Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/de-gruyter/categories-in-topics-i-9-a-new-plea-for-a-traditional-interpretation-uqqGKXPkOW
- Publisher
- de Gruyter
- Copyright
- © 2021 the author(s), published by De Gruyter.
- ISSN
- 0003-9101
- eISSN
- 1613-0650
- DOI
- 10.1515/agph-2019-0106
- Publisher site
- See Article on Publisher Site
Abstract
1IntroductionTopics I 9 has received much attention in recent scholarship. One reason is that it promises to offer a rationale for the ten katēgoriai, and consequently to illuminate Aristotle’s doctrine of the categories. It may seem that the treatise entitled Categories should be the first text approached by someone so interested. But even though it certainly contains important applications of the doctrine, this treatise is not particularly informative about it – it gives a list of ten kinds, but does not clearly say what a katēgoria is. The word katēgoria is used only in passing, and the title Katēgoriai, which eventually got the upper hand in the tradition of the text, is just one of several and cannot be trusted. Indeed it is not adopted in the latest critical edition, Bodéüs 2002, where the editor opts for the title Πρὸ τῶν τόπων, ‘Before the commonplaces’ (i. e. ‘Preliminaries to the Topics’), attested in the tradition, while enclosing Κατηγορίαι in square brackets. By contrast, Topics I 9 appears to provide a tidier and more principled account, because the notion of katēgoria is more intelligibly connected to Aristotle’s standard use of cognate words like katēgoreisthai and katēgoroumenon, in the sense of ‘to predicate’ and ‘predicate’, covering both linguistic and metaphysical predication. Moreover, in the Topics Aristotle deploys the doctrine of the categories as a source of dialectical topoi or commonplaces, and this may offer a clue about the origin of the doctrine itself and the ontological commitments it was meant to make in its original application. In particular, an emphasis on predication raises important issues concerning the role played by substance or ousia, and especially the primary substance of the Categories, which is the underlying subject of everything, being itself predicated of nothing. Why should primary substance be a category, if indeed the notion of category stems from the notion of predication?In order to decide whether the chapter keeps the philosophical promise, one must first resolve a very difficult exegetical problem which can be directly put as the question: what is the first category in the list at the beginning of Topics I 9? This paper criticizes previous attempts to solve this problem and tries to vindicate (with an important improvement) the traditional interpretation, stemming from Alexander of Aphrodisias, which has come seriously under fire in recent contributions. For the traditional interpretation see Mansion 1968, 198. This interpretation is still adopted in several more recent discussions resisting the new wave: Irwin 1988, 502 n17; Smith 1997, 75 f., who also claims that Aristotle distinguishes an absolute and a relative sense of ‘what it is’. While solving the exegetical problem, the paper aims to show the importance of the distinction between ‘predicate’ and ‘predication’ in the interpretation of the term katēgoria. The distinction was made in Michael Frede’s seminal contribution on the subject Frede [1981]/1987a, 29–48. and has been regularly invoked since then, but if I am right it has been misunderstood and consequently misapplied in interpreting the chapter.2The Traditional InterpretationHere is the list of katēgoriai laid down at the beginning of Topics I 9:Next, then, we must distinguish between the kinds of predications, in which the four above-mentioned are found. These are ten in number: what it is, quantity, quality, relative to something, where, when, to be in a position, to be in a state, to do, to undergo. (103b20–23)The list corresponds to the list of the Categories, save that the first member of the latter is not ‘what it is’ but substance:Of things said without interweaving, each signifies either substance or quantity or quality or a relative or where or when or to be in a position or to be in a state or to do or to undergo. (Cat. 4, 1b25–27)The traditional interpretation (established at least since Alexander’s commentary on the Topics and then transmitted to us, unchallenged until recently, through generations of interpreters) See Brunschwig’s apparatus at b22 (Brunschwig 1967, 13): ‘supra τί ἐστι add. ἤτοι οὐσία AVPcu2’. Of course, to say that the traditional interpretation was not challenged is an overstatement on my part. The problems of this chapter have always been perceived. See the variety of resourceful interpretative suggestions put forward in the medieval commentaries on Top. I 9 collected and edited in Hansen 2016. plays down the general idiom of the Topics and restricts ‘what it is’ to ousia. The identification is justified by noting that in other works, especially in the Metaphysics, the first category is sometimes referred to as ‘what it is’ and still Aristotle clearly means ousia. Most perspicuously in Met. Δ 7, 1017a24–27; Ζ 1, 1028a10–20. An explanation why ousia may have been characterized by this generic description is suggested by Alexander: “he [Aristotle] took ‘what it is’ instead of ‘substance’ because substance is being in the strictest sense, and in the strictest sense ‘what it is’ and ‘definition’ are of substance” (in Top. 65.17–19). Textual support for this suggestion can be found only in Metaphysics Z 4, where Aristotle explains that other categories too can be ‘what it is’, but ousia is a ‘what it is’ in a primary way. “Or is it that definition too is said in several ways, like the ‘what it is’? For the ‘what it is’ in a way signifies substance and a this, in another way each of the predicates: quality, quantity, and all the others of this sort. For as ‘is’ belongs to all but not in the same way, but of one in a primary way and of the others secondarily, in the same way also the ‘what it is’ holds simpliciter of substance, and in a certain way of the others. For, indeed, of a quality we may ask what it is, and so also a quality is a ‘what it is’, but not simpliciter” (Met. Ζ 4, 1030a17–25). In Z 4 Aristotle envisages two approaches: either we distinguish different manners or ways of being a what-it-is, or different meanings or senses of the phrase ‘what it is’ (1030a32–35). Since the distinction does not affect our present discussion, I shall leave the alternative open. In other words, although the question ‘what is it?’ can be asked about everything, it is a question that pertains to substances in a special and more fundamental manner. The reference to this passage of the Metaphysics seems reasonable and unproblematic. Interpreters have linked this passage to the Topics passage without hesitation, and no doubt it was in the back of Alexander’s mind when he wrote the passage quoted above. It certainly was in the front of Ps.-Alexander’s mind when he commented on Metaphysics Z 4, since he cites Topics I 9, 103b29–35 in full. Alexander, in Met. 473.3–10 Hayduck. For a similar move see Kosman 1967, 490 and 2013, 130–32.It is also clear that the interpreter who restricts the ‘what it is’ to substance is at the same time compelled to postulate different senses or manners of ‘what it is’ by what follows in the chapter. For, just a few lines forward, Aristotle says:But it is clear from the things themselves that a man who signifies the what it is sometimes signifies a substance, sometimes a quantity, sometimes a quality, and sometimes some of the other predications. (b27–29)The ‘what it is’ here cannot be restricted to ousia, for it clearly covers the other predications as well. This is not a problem, provided that we are prepared to grant that Aristotle has switched to a different and more general sense or use of ‘what it is’.Yet, as Michael Frede has pointed out, such a sudden change of meaning or use of the expression ‘what it is’ is implausible. And although ultimately not successful, Frede’s ingenious way of avoiding it deserves rehearsing. But in order to introduce and assess it, we must first read the whole chapter, set it in its context, and signpost its most important parts for ease of reference. Before we leave for a moment the traditional interpretation, however, let us note that it has two components:1)In Top. I 9 the title ‘what it is’ only refers to substance predicates.2)In Top. I 9 ‘what it is’ has different meanings (or can be attributed in different ways).In due course we shall see that the two points are best kept separate, because it will turn out that we can retain 1) while rejecting 2).3Topics I 9: Translation, With Context and a List of (Categorial) ListsHere is a translation of the chapter, followed by the Greek text as edited in Brunschwig 1967, 13 f.Next, then, we must distinguish between the kinds of the predications, in which the four above-mentioned are found. These are ten in number: what it is, quantity, quality, relative to something, where, when, to be in a position, to be in a state, to do, to undergo. For the accident, the genus, the proprium, and the definition will always be in one of these predications.For all the propositions ‘Proposition’ translates πρότασις, even though in the next chapter, Top. I 10, 104a8, a πρότασις is defined as a (yes/no) question. According to APr. I 1, 24b10–15, in dialectic there is the πρότασις-as-question and the πρότασις-as-statement. Since the latter is typically assumed as a syllogistic premiss as soon as the corresponding question has been answered, the two characterizations are only meant to illuminate two aspects of the same notion. coming about through these signify either what it is, or quantity or quality or one of the other predications. But it is clear from the things themselves that a man who signifies the what it is sometimes signifies a substance, sometimes a quantity, sometimes a quality, and sometimes one of the other predications.For when human In translating ἄνθρωπος, I use the gender neutral ‘human’, rather than ‘man’, although it is most prominently an adjective. Moreover, with or without the definite article the Greek noun is sometimes meant to introduce the species. Hence I avoid adding the indeterminate article and leave the matter open. is put up for attention and the speaker says that what is put up is human or an animal, he states what it is and signifies a substance; but when white colour is put up for attention and he says that what is put up is white or is a colour, he states what it is and signifies a quality. Likewise, also, if the one-foot-long magnitude is put up for attention and he says that what is put up is one-foot long or a magnitude, he will state what it is and signify a quantity.Likewise, in the others too; for each of such items, if either it is said of itself or its genus is said of it, signifies what it is; but when it is said of another thing it does not signify what it is, but quantity, or quality, or one of the other predications.Therefore these and so many are the things the arguments are about and come from. Next we must say how we shall acquire these things and the means by which we shall be well equipped with them.103b20Μετὰ τοίνυν ταῦτα δεῖ διορίσασθαι τὰ γένη τῶν κα-τηγοριῶν, ἐν οἷς ὑπάρχουσιν αἱ ῥηθεῖσαι τέτταρες. ἔστι δὲταῦτα τὸν ἀριθμὸν δέκα, τί ἐστι, ποσόν, ποιόν, πρός τι,ποῦ, ποτέ, κεῖσθαι, ἔχειν, ποιεῖν, πάσχειν. ἀεὶ γὰρ τὸσυμβεβηκὸς καὶ τὸ γένος καὶ τὸ ἴδιον καὶ ὁ ὁρισμὸς ἐν25μιᾷ τούτων τῶν κατηγοριῶν ἔσται· πᾶσαι γὰρ αἱ διὰ τού-των προτάσεις ἢ τί ἐστιν ἢ ποσὸν ἢ ποιὸν ἢ τῶν ἄλλων τινὰκατηγοριῶν σημαίνουσιν. δῆλον δʹ ἐξ αὐτῶν ὅτι ὁ τὸ τί ἐστισημαίνων ὁτὲ μὲν οὐσίαν σημαίνει, ὁτὲ δὲ ποσόν, ὁτὲ δὲ ποιόν, ὁτὲδὲ τῶν ἄλλων τινὰ κατηγοριῶν. ὅταν μὲν γὰρ ἐκκειμένου ἀνθρώ-30που φῇ τὸ ἐκκείμενον ἄνθρωπον εἶναι ἢ ζῷον, τί ἐστι λέγει καὶοὐσίαν σημαίνει· ὅταν δὲ χρώματος λευκοῦ ἐκκειμένου φῇ τὸἐκκείμενον λευκὸν εἶναι ἢ χρῶμα, τί ἐστι λέγει καὶ ποιὸνσημαίνει. ὁμοίως δὲ καὶ ἐὰν πηχυαίου μεγέθους ἐκκειμένουφῇ τὸ ἐκκείμενον πηχυαῖον εἶναι <ἢ> μέγεθος, τί ἐστι ἐρεῖ καὶ35ποσὸν σημαίνει. ὁμοίως δὲ καὶ ἐπὶ τῶν ἄλλων· ἕκαστονγὰρ τῶν τοιούτων, ἐάν τε αὐτὸ περὶ αὑτοῦ λέγηται ἐάν τετὸ γένος περὶ τούτου, τί ἐστι σημαίνει· ὅταν δὲ περὶ ἑτέρου, οὐτί ἐστι σημαίνει ἀλλὰ ποσὸν ἢ ποιὸν ἤ τινα τῶν ἄλ-λων κατηγοριῶν.ὥστε περὶ ὧν μὲν οἱ λόγοι καὶ ἐξ ὧν,104a1ταῦτα καὶ τοσαῦτά ἐστι· πῶς δὲ ληψόμεθα καὶ διʹ ὧνεὐπορήσομεν, μετὰ ταῦτα λεκτέον.I consistently translate katēgoria with ‘predication’, so to get as close as possible to the literal meaning, which suggests an act of attributing a predicate to a subject. However, we shall see that ‘predicate’ also captures an important nuance of katēgoria, and that the distinction between predicate and predication is crucial to assessing the weight of each occurrence of the Greek word in our chapter.The four kinds of predication mentioned in b23–24 and mapped onto the ten kinds of katēgoriai are the predicables or praedicamenta. They are implicitly called here katēgoriai but in the previous chapters they are called katēgoroumena. One scholar thinks that the different terms point to a difference of nature, but I doubt that Aristotle consistently distinguishes the meaning of these two words. Ebert 1985, 118. In fact, to see how Aristotle can blur the distinction cf. Met. Z 4, 1030a20. The predicables are defined in chapter 5 as follows:Definition: “A definition is a formula which signifies the ‘what it is to be’”. (Top. I 5, 101b38)Proprium: “A proprium is what does not exhibit the ‘what it is to be’ of the object but belongs only to it and counterpredicates with it.” (5, 102a18–19)Genus: ‘A genus is what is predicated in the what it is of many things which are different in species.’ (5, 102a31–32)Accident: ‘An accident is something which is none of these – not a definition, a proprium, or a genus—but yet belongs to the object. Or: what may possibly belong and not belong to any one and the same thing whatever.’ (5, 102b4–5)Even on the most superficial reading, we perceive that Topics I 9 elaborates on the notion of the ‘what it is’ (ti estin is repeated seven times in twenty lines), but as we proceed it also appears that the ‘what it is’ does not just pertain to the tenfold list of categories given at the outset, but also (whether or not in the same sense) to the fourfold list of predicables. Indeed, in the Topics, the definition and the genus are often said to be ‘predicated in the what it is’. This locution is defined as follows:Let us say that all such things are predicated in the what it is [ἐν τῷ τί ἐστι κατηγορεῖσθαι] which it would be appropriate to give as answers when asked what the thing in question is, as it is appropriate in the case of a human, when asked what it is, to say that it is an animal. (Top. I 5, 102a31–36)From the examples set out in lines b27–35, it is clear that ‘to be predicated in the what it is’ is what Aristotle has in mind:Human is picked out and the speaker says that it is human or an animalThe colour white is picked out and the speaker says that it is white or a colourThe one-foot-long magnitude is picked out and the speaker says that it is one-foot long or a magnitude.Or perhaps it is not so clear, if we note that the three examples given in b27–35 do not exactly match the predicables predicated in the ‘what it is’. True enough, predications of the genus involve the officially recognizable predicable, but what about predications in which the predicate exactly repeats the subject (self-predications of the kind ‘Human is human’)? The best candidate would be definition, but a definition is a formula – not a one-word predicate. However, something Aristotle says in Top. I 5 helps to allay this worry:But as for those who, in whatever manner, render [the definition] with one word, it is clear that they are not rendering the definition of the object, since every definition is a certain phrase. We should, however, class as definitory something like ‘the beautiful is the appropriate’, and similarly the question whether perception and knowledge are the same or different. For indeed, in connection with definitions, the greatest effort is about whether things are the same or different. To put it simply, let us call all those things definitory which fall under the same method as definitions. (Top. I 5, 102a4–10)So self-predication is a definitory predication after all. Alexander, in Top. 67.8–11 Wallies. Malink 2007, 280. And we can understand why Aristotle avoids offering examples featuring the full definitory formula. In discussing the system of categories he is very likely to concern himself only with the basic building-blocks of predicative propositions. Compare Categories 4, where the items classified by the list are ‘things said without interweaving’. Lines b25–35 are therefore likely to deal with both the predicables predicated in the ‘what it is’ even though no example of the full definition is provided. As for the other occurrences of the phrase ‘what it is’, we shall deal with them case by case.According to Alexander, the default subject of a predicable, described by Aristotle as ‘the object’, to pragma, is a species. Alexander, in Top. 39.1–5 Wallies. This makes sense, at least insofar as it is unlikely that dialectical problems should concern individual items, like Callias or his paleness. Moreover, if individuals were possible subjects, we would expect the species to be recognized as a fifth predicable. True enough, in lines b27–35 the subject is picked out in a way that suggests that it is a particular item, like Coriscus or this patch of white. The ekkeimenon, however, is the result of an ekthesis, and the latter, if cautiously performed, can isolate a universal item without warping its nature. Aristotle deals with the problem in SE 22, 178b36–179a10, in his discussion of the ‘third man argument’. Since a soft ekthesis does not necessarily require that the item picked out is a particular (a tode ti), nothing prevents our subject from being a universal. And a universal subject is a better fit for the predication of a predicable in a dialectical problem or proposition. On what I call ‘soft ekthesis’ in SE 22 see Fait, forthcoming.The chapter contains several lists of katēgoriai which must be made to cohere. I have given them a name for ease of reference. The name of the second list will be explained in due course. The others are self-explanatory.Official List (b21–23):These are ten in number: what it is, quantity, quality, relative to something, where, when, to be in a position, to be in a state, to do, to undergo.ἔστι δὲ ταῦτα τὸν ἀριθμὸν δέκα, τί ἐστι, ποσόν, ποιόν, πρός τι, ποῦ, ποτέ, κεῖσθαι, ἔχειν, ποιεῖν, πάσχειν.Interim List (b25–27):All the propositions coming about through these signify either what it is, or quantity or quality or one of the other predications.πᾶσαι γὰρ αἱ διὰ τούτων προτάσεις ἢ τί ἐστιν ἢ ποσὸν ἢ ποιὸν ἢ τῶν ἄλλων τινὰ κατηγοριῶν σημαίνουσιν.Ousia List (b28–29):Sometimes signifies a substance, sometimes a quantity, sometimes a quality, and sometimes one of the other predications.ὁτὲ μὲν οὐσίαν σημαίνει, ὁτὲ δὲ ποσόν, ὁτὲ δὲ ποιόν, ὁτὲ δὲ τῶν ἄλλων τινὰ κατηγοριῶν.Acephalous List (b38–39): but when it is said of another thing it does not signify what it is, but quantity, or quality, or one of the other predications.ὅταν δὲ περὶ ἑτέρου, οὐ τί ἐστι σημαίνει ἀλλὰ ποσὸν ἢ ποιὸν ἤ τινα τῶν ἄλλων κατηγοριῶν.4Michael Frede’s Criticism of the Traditional InterpretationAgainst the traditional interpretation Frede levels the following arguments:1)The Topics cannot presuppose a piece of doctrine found in the Metaphysics, because Aristotle has not yet developed that more mature doctrine when he writes the Topics. By importing a distinction of the Metaphysics into the Topics, supporters of the traditional interpretation argue in a vicious hermeneutical circle.2)Apart from the Official List at the outset, the rest of the chapter seems to presuppose that the ‘what it is’ has a uniform use across all sorts of entities. The phrase has been employed in the preceding chapters and will be used in the rest of the Topics in this general and neutral way. Including Top. VI 6, 144a17–18, a passage ostensibly using the doctrine of the categories. See also SE 22, 178a4–13, although this passage is extremely difficult. On one interpretation (Ebert 1985, 128 f.), ‘what it is’ should be taken as neutrally including all kinds of predications, because a general description of the fallacy is said to be illustrated by an example which does not concern substance, but the categories of doing and undergoing (cf. ‘like in this argument’, b9). Since, however, the general description does not fit the example at all, lines a6–8 may be a parenthetical remark, in which case ‘what it is’ can be understood as restricted to substance. See Fait 2007, 188 f. It is plausible, therefore, that the first occurrence was meant to be taken in the same way.The first argument may appear outdated. A few decades after Frede’s article not many interpreters would rule out a doctrinal point merely on account of a developmental argument. Moreover, Frede thinks that the distinction made in Z 4 presupposes the full doctrine of ‘focal meaning’, which is a questionable contention. Indeed, since the passage in Z 4 occurs amidst a ‘logical’ discussion, Aristotle may well be rehearsing doctrines of the Organon. On the λογικῶς style of enquiry in the Metaphysics see Burnyeat 2001, 19–24, 88 f.; Peramatzis 2016. The doctrine labelled ‘focal meaning’ or ‘core dependence’ is expounded in Met. Γ 2 and is also alluded to at Z 4, 1030a34–b3. However, since this doctrine presupposes a genuinely metaphysical approach revealing ‘how things are’ (1030a28), Aristotle keeps it separate from the ‘logical’ ways of distinguishing the senses or manners of ‘what it is’, which only tell us ‘how we ought to speak’ (a27). Several interpreters agree that the ‘logical’ distinctions of Z 4 come from the Organon.Still, we may renew Frede’s argument by forsaking chronological priority and invoking didactic priority in its place. The first book of the Topics is an introduction to a relatively self-contained textbook which is a prerequisite to the study of the philosophical sciences. Top. I 2, 101a25–28 and 101a34–b4. This means that we should not appeal to a common background of distinctions Aristotle’s pupils were supposed to share, because the latter are neophytes or nearly so. Moreover, the first book is a remarkably precise piece of work (relative to Aristotle’s standard) and all of its doctrinal definitions are very carefully crafted. The shift of meaning conjectured by the traditional interpretation would have confused Aristotle’s pupils as it has puzzled modern interpreters; Aristotle would have soon realized that.As for Frede’s second argument, at least prima facie its strength seems undeniable. It is, therefore, a crucial desideratum of any interpretation to be able to interpret the ‘what it is’ of the Official List on all fours with the following occurrences of the phrase.Frede thought that if we take ‘what it is’ in the same general and inclusive sense adopted in the rest of the chapter, the list makes excellent sense. His interpretation nonetheless finds an insurmountable obstacle in explaining the Ousia List at b28–29.5Frede’s InterpretationAccording to Frede we should not try to read the list of the Categories into our Official List, since the category of substance plays no role whatsoever in the latter. There is no place for the Categories’ primary substance, for, since it is an individual, from primary substance there is no predication. But there is no dedicated place for substance kinds either, i. e. for the Categories’ secondary substances, because the ‘what it is’ slot collects essential predications of all sorts of entities (we can ask what it is about objects, colours, sizes and what have you). The radical conclusion Frede draws from these remarks is that there is no category of substance in the Topics. The Official List is a collection of ‘free-floating’ predications. That is to say: a list of predications entirely defined by the manner in which the item predicated attaches to a subject in a proposition, regardless of the nature of that subject, which is left unspecified. Frede introduces here the crucial distinction between predications (i. e. acts of predication in a proposition) and predicates (components of the proposition which can be used to make predications), which we shall develop and apply later on. He then argues that the Official List is a collection of kinds of predication by showing case by case that the Topics has less interest in the corresponding predicates.The use of this list of predications in the Topics can be gleaned, according to Frede, from certain topoi employing these categorial distinctions in order to establish or demolish a conclusion. For example, a proposition stating what the subject is like (its quality) cannot be the attribution of a genus, as the genus can only “signify the ‘what it is’”. Moreover, the system is useful in order to distinguish the several meanings of a term. Notice in particular (although Frede mentions this only in passing) that in the Sophistical Refutations, a treatise often described as the last book of the Topics, a linguistic fallacy discussed under the title ‘form of expression’ (schēma lexeōs) precisely trades on category mistakes. These are errors caused by certain expressions misleadingly suggesting that the item signified belongs to a wrong category of predication. Such is, according to Frede, the environment where the doctrine of categories saw the light as a classification of kinds of predication. This is a plausible and attractive suggestion. But is it consistent with the chapter?As anticipated, Frede’s interpretation faces a serious problem when we come to the Ousia List at b28–29 (ὁτὲ μὲν οὐσίαν σημαίνει, ὁτὲ δὲ ποσόν, ὁτὲ δὲ ποιόν, ὁτὲ δὲ τῶν ἄλλων τινὰ κατηγοριῶν). Ousia is the first item and is one of the katēgoriai like all the items that follow, as implied by the phrase ‘one of the other katēgoriai’. So it looks as though ousia is back as the first katēgoria. Frede is not worried by the occurrence of the word ousia, and ingeniously