Logic and Art: Essays in Honor of Nelson Goodman ed. by Richard Rudner, Israel Scheffler (review)

Books 157 A machine that produces stereoscopic shadow-images (Ch. 21) for groups of observers is another source of interesting perceptual experiments in art classes. Other instruments of Gregory’s design may not be so interesting to artists and art teachers and many chapters of the book deal with rather technical points in the psychology of perception, but the book is one that those in the visual arts should at least look into for themselves. Pattern Recognition: Introduction and Foundations. Jack Sklansky, ed. Dowden, Hutchinson & Ross, Stroudsburg, Penn., 1973. 421 pp., illus. $22.00. Reviewed by Michael Thompson* This is a collection of 27 papers published mainly in the 19603, but dating back as far as 1943. They represent the efforts of authors to introduce new research and have a directness and clarity often lacking in papers that deal with technical aspects of a well developed area of research. (The editor has included a paper of his own (page 266) as an example of the latter.) Most of the authors strike a good balance between theory and results of practical experiment. The book contains a good deal of mathematics , however most of it is descriptive. As in most texts on pattern recognition, about half the material does not deal with pictures at all, but with statistical or linguistic methods of processing data that have been extracted from pictures ‘by an appropriate method’. Image processing is the transforming of pictures to forms more accessible to machines or humans. They may be cleaned and sharpened, compressed for storage, changed from halftone into line drawings and then analyzed to produce a description. Eight papers are offered, most quite approachable for the non-specialist. It is a pity that no papers deal with 3-dimensional scenes. A major characteristic of most data processing is that it is sequential (e.g., sort customer-information into ascending order of account number and process each account one after the other). Dealing with a picture this way gives rise to an arbitrary choice of sequence that denies the simultaneous quality of the picture elements. This suggests that the same operation might be carried out on every element at the same time. Three papers are presented describing such parallel processing. Two papers are included, which, although outside the mainstream of research, have been very influential intellectually and are often quoted. The first is‘The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain’ by Rosenblatt. The model allows for nerve cells that are used to gain in strength at the expense of the more idle ones and this ingeniously simple process is shown to be sufficient to enable recognition of patterns through training and ‘spontaneous’ concept formation. The other paper is Turing’s ‘Computing Machinery and Intelligence’, which poses the question ‘Do machines think?’ a n d then immediately replaces it by a practical ‘imitation game’ in which a computer must make it as difficult as possible for the other player to detect that his adversary is a computer and not a human being. Turing wrote this philosophical paper nearly 25 years ago when computers were quite primitive and he expected that it would take 50 years before they would be able to put up a good showing at the game. Are we halfway there? Well, I read the papers in the book chronologically instead of in the subject order in which they are presented and it was quite obvious that after promising starts, research gets bogged down. There is a gap of 14 years .between the paper entitled ‘Programming Pattern Recognition’ and the practical application of Hilditch on automatic analysis of chromosome spreads. Also, Chomsky’s paper dated 1956 is followed eight years later by a very primitive practical application by Kirsch on the interpretation of English text and picture patterns. *I Remez Road, Kadimah, Israel. Logic and Art: Essays in Honor of Nelson Goodman. Richard Rudner and Israel Schemer, eds. Bobbs-Merrill, New York, 1972. 330 pp. $10.00. Reviewed by Michael Thompson** Theremarkable variety of the 16essays that constitutes this tribute to Nelson Goodman reflects the diversity of his own achievements. The psychology of strategy formation, the subject-predicate model of mind, utilitarianism, pragmatics...