Book reviews A Missing Link in Cybernetics: Logic and Continuity

Citation

Mitchell, R. (2010), "Book reviews A Missing Link in Cybernetics: Logic and Continuity", Kybernetes, Vol. 39 No. 2, pp. 383-383. https://doi.org/10.1108/03684921011021552

The missing link here is that between the “processing of continuous variables” and the “manipulation of discrete concepts” which, it is argued, has contributed to the “relative failure” of the analysis and modelling of intelligence. This is an interesting and well‐written book, by a former Lecturer in Cybernetics at the University of Reading, who knew some of the early cyberneticists and is well aware of the work of many other workers in the field.

The book begins with an interesting perspective of the origins of cybernetics, from that of McCulloch rather than that of Wiener. Chapter 2 introduces the study of intelligence and understanding, arguing that systems with continuous variables should be used when first studying intelligence. The viewpoint is then defended in Chapter 3 where various examples are given of where continuity is important. Chapter 4 then considers adaptation, self‐organisation and learning, which by also considering emergence, covers Lovelock's Daisyworld. Chapter 5 discusses learning in artificial neural networks, including backpropagation, and covers the author's work on significance feedback for which (he says) he should be added to the list of independent developers of backpropagation. This is followed by two short chapters on consciousness and fractal intelligence. The concluding chapter justifies the motivation for the book and looks at the question of whether artificial intelligence (AI) is possible. Each chapter has a useful summary at the end.

I thoroughly enjoyed reading this book. It is very well‐researched, with numerous references to the literature, and to some important debates, over the last 70 years. It provides an interesting and perhaps unconventional view on a wide range of aspects of intelligence, covering machine intelligence, and biological plausibility, culminating in the suggestion that intelligence might also have fractal‐like characteristics. I feel that the book would be improved were Chapters 6 and 7 to be longer, in particular exploring further the ideas of fractal intelligence. Although there is a little mathematics in here, most of the book is descriptive and readily understandable.

I happily recommend the book to relevant researchers in cybernetics, AI, and related fields.