argues that ousia here or elsewhere in the Topics is not a katēgoria. Frede [1981]/1987a. As a matter of fact, Frede generalizes this claim to the whole Organon. He thinks that the Ousia List should be understood with a pragmatic pinch of salt. Think of the following case: if I say ‘you can have beer or ginger ale or coke or any other soft drink’, I am not implying that beer is a soft drink. Rather, the real list starts only with the second member. Although slightly infelicitous, my sentence would be understood sano modo by any collaborative interlocutor. It remains unclear, though, whether Aristotle’s reader should be expected to be collaborative in this way, for the analogy between beer and ousia is far from perfect. It would be plausible to guess that the first item is not part of the list only if it was clearly different from the items that follow it. Is this the case in our passage? Frede thinks ousia simply means ‘object’, while the other members of the list are ‘features’. Frede [1981]1987a, 39: “Assume, e. g., that we had objects or substances and various kinds of features, namely, qualities, quantities, and whatever else there may be. Assume also that all the names we had were names either of objects or of features. In this case it would be true to say: ‘A name names an object or a quality or a quantity or another kind of feature’. Now, no one would take this sentence to mean that objects are among the kinds of features we find in the world”. Two remarks. First, I fail to see why we should import names and naming into our interpretation of the Ousia List. Second, Frede seems here to invoke an unemphatic use of οὐσία which I find very problematic. Aristotle’s word for the subject of predicables is πρᾶγμα and, whatever it means, οὐσία is never a nondescript concept. But nothing prevents us from taking ousia as secondary substance, i. e. as a substance predicate, like the species or a genus of a particular substance. Like Frede, Ebert 1985, 135, too denies that substance in the Ousia List can be one of the γένη τῶν κατηγοριῶν, and argues that to say that substance is a katēgoria does not amount to saying that it is a kind of katēgoria. His reason is that Aristotle is classifying kinds of items predicable of a range of subjects, but substance, unlike items in the other kinds, does not apply with respect to a range of subjects: a substance is not the substance of something. Yet, even accepting Ebert’s criterion, Aristotle could well include substantial predicates (species and genera of primary substances) in his classification. Those would be predicates of primary substances exactly as items in the other categories are predicated of primary substance. Indeed in this context it is clear the ousia is not the subject, but the predicate of the proposition saying what something is. In this case ousia could well constitute a homogeneous class of predicates; which means that we would have no reason to expel it from the list of katēgoriai (and from the chapter).Roberto Granieri Granieri 2016, 12 f. has noted that elsewhere in the Topics Aristotle mentions substance together with other titles suggesting a classification of categories: ‘Moreover, see whether the genus and the species are not found in the same division [ἐν τῇ αὐτῇ διαιρέσει], but the one is a substance while the other is a quality, or the one is a relative while the other is a quality’ (Top. IV 1, 120b36–37; cf. 121a5–9). Granieri justly complains that Frede fails to mention the passage; still I think Frede’s interpretation is not vulnerable to this kind of counterevidence. The ‘divisions’ are the divisions of kinds of being that are expounded in the Categories; and these, according to Frede, could not offer an answer to a question about the original scheme of the katēgoriai, because they are not the categories themselves, but rather a classification of entities arrived at through the doctrine of the categories. Indeed this is the reason why the Categories itself, according to Frede, is not the text we should interrogate for an answer to our questions about the original scheme. Granieri is right in pointing out that, pace Frede, in Cat. 8, 10b18–25, katēgoriai must be taken in a technical sense, but the passage is compatible with Frede’s point that the Categories does not articulate the original scheme of categories, but takes it for granted (see Frede [1981]/1987a, 48); nor does the passage suggest that substance is a katēgoria.Even confining the discussion to our chapter, though, we see that it is difficult if not impossible for Frede to deny that substance predicates are katēgoriai. So substance still lurks in the chapter, and we must reject Frede’s radical interpretation. We may agree with him that in the Topics there is no mention of primary substance, the particular objects of the Categories, but that can be explained, as we shall see at the end of this paper.6Two Lists of Categories?Several interpreters grant that in the chapter Aristotle distinguishes two lists of categories. Malcolm 1981, Malink 2007, Bolton 2013, Granieri 2016. Most conspicuously, Malink retains Frede’s claim that the ‘what it is’ in the Official List is not restricted to substance, but rejects his radical claim that ousia in the Ousia List is not referring to a category. Rather, the Ousia List presents the ten categories of Categories 4 (C-categories), while the Official List introduces a different list of categories (T-categories) as follows:Quoted from Malink 2007, 291, and 2013, 148 f. (NB essence = ‘what it is’)How does Malink set up the second table? The Official List is of course the inaugural list of T-categories: an exhaustive classification of predicates or predications in which each item must fall under precisely one column. In the Topics chapter, Aristotle intends to show that the first T-category collects items from all the C-categories, hence the Ousia list is introduced as a familiar piece of doctrine. It is not implausible that the Topics presupposes the Categories. Husik [1904]/1952; Menn 1995; Burnyeat 2001, 110, defend this suggestion very convincingly. For instance, Top. II 8, 113b15, takes for granted the fourfold classification of the manners of oppositions developed in Categories 10. Husik [1904]/1952, [523]/106; Menn 1995, 319 f.; and Granieri 2016, 11, also argue that Top. I 9 in particular presupposes the Categories, which is also plausible. The first step of the argument (the first gar clause, b29) shows that whenever predicables are predicated in the ‘what it is’, they also signify one of the C-categories. The second step of the argument (the second gar clause, b36) provides criteria for belonging to the T-categories. I discuss a serious problem in Malink’s interpretation of the second gar clause in Appendix A, but at this stage of our inquiry it is more important to understand what it may mean, quite in general, to have two lists of categories back to back in the same chapter. Frede had an excellent explanation of the role of the Official List as a mere classification of predications in the dialectical (sometimes turning sophistical) exchanges discussed in the Topics (+ Sophistical Refutations). It would also be easy to explain how the ontologically neutral Official List could have morphed into the list of the Categories with the restriction of the first category to substance. By contrast, on Malink’s account no similar explanation is possible. We may perhaps imagine that the list of the Categories had already been discovered by Aristotle and that the Official List was developed at a later moment by demoting substance and reorganizing the other C-categories to some purpose. See also Granieri 2016, 11, who accepts Malink’s distinction between C- and T-categories and claims that the second system is derived from the first: “T-categories are an application of C-categories to the logical field of propositions”. But which purpose? What the table of T-categories shows is that the crucial distinction in the T-categories is the distinction between synonymous and paronymous predication. Subjects of synonymous predication (e. g. justice) belong to the first T-category, while predicates of paronymous predication (the just or being just) belong to one of the other T-categories. Notice that for echein and keisthai there are no examples in Malink’s table of C- and T-categories. According to an interpretation of Cat. 7, 6b2–14 (already suggested by the interpolated passage at 9, 11b10–13 and endorsed by several interpreters) keisthai is only a paronymous entity derived from a relative. Hence it would be a T-category but not a C-category. For a different interpretation, however, see Crivelli 2017, 552. To be true, Aristotle often deploys the distinction synonymous vs. paronymous in the Topics, Top. II 2, 109a34–b12; II 4, 111a33–b4; II 8, 114a26–b5; III 1, 116a23–8; III 3, 118a34–9; IV 3, 124a10–14; VII 1, 151b28–33; VII 3, 153b25–35; VIII 1, 156a27–b3. but why should he set up a new classification of categories based on this distinction? Contrast the Categories, where primary substance is the ultimate subject and all the others categories are ‘said of’ or ‘inhere in’ it as a subject. As we glean from Cat. 2 and 5 with some plausible contextual assumptions. The fundamental relation of the C-categories is then a relation of predicative dependence on primary substance; a point of view from which it would make no sense to separate, e. g. justice, from its cognates derived by paronymy. The difference between an item and its paronymically inflected forms is a difference in linguistic ending which does not imply that they belong to different categories. Wouldn’t it be confusing then to set up an alternative list which re-uses nine out of the ten titles, and yet rests on an entirely different foundation?Malink would answer that Aristotle’s list of T-categories has a precise raison d’être as a classification of terms, and that it provides a principled system of semantic categories, and one compatible with the ontology of the Categories, although not designed to emphasize the same ontological relations. Malink uses these categories, especially the category of ‘essence term’, to interpret Aristotle’s modal syllogistic. Malink 2013, esp. chs. 8 and 9. This is certainly a legitimate use of an Aristotelian text, and Malink can claim that you know the tree by its fruit – in this case by the interpretation of modal logic that it enables. In the present paper I shall not further discuss the merits of this systematic use of the T-categories, for my main interest is the Aristotelian notion of category in general, and, at any rate, the correct interpretation of the text must first be assessed as a matter independent of any application. For criticism of Malink’s use of the Topics to illuminate the modal syllogistic of the Prior Analytics see Angioni 2018. And, as I said, I shall discuss Malink’s interpretation of the text in Appendix A.For another defense of the two-list account, note that a version of the T-categories (probably in a rudimentary shape only including a couple of headings) was already in use before the development of the list of the Categories, since it is already deployed by Plato in the Meno, when Socrates distinguishes, regarding virtue, between stating what it is and stating what it is like (71b3–4). Other applications of this distinction appear in Plato’s Laches 189e3–190c2, Protagoras 312c1–4 and 360e6–361d6, and Gorgias 463c3–5. There is some evidence in the Topics that Aristotle mentions a categorization which opposes the ‘what it is’ of virtue (hence of a non-substance) and its quality. Top. VI 6, 144a17–19: “Moreover, state signifies what virtue is, whereas good signifies not what it is but a quality; and it appears that the differentia signifies a quality [ἔτι ἡ μὲν ἕξις τί ἐστι σημαίνει ἡ ἀρετή, τὸ δ’ ἀγαθὸν οὐ τί ἐστιν ἀλλὰ ποῖον· δοκεῖ δ’ ἡ διαφορὰ ποιόν τι σημαίνειν]”. Top VI 5, 142b26–29, may also be relevant. Moreover, if inspired by Plato, the list would probably not be metaphysically neutral, because, from the standpoint of Platonic dialectic, the ‘what it is’ affords criteria of reality and fundamentality that are more reliable than the criterion of subjecthood of the Categories. According to this narrative, the Official List of Topics I 9 would be inspired by Platonic views about the centrality of the ‘what it is’ question and could be seen as a potential competitor of the Ousia list, since the latter harks back to the Categories, where the ultimate subject, the particular substance, plays a more important role than any predicate. In fact, in the Categories secondary substances (substance species and genera) deserve the title of substance not just because they are subjects, but also because they reveal what primary substances are (5, 2b8–10). Even so, there is no possible Platonic overtone in the Categories, because both criteria are subservient to primary substance as the ultimate subject. The power of revealing the ‘what it is’ confers on a predicate the status of substance precisely because it is the ‘what it is’ of primary substance. It might then be part of