You are on page 1of 8

The 'Analogia Entis' according to Franz Brentano

A Speculative-GrammaticalAnalysis of Aristotle's "Metaphysics' E n z o Melandri

Die Kategorieneintheilung des Aristoteles hat in einer wunderbaren Weise dem Wechsel tier Zeiten getrotzt. --


1. In the modern age the problem of individuation was solved through a physicalistic reference. Usually the solution was presupposed implicitly, but Kant made the attending reduction explicit. Space and time, as forms of intuition, are sufficient to provide every existing thing with a unique indexing (x~, x2, x3, x4 ---- t). Thus space and time are not c a t e g o r i e s any more, but indexical p a r a m e t e r s which are part of the cognitive apparatus. Going back to the origins of this turning point in the history of thought, we see that the postulate of the impenetrability of bodies was connected with the extensional individuation of physical reality: in the res e x t e n s a every space-time point must be outside every other. Every existing thing, if it is to be individuated, must obey such a postulate. This has the advantage of reducing the metaphysical problem of the i n d i v i d u a t i o n of a being to the easier problem of the i d e n t i f i c a t i o n of a body through mathematico-physical coordinates. Obviously, the reduction is not automatic: there must be a Copernican revolution in the viewpoint. But this metaphysical shift is not part of science, it remains insensible and subterranean: it is enough to say that the tree of knowledge has its roots deep into the earth. If two different bodies could coexist within the same space-time point, clearly the spatio-temporal indexing would not be sufficient for individuation. At the time of classical physics, from Descartes at least until Laplace, one could not predict that such a condition might not hold a priori, that is in a necessary and universal way. But for more than a century now two theoretical considerations, marginal at first, have gradually become so crucial as to shake our confidence in the constituting presuppositions of this m a t h e s & u n i v e r s a l i s : a clear sign that a 'deus deceptor, valde callidus sive malignus' may induce us in error even when the evidence is most reliable.
Topoi 6 (1987), 51--58. 9 1987 by D. Reidel Publishing Company.

The first of these two considerations was the discovery of non-Euclidean geometries, or rather the proved possibility to give them a coherent physicomathematical interpretation. This is a well-know fact and we need not linger upon it. But it is worth making a point which is important for our argument. The spatiotemporal indexing, once non-Euclidean geometries are admitted, is no longer sufficient to ensure individuation, unless we specify the kind of geometry that is in question, that is, the sign and degree of curvature of the space being considered. And notice that the whole question is independent of relativity theory, since it originated before it and therefore can be understood in its absence. Gauss was already aware of it. The second consideration, less well-known in epistemology, goes back to Lotze and to his researches in the physiology of spatial perception. 2 Lotze was able to prove for the first time that the space perceived, that is, spatial consciousness, is not based on some preexisting form of intuition, but is itself the contingent result of an operation of logical reorganization of local sensations. So there is no innate form of spatial intuition. For if there were one, it would be universal and necessary, whereas from a genetic point of view topical sensations, though localized, do not predetermine any unique and logically cogent result. Which naturally does not exclude that at some level of development of a spatial consciousness there might be a convergence on Euclidean geometry, say through a process of equifinality. But this would be a fact to be explained, not an a priori binding condition. Therefore proceeding from the bottom up we get to the same result: the psychic space is as pluralistic, in principle, as the geometric-axiomatic one. Not less important will be the researches on time, on the temporal flux, and on duration conducted by the experimental psychologists of the turn of the century, among whom are important Meinong and the Graz school. Even without recurring to later theories, like special and general relativity, or to indeterminacy in microphysics, we have enough justification here to talk about


