Prosodic Plot of Dialogues: A Conceptual Framework to Trace Speakers’ Role

The Silent Treatment? Changes in patient emotional expression after silence

“…Cumulative statistics also indicated about 70% of silences were under 1 s, and about 40% were under 0.5 s, whereas, on average, 75% of conversational pauses are typically under 0.5 seconds (Levinson & Torreira, 2015) and between 82% and 95% are less than 1 s (Heldner & Edlund, 2010). As such, therapists may pause more often than a speaker engaged in typical colloquial conversation (see Silber-Varod & Lerner, 2017), supporting prior claims that therapists may try to be more intentional about using silence to observe the client, process therapeutic content and demonstrate their interest in the client (Hill et al, 2003).…”

“…Using linguistics research as a model to study psychotherapy may provide additional information about how all types of silence length are related to psychological processes. For instance, Silber‐Varod and Lerner (2017) compared silences ranging from 0.2 to 10 s from four colloquial conversations and four therapy sessions. Results indicated that the number of between‐speaker silences—one speaker stops talking, there is silence and the other speaker starts talking—was significantly different across colloquial and therapeutic conversation.…”

“…Results indicated that the number of between‐speaker silences—one speaker stops talking, there is silence and the other speaker starts talking—was significantly different across colloquial and therapeutic conversation. Although Silber‐Varod and Lerner (2017) began to explore how silence length relates to psychotherapy process, quantitative research is sparse and has yet to investigate the connections between silence length and emotional expression. Given the additional importance of emotional processing during psychotherapy (Greenberg & Safran, 1989) and the contradictory evidence regarding silence length and client emotion (Levitt, 2001a; Cuttler et al, 2019), a deeper quantitative study of all silence lengths and client emotional expression could be crucial to further understanding psychotherapy processes.…”

Using Complexity-Identical Human- and Machine-Directed Utterances to Investigate Addressee Detection for Spoken Dialogue Systems

“…Regarding H-M dialogues, humans usually have some negative experience in their daily living to adapt themselves to the limitations of technical systems. This results in a slower, more pronounced but less modulated way of speaking with a limited vocabulary and being mostly non-spontaneous [15,38]. Given these considerations, we claim that a real benchmark data set for AD systems should take into account future developments where H-H and H-M dialogues are getting more similar due to the growing capabilities of future technical systems.…”