Aristotle’s strategy to offer the Official List as a concession to the Platonic approach (or simply to recycle a Platonic relic), only to neutralize it by showing in b29–35 that it perfectly dovetails with the Ousia List, because what the first category of the Official List collects is not a special class of (Platonic) entities, but a mixed bag still waiting to be ontologically partitioned according to the Aristotelian C-categories. For another interpretation recognizing the two lists and attempting to reconcile them, see Bolton 2013: taking a strictly anti-developmental approach, he argues that the two lists can coexist in Top. I 9 because they reflect two different approaches: one ‘according to opinion’, the other ‘according to truth’. The first is dialectical and deals with mere predications, the second is scientific and centres on substance.I have contrived this story in order to consider all the reasons for adopting the two-list interpretation. Doubtless, it is not without some appeal; but it amounts to little more than a fantasy. The evidence for a Meno-style categorization only consists of a single passage in the Topics and is just the maxim of a topos, so nothing Aristotle personally commits himself to. Moreover, the example at Top. VI 6, 144a17–19, involves the problematic case of specific differentia, which is a notorious anomaly (Irwin 1988, 64). For example, a differentia is predicated in the ‘what it is’ of the subject but does not signify the ‘what it is’, but quality (Top. IV 2, 122b12–17; 6, 128a26–29; VI 6, 144a16–19, 20–23). The best explanation (Irwin 1988, 64 f.; Perin 2007, 130) is that a differentia always presupposes the genus, which is the reason why it is ‘predicated in the what it is’ as well, and does not constitute a predicable in itself: Top. IV 6, 128a20–29 (cf. I 4, 101b18–19). The genus is the most important component of the definition (Top. VI 1, 139a29–30; 5, 142b26–28), while differentiae are just qualities of the genus. So even when predicated alone, differentia encapsulates the whole definition. See also footnote 51 below. Moreover, the full ten-item Official List, as well as the Ousia List – which I shall argue to be identical – strongly suggest a different story. By taking account of this alternative story, we shall be able to explain the entire chapter, and to do so by appealing to only one scheme of categories. This will undercut the reasons for reading the Official List as a Meno-style categorization. The alternative account I am alluding to will be used in the next section to interpret the chapter; once fully expounded in Section 8, Point 1, it will reinforce the proposed interpretation.7Interpreting the ChapterI shall offer my interpretation as a running commentary. In order to avoid excessive stage setting and unnecessary suspense, I shall sketch my argument quite dogmatically and postpone the discussion of three important assumptions until the next section.Next, then, we must distinguish between the kinds of the predications, in which the four above-mentioned are found. These are ten in number: what it is, quantity, quality, relative to something, where, when, to be in a position, to be in a state, to do, to undergo.Μετὰ τοίνυν ταῦτα δεῖ διορίσασθαι τὰ γένη τῶν κατηγοριῶν, ἐν οἷς ὑπάρχουσιν αἱ ῥηθεῖσαι τέτταρες. ἔστι δὲ ταῦτα τὸν ἀριθμὸν δέκα, τί ἐστι, ποσόν, ποιόν, πρός τι, ποῦ, ποτέ, κεῖσθαι, ἔχειν, ποιεῖν, πάσχειν. The Official List is a list of predicates. According to a sound piece of traditional scholarship I shall rehabilitate in Section 8, Point 1, this list of katēgoriai is a repertoire of the kinds of predicate of a particular substance and has been collected by imagining all the possible questions that can be asked about a particular substance. ‘What it is’ predicates are therefore contextually restricted to substance predicates. Crucially, this does not require a special meaning or use of ti estin.Moreover, the katēgoriai of the Official List are likely to be stable predicates rather than occasional and possibly context-dependent predications. This is because they are attributed to their natural, typical, subject: particular substance, and only these predications exhibit the permanent nature of the predicate. Keeping in mind that for Aristotle a predicate is not just a linguistic expression, but an entity (possibly as expressed in a certain way), let us explain this distinction as follows:A Predication (the literal translation of katēgoria) is the attribution of a predicate to a subject in a proposition.A Predicate is an independent item which can be attributed to a subject in a proposition.I shall expand on the distinction in Section 8, Point 2. In particular, I will show that not every feature of a predication is automatically an absolute feature of the predicate involved in the predication. For a foretaste of what I shall argue then in more detail, consider the two sentences, which I am offering only as an illustration:(1)John is a man(2)John’s brothers are menWe may want to say that (1) is a singular predication and (2) is a plural predication, but it would not make sense to say that ‘man’, the predicate shared by these propositions, is a singular predicate, or that it is a plural predicate. On the other hand, it is true to say both that (1) and (2) are masculine predications and that the predicate ‘man’ itself is masculine. For it retains this feature in all its predications. On what I have here dubbed ‘masculine predication’ of nouns see the discussion on page 25 and the cautionary remarks in footnote 50.Maybe Aristotle avoids the word ‘substance’ in the Official List because the reader would immediately think of primary substance, which does not have a place in this formulation of the system. Aristotle does not include particular substance simply because particulars are never predicated, and in the Topics they have no systematic role to play in the problems debated in dialectical discussions. The Official List coincides then, as a matter of fact, with the catalogue of the Categories minus particular substance. This is in line with the traditional interpretation, except that we do not need to invoke a special meaning of ‘what it is’, nor to import a doctrine of the Metaphysics into our chapter.For the accident, the genus, the proprium, and the definition will always be in one of these predications.ἀεὶ γὰρ τὸ συμβεβηκὸς καὶ τὸ γένος καὶ τὸ ἴδιον καὶ ὁ ὁρισμὸς ἐν μιᾷ τούτων τῶν κατηγοριῶν ἔσταιIn this sentence Aristotle puts forward the thesis he will defend in the rest of the chapter: the accident, the genus, the proprium, and the definition will always be in one of these katēgoriai, i. e. in one of the kinds of predicate introduced by the Official List. Chapter 8 has argued inductively and deductively that the system of the predicables is exhaustive: each concrete predicate will exemplify one and only one predicable. Given Aristotle’s systematic preoccupations, it makes sense to show that the system of the four predicables does not include any item which is not matched by the system of the categories. In this way, provided they have different jobs to do, both lists can apply to the same domain of items predicated.At this stage we could enter into speculations about the ontological commitments of the different lists. Is Aristotle worried that someone may think that the predicables constitute an alternative system of ontological categories? I refrain from pursuing this kind of conjecture here and move to the next sentence.For all the propositions coming about through these signify either what it is, or quantity or quality or one of the other predications.πᾶσαι γὰρ αἱ διὰ τούτων προτάσεις ἢ τί ἐστιν ἢ ποσὸν ἢ ποιὸν ἢ τῶν ἄλλων τινὰ κατηγοριῶν σημαίνουσιν. This sentence contains just the first part of an argument not to be completed until b39. The Interim List presented in these lines is a list of all the possible predications generated by the four predicables. All items belonging to the predicables not predicated ‘in the what it is’ (proprium and accident) will signify one of the nine predications different from the ‘what it is’. Conversely, all predications obtained through the remaining predicables (definition and genus), must be classed as signifying ‘what it is’. Where else could we fit such predications? So, in the Interim List the ‘what it is’ is a mixed bag, for every sort of entity can be defined by suitable items in the predicables of genus and definition, and here we see that all such items fall in the same class. Here the kind ‘what it is’ is indeed the inclusive kind Frede thought it was in the Official List. But this is not a permanent characterization of the ‘what it is’, as Aristotle will argue in the next step of the argument. The ‘what it is’ is a makeshift soon to be dismembered (which is why I have dubbed the list in this passage ‘Interim List’).Importantly, only in this passage can we be entirely sure that katēgoria means ‘predication’, because the grammatical subject of the verb ‘to signify’ is the proposition constituted by a predicable (cf. hai dia toutōn protaseis). So, what is signified by one of the items on the Interim List is just the kind of an act of predication. This is the reason why this list can only form an impermanent classification. Again: more on this distinction in Section 8, Point 2.But it is clear from the things themselves that a man who signifies the what it is sometimes signifies a substance, sometimes a quantity, sometimes a quality, and sometimes one of the other predications.δῆλον δʹ ἐξ αὐτῶν ὅτι ὁ τὸ τί ἐστι σημαίνων ὁτὲ μὲν οὐσίαν σημαίνει, ὁτὲ δὲ ποσόν, ὁτὲ δὲ ποιόν, ὁτὲ δὲ τῶν ἄλλων τινὰ κατηγοριῶν.This sentence does not contribute a fresh remark, but continues Aristotle’s argument. The hypothetical person who signifies the ‘what it is’ is the same person who has uttered the propositions mentioned in the previous passage (b27). On different occasions regarding subjects of all sorts, this person intentionally signifies ‘what it is’, but at the same time also signifies another katēgoria belonging to the Ousia List.It is already possible to see what Aristotle is aiming to achieve: even when a speaker uses a proposition to signify what the subject is (thereby engaging in a ‘what it is’ predication), the item predicated – as a predicate and independently of that predication – signifies a kind in the Ousia List. The Ousia List is then a classification of predicates. But we have argued above that the Official List is precisely a list of predicates whose first member is contextually restricted to substance predicates. It is therefore quite plausible that the Ousia List is nothing else than the Official List with the name of the first member changed. This suggestion affords a very reasonable explanation of what Aristotle is doing: the previous sentence has left unfinished business with regard to the Interim List. Aristotle needs now to sort through all the predicated items provisionally lumped in the interim ‘what it is’ class, and to attribute each case to its real kind of predicate in the Ousia (alias the Official) List. The change of the first title is necessary, at this stage, because the label has also been used in the Interim List, where the ‘what it is’ is free of any contextual restriction. It would be very confusing, from now on, to refer to substance predicates with the title ‘what it is’. But nomenclature apart, Aristotle is working with only one list of the kinds of predicate.If this is the case, the argument shows in two steps that every predicable belongs to only one kind in the Official List. The first step shows that for all propositions attributing a predicable not predicated in the ‘what it is’ (proprium and accident) the predication exemplifies one of the nine kinds of predication other than the ‘what it is’, whereas for all propositions attributing a predicable predicated in the what it is (genus and definition), the predication is collected into an interim ‘what it is’ kind. In the second step Aristotle ‘unpacks’ the interim ‘what it is’ kind and assigns each predicate involved in a ‘what it is’ predication to its own kind in the Official List. Of course there is no need for a similar move as regards the other nine kinds of predications on the Interim List. In fact, in their case any predication belonging to one of them immediately identifies a predicate belonging to the corresponding kind on the Official List.Let us now see how the rest of the chapter confirms this interpretation.For when human is put up for attention and the speaker says that what is put up is human or an animal, he states what it is and signifies a substance; but when white colour is put up for attention and he says that what is put up is white or is a colour, he states what it is and signifies a quality. Likewise, also, if the one-foot-long magnitude is put up for attention and he says that what is put up is one-foot long or a magnitude, he will state what it is and signify a quantity.