Abstract
This article discusses a less-studied aspect of repair sequences in conversation, that is, their exit phases. It will be argued that while the most common way of exiting is a resumption of the main activity straight after requested repair, sometimes specific receipt objects are also needed. The focus of the article is on the use of these repair receipts. Two types of motivation for using them as exit devices are discussed: prolongation of the repair sequence and the repairers’ critical stance toward the repair initiation. The article will also consider the use of different change-of-state tokens as repair receipts in Finnish conversation. It will be argued that a claim of now-understanding (aa) is the repair receipt proper, enabling sequence closure and resumption of the main activity, while news receipts target the newsworthiness of the information provided in the repair turn, enabling sequence expansion.
Introduction
In Conversation Analytic research, other initiations of repair – the beginnings of repair sequences – have attracted a lot of attention over the past few years. For example, scholars have been interested in the comparison of different formats of repair initiation (e.g. open class, question word, (partial) repeat; see Schegloff et al., 1977: 367–368) in different languages (see, for example, Haakana et al., 2016 for Finnish; Benjamin, 2013 for English; Rossi, 2015 for Italian). These different formats have also been studied cross-linguistically (Dingemanse and Enfield, 2015; Dingemanse et al., 2014).
The study focuses on a less studied aspect of the repair sequence, that is, its exit phase. In the literature, this phase has not traditionally been considered as a part of the repair sequence (however, see Floyd et al., 2016: 178–180; also Dingemanse et al., 2014 fn 1; Couper-Kuhlen and Selting, 2018: 139). Instead, a typical way of presenting the structure of a repair sequence is as follows: (1) trouble source turn (T − 1), (2) repair initiation (T0) and (3) repair (T + 1) (see Dingemanse et al., 2014; Enfield et al., 2013: 345–346). In this article, the starting point is an observation that there are (at least) two alternative ways of exiting a repair sequence: 1) the resumption of main activity straight after the requested repair, without a sequence-expanding receipt turn (see Haakana, 2001: 53; Heritage, 1984: fn 18; Schegloff, 2007: 116; on resumption and continuation in other type of context, see Mazeland, 2007) or 2) the resumption of the main activity after a specific receipt item produced by the repair initiator (see Heritage, 1984 on oh). Consider the following example from Finnish conversation for the first alternative. The example starts with a question which is the trouble source turn (T − 1) and is followed by a repair initiation (T0) and a repair which is a repeat of the problem source turn (T + 1):
In this example, the open class repair initiator (Drew, 1997) mitä (‘what’) targets the prior question. As a solution to the apparent hearing problem, the recipient repeats the question (line 4). After that, the repair initiator answers the pending question and thus resumes the base sequence (line 5). Sequentially speaking, the adjacency pair [repair initiation + repair] is inserted between the base first pair part and base second pair part, thus forming a post-first insert expansion (Schegloff, 2007: 101–102). The resumption then constitutes the exit phase. That is, providing the pending answer shows that the problem has been successfully repaired and no extra assurance of the successfulness of the repair is needed. Overwhelmingly, repair sequences that occur as post-first insert expansions (that is, between the original question and the answer) are not closed with a receipt item.
In the following example from a dinner table discussion between six young men, the repair sequence is organized differently:
In this case, the ongoing activity is different: Lauri is telling a story. At the point where the multiunit turn (the story) is recognizably incomplete – after a reported first pair part (lines 5–6) – Taavi initiates repair with question word kuka (‘who’, line 8) thus claiming difficulty in recognizing a person reference in Lauri’s story. The repair initiation thus suspends the ongoing main activity. The repair turn involves a recognitional description (se Antin tyttöystävä ‘that Antti’s girlfriend’, line 10). Immediately upon the completion of the requested repair, Taavi claims recognition with the particle aa (line 11). The particle can be considered a third position receipt (a sequence closing third, Schegloff, 2007) with respect to the repair initiation. It is only after this receipt when Lauri resumes his unfinished story (line 13). In this case, a repair receipt from Taavi is needed to assure Lauri for the sufficiency of the repair turn. In contrast to example 1, then, the repair initiator is not the one that can resume the main activity straight after the successful repair. Instead, it is the other party that needs an explicit signal of problem resolution in order to resume the suspended activity. The suspended activity thus ‘belongs’ to the original speaker, not the repair initiator.