ENZO MELANDRI 2. For Aristotle space and time are categories, not parameters. This means that their combination does not support a principle of individuation, because they lack inherence in a substance. It would be like saying that there is a coincidence of 'musical' and 'white' without mentioning 'who or what' is musical and white. The problem of individuation is solved by saying what/s that we talk about, and this is an ontological problem. The question of ontology cannot be solved by a "science of being", 6wo-2oTia in general, as is the case for those particular sciences which deal with a special ontic determination, such as psychology, physiology, cosmology, and so on: Aristotle says that 'being is not a genus', that is, it is not a particular object, which can be circumscribed categorially. Thus the problem of ontology becomes the converse problem, that is, that of establishing what it means to be in discourse. The question can be reformulated as follows: can language understood as 26yoq d~o~avroc6q answer by its means the question rt" gar,? -- concerning the quid, the "what" we talk about? And this not only in the sense of the category, of the predication which answers the question by making being fall within an ontic genus, but also and especially in the stronger, ontological, sense, which expresses the modality of its relation with the "ens per se", ~v raOdvrr, and determines its individuation as well as its identification? It is to Brentano's credit that he, developing a well-known thesis of Trendelenburg, radicalized the indirect way in which Aristotle addresses the ontological problem, to repropose it in terms which it is not abusive to define of speculative grammar. Trendelenburg 4 would have been the first one to notice, among the moderns, that in Aristotle if one thing is essentiallypredicated of another so that name and concept of the predicate applies to it, then this occurs in a grammatically different form than if the predicate merely gives its name to the subject without being of the essence of the subject.5 And it is Brentano himself who speaks of Trendelenburg's peculiar ability to exploit the "speculative content" of the ancient thinkers by starting with the affinity that such content often has with the grammatical peculiarities of "linguistic forms". 6 Let us proceed carefully. What are categories in the Aristotelian sense? (We know that the term is generally used in a very vague way.) They are the highest or most general predications, which must be understood both in

a crisis of the foundations of contemporary knowledge (not only of physics), which reopens in a completely general way (and not only within metaphysics) that problem of individuation that had previously been forgotten because it was considered solved. This is borne out by the research themes of the turn of the century: the theory of judgment, the thesis of the primacy of the proposition on the terms, the psychology of acts, the independence of objectivity from existence and in general the reappearance of an ontological formulation of cognitive problems, not to mention the great development of logic. So when looking back more than a century later we see the emergence, under various titles and beginning from different disciplines, of an area of related interests within which in my opinion must be located the "Aristoteles-renaissance", too. The development of these studies (it will be enough to refer here to Zeller, Prantl, and Trendelenburg) takes off from an interest in the history of ideas, to be reconstructed with a philologico-critical method, but issues, especially with Brentano, in theoretical results which without representing novelties still constitute at the time very important suggestions for a critical reflection. On the other hand, in all the "renaissances", which Toynbee interprets as contacts of a civilization with itself at different times, the rediscovery of the ancient on the part of the new becomes frequently the reason of unpredictable developments. Let us summarize. For Aristotle a principle of individuation based only on the identification of a space-time point would not have been sufficient. For such a coincidence could have occurred through the combination of two categories, space and time, in a purely accidental manner, without any implication of inherence in a substance. To such a space-time point could belong many other categories, whereas the essential implication of substance excludes such distributivity. Aristotle's problem, from a metaphysical point of view, was that of seeing whether with the means of declarative discourse it was possible to identify what the discourse itself talks about. The solution of this problem is the theory of substance. But the identification of a space-time point does not support such a construction, and does not exhaust referentiality in the Aristotelian sense. The problem is not one of "recovering" substance, but of beginning to reflect on the meaning (perhaps new for us) of some very old questions?

THE 'ANALOGIA ENTIS' ACCORDING TO FRANZ BRENTANO the ontic sense of greatest genera, 12g~tara ygvrl, which circumscribe a certain region or section of being, and in the logical sense of predicates which cannot themselves be subjects, that is, ~carq~ooiat with respect to which no further predication is possible. The purely logical conception of categories is the simplest and also the least deceptive, but this reduction to mere predicates is rejected by Brentano because Aristotle's logic -- or better, analytic -- is not part of metaphysics or ~rOo)rr / gartar0/ar/ (which is real science), but as formal science deals only with fictitious entities like the "ens tamquam verum', 6v aSg 6.2r/0gq, and not with real predication. This remark will be better appreciated later, when Brentano deals with trans-categorial or analogic predication, the 6/uwvf~pwg 7r For the moment, it will be enough to remember the fact that logic and metaphysics are equally interested in the division into categories. 7 Categorial predication is thus part of that way of relating to being which Aristotle calls of "synonymy", ovvwv~l~wg ravqTooe~aOat , and which consists in subsuming the t~v under its most general ontic determinations. Synonymy must be understood as internal to each of these highest genera. 8 For the distinction and correlation of the logical and the metaphysical sense in Aristotle see e.g. the corresponding use of the apophantic 'tamquam', which works synonymously (tO ~v ,~ 6v), and of the existentialindividual 'per se', which works transversally (r6 6v


