Using Full Repeats to Initiate Repair on Others’ Questions

Abstract

This article is a conversation-analytic investigation of a previously undescribed practice of other-initiation of repair designed to locate an entire action as the repairable. This practice is implemented by a full, virtually identical, final-rising-intoned repeat of a sentential turn-constructional unit (a full repeat). This article focuses on how the full repeat is deployed to deal with one type of action, the question (i.e., a first-pair part whose primary action is one of requesting information in the vernacular sense). The full repeat delimits the nature of the trouble by working to rule out that of hearing-what-was-said and understanding-that-a-question-was-asked. The full repeat tacitly claims that a questioning action is somehow problematic. The nature of this problem can be one of either understanding the questioning action (i.e., trouble understanding the thrust of the question-as-a-whole, or what it is getting at, or what is meant by its asking), or accepting the questioning action (e.g., characterizing it as being ridiculous). Data are approximately 80 hours of audio- and videotaped mundane conversation between adult friends and family members.

The authors thank Galina Bolden, John Heritage, Jenny Mandelbaum, and Geoff Raymond for comments on previous drafts, and Deborah Tannen for giving us permission to retranscribe and use several of her data fragments.

Notes

¹ The prosodically astonished (Selting, 1988) form of What? can work to delimit the nature of the trouble to a problem of speaking (Schegloff, Jefferson, & Sacks, 1977) or “acceptability” (Svennevig, 2008), which can involve disagreement. The form Sorry? can communicate a stance that responsibility for the trouble belongs to the person who initiates repair, rather than to their addressees (Robinson, 2006).

² The prosodically astonished (Selting, 1988) form of What do you mean? can work to delimit the nature of the trouble to a problem of speaking (Schegloff, Jefferson, & Sacks, 1977) or acceptability (Svennevig, 2008), which can involve disagreement.

³ We recognize the possibility that some contexts might exist in which full repeats of questions do not work to rule out hearing trouble (for one such context, see Goldberg, 1975). However, neither our data nor Goldberg's actually contain any such cases. We do have one case (Extract A) in which hearing does appear to be the trouble, but note that the repeat is not identical, and thus not what we consider to be a full repeat:

Extract A: LOW FAT [UTCL:DO8:04]

Table

Carrie's “‘T's your speech o:n” (line 8) involves a massively truncated/clipped version of What's (i.e., “‘T's”)—which is partially overlapped by Rich's “°Ye:ah.°” (line 7), which can make hearing trouble possibly relevant (Schegloff, 2000a)—and Carrie places stress on “o:n” (symbolized in the transcript by the underline and colon). In contrast, Rich's repeat, “ What's my speech on?” includes a fully formed “ What's,” and Rich places stress on “speech.” Carrie responds with mere confirmation, “>Uh huh?<” (line 11), which orients to possible hearing trouble. There is no evidence that, for Rich, the trouble was anything more than a possible mishearing.

⁴ Differences in turn-final intonation affect both the function and nature of repeats as practices of other-initiation of repair. Not only can turn-final intonation affect the type of trouble that a repeat makes relevant, but it can affect its status as a genuine practice of other-initiation of repair. For example, see Extract B. Dan has called Sam as part of a research study designed to collect naturally occurring conversation.

Extract B: PAID [CallFriend:6869]

Table

Dan's repeat, “Do you get pa:id” (line 3), ends with final-falling intonation (symbolized in the transcript by a period), which our data suggest works to delimit the nature of the trouble to a class described by Svennevig (2008) as involving non-acceptance (see also Schegloff et al. 1977). As evidence, after Sam responds by laughing “hh .hhh h h h” (line 4), Dan chastises Sam, albeit in a good-humored manner: “you're doing a service to sci:ence my man” (lines 5–6). Here, Dan orients to his repeat as claiming that Sam's trouble-source question embodied impudence. Neither Sam nor Dan treat Dan's repeat as a genuine practice of other-initiation of repair.

As seen in Extract C, turn-final-level or continuing intonation (symbolized in the transcript by the blank underscore at line 3) can project the lack of turn completion, and thus turn continuation.

Extract C: STUDYING [UTCL:D08:04]

Table

After his repeat (line 3), Rich continues to answer Carla's question with “speech.” Again, neither participant treats Rich's repeat as a genuine practice of other-initiation of repair.

Even repeats with final-rising intonation can be produced as answer prefaces (Bolden, 2009; Schegloff, 1997), and thus not as genuine practices of other-initiation of repair. For example, see Extract 12 (in main text).

⁵ By indexing trouble with the pronominal subject, the insurance agent's partial repeat (at line 3) indexes trouble with that aspect of the surgeon's questioning action (at line 1), thereby indexing “non-acceptance” (Svennevig, 2008) of merely one component of the questioning action, rather than with the questioning action as a whole.

⁶ Although speculative, Extracts 1 and 2 may be evidence for a pragmatic rule informed by Levinson's (2000) work on interactional heuristics. Expanding on Grice's (1975) maxim of quantity (i.e., “Make your contribution as informative as is required”), Levinson presented evidence for what he called the Q-heuristic: “What you do not say is not the case.” At least regarding partial-repeat-based, other-initiation of repair, the Q-heuristic may operate as follows: “What the practice of repair initiation does not locate as troubling is not troubling for the person initiating repair.”

⁷ For example, in Extract 15 (in text), Marsha's “What happened today” (line 5) refers to the presence of police at a department store (data not shown).

⁸ We use the qualifier “ultimately treat” because Rich appears to initially design his full repeat as a genuine practice of repair initiation and not, for instance, as an answer-preface (Bolden, 2009; Schegloff, 1997; see Extract 12 in main text). That is, rather than immediately progressing from his full repeat (at line 23) to an answer of Carla's question (at line 25), Rich waits approximately 0.2-tenths of a second, in which Carla draws breath (at line 24), and thus Rich provides Carla with an opportunity to respond to his full repeat.

⁹ Pomerantz (1984) actually uses the term agreement (vs. acceptability) trouble, but here we employ the more general notion of acceptability trouble (Svennevig, 2008) in order to sidestep the related, but highly complicated, issue of “disagreement.”