Hume's definition of a miracle as a violation of a law of nature, is the best ever given, and it really is as perfect as such a definition can be. It has been carped at by Christian scribblers, and criticised by superior theologians like Mozley. But, to use Mr. Gladstone's phrase, it keeps the field. Even the criticisms of Mill and Huxley leave its merit unimpaired. The ground taken by these is, that to say a miracle is a violation of a law of nature is to prejudge the question, and to rule out all future facts in the interest of a prepossession. Mill, however, allows that a miracle is a violation of a valid induction, and as a law of nature means nothing more it is difficult to understand why he takes any exception to Hume's statement of the case. It is perfectly obvious that Hume's argument is not metaphysical, but practical. He does not discuss the _possibility_ but the _probability_ of miracles. He reduces the dispute to a single point, namely, whether the person who relates a miracle (for to the world at large the question is necessarily one of testimony) is deceived or deceiving, or whether the otherwise universal experience of mankind is to be disbelieved; in other words, whether he or the rest of the world is mistaken. One man may, of course, be right, and all the human race opposed to him wrong, but time will settle the difference between them. That _time_, however, simply means general experience through long ages; and that is precisely the tribunal which Hume s argument appeals to. Quarrelling with Hume's definition is really giving up miracles altogether, for, except as supernatural evidence, they are no more important than shooting stars. The very nature of a miracle, in whatever formula it may be expressed, is superhuman, and having a purpose, it is also supernatural; in other words, it is a special manifestation of divine power for a particular object. Whether, being so, it is a violation, a contravention, or a suspension of the laws of nature, is a mere question about words. We may say that a miracle has three elements. It is first a fact, unaccountable by science; secondly, it requires a conscious agent; and thirdly, it results from the exercise of a power which that agent does not naturally possess. Let us descend to illustration. Huxley takes the following case. Suppose the greatest physiologist in Europe alleged that he had seen a centaur, a fabulous animal, half man and half horse. The presumption would be that he was laboring under hallucination; but if he persisted in the statement he would have to submit to the most rigorous criticism by his scientific colleagues before it could be believed; and everybody would feel sure beforehand that he would never pass through the ordeal successfully. The common experience, and therefore the common sense, of society would be dead against him, and probably he would be refused the honor of examination even by the most fervid believers in ancient miracles.