This research project aims to establish the gestural and verbal profile of very old people in normal aging, looking at their pragmatic competence from a naturalistic and ecological perspective. To reach this goal, a multimodal corpus (integrating text, sound, and gesture) will serve to carry out corpus-based analyses with the aim of testing several hypotheses: (i) a change in the use of (inter)subjective verbal markers (e.g., tu sais/vous savez ‘you know’) could be an indicator of a change in the empathic ability of the elderly; (ii) we expect a change in the use of gesture (incl. hand gesture, facial expression, and eye contact) associated with increasing age, which could be an indicator of an adaptive strategy used to reduce the cost of gestural production during interaction; (iii) we expect a change in the concurrent use of verbal and non-verbal pragmatic units considered within temporality (synchronization or heterochrony between modes), which could be an indicator of a change in the pragmatic competence of the elderly’s language. The dual approach, combining parametric and functional analysis of both verbal and non-verbal modes, will allow quantitative and statistical processing of language data, while guaranteeing maximum reproducibility of the analysis results. In response to European socio-economic concerns about the aging population, the project’s contribution is multiple: (i) enrichment of the discussion of the concrete strategies to be implemented to promote successful aging, in particular by improving care of the elderly; (ii) enrichment of multimodal interaction annotation systems, which is a key issue for linguistic description and more generally for understanding language mechanisms; (iii) provision of a reference multimodal corpus of the language in healthy very old people that may serve as a control corpus for further comparative studies (e.g., as an aid to early detection of dementia).