THE GENERALISATIONS OF SCIENCE. 89

score of their containing what he terms an "ascertainment- clause ". " It is perfectly accurate," lie says, " to describe all known natural laws as observed uniformities of process : but surely the essence of the law is its uniformity, and not the accidental fact that it has been observed. Science is perpetually adding to the number of discovered laws ; but these laws existed from the time when the operations of nature began, and the mere fact of their discovery does not add a tittle to their validity. In short, ascertainment is necessary to our knowledge of natural laws, but it is not the least necessary to their existence " (p. 564). And, after elaborating his view, he says in conclusion : " If this be so, the case against the ascertainment-clause is made out. If we believe Natural Law to prevail universally, it is incorrect to define it as an order which is limited limited, that is, by the condition of previous observation. If, on the other hand, we desire to restrict its meaning to observed uniformities of process, it is inaccurate to call it Natural Law ; seeing that, ex hypothesi, it does not extend to the whole of nature, but only to that small part of it which has fallen under human observation" (p. 569). Now there is much in Mr. Pearson's paper with which I am glad to find myself in agreement ; but there is perhaps more in which I cannot concur. I am in agreement with him in believ- ing that there is in a so-called Law of Nature something beyond a mere generalisation from experience. But I differ from him as to what that something is ; and I wholly part company with him when he draws a distinction between our knowledge of Natural Laws and their existence. Every Natural Law comprises, besides the generalisation from experience on which it is based, the hypothesis or assumption that it holds good not only in those cases which have been actually observed, but in all cases of like nature under like condi- tions. Laws of Nature are, as I have elsewhere expressed it, 41 well-proven and oft-verified inferences from known facts, and also, as we believe, generalised statements of all the facts of like nature, whether we have observed them and verified the law in their case or not " (Springs of Conduct, p. 70). I therefore fully agree with Mr. Pearson that to restrict the meaning of Natural Law to observed uniformities of process, and to limit it by the condition of previous observation, would be in the highest degree unsatisfactory and unwise. It would totally change the meaning which we attach to the oft-misunderstood term Natural Law. But I do not think that this would justify us in abandoning the " ascertainment-clause": nay, it would rather justify us in adding thereto an 'inference-clause,' at present implied but not expressed. Mr. Pearson would, however, draw a far more fundamental dis- tinction between Natural Law and notation of observed facts than that which I have briefly sketched. He holds that Natural Laws are not merely human products, the result of scientific generalisa- tion and inference, but that they have an independent existence, separate from and holding jurisdiction over the facts, and only