en.wikipedia.org/wiki/Blockhead_(thought_experiment)
1 correction found
there are a finite number of syntactically and grammatically correct sentences that can be used to start a conversation.
This misstates Block’s premise. Block’s 1981 argument relies on a fixed time limit and the finiteness of *typeable strings* in that interval, not on natural language having only finitely many grammatical sentences.
Full reasoning
In Block’s paper, the finiteness claim is explicitly time-bounded. He defines a “typable string of sentences” as a string that can be typed by a human “in an hour or less,” and then says the set of all such strings is finite because English has a finite vocabulary and there are only finitely many letter strings a typist can produce in that period. He does not claim that natural language as such contains only finitely many syntactically or grammatically correct sentences.
That distinction matters, because standard linguistic accounts treat natural languages as productive/recursive systems that can generate an infinite number of sentences from a finite lexicon. So the article’s wording incorrectly turns Block’s limited, practical finiteness assumption into a broad claim about language itself.
A more accurate paraphrase would be that, for any fixed-duration Turing test, there are only finitely many sentence-strings that could be typed during the test, allowing Block to imagine a giant lookup-table system for that bounded interaction.
2 sources
- Ned Block, “Psychologism and Behaviorism”
Block defines a “typable string of sentences” as one that can be typed “in an hour or less,” then says this set is “very large, but nonetheless finite.”
- Syntax — Open Encyclopedia of Cognitive Science
“Speakers of any language are able to produce an infinite number of sentences from a finite number of words.” The article also notes that recursion is “limitless” in theory.