Searle's Chinese room

In 1980 the philosopher John Searle published Minds, Brains, and Programs in the journal Behavioral and Brain Sciences, with a set of replies. The paper contains the Chinese room argument. Searle, who knows no Chinese, imagines being locked in a room with baskets of Chinese symbols and a rulebook, in English, that tells him which symbols to hand out in response to which symbols are passed in. People outside, who write in Chinese, pass in questions. Following the rules, he sends out answers that are indistinguishable from a native speaker's. Yet he understands nothing of what the symbols mean.¹

Searle's conclusion is that a computer running a program is in the same position. It manipulates symbols by their form, which is syntax, but understanding requires meaning, or semantics, and syntax alone does not produce semantics. He distinguishes strong AI, the claim that an appropriately programmed computer literally has a mind, from weak AI, which treats computers as useful tools for studying minds, and he rejects the strong version. The argument was aimed at programs such as Roger Schank's story-understanding systems, and at the claim, which goes back to Alan Turing's 1950 test, that behaving intelligently is enough for thinking.¹

The replies in the same issue and since are many. The systems reply says that the whole room, not the man, understands. Searle answers that the man could memorize the rulebook and still not understand. The robot reply, that connecting the system to the world would give meaning, and the brain-simulator reply have also been debated. It is one of the most discussed arguments in philosophy of mind, and it has returned in discussions of whether large language models understand language.¹