While the most common way of exiting a repair sequence appears to be the one illustrated in Extract 1 (Haakana, 2011: 53; Schegloff, 2007: 116), this article will focus on cases like Extract 2, that is, the role of repair receipts (Heritage, 1984) or repair uptake (Golato and Betz, 2008) in exiting repair sequences. This concept was introduced by Heritage (1984: 315–318) who claimed that one of the main sequential environments for English oh is its use as a repair receipt in other-initiated repair sequences. In this position oh proposes a change-of-state of information and, by implication, problem resolution. I will suggest that even though the use of a receipt is not the most frequent exit strategy (at least not in the current data), in some repair sequences its use is actually mandatory and its absence would prevent the progression of the main activity (see also Betz et al., 2013: 152). I will consider the use of repair receipts in Finnish, and more specifically, Finnish change-of-state tokens as repair receipts.
There are two main issues to be discussed:
The use of repair receipts in repair sequences (in Finnish conversation): what motivates their use?
The linguistic form of a repair receipt: how do the different Finnish particles used in this context differ from one another with respect to sequence development?
Data
As data, I am using a collection of repair sequences compiled for the project ‘Repair Practices and Understanding in Interaction’, led by Markku Haakana and Salla Kurhila (funded by the University of Helsinki during 2011–2013). The project studied formats of repair initiation in Finnish everyday conversation (see Haakana et al., 2016). The data for this project came primarily from the Conversational Data Archive at the University of Helsinki, recorded during a longer period of time (1980–2000), but it also entails fairly recent video recordings from the early 2010, recorded for the purposes of the project. The database consists of both telephone and face-to-face conversation. 1 The total amount of data is 37.5 hours of conversation involving 461 cases of repair initiation (on the distribution of different formats of repair initiation in this collection, see Haakana et al., 2016). 2
In this data, I found 95 sequences that involved a third position turn produced by the repair initiator. When making this collection, I included all cases where the repair initiator produced a particle response (or sometimes even a more substantial receipt turn involving repetition of the previous turn, see Extract 4, or an account for initiating repair, see Extract 5) to the repair turn. Roughly, these different third position objects vary in terms of whether they function to close the sequence or whether they enable more talk on the topic. I will specifically focus on what I call repair receipts proper. These are mainly claims of now-understanding or recognition that are used to close the repair sequence and resume the main activity. Other particle responses in the data – while similarly signaling problem resolution – are different in the sense that they may also target the newsworthiness of what is said in the repair turn and thus promote sequence expansion. I will term them other receipt objects. This distinction between repair receipts proper and other receipt objects will be further elaborated in section 4. The general terms repair receipt and third position turn are used to refer to all cases included in the collection. The table below summarizes the numerical information relevant for this study (Table 1).
Repair receipts in the data.
Considering the relative scarcity of the use of third position turns in the data, one needs to ask what motivates or accounts for their use. In terms of type of repair initiation, the most typical problem associated with their use is problem of recognition (e.g. unclear person reference as in Example (2)), that is, a repair receipt such as a claim of recognition may be needed to indicate problem resolution in a context where the repair initiator has initially failed to recognize a referent. However, there are also more general explanations for the motivation of repair receipts, which encompass problems with recognition and other types of problems indicated with different formats of repair initiation. I will now turn to the two most prominent motivations, prolongation of the repair sequence (section 3.1.) and the co-participant’s resistance toward the act of initiating a repair (section 3.2.).
What motivates the use of repair receipts?
Prolongation of the repair sequence and other progression-related troubles
The most common clearly distinguishable motivation for the use of a repair receipt as an exit device is the prolongation of the repair sequence, that is, situations where the first attempt to solve the problem does not solve it but an additional attempt(s) is (are) required (see also Floyd et al., 2016). This definition covers both cases where the repair initiator resorts to more than one repair initiations and cases where there is noticeable absence of a sequentially relevant next turn after the repair, resulting in another attempt to repair the problem – that is, without an explicit second repair initiation (see Extract 6). According to Heritage (1984: 318), repair receipts can be found both in simple repair sequences and in the extended ones, but in light of my data, prolongation seems to be a specific motivation: out of 64 cases involving a repair receipt proper, 28 occur in prolonged repair sequences. There are also other, less frequent troubles that relate to progressivity. One is lateness of the repair initiation in relation to the trouble source. The other is the position of the repair initiation within the ongoing activity, that is, suspensions of some recognizably unfinished activities such as story-telling may call for a repair receipt that explicitly allows the original speaker to resume the unfinished activity. Extracts (2) and (6) provide an example of this type of situation. It should also be noted that these different interactional contingencies also occur in combinations.
