Social interactions can be viewed as “dialogs” during which messages between partners can be multimodal, consisting of several simultaneous or sequential signals from different sensory modalities. The nature of a message and the way it will be interpreted may depend on the signals, their order, but also the time intervals between them. This study analyzed the sequential organization of multimodal interactions with and without vocalizations to determine the influence of vocalizations on the behavioral structure of these interactions; to this purpose, we used Theme software which extracts behavioral structures (“patterns”) within behavioral sequences. These behavioral structures are characterized by their temporal organization. Our results on a captive group of red-capped mangabeys (Cercocebus torquatus torquatus) showed that during sociosexual interactions, vocalizations were signals associated with decision-making process of the caller, to initiate contact with a partner some distance away. During grooming interactions, in contrast, vocalizations were not associated with the decision-making process of initiating an interaction. Patterns extracted from interactions with and without calls differed in the number, nature, and position of the decision points they included. So, information conveyed by vocalizations was neither fully original nor redundant with that conveyed by other sensory modalities. Calls seem to reduce complexity of interactions and to decrease the number of decision points partners used to mutually adjust their behaviors. Calls also play a role in structuring interactions as they may occur in patterns either at the beginning or in the “middle” of the interactions but never at the end of them