Page:Mind (New Series) Volume 9.djvu/308

 294 G. E. MOORE : events that are thus related and two ideas that have the relation of similarity. In so far as he appeals to this differ- ence he may be taken as allowing that there is an idea of necessary connexion which is not identical with that of constant succession ; and this idea may be that which we assert of a cause and its effect, whether it really does apply to them or not. Only by his denial that this is the case by his assertion that there is nothing in common between the idea of a necessary truth and the idea of a causal connexion does Hume really contribute anything to the question what the latter means. We have it, then, suggested that there are two forms of connexion commonly called necessary, and that there is nothing in common between these two ; and this view seems still to be held by those who oppose a ' real ' to an ' ideal ' necessity. In order to decide whether it be a true view, it will be necessary to discuss at some length each of these two forms of necessity, which are at first sight so different the necessity of necessary truths and the necessity of real causes. Now the line which Kant took in answering Hume was based, in part at least, on a denial that they were so different as Hume had thought. Kant pointed out that truths, which Hume had allowed to be necessary, on the ground that they were analytic, were, like the relation of cause and effect, synthetic. The truths of arithmetic were both synthetic and necessary, and, if Hume had considered this, it would have destroyed his reason for allowing no common element between ideal and real necessity. Kant, however, does still allow that there are such things as analytic truths, and that they are necessary. Though, therefore, he classes together, as having a common element, two forms of necessity, which Hume had separated, he still allows another form, which may or may not be different in meaning from this. He does not decide the question : In what sense are analytic truths necessary? Now, if we take the view that the sense is different from that in which synthetic truths are necessary, there would seem to be two alternatives open. Either (1) it may be said that ' necessary ' here merely means ' analytic ' ; that the two conceptions are identical. In this case it be- comes an analytical truth that analytical truths are neces- sary ; and no exception can be taken to the separation of this meaning of necessity from all others, if only there be any meaning in analytic truths. But, at the same time, this necessity becomes utterly unimportant. It is impossible to draw from it any inferences with regard to the truths that