20 THE WORLD AS IDEA.
is, however, one other possible origin of this question, quite different from the purely speculative one which we have considered, a specially empirical origin, though the question is always raised from a speculative point of view, and in this form it has a much more comprehensible meaning than it had in the first. We have dreams; may not our whole life be a dream? or more exactly: is there a sure criterion of the distinction between dreams and reality? between phantasms and real objects ? The assertion that what is dreamt is less vivid and distinct than what we actually perceive is not to the point, because no one has ever been able to make a fair comparison of the two; for we can only compare the recollection of a dream with the present reality. Kant answers the question thus: "The connection of ideas among themselves, according to the law of causality, constitutes the difference between real life and dreams." But in dreams, as well as in real life, everything is connected individually at any rate, in accordance with the principle of sufficient reason in all its forms, and this connection is broken only between life and dreams, or between one dream and another. Kant's answer therefore could only run thus: — the long dream (life) has throughout complete connection according to the principle of sufficient reason; it has not this connection, however, with short dreams, although each of these has in itself the same connection: the bridge is therefore broken between the former and the latter, and on this account we distinguish them.
But to institute an inquiry according to this criterion, as to whether something was dreamt or seen, would always be difficult and often impossible. For we are by no means in a position to trace link by link the causal connection between any experienced event and the present moment, but we do not on that account explain it as dreamt. Therefore in real life we do not commonly employ that method of distinguishing between dreams and reality. The only sure criterion by which to dis-
