Request for confirmation sequences in Korean

4.1 Syntactic design

The majority of the RfCs take the clausal form, marked with polar declarative endings (n = 168/200, 84%), followed by the clausal form with polar interrogative endings (n = 19/200, 9%), the clausal form with connectives (n = 8/200, 4%), and the phrasal form (n = 5/200, 3%). Table 1 shows the distribution of RfCs in terms of syntactic design.

In Korean, questions are predominantly constructed as declarative questions (Yoon 2010), marked by a declarative ending with a rising intonation (Note 7). About a half of the clausal RfCs marked with SESs are declarative questions constructed as such; there are 79 out of 168 cases (47%) where they are marked by the informal ending suffix -a/e (or -ay in the case of the verb ha ‘do’) or with the politeness marker -yo.

Other grammatical resources frequently used for formulating RfCs in Korean include ‘pseudo-tags’, marked by the ‘committal’ suffix -ci or its related form -cianha (ci:NEG).^[4] RfCs with pseudo-tags constitute about 40% (n = 68/168) of the RfCs taking the form of a polar declarative question (Section 4.6).^[5]

The remainder of the declarative RfCs (n = 21/168, 13%) are formatted with an array of SESs that include -ney (‘factual realization/noticed’), -tay (‘hearsay’), -takwu (‘quotative’), -kwuna (‘inferred’), -(ta/la)may (‘hearsay’), -ta (‘declaration’), and -tela (‘retrospective’).^[6]

There are 19 instances in the data of RfCs constructed with interrogative endings (n = 19/200, 9%), which include dubitative particles (-na/nka/lkka) and interrogative particles indexing intimacy (-ni/nya) (Yoon 2010).

RfCs marked with a clausal connective (n = 8/200, 4%) and phrasal RfCs (n = 5/200, 3%) are both produced in the form of an ‘appendor question’ (Sacks 1992b, 559), tied back to the recipient’s prior turn (Goodwin 2013; Extracts (4) and (5)).

The following shows types of RfCs in terms of syntactic design.

4.2 Polarity

There are 139 instances of positively formatted RfCs (n = 139/200, 70%) and 61 negatively formatted RfCs (n = 61/200, 30%).

Negatively formatted RfCs constitute a rich set of grammatical resources in Korean used for organizing requesting actions,^[9] in a way that exhibits the questioner’s delicately nuanced, obliquely asserted, or normatively grounded stances. Among the negatively formatted RfCs are 43 RfCs formatted with ‘verbal negation’, which include pre-verbal negation (an/mos V) (n = 13/43, 30%), post-verbal negation (V-ci anh (n = 21/43, 49%)), and post-nominalization negation (V-nun-ke ani (n = 9/43, 21%)). As Kim (2023a) shows, these three forms of verbal negation shape RfCs into distinctive interactional resources. For instance, post-verbal negation RfCs, formatted with the post-verbal negation -ci anh (‘committal’ suffix -ci + auxiliary negative verb anh), index the questioner’s ‘problematizing’ stance toward the recipient’s (or a shared) domain. Consider Extract (6), where Jiho uses the post-verbal negation RfC topic-initially in problematizing a test question made by Hani, his fellow TA (‘Question No. 5 is a little strange, isn’t it?’):

Extract (6): NC Talk: TA Meeting 16:47, No. 130

With his post-verbal negation RfC, Jiho seeks agreement, prodding Hani, the epistemically privileged recipient (as the one who made the test question at hand), to be acquiescent to his terms (also Extract (7); Section 4.3).

This feature renders post-verbal negation RfCs distinct from the other two types of negatively formatted RfCs, which are marked with pre-verbal negation (an ‘not’ or mos ‘not able’) and post-nominalization negation (-nun-ke ani ‘ATTR-thing not:COP’). Unlike their post-verbal counterpart, RfCs with pre-verbal negation tend to be more distinctively ‘other-attentive’, to the effect that the recipient is prompted to attend to his/her own epistemic domain in regard to what the questioner inferentially formulated as a ‘negative event’ therein (Pomerantz 1988, Schegloff 1988, Heritage and Raymond 2021; Extracts (1), (2), and (12)). The use of post-nominalization negation RfCs has a deontic and normative character, bringing a ‘deviant’ or otherwise ‘normatively discrepant’ aspect of the confirmable to the attention of the recipient as an object to be co-assessed on shared normative reasoning (Kim 2015b, Kim 2023a; see Min’s RfC in Extract (17)).

