Second installment. Presents Searle’s Chinese Room thought experiment — a human following an algorithm with Chinese symbol baskets never learns Chinese — and extends it to a ‘nonsense symbol room’ to show the computer’s additional disadvantage: no ability to even recognize symbol type. Dismisses Churchland’s electromagnetic ‘light room’ counterargument. Examines Proposition 3 (syntax alone cannot create semantics) through three cases: computer translation failures, the Bell Labs voice typewriter homonym problem (97% syntactic accuracy but 0% semantic cognition), and Gödel’s Theorem establishing hard limits on universal algorithms.
The Limits of Computer Intelligence II
See all articles from SMUG Bytes v7 n10