Let us now look at an example of a prolonged repair sequence. In Extract (3), the trouble source turn involving an unidentifiable referent triggers three separate repair initiations and several attempts to solve the problem of recognition. A middle-aged man named Raimo and a young man named Teemu (probably a father and a son) are discussing an invoice that Raimo received in his address but that was addressed to Teemu. The reason for Raimo’s call is to inform Teemu about the invoice (lines 6–7):
Teemu’s first reaction to Raimo’s informing is a slightly delayed and hesitantly produced joo (line 9). It receipts the informing but does not display any orientation to why this informing was issued – at this point it can of course be unclear. After a silence in line 10 Raimo extends his turn (turn-initial ja ‘and’) by saying that the invoice belongs to Teemu. It is left for Teemu to decide how to respond to this information: he could, for example, offer to pick up the invoice. Instead of orienting to the implications of the informing, Teemu initiates repair (line 13). Initiating repair at this point is in fact delayed – the problematic referent was first treated as understandable with joo (‘right’) in line 9 and thus the opportunity to initiate repair was initially passed (Schegloff, 1982: 87–88; see also Sorjonen, 2001: 26). The format of the repair initiation is question word + repeat (minkä lasku ‘which invoice’). Latched onto that is another repeat (ÄsGeen ‘SG’s’) which can have two interpretations in this context: either it asks for confirmation for Teemu’s hearing or hints at a trouble of understanding, or more specifically, recognition (see Haakana et al., 2016: 264–267). Raimo treats the repeat as a candidate hearing, now producing a confirmation (line 15).
After the confirmation, there is a clear place for claiming understanding or moving on with the conversation. However, a silence develops (line 16). Teemu’s failure to close the repair sequence results in its expansion: Raimo treats Teemu’s silence as a failure to recognize the invoice by providing another specification (line 17). However, this does not solve the problem either but leads to another two rounds of repair initiations, in the form of a question word (minkä ‘which one’, line 19) and a repeat (↑ tietokonejuttuja. ‘computer stuff’, line 30) and several attempts to provide a specification that would make the invoice recognizable. In line 32 Raimo produces another company name that finally triggers a claim of recognition from Teemu in line 34 – though after a pause. The first turn-constructional unit (TCU) of this turn is composed of the particle aa (in a lengthened form and with some rise-fall pitch movement) and multiple sayings of joo. Aa claims now-understanding and thus problem resolution (Koivisto, 2015b) and the multiple joos target the prolonged repair sequence as a whole and clearly propose sequence closure (Stivers, 2004). After the particle response, Tommi continues with an explicit claim of recognition (n
Extract (3) illustrates that the pressure for an explicit sign of problem resolution, such as a claim of recognition, may build up as a result of an initial failure or failures to repair the problem. Prolongation of the repair sequence is one possible explanation for the need of using a repair receipt and thus providing a clear exit from the problematic repair sequence. It can be argued that the repair initiator orients to the accountability for the problematic repair sequence by claiming now-understanding and thus allowing resumption of the main activity (see also Koivisto, 2015b). I will now turn to cases where the claim of recognition is motivated by the way in which the producer of the requested repair treats the repair initiation.
Co-participant’s resistance toward the initiation of a repair
The common feature of cases to be discussed in this section is that the producer of the repair treats the repair initiation as in a way or another obvious or unnecessary, that is, inapposite (on indicating inappositeness from second position see also Halonen and Lappalainen, 2015; Heinemann, 2009; Heritage, 1998; Stivers, 2011). The implication is that the repair initiator should have, for example, recognized the referent without any further clarification. In the repair receipt collection, clearly observable resistance is found in 15/64 repair receipt proper cases, often coupled with prolongation.
