Selting (2000) describes the dynamics in turn interaction as an "interplay of syntax and prosody in their semantic, pragmatic, and sequential context. In the same way, Ford (2004), putting "contingency" at a central feature for the study of turn construction, highlights how grammar, prosody, action sequential and gestures contribute to the creation of talk-in interactions in face-to-face conversations. In this paper, I will also focus on turn constructions, and in particular, on overlaps, in 11 different invitations made through phone calls and between speakers of different L1. I will show that speakers, when unable to rely on gestures (phone calls do not provide any visual aid to the conversation) and on prosody (different L1s have substantial different prosodic units), relay mostly on action and action projection almost in the same rate than of grammar and action together. These results, on the one hand, challenge the supposed importance of grammar for the interpretation of possible turn completion and, on the other hand, highlights the main role of action and action projection in the construction of turns. The results of this research also emphasize the role of the speakers and how they "constantly detect patterns of conversations, extract probabilistic information about frequency of occurrence and have expectations about how the talk will proceed" (Larsen-Freeman and Cameron 2009). This holds true, especially when speakers cannot rely on contributors such as prosody and gestures, and suggests, in more general terms, that "action" is, more than any other, the central feature for the construction of turns in conversations.