there still remains a residue from the first portion of sensation A, and with this residue the second portion of sensation A coalesces. The mixed state formed by the coalescence of the second portion of sensation A with the residue of the first portion of sensation A is Barratt's idea of resemblance. He apparently calls this mixed state also a sensation of resemblance, but prefers calling it a perception of resemblance. He further describes this perception of resemblance as formed by the combination of two sensations of difference. What are these two sensations? The first of them is the simultaneous excitation of A and B, in which consists the change from sensation A to sensation B. The second of them would seem to be the simultaneous excitation of B and A, in which consists the change from sensation B to sensation A. This second change is said to neutralise the first change.

There are now present the second sensation A, the residue of first sensation A, and also the residue of B; since if the first sensation A can leave a residue, much more can the sequent B. Call the residues a and b. Now the perception of resemblance is formed by the coalescence of A and a, and coalescence means simultaneous excitement (Physical Ethics, p. 334, note). But it is also formed by the combination of A + B and B + A, if we take these expressions to represent the two sensations of difference, as Barratt would apparently take them. The series of sensations ought to be from Barratt's standpoint

But here we find four sensations of difference, if the changes in consciousness are strictly regarded and no favouritism shown to the residues Barratt requires. If we are partial and admit his erroneous plea, we get a series as follows:

A

A+ B

B

B+ A

A + a

where the order is, in Barratt's terminology,

Simple sensation (A).

Mixed sensation of difference.

Simple sensation (B).

Mixed sensation of difference.

Mixed sensation of resemblance.

It seems to me hardly legitimate to speak of this final mixed sensation as formed by the combination of the preceding sensations of difference, and still less legitimate to speak of it both as

formed by that combination and as formed by the coalescence (or simultaneous excitement) of a present sensation and the residue of a preceding one. It is one thing to say that there cannot be a sensation (perception) of resemblance without there having been two sensations of difference. It is another thing to say that the sensation of resemblance is formed by the combination of those two sensations of difference; which two sensations may obviously be conditions, without being the constitution, of the sensation of resemblance.

Again, how can there be a simultaneous excitement of a present sensation and the residue of a preceding one when this residue consists in a weaker action of the same nerve-centres as are stimulated in the case of the sensation itself? This is a doctrine which Barratt accepts: "Idea is thus exactly the same physical and conscious state as its corresponding sensation, but of a less intensity" (p. 334, note). The only meaning, then, we can give to his "simultaneous excitement" here must be that the present sensation is more vivid than it otherwise would have been, and the perception of resemblance is reduced to a sensation of greater vividness. I think Barratt would be unwilling to adopt this position, and, moreover, he adds that the perception of resemblance "arises only from that particular form of change which we call reversal, of which one term is equal and opposite to the other". Barratt's expressions, in short, concerning the origin of the relation in question cannot be made to agree; but it appears that the dominant view in his mind was analogous to Mr. Spencer's, and that his perception of resemblance involved the reversal or neutralisation of one sensation of difference by another, just as Mr. Spencer's relation of likeness involves the neutralisation of one relation of unlikeness by another-of, in the lettering above, z by . But how does this doctrine comport with the view that all knowledge is classification of like to likeis assimilation of feelings to feelings and relations to relations? When Mr. Spencer writes that "a relation of likeness consists of two relations of unlikeness which neutralise each another," we must not suppose the relation of likeness to be constituted by the mere feeling of neutralisation any more than we must suppose it to be constituted by the mere presence of the two relations of unlikeness. The description of these is the description, from an analytic reflective point of view, of the mental processes by which the relation of likeness is disclosed (Prin. of Psych., ii. 283). When we assert that any two primary feelings are alike in kind, "we express an intuition of which we can say nothing further than that we have it. Though, as will by and by be seen, the intuition may be otherwise expressed, it cannot be decomposed" (Ib., p. 280). "That two changes in consciousness are of like kind is a fact of which we can give no account further than that we perceive it to be so. When two transitions in consciousness produce in us two like feelings, we know nothing more than that we have the like

feelings. It is true, as will be shown in a subsequent chapter, that it is possible to say specifically what we mean by asserting the likeness of these feelings. But beyond this it is impossible to go" (ib.). When, therefore, Mr. Spencer speaks of the primordial relation of unlikeness as consisting of two states only, he means that when two states such as he describes are given a relation of unlikeness is established; that when two relations of unlikeness such as he describes are given, a relation of likeness is established. The principle here involved is too frequently forgotten in dealing with mental evolution. Another precisely parallel instance may be given. The relation of coexistence is a relation said to be disclosed by experience; it is a relation between two particular relations of sequence; it is neither the one relation of sequence nor the other nor the mere both; but when these two particular relations of sequence are established, the relation of coexistence is established. To ask why, is to ask why relation should be the form of thought; further interpretation cannot be given: we have come to the unknowable (see Brown, Phil. of Human Mind, Lect. xxxiii., p. 211, and x., p. 61). I can analyse my experience, and I may determine the order of the relations established in my consciousness; I may show that certain relations have arisen for me only after the establishment of certain other relations. To trace the series of relations throughout, from the most complex (i.e., as requiring the previous establishment of other relations) to the most simple, is to exhibit in one of its aspects the process of evolution.. But in no case are the earlier stages to be considered as producing the later ones, any more than the walls of a house are to be considered as producing the roof.

