ference is inductive (deduction being only the interpretation of formulæ originally established by Induction), and is, in its ultimate nature, simply the result of past association. A is a mark of B, because A and B have been conjoined in our experience; that is, the one suggests the other. Reasoning therefore is always from particulars to particulars (so, however, that wherever one inference is warranted, a general inference as to a class resembling that particular in all essential respects, is equally warranted-thus, from the proposition, All our ancestors without exception have died, we infer that the Duke of Wellington will die, but have an equal right to infer that any man―i. e. that all men—now living will die); and hence, "nothing is required to render reasoning possible, except senses and association; senses, to perceive that two facts are conjoined; association, as the law by which one of these two facts raises up the idea of the other."—(Logic, Vol. II. Book iv. c. iii. § 3.) And the canon of probable inference, for no certain inference is, in the first instance, possible, which alone is applicable to all kinds and sorts of reasoning, especially the primary universal inferences, which support all strict laws of belief, is this-" Certain individuals have a given attribute; an individual or individuals resemble the former in certain other attributes; therefore they resemble them also in the given attribute." —(Vol. I. Book ii. c. iii. § 7.) To apply this theory to the establishment of any laws of reasoning, it is requisite to observe, that before any phenomenon a, we usually observe the same antecedent A, so that A becomes to us a sign of a, and when we see A, we infer a;—this was not always universally the case, so that there was a time when no one would have dared to assert that before any change, a, there must have been a uniform antecedent, A, of such a kind that whenever A had been seen, a might be infallibly predicted. But further observation has detected before every phenomenon that can be examined at all, some such unconditional antecedent, or at least traces of it, and if there had been any case where it was not so, we must have known it, because changes in every part of our universe have been observed every day since man was upon the earth, and never was it yet found [how could it have been observed even if it had really happened?] that there was a

phenomenon without such an antecedent as, if repeated, would be again followed by the same event. Hence, in the present state of the world at least, uniform and uncontradicted experience (which must, says Mr. Mill, have been contradicted, were there any experience of the opposite kind in existence) has shown us that where a change has happened, there has been some antecedent which would be invariably and unconditionally followed by that change; [*now any future change or phenomenon resembles all former changes in being a change (i. e. in one respect), therefore it (probably) resembles them also in this other respect, and will have some invariable and unconditional antecedent, i. e. will have a cause ;] hence, the proposition that every change will have a cause (i. e. antecedent invariably and unconditionally followed by the same change), has the highest degree of certainty which human experience can give it. This principle established, Mr. Mill deduces without difficulty his inductive canons; but we must pause to examine somewhat minutely the important step he has made.

The theory of inference here adopted by Mr. Mill, states that the mind is carried on in thought from the evidence, to the thing evidenced, that this is the whole and only operation that if the mind be thus irresistibly carried on, the inference is certain; if only prominently, so that it remain partially open to different successions in its place, the inference is only probable; and that these different degrees of mental connection correspond to different numbers of repetitions in past experiences. What becomes then of Mr. Mill's lofty and reiterated censure of those who "adduce as evidence of the truth of fact in external nature any necessity that the human mind may be conceived to be under of believing it?" The law of association is simply and purely a necessity of the human mind, in reproducing and forcing upon it again, conjunctions which have once or oftener taken place within its observation, and any belief which rests upon this law, which expects a, after seeing A, because A and a have been previously connected in observation, does

*We put this in brackets because it is a link in the argument, apparently not supplied but only implied by Mr. Mill, yet obviously quite necessary to his reasoning.

+ See Vol. I. Book iii. c. iii. and v.; and especially Vol. II. c. xxi.

