228. His account of mathematical truths, as contained in Part II. of the First Book of the ‘Treatise on Human Nature,’ cannot be fairly read except in connection with the chapters in Part IV. on ‘Scepticism with regard to the Senses,’ and on ‘the Modern Philosophy.’ The latter chapter is expressly a polemic against Locke’s doctrine of primary qualities, and its drift is to reverse the relations which Locke had asserted between them and sensations, making the primary qualities depend on sensations, instead of sensations on the primary qualities. In Locke himself we have found that two inconsistent views on the subject perpetually cross each other. [1] According to one, momentary sensation is the sole conveyance to us of reality; according to the other, the real is constituted by qualities of bodies which not only ‘are in them whether we perceive them or not,’ but which only complex ideas of relation can represent. The unconscious device which covered this inconsistency lay, we found, [2] in the conversion of the mere feeling of touch into the touch _of a body_, and thus into an experience of solidity. By this conversion, since solidity according to Locke’s account carries with it all the primary qualities, these too become data of sensation, while yet, by the retention of the opposition between them and ideas, the advantage is gained of apparently avoiding that identification of what is real with simple feeling, which science and common sense alike repel. [1] See above, paragraph 99 and following. [2] See above, paragraph 101. It will not do to oppose bodies to our feeling when only feeling can give idea of body.