standards of comparison. As we have seen, the validity of this assumption could never have been known _a priori_; it can only be said to be actually valid where actual use has justified it. At the same time, it is clearly a principle of method to assume the universal applicability of our practical postulates wherever it is to our interest that they should be applicable, as explained in the last paragraph. This is why we rightly assume the applicability of the postulates in spheres where the successful establishment of general uniformities has not hitherto been effected, so long as no positive reason can be shown why they should not apply. We shall find this last reflection suggestive when we come to deal with the ethical difficulties which have been felt about the application of the postulates of Uniformity and Causality to voluntary action. It is of course clear that our reduction of Uniformity to a mere practical postulate does not introduce any element of pure “chance” into the actual order of existing things. “Chance” is a term with more than one meaning, and its ambiguity may easily lead to misapprehensions. Chance may mean (_a_) any sequence for which our actual knowledge cannot assign the ground. In this sense chance, as another name for our own mere ignorance, must of course be recognised by any theory which does not lose sight of the fact of human ignorance and fallibility. Or again, chance may mean (_b_) a sequence of which the ground is partially understood. We may know enough of the ground of the sequence to be able to limit the possibilities to a definite number of alternatives, without knowing enough to say which alternative completely satisfies the conditions in a special case. It is in this sense that we speak of the “chances” of any one of the alternative events as capable of computation, and make the rules for their computation the object of special mathematical elaboration in the so-called “Theory of Probability.”