Conversation analysis
Repair in ordinary conversation
A 1977 paper described a system by which speakers handle trouble in talk. Almost forty years later, a twelve-language study asked whether that system was a feature of English or a feature of conversation.
In 1977 Emanuel Schegloff, Gail Jefferson and Harvey Sacks published a paper in Language on what they called repair. The term was chosen carefully. Correction implies an error, and much of what they were describing involved no error at all: a speaker pausing to search for a name, restarting a sentence halfway through, or replacing one word with another that fits better. Repair covers all of it, and treats it as ordinary machinery rather than as failure.
Their central observation concerned who does the repairing. Trouble can be flagged by the speaker who produced it or by someone else, and it can be fixed by either party. Across their recordings, the overwhelming majority of repairs were both initiated and carried out by the original speaker. When another participant did intervene, the intervention usually arrived in the very next turn, and it usually stopped short of supplying a correction. A questioning repeat, a request to say that again, a puzzled particle: these hand the trouble back rather than resolve it. The authors described this arrangement as a preference, meaning a structural bias observable in the recordings, not a matter of politeness or personal taste.
The cross-language test
The 1977 corpus was American English. In 2015 a team led by Mark Dingemanse, with Seán Roberts, Julija Baranova, Joe Blythe, Paul Drew, Simeon Floyd, Rósa Gísladóttir, Kobin Kendrick, Stephen Levinson, Elizabeth Manrique, Giovanni Rossi and Nick Enfield, published a study in PLOS ONE covering twelve languages from five continents, including several with no written tradition. They examined informal conversation recorded on video and coded every case in which one participant signalled trouble with another's turn.
The same three formats appeared in every language sampled. An open request signals trouble without locating it, of which the English huh? is one instance. A restricted request locates the trouble, typically by questioning a specific element. A restricted offer proposes a candidate understanding for confirmation. The authors also reported that the open format was strikingly similar in sound across the sample, and argued that this similarity reflects the constraints of the position it occupies in conversation rather than shared ancestry between the languages.
What the design can and cannot support
The twelve languages were not a random sample of the world's roughly seven thousand, and the authors do not present them as one. The corpora are modest in size, and they consist of informal talk among people who already know each other, recorded with consent and a camera present. Institutional talk, talk between strangers and talk under duress are largely absent. A convergent result across a deliberately varied sample is a stronger argument than a result from one language, and it remains an argument from convergence rather than from representativeness.
A second objection comes from inside conversation analysis itself. Schegloff argued in 1993 that counting instances in conversation is harder than it looks, because a rate requires a denominator, and the denominator has to be the number of opportunities for the practice to occur rather than the number of minutes recorded. Establishing what counts as an opportunity is itself an analytic decision. That argument has not been settled. Quantitative cross-language work of the Dingemanse kind and the case-by-case tradition Schegloff defended continue side by side, and researchers on both sides describe the tension in print.
What the literature supports, then, is narrow and reasonably firm: conversation contains a recurrent set of practices for handling trouble, those practices favour self-repair, and closely comparable formats have been documented in every language a research team has looked at so far. Claims beyond that, about what any particular exchange means or about how a given conversation will go, are not part of this body of work.
Sources
- Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction in the organization of repair in conversation. Language, 53(2), 361–382.
- Dingemanse, M., Roberts, S. G., Baranova, J., Blythe, J., Drew, P., Floyd, S., Gísladóttir, R. S., Kendrick, K. H., Levinson, S. C., Manrique, E., Rossi, G., & Enfield, N. J. (2015). Universal principles in the repair of communication problems. PLOS ONE, 10(9), e0136100.
- Schegloff, E. A. (1993). Reflections on quantification in the study of conversation. Research on Language and Social Interaction, 26(1), 99–128.
- Jefferson, G. (2004). Glossary of transcript symbols with an introduction. In G. H. Lerner (Ed.), Conversation Analysis: Studies from the First Generation. John Benjamins.