ὅταν μὲν γὰρ ἐκκειμένου ἀνθρώπου φῇ τὸ ἐκκείμενον ἄνθρωπον εἶναι ἢ ζῷον, τί ἐστι λέγει καὶ οὐσίαν σημαίνει· ὅταν δὲ χρώματος λευκοῦ ἐκκειμένου φῇ τὸ ἐκκείμενον λευκὸν εἶναι ἢ χρῶμα, τί ἐστι λέγει καὶ ποιὸν σημαίνει. ὁμοίως δὲ καὶ ἐὰν πηχυαίου μεγέθους ἐκκειμένου φῇ τὸ ἐκκείμενον πηχυαῖον εἶναι <ἢ> μέγεθος, τί ἐστι ἐρεῖ καὶ ποσὸν σημαίνει.Sometimes Aristotle supports a general claim ‘from the things themselves’ (b27), i. e. with an illustration case by case. Cf. Top. I 5, 102a11; 102b20–21. In each of the three examples, Aristotle distinguishes the act of saying and the act of signifying. At the same time, the person who says what something is is the very person who, in b26–27, was described as ‘signifying’ the ‘what it is’. I suggest that, by using the verb legein, Aristotle simply alludes to the predication performed by the speaker and their statement. His point is that the predication may well state (legein) what the subject is, but the item predicated still signifies a kind in the Ousia List. What the three examples show, then, is that a ‘what it is’ predication does not cancel out or neutralize the true categorial nature of the predicate. In Section 8, Point 3, I shall criticize a widespread misunderstanding of this passage. There I shall also explain that the verb ‘to signify’ can have different nuances. An item predicated can signify a kind of predicate, a kind of predication, a category (whatever this means), and a kind of being. We must deal with it case by case.Likewise, in the others too; for each of such items, if either it is said of itself or its genus is said of it, signifies what it is; but when it is said of another thing it does not signify what it is, but quantity, or quality, or one of the other predications.ὁμοίως δὲ καὶ ἐπὶ τῶν ἄλλων· ἕκαστον γὰρ τῶν τοιούτων, ἐάν τε αὐτὸ περὶ αὑτοῦ λέγηται ἐάν τε τὸ γένος περὶ τούτου, τί ἐστι σημαίνει· ὅταν δὲ περὶ ἑτέρου, οὐ τί ἐστι σημαίνει ἀλλὰ ποσὸν ἢ ποιὸν ἤ τινα τῶν ἄλλων κατηγοριῶν.I confine the discussion of the grammatical difficulties of this sentence to Appendix A. On what I there argue to be the best grammatical interpretation, the sentence claims that the same non-substantial items can signify different katēgoriai at different times. Aristotle’s claim is restricted to non-substantial items, because hekaston tōn toioutōn is likely to refer only to the ‘amphibious’ items, those that signify different things in different predications, and these are only non-substantial items. Irwin 1988, 106 n36. Homoiōs at b33 and 35 precisely indicates this double-edged behaviour; and the gar sentence explains why it is common to all non-substantial cases, which are referred to as ‘each of such items (toioutōn)’. On the other hand, double signification does not affect the first kind, because it is not the case that ousia is ever said of something else. This means that ὅταν μὲν […] ὅταν δὲ at 103b29–31 should be read as indicating a strong contrast (μὲν […] δὲ) between substance and the rest. The only kind of predication a substance predicate can engage in is ‘what it is’ predication. This is why I called the list contained in this sentence ‘acephalous’: the first kind is not involved at all. In fact, it is not even included in the first part of the sentence, where Aristotle mentions the ‘what it is’.As far as this sentence goes, ‘what it is’ predication and the other kinds of predication occur at different times. It is not the case that (for instance) at the same time an item predicated signifies the ‘what it is’ and a quality. However, this seems to contradict the previous sentence where the examples showed that, by the same token, one can say or signify what something is, and a quality. A reasonable explanation of this discrepancy is that, although in the final explanatory sentence Aristotle only points out that the same item can signify two different kinds of katēgoria in different sentences, the reader is expected to understand that when a non-substantial item is predicated of something else, namely a substance, this kind of predication permanently characterizes the predicate, This explains why, as far as the nine non-substantial kinds are concerned, Aristotle does not need to modify the Interim List of predications to get the correspondent classification of predicates. whereas when it is predicated as a ‘what it is’, it is only an inconsequential predication. We may say that the permanent categorial nature of the predicate trumps an occasional ‘what it is’ predication and surfaces even when this predication is performed. Only in this way can the reader accept that in the three examples above the same item predicated can signify at the same time the ‘what it is’ (i. e. as engaged in a superficial predication), and another category (i. e. insofar as, however employed in the proposition, the predicate still exhibits its belonging to a kind different from the ‘what it is’). This crucial point is made by Delcomminette 2018, 58 (despite the reservations I express below in Appendix B).Let us take stock and harvest a first conspicuous conclusion from this discussion. The final sentence of the chapter (103b40–104a2) winds up the section of Topics I devoted to the ingredients of dialectical propositions and problems. It does not require discussion. The chapter does not offer two candidates for the doctrine of categories: there is only one candidate and this is the Official List; and this turns out to be identical to the Ousia List. Furthermore, this is a classification of predicates rather than predications.I am aware that not every detail of the proposed interpretation can be backed by independent arguments. For example, when Aristotle mentions the case of items predicated ‘of something else’, I identify this subject as primary substance, whereas the text is not explicit on that. Yet I take the strength of the interpretation to lie in the coherence it lends to Aristotle’s argument. Indeed, it seems to me that previous interpretations do not even see an argument in the chapter. Moreover, this interpretation takes every occurrence of the phrase ‘what it is’, ubiquitous in the chapter, in exactly the same sense.8Three General Background AssumptionsThe interpretation proposed in the previous section rests on three assumptions. To them we now turn.1. Particular substance as the default subject of katēgoriai. This is the claim that Categories 4, Topics I 9, and possibly some other lists in the Aristotelian corpus, classify all the possible predicates of the same subject – an individual substance (typically a human being). So much is admittedly never stated in the texts, but the claim can be backed with strong indirect evidence. Several category labels are indefinite pronouns that can be turned into interrogatives just by a change of accent. Ademollo (forthcoming) notes that editions of Aristotle’s texts waver between interrogative and indefinite accentuations, and argues that the first six labels are all interrogatives: τί ἐστι, πόσον, ποῖον, πρὸς τί, ποῦ. This suggests that Aristotle is interested in all the range of questions that can be asked about the same object. Someone points to Callias or Coriscus and asks ‘what is it?’, ‘what is it like?’, ‘How tall is it?’, ‘Where?’, ‘When?’, etc. The idea that the subject is a particular human being is confirmed by the last two categories being in a position (keisthai) and being in a state (echein), which are highly specific, typically human, and likely to characterize some individual human in particular, rather than humans in general. Incidentally, this anthropocentric perspective may be due only to the circumstance that a human being, compared to another natural object or to an artifact, is (in Aristotle’s world) the most complete entity and enjoys the richest range of features. Another advantage of selecting a focal particular subject is that only by keeping it thus fixed could Aristotle hope to cover the complete range of predicates without overlaps. I am relying on Gillespie’s [1925]/1979 still illuminating exposition of this hypothesis. See also Kapp [1920]/1968, 244.Notice also that Aristotle denies that an accident of a substance can be the subject of another accident, and explains apparent counterexamples as cases where both accidents are predicated of an underlying genuine subject. He might be wrong in that; after all it is perfectly reasonable to ask whether virtue, a quality of substance, is itself teachable. But this is what Aristotle says. APo. I 22, 83a36–39; Met. Γ 4, 1007b2–5. Essential predicates can be an exception, but not in APo. I 22. This thought too tells against the possibility that the list of the categories could be established without any restriction concerning the nature of their subject. If that were the case, how could Aristotle rule out qualities of qualities and quantities of quantities?This approach to the list of katēgoriai is famously prefigured in Plato’s Sophist:We speak of a human being for example, and apply many names to him, attributing colours to him and figures and sizes and vices and virtues, and in these and thousands of other ways we say not only that he is a human being but also that he is good, and an infinity of other things; and in other instances too on the same principle we thus posit each thing as one and again call it many things and by many names. (Sophist 251a8–b3)This interpretative hypothesis about the origins of the list is part of traditional wisdom and has been repeated by generations of commentators. Here is Ockham, invoking the authority of Averroës: “And the distinction made by the categories is drawn from the distinction of questions about substance or about an individual substance, just as the Commentator suggests (Metaphysics VII). Hence, as the answer is given to diverse questions about substance through diverse simple terms, so those simple terms are organised under diverse categories” (Ockham, Summa Logicae, I 41, as quoted and translated by Kahn). The idea that a substantial subject is the focus of all katēgoriai is sometimes taken as a rigorous criterion for collecting all the categories. As such it has been found wanting especially because it cannot explain why particular substance is itself on the list, as it is in the Categories. See especially Malcolm 1981, and Frede [1981]/1987a, 38, who objects that the hypothesis is falsified by Topics VI 6, 144a17–18, quoted above, footnote 31. But, as I have said, the passage in question is just a topos and must be handled with care. However, there is no indication that this method was taken by Aristotle as an ambitious criterion. It seems rather an effective but very modest empirical way of surveying the range of possible predicates. Granieri 2016, 5 n17. Kahn 1978, 240, correctly calls it the ‘stage zero’ of the doctrine of categories, and points out that ‘as a classification of predications or predicates, the scheme does not initially apply to the subjects of discourse as such. A concrete subject is taken as given; predication in the first category simply specifies what it is’. It only matters that Aristotle and his audience should take for granted that the katēgoriai were supposed to refer to a particular substantial subject. It might be suggested that in Top. I 9 the default subject is not the primary but the secondary substance. This might seem an attractive proposal, because the default subject of the four predicables is the species; so, to simplify the comparison of the two systems, Aristotle may assume that also the subject of the categories is the substance species. After all, in the Categories Aristotle accepts that secondary substances too are subjects (5, 2b37–3a6), and according to several interpreters they constitute a genuine class of objects even though they are universal (Frede [1978]/1987b, 56 n3; Perin 2007, 140; Kohl 2008). However, in that very passage of the Categories Aristotle makes clear that, as a subject, the secondary substance plays second fiddle. The system of categories cannot rest on a penultimate subject.2. Predications and predicates. The distinction introduced by Frede is by now sufficiently familiar, but let us make it more precise. I have already noted that predications always take place in a proposition, or by means of a proposition, and involve a specific subject. However, predications are not propositions. This is ruled out by the last clause of Top. I 9 (and cf. 4, 101b26–28: “No one should take us to mean that each one of these, uttered by itself, is a proposition or a problem, but instead that it is out of these that propositions and problems arise”). Kinds of predicates can correspond to kinds of predication, but a kind of predication does not automatically generate a kind of predicate. This only happens when the predication is relevantly characteristic of the item predicated with respect to all possible subjects, or to some very typical subject. Let us begin to clarify this point with a very simple example. In languages like Greek, Latin or Italian, adjectives concord in gender with the noun of which they are predicated. So we may say that: Giovanna è pazza is a feminine predication. However, it would be rash to conclude that ‘pazza’ is a feminine predicate. This is because: Giorgio è pazzo features the same predicate, but now engaged in a masculine