privilege of adding to itself a reproduction, this time relative and variable, of all the others. It is true that you afterwards pretend to attach no importance to this representation, to see in it a mere phosphorescence which the cerebral vibrations leave behind them: as if the cerebral matter and cerebral vibrations, set in the images which compose this representation, could be of another nature than they! All realism is thus bound to make perception an accident, and consequently a mystery. But, inversely, if you posit a system of unstable images disposed about a privileged centre, and profoundly modified by trifling displacements of this centre, you begin by excluding the order of nature, that order which is indifferent to the point at which we take our stand and to the particular end from which we begin. You will have to bring back this order by conjuring up in your turn a deus ex machina; I mean that you will have to assume, by an arbitrary hypothesis, some sort of pre-established harmony between things and mind, or, at least (to use Kant's terms), between sense and understanding. It is science now that will become an accident, and its success a mystery.—You cannot, then, deduce the first system of images from the second, nor the second from the first; and these two antagonistic doctrines, realism and idealism, as soon as they decide to enter the same lists, hurl themselves from opposite directions against the same obstacle.
If we now look closely at the two doctrines,