modify the substance and is the most distant from it, we have the category of rd arO6g rt, the external relations. 11 Thus for Brentano the Aristotelian categories are exactly 2 3 ----8, as the result of three applications of the 8tafoeatg. le And in fact in Aristotle, too, occurs this number, in inclusive alternative with ten. The two residual categories, in this case, that is, having (~xetv) and posture (regaOat), may be regarded as redundant and subordinated to the ~ctvgaetr Note in passing how this reformulation makes it easier to understand the origin of the distinction, not much used today, between relations proper, that is external, and affections altering the substance, that is internal: the latter are the last six ztdOr/, excluding o~)aia and v~t z~O6g ft. And to go back to our previous argument, it will be noticed that space and time are here, consistently, not relations but affections of things. Space and time are not parameters because they react on what they contain. Resuming the exposition of Brentano, we see now the strict correspondence subsisting between the categorial schemes and the grammatical forms of the "nomen substantivum" (substance), the "nomen adiectivum" (quality and quantity), the "verbum activum etpassivum" (to act and to be acted on), the "adverbia loci et temporis" (where and when) and the "praepositiones" (external relations), respectively; the two residual categories of having and posture may correspond to the


"participium praesens et perfectum" ( habens, habitus). 14

This correspondence, of course, does not become a precise one in the absence of an adequate speculative commitment, which could convincingly interpret as anomalies the morphological deviations from the ideal form, and which could continue to the end in the direction of rational radicalization thus opened. E.g. an abstract noun like 'whiteness' will have to be understood not as a noun, nor as an adjective, but as a pure relation extrinsic to the unity that joins together the innumerable white things) 5 We can thus easily reach the conclusion, from another point of view, that the eight (or ten) Aristotelian categories completely exhaust the kinds of questions that can be meaningfully asked about ontic determinations, that is, about the 6Oocd of the r{ eo~/,.16 In fact, this seems to be the way to get there genetically, starting from the Socratic etoove(a. From this angle, too, the exhaustivity of the categories does not force one to regard their possible redundance as a capital defect. From a logical point of view, the syllogistic principle

As against Kant and Hegel, who wanted to see in the Aristotelian table of categories nothing more than a merely inductive list without any systematic justification, Brentano can establish how to get to this table by a consistent and complete 6tas162 l~ The systematic criterion coordinated with the procedure of division by two and reutilization of the residue follows the order of the growing closeness to o~)aia. Proceeding in the reverse order, from the top down, we find first of all the great division between the category of of~aia, the prime substance relative to the individual, the fvroreipevov, and all the other not substantial but accidental ontic determinations, the ovlufleflrlr6m. Among these we distinguish, according to the categories of inherence, the gwztd~xovra, that is how much (aroaOv) and what (z~ot6v); then come the dynamical affections, the Ktvgaetg of acting (aroteCv) and suffering (zdo'zetv); in a similar way, these are followed by the circumstances, rd ~v rtvt', where (arot3) and when (~roz~); and then, as the only category not of affection or ardOog, which does not


