[1] P. 483 and following, and p. 486 [Book I, part IV., sec. II.] [2] P. 516 [Book I, part IV., sec. IV.]. [3] Pp. 342, 343 [Book I, part II., sec. III.]. [4] P. 342 [Book I, part II., sec. III.] The fact that colours mix, not to the purpose. 248. This plausibility depends chiefly on our reading into Hume’s doctrine a physical theory which, as implying a distinction between feeling and its real but unfelt cause, is strictly incompatible with it. Is it not an undoubted fact, the reader asks, that two colours may combine to produce a third different from both--that red and yellow, for instance, together produce orange? Is not this already an instance of a compound impression? Why may not a like composition of unextended impressions of colour constitute an impression different from any one of the component impressions, viz. extended colour? A moment’s consideration, however, will show that no one has a conscious sensation at once of red and yellow, and of orange as a compound of the two. The elements which combine to produce the colour called orange are not--as they ought to be if it is to be a case of compound impression in Hume’s sense--feelings of the person who sees the orange colour, but certain known causes of feeling, confused in language with the feelings, which separately they might produce, but which in fact they do not produce when they combine to give the sensation of orange; and to such causes of feeling, which are not themselves feelings, Hume properly can have nothing to say. How Hume avoids appearance of identifying space with colour, and accounts for the abstraction of space.