g., GA 731a24, PA 652a31, 658a32.

[60] Bonitz's index, which is not exhaustive, cites twenty-three instances from the Corpus, 836b28ff. Cf. the discussions of Ulmer 1953; Solmsen 1960, pp. 102ff.; Bartels 1966; Fiedler 1978.

[61] E.g., Ph. 199a33ff., b26ff.

― 189 ―

just as an axe, to be used for chopping, must be made of a hard material such as iron or bronze, so each of the parts of the body must be of a material suitable for the function it is to perform.^[62]

But if, in general, we can see why art may be invoked as an analogue to nature, this does nothing to explain why in any given case a particular technological analogy should be used, let alone guarantee that it will not mislead. The crisscrossing of the blood-vessels may suggest wickerwork, but it does not show that they do indeed have the function of binding the front and the back of the body.^[63] Moreover, in this instance there is a fairly obvious negative analogy (or difference) that might have given Aristotle pause, in that the texture of the blood-vessels, the veins especially, might be thought ill suited to serve a binding function.

An even more disastrous example is Aristotle's theories concerning the role of the testicles, which he several times compares to the weights on looms.^[64] He believes their function to be, not to produce the semen but, rather, simply to keep the seminal vessels taut. It is true that he believes he has independent evidence that even after castration bulls can fertilise cows successfully, a supposed fact that he took to suggest that the testes do not produce seed.^[65] The tension of the seminal vessels, on the other hand, would—he thought—be released only gradually after the excision of the testes. The loom-weight idea offered the basis of an alternative theory, though the more immediately visible similarity it appealed to was—we should say—superficial.

Again, the general doctrine of the adaptation of the parts of living creatures to ends is expressed by Aristotle with the help both of particular comparisons with

, tools or instruments, and of the term

, instrumental, applied to such non-uniform parts as the hand. When he speaks of the organs of the body, the technological model plays an active heuristic role. A single text will serve to illustrate

[62] PA 642a9ff.; cf., e.g., 639b23ff., 646a24ff., b3ff.

[63] PA 668b21ff. Aristotle does, however, also hold that the blood-vessels serve to convey nourishment to the parts of the body, e.g., 668a12–21.

[64] E.g., GA 717a34ff., 787b19ff., 788a3ff.

[65] GA 717b3f.; cf. HA 510b3f.

― 190 ―

the doctrine: "since every instrument (

) is for the sake of something, and each of the parts of the body is for the sake of something , that is to say, some action, it is clear that the body as a whole arose for the sake of some complex action. Just as the saw came to be for the sake of sawing, and not sawing for the sake of the saw . . . so the body exists for the sake of the soul in a way and the parts of the body for the sake of the functions that each of them naturally fulfils."^[66]