ENZO MELANDRI category, but only in the residual ontic sense of a

intervenes as a test of the normality of categorial predication. It is in this direction, toward the atEartg &d ovMtoyto/~oO, that not so much Aristotle's logic as his analytic, that is, his speculative grammar, developed. E.g. in 'Socrates is white, and white is a color; therefore Socrates is a color' the reduction to syllogistic morphology proves deceptive. But theoretical reflection shows that 'white' occurs the first time as a qualitative adjective (inherence) and the second time as an abstract noun (relation to one); finally the conclusion shoultl hold again as an inherence. One can only conclude downward: 'therefore Socrates is a colored thing.' One can only go from the top down, or stay on the same level, and nothing else: the syllogism does not allow us to increase our knowledge ex terminis. But it has another function, too: it is a test of order and regularity in categorial predication. The reduction in syllogistic form, however, requires as a supplement an independent theoretical conception of grammatical morphology. When this reduction is successful, this means that the categories have a homogeneous application, without holistic complications: a "mereological" application as it were. The principles of Aristotelian "synonymy", that is, of pure categorial predication, can finally be reduced to two: the principle of excluded contradiction and the principle of relevance, that is, of categorial order. The first one excludes that there be contradictory predicates, or which is the same, infinitely many predicates in the definition of the object; the second one entails the necessity to follow a fixed order, from the top down or viceversa, when considering the succession of the various categories. Traditionally one always followed the order from the top down, from substance to relation. Aristotle himself does it that way. This order is commonsensical and intuitive, since it is easier to think downward than upward. But this procedure has the drawback that it distracts attention from the fact that the first division, the one between substance and accidents, is radically different from all the others, those between the various accidents. The order in terms of a maximum and minimum of substantiality or a maximum and minimum of relationality is not sufficient to fill the gap, because the first division is between ontological and ontic, whereas the others concern the ontic only. Going from the bottom up, on the other hand, from the relational to the substantial, we can see how the process of integration of ontic determinations never reaches the summit of oOaEa. Substance as ztOtbrq o~)oia may be a

3. The problem of being must then receive a nonsynonymic formulation. This is what is meant first and foremost by 'being is not a genus'. At the beginning of the Categories Aristotle prefixed the distinction between "homonymous" and "synonymous" terms. We have already discussed the synonymous ones: they are the categories understood as highest genera and ultimate predicates. Of the homonymous (or equivocal) ones is given a purely grammatical exemplification. But elsewhere (Eth. Nicom., 1, iv, 1096b) he says, after excluding the case of "fortuitous homonymy", that the most important theoretical question is that of the "homonymy by analogy". Thus predication can be synonymic or categorial, but also homonymic or analogical. The latter, too, is a predication, though its terms do not fall within the categories. It concerns the ontological, not the ontic, sense of being, and this explains why Brentano, following Aristotle, does not allow one to speak of the categories in terms of purely logical predicates. The disjunction would seem complete, if it were not for paronymy. In the introduction of the Categories paronymy comes last and is treated hurriedly, so much so that it never raised much attention. Paronymic predication concerns the relation between the casus rectus and the other casus obliqui, or between the meaning of a word or root and that of the other words which derive etymologically from it. To put it in more speculative terms, paronymy concerns the relation between substance, which normally should be expressed by the noun in the nominative, and accidents, which as we saw should be expressed by other parts of speech and different cases. Now the same thematic root can work as a noun, an adjective, a verb, etc. To us, the question seems strange, especially because Aristotle himself tells us that the noun signifies by convention, ~ a r d ovvO~rrlv, and not by some intrinsic virtue, or (~(met (De Int., ii, 16a). But we must consider the special theoretical context of speculative grammar, the antiquity of the problem and its connections with other areas of knowledge. Its origins are in the theme of the voktoO~og, of the creator of words and their derivations from the grvkta, as we find it for example in Plato's Cratylus, which for Aristotle is related with the analysis of metaphors and analogies, treated in books I and II of the Rhetoric. But the whole question can (in fact, must)

