The main area of his research takes a cognitive and developmental perspective related to educational issues. His focus is brought to highlighting and analyzing of the role of prior knowledge in comprehension and reasoning activities, not neglecting situations in which knowledge is acquired. He proposes a theory of the cognitive mechanisms involved in these activities as well as their domain of validity, tested through experimental protocols and by the construction of models. His research is particularly rooted in the field of analogy, which he transposes to the psychology of school learning in order to open innovative perspectives in education. In particular, he seeks to model the interpretative aspects involved in solving arithmetic word problems, whose semantic aspects, irrelevant for providing an expert solution, induce a structure of the problem that is not the expert structure, and replaces it. He seeks to construct situations and analyze the solving processes which make it possible to specify the nature of the conceptualizations and thus to answer current educational and didactical issues. Finally, taking into account content effects of problem statements in order to foster knowledge acquisition, providing conditions for conceptual development and methods for its evaluation is a an important object of his research. Some of his work focuses on intelligent tutor learning, as is the case of DIANE (Computer Diagnostics on Arithmetic at the Elementary Level) which operates a fine diagnosis of solving strategies and errors elementary school students make when solving arithmetic problems.