In terms of types of responsive action, the majority of positively formatted RfCs engendered confirmation (n = 85/139, 61%), with only 17% being responded to by disconfirmation (n = 24/139, 17%). Negatively formatted RfCs were found to be less likely to be responded to with confirmation (n = 31/61, 50%), and slightly more likely to be responded to with disconfirmation (n = 15/61, 25%; Section 5.1).

Among the three types of negatively formatted RfCs, post-nominalization negation RfCs were less likely to be responded to with confirmation (n = 2/9, 22%) than pre-verbal negation RfCs (n = 7/13, 54%) or post-verbal negation RfCs (n = 14/21, 66%). Even though the number of tokens is small, 67% of post-nominalization RfCs engendered disconfirmation or ‘neither’ (confirmation nor disconfirmation)-type response (n = 6/9), whereas 46% of pre-verbal negation RfCs (n = 6/13) and 24% of post-verbal negation RfCs (n = 5/21) did so. A preliminary observation suggests that this skewing may be attributed to the function of the nominalizer -nun-ke in post-nominalization negation RfCs, which renders them subject to distinctively ‘outer/metalinguistic (as opposed to inner)’ reading (‘Isn’t it/shouldn’t it be the case that …?’ [Ladd 1981, Reese 2007]),^[10] often being constitutive of a ‘mock-action’ (e.g., mock-tease) that can be assumed to be more likely to be resisted/disconfirmed than confirmed (Extract (17); Kim 2023a).

Other forms of negatively formatted RfCs identified in the data set involve post-nominal negation (Extract (25): ‘Isn’t it (called) the summer Bible school these days?’; n = 15/61, 25%), and the form -(u)lswu ep ‘cannot’ (tasi ssu-lswu-nun ep -nun-ke-ci-yo? “He cannot try re-writing it, can he?”; n = 2/61, 3%).

4.3 Modulation

In the data, 44 RfCs involve the use of a modulating device (n = 44/200, 22%), as shown in Table 2.

While adverbials are resources primarily used to modulate action, interrogative SESs indexing ‘uncertainty’, such as -nka, -na, or -lkka (‘dubitative’ (‘I wonder’)) (Yoon 2010), constitute a set of members through which action modulation is accomplished. They render the RfC they mark a form of ‘self-directed’ musing, geared to mitigating the questioner’s claim of epistemic rights. Extract (7) shows a case where the dubitative interrogative SES -na is used in a post-verbal negation RfC as a modulating device:

Extract (7):NC Talk 3 39:03, No. 74

With the use of the dubitative -na, Byen’s RfC is formulated as a self-addressed inquiry. It is implemented in the form of ‘musing aloud’ (‘I wonder …?’), enacting a context where the recipients are positioned as ‘overhearers’ (Goffman 1979). Note that this practice is reciprocated by the recipients. In line 4, Yun produces a na-marked utterance of his own in overlap, and in line 5, Ari’s confirmation is formulated with the SES -kwuna, rendering her uptake a form of self-addressed inference-making (“(I’ve just inferred that) it is indeed so.”).

Note that, in Extract (7), the SES -na, as a modulating device, is embedded in an RfC formatted with post-verbal negation (-ci anh). As Table 2 shows, there are 15 (out of a total of 21) post-verbal negation RfCs that are used with some form of modulating device (71%). This points to the assertive character of post-verbal negation RfCs (Kim 2023a), designedly constituted and oriented to as the ‘mitigatable’, i.e., an object to be mitigated ipso facto (Section 4.2).