THE 'ANALOGIA E'NTIS" ACCORDING TO FRANZ BRENTANO be understood in a more strictly theoretical sense, and then it becomes, as we will see, an integral part of the analysis of the ~ctOtov~/~t9 rare~.'oQegaOat, that is, of ~ accidental predication par excellence.1 s To return to Brentano, he wants first of all to accept the thesis of the complete dichotomy of synomymic and homonymic predication, by including paronymic predication under the homonymic one) 9 Homonymy is the Aristotelian thesis of the ~022aZt~ ~ 2ey61aevov, which applies properly to the "ens" (rv) and to the "unum" (fv), but then extends easily to all transcendental, that is, transcategorial concepts. To be precise, what Aristotle says repeatedly is that "being is said in many ways, but not equivocally (o~)z 6/xa)v~ptog): rather, by analogy and relative to a one and to a unique nature (d22dt rar'ava2oTs rat ~Q6~ ~v nQrff i~[av rtvdt r Brentano interprets the exclusion of equivocity only as irrelevance of fortuitous homonymy (which is fine), but analogy as inclusive not only of proportion and proportionality but also of the relation to an identical term? ~ All depends on how one expands analytically the famous Aristotelian phrase zto22azt~ q 2ey6tzevov. To say 'by analogy and with respect to a one etc.' could be a hendiadys (fv 6t& 6vogv), which expresses one and the same meaning by two words..In this case the legitimate homonymy is the one regulated by analogy, which also includes what the Scholastics used to call 'attributive analogy'. Attributive analogy has three terms and includes the reference to a one. But the phrase might also not be a hendiadys, and then the analogy does not include all cases, but only some of them. Among those left out is precisely the one of the contingent relation to a one, the zcObg gv ~cai ~rObg /xCav rtvdt (~Oatv. The question could turn out to be purely nominal, but it is worth mentioning. The Scholastics used to distinguish three kinds of analogy. The first one is called of proportionality, and in it falls certainly the Aristotelian ta6~g 26ytov. The equality of ratios is to be understood not only in a quantitative or mathematical sense, but also in a comparative one. According to proportionality, this holds whenever a/b = c/d. This structure includes also the so-called continuous analogy, avvr/zgg, where the middle term sets the proportionality, as in a/b = b/c. The Aristotelian ta6rqg ~67tov bears traces of the many Pythagorean, Platonic, and Eudoxian speculations on this subject, and is also the reason why many Aristotelian commentators want the r~oarv to precede


the ~rot6v in the descending order of accidental predications starting with the ogas Which, as we will see, could have some intrinsic reasonableness: it is not to be taken for granted that the qualification must precede the distinction. Is it not the case that the "quale" has its foundation in the "quantum"? Or is it only viceversa? Then there is the purely proportional analogy, without either quantitative or comparative implications. E.g. 'this thing is as hot as that thing is white'. A proportion always has four terms, but here the analogy should we then call it continuous? -- uses a single mode of comparison, that of qualitative attribution. Is the principle of the ta6rq~ 267toy satisfied here? I would think it is, provided that we understand it as relative to a dSog, which qualitatively is something more than a ~v s r~o22~v. We agree with Brentano on the extreme importance of this non-quantitative form of equality of ratios, which institutes some kind of "preestablished" (though not always decidible) harmony among all the forms of predication of being. Not only quality but also all the other categories are subject to this proportional modification of their ultimate meaning. We are thus left with attributive analogy, which is a relation among three terms, that is, of two in relation to one. In the Scholastic conception analogy is understood from the point of view of a creature, that is, of the dependence relation that each being has with the "ens increatum", intrinsic origin and end of all things. All beings therefore depend on the god, or on the nature at once original and teleological, and this dependence is related with their distance from the origin or the end: it is a function of their proportionate imperfection. It is doubtful whether we can still speak of analogy here, or of analogical analogy, and so on, unless this concept becomes itself equivocal, and in a derogatory sense. The reference to a one is for Aristotle identical per se with the reference to being, though the one and being do not coincide in the in se: from an objective point of view they are the same, but in the concept they are distinct. He says that 30 ~v ~ca~rO 8v TavTOv 7ca112ia ~Oatg, but from a logical point of view they are not the same, ot~ z r ~vi ,~Oyqg. Brentano puts this kind of relation to 21 the one, too, which seems transanalogical besides being transcategorial, under the general title of the 6gogv~lttog rarqTopefaOat. He sees that case, too, supported by a principle of hidden analogy. But it is likely that he does so not to be forced to admit, in addition to the first two


ENZO MELANDRI Perhaps the ;tt)7og d:to~avur6g can individuate the being of which it guards the secret only if the speaker is given the faculty of proceeding to infinity. The Brentano of the essay on Aristotle would certainly not agree with these conclusions of ours. But even without considering the later development of his thought, which tends to give to being a fundamentally univocal meaning, what we get in the end from his interpretation of Aristotle is that the analogia entis is ultimately a lucus a non lucendo.

