Computer-based cognitive tools may have an important role to play in making widespread improvements in history teaching. Scholars agree that one important way to help students understand history is to involve them in historical interpretation, and there have been promising developments in the design of tools that scaffold students' interpretation of historical sources. However, some of the researchers themselves have pointed out important limits to this approach. Using participant-observer data from a classroom project aimed at improving students' grasp of "metahistorical" ideas, this article further illuminates the challenge of helping students to understand historical interpretation, and sketches the rough outlines of a library-based system that would augment existing cognitive tools. (Contains 2 footnotes.)