Review of 'The Cambridge Quintet and A Beautiful Mind'
- By Paul Trachtman
- Smithsonian magazine, May 1999, Subscribe
The Cambridge Quintet
After nearly half a century of living with computers, quite rational people still argue about whether such machines can really think. Two people can look at the same computer and where one sees artificial intelligence, the other sees a mindless embodiment of logical rules. The same two observers, of course, might very well disagree about another human being's state of mind, one seeing a madman where the other sees a visionary or a prophet. At the heart of such differences lie our deepest beliefs, and fears, about the nature of intelligence, rationality and the human mind. Two recent books explore such issues, one as a matter of pure reason, the other as a story of genius, insanity and love. John Casti's The Cambridge Quintet is a brilliant literary invention and intellectual feast. It is the account of an imagined dinner party at Cambridge University just after World War II, hosted by the famed writer C. P. Snow to pick the brains of other great minds about the prospects of actually constructing a "thinking machine." The guests are mathematician Alan Turing, who did pioneering work on such a project, along with biologist J.B.S. Haldane, quantum physicist Erwin Schrödinger and philosopher Ludwig Wittgenstein.
The chapters of the book are divided into the various courses of the meal: the sherry, the soup, the fish and so on. And the conversation is as rich as the menu. As the guests argue for and against the possibility of a thinking machine, the eavesdropping reader gets a deliciously clear and engaging account of the problem. Turing explains lucidly how, by manipulating the symbols 0 and 1, a machine can perform logical operations; and how this is analogous to the brain, where each individual neuron is switched on or off by its stimuli. "Moreover," Schrödinger offers, "both the computer and the brain process these data into patterns. Is that the analogy you're pursuing, Turing?"
"Precisely," Turing replies. "The brain stores its data in the form of patterns created by the firing of its neurons. Each such pattern is just a listing of which neurons are on and which are off at any given moment. These patterns are then associated with what we call 'thoughts' in ways that no one yet really understands." Brushing aside his own caveat — that all this occurs in a manner that hasn't yet been elucidated &$0151; Turing holds that "intelligence is just a matter of following the right kind of rules." Several generations of scientists have followed him down this path.
At Snow's dinner, however, it is Wittgenstein, the philosopher of language, who points to the flaw in Turing's reasoning. A pattern of symbols, or language, has no meaning, he says, apart from the shared experience of people using those symbols. "To have language is to have a way of life," he insists. "Everything we say is totally bound up with what we do." Whether the essence of language is syntax (the rules) or semantics (the meaning) continues to divide linguists as well as computer scientists and philosophers, but Casti's dinner debate is as clear as it is entertaining. Because Turing and his heirs define human thinking in abstract, mechanical terms, they believe machines can think. For Wittgenstein and his heirs, thinking follows the experience of life, not just the rules of logic, and machines have it not.
Perhaps, Casti concludes, the answer will lie in a divergence we can only guess at. "This half a century of research," he observes, "has shown that these are two distinct forms of intelligence, and that for a short period they will peacefully coexist. After the current, but brief interregnum, machines and humans will go their separate ways, much as humans and dolphins parted company many millennia ago."
Subscribe now for more of Smithsonian's coverage on history, science and nature.
Related topics: Book Reviews