But the empiricist must deal in some manner with the products of reflection. The atoms of chemistry and the elements of the psychologist are not experienced facts, but still they play a valuable, indispensable role in the technique of the sciences. What is to be done with them? It must be made to appear that they are valid within knowledge, but invalid elsewhere. This leads to a separation of knowing from other modes of experiencing, and the descriptive method is depended upon to maintain the empirical validity of the separation. It has been seen how Dewey's attempt to interpret knowledge led gradually to a distinction between the 'cognitional' and the 'non-cognitional' processes of experience. The completed theory of knowledge depends for its validity upon the distinction thus established between knowing (as reflective thought) and the practical attitudes of life. The concepts, elements, and other apparatus of reflection are employed, it is said, only when there is thinking,--and this is only occasionally. Theory is an instrument to be used in connection with that special activity, reflective thought, the general aim of which is the furtherance of the practical ends of life. One fairly obvious difficulty with this separation of reflection from the other life activities is that the 'direct descriptive method,' as here employed, is itself reflective. How does it come, then, that this particular method achieves such an effective hegemony over the other modes of reflection? The 'descriptive method,' as the method of pure experience, is made to determine or supplant all other methods. It defines the limits and aims of conceptual systems; it marks out the limits, aims, and tests of reflective thought in general. How, it may be asked, does the 'direct descriptive method' escape the limitations which it imposes upon the other forms of reflective thought? It has been seen that in Dewey's view logic is subsidiary to psychology. But psychology (his psychology) results from the application of the 'descriptive method' to experience. The 'descriptive method,' it may be inferred from this, is not subject to logical criticism. On the contrary, it is the basis of all logic. Logic, as the criticism of categories, is confined to the study of the instrumental concepts as functioning within the knowledge experience, and its limits are set by descriptive psychology. There is, apparently, no means by which the 'direct descriptive method' can itself be brought under criticism.