grids, the synonymic one and the homonymic-analogic one, a third one, paronymic, which would regenerate the whole problem from which we started. 22 Let us summarize. In the homonymic predication there are three grids: one is the analogy of proportionality, the second is the simply proportional one, and the third we can call equally well attributive analogy or paronymic predication. The analogy of proportionality is a system of quantitative or comparative relations, which is decidable by the understanding. Given three terms, the fourth one is always contained, even if for the moment it is unknown. Sometimes, as in the continuous analogy, one can deduce two terms over four. But analogy besides being continuous can be itself analogical, become an analogy of the analogy, move to qualitative inherences by going from the bottom up. It can bring us to saying, always respecting critical reason, 'this is as hot as that is white'. This a mere proportional analogy in the predication of inherence, qualitative in this case, of the second order. All of this is still straightforward, and it holds for all cases of analogicoproportional predication, not only the qualitative ones. But when we pass on to attributive analogy, to the pure and simple relation to a one, the one understood thematically, how can we avoid noticing that we are dealing with no restriction of the contingently such, of what at any one time is understood thus-and-so, but is relational in an abstract and ultimately arbitrary sense? The third and last grid, the one of attributive analogy, shows a decided inversion of tendency and involves a regress to the fortuitous. The first two analogies could perhaps be understood as successively upward, in direction of a progressive ontic convergence on the ontological. The third one cannot. It tends irresistibly to trivialize itself into the c o m m o n sense of what is one at a given time, according to the occurrence, the occasion, the changeable and impermanent and ephemeral sense of communicative discourse in act. At any rate, it is fatal that the highest analogy, the attributive or paronymic one, cannot be expressed if not in the same terms of the lowest category, the least substantial, ~lrtO'ra of)os the pure and simple relation :r06g ft. There is of course the possibility of a recourse to the One or the G o d as an object of the eminent reference. But even admitting that we can reach such a persuasion, how could this convince us that we have offered an explanation of the problem of being? A theism born out of desperation cannot suppress even in the believer the deepest atheism of the purely intellectual or~pLg.

4. Aristotle's "transcendentalia" are not concepts derivable from judgments of reflection, but simply properties of being that are ubiquitous because they are analogical. Unless we say that the very language by which he tries to fix the sense of the reference to things of this world, be they substances or accidents, is in its turn an affection of the substance which we want to express. But to objectivize a reflection in this sense, that is, in the sense of a retroaction of being on its individuation, we need different -- and later -theoretical concepts. In contrast with the above, a coherent analogical interpretation of being will be given by Plotinus, though at the cost of annihilating being. Plotinus refers to the analogy that will later be called attributive, solving in his own way the difficulties of Plato's Parmenides, by assigning a strong sense to the relation with the one. This is possible provided that we locate this one within absolute transcendence, that is beyond being, g:tgretva ~ o~)txs and it is only by virtue of the dependence on the one that being becomes what it is. The one and being, distinct in the concept, do not coincide even objectively any longer. The relation of participation in the one is what gives substantiality to beings, and the preeminence of this relation becomes the measure of the attributive analogy. The converse relation, which proceeds from the one to the world, becomes the emanation from the ineffable center into the various, degrading hypostases. All of this is perhaps Platonic, but certainly not Aristotelian. Aristotle criticizes the doctrine of ideas because of its separateness, the hypostases, the so-called third man argument etc. More significant, in my opinion, is his unwillingness to admit a ktil~latg or alternatively a ktg0e~tg as relations founding the dependence of being from a single principle. Thus we can make sense of the fact that the philology of the last century classified

