Intelligence, as Kant represents it to us, is bathed in an atmosphere of spatiality to which it is as inseparably united as the living body to the air it breathes. Our perceptions reach us only after having passed through this atmosphere. They have been impregnated in advance by our geometry, so that our faculty of thinking only finds again in matter the mathematical properties which our faculty of perceiving has already deposed there. We are assured, therefore, of seeing matter yield itself with docility to our reasonings; but this matter, in all that it has that is intelligible, is our own work; of the reality "in itself" we know nothing and never shall know anything, since we only get its refraction through the forms of our faculty of perceiving. So that if we claim to affirm something of it, at once there rises the contrary affirmation, equally demonstrable, equally plausible. The ideality of space is proved directly by the analysis of knowledge indirectly by the antinomies to which the opposite theory leads. Such is the governing idea of the Kantian criticism. It has inspired Kant with a peremptory refutation of "empiricist" theories of knowledge. It is, in our opinion, definitive in what it denies. But, in what it affirms, does it give us the solution of the problem? With Kant, space is given as a ready-made form of our perceptive faculty--a veritable _deus ex machina_, of which we see neither how it arises, nor why it is what it is rather than anything else. "Things-in-themselves" are also given, of which he claims that we can know nothing: by what right, then, can he affirm their existence, even as "problematic"? If the unknowable reality projects into our perceptive faculty a "sensuous manifold" capable of fitting into it exactly, is it not, by that very fact, in part known? And when we examine this exact fitting, shall we not be led, in one point at least, to suppose a pre-established harmony between things and our mind--an idle hypothesis, which Kant was right in wishing to avoid? At bottom, it is for not having distinguished degrees in spatiality that he has had to take space ready-made as given--whence the question how the "sensuous manifold" is adapted to it. It is for the same reason that he has supposed matter wholly developed into parts absolutely external to one another;--whence antinomies, of which we may plainly see that the thesis and antithesis suppose the perfect coincidence of matter with geometrical space, but which vanish the moment we cease to extend to matter what is true only of pure space. Whence, finally, the conclusion that there are three alternatives, and three only, among which to choose a theory of knowledge: either the mind is determined by things, or things are determined by the mind, or between mind and things we must suppose a mysterious agreement.