2. Mr. Balfour impugns what he takes to be 'the most ordinary view of scientific philosophy, ... that science, in so far as it consists of a statement of the laws of phenomena, is founded entirely on observation and experiment,' which 'furnish not only the occasions of scientific discovery, but also the sole evidence of scientific truth--evidence, however, which is considered by most men of science not only amply sufficient, but also as good as any which can be well imagined.'[9] In this statement there are obvious laxities, which may serve as openings for idle dispute. No man of science, surely, holds that all statements of the laws of phenomena are equally well 'proved' by observation and experiment. They do hold that such a proposition as that of 'the uniformity of nature,' considered as a 'law of phenomena,' is founded on observation and experiment, as fully as any proposition of natural mode can be. But there is obvious room for ambiguity, again, in the expression 'laws of phenomena.' Let us consider, for instance, 3. Mr. Balfour's contention that the 'law of universal causation' is incapable of proof, and cannot properly be said to be founded on observation and experiment. Here the rationalist may safely grant him his whole case--at least the present writer does. He is right, I submit, in his criticism of Mill's ostensible attempt to prove that the so-called 'law of universal causation' is deduced from observation and experiment. I will further waive the question whether he rebuts the proof offered by Kant for his proposition that 'the judgment of sequence cannot be made without the presupposition of the judgment of causality,' which, like many of Kant's formulas, seems to me very awkwardly phrased. But I advance without hesitation the proposition that all reflection upon events involves the conception of universal causation, and that all reflection upon things involves the conception of them _in eventu_.[10] And this necessary assumption is not as such a product of observation and experiment, though we can never exactly say how far experience may condition[11] our manner of making the assumption. It is quite needless to trace the history of it in human experience, for it is clearly pre-human. If from a tree you fire at and wound a tiger who sees you, he will try to get at you, plainly regarding you as the cause of his wound, though he may never have been shot or seen a shot fired before. The accuracy of his inference is worth noting, though he might chance, of course, to have been wounded by a shot fired by an unseen companion of yours. It may 'reasonably' be 'inferred' (to use terms which Mr. Balfour would probably censure), that man has always obeyed the law of _thought_ thus illustrated; and no number of wrong particular inferences can affect the inevitableness of his assumption that any event has a cause. The _concept_ of cause roots in primary animal habit.