THE 'ANALOGIA ENTIS' ACCORDING TO FRANZ BRENTANO Plotinus among the "Neoplatonists", a historiographic category which has some intrinsic systematic value. F r o m Plotinus descend the Arabic Aristotelians and then the Scholastics. A n d perhaps it was just an echo of the Scholastic, which he knew very well in its thomistic formulation, that induced Brentano to classify paronymy under the title of attributive analogy, or rather of analogy of relation to a single term. e3 We must say that Aristotle is rather stingy in exemplifying the paronyms. One of the clearest paradigms concerns medicine and health. A p r o p o s the ~to22azd)g 2eyrltevov, he notes that if the word is used equivocally and in virtue of nothing common to its various uses, being does not fall under one science (for the meanings of an equivocal term do not form one genus); but if the word is used in virtue of something common, being will fall under one science. The term seems to be used in the way we have mentioned, like "medical" and "healthy". (Met.,K, ii, 1060b) 2a Or that "being is said in many ways, but always with reference to a one and to some one nature". Except for proportional analogy, all that "is" is related to one central point, one definite kind of thing, and is not said to "be" by a mere ambiguity. Everything which is healthy is related to health, one thing in the sense that it preserves health, another in the sense that it produces it, another in the sense that it is a symptom of health, another because it is capable of it. And that which is medical is relative to the medical art, one thing being called medical because it possesses it, another because it is naturally adapted to it, another because it is a function of the medical art. And we shall find other words used similarly to these. (Met.,F, ii, 1003a--b) 25 I do not know whether or not it is legitimate to generalize on the basis of these and the not many other similar exemplifications which one finds on the subject. But be that as it may, we can conclude that dealing with being in the sense of paronymic predication belongs to, and hence falls within the powers of, gerto~t~ 1. Therefore the reference to the one does not indicate a transcendence in the theological sense, but is something accessible to science, even if to a science which among its principles admits the teleological one. One will never insist enough on the fact that that theoretical science which Aristotle puts on the same level as physics and mathematics, and which he calls Oeo2oyt~cO,being like the other two subordinated to the ztOtbrq gztto~prl, should be translated as teleology. A fortiori then the science of substance, since it is part of the etQagr~/


gmtrrOl~l, must be competent to determine the :t06 q gv rag ~rOOqpear vtvd ~(~atv.
F r o m a scientistic point of view, at this point one would have to blame Aristotle for not having fully developed a coherent language of science in general. For certainly his acute observations on synonyms, homonyms, and paronyms, though giving an overall impression of great systematicity from an architectonic point of view, are still insufficient, and hence unsatisfactory, from the point of view of method. In sum, the theory of categories leaves behind itself a residue: the ontological notion of substance. The analysis of the residue is done through the analogy, a more powerful principle, which retroacts on the categories and modifies their theory. A n d in its turn, the analogical theory of the equality of ratios leaves behind itself another residue, be that called attributive analogy or paronymy: that of the ultimate reference to being or to the one. This last residue, too, if it could be theorized, would retroact on the two previous theories. But there is in Aristotle no theory of the third predicative level which does not simply consist of a reference to 'at a given time'. So here arises a problem which, not being solved, reacts with deep destabilizing effects on the whole structure of Aristotelian gnto~p~1. The science of substance remains a presupposition, a ,3:r62e~ptg which cannot be reconstructed with the means of the 2670 ~

There has been discussion for a long time on whether or not for Aristotle it was legitimate to speak of supersensible substances. This pseudoproblem is certainly of Aristotelian origin, but it is discussed in the aporiae (Metaphysics B). Now the simple fact here, leaving along the sterile exercise of "reading Aristotle's thought", is that in Aristotle there are not even sensible substances. Substance is something sublime, indeterminable and inexistent. But it would be a mistake to address a scientistic objection to Aristotle. F o r Aristotle g~to~flr] is not based on an exact/udO~]at~, since it is nothing else than a 66~a extended to a potentially general consensus. F r o m this point of view the result he obtained must be considered perfect. 5. A scientific language should be based on a semiotics which is coherent, complete, and uniform. If some theory had to precede this semiotics to distinguish categories of signs and different semiotic functions the result would not be scientific, unless the latter


