subject, a universal its predicate; and they have in common the Aristotelian metaphysics, which differs greatly from the modern logic of subject and predicate. Subject (ὑποκείμενον) originally meant a real thing which is the basis of something, and was used by Aristotle both for a thing to which something belongs and for a name of which another is asserted: accordingly “predicate” (κατηγορούμενον) came with him to mean something really belonging (ὑπάρχον) to a substance as real subject, as well as a name capable of being asserted of a name as a nominal subject. In other words, to him subject meant real as well as nominal subject, and predicate meant real as well as nominal predicate; whereas modern logic has gradually reduced both to the nominal terms of a proposition. Accordingly, when he said that a substance is a subject, he meant a real subject; and when he said that a universal species or genus is a predicate, he meant that it is a real predicate belonging to a real subject, which is always some individual substance of the kind. It follows that Aristotelianism in the Categories and in the Metaphysics is a realism both of individuals and of universals; of individual substances as real subjects, and of universals as real predicates.
Lastly, the two works agree in reducing the Categories to substance and its belongings (ὑπάρχοντα). According to both, it is always some substance, such as Socrates, which is quantitative, qualitative, relative, somewhere, some time, placed, conditioned, active, passive; so that all things in all other categories are attributes which are belongings of substances. There are therefore two kinds of belongings, universals and attributes; and in both cases belonging in the sense of having no being but the being of the substance.
In brief then the common ground of the Categories and the Metaphysics is the fundamental position that all things are substances having belonging to them universals and attributes, which have no separate being as Plato falsely supposed.
This essential agreement suffices to show that the Categories and the Metaphysics are the result of one mind. Nevertheless, there is a deep difference between them in detail, which may be expressed by saying that the Categories is nearer to Platonism. We have seen how anxious Aristotle was to be considered one of the Platonists, how reluctant he was to depart from Plato’s hypothesis of forms, and how, in denying the separability, he retained the Platonic belief in the reality and even in the unity of the universal. We have now to see that, in writing the Categories, on the one hand he carried his differences from his master further than he had done in his early criticisms by insisting that individual substances are not only real, but are the very things which sustain the universal; but on the other hand, he clung to further relics of the Platonic theory, and it is those which differentiate the Categories and the Metaphysics.
In the first place, in the Categories the belonging of things in other categories to individual substances in the first category is not so well developed. A distinction (chap. 2) is drawn between things which are predicates of a subject (καθ᾽ ὑποκείμενον) and things which inhere in a subject (ἐν ὑποκειμένῳ); and, while universals are called predicates of a subject, things in a subordinate category, i.e. attributes such as colour (χρῶμα) in the qualitative, are said to inhere in a subject. It is true that the work gives only a negative definition of the inherent, namely, that it does not inhere as a part and cannot exist apart from that in which it inheres (1 a 24-25), and it admits that what is inherent may sometimes also be a predicate (chap. 5, 2 a 27-34). The commentators explain this to mean that an attribute as individual is inherent, as universal is a predicate. But even so the Categories concludes that everything is either a predicate of, or inherent in, a substance; and the view that this colour belongs to this substance only in the sense of being in it, not of it, leaves the impression that, like a Platonic form, it is an entity rather in than of an individual substance, though even in the Categories Aristotle is careful to deny its separability. The hypothesis of inherence gives an inadequate account of the dependence of an attribute on a substance, and is a kind of half-way house between separation and predication.
On the other hand, in the Metaphysics, the distinction between inherence and predication disappears; and what is more, the relation of an attribute to a substance is regarded as so close that an attribute is merely the substance modified. “The thing itself and the thing affected,” says Aristotle, “are in a way the same; e.g. Socrates and Socrates musical” (Met. Δ 29, 1024 b 30-31). Consequently, all attributes, as well as universals, belong as predicates of individual substances as subjects, according to the Metaphysics, and also according to the most authoritative works of Aristotle, such as the Posterior Analytics, where (cf. i. 4, 22) an attribute (συμβεβηκός) is said to be only by being the substance possessing it, and any separation of an attribute from a substance is held to be entirely a work of human abstraction (ἀφαίρεσις). At this point, Plato and Aristotle have become very far apart: to the master beauty appears to be an independent thing, and really separate, to the pupil at his best only something beautiful, an attribute which is only mentally separable from an individual substance. The first difference then between the Categories and the Metaphysics is in the nature of an attribute; and the theory of inherence in the Categories is nearer to Plato and more rudimentary than the theory of predication in the Metaphysics. The second difference is still nearer to Plato and more rudimentary, and is in the nature of substance. For though both works rest on the reality of individual substances, the Categories (chap. 5) admits that universal species and genera can be called substances, whereas the Metaphysics (Ζ 13) denies that a universal can be a substance at all.