Other modulating devices include modal expressions, such as -keyss (‘suppositive/affect attribution’; Suh and Kim 1991), and the SES -ney (‘factual realization/noticing’; Lee 1993, Kim 2004), which often co-occur (Suh and Kim 1991). Extract (8) shows a case where -keyss and -ney are used together (in Inho’s RfC in line 1):^[11]

Extract (8): SB Post-Service Gathering 47:01, No. 43

As modulating resources, the use of the modal marker -keyss and the suffix -ney index Inho’s orientation toward mitigating his epistemic claim; his RfC is shown to be grounded in his supposition (with the modal marker -keyss), and the confirmable it proposes is formulated as something he has ‘noticed’ (with the suffix -ney; also see Extracts (19) and (20)).

Inho’s RfC becomes a trouble source turn by Yun’s repair initiation in line 2 (yey? ‘Pardon?’), which prompts Inho to repeat his RfC in a slightly modified form in line 3 (“I probably won’t be seen.”). Inho’s repair turn in line 3 is again met with Yun’s repair initiation in line 4, mwe-yo. (‘What.’), which, produced with falling intonation, indexes that the trouble has resulted not from a problem of hearing or understanding, but from that of referent identification (Schegloff 1997).

Note that the suffix -ney (‘factual realization/noticing’), initially used to format Inho’s RfC in line 1, is not recycled in his repair turn in lines 3 and 5, where -ney is replaced by -e (‘Informal ending’) and -eyo (informal politeness marker), respectively. This points to the ‘positionally sensitive’ character of the suffix -ney as a modulating device sequentially embedded in the ‘initially responsive’ position (Schegloff 1987, Kim 2001).^[12]

4.4 Inference marking

Inference marking of various types was used in 24% of the RfCs examined (n = 48/200). Among these are 20 cases where two or more inference markers are used in a single RfC (n = 20/48, 43%). Table 3 shows the distribution of inference markers, served by a wide range of grammatical forms.

Note that the list contains the nominalization marker -nun-ke, composed of the attributive/adnominalizer -nun and the defective/general noun -ke (‘thing’), which is agglutinated to the verb stem, with the meaning, “Is it the case/Should it be the case that […]?”. This form, which frequently co-occurs with other inference markers, is used in the context where the speaker inferentially invokes an aspect of shared understanding on the basis of which the noted event/item is to be evaluated (Kim 2023a). Extract (9) shows a case where -nun-ke is used with two other inference markers: the change-of-state token/realization marker ah and the inference-marking discourse connective kulem ‘then’ ( ah kulem hayngsi chi-si-n-ke ^[13] -eyyo? “Oh, then, is it that (your farther) took (and passed) the public officer qualification examination?”):

Extract (9): SB2 15:20, No. 10

In the preceding context, Sehi said that her father is a public official, and in line 1, Orin, with his RfC, is asking if he passed the (highly competitive) public officer qualification examination. This is something that Koreans may normatively infer since high-level public officials in Korea are often those who have passed this examination. With his RfC, Orin exhibits normative orientation toward constituting the target referent (the recipient’s father) as someone whose career is praiseworthy (i.e., as someone who passed the highly difficult public officer qualification test). This is demonstrably oriented to by the recipient, Sehi, whose markedly self-deprecating disconfirmation works to counter Orin’s move to constitute her father as a praiseworthy stance object (“No, Not in the world hhuhh.hh”). In this process, the nominalizer nun-ke plays a significant role as an inference marker, rendering the confirmable anchored in normative or otherwise interpersonally shared reasoning (Kim 2023b; also see Extract (15)).^[14]

Other inference markers identified in the data involve the SES -takwu (‘quotative’; Extract (10)), -kwuna (‘inferred discovery’; Lee 1993, Kim 2004, Kim and Suh 2021), the clausal connective -nuntey (‘circumstantial’), used to formulate an inference about the prior speaker’s ‘my-side-telling’ (Pomerantz 1980, Park 1999, Kim 2018 [Extract (16)]), and the clausal connective -llakwu (‘intention’), used as a resource for inferentially attributing a particular intention to the recipient (e.g., ‘Why? Because you want to have fun?’; Levinson 2013).

4.5 Connectives

In the data examined, 36% of the RfCs (n = 72/200) were used with some form of connective, with many of which also serving as inference markers, as shown in Table 4.

