Interactive repair and the foundations of language

Dingemanse, Mark; Enfield, N.J.

doi:10.1016/j.tics.2023.09.003

Cited by 10 publications

(5 citation statements)

References 93 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Seminal work in psycholinguistics and conversation analysis suggests that “language use is fundamentally a joint activity” (Clark, 1994, p. 244; Clark & Schaefer, 1989; Clark & Wilkes-Gibbs, 1986; Schegloff et al, 1977): While speaking and listening are individual actions, they contribute toward joint “problems”—errors in achieving shared understanding—and dealing with these “problems” requires joint management, strategies that can help both prevent and repair glitches in shared understanding (Clark, 1994). Taken together, individuals use different verbal, nonverbal, and paralinguistic cues to coordinate their shared understanding on a turn-by-turn basis, and doing so is a uniquely human ability (Dingemanse & Enfield, 2023).…”

Section: Previous Research On Listening—a Conundrum For Determining A...mentioning

confidence: 99%

Conveying and detecting listening during live conversation.

Collins,

Minson,

Kristal

et al. 2024

Journal of Experimental Psychology: General

View full text Add to dashboard Cite

Across all domains of human social life, positive perceptions of conversational listening (i.e., feeling heard) predict well-being, professional success, and interpersonal flourishing. However, a fundamental question remains: Are perceptions of listening accurate? Prior research has not empirically tested the extent to which humans can detect others' cognitive engagement (attentiveness) during live conversation. Across five studies (total N = 1,225), using a combination of correlational and experimental methods, we find that perceivers struggle to distinguish between attentive and inattentive conversational listening. Though people's listening fluctuated naturally throughout their conversations (people's minds wandered away from the conversation 24% of the time), they were able to adjust their listening in line with instructions and incentives-by either listening attentively, inattentively, or dividing their attention-and their conversation partners struggled to detect these differences. Specifically, speakers consistently overestimated their conversation partners' attentiveness-often believing their partners were listening when they were not. Our results suggest this overestimation is (at least partly) due to the largely indistinguishable behavior of inattentive and attentive listeners. It appears that people can (and do) divide their attention during conversation and successfully feign attentiveness. Overestimating others' attentiveness extended to third-party observers who were not immersed in the conversation, listeners who looked back on their own listening, and people interacting with partners who could not hear their words (but were incentivized to act like they could). Our work calls for a reexamination of a fundamental social behavior-listening-and underscores the distinction between feeling heard and being heard during live conversation. Public Significance StatementThis research reveals that there is a difference between being and feeling heard. People's listening fluctuated naturally throughout their conversations (with mind wandering reported 24% of the time)-and their conversation partners were often unable to detect the rise and fall of their partner's attentiveness (whether via natural fluctuation or via our experimental interventions). Across a diverse set of studies, we find support for three key results: (a) perceptions of conversational listening often do not align with listeners' cognitive attentiveness; (b) perceptions of listening are often inaccurate due to a lack of diagnostic behavioral cues displayed by listeners-inattentive listeners behave similarly to attentive listeners; and consequently (c) perceivers primarily overestimate the extent to which their conversation partners are listening to them. These results emphasize the importance of recognizing that moments of inattentiveness happen in conversation-encouraging conversants to acknowledge, forgive, and repair these moments to achieve relational and informational success.

show abstract

Section: Previous Research On Listening—a Conundrum For Determining A...mentioning

confidence: 99%

Conveying and detecting listening during live conversation.

Collins,

Minson,

Kristal

et al. 2024

Journal of Experimental Psychology: General

View full text Add to dashboard Cite

show abstract

“…With the increasing availability of sign language corpus data produced in naturalistic settings, there has been an increase in studies examining everyday face-to-face interaction, the primordial setting of human communication [69]. Recent research reveals that human communicative interaction strives for efficiency, clarity, and minimization of breakdown or communicative trouble [57,[96][97][98]. Turn transitions happen at a very rapid rate [99].…”

Section: Explanationmentioning

confidence: 99%

“…Communicators even start planning their own response while still perceiving the previous turn [100]. Non-manual cues have been reported to largely contribute to this fast dynamic and to ensure progressivity in interaction [52,95,98,[101][102][103][104]. Visual cues appear to be best suited for this, as "the core of social interaction perception is visual" [105, p. 1165].…”

Section: Plos Onementioning

confidence: 99%

Phonetic differences between affirmative and feedback head nods in German Sign Language (DGS): A pose estimation study