It is evident that in the category of substance, as Aristotle perceived, substance is predicate of substance, e.g. Socrates (οὐσία) is a man (οὐσία), and an animal (οὐσία). The question then arises, what sort of substance can be predicate; and in the Categories Aristotle gave an answer, which would have been impossible, if he had not, under Plato’s influence, accepted both the unity and the substantiality of the universal. What he said in consequence was that the substance in the predicate is not an individual substance, e.g. this man or this animal, because such a primary substance is not a predicate; but that the species man or the genus animal is the substance which is the predicate of Socrates the subject (Cat. 5, 3 a 36 seq.). Finding then that substances are real predicates, and supposing that in that case they must be species or genera, he could not avoid the conclusion that some substances are species or genera, which were therefore called by him “secondary substances,” and by his Latin followers substantiae universales. It is true that this conclusion gave him some misgivings, because he recognized that it is a characteristic of a substance to signify an individual (τόδε τι), which a species or a genus does not signify (ib. 5, 3 b 10-21). Nevertheless, in the Categories, he did not venture to deny that in the category of substance a universal species (e.g. man), or genus (e.g. animal), is itself a substance. On the other hand, in the Metaphysics (Ζ 13), he distinctly denies that any universal can be a substance, on the ground that a substance is a subject, whereas a universal is a predicate and a belonging of a subject, from which it follows as he says that no universal is a substance, and no substance universal. Here again the Categories forms a kind of transition from Platonism to the Metaphysics which is the reverse: to call universals “secondary substances” is half way between Plato’s calling them the only substances and Aristotle’s denial in the Metaphysics that they are substances at all.
What conclusion are we to draw from these differences between the Categories and the Metaphysics? The only logical conclusion is that the Categories, being nearer to Plato on the nature of attributes, and still nearer on the relation of universals to substances, is earlier than the Metaphysics. There are difficulties no doubt in drawing this conclusion; because the Metaphysics, though it denies that universals can be substances, and does not allow species and genera to be called “secondary substances,” nevertheless falls itself into calling a universal essence (τὸ τί ἦν εἶναι) a substance—and that too in the very book where it is proved that no universal can be a substance. But this lapse only shows how powerful a dominion Plato exercised over Aristotle’s soul to the last; for it arises out of the pupil still accepting from his master the unity of the universal though now applying it, not to classes, but to essences. The argument about essences in the Metaphysics is as follows:—Since a separate individual, e.g. Socrates, is a substance, and he is essentially a rational animal, then his essence, being what he is, is a substance; for we cannot affirm that Socrates is a substance and then deny that this rational animal is a substance (Met. Ζ 3). Now, according to the unity of a universal asserted by Plato and accepted by Aristotle, the universal essence of species, being one and the same for all individuals of the kind, is the same as the essence of each individual: e.g. the rational animal in the human species and in Socrates is one and the same; “for the essence is indivisible” (ἄτομον γὰρ τὸ εἶδος, Met. Ζ 8, 1034 a 8). It follows that we must call this selfsame essence, at once individual and universal, substance—a conclusion, however, which Aristotle never drew in so many words, though he continued always to call essence substance, and definition a knowledge of substance.
There is therefore a history of Aristotle’s metaphysical views, corresponding to his gradual method of composition. It is as follows:—
(1) Negative rejection of Plato’s hypothesis of forms and formal numbers, and reduction of forms to the common in the early dialogue περὶ φιλοσοφίας and in the early work περὶ ἰδεῶν.
(2) Positive assertion of the doctrine that things are individual substances in the Categories, but with the admission that attributes sometimes inhere in substance without being predicates of it, and that universal species and genera are “secondary substances.”
(3) Expansion of the doctrine that things are individual substances in the Metaphysics, coupled with the reduction of all attributes to predicates, and the direct denial of universal substances; but nevertheless calling the universal essence of a species of substances substance, because the individual essence of an individual substance really is that substance, and the universal essence of the whole species is supposed to be indivisible and therefore identical with the individual essence of any individual of the species.
2. The De Interpretatione.—Another example of Aristotle’s gradual desertion of Plato is exhibited by the De Interpretatione as compared with the Prior Analytics, and it shows another gradual history in Aristotle’s philosophy, namely, the development of subject, predicate and copula, in his logic.
The short discourse on the expression of thought by language (περὶ Έρμηνείας, De Interpretatione) is based on the Platonic