third form, so as to make it quite clear why theAnd might lead to an empirical solution of metaphysical problems. problem of memory is in our eyes a privileged problem. From our analysis of pure perception issue two conclusions which are in some sort divergent, one of them going beyond psychology in the direction of psycho-physiology, and the other in that of metaphysics, but neither allowing of immediate verification. The first concerns the office of the brain in perception: we maintain that the brain is an instrument of action, and not of representation. We cannot demand from facts the direct confirmation of this thesis, because pure perception bears, by definition, upon present objects, acting on our organs and our nerve centres; and because everything always happens, in consequence, as though our perceptions emanated from our cerebral state, and were subsequently projected upon an object which differs absolutely from them. In other words, with regard to external perception the thesis which we dispute and that which we substitute for it lead to precisely the same consequences, so that it is possible to invoke in favour of either the one or the other its greater intelligibility, but not the authority of experience. On the contrary, the empirical study of memory may and must decide between them. For pure recollection is, by hypothesis, the representation of an absent object. If the necessary and sufficient cause of perception lies in a certain activity of the brain, this same cerebral activity,
