Languages of Art: An Approach to a Theory of Symbols by Nelson Goodman, and: Laws of Form by C. Spencer Brown (review)

Books 175 possible between a man typing questions, assertions and commands and the machine that prints its replies on a graphics display terminal. The subject matter is, however, quite narrow and appears as a line drawing on the display. It is a set of toy blocks of different colours and simple shapes (oblong, cube, pyramid) that lie on a table or on top of each other and can be taken out of or put into a box. This ‘blocks world’ has quite sufficient complexity to test the computer’s understanding of English sentences. The book is a reprint of a previously published paper [I] and it is a pity that it was not expanded a little to include some less impersonal material, e.g. who was involved and what difficultieswere overcome. The first third of the book is easily accessible to the general reader and includes a description of the system, a comparison with other systems and an outline of the English grammar that it uses. The rest of the book is more technical and readers must be able to understand the LISP programming language. Those experienced in a language other than LISP will probably be able to follow the text, others will need to read an introduction to list processing [2]. Few sentences seem ambiguous to humans when first read, as they are guided by understanding to choose a single meaning for each phrase more or less as they come to it. Winograd insists that a computer should follow such an ‘intelligent’ approach. For instance, whilst it is parsing a sentence, it may need to call on one of the semantic programs to decide a point of ambiguity. Indeed, the meaning of the words themselves may not be sufficient to do this and the semantic program may in turn have to call upon a deductive program before the point is settled and the machine can return to parsing. The computer’s ability to switch rapidly between syntax, semantics and deduction is enhanced by having every part of the system written as a procedure to be followed. Even knowledge consists of procedures. Below is an example (fictitious to avoid lengthy explanation of Winograd’s methods but I feel in the same spirit). The word ‘ball’ has been encountered and to understand it a program BALL is called upon. This has been written to choose between ‘spherical object’ and ‘social assembly for dancing’. BALL may call upon other semantic and deductive programs to access immediate context and subject of discourse as a guide. Rather than debating the philosophy of distinctions betweeen ‘specific objects’, ‘properties’, ‘relationships’, ‘events’ etc., Winograd is looking for a usable way of representing knowledge. Does not this suggest directions for computer art? Incorporation of ‘intelligent’ discrimination and ‘knowledge ’ of aesthetics in procedures is probably vital for a computer to ‘understand’ and, hence, deal effectively with visual problems. Note also that syntax, semantics, deduction and knowledge are dealt with in the system as a whole and that the success of a system is a matter of subjective judgement. The approach is not foreign to the artist. References 1. T. Winograd, Understanding Natural Language, Cognitive Psychol. 3 (No.1, 1972). 2. McCarthy el at. LISP 1.5 Programmer’s Manual (Cambridge, Mass: Mass. Inst. of Tech. Press, 1962). Those wishing to understand list processing in general would do better to read the short book by J. M. Eoster, List Processing (London: McDonald, 1967). Climate for Creativity. Calvin W. Taylor, ed. Pergamon Press, Oxford, 1972. 304 pp. f5.75. Reviewed by: Florence Vidal* This collection of papers by psychologists was presented at the Seventh National Creativity Research Conference (1966.) Like all collections of this type, the book lacks unity, nevertheless it is of particular value, because it presents diverse viewpoints of this complex subject. ‘12 rue Denfert-Rochereau, 92100 -Boulogne sur Seine, France. The title of the book is very suggestive. It implies that human creativity is not dependent upon mere chance of genetics. Rather, it depends very much on the sociocultural environment, which may encourage or discourage its expression. If creativity in the modern world is ‘a psychological uranium’ and if, according to Toynbee, it is even a ‘question of...