Purdue University Graduate School
Yamamoto Dissertation 06.16.2022.pdf (1.94 MB)


Download (1.94 MB)
posted on 2022-06-16, 17:43 authored by Mai YamamotoMai Yamamoto


It is well established that people are highly sensitive to timing in interaction and that delay of response is consequential in shaping the interaction (Davidson, 1984; Jefferson, 1989; Pomerantz, 1984; Roberts, Francis, & Morgan, 2006). There is also crosslinguistic evidence that people of different language backgrounds orient to these delays as problematic, though with different levels of tolerance (Roberts, Margutti, & Takano, 2011). However, few studies, if any, have focused on the timing of responsive actions such as laughter or open class repair initiators (OCRI; e.g., English “what?” or Japanese nani). Both types of actions fill the slot of responses but are semantically-underspecified. They provide no overt comment on the speaker’s talk; although laughter does have a sociocultural association with humor and positive affect. In effect, laughter and OCRI are “agnostic” responses whose interpretations are highly dependent on contextual factors. This dissertation explores the phenomena of laughter and OCRIs in relation to timing and delay and the possibility of crosslinguistic variation by examining delayed laughter and delayed OCRI in American English and Japanese via two inter-related studies.

The first phase examined delayed laughter and delayed OCRI in naturally-occurring conversation (telephone calls) using the analytic framework of Conversation Analysis (CA). The data for this analysis came from the CallFriend – English (Northern US) and CallFriend – Japanese corpora of audio-only telephone call conversations between friends and are accessible through the TalkBank database (MacWhinney, 2007). The English corpus contains 31 half-hour conversations, and the Japanese corpus consisted or 32 half-hour conversations. The results suggest little crosslinguistic variation in the sequence and function of delayed laughter and delayed OCRI. However, the results of this phase indicate that delayed laughter is often found as a hearer’s response to the speaker’s delicate and is typically oriented to by the speaker as an affiliative signal, unless additional indications of disaffiliation also co-occur in the interaction. Delayed laughter also functions as a forecast of an upcoming action that may spark trouble in the interaction or pose a face threat (such as an imperative.) On the other hand, delayed OCRI seem to indicate the hearer’s surprise at some sequential trouble, such as a disjunctive topic, or in a sequence of disaffiliation.

The second phase takes these insights gleaned from the CA phase and investigates them experimentally by asking native speakers of English and Japanese to report their perceptions of laughter or OCRI at one of three different delay lengths. Participants for this study were recruited via the online crowdsourcing platforms Amazon Mechanical Turk (for English native speakers; n = 413) and Crowdworks (for Japanese native speakers; n = 240). Participants were asked to complete an online Qualtrics survey in which they listen to short, simulated telephone calls and rate their perceptions of agreement, surprise, and avoidance in the context of a recipient’s delayed laughter or OCRI. The experiment was conducted as a between-groups design, so for each language group, participants were exposed to either the laughter or the OCRI condition at one of three delay levels (0ms, 400ms, or 600ms). The results support the CA findings, as the participants reported that laughter expressed more agreement and avoidance than OCRI, and OCRI expressed more surprise than laughter, regardless of delay length. As with the CA findings, there was little crosslinguistic variation, but there was indication that native English speakers were more likely than Japanese speakers to perceive laughter as an expression of agreement and OCRI as an expression of surprise.

The overall results suggest that delayed laughter is typically oriented to as an affiliative action by the co-participant while OCRI is more of a neutral action that has the potential to be a face-threatening action. Because these non-linguistic features are understudied, this dissertation contributes to our understanding of how the perceptions of seemingly agnostic responses are shaped by temporal characteristics and interactional sequence. 


Degree Type

  • Doctor of Philosophy


  • Linguistics

Campus location

  • West Lafayette

Advisor/Supervisor/Committee Chair

Felicia Roberts

Additional Committee Member 2

Lori Czerwionka

Additional Committee Member 3

Atsushi Fukada

Additional Committee Member 4

Myrdene Anderson

Usage metrics



    Ref. manager