All the parts of a picture do not have that relation of mutual causal dependence, which certain parts of the body have, and yet the existence of those which do not have it may be absolutely essential to the value of the whole. The two relations are quite distinct in kind, and we cannot infer the existence of the one from that of the other. It can, therefore, serve no useful purpose to include them both under the same name; and if we are to say that a whole is organic because its parts are (in this sense) ‘means’ to the whole, we must not say that it is organic because its parts are causally dependent on one another.
22. But finally (3) the sense which has been most prominent in recent uses of the term ‘organic whole’ is one whereby it asserts the parts of such a whole have a property which the parts of no whole can possibly have. It is supposed that just as the whole would not be what it is but for the existence of the parts, so the parts would not be what they are but for the existence of the whole; and this is understood to mean not merely that any particular part could not exist unless the others existed too (which is the case where relation (1) exists between the parts), but actually that the part is no distinct object of thought—that the whole, of which it is a part, is in turn a part of it. That this supposition is self-contradictory a very little reflection should be sufficient to shew. We may admit, indeed, that when a particular thing is a part of a whole, it does possess a predicate which it would not otherwise possess—namely that it is a part of the whole. But what cannot be admitted is that this predicate alters the nature or enters into the definition of the thing which has it. When we think of the part itself, we mean just that which we assert, in this case, to have the predicate that it is part of the whole; and the mere assertion that it is a part of the whole involves that it should itself be distinct from that which we assert of it. Otherwise we contradict ourselves since we assert that, not it, but something else—namely it together with that which we assert of it—has the predicate which we assert of it. In short, it is obvious that no part contains analytically the whole to which it belongs, or any other parts of that whole. The relation of part to whole is not the same as that of whole to part; and the very definition
