the authority of a philosopher who perfectly agreed with those who disbelieve in the possibility of unconscious knowledge,' and contemptuously rejected Plato's opinion to the contrary. Nor is this all. The doctrine that reasoning is from particulars to particulars, even when it passes through general propositions, may be rigorously deduced from Aristotle's own admissions. If nothing exists but particulars, and if knowledge is of what exists, then all knowledge is of particulars. Therefore, if the propositions entering into a chain of reasoning are knowledge, they must deal with particulars exclusively. And, quite apart from the later developments of Aristotle's philosophy, we have his express assertion, that all generals are derived from particulars, which is absolutely incompatible with the alleged fact, that 'all knowledge, all thought, rests on universal truths, on general propositions; that all knowledge, whether "deductive" or "inductive," is arrived at by the aid, the indispensable aid, of general propositions.' To Aristotle the basis of knowledge was not 'truths' of any kind, but concepts; and in the last chapter of the Posterior Analytics he has explained how these concepts are derived from senseperceptions without the aid of any 'propositions' whatever.

We are here confronted with an important and much disputed question, Was Aristotle an empiricist? We hold most decidedly that he was, if by empiricist is meant, what alone should be meant-one who believes that the mind neither anticipates anything in the content, nor contributes anything to the form of experience; in other words, who believes knowledge to be the agreement of thought with things imposed by things on thought. We have already shown, when discussing Sir A. Grant's view to the contrary, that Aristotle was in no sense a transcendental idealist. The other half of our position is proved by the chapter in the Posterior Analytics already referred to, the language of which is primâ facie so much in favour of our view that the burden of proof That is, knowledge which has never been actualised.

rests on those who give it another interpretation. Among these, the latest with whom we are acquainted is Zeller. The eminent German historian, after asserting in former editions of his work that Aristotle derived his first principles from the self-contemplation of the Nous, has now, probably in deference to the unanswerable arguments of Kampe, abandoned this position. He still, however, assumes the existence of a rather indefinable à priori element in the Aristotelian noology, on the strength of the following considerations:-In the first place, according to Aristotle, even sense-perception is not a purely passive process, and therefore intellectual cognition. can still less be so (p. 190). But the passages quoted only amount to this, that the passivity of a thing which is raised from possibility to actuality differs from the passivity implied in the destruction of its proper nature; and that the objects of abstract thought come from within, not from without, in the sense that they are presented by the imagination to the reason. The pure empiricist need not deny either position. He would freely admit that to lose one's reason through drunkenness or disease is a quite different sort of operation from being impressed with a new truth; and he would also admit that we generalise not directly from outward experience, but from that highly-abridged and representative experience which memory supplies. Neither process, however, constitutes an anticipation of outward experience or an addition to it. It is from the materialist, not from the empiricist, that Aristotle differs. He believes that the forms under which matter appears are separable from every particular portion of matter, though not from all matter, in the external world; and he believes that a complete separation between them is effected in the single instance of self-conscious reason, which again, in cognising any particular thing is identified with that thing minus its matter. Zeller's next argument is that the cognition of ideas by the Nous is immediate, whereas the process of generalisation from experience described by Aristotle

is extremely indirect. Here Zeller seems to misunderstand the word aμeσos. Aristotle never applies it to knowledge, but only to the objective relations of ideas with one another. Two terms constitute an 'immediate' premise when they are not connected by another term, quite irrespective of the steps by which we come to recognise their conjunction. So with the terms themselves. They are 'immediate' when they cannot be derived from any ulterior principle; when, in short, they are simple and uncaused. Finally, the objection that first principles, being the most certain and necessary of any, cannot be derived from sensible experience, which, dealing only with material objects, must inherit the uncertainty and contingency of matter,-is an objection, not to the empiricist interpretation of Aristotle's philosophy, but to empiricism itself; and it is not allowable to explain away the plain words of an ancient writer in order to reconcile them with assumptions which he nowhere admits. That universality and necessity involve an à priori cognition or an intellectual intuition, is a modern theory unsupported by a single sentence in Aristotle. We quite agree with Zeller when he goes on to say that in Aristotle's psychology 'certain thoughts and notions arise through the action of the object thought about on the thinking mind, just as perception arises through the action of the perceived object on the percipient' (p. 195); but how this differs from the purest empiricism is more than we are able to understand.

