This chapter tests the usefulness of the category intonation phrase for the analysis of natural conversation. It asks whether the intonation phrase is a relevant unit for participants, and if so, whether it is a prosodic, or indeed an interactional category. The data show that while participants do divide their speech into intonation phrase-like chunks, these chunks are not defined by intonation alone. Instead, participants draw on a variety of interactional modes in their production of speech chunks, which are defined here as building blocks for turns and Turn Constructional Units. Chunks are shown to be employed as interactional units below the turn, and potentially below the Turn Constructional Unit; therefore the term Turn Constructional Phrase is suggested.

During the pursuit of learnables in vocal master classes, masters frequently produce lengthy clusters of directives, while students and accompanists orient to early opportunities for putting directives into practice. Participants are therefore faced with a continuous necessity to negotiate whether a directive is to be put into practice “now” or “not now.” Accompanists typically initiate musical (re)performances and are therefore the first to respond to a master's instruction completion, often preempting it early on in a master's potentially final turn constructional unit. Master class participants have to coordinate two action types: masters orient to giving instructions through talk; students and pianists orient primarily to restarting the musical performance.

The majority of research on prosody in conversation to date has focused on exploring the role of individual prosodic features, such as certain types of pitch accent, pitch register or voice quality, for the accomplishment of specified social actions. From this research the picture emerges that when it comes to the implementation of specific actions at specific sequential locations conversationalists employ prosodic features systematically, but also with considerable variation, and indeed flexibility. This paper suggests a further line of enquiry, which pursues a wider, more fundamental role of prosody for interaction, and which does not focus on individual prosodic practices or features, but on participants’ collaborative use of prosody for the implementation of one of the most basic interactional decisions: whether to continue a previously established action trajectory, or whether to start a new one. The data and findings of recent research make it clear that prosody, and in fact talk-in-interaction as such, is not appropriately defined by reference to individual features, speakers, locations and actions alone, but must be approached as a resource and negotiating strategy for social interaction. Prosody, therefore, must be described according to its role for both the accomplishment, and the coordination of actions across turns and participants.