384 F. c. s. SCHILLER: Verbal (?) contradictions of this sort seemed to me to run through all his utterances : on the one hand he was strenuous on behalf of ' pure ' thought, on the other he emphasised the importance of the emotions which sustained the intellectual life. This could not, I thought, be pure inadvertence. So I tried to elucidate it on the analogy of Kant's refusal to regard the ' pure respect for the moral law ' as psychologically classifiable with the other feelings ; but I confess I do not really know what Prof. Taylor can mean. III. We now come to what is theoretically the most important part of Prof. Taylor's paper, and also that most welcome to me as being the very thing I had long asked for, viz. his illustrations of absolutely useless knowledge. I must begin by expressing my gratification that I should have persuaded him of what I had to point out in N.S., 54, p. 238, viz. that there is a good deal of the truistic about pragmatism. Its fundamental principle is almost a truism, because in 99 out of 100 cases of recognised truth the test whereby this is established is pragmatic. But not quite, as Prof. Taylor supposes. For in the 100th case, it turns out that the alleged ' truth ' is no truth at all. Now I had rashly supposed that the bitterness of the absolutists' opposition to the new theory was due to their perception of the deadliness of the pragmatic test in the 100th case, and that they denied its applicability to the 99 in order to preclude its application to the 100th. But now Prof. Taylor has made it clear that they have actually never understood its application to the 99 ! This astonishing fact is revealed by his choice of examples. It comes out best in his second example, which I will therefore take first. It is just the sort of simple and elementary example one might choose to illustrate the working of the pragmatic test of truth ! The impossibility of answering truly the question whether the 100th (or 10,000th) decimal in the evaluation of IT is or is not a 9, splendidly illustrates how impossible it is to predicate truth in abstraction from actual knowing and actual purpose. For the question cannot be answered until the decimal is calculated. Until then no one knows what it is, or rather will turn out to be. And no one will calculate it, until it serves some purpose to do so, and some one therefore interests himself in the calculation. And so until then the truth remains uncertain : there is no ' true ' answer, because there is no actual context in which the question has really been raised. We have merely a number of conflicting possibilities, not even claims to truth, and there is no decision. Yet a decision is possible if an experiment is performed. But this experiment presupposes a desire to know. It will only be made if the point becomes one which it is practically important to decide. Normally no doubt it does not become such, because for the actual purposes of the sciences it makes no difference whether we