predication. So, with adjectives it makes little classificatory sense to introduce a gendered predicate in correspondence with the gender of the predication: you don’t need two entries in the dictionary, one for ‘pazza’ and another for ‘pazzo’. Conversely, for the most part, nouns retaining their gender across all their predications and with the whole range of their subjects can usefully be classified as masculine or feminine per se. Thus for example ‘man’ is a masculine predicate, i. e. a masculine noun so classified in dictionaries. These grammatical examples are merely illustrative of the crucial distinction between predication and predicate. In presentations of this paper, I noted that these examples are particularly helpful, because the morphological features displayed in a predication but not inherited by the citational form in isolation are easy to identify. The comparison, however, is not liable to the charge of anachronism. Although in Top. I 9 Aristotle does not apply the distinction to grammatical features like gender and number, his grammatical classifications would enable him to do so. He defines nouns and their indirect cases in Poet. 20, 1457a12–23, and in De Int. 2, and even though he lacks a slot for adjectives, he could describe the grammatical behaviour of an adjective through his notion of a paronymous application of a noun (ὄνομα) involving a modification of its ending (πτῶσις) in concordance with a grammatical subject (Cat. 1, 1a12–15); see also the discussion of Cat. 5, 3b10–23, below. Grammatical accidents such as gender and number may or may not be relevant from an ontological point of view, but this should not detain us in the present discussion. A reader interested in exploring this possibility may start from SE 4, 166b10–16, where Aristotle broadens his list of categories to include masculine, feminine and neutral entities, so to explain how grammatical gender can induce a category mistake in connection with sex and gender roles. Indeed in SE 14, 174a5–9, he notes that a wrong concordance is just a solecism (normally an inconsequential transgression of the rules of language) when it concerns words, but it becomes a fallacy of ‘form of expression’ (i. e. a category mistake: a more serious confusion) if it induces an error regarding the gender of the objects signified. See Fait 2007, 171, and Di Lascio 2007, 202.Think now, for an intermediate case, of the traditional gender of certain nouns for occupations in Italian. A noun like ‘ministro’ is still registered in some Italian dictionaries as masculine, despite the fact that both morphologically and in current usage the substantive can concord with the subject as in ‘Jacinda Ardern è la prima ministra della Nuova Zelanda’. One could only insist on the validity of this lexicographical classification (in fact, we like to think it to be only a vestige) on the assumption that the ‘proper’ gender of the predicate ‘ministro’ is determined by certain special predications, i. e. predications in which the subject stands for a man. Thus, although gender concord is grammatically possible, the grammatical gender of a predicate would be taken as fixed in virtue of the fact that a gendered predication is perceived as the typical or the only ‘right’ case. Mutatis mutandis, this is also Aristotle’s attitude when he selects a certain kind of predication as determining the corresponding kind of predicate. As we have described it in Section 7, Aristotle’s approach shows a clear analogy with the sexist classification of ‘ministro’. Starting from the classification of ‘ministro’ as masculine, one might even legislate on linguistic usage and prescribe the masculine with a feminine subject too: ‘Jacinda Ardern è il primo ministro’ (also still very frequent in Italian). For an interestingly similar attitude in Aristotle, see Cat. 5, 3a28, where he seems to avoid using the masculine required by gender concord, because the neuter is metaphysically more appropriate and reveals the nature of the predicate. He uses a sentence like ‘Human [ἄνθρωπος] is footed [πεζόν]’, in which the neuter predication does not agree with the masculine subject. Although this lack of agreement is not uncommon, in this case it probably has a reason: the subject implicit in ‘footed’ is a neuter ‘thing’ or ‘object’ (see EN IX 9, 1169b18–19, for another example). Perhaps one may even venture to say that the predicate is meant to concord with the neuter genus word ‘animal’ (ζῷoν), and conclude that the differentia always brings about the genus, as suggested by Irwin 1988, 64 f., and 508 n46 (cf. also Ackrill 1963, 86). At any rate, Aristotle’s grammatical choice helps to explain why a differentia (like ‘footed’) is more akin to a noun than to an adjective and, unlike genuine adjectival or paronymous predicates, is ‘said of’ the subject, i. e. predicated in its essence (Cat. 5, 3a22–28). We may say that for ‘footed’ to be neuter is an important, indeed indelible, feature of the predicate; one that does not change depending on the different predications in which the predicate is involved. For a somewhat similar case see APr. I 2, 25a1–10: πᾶσα ἡδονὴ ἀγαθόν. Here the adjective, in the neuter, is not made to agree with the feminine subject, and the context suggests that this is in order for the two terms of a proposition to be treated as independent items capable of exchanging syntactic role and taking a suitable quantification. Indeed, Aristotle is here explaining term conversion, i. e. the family of immediate inferences obtained by swapping the subject and the predicate and adjusting the quantity of the proposition. For instance, from the proposition ‘Every pleasure is (a) good (thing)’ one can infer ‘Some good (thing) is a pleasure’. Again, the gender becomes a permanent characteristic of the predicate. – Thanks to Paolo Crivelli for pointing out this latter passage to me, and to an anonymous reader for helping me to sharpen up this and the previous footnote.Aristotle’s awareness of this difference is especially keen with the four predicables; and most explicit when he recognizes temporary and relative propria (ποτὲ ἢ πρός τι ἴδιον, Top. I 5, 102a25–30; and V 1). For instance, being biped for a human can be a proprium in relation to a horse or a dog. But absolutely speaking, biped is not a proprium of human. Some scholars think that the classification of a predicate under a predicable is always relative to a subject, but the distinction made by Aristotle between absolute and relative propria shows that every predicate, by nature and absolutely, belongs only to one of the predicables. Furthermore, the attribution of a predicate to its predicable does not prescind from any subject. As I said, I follow Alexander in taking the species as the default subject-kind (typically referred to as the πρᾶγμα) of the four predicables (see footnote 12 above). Accident too seems to be defined with reference to a subject-kind (Top. I 5, 102b5), although it is not straightforwardly assessed with respect to the kind, but at the level of the particulars under the kind. An accident is what can belong and not belong to ‘any one and the same thing whatever’ (Top. I 5, 102b6), and of course it is only by considering a particular that you can see whether a subject remains one and the same when acquiring or losing an accidental property; for the subject-kind itself can well be F and not F merely because some of the particulars under it are F and others are not F. For a similar assessment of the possibility of belonging and not belonging see Cat. 5, 4a10–21, where substance, i. e. not just particular substance but substance in general, is claimed to be one and the same while receptive of contraries because (as the examples show: 4a18–19) the particular substances are. Rather, it is on display when the predicate is attributed to its characteristic and representative subject (or one of such subjects). This is straightforward when the predicables are taken in their exclusive sense (for the exclusive vs. inclusive interpretation of the predicables see Brunschwig 1967, LXXVI–LXXXIII). So they are defined in Top. I 5 and in the following chapters, where Aristotle argues at length that one item belongs to exactly one of the predicables. In Top. II–VII, the accident and the proprium are sometimes best interpreted in an inclusive sense whereby the definition of a subject is also a proprium of the same subject and the genus is also an accident. In such cases, whenever a single item also exemplifies a more restrictive predicable (genus, definition), the more inclusive predicable (accident, proprium) probably applies only in a looser sense.Putting the distinction predication/predicate to work on another Aristotelian example (Top. I 5, 102a22–26), we may say that although being asleep could be a proprium of human in certain predications (if only humans are asleep), as a predicate ‘being asleep’ is an accident, because the predication whereby it is attributed to its representative subjects qualifies the predicate as such and permanently. Relative and temporary predicables are thus just kinds of predications, whereas only absolute predicables form a genuine class of predicates. It is misleading then to claim that the difference between the four predicables and the ten categories in Topics I 9 is a difference between ‘predicated predicates’ (= predications) and ‘predicable predicates’ (= predicates). So Ebert 1985, 117–124, followed by Malink 2007, 279. In fact, both schemes, the predicables and the categories, classify predicates, not just predications. Only in some circumstances are they mere predications.3. Signifying a category. When dealing with the categories, Aristotle systematically uses the verb sēmainō. Malink 2007, 277. He does not say that X is a substance or a quality, but that X signifies a substance or a quality. It is tempting to take the grammatical subject of the verb as a linguistic entity. For example, where Aristotle says:Of things said without interweaving, each signifies either substance or quantity or quality […] To give an idea, examples of substance are: human, horse […]. (Cat. 4, 1b25–29)He uses the verb sēmainō to indicate that a certain linguistic expression (for things said without interweaving are basic constituent parts of declarative sentences: Cat. 2, 1a16–19; 4, 2a4–10) signifies something that happens to belong to a category. E.g. ‘horse’ signifies horses and we know that horses are substances, so ‘horse’ also signifies a substance. However, if this were the only relevant sense of sēmainō, linguistic considerations would be largely irrelevant to the determination of the categories: Aristotle could simply say that a horse is a substance. It is easy to see that this use of the verb is not the rule, because very often, in using sēmainō, Aristotle does not suppose that we can assign a category to a linguistic expression just by inspecting and classifying the object denoted by this expression. Rather, the signification more often suggested by the verb sēmainō depends on phenomena that we may broadly speaking describe as the grammatical behaviour of an expression: kinds of word used, inflections, manners of predication etc. And what is signified, the object of the verb sēmainō, is what a linguistic item shows through some morphological, syntactical or semantical feature.As for the subject of the verb sēmainō, it is very important to register a complication. If occasionally Aristotle makes explicit that the grammatical subject of ‘signifies’ is a linguistic entity, See for instance SE 22, 178b37–39: “But ‘human’ [τὸ ἄνθρωπος], and indeed every common [predicate], signifies not a this but some such-like or a so-much or a relative to something, or something of that sort”, where the neuter article indicates that the word is mentioned rather than used. in many cases he expresses himself as if an entity itself signified a category straightaway. However, it is not difficult to see that language still makes an important contribution.A crucial passage of the Categories will illustrate what I have just stated in general terms:Every substance seems to signify a this. As regards the primary substances, it is indisputably true that each of them signifies a this; for the thing indicated is indivisible and numerically one. But as regards the secondary substances, though it appears from the manner of denomination – when one says ‘human’ or ‘animal’ – that a secondary substance likewise signifies a this, this is not really true; rather, it signifies a quality – for the subject is not one, as the primary substance is, but ‘human’ and ‘animal’ are said of many things. However, it does not signify simply a quality, as ‘white’ does. ‘White’ signifies nothing but a quality, whereas the species and the genus mark off the quality regarding substance – for they signify a substance of a certain quality [τὸ δὲ εἶδος καὶ τὸ γένος περὶ οὐσίαν τὸ ποιὸν ἀφορίζει, – ποιὰν γάρ τινα οὐσίαν σημαίνει]. One draws a wider boundary with the genus than with the species, for in saying ‘animal’ one includes more than in saying ‘human’. (Cat. 5, 3b10–23)Primary