In arguing this dialectic thesis, one must avoid slipping from the logical into the physical point of view. It would be easy, in taking a concrete example to fix one's ideas by, to choose one in which the letter M should stand for a collective noun of some sort, which noun, being related to L by one of its parts and to N by another, would inwardly be two things when it stood outwardly in both relations. Thus, one might say: 'David Hume, who weighed so many stone by his body, influences posterity by his doctrine.' The body and the doctrine are two things, between which our finite minds can discover no real sameness, though the same name covers both of them. And then, one might continue: 'Only an absolute is capable of uniting such a non-identity.' We must, I say, avoid this sort of example; for the dialectic insight, if true at all, must apply to terms and relations universally. It must be true of abstract units as well as of nouns collective; and if we prove it by concrete examples, we must take the simplest, so as to avoid irrelevant material suggestions.
Taken thus in all its generality, the absolutist contention seems to use as its major premise Hume's notion 'that all our distinct perceptions are distinct existences, and that the mind never perceives any real connexion among distinct existences.' Undoubtedly, since we use two phrases in talking first about 'M's relation to L' and then again about 'M's relation to N,' we must be having, or must have had, two distinct perceptions;—and the rest would then seem to follow duly. But the starting-point of the reasoning here seems to be the fact of the two phrases; and this suggests that the argument may be merely verbal. Can it be that the whole dialectic achievement consists in attributing to the experience talked-about a constitution similar to that of the language in which we describe it? Must we assert the objective doubleness of the M merely because we have to name it twice over when we name its two relations?
Candidly, I can think of no other reason than this for the dialectic conclusion![1] for, if we think, not of our words, but of any simple concrete matter which they may be held to signify, the experience itself belies the paradox asserted. We use indeed two separate concepts in analyzing our object, but we know them all the while to be but substitutional, and that the M in L—M and the M in M—N mean (i.e., are capable of leading to and terminating in) one self-same piece, M, of sensible experience. This persistent identity of certain units, or emphases, or points, or objects, or members—call them what you will—of the experience-continuum, is just one of those conjunctive features of it, on which I am obliged to insist so emphatically. For samenesses are parts of experience's indefeasible structure. When I hear a bell-stroke and, as life flows on, its after-image dies away, I still hark back to it as 'that same
[Footnote 1: Technically, it seems classable as a 'fallacy of composition.' A duality, predicable of the two wholes, L—M and M—N, is forthwith predicated of one of their parts, M.]
bell-stroke.' When I see a thing M, with L to the left of it and N to the right of it, I see it as one M; and if you tell me I have had to 'take' it twice, I reply that if I 'took' it a thousand times, I should still see it as a unit.[1] Its unity is aboriginal, just as the multiplicity of my successive takings is aboriginal. It comes unbroken as that M, as a singular which I encounter; they come broken, as those takings, as my plurality of operations. The unity and the separateness are strictly co-ordinate. I do not easily fathom why my opponents should find the separateness so much more easily understandable that they must needs infect the whole of finite experience with it, and relegate the unity (now taken as a bare postulate and no longer as a thing positively perceivable) to the region of the absolute's mysteries. I do not easily fathom this, I say, for the said opponents are above mere verbal quibbling; yet all that I can catch in their talk is the substitution of what is true of certain words for what is true of what they signify. They stay with the words,—not returning to the stream of life whence all the meaning of them came, and which is always ready to reabsorb them.
[Footnote 1: I may perhaps refer here to my Principles of Psychology, vol. i, pp. 459 ff. It really seems 'weird' to have to argue (as I am forced now to do) for the notion that it is one sheet of paper (with its two surfaces and all that lies between) which is both under my pen and on the table while I write—the 'claim' that it is two sheets seems so brazen. Yet I sometimes suspect the absolutists of sincerity!]
IV
For aught this argument proves, then, we may continue to believe that one thing can be known by many knowers. But the denial of one thing in many relations is but one application of a still profounder dialectic difficulty. Man can't be good, said the sophists, for man is man and good is good; and Hegel and Herbart in their day, more recently H. Spir, and most recently and elaborately of all, Mr. Bradley, inform us that a term can logically only be a punctiform unit, and that not one of the conjunctive relations between things, which experience seems to yield, is rationally possible.
Of course, if true, this cuts off radical empiricism without even a shilling. Radical empiricism takes conjunctive relations at their face-value, holding them to be as real as the terms united by them. The world it represents as a collection, some parts of which are conjunctively and others disjunctively related. Two parts, themselves disjoined, may nevertheless hang together by intermediaries with which they are severally connected, and the whole world eventually may hang together similarly, inasmuch as some path of conjunctive transition by which to pass from one of its parts to another may always be discernible. Such determinately various hanging-together may be called concatenated union, to distinguish it from the 'through-and-through' type of union, 'each in all and all in each' (union of total conflux, as one might call it), which monistic systems hold to obtain when things are taken in their absolute reality. In a concatenated world a partial conflux often is experienced. Our concepts and our sensations are confluent; successive states of the same ego, and feelings of the same body are confluent. Where the experience is not of conflux, it may be of conterminousness (things with but one thing between); or of contiguousness (nothing between); or of likeness; or of nearness; or of simultaneousness; or of in-ness; or of on-ness; or of for-ness; or of simple with-ness; or even of mere and-ness, which last relation would make of however disjointed a world otherwise, at any rate for that occasion a universe 'of discourse.' Now Mr. Bradley tells us that none of these relations, as we actually experience them, can possibly be real.[1] My next duty, accordingly, must be to rescue radical empiricism from Mr. Bradley. Fortunately, as it seems to me, his general contention, that the very notion of relation is
[Footnote 1: Here again the reader must beware of slipping from logical into phenomenal considerations. It may well be that we attribute a certain relation falsely, because the circumstances of the case, being complex, have deceived us. At a railway station we may take our own train, and not the one that fills our window, to be moving. We here put motion in the wrong place in the world, but in its original place the motion is a part of reality. What Mr. Bradley means is nothing like this, but rather that such things as motion are nowhere real, and that, even in their aboriginal and empirically incorrigible seats, relations are impossible of comprehension.]