Le dialogue est une activité conjointe au cours de laquelle au moins deux locuteurs interagissent afin d’atteindre un but commun. Tout au long de cette interaction, chaque locuteur met en œuvre un certain nombre de processus qui contribuent à la réussite du dialogue. L’objectif de cette revue est de présenter les approches théoriques qui ont été développées dans le but de rendre compte de ces processus. La première partie est consacrée à l’approche collaborative, qui considère que le dialogue est une activité au cours de laquelle chacun des locuteurs s’adapte à son partenaire. La deuxième partie est consacrée à l’approche égocentrique, qui précise que, dans certaines situations, les partenaires parviennent à se comprendre sans qu’ils ne se comportent de façon adaptative. Enfin, la dernière partie est consacrée à la présentation d’approches alternatives aux deux précédentes et à la façon dont ces alternatives permettent, dans une certaine mesure, de dépasser l’antagonisme entre collaboration et égocentrisme.

Dialogue is a joint activity during which at least two partners interact in order to reach a common goal. Throughout the interaction, each speaker resorts to a number of processes, which contribute to dialogue success. Several theoretical approaches have been developed in order to account for these processes. Among these approaches, the collaborative approach can be considered as the most influent one in the field. This approach offers a global framework for dialogue by pointing out that interactively building up mutual knowledge allows partners to adapt to each other during the remainder of the interaction. This implies that speakers frequently resort to costly inferential processes to determine what their partner is likely to know and what he or she is also likely to ignore. However, this approach has seriously been challenged by the egocentric approach, which suggests that speakers sometimes fail to take into account their partners’ dialogic needs. Furthermore, recent work within this approach has pointed out that adaptation is not systematically necessary for dialogue to succeed – rather, individual processes at play within each speaker incidentally contribute to mutual understanding. The probabilistic approach offers a means to overcome the apparent discrepancy between collaboration and egocentrism by showing how the processes at play during reference production and comprehension may be simultaneously affected by both kinds of influences. Moreover, the relative weighting of each kind of influence depends on the characteristics of the dialogue situation per se: speakers are more or less likely to act collaboratively depending on the situation they find themselves in. Finally, the interactive alignment model offers a qualitatively different framework for dialogue, as it suggests that most dialogic behaviors can be accounted for by one single automatic process. The aim of the current review is to provide a thorough presentation of each of these approaches, and to determine to what extent they are compatible with each other, on one hand and how they differ, on the other. We also raise the question of the apparent easiness of dialogue, looking at how – according to each approach – speakers deal with the individual and collective costs inherent in dialogue. We finally point out a number of issues that remain to be addressed in this research field.