substances uncontroversially signify a this because they are introduced by words – almost certainly Aristotle has in mind proper names – that signify items (‘the thing indicated’) which are indivisible and one. Aristotle talks about entities but is acutely interested in the manners in which they are expressed in language. This is confirmed throughout the passage: the reason why secondary substances, the species and genera of substances, appear to signify a this is their linguistic form or ‘the manner of denomination’. This error, i. e. taking every predicate to signify a this is the most important cause of the fallacy of ‘form of expression’ (SE 7, 169a30–36; 8, 170a15; 22, 178a4–8). It is most notably committed in the ‘third man argument’ (SE 22, 178b35–179a10). See Fait, forthcoming. The reason why they rather signify a quality is that they are ‘said of many things’. And the reason why such items signify a special quality is that they circumscribe a kind or class of individual substances.Signification is here a certain exhibition of a categorial feature which depends on the grammatical behaviour of the current linguistic expression of an entity: proper names, which characteristically introduce primary substances, can only stand for items that are individual and numerically one, while common nouns, which characteristically introduce species and genera of substances, can be predicated of several subjects. This is why primary substances signify a this while secondary substances signify a quality. Finally, species and genera of substances are not plain qualities, because unlike simple qualities they demarcate a class of objects. How does language work in this case? This is a difficult and potentially controversial point, but it will suffice here to say that Aristotle is somewhat envisioning the distinction between an adjective like ‘white’ and a substantive like ‘horse’. One may object that ‘white’ marks off a class as well, but it will be admitted that to do so it has to lean predicatively on another term like ‘object’. ‘Horse’, on the other hand, marks off its class as an independent term, and can be the genuine subject of a predication. ‘Horse’ is in this respect akin to the proper names of primary substances: although ‘Bucephalus’ denotes one single object and ‘horse’ a plurality of objects, they are both substantives.Let us now bring this all too sketchy analysis of the categorial use of the verb sēmainō to bear on our chapter of the Topics.First, even if Aristotle aims to discover the ultimate nature of real-world entities, we should not expect that when he says that a certain item signifies a substance or a quality, he automatically means that it is one of these categories. For the logico-linguistic enquiry to which the doctrine of categories belongs does not attempt to inspect the nature of things straightforwardly, but rather relies on indirect methods of conceptual analysis which examine what is signified, i.e the clues offered by the linguistic clothing of an item. This kind of consideration seems to me to rule out an influential reading of Topics I 9, 103b27–35, according to which the action of signifying described in sentences like ‘he states what it is and signifies a substance/quality/quantity’ is the completion of a process whereby, starting from any item under consideration, we can iterate the question ‘what is it?’ until we reach this item’s category. What is this? A human being. What is a human being? An animal. What is an animal? A living body. Until you give the final answer: a substance, thereby revealing the category. And likewise, what is that? Crimson. What is crimson? A kind of red. What is red? Etc. Several interpreters take our chapter to describe this method, Most notably Ackrill 1963, 79: “Repeating the same question with reference to the species, genus, or higher genus mentioned in answer to the first question, and continuing thus, we shall reach some extremely high genera”; Malcolm 1981; Delcomminette 2018, 55. Notice that by this method you will never tell a primary from a secondary substance. Indeed Socrates is an animal in the same way in which the species human is an animal; hence Socrates is a substance in the same way in which the species human is a substance. Thus, the fact that Socrates is a this is not expressed by any defining kind or feature, but only signified by the item itself. but a) this iteration of the ‘what is it’ question is not borne out by the text; and b) the context shows that we do not learn the category of an item by broadening its definition, Establishing the category of the definiendum is normally a preliminary step towards definition, not the final stage (see De An. I 1, 412a15–21, with Menn 1995, 328; APo. II 13, 96b15–21). Perhaps the process of inductive generalization envisaged at APo. II 19 might be thought to be an objection to this claim. But that process seems rather an idealization: a human being, using perception as the sole discriminative power, might be able to grasp all the kinds of increasing generality; and the process would culminate in the acquisition of summa genera (cf. in particular 100b1–5). This probably vindicates the theoretical possibility of grasping all the concepts without the help of a special mental power, but is not necessarily meant by Aristotle to reflect the order of real-life learning, which can take many shortcuts simply by exploiting information embodied in human language. but by some manner of exhibition which Aristotle alludes to by the verb sēmainō.In Topics I 9, the grammatical subject of sēmainō is in some cases a katēgoria, i. e. a predication or a predicate, but interpreters agree that the masculine article rather indicates a human being at b27 and consequently at b28 and b31. This person signifies (the) what it is by saying what a certain item is. For instance, by saying that human is human or an animal. Here we have an act of intentional signification. By the same speech act, however, the speaker also signifies a substance (b28, b31), although in this case this is not by performing the kind of predication she has performed, but rather as the consequence of employing a certain predicate.Malink denies that a person who says (verb legō) what something is (e. g. ‘human is human’ or ‘human is an animal’) and one who ‘signifies (the) what it is’ (b27–28; b37) Or ‘the what it is’: the addition (b27) or omission (b37) of the definite article is not significant: Malink 2007, 277. can be doing the same thing. This is because he thinks that ‘saying what it is’ is not a claim about the categories but only about the two essential predicables, genus and definition, whereas the verb ‘to signify’ always indicates the attribution of a category. Malink 2007, 277, followed by Granieri 2016, 15. On the contrary, in Top. I 4, 101b20–22; 5, 101b38; 8, 103b9–12, the direct complement of a relevantly similar use of σημαίνω is τὸ τί ἦν εἶναι; hence not a category but the defining feature of a predicable. But what I have argued about Aristotle’s use of ‘to signify’ should persuade us that he may well use the verb ‘to signify’ just to say that this person’s speech act exhibits the kind of predication whereby we say what something is. It can be the most ephemeral exhibition, but it arguably plays a role in Aristotle’s argument. And surely, signifying something in this way is not a definitive categorial criterion insofar as it does not need to be a display of the ultimate ontological category. We have just seen, for example, that in the Categories to say that the species and the genera of substances signify a quality does not mean that they belong to the category of qualities.This also confirms that there is no distinction between ‘what it is’ in the sense of the predicables and in the sense of the categories. In sketching out my interpretation, in the previous section, I took for granted that in different ways a sentence and, indirectly, a person can signify a kind of predication (‘what it is’) and a kind of predicate (substance, or quality, etc.). I hope these remarks suffice to show that it was a reasonable move to make.9From Predicates to CategoriesIf the proposed interpretation is correct, non-substantial predicates are defined by their most characteristic and representative kind of predication, which is the manner in which they are predicated of a particular substance. Most certainly Aristotle believes that this predication reveals their metaphysical nature and thus can be regarded as the katēgoria, i. e. the ontological category. On the other hand, in the case of non-substantial items ‘what it is’ predication fails to provide any ontologically useful information and yields no category. Conversely, substance predicates can only be predicated in the ‘what it is’. But it is easy to see that, in their case, this kind of predication does reveal their fundamental ontological condition; hence ‘what it is’ provides a criterion for establishing an ontological category.At the same time, if we keep in mind the special role of predications signifying the ‘what it is’ in the case of substance, we can readily see why in Metaphysics Z 4 Aristotle can claim that substance is a ‘what it is’ in a primary way. Only when substantial predicates are said of their subject does the ‘what it is’ predication display its full force. The other cases are presumably derived by analogy. Analogically, everything can be isolated and defined, but when we single out a genuine this (a tode ti) and articulate its true essence, we make an ontologically basic claim.Who decides that a kind of predication is more interesting and relevant than another? Different kinds of predication, as well as other logical and linguistic features, are interesting starting-points, but the metaphysician’s critical assessment has always the last word. No linguistic or logical criterion can automatically yield a category. The definition of a category must be worked out by serious analytical work. Maybe some form of intuition or insight is also involved, but Aristotle does not expand on that.Frede claimed that at b25 katēgoria occurs in its technical sense. Indeed, if in the phrase genē tōn kategoriōn the word katēgoria simply means ‘predicate’ or ‘predication’, when the word is used alone in the precise sense of ‘kind of predicate/predication’, as in b25, it must have already acquired the status of a technical term. Frede thought that in the technical sense the term means ‘kind of predication’, but we have shown that, since at b25 katēgoria refers back to the Official List, it must signify ‘kind of predicate’, as it does in the Ousia List (b27), where it certainly means ‘kind of predicate’.In a sense, however, we are prepared to accept one of Frede’s conclusions: since primary substance is not a predicate, it is not a category in the technical sense. It cannot be included in the catalogue, unless it is clear that we are not classifying categories in the strict sense. So, in order to include primary substance, the Categories collects the ten kinds not as kinds of predicate but as the kinds of ‘things said without interweaving’, where such legomena are the parts of declaratives sentences. But this of course includes the subjects along with the predicates of such sentences. Moreover, in Categories 2, 1b3–6, Aristotle includes a kind of being that is not predicated in either of the two forms of metaphysical predication envisaged (essential predication and inherence). Assuming, for the sake of simplicity, that the technical notion of ‘being in a subject’ is tantamount to a form of predication. This kind of being is then fleshed out as the ultimate metaphysical subject, primary substance (Cat. 5, 2a11–b6c). What I have just said about the Categories is in substantial agreement with Ebert 1985, 138. Of course, since in the Topics primary substance is the default subject of such predicates, we cannot say that the classification adopted in this work entirely prescinds from primary substance, even if the latter is not part of the list. But it involves no paradox to say that, as a system, the doctrine of categories hinges on two fundamental ingredients: metaphysical subjects and metaphysical predicates. In order properly to use it, we need two equally fundamental capacities: we must single out objects and distinguish their features. And if in certain applications of the very same system, like those we find in the Topics or the Posterior Analytics, See APo. I 22. As I interpret this chapter, and pace Frede [1981]/1987a, 44 f., ‘what it is’ (or ‘what is in the what it is’: 82a36; 83a21; b5) includes the species, the genus the difference and the definition, and is contextually restricted to substance (notice ‘those in the substance’, 83b15). See Barnes 1994, 176 f. Unlike Top. I 9, this chapter does not contemplate picking out a non-substantial item and asking what it is. In the sort of predication declared in the chapter as canonical or regular (probably with regard to scientific discourse), non-substantial items are never in subject position, because such items can only be said of something else (i. e. of substance); hence