dares in an ignorant and foolhardy manner is not to be called courageous, we must change the definition of courage so as to include the element of knowledge. The judgment that the foolhardy person is not to be called courageous can be taken as more certain than the definition, only because one's power of applying a concept in recognizing or excluding is more certain than one's power to express it in terms of predicates. It must be assumed that one knows what courage is, for the purposes of these recognitions, in order that the dialectical apparatus shall have a fixed ground to operate from. Yes, one must know what courage is, that is, one must actually know the connotation, in order to effect these judgments of denotation. But this knowledge of the essence as an inaccessible knowledge may be called relatively subconscious; one can reach it for purposes of expression only by a succession of these dialectical efforts or experiments.
Now this process, which is applied by Plato chiefly to the task of learning what we think, is also quite spontaneously applied by all of us to the task of learning what we want. For all assertions of the form 'I wish X' may be regarded as essays at definition, namely the definition of a wish in terms of its objects. And all such definitions, which children and others are inclined to put forth with a high sense of dogmatic certainty, are seen in the course of experience to be, in truth, highly hypothetical. They are, in effect, hypothetical interpretations of a wish, which in its completeness remains unknown in quite the same way as the nature of justice or courage is unknown. And the general effect of experience is to lead to revisions of the assumed definition. Not all learning by experience, however, is dialectical in character; indeed the most conspicuous examples are not so, and partly perhaps for this reason this analogy, so far as I know, has not been pointed out in current discussions of the learning process.
For in the common processes of motor learning, in which pleasures and pains, or the 'original satisfiers and annoyers' of which Professor Thorndike speaks, furnish the definitive 'yeses' and 'noes' for our active experiments, the revisions that take place affect not so much our understanding of our wishes as our under-
