Molyneux problem

2. Philosophical responses

Philosophers’ answers to Molyneux’s problem and their conception of its significance reflect their general epistemology. Both Locke’s empiricism and his requirement that any idea is such as the subject perceives it to be rule out the possibility of the sort of innately guided, unconscious reasoning postulated by Descartes. The perception of distance is therefore an acquired ability and the answer to Molyneux’s question is negative. This reply may seem inconsistent with Locke’s doctrine that ideas of primary qualities, such as shape or motion, resemble their causes: visual and tactual ideas of such qualities have common causes, which implies that they resemble each other. However, Locke saw Molyneux’s problem simply as showing that ‘the Ideas we receive by sensation are often in grown People alter’d by the Judgement, without our taking notice of it’ (1689: ch. II, §8). The ideas we receive by sight are allegedly only two-dimensional patterns of light and colour. In order to be able to recognize a three-dimensional shape or body in these patterns, we need to know how to correlate them with information provided by touch. The visual idea is like a word which induces in a hearer the associated idea, but is itself hardly noticed (see Locke, J. §§2, 4).

Leibniz (1765) accepted that the connection between the visual and the tactual sensations is not innate. However, if the blind person has been given the information that a cube and a globe lie in front of him, they can work out which is which from the geometric features the visual and the tactual ideas share. In the cube there are eight points distinguished from all the others, whereas in the globe there are none. This shows, Leibniz concludes, how important it is to distinguish sensory images from ‘exact ideas constituted by definitions’ (see Leibniz, G.W. §8). Another of Locke’s contemporaries, Edward Synge (1695), insisted on a somewhat similar distinction between image and sensation on the one hand, and idea and perception on the other. He argued that, although the images acquired by sight and touch are different, the ideas are identical. In recognizing a cube by sight and by touch only one conceptual ability is involved.

Berkeley took a contrary view, rejecting Locke’s account of primary qualities and arguing that ideas of sight and touch are radically heterogeneous, connected only by contingent correlations known through sense experience. His reasons for endorsing the negative answer to the problem, and for his enthusiastic development of the analogy with language, are thus far more radical than Locke’s (see Berkeley, G. §4).

The Molyneux problem was widely discussed in the eighteenth and nineteenth centuries (indeed Ernst Cassirer has claimed that it was the fundamental question for both continental and Anglo-saxon epistemology and psychology in the eighteenth century). Among the contributors were Condillac and Diderot whose Lettre sur les aveugles (Letter on the blind) contains a fascinating description of the difference between the ‘life-world’ of the blind and that of the sighted. Various attempts have been made to answer the question empirically, the first being ‘Cheselden’s case’, named after the surgeon who, in 1728, reported an operation performed on a boy with congenital cataract. Since it took some time for the boy to perceive and recognize objects by sight, Cheselden’s case was supposed to confirm the negative reply. Yet even Leibniz had foreseen that the subject might be ‘dazzled and confused by the strangeness’. In any case, since blindness is a symptom resulting from different diseases of the eyes and along the optical tract, it is difficult to assess what bearing empirical data have on the Molyneux problem.

On any interpretation, the Molyneux problem raises the issue of the relation between the perceptual representation of space attributable to the blind and that made available by sight. Berkeley’s radical assertion of heterogeneity assigned primary spatial experience to touch, treating visual ideas as mere signs of impending or available tactual ideas. Conversely, it might be questioned whether the blind possess genuine spatial concepts, on the grounds that they only have successions of tactual experiences. Possession of spatial concepts, however, involves thought of distinct objects existing not in succession but simultaneously. Yet both these views ignore the fact that there is only one behavioural space for a subject to move around in. Simply to act, a subject – whether blind or sighted – must have experience of egocentric space involving an ability to locate what is perceived, however perceived, in relation to their body. If that is correct, the chief question remaining is how readily a blind subject, made to see, is able to generalize spatial concepts to visual experiences.