Abstract

This paper, which has both a historical and a polemical aspect, investigates the view, dominant throughout the eighteenth and nineteenth centuries, that the sense of sight is, originally, not phenomenally three-dimensional in character, and that we must come to interpret its properly two-dimensional data by reference to the sense of 'touch'. The principal argument for this claim, due to Berkeley, is examined and found wanting. The supposedly confirming findings concerning 'Molyneux subjects' are also investigated and are shown to be either irrelevant or disconfirming. Recent investigations on infant and neonatal perception are discussed and are also found to be disconfirming. An innatist version of the theory is then considered and is shown to be undermined by the largely 'Gibsonian' character of early space-perception. Finally three recent arguments in favour of the theory - two from psychologists, one from a philosopher - are considered and answered