not "adapt itself to the realities of things," but only to its own inward necessity. Mr. Mill entirely fails to point out any distinction between the law of association and any other laws of mind (proved to be such) justifying us in accepting the one as more objective than any other. A casual reader would be led to suppose, from his mode of writing in these and other places, that Mr. Mill had really dispensed with so ungrounded an hypothesis as laws of mind altogether, and had succeeded in identifying his mind, in every respect, with the outward world and its laws. He might perhaps reply that the superior safety of the law of association lies not in its always ruling its expectations by past facts (which obviously need not necessarily be a good guide to the future), but in its self-verification in generally coming out right. But this must be the case with any law of evidence that could be shown to be even plausible, much more true; there is no question about inventing an imaginary law of evidence, but only about determining the true nature of the existing law: whatever it is, or has been, it must agree with the facts of human inference; and the point now at issue is not whether the associative law or any other law most accurately corresponds with the facts of human belief, but what reason Mr. Mill has for asserting that a law which compels us to expect the future to resemble the past, is, per se, in the least less liable to the charge of adducing subjective necessity as evidence for the truth of an objective fact, than any other law of mind whatever, whether a priori or otherwise. We are not at present arguing against the actual law assumed by Mr. Mill, but only against the unfair attempt to establish it, not by showing its more perfect agreement with the facts of belief, but by arbitrarily asserting, that it alone, of all laws, does not bring any internal necessity as evidence of an external truth. The attempt, so often made, to disclaim any law of belief different from those of external occurrence, could only come to disclaiming any possibility of belief not warranted by actual observation; so that no belief at all could be possible before an event: if the expectation of future facts is to be formed upon the experience of past facts, then this is, in itself, a law of mind, and of mind only, which brings as evidence of an outward fact" a necessity which the human mind is believed to be

under, of believing it ;" and is therefore protested against by Mr. Mill.

We think then that we have proved that we may, or rather must, put faith in the laws of our own mind, whatever these may be, if we are to believe anything about future events at all, even of those of external nature; and that if Mr. Mill's protest is worth anything, it is directed against his own theory of evidence, quite as strongly as against any other. The next remark we have to make upon this rationale of the process of evidence, is, that its whole basis, viz. that every change has (to a moral certainty) an invariable and unconditional antecedent, i. e. one which will, in every case, be followed by the same change, is obviously founded on the theory of probabilities, and is, in truth, simply an application of that theory to a particular case. We have heard it asked why Mr. Mill has not put his chapter on Chances before his chapter on the ground of Induction; and certainly we found it on consideration quite impossible to explain. He would obviously have given much greater unity to his book by so doing, and would have relieved his readers of much of the mystification which we felt, when we were suddenly introduced to the law of universal causation, without a word of explanation about its evidence, except that it was a truth grounded on familiar observation. This of course assumes that observations of the past are (according to their degree of universality and familiarity) grounds of presumption for the future; but nowhere does Mr. Mill explain why this is the case, or even refer it explicitly to the law of association (except incidentally in a chapter of the Fourth Book quoted above, on operations subsidiary to Induction, when all the theory of the process itself is completed), nor does he explicitly refer his doctrine of probabilities to the same law, but only, in general terms, grounds them both on experience (or the absence of all experience).—Surely he ought to have considered and classed together our inferences from uniform or nearly uniform experience (moral certainty), and from occasional experience (moral probability); the law of causation, he says, rests on the one, and we therefore expect an invariable unconditional antecedent for every change with perfect confidence; but that does not alter the principle of our expectation, which is identical with that of a probable

guess; and no one who considers the question for a moment, will deny that Mr. Mill ought to have prefixed to his book on Induction, an analysis of the principle of expectation, and then immediately grounded upon this his doctrine of probability, of analogy, and of certainty; and so made us to understand why he considers the axiom of universal causation to be morally certain. Assuredly to have done so, would materially have altered the whole of his book on Induction. Let him once have stated clearly, that the association of A with a, in fact, is the only ground for our expecting its association in similar cases, and that the strength of such past association is a minimum measure of its probability in every case (though of course no sufficient measure where A can be shown to be related to a, as cause to effect, because then the special probability reaches the same certainty as the general truth that every change is preceded by some invariable and unconditional antecedent), and we should not have had the confusion, of founding the whole doctrine of chances and of causation upon analogy, and then making analogy to depend entirely upon causation. Certainly when we found Mr. Mill stating as the ultimate principle of all inference that because an individual resembles certain other individuals in some one or more ways, therefore it will resemble them also in other ways (which is the principle of analogy), we were not prepared to find him asserting that analogy is worthless, unless we have reason to suspect that the inferred resemblance is connected by causation with the perceived resemblance. Lest our readers should doubt that so clear-headed a writer could have so slighted and reversed his own theory, we quote his own words: he says,—

"We conclude that a fact m, known to be true of A, is more likely to be true of B, if B agrees with A in some of its properties (even though no connection is known to exist between m and these properties), than if no resemblance at all could be traced between B and any other thing known to possess the attribute m. To this argument it is of course [] requisite, that the properties common to A with B shall be merely not known to be connected with m; they must not be properties known to be unconnected with it. If, either by processes of elimination, or by deduction from previous knowledge of the laws of the property in question, it can be concluded that they have nothing to do with m, the argument of analogy is put out of court. The sup