by Umberto Eco
As Aristotle sees it, we find ourselves faced with an example of accidental equivocation (in the Middle Ages they would have said it was due to penuria nominum). We have synonymity or univocality when the term corresponds to a single definition (when, that is, zoon is said of a man or an ox). And finally, we have paronymity when things are designated by the same term but with a different grammatical ending (“the grammarian” [grammatico] when it stands for “grammar” [grammatica]). Owen (1951) makes it clear that Aristotle considers equivocity or univocality to be properties, not of the term itself, but of the things for which a single term is used.33 Thus, we have univocality when a single term is used for what is expressed by a single definition, and equivocity when we have a single term for two things that correspond to two different definitions.
Different uses of a term are broadly discussed in the Topics (I, 15, 106a 1–8), where Aristotle takes on for the first time the question of a twofold way of employing terms: it is one thing to say that justice and courage are called “good” univocally (because goodness is part of the definition of both) and it is another to say in various ways that what is conducive to health is good. The allusion here is to the original Aristotelian example, widely discussed in the Middle Ages, according to which both people in good health and the medicine conducive to good health, not to mention urine as a sign of good health, are dubbed “healthy.”
In the Nicomachean Ethics (I, 6, 1096b 23–29) the question of why honor, wisdom, and pleasure are called “goods” comes up again. The three things are different, and yet the use of the term is not an example of casual equivocation. Are they called “good” because they depend upon a single cause (“aph’enos”) or because they are directed toward the same end or good (“pros hen”)? Or is it by analogy, following the example of sight that is good for the body just as the intellect is for the soul? Here Aristotle clearly distinguishes the first case from true analogy, which sets up a proportion among four terms.34
In Boethius’s Latin translation of Porphyry’s Isagoge, “aph’enos” and “pros hen” are rendered respectively as “ab uno” (the term “medical” used both for the doctor himself and for the doctor’s potions and instruments) and “ad unum” (the classical example of “healthy” said of the body, the medicine, and the urine). Clearly, however, the first example is a relatively weak one, since it could be reduced to a case of paronymy. In fact the concept that remains central in Aristotle is that of pros hen. Briefly put, to be named for the cause one proceeds from or for the end toward which one tends is to all intents and purposes the same thing (we could say that the relationship is based on a common cause, whether it be efficient or final). What we have, then, are two forms of equivocity, pros hen, which the scholastic tradition will dub analogy of proportion (and, in the case of Cajetanus, of attribution), and that by analogy, which the scholastic tradition will dub analogy of proportionality. For convenience sake, from now on we will use the two terms attribution, which for Aristotle was not a form of analogy, and proportionality, which for Aristotle was the only form of analogy.
Aristotle explains the attribution in the Metaphysics (K. 3, 1060b 36–1061a 7) where he takes as examples of speaking “in several ways” the adjectives medical and healthy: they are used in reference to (pros) the same thing: a medical discourse and an instrument are both called “medical” because the medical discourse proceeds from medical science and the instrument is useful to that science; in like manner, things that are signs or causes of health are termed “healthy.” Now, health is something that is only found in a body and is not present in the color of the urine or that of the medicine (we ought to speak, then, of a patently equivocal situation in which a single term is referred to things that have different definitions). Both the urine and the medicine, however, refer to health. Just as the term being is used in various senses but with reference to one central idea (pros hen), and is therefore not equivocal, the same goes for the term “healthy.” Both express a common notion (legonthai kath’en).
Attribution is a relationship involving two terms: medicine is healthy because it causes health, and we cannot say that medicine is to the sick body as health is to the healthy body. The case of analogy is different. Here four terms are required, as we are also reminded in the Poetics and the Rhetoric. The stone is shameless because it is to Sisyphus as the shameless man is to his victim (Rhetoric, III, 11, 1412a 5, in Bollingen ed., p. 2253). Now, whereas the examples of attribution are always given as instances of the stereotyped use of language (healthy medicine, healthy urine), for Aristotle the analogy is an instrument of knowledge, and he makes use of it, when it serves him, in his books on nature too. “The underlying nature can be known by analogy” (Physics I, 7, 191a, in Bollingen ed., p. 326).
At this point let us reconsider the very nature of metaphor. As proposed in Eco (1984a: sect. 3.8.3), let us suppose that metaphor and metonymy can be explained on the basis of a componential analysis in the form of an encyclopedia which in the definition of a given term includes its form (or morphological aspect), cause, matter, and end (or function).
Property 1, form
Sememe A
Property 2, cause
Property 3, matter
Property 4, end or function
The idea was already present among the Scholastics: see, for instance, how Thomas (De principiis naturae, 6) admits that sometimes the like properties are predicated with respect to the cause, and at others with respect to the end.
