Page:Essays on the Principles of Human Action (1835).djvu/115

 which they were associated. Now however we may reconcile it with the foregoing reasoning, it is certainly a fact that it does do so. And I conceive it will not be difficult to account for this, according to the explanation above hinted at of the principle of association: for we may in general suppose any similar state of mind to be favourable to the re-admission, or recollection of the ideas already associated with such a state of mind, whether the similarity is produced by a revival of the old idea, or by the recurrence of a similar external object. In this case, however, we must suppose that association is only a particular and accidental effect of some more general principle, and not the sole-moving spring in all combinations which take place between our ideas: and still more, that similarity itself must be directly a very strong source of connection between them, since it extends beyond the similar ideas themselves to any ideas associated with them. This conclusion is still more clear on the Hartleian theory of association as carried on by the connection of different local impressions, which alone makes it difficult to admit similarity as a distinct source of connection between our ideas. Now I am utterly unable to conceive how this effect can ever take place, for I contend that there must be in this case a direct communication between the new impression, and the similar old one, before there can be any possible reason for the revival of the associated ideas, and then the same difficulty will return as before, that is, why one similar impression should have a natural tendency to excite another, which tendency cannot be accounted for from association, for it goes before it, although on this hypothesis it is absolutely necessary to account for it.—For evidently whatever relates to local connection must be confined to the individual impression and cannot possibly extend to the class or genus. Suppose association