ENZO M E L A N D R I 2 H. Lotze, Medicinische Psychologie oder Physiologie der Seele, Leipzig, 1852. 3 M. Heidegger, in his 'Vorwort' to W. J. Richardson's Heidegger, Through Phenomenology to Thought, The Hague 1963, says that Brentano's work was his first livre de chevet during his University days (pp. ix--xi). And the affinity is clear between Heidegger's thesis of the Seinsvergessenheit and the result of Brentano's work. 4 A. Trendelenburg, Geschichte der Kategorienlehre, I, Berlin 1846, in F. Brentano, op. cit., Ch. V, w15. 5 F. Brentano, op. cit., p. 185. George, op. cit., pp. 123--24. 60p. cit., ib. 7 0 p . cit.,p. 194. 8 0 p . cit., Ch. V, w 9 0 p . cit.,Ch.V,w lo Op. cit.,Ch.V, 915;p. 184. 11 Op. cit., Ch. V, 913; see tables on pp. 175 and 177. 12 Op. cit., Ch. V, 91; see also g l 3. 13 Op. cit., ib. 14 Op. cit.,Ch.V, 913. 15 Op. cit., lb., pp. 186--87. 16 Op. cit., Ch. V, 98. 17 On this point Brentano's opinion is expressed very cautiously; see Ch. V, 913, pp. 148--50. It is interesting to consider Brentano's position in the light of the successive development of his thought: first his work on Aristotle's De anima (Die Psychologie des Aristoteles, insbesondere seine Lehre vom 'nous poietikos', Mainz 1867), and then the celebrated Psychologie vom empirischen Standpunkt, Leipzig 1874, where occurs the theory of the "intentional" or "inexisting object", by now entirely independent of Aristotle. 18 F. Brentano, Von der mannigfachen Bedeutung... , Ch. V, w pp. 85 and 98. 19 Op. cit.,Ch.V, 93. 20 Op. cit.,Ch.V, 96, pp. 108--13. 2~ Op. cit., Ch. V, 99, p. 126, footnote 161 (Met.,F, ii, 1003b). 22 Op. cit., Ch. V, w p. 98, see also above, Note 17. 23 Op. cit.,Ch.V, 911, pp. 143--44. 24 The translation is from The Works of Aristotle, ed. by W. D. Ross, Oxford, Clarendon Press 1908, Vol. VIII. 25 lb. See F. Brentano, op. cit., Ch. V, w p. 96. 26 We refer here to the distinction between "colloquial language" and "formalized language" in A. Tarski, 'The concept of truth in formalized languages', in Logic, Semantics, and Metamathematics: Papers from 1923 to 1938, Oxford University Press, 1956, pp. 152-278.

theory had in turn a coherent, complete, and uniform semiology. But there is no semiotic r e p r o d u c t i o n which is so indefinitely extended as to include the modifications of semiotics. If these are not present f r o m the beginning, the result can be called in various ways, the least appropriate of which seems to be semiotics, since its n a m e is Erkenntnistheorie. T h e objection to Aristotle, anticipating the modisti, would be the following. T h e different parts of speech, that is, noun, adjective, verb, etc., when expressed in a speculatively correct manner, have (in agreement with the modisti) different m o d i significandi. N o w it is true that the diversity of these m o d e s can be m a d e precise, as we saw, for the various species of categories, such as substance, inherence, m o v e m e n t , etc. But the definition c a n n o t be purely morphological, that is, ontic. It must take into a c c o u n t the modi-fying factor which intervenes in the regress to h o m o n y m y and p a r o n y m y of the individuating, that is, ontological, signification. T h e residue re-modifies the determinations already acquired, and we are not at all certain that this guarantees a process of approximation. Since p a r o n y m y is modistically undecidable, and therefore the system remains always open, we must be ready to accept its unpredictable catastrophe. Only in a closed system, which tends artificially to univocity, coherence, and completeness within itself, would it be possible to have a perfectly formalized language and hence a conclusive science. But f r o m the point of view of language, that is, of communicative discourse, this would be a solipsistic language, which cannot speak to others, not about the world. 26 In spite of its inadequacies, we prefer the clear c o m p e t e n c e of Aristotle in his Sprachlogik, the logic of speech. A n d we would like to c o n c l u d e by quoting a graffiti which a p p e a r e d at one time at the Sorbonne:

q u a n d on indique la lune du doigt, l'imb~cile regarde le doigt.

1 F. Brentano, Von der mannigfachen Bedeutung des Seienden nach Aristoteles, Freiburg im Breisgau 1862, p. 193 (Olms, Hildesheim 1960). The book is dedicated to Trendelenburg, "meinem verheresten Lehrer', and its motto is r6 dv 2~erctt ~to22ctz~~. For the English translation I used F. Brentano, On the Several Senses of Being in Aristotle, ed. and transl, by R. George, Berkeley, Los Angeles & London, University of California Press, 1975.

Dipartimento di Filosofia Universit& di Bologna via Z a m b o n i 38 Bologna, Italy