Can a Machine Think Like Us?

by Fred Copley

Machines regularly out-pace human beings in all sorts of ways. They are capable of greater strength, speed, and endurance than we, or any animal, can muster. We not only accept this fact but are usually appreciative to boot. Who among us could run from New York to Boston in a matter of hours? Considering that machines act as extensions of our own intentions—as John Searle perhaps said best but by no means first—we are right to appreciate them.

Even as we appreciate the work a computer can do, why, then, must we also begrudge its powers of computation, always carefully delineating them from us?

Clearly, we wish to defend our own uniqueness, as if that were in question. Perhaps we cling to the idCan aea of the mental as something entirely separable from the physical capabilities of machines. What if thinking is necessarily a feature of bodies, and what if we are merely a very special kind of machine? Will machines ever replicate and even surpass us in our kind of thinking?

A.M. Turing’s conviction that machines would be recognized as capable of thought, and even of fooling humans, by the year 2000 is represented by his famous Turing machine. Dispensing with the question of whether machines can think—which only leads to questions about what thinking is—Turing asked whether a human being would reliably think that a machine could think.

More precision is needed. Machines admittedly compute more powerfully than humans in many, if not most respects. However, thinking and computing are not the same thing. Thinking is understanding, not mere processing. This doesn’t mean machines can’t think. Donald Davison said, “I do not doubt that an artificer could, at least in principle, manufacture a thinking machine. The problem, for philosophy anyway, is what to aim for; what would show that the artificer had succeeded.” [See Note 1.]

For Davidson, thinking minimally requires the ability to think propositionally (i.e., that something is the case) and to recognize one’s errors. This involves history and a physical existence. Searle contends that a thinking machine would have to have the same bio-chemical causal relations that a brain does.

Searle and Davidson both believed that in principle machines could think, but in order to do so the machines would need to be an awful lot like us.

Perhaps this is less scary than it seems.

[ Note 1: From Donald Davison, “What Thought Requires” in Problems of Rationality (Oxford: Oxford University Press, 2004), p.136: “We are just machines that are complex in ways flies are not, so the problem isn’t one of transcending mere physical devices. I do not doubt that an artificer could, at least in principle, manufacture a thinking machine. The problem, for philosophy anyway, is what to aim for; what would show that the artificer had succeeded?” ]

(Image of book cover from computer history museum from flickr.)