Trends in emphases in the study of clinical reasoning are examined, with attention to three major branches of research: problem-solving, knowledge engineering, and propositional analysis. There has been a general progression from a focus on the generic form of clinical reasoning to an emphasis on medical content that supports the reasoning process. This has led to investigation of the nature of physicians' understanding of medical problems and the role of the basic biomedical sciences in clinical thinking. Studies of expert and novice diagnosis in pediatric cardiology have shown the importance of: medical knowledge in clinical thinking; and hypothesizing in case solution. Other research has assessed problem representation in a medical field and how it influences radiological diagnosis. Bill Clancey, a researcher in artificial intelligence at Stanford, has studied expert diagnostic performance rules and traced back underlying cognitive structures that give rules their conceptual meaning. Clancey used MYCIN heuristic programming in studying clinical diagnosis. Patel and colleagues at McGill University have focused directly on clinical understanding (based on theories of comprehension and the field of discourse processing) and have examined inferences by experts aimed at supporting clinical actions. Other lines of research are also noted. Four pages of references are included. (SW)