It is remarkable that Aristotle, after repeatedly speaking of induction as an ascent from particulars to generals, when he comes to trace the process by which we arrive at the most general notions of any, does not admit the possibility of such a movement in one direction only. The universal and the individual are, according to him, combined in our most elementary sense-impressions, and the business of scientific

It is a mistake to translate vónois, as the Germans do, by Anschauung. The Nous does not intuite ideas, but is converted into and consists of them.

experience is to separate them.

Starting from a middle

Moreover,

point, we work up to indivisible predicates on the one hand and down to indivisible subjects on the other, the final apprehension of both extremes being the office, not of science, but of Nous. This theory is equally true and acute. The perception of individual facts is just as difficult and just as slowly acquired as the conception of ultimate abstractions. the two processes are carried on pari passu, each being only made possible by and through the other. No true notion can be framed without a firm grasp of the particulars from which it is abstracted; no individual object can be studied without analysing it into a group of common predicates, the idiosyncrasy of which-that is, their special combinationdifferentiates it from every other object. What, however, we wish to remark is the illustration incidentally afforded by this striking aperçu of Aristotle's analytical method, which is also the essentially Greek method of thought. We saw that, for our philosopher, syllogism was not the subsumption of a particular case under a general law, but the interpolation of a mean between two extremes; we now see that his induction is not the finding of a law for the particular phenomenon, but its analysis into two elements-one universal and the other individual-a solution of the mean into the extremes. And the distinctive originality of his whole system was to fix two such extremes for the universe—a self-thinking thought in absolute self-identity at one end of the scale, and an absolutely indeterminate matter at the other; by combining which in various proportions he then re-constructed the whole intermediate phenomenal reality. In studying each particular class of facts, he follows the same method. The genus is marked by some characteristic attribute which one species the prerogative species, so to speak-exhibits in its greatest purity, while the others form a graduated scale by variously combining this attribute with its opposite or privation. Hence his theory, since revived by Goethe, that

the colours are so many different mixtures of light and darkness.

It has, until lately, been customary to speak as if all that Aristotle knew about induction was contained in a few scattered passages where it is mentioned under that name in the Analytics. This, no doubt, is true, if by induction we mean simple generalisation. But if we understand by it the philosophy of experimental evidence the analysis of those means by which, in the absence of direct observation, we decide between two conflicting hypotheses-then the Topics must be pronounced as good a discussion on the subject as was compatible with his general theory of knowledge. For he supposes that there are large classes of phenomena, including, among other things, the whole range of human life, which, not being bound by any fixed order, lie outside the scope of scientific demonstration, although capable of being determined with various degrees of probability; and here also what he has in view is not the discovery of laws, but the construction of definitions. These being a matter of opinion, could always be attacked as well as maintained. Thus the constant conflict and balancing of opposite forces, which we have learned to associate with the sublunary sphere, has its logical representative no less than the kindred ideas of uncertainty and vicissitude. And, in connexion with this side of applied logic, Aristotle has also to consider the requirements of those who took part in the public debates on disputed questions, then very common among educated Athenians, and frequently turning on verbal definitions. Hence, while we find many varieties of reasoning suggested, such as Reasoning by Analogy, Disjunctive Reasoning, Hypothetical Reasoning (though without a generalised expression for all its varieties), and, what is most remarkable, three out of Mill's four Experimental Methods,' we do not find that any interesting or

For Analogy, see Top., II., x., sub in.; Disjunction, II., vi., sub in.; Hypothetical Reasoning, II., x., p. 115, a 15; Method of Differences, II.,