This research contributes to the field of teaching and learning French as a Foreign Language and computer-mediated communication (CMC). Acquiring a second language (L2) is studied from an interactional perspective, according to wich learning a second language is facilitated through interpersonal interactions that occur during the execution of a task. The research builds on the teaching-learning project of the FFL " Français en 1ère ligne ," which connects tutors in France and American learners of Berkeley via a desktop videoconferencing tool. The aim of the study is to evaluate the potential of desktop videoconferencing through the analysis of the metalinguistic sequence triggered during the interactions. These sequences are episodes in which interlocutors address code problems related to mutual comprehension and language production. Based on the principles of conversation analysis with an ethnomethodological inspiration, our study is primarily empirical and descriptive. The categorization of metalinguistic sequences, the description of the situational context of appearance and interactional patterns, and the analysis of the discursive nature of the processes used to solicit the help of partner and repaired interactively language barriers, indicate that interactions through videoconferencing may trigger metalinguistic sequences, similar to those produced in face-to-face. The analysis of multimodality, focusing on how interlocutors use different modes of communication to solve communication problems, shows the positive effects on learning a second language. Indeed, multimodality allows the learner to better focus on form as well as a better awareness of the differences between his/her interlanguage and the target language.