Each year since 1982 Adelphi University has sponsored a conference on applied experimental psychology. The first six conferences were broad in scope in that each included a number of topical sessions. Around 1986, however, we started to think about holding a series of specialized or topical conferences, with some connection to applied experimental psychology (to be sure) but also with a workshop atmosphere in that groups of researchers would be collected so that they could hash out important issues. One such conference was on the perception of illusory contours ( Petry & Meyer, 1987). Another marked the centennial of the psychology of learning ( Gorfein & Hoffman, 1987). Another focused on behavior and social attitude change ( Curtis & Stricker, 1990). Yet another focused on research on semantic ambiguity ( Gorfein, 1989).
For 1988, we decided to focus on expert knowledge and the application of experimental psychology to expert system development. It had become clear that there was a need to get certain people together. Some of these "certain" people were computer scientists and some were psychologists, but all shared an empirical approach to the problem that artificial intelligence (AI) has encountered. Many had conceived of conducting empirical (if not experimental) analyses of knowledge elicitation issues, and most had gotten their hands dirty in knowledge elicitation.
The conference titled "Expert Systems and The Psychology of Expertise" was held at Adelphi on May 5, 1989. There were 14 participants, including Keynote Speaker Robert Sternberg. Each participant gave a 30-minute talk, followed by discussions. The conference was small (about 40 attendees) and had a workshop atmosphere. Actually it was rather intense. It was certainly jam-packed with discussions.
Many of the debates were stimulated and nurtured by Stephen Regoczei of Trent University. He came to the conference feeling hesitant to speak up, being one of a few fish out of water. But his questions were perhaps more novel and challenging than those that might be proffered by one of the "experts" on expertise.
I should point out here that the focus on experimental psychology and empirical approaches did not exclude the work by computer scientists on automated knowledge acquisition (cf. Gaines & Boose, 1988). Such work was described at the conference and is represented in this volume. How-