But suppose that the realist accepts the traditionary psychology according to which every event in the way of a perception is also a case of knowing something. Is the way out now so simple? In the case of the doubled fingers or the seen light, the thing known in perception contrasts with the physical source and cause of the knowledge. There _is_ a numerical duplicity. Moreover the thing known by perception is by this hypothesis in relation to a knower, while the physical cause is not. Is not the most plausible account of the difference between the physical cause of the perceptive knowledge and what the latter presents precisely this latter difference--namely, presentation to a knower? If perception is a case of knowing, it must be a case of knowing the star; but since the "real" star is not known in the perception, the knowledge relation must somehow have changed the "object" into a "content." Thus when the realist conceives the perceptual occurrence as an intrinsic case of knowledge or of presentation to a mind or knower, he lets the nose of the idealist camel into the tent. He has then no great cause for surprise when the camel comes in--and devours the tent. Perhaps it will seem as if in this last paragraph I had gone back on what I said earlier regarding the physical explanation of the difference between the visible light and the astronomical star. On the contrary, my point is that this explanation, though wholly adequate as long as we conceive the perception to be itself simply a natural event, is not at all available when we conceive it to be an attempt at knowing its cause. In the former case, we are dealing with a relation between natural events. In the latter case, we are dealing with the difference between an object as a cause of knowledge and an object as known, and hence in relation to mind. By the "method of difference" the sole explanation of the difference between the two objects is then the absence or presence of relation to a knower.