Bauer,

Kuder,

Schulder

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

This study investigates head nods in natural dyadic German Sign Language (DGS) interaction, with the aim of finding whether head nods serving different functions vary in their phonetic characteristics. Earlier research on spoken and sign language interaction has revealed that head nods vary in the form of the movement. However, most claims about the phonetic properties of head nods have been based on manual annotation without reference to naturalistic text types and the head nods produced by the addressee have been largely ignored. There is a lack of detailed information about the phonetic properties of the addressee’s head nods and their interaction with manual cues in DGS as well as in other sign languages, and the existence of a form-function relationship of head nods remains uncertain. We hypothesize that head nods functioning in the context of affirmation differ from those signaling feedback in their form and the co-occurrence with manual items. To test the hypothesis, we apply OpenPose, a computer vision toolkit, to extract head nod measurements from video recordings and examine head nods in terms of their duration, amplitude and velocity. We describe the basic phonetic properties of head nods in DGS and their interaction with manual items in naturalistic corpus data. Our results show that phonetic properties of affirmative nods differ from those of feedback nods. Feedback nods appear to be on average slower in production and smaller in amplitude than affirmation nods, and they are commonly produced without a co-occurring manual element. We attribute the variations in phonetic properties to the distinct roles these cues fulfill in turn-taking system. This research underlines the importance of non-manual cues in shaping the turn-taking system of sign languages, establishing the links between such research fields as sign language linguistics, conversational analysis, quantitative linguistics and computer vision.

show abstract

“…It cannot, for example, clarify the aim of a request while the request is in progress, or modify its response as it is speaking based on the facial expression of the requester. Both of these are kinds of interactional repair, a phenomenon humans engage in frequently and fluidly in naturalistic conversation (Dingemanse & Enfield, 2024), and which plays a key role in effective communication. In short, the mechanisms of LLMs place limits upon its functional similarity with humans.…”

Section: Multimodality and Interactionmentioning

confidence: 99%

“…Even assuming models may eventually reach human-like functional competence in some pragmatic, gestural, or prosodic areas, the functional capacity for language in humans is not limited to the production of more or less well-formed strings for broadcast transmission. There are key functional elements of language that are not available to or emergent in LLMs, for example, turn taking (e.g., Casillas et al, 2016;Levinson, 2016;Stivers et al, 2009), co-speech gesture and multimodality (e.g., Goldin-Meadow & Brentari, 2017;Kita et al, 2007;Rasenberg et al, 2022), repair (e.g., Dingemanse & Enfield, 2024;Dingemanse et al, 2014Dingemanse et al, , 2015Hayashi, Raymond & Sidnell, 2013), and common ground negotiation (e.g., Brennan & Clark, 1996;Brown-Schmidt & Duff, 2016;Clark & Wilkes-Gibbs, 1986) are all essential parts of natural language in interaction. This also blurs into other aspects of our broader socio-cognitive suite including ostensive inference, perspective taking and joint attention (e.g., Heintz & Scott-Phillips, 2023;Tomasello et al, 2005).…”

Section: Multimodality and Interactionmentioning

confidence: 99%

The limitations of large language models for understanding human language and cognition

Cuskley,

Woods,

Flaherty

2024

Preprint

View full text Add to dashboard Cite

Researchers have recently argued that the capabilities of Large Language Models (LLMs) can provide new insights into longstanding debates about the role of learning and/or innateness in the development and evolution of human language. Here, we argue on two grounds that LLMs alone tell us very little about human language and cognition in terms of acquisition and evolution. First, any similarities between human language and the output of LLMs are purely functional. Borrowing the “four questions” framework from ethology, we argue that what LLMs do is superficially similar, but how they do it is not. In contrast to the rich multimodal data humans leverage in interactive language learning, LLMs rely on immersive exposure to vastly greater quantities of unimodal text data, with recent multimodal efforts built upon mappings between images and text. Second, turning to functional similarities between human language and LLM output, we show that human linguistic behavior is much broader. LLMs were designed to imitate the very specific behavior of human writing; while they do this impressively, the underlying mechanisms of these models limit their capacities for meaning and naturalistic interaction, and their potential for dealing with the diversity in human language. We conclude by emphasising that LLMs are not theories of language, but tools that may be used to study language, and that can only be effectively applied with specific hypotheses to motivate research

show abstract

Interactive repair and the foundations of language

Cited by 10 publications

References 93 publications

Conveying and detecting listening during live conversation.

Conveying and detecting listening during live conversation.

Phonetic differences between affirmative and feedback head nods in German Sign Language (DGS): A pose estimation study

The limitations of large language models for understanding human language and cognition

Contact Info

Product

Resources

About