To formulate the metonymy drink a glass (container for content), it is not necessary to compare two terms: one identifies in the encyclopedic definition of the glass the fact that it contains wine; the substitution is therefore one of semic interdependence within the same sememe. To call the glass the shield of Dionysus on the other hand I must compare the properties of Dionysus and the god of war Ares, recognize that in both the same morphological property appears (a typical instrument or emblem), identify a property that the two instruments have in common (both being round and concave in form), and activate the exchange. In both cases the substitution first occurs on account of the semic identity among sememes, then two sememes are crossed with two semes.
Now, it appears that metaphor imposes a comparison between two entities that were previously separate, thereby increasing our knowledge, whereas metonymy presumes prior knowledge of the thing played upon. Hence the greater cognitive power of metaphor.
Attribution seems to be akin to metonymy: we call medicine “healthy” because we already know that the property of medicine is to procure health. But if this is the case, then many of the metaphors cited by Aristotle, from genre to species and vice versa, are in fact forms of metonymy or synecdoche, given the fact that genre ought to be a property of the species. Just as being an animal is a property of mankind, and makes it possible for Francesca da Rimini to address Dante with the vocative “O aminal grazioso e benigno,” similarly standing still is a property of being at anchor. We have only to look at Emanuele Tesauro’s Cannocchiale aristotelico ([1670]1968: 284). He has no qualms about calling metaphors from genre to species and vice versa “analogiae attributionis.”
On the other hand, when Aristotle calls the stone shameless, he is attributing to it a property (certainly justified by the context) that had not previously been recognized. Let us take another look at the example calling pirates commercial purveyors. First of all, a four-term analogy is set up: the pirates are to the transportation of stolen property as merchants are to that of the goods they acquire. The impression of identifying a genre X, of which pirates and purveyors are both species, is a consequence of the analogical operation. In fact it takes two independent sememes and identifies in them a common property (that of being transporters of goods). Only when we have understood the metaphor can we say that pirates and purveyors belong (unexpectedly) to the same genre, or the same whole. The property they share, surprisingly brought to the fore, becomes a common genre.
purveyors
transporters
The entire Scholastic discussion of the analogia entis (despite the great variety of its outcomes) is fundamentally based on a choice between analogy of attribution and analogy of proportionality, and the examples are similar to those given by Aristotle when it comes to finding attributions or proportions between medicine and health and meadows and smiles.
The real problem, already looming in Dionysius, arises when the divine names come into play. When we say that medicine is healthy is that the same kind of attribution as saying that God is Good? We recognize the properties of health and we are familiar with the properties of both medicine and urine (one causes health, the other reveals it). Combining together known properties, we perform the attribution. What happens, however, in the case of the divine names?
There are only two possible solutions.35
(i) We know the goodness of things per prius and we infer per posterius that the cause of this goodness must exist in God. But what we have at this point is an inference from something known to something that must exist, but whose nature is unknown to us. And it is not enough to suppose that the cause must somehow resemble the effect. All the more since, in the course of his discussion of analogy, Thomas (in Summa Theologiae I, 45, 7, for example, following the lead of Augustine) distinguishes two types of likeness between cause and effect. The effect may represent “quantum ad similitudinem formae” (“by reason of the similarity of its form”) and this is the case with the “repraesentatio imaginis” (“the representation of an image”), in other words, of the statue of Mercury that resembles Mercury. But it can also represent by “causalitas causae” (“the causality of the cause”), in which case there is no morphological likeness but rather “repraesentatio per vestigium” (“representation by way of a visible trace”), as occurs both in the relationship between smoke and fire and that between a man and the footprints he leaves behind him. (Thomas—following Albertus Magnus—grants that the footprints may resemble the form of the foot, but he points out that the imprint of the foot is not similar to the man who left it and therefore cannot tell us who that man was.36 In Scriptum super libros Sententiarum I, 8, 1, 2 he gives the example of the sun, which produces heat but is not hot in itself.) If then we go back from the goodness of things to their divine cause, we do so out of causalitas causae, but we have no idea of what this goodness is like. We call Goodness the cause of goodness merely to make up for the penuria nominum, and hence a case of equivocation. It is as if, seeing smoke and not knowing anything about the fire that caused it is, we were to name this unknown quantity Hypersubstantial Smoke, thinking that what we were faced with was an example of repraesentatio imaginis and not repraesentio per vestigium. Let us consider the disturbing consequences of such a solution: if the mechanism of attribution were still valid, given that, among our actions and among the events of the world, some things are bad (a crime, rotten food, an illness), why do we not attribute the cause of these things to God, thereby making him responsible for Evil? Because we know a priori that there is no Evil in God (whereas there is Goodness). But if we already knew that, there was no need to look for an analogy. All that remains, then, is the second conclusion.
(ii) We know (by faith or revelation) the attributes of the Divinity, and it is therefore per prius, on the basis of these attributes, that we predicate per posterius the goodness of terrestrial things. We know, in other words, that God is ontologically Good per prius and that things are good per posterius, insofar as they share in the goodness of the Divinity. The attribute “good” characteristic of a certain thing is the equivalent of the attribute “animal” that characterizes a cat. We understand that a cat is an animal because we already know what an animal is. Thus we have a predication of a metonymical type from one known thing to another known thing: the attribution does not lead us to discover anything we did not already know.
