The many authors debating whether computers can understand often fail to clarify what understanding is, and no agreement exists on this important issue. In his Chinese room argument, Searle (1980) claims that computers running formal programs can never understand. I discuss Searle's claim based on a definition of understanding that is empirical, in the sense of being a theoretical statement leading to several testable hypotheses. I argue that an empirical definition with experimental support has a more solid footing than other definitions, and go on to discuss three hypotheses of particular relevance to the question of machine understanding, which turn out to have considerable support in the literature from experimental psychology, linguistics, and other fields. The hypotheses claim that (1) knowledge may not be needed to understand, (2) one may know all there is to know about X, and still not understand X, and (3) understanding tends to reduce the search for relevant information. I also show that understanding probably does not presuppose the causal powers of the human brain, and that computers can have intentionality in Searle's sense of the word. The article concludes that what Searle discusses is not what speakers of English refer to when they say they understand, and even if it were, his arguments would be unsound.
"Empirically Understanding Understanding Can Make Problems Go Away: The Case of the Chinese Room,"
The Psychological Record:
4, Article 6.
Available at: http://opensiuc.lib.siu.edu/tpr/vol55/iss4/6