AN ALGORITHM FOR THE SEGMENTATION OF AN ARTIFICIAL LANGUAGE ANALOGUE

Hayes & Clark (1970) have demonstrated that adult subjects are capable of identifying the beginnings and ends of ‘words’ in artificial ‘speech’ when all clues from pause, stress or intonation are absent. This paper describes the mechanism and properties of a computer program which can perform an analogous segmentation of letter strings. Although the program contains artificialities, it seems to model other aspects of cognition besides perceptual segmentation: the use of redundancies to effect economies in storage and retrieval of information, induction, and the importance of context in recognition. The behaviour of the program and other evidence suggests that the joint probability of two perceptual elements provides a better definition of association than transition probability. The possible relevance of the program to the learning of grammatical patterns is briefly discussed.