Conventional symbolic gestures like pointing appear at the end of children's first year. Gesture-word combinations can be observed in their second year as the transition towards two-word speech. But before gestures are combined with words, gaze and vocalizations complement them. These combinations might already carry pragmatic, social and semantic functions. In this paper, we explore the issue of (dis)continuity between gestures and words and pointing/vocalization/gaze combinations with the data taken from two longitudinal follow-ups of French speaking children aged 8 months to three, filmed at home with their parents once a month. Our analyses show that gaze alone, or prosody alone, do not seem to have a discriminating role in differentiating the traditional functions of pointing (proto-imperative versus proto-declarative). Prosody in gesture/vocal combinations seemed to be an earlier, more discriminating feature than gaze in indicating children's positioning to their interlocutor and their use of differentiated social-pragmatic functions. Our results show that no feature alone can help interpret “pointing events” but that fine-grained correlations between pointing gestures, prosody, gaze, nature and position of the target, position of the participants, interpretation of the adult could help distinguish between the different functions of “pointing events”. Our first analyses of the vocal and verbal contents in the data show that pointing gestures are accompanied by vocalizations from the start. Vocal and gestural modalities are therefore associated and complement each other from the very onset of pointing. We extract and categorize all pointing gestures, vocalizations and demonstratives in order to analyze their functions in context. Different “stages” could be drawn. 1) At first, the children vocalize as they point. In order to interpretate the “pointing events”, we need to take various parameters into account (gaze, prosody, nature of the object, position of the adult...) and it is often quite difficult to understand the child's intention. 2) The children then use the demonstrative “ça” (“that”) combined with a pointing gesture at 1;02 simply to localize the object with the gesture and refer to it with a pronoun simultaneously. There is no additional semantic information in the demonstrative, which directly refers to the present object. But the children are now using language as part of their act of reference. 3) Around 1;05, the children replace the demonstrative “ça” with a noun and therefore name the object they are pointing. They combine a gesture to localize an object and a word with a referential value. 4) Around 1;07, they start combining the spatial demonstrative “là” with a noun (“un cochon là” / “a pig there”). The demonstrative “là” is a spatial localizer, which simply reduplicates the pointing gesture and does not give any semantic information. It is therefore informationally redundant but seems to express the children's true appropriation of language : they have fully entered the verbal mode. At the end of our data, the children are making syntax and combine two words in addition to the pointing gesture, a lexical word and a deictic word, which enables them to anchor their production in the here and now. The pointing gesture seems to be the pivot of this developmental path. It enables the interlocutor to understand the children's communicative intent. At the end of the process, pointing simply helps disambiguate the reference of the target, which is clearly referred to and localized in the verbal mode.