Criminal law, like all other disciplines, must ask under what conditions and when we are entitled to say “we know.” The answer is far from being perennially identical, though it might have been expected that the conviction of knowledge would be ever united with identical conditions. The strange and significant difference is determined by the question whether the verdict, “we know,” will or will not have practical consequences. When we discuss some question like the place of a certain battle, the temperature of the moon, or the appearance of a certain animal in the Pliocene, we first assume that there is a true answer; reasons for and against will appear, the former increase in number, and suddenly we discover in some book the assurance that, “We know the fact.” That assurance passes into so and so many other books; and if it is untrue, no essential harm can be done. But when science is trying to determine the quality of some substance, the therapeutic efficiency of some poison, the possibilities of some medium of communication, the applicability of some great national economic principle like free trade, then it takes much more time to announce, “We know that this is so and not otherwise.” In this case one sees clearly that tremendous consequences follow on the practical interpretation of “we know,” and therefore there is in these cases quite a different taxation of knowledge from that in cases where the practical consequences are comparatively negligible. Our work is obviously one of concrete practical consequences. It contains, moreover, conditions that make imperfect knowledge equivalent to complete ignorance, for in delivering sentence every “no” may each time mean, “We know that he has not done it” or again, “We know that it is not altogether certain that he has done it.” Our knowledge in such cases is limited to the recognition of the confusion of the subject, and knowledge in its widest sense is the consciousness of some definite content; in this case, confusion. Here, as everywhere, knowledge is not identical with truth; knowledge is only subjective truth. Whoever knows, has reasons for considering things true and none against so considering them. Here, he is entitled to assume that all who recognize his knowledge will justify it. But, when even everybody justifies his knowledge, it can be justified only in its immediacy; to-morrow the whole affair may look different. For this reason we criminalists assert much less than other investigators that we seek the truth; if we presume to such an assertion, we should not have the institutions of equity, revision, and, in criminal procedure, retrial. Our knowledge, when named modestly, is only the innermost conviction that some matter is so and so according to human capacity, and “such and such a condition of things.” Parenthetically, we agree that “such and such a condition of things” may alter with every instant and we declare ourselves ready to study the