nothing but the sensations, ideas, clusters,
and trains. The attempt to base all truth upon experience, to bring
philosophy into harmony with science was, as I hold, perfectly right.
Only, upon these assumptions it could not be carried out. Mill had the
merit which is implied even by an unsuccessful attempt to hold by
fact. He raises a number of interesting questions; and I think that it
is more remarkable that so many of his observations have still an
interest for psychologists than that so much is obviously wrong. Mill,
it may be said, took an essay upon association for a treatise upon
psychology in general. He was writing what might be one important
chapter in such a treatise, and supposes that he has written the
whole, and can deduce 'philosophy' from it, if, indeed, any philosophy
can be said to remain. Meanwhile, I may observe, that by pushing his
principles to extremes, even his 'association' doctrine is endangered.
His _Analysis_ seems to destroy even the elements which are needed to
give the simplest laws of association. It is rather difficult to say
what is meant by the 'contiguity,' 'sequence,' and 'resemblance,'
which are the only conditions specified, and which he seems to explain
not as the conditions but as the product of association. J. S. Mill
perceived that something was wanting which he afterwards tried to
supply. I will just indicate one or two points, which may show what
problems the father bequeathed to the son. James Mill, at one place,
discusses the odd problem 'how it happens that all trains of thought
are not the same.'[553] The more obvious question is, on his
hypothesis, how it happens that any two people have the same beliefs,
since the beliefs are made of the most varying materials. If, again,
two ideas when associated remain distinct, we have Hume's difficulty.
Whatever is distinguishable, he argued, is separable. If two ideas
simply lie side by side, as is apparently implied by 'contiguity,' so
that each can be taken apart without change, why should we suppose
that they will never exist apart, or, indeed, that they should ever
again come together? The contiguity does not depend upon them, but
upon some inscrutable collocation, of which we can only say that it
exists now. This is the problem which greatly occupied J. S. Mill.

The 'indissoluble' or 'inseparable' association, which became the
grand arcanum of the school, while intended to answer some of these
difficulties, raises others. Mi