Alternatively, predication in divinis implies an analogy of proportionality. But, in the case of the Aristotelian analogy, we discover an identity of properties between two things both of whose properties are known (the discovery involves the unsuspected relationship established between two known things). In an analogy extended in divinis, on the other hand, the trick would be to identify (and this would be truly unsuspected) an identity of properties between something about which we know everything and something about which we know nothing. In other words, the proportion established is not (as was the case with the shield of Ares and of the cup of Dionysus) A:B = C:D, but A:B = x:y, where x and y are unknown properties. This would in fact be the proportionality according to which we could say that human knowledge bears the same relation to the human mind as divine knowledge does to the divine mind. The most one could hazard is that between divine knowledge and the divine mind (both unknown) a relationship is established in some way similar to the one established between human knowledge and the human soul. But similar how? By repraesentatio imaginis or by causalitas causae? The comparison established between Achilles and the lion works as long as we already know what the wrath of Achilles is like, as well as the fierceness of the lion, and only then does the wrath of Achilles appear more convincing. But saying that divine Knowledge is to the divine Mind what human knowledge is to the human mind teaches us less than the comparison does about Achilles. In the second case, the wrath of the warrior, of which we already have some inkling, is reinforced through the comparison with the lion with the attributes of fierceness and courage. We learn something new. In the case of predication in divinis, we learn that something, we don’t know what, bears a pale resemblance to human intelligence. Accordingly, if predication in divinis were analogy of proportionality it would teach us less than a good metaphor teaches us.
Unless we already know what God is and what his qualities are, in which case the analogy would tell us something interesting about whatever is compared to God, not about God, about Whom we already know all there is to know.
It could be argued that the cases in which God is truly spoken of metaphorically are exempt from this criticism. The poetic metaphors of the Bible that speak of a God raging like a lion or as persistent as a woodworm tell us something about his wrath or his obstinacy. Granted. But these metaphors are not designed to reveal to us God’s nature, which is unknown, but the effects of his operations, which we already know. They do not posit an unknowable God but a God already anthropomorphized, like the pagan gods. Proceeding from the known to the known, these metaphors place something before our eyes, but in the mode of a simile. We are on this side of, or in any event outside of, an analogical discourse in divinis.
This is the fundamental weakness with any discussion of the analogia entis, and in fact all it permits the philosopher to discover is what the philosopher already knew on faith. It is no accident that discussions of the analogia entis engender prodigies of subtlety, but end up dissipating with the Scholasticism of the Post-Reformation. In fact, whenever we have to speak of the divine attributes, if we assume a Platonic-Augustinian position, then we already know everything about God for innate reasons, and only because we have this knowledge of the divine can we say that something shares (pallidly) in His Goodness or another of the transcendental properties of being. These appear to be the terms in which authors like Alexander of Hales, who speaks of the soul as “imago Dei,” or Bonaventure, for whom the soul possesses “principia per se nota,” handle analogy. And analogy is not so much a pathway to knowledge as a proportion known by illumination (see Lyttkens 1952: 123–153).
Otherwise we must take experience as our starting point, in which case the analogia entis is reduced to the rational demonstration of God’s existence, or to the formula that basically reiterates Thomas’s five ways: given a chain of cause and effect in the world, ergo there must exist a causeless first cause. Apart from the fatal weakness of the argument (the ergo that leads up to the final conclusion is exactly what was supposed to be proved—that is, just as the things of the world suppose a chain of causes and effects, so the chain of causes and effects of t
he world supposes an otherworldly cause—an argument that fails to withstand Kant’s criticism), we should note that what the five ways tell us at the most is that God must exist, not what God is like.
In point of fact, any discussion of analogy only serves to remind us that all we can predicate of God is Goodness, Truth, Fullness of Being, Unity, Beauty, but nothing further. And it can only come up in a culture that already assumes that God is Goodness, Unity, Truth, and Beauty.
Precisely on account of this dramatic impasse, which will lead to its collapse, the analogia entis has less cognitive value than a good metaphor.
3.8. Conclusion
The poetry and prose of the Middle Ages abound in metaphors, while contemporary theory, be it philosophical or poetic and rhetorical, is inadequate to account for this richness. This should not surprise us, as it is a commonplace that the culture of the time frequently shows a dichotomy between theory and practice. The typical example is music, a field in which the doctrinal discussion is extremely abstract, based on Pythagorean models, relicto aurium iudicio (“setting aside the judgment of the ears”), as Boethius remarked, and as a result deaf to the evolution of musical practice (see Eco 1987 and Dahan 1980: 172). But at least in the case of music there is an explanation, which is, as we mentioned, the weight of the Pythagorean tradition as transmitted by Boethius. Can we find a similar reason in the case of the theory of metaphor?