Can Computers Understand Language?

Title: Can Computers Understand Language?

Research Question: Wolfram Schmied, in his paper "Demolishing Searle's Chinese Room," challenges John Searle's argument that computers cannot understand language. Schmied presents two versions of Searle's Chinese Room argument and demonstrates that both versions fail for different reasons.

Methodology: Schmied reconstructs Searle's argument, presenting the original version and an internalized version. In the original version, a person, Searle, sits in a room and follows a set of rules to respond to Chinese characters. Searle claims that he doesn't understand Chinese, but his responses are indistinguishable from those of a native Chinese speaker. In the internalized version, Searle memorizes the rules and calculations and does everything in his head, claiming that he still doesn't understand Chinese.

Results: Schmied points out that in the original version, Searle is not the same as the computer running the Chinese program. Therefore, the computer can understand Chinese, something Searle cannot. In the internalized version, Searle fails to provide an reason why he, despite speaking Chinese, cannot understand Chinese.

Implications: Schmied's critique suggests that Searle's Chinese Room argument is not a valid way to prove that computers cannot understand language. It also highlights the importance of clearly defining what it means to "understand" language.

Link to Article: https://arxiv.org/abs/0403009v2 Authors: arXiv ID: 0403009v2