If multidisciplinarity is necessary, first, for studying the widest possible set of communication phenomena (organizational, in groups, interpersonal, media, computer-mediated communication...) and, secondly, for grasping the complexity of the different moments of the same phenomenon of communication (production, content, reception, circulation ...), methodological pluralism is also important. However, French research in communication sciences leaves in the shade a number of phenomena and moments of communication that could be better understood thanks to the experimental method. We show that the epistemological issues related to rational use of the experimental method in communication sciences are not negligible: it allows the study of objects that cannot be investigated with other methods and offers the opportunity to build knowledge by the refutation of hypotheses and theoretical propositions. We clarify some epistemological misunderstandings concerning this method. First, it is actually a method of studying complex systems and communication processes. Secondly, its use is not incompatible with constructivism.
