"Where are you?". This recurring question when opening a mobile conversation attests the importance of spatialisation in social interactions. More and more applications support mutual location-awareness: they enable members of a group to locate their partners both in the physical environments and virtual worlds. This thesis contributes to the research in the field of Computer Supported Cooperative Work (CSCW) by examining how location awareness tools influence collaboration and what interpretations are drawn upon them in a collaborative context. Our research question concerns the effect of location-awareness on group cognitive processes such as communication and the modeling of others' intents (a process we referred to as "mutual modeling"). After a critical review of the existing mutual location-awareness interfaces, a theoretical framework grounded in psycholinguistics is introduced. It describes location-awareness as a "coordination device" that allow members of a group to have a shared understanding upon which they could mutually infer their respective intentions. Three studies have been realized in the form of semi-controlled experiments in two multi-user games; one in a 3D virtual world and two others in the physical environment, based on a location-based game. The first study demonstrated how the presence of a location-awareness interface did not necessarily imply its use. More surprisingly, in the second experiment, this type of interface had inhibiting effects on communication within groups and on the recall of partners past positions. It also made the group more passive than those who did not have this interface. Our third study showed how location-awareness is integrated among a set of coordination devices, namely the plan players established before the game. In addition, the three studies brought forward the various roles of mutual location-awareness ranging from a resource for division of labor to the facilitation of situation understanding or the use of past positions to draw hypotheses about the partners future behavior. We also developed visual representations of coordination to depict the effects of location-awareness tools on collaboration. These visualizations allowed us to represent these negative effects and to refine our understanding of their influence over time. Such results finally allowed us to discuss how automating location-awareness can be detrimental to group collaboration in certain situations. As most of the research in human-computer interaction has rather focused on optimizing the accuracy of location sensing and representation, the work done in this thesis is meant to ponder this agenda by bringing forward unexpected issues regarding how collaboration can be undermined by location-awareness.