We have now to notice that Barratt's view concerning the nature of feelings of relation resembles the doctrine of "transformed sensations" offered by Condillac in the last century. Barratt holds that "there is nothing in the relation beyond its two members, the change is merely a short simultaneous consciousness of the two sensations" (Physical Ethics, p. 47). Condillac, in Traité des Sensations, writes, pp. 16-17-referring to two sensations, one which we have had, and the other which we have " Nous les apercevons à la fois toutes deux. Apercevoir ou sentir ces deux sensations, c'est la même chose. La mémoire n'est donc que la sensation transformée"; p. 50-"Le jugement, la réflextion, les désirs, les passions, etc., ne sont que la sensation même qui se transforme différemment"; p. 121-"La sensation renferme toutes les facultés de l'âme ". Condillac's doctrine has been ably criticised by that keen but sadly neglected thinker, Dr. Brown, in his Philosophy of the Human Mind, Lect. xxxiii. He urges that Condillac's great error "consists in supposing that, when he has shown the circumstance from which any effect results, he has shown this result to be essentially the same with the circumstance which produced

it," and displays great analytical acumen in exposing the fallacy underlying Condillac's position. Brown's argument is fatal to Barratt's view, as much as to that system, which he otherwise describes as supposing "our comparison to be the ideas compared, and nothing more, as if these had flowed together into one ".

66

"Because two affections of mind are followed by a third, he considers this third to be the two former co-existing, or as he terms it, transformed." "They do not involve or constitute, they merely give occasion to this third state, and give occasion to it, merely in consequence of the peculiar susceptibilities of the mind itself as formed, by its divine author, to be affected in this particular manner, after being affected in those different manners which constitute the separate perceptions, as sensation itself, the primary feeling, was made to depend on some previous organic affection produced by an external object. It is not, therefore, as being susceptible of mere sensation, but as being susceptible of more than mere sensation, that the mind is able to compare its sensations with each other."

Finally, I observe that more than one mistake is made by Prof. James in the article I have mentioned. He appears to think that Mr. Spencer was the first to use the phrase "feeling of relation": he appears to think also that Mr. Spencer has not "seen very deeply into the doctrine". But this doctrine was put forward at least as early as Brown, who uses the very phrase to which Prof. James refers. It is true that Brown's doctrine is much less evolved than Mr. Spencer's, but it is substantially the same in foundation. Reference may be made for various expressions of it to Lectures x., xxxiii., xli., xlv., xlix., 1., li. He speaks continually of the "feelings of relation. The praise, then, which Prof. James bestows upon Mr. Spencer is undeserved. The blame which he bestows upon Mr. Spencer is equally undeserved.

"Mr. Spencer," he says, "tries to reduce the number of relations among things to a minimum; and in other passages says they are limited to likeness and unlikeness, co-existence in space and sequence in time. Whether this be true of real relations, does not here concern us. But it is certainly false to say that our feelings of relation are of only these four kinds.” Now, I am surprised to learn that Mr. Spencer has reduced the number of" real relations" to no less than four, and I am also surprised to learn that he asserts "our feelings of relation" to be "of only these four kinds -as I understand Prof. James to mean the expression. Prof. James cannot have attended to the rest of the chapter in the Pyschology where § 65, to which he makes reference, occurs, or even to the note appended to that section itself, which runs thus:

"It will perhaps be objected that some relations, as those between things which are distant in space or in time, occupy distinguishable portions of consciousness. These, however, are not the simple relations between adjacent feelings which we are here dealing with. They are relations that bridge over great numbers of intervening feelings and relations; and come into existence only by quick transitions through these intervening states, ending in the consolidation of them."

It is rather Prof. James who has not "seen very deeply into the doctrine". The last paragraph of § 73 should justify my statement. See also Part iv. (Special Analysis') throughout, on the varieties of our numerous feelings of relation. Need I do more than ask what is suggested, say, by Mr. Spencer's description of the perception of softness as "the establishment in consciousness of a relation of simultaneity between three series of sensations—a series of increasing sensations of pressure; a series of increasing sensations of tension; and a series of sensations of motion"? Do we find here suggested that Mr. Spencer regards "our feelings of relation" as of only four kinds? Or is it suggested by the statement that "the term Perception is applied to mental states infinitely varied, and even widely different in their natures";-or that "a perception may vary indefinitely in complexity, in degree of directness and in degree of continuity";-or that "in all their various kinds and compounds, what we call relations can be to us nothing more than the modes in which we are affected by bringing together sensations or remembered sensations or both: hence what we have next to do is, first to resolve the special kinds of relations into more general kinds, ending with the primordial kinds; and then to ascertain what are the ultimate phenomena of consciousness which these primordial kinds express"?

Analysis brings us evidently down to the single primordial relation which is a change in consciousness, one aspect of which is the relation of unlikeness and the other aspect a relation of sequence.

MR. F. H. BRADLEY ON FACT AND INFERENCE.

By B. BOSANQUET.

I thought that if there was one doctrine that European philosophy had fairly made its own, it was that of the inferential character of fact. According to Mr. Bradley (Principles of Logic p. 74), "Events past and future, and all things not perceived, exist for us only as ideal constructions connected, by an inference through identity of quality, with the real that appears in present perception". Here we have a clear, though in one point it seems to me an inadequate, statement of the doctrine which I understand to be the basis of modern European thought, and to be in a peculiar sense the inheritance of the English experiential school. Whatever other opinions an English writer may hold, he has seldom from the time of Locke failed to lay stress on the relativity of knowledge, and on the inaccessibility of fact to immediate cognition. Mill in his "Psychological Theories of the External World and of Mind" has pushed this view into extremes. I was therefore unprepared to find, in so advanced a writer as Mr. Bradley, the artificial or manufactured character of fact con