We said, indeed, that all order necessarily appears as contingent. If there are two kinds of order, this contingency of order is explained: one of the forms is contingent in relation to the other. Where I find the geometrical order, the vital was possible; where the order is vital, it might have been geometrical. But suppose that the order is everywhere of the same kind, and simply admits of degrees which go from the geometrical to the vital: if a determinate order still appears to me to be contingent, and can no longer be so by relation to an order of another kind, I shall necessarily believe that the order is contingent by relation to an _absence of itself_, that is to say by relation to a state of things "in which there is no order at all." And this state of things I shall believe that I am thinking of, because it is implied, it seems, in the very contingency of order, which is an unquestionable fact. I shall therefore place at the summit of the hierarchy the vital order; then, as a diminution or lower complication of it, the geometrical order; and finally, at the bottom of all, an absence of order, incoherence itself, on which order is superposed. This is why incoherence has the effect on me of a word behind which there must be something real, if not in things, at least in thought. But if I observe that the state of things implied by the contingency of a determinate order is simply the presence of the contrary order, and if by this very fact I posit two kinds of order, each the inverse of the other, I perceive that no intermediate degrees can be imagined between the two orders, and that there is no going down from the two orders to the "incoherent." Either the incoherent is only a word, devoid of meaning, or, if I give it a meaning, it is on condition of putting incoherence midway between the two orders, and not below both of them. There is not first the incoherent, then the geometrical, then the vital; there is only the geometrical and the vital, and then, by a swaying of the mind between them, the idea of the incoherent. To speak of an uncoördinated diversity to which order is superadded is therefore to commit a veritable _petitio principii_; for in imagining the uncoördinated we really posit an order, or rather two. * * * * *