they are not genuine subjects. In particular, they are not genuine subject of ‘what it is’ predications. This measure suffices to rule out the doctrine of Platonic forms, or at least to ban it from the domain of Aristotelian science (83a32–35). Indeed, if there is no ekthesis of non-substances, it is impossible to mistake them for substances. The cause of sin is rooted out. Notice finally that at APo. I 24, 85b20 the category referred to as τί (equivalent to τί ἐστι) must mean substance even without contextual restriction. Similar interpretation for EN I 6, 1096a19–20 (τί ἐστι) and 24 (τί), and EE I 8, 1217b27–28 (τί ἐστι) and 30 (οὐσία). In both texts the ‘what it is’ is not contextually, but conventionally, restricted to substance. Aristotle appears to neglect primary substance, this is fully explained by the marginal role primary substance plays in dialectic and in science.Appendix A: On the Grammar of 103b35–39ἕκαστον γὰρ τῶν τοιούτων, ἐάν τε αὐτὸ περὶ αὑτοῦ λέγηται ἐάν τε τὸ γένος περὶ τούτου, τί ἐστι σημαίνει· ὅταν δὲ περὶ ἑτέρου, οὐ τί ἐστι σημαίνει ἀλλὰ ποσὸν ἢ ποιὸν ἤ τινα τῶν ἄλλων κατηγοριῶν.According to Malink ἕκαστον τῶν τοιούτων refers back to the previous examples, human, colour, one-foot-long magnitude, etc. All these examples propose self-predication and genus-predication as ways of saying what the subject is. Since the coordinate sentences ἐάν τε […] ἐάν τε precisely mention these two kinds of predication, it obviously refers back to those cases. Malink thinks that ἕκαστον τῶν τοιούτων must refer to the subject of those previous examples (and similar examples completing the list). Malink agrees that the person who says what it is in b27–35 predicates something of a subject. But since Malink sharply contrasts ‘saying what it is’ and ‘signifying (the) what it is’, and reads the first as the attribution of a predicable and the second as the attribution of a category, he claims that in b35–37 Aristotle switches to the subjects of the predications exemplified in the previous lines. I have criticized the reasons for this move at the end of section 8. It could be the predicate according to the clause ἐάν τε αὐτὸ περὶ αὑτοῦ λέγηται, where the anaphoric αὐτὸ is the predicate, but it must be the subject in the coordinate clause ἐάν τε τὸ γένος περὶ τούτου, because here the pronoun responding to ἕκαστον is in subject position. Other interpreters take ἕκαστον τῶν τοιούτων as an adverbial phrase (‘in each of such cases’), but Malink is right in rejecting this option. As he argues, it is also grammatically preferable that the anaphoric αὐτὸ (b36) pick up the object signified by ἕκαστον τῶν τοιούτων, something that many translations end up avoiding. An alternative, proposed by Ebert 1985, 132, followed by Delcomminette 2018, 57, would be to adopt the reading τὸ αὐτὸ attested by some manuscripts and Boethius’ Latin translation. Of course, preceded by the article, αὐτὸ would not be anaphoric and the whole sentence ἐάν τε τὸ αὐτὸ […] ἐάν τε τὸ γένος would read very smoothly (perhaps suspiciously so). However, I think the explanation I offer in the next paragraph of the main text is an effective defense of the text printed by the editors. So I stick to it. Malink then argues that the ἐάν part provides sufficient criteria for belonging to the first T-category, while the ὅταν part (ὅταν δὲ περὶ ἑτέρου) provides sufficient criteria for belonging to one of the other T-categories. On his reading, ‘each of such items’:1)if it is the subject of itself or of a genus, it signifies the ‘what it is’, i. e., it belongs to essence, the first of the T-categories;2)if it is predicated of something else, it belongs to one of the other T-categories.Notice that, on this reading, each of conditions 1) and 2) is sufficient for assigning an item to a T-category. Moreover, conditions 1) and 2) cannot be satisfied by the same item, otherwise the same item would belong to different T-categories.I think that Malink’s reading of this sentence exaggerates the importance of an untidiness in Aristotle’s formulation that is not as crippling as it may seem.Malink does not do justice to the correspondence between ἐάν τε αὐτὸ περὶ αὑτοῦ λέγηται and ὅταν δὲ περὶ ἑτέρου, which strongly suggests that the same αὐτό must be understood in the second half of the sentence as well; and there the pronoun is clearly in predicate position. Although the intrusion of the clause ἐάν τε τὸ γένος περὶ τούτου, with a swap of subject and predicate, may be felt to upset the balance between the two sentences, I think a simple explanation is available. Aristotle takes the liberty to make this variation on the assumption that genus-predication is just a weaker form of self-predication. Otherwise, if indeed the genus were a genuine case of other-predication, how could we avoid classifying it among the cases of περὶ ἑτέρου predication as well? Aristotle must have thought that moving the same item in subject position to accommodate the predication of the genus would not blur the point that the genus is the same as its subject, albeit in a weaker sense. A paraphrase like ‘if, either as itself or as genus, it is said of itself’ renders what Aristotle must be saying.If we accept that περὶ αὑτοῦ […] περὶ ἑτέρου both refer to the predicate, we can look back at the previous sentence (b27–35): there several times an item is picked out and a speaker says what it is and signifies a C-category. The act of the speaker is an act of attributing a predicate. Which suggests that if in the following gar sentence ἕκαστον τῶν τοιούτων refers to the same items, they must be predicates.On the other hand, if Malink were right in saying that ἕκαστον τῶν τοιούτων refers back to the subjects of the previous sentences picked out for definition (human, white colour, one-foot-long magnitude, etc.), those could only be subjects of self/genus-predication, for these subjects are never paronymous terms and as such are not said of something else. Every tie with the second half of the sentence would be severed. Rapp/Wagner 2004 deliberately separate the second part of the sentence and start a new paragraph with ὅταν δὲ. The new paragraph would briefly complete a long thought on the ‘what it is’ started at b27: δῆλον δʹ ἐξ αὐτῶν and ended at b37: τί ἐστι σημαίνει. Malink does not seem to acknowledge this problem, because he explicitly agrees that in the second part of the gar clause at 103b35–39 the ‘grammatical subject of both the main and the subordinate clause at 103b35-39 is the same as in the first part of the […] gar clause (103b35–37), namely, ἕκαστον τῶν τοιούτων’ (Malink 2007, 287 n54). According to Peramatzis 2016, 104 f., ἕκαστον τῶν τοιούτων refers to ‘what-it-is predicables’, and the sentence partitions them into two classes: those which are capable of self-predication and those which are said of another thing. Peramatzis argues that only substances can belong to the former kind, while in the other cases the predicate can only indicate non-substance categorial beings. In this way he finds in the Topics the primacy of substance established in Met. Z 4, and so reconciles the two works. I think Peramatzis’s interpretation too suffers from unduly separating the two parts of the sentence, especially if, as I am to argue, the sentence is best interpreted as considering what happens to the same item at different times. Although ἐάν and ὅταν can have a conditional as well as a temporal force, the two subordinate sentences starting with ἐάν and ὅταν have been taken by most interpreters to apply to the same item at different times: ‘each of such items, when […] whenever […]’. Indeed, if in our sentence the conjunctions ἐάν and ὅταν introduce different conditions in which the same predicative item can be found, they must be conditions that repeatedly occur over time, rather than a non-temporal condition whose satisfaction would warrant membership of a T-category (as per Malink’s criterion) on the assumption that one item satisfies one and only one condition. The items to be tested do not signify ‘what it is’ once for all, but whenever they are involved in self-predication; and they signify one of the other katēgoriai whenever they are predicated of something else. For this temporal interpretation allows the same item to satisfy different conditions at different times. As I have explained in Section 7, these items can engage in self/genus-predications or in other-predications, depending on the subject they are applied to on different occasions, and can do so while remaining the same because this is a prerogative of non-substantial predicates.Appendix B: On a Recent InterpretationThe interpretation defended in this paper has important points of contact with Delcomminette 2018 (see also Delcomminette 2009). Delcomminette believes that predications are ‘predicative functions’ that can be actualized in an act of predication or integrated in the predicate even when the latter is ‘detached’ and is not performing the predicative function in a proposition (normally the predicative function is incorporated in the copula, but it can be transferred to the nominal part of the predicate). Delcomminette thinks that the ten fundamental predicative functions refer to a given ‘subject’, which he refrains from immediately identifying with primary substance and suspends in inverted commas. However, it then turns out that as soon as this mysterious subject receives all its ‘what it is’ predicates, it reveals itself as a particular substance. Delcomminette’s interpretation of the chapter takes the Official List as a list of predications; these are then integrated in the predicate in lines b27–28. When we say what a non-substantial item is, we make its fundamental predicative function explicit. This is what Aristotle means when he claims that a person says what it is and signifies a substance or a quality or a quantity. To say what the subject is means to reveal its integrated predicative function.So interpreted, the scheme of categories is a system of predicative functions which does not owe anything to any independently given subject (primary substance). Rather, primary substance itself, according to Delcomminette, incorporates a predicative function, because it can be predicated of itself, and thereby determines itself. Here is, I suspect, a serious shortcoming of his proposal: if the subject is not a fundamental being, how can we know that the predicative function of a non-substantial item, when it is predicated of a ‘subject’, is its essential predicative function? It would appear that, until it is revealed as a substance, such a ‘subject’ has no metaphysical credentials. And since a non-substantial item, say a quality, can incorporate two predicative functions, i. e. quality and ‘what it is’, how can we determine which one is the more fundamental? We need, I think, an independent criterion for primary substancehood. Self-determination of primary substance by self-predication also beggars belief. I think Aristotle means what he says when he claims that from primary substance there is no predication (Cat. 5, 3a36–37).One of the reasons why Delcomminette believes that the subject is an indeterminate X that must be entirely determined by predication is that he thinks that in our experience ‘quelque chose comme un homme n’est jamais donné immédiatement à la sensation, qui n’a accès qu’à des couleurs, des sons, des odeurs, etc. L’homme n’apparaît que comme le sujet auquel on attribue les qualités correspondant au sensibles propres, sujet qui n’est pas lui-même perçu par la sensation mais “sup-posé” à titre de substrat’ (Delcomminette 2018, 63). Delcomminette adds that by providing dialectical definitions we refine our brute data of experience until we reach substance. At that point we are at the level of science. Although it is impossible to discuss these difficult issues here, I believe that this picture inverts the correct order. I take it that for Aristotle our experience is first and foremost an experience of objects, and we can single them out by deixis and description well before we can articulate a scientific definition of them. For helpful feedback I thank audiences at workshops in Campinas and Oxford, and at a conference in Vienna. Individuals I wish to thank warmly for their help include: James Allen, Lucas Angioni, Alex Bown, David Charles, Paolo Crivelli, Roberto Granieri, Pieter-Sjoerd Hasper, Doukas Kapantais, George Karamanolis, Marko Malink, Brother Bede Mullens, Michael Peramatzis, Breno Zuppolini, an anonymous reader, and especially Christof Rapp, who acted as discussant at the Vienna Conference. Special thanks go to Francesco Ademollo for exchanging with me countless emails on virtually every problem raised by Topics I 9 at the time when my interpretation of the chapter was taking shape. His constructive criticism has been a great stimulus to improve the paper.Cat.CategoriaeDe Int.De InterpretationeAPr.Analytica PrioraAPo.Analytica PosterioraTop.TopicaSESophistici ElenchiDe An.De AnimaMet.MetaphysicaENEthica NicomacheaEEEthica EudemiaPoet.Poetica
Journal
Archiv für Geschichte der Philosophie – de Gruyter
Published: Mar 31, 2023