The analysis, however, is not yet complete. Just as reproduction conditions apprehension and both rest on recognition, so in turn recognition presupposes a still further condition, namely, self-consciousness. For it is obvious, once the fact is pointed out, that the recognition of reproduced images as standing for past experiences can only be possible in so far as there is an abiding self which is conscious of its identity throughout the succession. Such an act of recognition is, indeed, merely one particular form or concrete instance of self-consciousness. The unity of the empirical concept in and through which recognition takes place finds its indispensable correlate in the unity of an empirical self. Thus an analysis of our consciousness, even though conducted wholly on the empirical level, that is, without the least reference to the _a priori_, leads by simple and cogent argument to the conclusion that it is conditioned by complex synthetic processes, and that these syntheses in turn presuppose a unity which finds twofold expression for itself, objectively through a concept and subjectively in self-consciousness. So far I have stated the argument solely in reference to serial consciousness. Kant renders his argument needlessly complex and diminishes its force by at once extending it so as to cover the connected problem, how we become aware of objects. This occurs in the section on the synthesis of reproduction. An analysis of our consciousness of objects, as distinct from consciousness of the immediately successive, forces us to postulate further empirical conditions. Since the reproductive imagination, to whose agency the apprehension of complex unitary existences is psychologically due, acts through the machinery of association, it presupposes constancy in the apprehended manifold. “If cinnabar were sometimes red, sometimes black, sometimes light, sometimes heavy, if a man changed sometimes into this and sometimes into that animal form, if the country on the longest day were sometimes covered with fruits, sometimes with ice and snow, my empirical imagination would never even have occasion when representing red colour to bring to mind heavy cinnabar....”[885] This passage may be compared with the one which occurs in the section on the synthesis of recognition. Our representations, in order to constitute knowledge, must have the unity of some concept; the manifold cannot be apprehended save in so far as this is possible.