348 w. JAMES :

be ; if they had gone on definitely to ask and definitely to answer the question, What are the size and distance in their proper selves ? they would not only have escaped the present deplorable vagueness of their space-theories, but they would have seen that the objective spatial attributes ' signified ' are simply and solely certain other optical sensations now absent, but which the present sensations suggest. What, for example, is the slant-legged cross which we think we see on the wall when we project the rectangular after-image high up towards our right or left (Figs. 5 and 6)? Is it not in very sooth a retinal sensation itself? An imagined sensation, not a felt one, it is true, but none the less essentially and originally sensational or retinal for all that, the sensation, namely, we should receive if a ' real r slant-legged cross stood on the wall in front of us and threw its image on our eye. That image is not the one our retina now holds. Our retina now holds the image which a cross of square shape throws when in front, but which a cross of the slant-legged pattern would throw, provided it were actually on the wall in the distant place at which we look. Call this actual retinal image the ' square ' image. The square image is then one of the innumerable images the slant-legged cross can throw. Why should another one, and that an absent one, of those innumerable images be picked out to represent exclusively the slant-legged cross's ' true ' shape ? Why should that absent and imagined slant -legged image displace the present and felt square image from our mind ? Why, when the objective cross gives us so many shapes, as it varies its position, should we think we feel the true shape only when the cross is directly in front ? And when that question is answered, how can the absent and represented feeling of a slant- legged figure so successfully intrude itself into the place of a presented square one ? Before answering either question, let us be doubly sure about our facts, and see how true it is that in our dealings with objects we always do pick out one of the visual images- they yield, to constitute their real form or size. The matter of size has been already touched upon (p. 192), and no more need be said of it here. As regards shape, almost all the retinal shapes objects throw are perspective 1 distortions '. Square table-tops constantly present two acute and two obtuse angles ; circles drawn on our wall- papers, our carpets or on sheets of paper usually show like ellipses ; parallels approach as they recede ; human bodies are foreshortened ; and the transitions from one to another