Resistance can be indicated by laughing voice in the repair turn, certain prosodic means such as high onset (Halonen and Lappalainen, 2016), or by marking the referent mutually identifiable with a determiner (Laury, 1997). We already saw an example of the use of a determiner in Extract (2), se Antin tyttöystävä ‘that Antti’s girlfriend’. In Extract (4) the repair initiation is treated as unnecessary and thus unexpected with (disbelieving) laughter incorporated in the repair turn. This is an extract from a phone call where S and V, two men in their thirties/forties, are making plans for a get-together in a restaurant in Helsinki city center. The restaurant is located on a street called Kasarmikatu. S’s question in line 1 reveals his unawareness about the location of the street (even though he is the one that made the suggestion):
[
In response to S’ question V tries to illustrate the location of the street by using different landmarks. They do not turn out to be helpful: S initiates repair (line 4, 12, 20), claims lack of knowledge (line 6) and gives a vague/evasive answer (line 14). V checks S’ ability to recognize the landmarks by asking ‘does that mean anything (to you)’ several times (lines 10, 11, 19) but does not get a clear yes-answer. 3 The repair initiation of our interest takes place in overlap with the third occurrence of this question (line 20). It thus initiates a post-first insert expansion, arguably targeting a name of a church mentioned in line 18. The format of the repair initiation comes close to the format question word + repeat, but the ‘repeat’ part is actually the speaker’s own formulation of a more general nature, a superordinate concept, if you will. That is, instead of asking ‘which church’, S asks ‘which building’. The repair initiator is also possible to interpret as targeting a previously mentioned referent, ‘Soviet Union embassy building’ (line 17) which, in fact, contains the word ‘building’. Be that as it may, this formulation suggests that he picked up the fact that some building was mentioned but he did not hear or recognize what it exactly was.
V interprets the repair initiation targeting the just-prior mentioned referent St John’s church by repeating it (line 21). What is noteworthy is that the word kirkko (‘church’) is now produced with laughter. This suggests that V treats the act of initiating repair as unexpected in the sense that S should know the church and should thus not need to initiate a repair. It is also possible that at this point the laughter is a reaction to the fact that the explaining activity has been going on for a long time without a result. That is, while providing the requested repair, laughter indicates resistance toward the initiation of repair (cf. Haakana, 2002: 224–226).
The repair initiator’s own orientation to the repair initiation as unwarranted can be seen in the formulation of his third position turn in line 22. The turn starts with confirmatory particle nii (which is not easy to translate in this turn-initial position) and a repeat of the problem source (on repeats in third position in German, see Betz et al., 2013; see also Schegloff, 2007: 126). Applying the observations made of repetitional answers to questions (see Heritage and Raymond, 2005, 2012; Raymond, 2003; Schegloff, 1996; Stivers, 2005), one could argue that compared to a mere particle response, repeat implies more epistemic agency over the information provided by the repair turn (on this line of argumentation, see also Koivisto, 2013). The turn-initial particle nii and the high onset of the repetition seem to add to the impression of now-understanding/recognition, that is, S now claims in retrospect that he does know the church (and where it is located), thereby constructing the problem as a hearing problem.
The repetition is followed by the reduplicated particle joo. The first joo is also produced with high onset, which is a way of treating the prior turn as newsworthy (Kunnari, 2011), that is, while the first part of the turn (nii + repeat) is devoted to claiming recognition and epistemic agency, the second part of the turn (joo joo) seems to treat Johanneksen kirkko as new(sworthy) information in the sense that it finally helps S to locate himself ‘on the map’, which makes the church a successfully chosen landmark considering the street he is supposed to find. The first part of the composite turn thus has a local scope as a repair receipt and the second part targets the informing that was issued prior to the initiation of repair, marking it as understood. In other words, this case is a demonstration of the fact that sometimes a composite third position turn is needed to deal with all aspects of the preceding talk so that the repair sequence can be drawn successfully to a close and the main activity resumed (see Schegloff, 2007: 130).
Ways of resisting the relevance of repair may also be less subtle. Let us look at another example where two middle-aged sisters are chatting and having some drinks. This example involves both prolongation of repair sequence and very overt critique toward the co-participant’s inability to recognize/locate a referent:
The sequence starts with Jaana’s suggestion to have some liqueur (line 1) and Tuula’s agreeing response to that (line 2). Even though Tuula is the hostess, it is Jaana who starts to organize the service – probably because Tuula is holding a cat in her arms. Jaana asks Tuula about the location of suitable glasses (line 6) and gets instructions (lines 8, 10). Jaana then stands up to get the glasses; at this point she goes off camera, so the analyst is unable to detect her movements. At first, Tuula does not pay attention to Jaana’s attempt to find glasses but focuses on stroking her cat, thus orienting to the search activity as unproblematic.
