A Chinese Room that Understands
Herbert A. Simon, Stuart A. Eisenstadt · 2002
Abstract The thesis of this chapter is that a computer can be programmed (and has been programmed) to understand natural language, in the sense of being able to translate natural language text, to answer in natural language questions put to it in language, and to perform similar tasks which, if performed by a human, would be taken as convincing evidence that the human understood the same language. We will call this thesis Empirical Strong AI (SAI-E) to distinguish it from Logical Strong AI (SAI-L), the thesis that Searle refutes, and which asserts that ‘an appropriately programmed digital computer with the right inputs and outputs, one that satisfies the Turing Test, would thereby necessarily have a mind’ (Searle 1999: 115).