This study systematically analyzed social and cognitive processes that underlie the development of argumentative knowledge. Group discussions of controversial issues and explicit instruction in argumentation were expected to help students acquire a sense of the overall structure of an argument, or an argument schema. In a quasi-experiment, 128 fourth- and fifth-grade students from 2 schools completed the same argument-related tasks, after receiving different instructional treatments. In the first treatment condition, students engaged in group discussions of moral and social issues raised in their readings. In the second treatment condition, we supported group discussions with explicit instruction in abstract principles of argumentation. Students in the third condition received their regular reading instruction. Postintervention tasks included responding to an interview designed to elicit awareness of the criteria for a satisfactory argument, writing a reflective composition, and recalling an argumentative text. We quantified the data through assigning codes to oral and written text students produced. Next, we examined treatment differences using statistical models and discussed characteristic features of student responses. Findings revealed the complexity of learning and transfer in the domain of argumentation. Students who engaged in discussions with or without explicit instruction provided well-articulated responses to the interview questions. Student performance on the reflective essay was improved only by participation in discussions, although mean differences between some pairs of classrooms did not reach statistical significance. Recall of the argumentative text was generally insensitive to variations in treatment; however, the writings of some students suggested benefits from discussions and explicit instruction. (Contains 1 figure and 2 tables.)