However, Jaana’s turn in line 15 shows that she does not know exactly where to look. She first initiates candidate-understanding-type repair siis kulma (‘you mean corner’). Then she produces a deictic element täs (‘here’), probably pointing at a potential cabinet (line 15). Tuula starts to reiterate her instruction (kulma ‘corner’) but cuts off to give a disconfirming answer to Jaana’s proposal (line 16). This is followed by an explicit other-correction (‘that’s not in the corner at all’, line 18), or maybe, a complaint, and yet another, enunciated mention of ‘corner cabinet’. It is prefaced with the particle siis which has an explanatory function here (see Hakulinen and Couper-Kuhlen (2015) (line 20). It becomes apparent that the two women have a different understanding of what ‘corner cabinet’ means. In line 22 Jaana claims to have now understood what Tuula refers to by producing a turn composed of particle aa and another deictic element tää (‘this’). Again, a claim of understanding, a repair receipt, is needed for a ‘mutually ratified exit from the repair sequence’ (Heritage, 1984: 318).
The repair sequence is not only prolonged but treated as unnecessary or ‘stupid’: after Jaana’s claim of now-understanding, Tuula calls her sister ‘dummy’ (line 23). As a result, Jaana claims lack of knowledge to account for her inability to locate the right cabinet (lines 24, 26). Post-resolution talk such as accounts is not in fact uncommon in my data – it reflects the extent of the transgression oriented to by the participants. It is only after these ‘post-completion musings’ that do ‘not extend the sequence but reflect on it’ (Schegloff, 2007: 143) that Tuula resumes the main activity by instructing Jaana on the next step of finding the right glasses (line 27).
The examples in this section have shown that repair receipts are needed when there is something problematic in the progression of the repair sequence or the act of initiating a repair itself. Delay, suspension of main activity and prolongation are progression-related problems; initiation-based problems are co-participant’s orientation to the initiation as inapposite. Both kinds of problems relate to the transgression attributed to the repair initiator which, in turn, relates to the problems of progressivity caused by the problematic repair sequence. It has also become apparent that typically the trouble source is something other than a question and the type of problem is problem of recognition.
In this section, we have not systematically analyzed the composition of repair receipts used in the examples. In most cases, we see the particle aa that serves to claim now-understanding, but also partial repeats of the trouble source can be used to claim restored access to the information (Extract 4). We will now turn to the variety of possible particle responses available in Finnish language and how they differ in terms of sequential progression. While most particle responses may be classified as change-of-state tokens (as English oh, Heritage, 1984), there is a difference whether they claim now-understanding (repair receipts proper) or mark a receipt of new information (other receipt objects).
Different particle responses in repair sequences
This section will show that different particle responses used as repair receipts differ with respect to how they regulate sequence development after a successful requested repair. I will show that a particle claiming now-understanding effectively closes a repair sequence, while particles that are used for receipting new information have a double duty: they signal problem resolution but simultaneously also target the newsworthiness of the repair turn (or originally, in the trouble source turn) and thus work to expand the sequence.
For claiming now-understanding there is one particle that is specialized in this task, namely aa that has already been mentioned in this article (see Koivisto, 2015b). For treating something as new information, there are an abundance of particles such as ai, aijaa, aha(a), jaa, ja(a)ha, mhy, vai niin (see Hakulinen et al., 2004; Sorjonen, 1999). Except for the distinction between aijaa and aha(a), the division of labor between these particles is still unclear (Koivisto, 2016). The task of this article, however, is not to tackle those differences – we will operate on a bit more general level, focusing on the sequence-closing work of aa and sequence-expanding work of news receipts. It should be noted, though, that news receipts occur in repair sequences far less frequently than claims of recognition/now-understanding (31/95 cases of all third position turns).
Examples of aa as a repair receipt have already been seen in Extracts 2, 3 and 5. The clearest examples illustrating its central features, that is, signaling problem resolution and closing implicativeness, are those where some suspended activity is resumed after the production of aa. Let us look at one more example of this kind. Two young women are talking on the phone:
In the beginning of this extract, Veera is in the midst of telling a story about her bumpy bus ride. At a point where the story is recognizably incomplete, Silja initiates repair with question word + repeat (minkä yli ‘over what’, line 8). The trouble source is Veera’s choice of word kanttari (line 5), which is a less common word for curb in Finnish. In line 9 Veera first treats the problem as a hearing problem by repeating the problematic word, but when she does not receive any response (see pauses in line 10), she gives a synonym, now orienting to a problem of understanding. In overlap with this, Silja claims recognition with aa + joo (line 12). This seems to be a recognitional overlap (Jefferson, 1983) that indicates the exact moment when Silja realizes what Veera is talking about. Note, however, that there is no actual demonstration of the now-achieved understanding, merely a claim (cf. Schegloff, 1982: 78). Nevertheless, it is clear that this claim of now-recognition serves its purpose in enabling the resumption of the suspended activity: immediately after its production, Veera continues her story (see line 13). In other words, no more talk about the problematic word kanttari is needed.
