The purpose of this study is to develop from naturalistic field observation data a set of concepts and hypotheses related to the teaching of thinking. The study conceptualizes and relates thought processes and instructional strategies by analyzing data collected from classrooms in which the Washington University Elementary Social Science Curriculum Project was taught. Two units from this project on India and on changing neighborhoods were observed in two different settings in fourth and fifth grade classrooms. In the settings, an initial set of questions guided the collection and analysis of data. Common characteristics of lessons in both settings helped to suggest categories for reporting the results of the study. Some examples of the data analysis include concepts and hypotheses relating instructional strategies and the reviewing, clarifying, and justifying of arguments. Implications of the study include further development of instructional theory, aiding the teacher in making judgements about teaching behaviors, and recognizing the importance of naturalistic field observation as a methodological tool for theory building. (Author/KSM)