Foundations of Neurology

An important and indeed almost Utopian project has been launched successfully. The research workers at the Bell Telephone Laboratories show in this book how a machine can translate the sounds of speech into patterns which the eye can learn to interpret. A startling principle has been proved, and the imagination is roused by the possibilities which this opens up. The authors wisely do not exploit this emotional appeal; instead they describe in a careful and matter-of-fact way what has been achieved, how it was done, and how much more work will be required before the implications can be fully understood. "Visible speech " as described in this book is a representation of sound according to frequency, intensity, and time. That such a representation could retain all the intelligence contained in the sound has long been appreciated; it is in fact similar to the analysis presumed to take place in the inner ear. But what is new is the experimental proof that the visual representation can be learnt and interpreted. Some experimenters have achieved considerable competence in this, and the authors give illustrations and a carefully devised commentary to enable the reader to learn the first steps for himself. The reader can thus obtain a clear idea of the processes of learning this new symbolism and assess the difficulties involved. The authors discuss the possible application of this type of technique to problems of deafness, and these suggestions seem to promise much. Teaching the deaf has always been an arduous task, and this work may well become the foundation of an entirely new method. Though little work with the deaf is yet on record the authors mention the case of one congenitally deaf man who learnt by this technique to use a telephone. An extension of the technique may even result in the portable " translator " as a permanent aid for the deaf, but the technical problems involved are evidently still very great. The authors discuss other applications of the technique covering nearly all fields of sound analysis; linguistic and phonetic investigations, the study of animal calls, the suppression of mechanical noises, and numerous other problems may all benefit. A precise and permanent record that can be investigated at leisure mav well provide a better means of classifying and analysing sounds than the rapid and subjective judgment made by listening. It is evident that the most lavish experimental technique was used to produce that form of display which held out the greatest promise. The authors point out that there is nothing final about the present method, and they refer to some variants of it. But it would have been interesting to hear why a horizontal time scale, a vertical scale for frequency, and an intensity-oflight representation of intensity of sound were chosen throughout all experiments. THOMAS GOLD.