Particles that I group together as news receipts or news particles behave differently in terms of sequential progression after the repair solution. Let us look at the next example where a young woman (Jutta) and a young man (Tommi) are discussing Jutta’s husband’s/boyfriend’s work situation over the phone:
The trouble source turn is Jutta’s answer in line 6 that gives a disconfirming answer to Tommi’s question about whether her husband still works at the same place (lines 1 and 4). Tommi initiates repair after this answer and a slight delay (line 8). The format of the repair initiation is a question word (missä ‘where’, line 8), which does not reveal whether the problem is in hearing or recognition. In sequential terms, a repair initiated in this position – after an answer – is a non-minimal post-expansion (Schegloff, 2007: 149–151). This means that it comes at a point where Tommi could otherwise appreciate the answer in terms of, for example, its newsworthiness. This action is now delayed. As an attempted repair, Jutta does not merely repeat the problematic referent but also describes what the current work entails (line 10). In his response, Tommi does not claim now-understanding or recognition but treats the repair turn as having provided new information. He does this with the news particle jaa produced with high onset and accompanied with a claim of not-knowing (line 12). In line 14, he expands the sequence by asking a follow-up question. Jutta’s husband’s new job thus becomes a topic of conversation.
In the next example, the exit from the repair sequence is organized similarly. P and E, two young women, are trying to find a solution to E’s problem: how to get home from a party they are planning to go to:
.
In line 1 P offers a candidate solution to the problem by asking whether E can afford to take a taxi. Attached to the question is a candidate answer ‘no’, which means that P is strongly oriented to a negative answer. And indeed, E’s answer is negative, but it comes with an account. In her account she offers some information marked as mutually identifiable by using a locative demonstrative pronoun sinne as a determiner (‘to that Helsinki’) (see Laury, 1997: 145). This formulation is the trouble source. In line 3, P initiates repair with the format question word + repeat (‘which/what Helsinki’): she claims lack of access to what ‘that Helsinki’ stands for. In her response, E explains why she is going to Helsinki, again using a determiner (ne mielenterveysmessut ‘that mental health fair’, line 4) to suggest that the information is mutually shared. P, however, does not claim late recognition but rather resists the claimed common ground: she receipts the turn as having provided new and newsworthy information by producing the news particle aijaa.
Aijaa is the most commonly occurring news particle in Finnish, specialized in receipting newsworthy and thus topicalizable information (Koivisto, 2015a, 2016). High onset adds to this effect. Aijaa opens up a space for elaboration, but since E does not volunteer one, P asks a follow-up question (line 7). Moreover, in line 9 she evaluates the piece of news. The trip then becomes the topic of conversation – even though from the perspective of how much it is going to cost. Similarly to Extract 7, then, the news receipt and the follow-up question can be seen as sequentially fitted responses to the initial informing (here, line 2), only as postponed ones. This means that there is no return to the main activity, as in the aa cases, but the content of the repair turn becomes the main activity (i.e. topic).
In addition to functioning as news receipts, both Tommi’s jaa in Extract (7) and P’s aijaa in Extract (8) also signal that that the problem of recognition is now resolved. However, this kind of particle response – together with the following claim of not knowing and/or a follow-up question – does not specifically deal with the fact that the problem is now resolved. Rather, it provides the pending appreciation to the prior (initially problematic) informing, that is, it is a fitted response to the question-answer sequence. This double-duty is reminiscent of what Schegloff (2007: 76) calls ‘double-barreled’ actions: while treating the information provided in the prior turn as new information the news particle response simultaneously gives evidence for the fact that the problem of understanding is now resolved.
Conclusion and discussion
In this article, I have discussed repair sequences from the perspective of different exit practices. First I pointed out that the most common way of exiting (at least on the basis of the current Finnish data) is just resuming the suspended activity. This seems to be almost the exclusive way in repair sequences that are inserted between a question and an answer, as a ‘post-first insert expansion’ (Schegloff, 2007: 100–101). After a successful repair solution, the repair initiator may continue just by answering to the pending question. In contrast, specific repair receipts which function as post-expansions of the repair sequence are typically found in contexts where the trouble source turn is not a question. Instead, in the repair receipt cases the delayed turn/activity belongs to the original speaker who cannot resume without a token of understanding that licenses the resumption.
