eye: it does not inform us that we have an eye at all. Neither does touch inform us that our visual sensations are located in the eye. It does not lead us to associate with the eye any of the visual phenomena or operations in the first instance. If it did, it would (firstly) either be impossible for it afterwards to induce us to associate them with the presence of tangible bodies distant and different from the eye: or, (secondly), such an association would merely give birth to the abstract knowledge or conclusion, that these bodies were in one place, while the sensations suggesting them were felt to be associated with something in another place; colour would not be seen—as it is—incarnated with body: or, (thirdly), we should be compelled to postulate for the eye, as many philosophers have done, in our opinion, most unwarrantably, "a faculty of projection"[1] by which it might dissolve the association between itself and its sensations, throwing off the latter in the form of colours over the surface of things, and reversing the old Epicurean doctrine that perception is kept up by the transit to the sensorium of the ghosts or simulacra of things,
"Quæ, quasi membranæ, summo de corpore rerum
Dereptæ, volitant ultro citroque per auras."[2]
It is difficult to say whether the hypothesis of "cast-off films" is more absurd when we make the films come from things to us as spectral effluxes, or go from us to them in the semblance of colours.
But according to the present view no such incomprehensible faculty, no such crude and untenable hypothesis, is required. Before the touch has informed us that we have an eye, before it has led us to associate any thing visual with the eye, it has already taught us to associate in place the sensations of vision (colours) with the presence of tangible objects which are not the eye. Therefore, when the touch discovers the eye, and induces us to associate vision in some way with it, it cannot be the particular sensations of vision called colours which it leads us to associate with that organ; for these have been already associated with something very different. If it be not colours, then what is it that the touch compels us to associate with the eye? We answer that it is the abstract condition of impressions as the general law on which all seeing depends, but as quite distinct from the particular visual sensations apprehended in virtue of the observance of that law.
Nor is it at all difficult to understand how this general condition comes to be associated with the eye, and how the particular visual sensations come to be associated with something distant from the eye: and further, how this association of the condition with one thing, and of the sensations with another thing, (an association established by the touch and not by the sight,) dissolves the primary synthesis of seeing and colours. It is to be observed that there are two stages in the process by which this secernment is brought about—First, the stage in which the visual phenomena are associated with things different from the organ of vision, the very existence of which is as yet unknown. Let us suppose, then, the function of sight to be in operation. We behold a visible object—a particular colour. Let the touch now come into play. We feel a tangible object—say a book. Now from the mere fact of the visible and the tangible object being seen and felt together, we could not associate them in place; for it is quite possible that the tangible object may admit of being withdrawn, and yet the visible object remain: and if so, no association of the two in place can be established. But this is a point that can only be determined by experience; and what says that wise instructor? We withdraw the tangible object. The visible object, too, disappears: it leaves its place. We replace the tangible object—the visible object reappears in statu quo. There is no occasion to vary the experiment. If we find that the visible object invariably leaves its place when the tangible object leaves its, and that the one invariably comes back when the other returns, we have
