Noting that researchers in the fields of cognitive psychology, artificial intelligence, and linguistics are taking a constructivist view of reading comprehension, this paper undertakes a comparison of that view with views concerning comprehension that have been expressed at the Claremont Reading Conferences over the past 50 years. The first section of the paper lists four assumptions basic to the constructivist view of comprehension: (1) what one knows affects what he or she will learn from reading; (2) both concept-driven and data-driven processes are necessary in comprehension; (3) the deeper one processes text, the more it will be remembered; and (4) context influences what is recalled. This section points out that teaching practices following from this view call for activating those concepts (schemata) that are related to the reading material, predicting what will follow, engaging in mental imagery, monitoring predictions, and revising. The second section of the paper reviews the various views of reading comprehension that have been offered by speakers at the Claremont Reading Conference since its inception. It concludes that these views are remarkably compatible with the constructivist theory. The final section of the paper discusses practices that are consistent with the previously discussed views. (A computer program flowchart for reading comprehension is appended.) (FL)