The Snow of yesteryear
Nigel Spivey
THE CAMBRIDGE QUINTET by John L. Casti Little, Brown, £16.99, pp. 182 That some few scientists possess divulga- tive ambitions is cause for gratitude. That most of those few lack the basic literary capacity to realise those ambitions must be accepted with stoical calm. There are aca- demics who profess English literature, even media studies, equally incapable of explain- ing their expertise to extramural ears. As one who still picks up the telephone with a sense of technological awe, I was eager to see if John L. Casti (apparently affiliated to the Santa Fe Institute and the Technical University at Vienna) could divulge what basically happens in the guts of my word- processor. What, I would like to know, does a memory look like? Such an exposi- tion, successfully performed, would be a service to multitudes. It is not, alas, achieved in this book, but we may give con- solatory applause for trying.
Casti subtitles his narrative 'a work of scientific speculation'. That can hardly be what he means. He has attempted a romance of historical imagination in which the mirrors are turned rearward. The hunched myrmidons of the Microsoft empire may safely disregard such overt retrospection. Like most fiction (or history) it is virtual reality. Casti imagines that in 1949 — presumably under Cold War pressure — British government scientific advisers asked the erstwhile physicist, adept bureaucrat and popular novelist C. P. Snow to 'sound out the scientific com- munity' on the question of whether a machine could be made to think. Snow's response is to organise a private function at his old Cambridge college, Christ's. To this dinner he invites Alan Turing, mathemati- cian and wartime code-breaker at Bletchley Park; J. B. S. Haldane, geneticist; Erwin SchrOdinger, physicist; and Ludwig Wittgenstein, philosopher.
The ensuing symposium breaks down into chapter headings of 'The Sherry', 'The Soup', and so on. Casti captures a fraction No, Darren, Teddy isn't gay.' of the egotism that would have crackled on such an occasion, and none of the wit. In fact his capacity for characterisation is close to zero. The number of research sessions conducted in Cambridge cannot have been great: serving a green salad between meat and dessert seems peculiar, and retiring directly for cognac equally so (Snow's nose was notoriously for claret, in any case). Perhaps the most unconvincing part of Casti's tale is that the food is pronounced excellent throughout. A kindly Oxbridge don should have told Casti that the penance paid at High Table for fine wine is almost invariably repellent food. But these are petty and snobbish cavils. What does the hypothesised meeting of great minds reveal?
Old Snow totters off at the end of the congress more befuddled than illuminated, and wondering — without distress — how he shall weave the necessary equivocation of a report to mandarins. But we have the benefit of playing with the computers that (in part) derived from the 'thinking machines' experimentally tried by Turing before his suicide in 1954. Most of us recognise that there are some cognitive tasks which our humming computers can do with an ease beggaring astonishment. Yet most of us (I believe) also see some- thing pathetically incompetent in a com- puter's void of inductive skills and sensitivity (Tool. Anyone could see I didn't mean to do that . . . '). This commonsense consensus or intuition about 'artificial intelligence' has had influential buttressing from the likes of Roger Penrose and others. Here, in Casti's exchanges, its spokesman is Wittgenstein, who pits him- self implacably against Turing and argues for the exclusivity of human thought indeed, for what is now called 'emotional intelligence'. Following Wittgenstein's logic, human cognitive activity cannot be estranged from human language, and human language cannot be estranged from human life. (Perhaps this is why, when he was beaten at chess last year by a computer called Deep Blue-II, Garry Kasparov described his opponent's strategy as manifesting an 'alien intelligence').
The physicist and the geneticist impinge on the debate, of course: nowadays, rather more could be contributed by an archaeol- ogist of human origins. But fundamentally the dispute collapses when Turing sum- marises by saying that he is not in quest of fabricating the physiology of the brain, but rather of a machine which can replicate human thought processes — implicitly, a selection of such processes. Thus defined, the enterprise ceases to threaten. A chess- playing machine is like a singing dog — of passing interest, but in the end we prefer dogs who bark predictably, and playing chess with someone who can weep, giggle and sweat. The residual virtue of Casti's book is that it reaffirms, in its mediocre way, the daily marvel of belonging to humankind.
Previous page