The analysis has shown that a typical type of problem associated with the use of repair receipts is problems of recognition. In a previous work, I showed that the Finnish aa, a particle claiming now-understanding, is often used in repair sequences (Koivisto, 2015b). Correspondingly, when the problem indicated by a repair initiation has to do with recognizing something, aa offers an apt solution: it shows that late recognition is now achieved. Late recognition can also be done by repeating (a part of) the repair turn and thus claiming prior, now-restored epistemic access, as we saw in Extract 4. However, in order to claim that sufficient understanding has also been achieved – so that the repair sequence can be closed – particles claiming understanding are also needed. Mere repeat in third position may be equivocal in terms of whether it receipts information or initiates repair (e.g. Kurhila and Lilja, 2017; Schegloff, 2007: 126; Thompson et al., 2015: 60–64).
Besides sequential position and the nature of the repair initiation, I also discussed more general motivations for using repair receipts. It seems that each time a repair receipt is used/required, there are specific interactional contingencies that make it relevant. The most important ones are different kinds of disturbances in the progression of the repair sequence or the placement of the repair initiation in relation to the trouble source, that is, a pressure toward a sign of now-understanding increases if the repair sequence gets prolonged (i.e. the first attempted solution is not successful). Similarly, if the repair initiation clearly cuts off the ongoing main activity (e.g. story-telling) and/or the repair is initiated late with respect to the trouble source, an explicit sign of problem resolution may be in order. The other recurrent motivation for using repair receipts is the co-participant’s (i.e. the one who has been requested to produce a repair) stance display toward the initiation of repair, that is, for example laughter or verbal/prosodic markers of mutual, shared knowledge suggest that the repairer does not consider the repair initiation necessary but rather inapposite ( ‘stupid’) and thus unexpected. The subtleness of these cues of inappositeness may vary (see Extract 5 for very overt criticism), but the point is that they seem very effectively to trigger a claim of now-understanding at the earliest possible moment.
In the latter part of article, I compared two kinds of change-of-state tokens used as repair receipts in Finnish language. The division was made between a claim of now-understanding (aa) and a news receipt (such as aijaa, jaa, aha(a)). A claim of now-understanding (aa) signals problem resolution and functions as an exit device. Thus, it strongly promotes sequence closure. By using aa, the repair initiator orients to the delay of understanding as problematic, that is, claims responsibility for the delay in understanding/recognizing something in a timely manner (see also Koivisto, 2015b). It is thus the repair initiator who ‘takes the blame’, and aa is used as a means of restoring intersubjectivity.
By contrast, news receipts as responses to requested repair have the capability of topicalizing the information provided in the repair turn. We saw two examples where the repair was initiated as a ‘non-minimal post expansion’, that is, where a third position acknowledgment could have otherwise been produced. Instead of specifically dealing with the repair turn and its successfulness in repairing the problem, new receipts (together with follow-up questions) targeted the newsworthiness of the information provided in the repair turn (and originally in the problematic second position turn). This also means that the act of initiating a repair and the problem is backgrounded and not specifically dealt with. News receipts (are more likely to) promote sequence expansion. The discussion of the different particle responses thus shows that through a closing-implicative repair receipt the repair sequence may remain ‘local’, thus forming an activity-suspending insert expansion. Through a news receipt, by contrast, the repair sequence merges with the ongoing larger activity, and the line between ‘dealing with a problem of hearing/understanding’ and ‘receipting new information’ gets blurred.
This article has hopefully shed some light on a less-researched aspect of repair sequences, that is, its exit practices and, more specifically, the use of change-of-state tokens as repair receipts. In addition, the article has hopefully demonstrated that comparing different change-of-state tokens in a similar environment gives evidence for their distinct interactional meanings.
Footnotes
Acknowledgements
I would like to thank Scott Barnes and the reviewer Mark Dingemanse for useful and accurate feedback on previous versions of this paper. My special thanks to Markku Haakana and Salla Kurhila and their project members for giving me access to their data. I also thank Salla Kurhila for checking the final version.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was conducted in the Centre of Excellence in Research on Intersubjectivity in Interaction, University of Helsinki, funded by the Academy of Finland (Project Number 250099).