Friday, July 02, 2004

Beattie on Association of Ideas

The doctrine is not peculiar to modern philosophy. Aristotle, speaking of Recollection, or active remembrance, insinuates, with his usual brevity, that the relations, by which we are led from one thought to another, in tracing out, or hunting after (as he calls it) any particular thought which does not immediately occur, are chiefly three, Resemblance, Contrariety, and Contiguity. And this enumeration of the associating principles does not differ, in any thing material, from what is here gven. I reduced them to five, Resemblance, Contrariety, Nearness of Situation, the relation of Cause and Effect, and Custom or Habit. Now the three last may very well be referred to that one which Aristotle calls Contiguity. Nearness of Situation is nothing else. In its influence a Cause may be said to be, because it really is, contiguous to its Effect. And two things or ideas cannot be associated by Custom, so as that the one shall introduce the other into the mind; unless they have, once and again, or once at least, been in company together, or thought of at the same time.

James Beattie, "Essay on Memory and Imagination," in Dissertations Moral and Critical (1783), Of Imagination, chap. 2, sect. 5.

I find this passage fascinating. Beattie's principles of association are essentially Hume's, slightly modified (his biggest change is the addition of contrariety; he has some interesting and, I think, cogent arguments that Hume should have considered contrariety an associating principle, too). It is essentially independent of any Aristotelian thought on recollection, and is proposed for a different phenomenon, but here we have an approximation of the one doctrine to the other - the philosophical equivalent of what Whewell calls a 'Consilience of Inductions', where two different fields 'jump together' in investigation. Like any consilience of inductions, this one suggests that these taxonomies of association-principles are capturing something real and definite. One doesn't see much work on association any more; but the above passage would make, I think, a good Exhibit A for why we should.