Computers that think in Japanese?

Press, 12 May 1984, Page 20

Computers that think in Japanese?

The Fifth Generation: Artificial Intelligence and Japan’s Computer Challenge to the World. By Edward A. Feigenbaum and Pamela McCorduck. Michael Joseph, 1984. 274 pp. $35.75.

(Reviewed by

Neil Birss)

“Fifth Generation” is a call to arms by a prominent American computer scientist (Feigenbaum), and a scientific writer and chronicler of artificial intelligence (McCorduck). They fear their country, with its great traditional smokestack industries such as steel and car making taking a hammering, will soon also lose its dominance of the world’s growth industry, computing. The book has been endorsed by Sir Clive Sinclair, a leading British electronics industrialst, as “Essential reading for anyone concerned with computers in Britain.” Sir Clive says the Japanese quest for fifth-generation supercomputers is the greatest battleground of the century. “If we lose, we are out of the game,” he writes.

Japan, in a 10-year plan co-ordinated by the Ministry of International Trade and Industry, is driving to develop a new “fifth-generation” of computers. These will rely heavily on the concepts of artificial intelligence. One definition of this intelligence is: “computer programs that perform tasks, which

when carried out by human beings, require intelligence.” If successful, the project could easily make Japan the world’s dominant industrial power. At first the Japanese quest was regarded with considerable scepticism by the West’s computing establishment, but this is no longer so. For example, the authors quote Gordon Bell, the mini-computer pioneer (architect of the DEC PDP-11), as being alarmed by the Japanese scheme. Feigenbaum and McCorduck argue for a centrally-planned national development drive, on the scale of the space programme, so that America may harness artificial intelligence for its own fifth-generation computers. By implication, the book accepts the growing argument that the industrial decline of first, Britain, and now, America, is a cultural problem. The first symptoms of the problem, the argument goes, were when the iron masters of the First Industrial Revolution, sent their sons to the schools of the English Establishment. There, they were “gentrified,” later buying retreats in the country, and affecting a distaste for industry, though still drawing their incomes from it, of course. The organised settlements brought some of this spirit to New Zealand, where some of us feel that, somehow, investment in agriculture is more

wholesome than investment in industry. The outcome is a pitiful infrastructure for industrial growth. Whether this theory is correct or not, it is becoming clear that the traditional exports are no longer sufficient to maintain New Zealand’s standard of living. One field that the Industrial Development Commission has recognised as having much potential for New Zealand is industrial electronics. To develop this field, the new field of artificial intelligence will have to be brought to bear on software problems in this country. New Zealand has a handful of experts on artificial intelligence who have soldiered on unsung, unhelped, and little appreciated over the years. With burgeoning interest in hobby computing, new interest is being generated in the subject, but official help, compared even with that slated by Feigenbaum and McCorduck as grossly inadequate in America, has been absent. This book is not without flaws. The authors constantly interpose their views, quoting themselves in the third person. This style jars at first. And they provide insufficient technical explanation for some readers’ tastes. But the work is a polemic, rather than a treatise on artificial intelligence, and its aim is achieved.