Standard models of language argue that speakers encode messages individually, but utterances in interaction are composed jointly (Goodwin 2018), dialogically designed with and for others (Linell 2009). This project investigates the practice of “sounding for others”, whereby one person vocalizes to enact someone else’s ongoing bodily experience. Pilot cases include vocalizing with others’ pain or emotion, and parents sounding an ‘mm’ of gustatory pleasure on their infant’s behalf (Wiggins & Keevallik 2020). Vocal resources, including loudness and voice quality, are furthermore used with both lexical and non-lexical items for socializing bodily skills, such as when instructors reinforce sensations during students’ real-time performance. The ‘sounding for others’ practice calls for us to investigate both how people establish synchronous co-experiencing through language and multimodal resources and the consequences for linguistic theory. Its existence challenges the conceptualization of communication as a linear transfer of information, and agency as an individual capacity. The project is designed to a) determine the means by which participants can sound for others; b) investigate how these means function in sensory displays, empathy, and pedagogical scaffolding; and c) re-conceptualize the notion of a linguistic and sensing agent. Using evidence from this project, we will be able to provide an empirically grounded theory of language as a multisensory and distributed phenomenon.