The most frequently used type of connectives used with RfCs are ones that contain the indexical component ku(le), e.g., kulem (‘then’) in Extract (9) and kulenikka (‘so’) in Extract (20). They constitute more than a half of the connective identified in the data (n = 41/72, 57%), followed by the particles ah/eh functioning as a change-of-state/realization token (‘Oh I see’ [n = 13/72, 18%]), clausal connectives (n = 7/72, 10%), response cries (n = 6/72, 8%), and other discourse markers (n = 5/72, 7%).

Extract (10) shows an example where connectivity to the immediately preceding context is marked by the discourse marker ani (‘no/well/wait’; Kim 2015), which prefaces Yuli’s RfC (lines 8–9). In response, Sohi produces the interjection ney (‘yes’) at line 10, a response token indexing politeness (Section 5.2).

Extract (10): SB2 00:20, No. 9

There are seven instances in the data where two connectives are used in an RfC. They involve 6 out of 11 kule-form connectives, which are prefaced by ah/eh or a response cry. An example is in Extract (9), introduced in Section 4.4, where we find the change-of-state token ah is used with the inference connective kulem (“Oh, then, is it that (your father) took the civil service examination?”). Another case involves the clausal connective -ese (‘cause’) prefaced by the change-of-state token ah, which is found in Extract (4), introduced in Section 4.1 (“Oh, (he decided to drop the class) because he was not good enough?”).

4.6 Tags

In Korean, question tags typically take the form of the tag-type clause kuleci or one of its contracted forms, kuchi/kuci (‘It is (like) that?’), which involve the ‘committal’ SES -ci. Their use is only sporadically observed as a feature of RfCs. In the data, there are only five instances of RfCs with a question tag (n = 5/200, 2.5%).

Extract (11) is a case in point. In the preceding context, Hijun claimed that he does not use honorific expressions when referring to American professors. His friends also pointed out that Hijun similarly did not use honorific expressions when talking about Korean professors with whom he is on close terms. In lines 1–2, Hijun produces a kuci-marked RfC addressed to his wife, one of the co-participants. With his wife, he appears to be making a backdown in collusion, as he adds a new observation that serves to revise or granularize his earlier claim (“But when we talk about Korean professors – when we talk about Korean professors, even when we do not know them, we don’t talk like that ( = We do use honorific expressions), right?”). While there is no hearable uptake from his wife, Kyeng comes forward with a response in lines 5–6, which constitutes partial (dis)confirmation (Section 5.1):

Extract (11): Lunch Discussion NC 54:00, No. 104

Extract (11) illustrates the ‘remedial’ function often associated with the use of the tag kuci, which is devoted to pursuing recipiency while managing the speaker’s face and affective stance.

In this article, a large number of questions marked with the tag kuci are not treated as RfCs and are thus excluded from the target data, because their use does not make confirmation/disconfirmation a relevant next action. As suggested in Extract (11), tags are specialized for ‘affect-displaying’ functions (Hepburn and Potter 2010), incrementally produced add-ons geared to pursuing recipiency (Ford et al. 2002). They emerge in the context of managing both the questioner’s and the recipient’s face in regard to the (often face-threatening or face-implicative) action of the host TCU, sometimes as an RfC, but more often as a form of question that presupposes the recipient’s confirmation (Kim 2023c).

The small number of tags would also be attributed to the availability of ‘pseudo-tags’, which function like tags. Pseudo-tags, -ci and -cianha, are constructed with the ‘committal’ SES -ci, which is agglutinated to the verb stem, rather than ‘tagged’ onto the host TCU. While pseudo-tags are not included in the category of ‘tag’ in this article (due to their status as an element agglutinated to the verb stem), they comprise about one-third of the RfCs with no tag (n = 66/195, 34%). Note that even two of the five RfCs used with a tag involve the use of the pseudo-tag -ci or -cianha. Excluding these, we have only three RfCs used with a question tag alone (1.5%). The distribution of RfCs with ‘tags’ is shown in Table 5.

Extract (12) shows an RfC formatted with the pseudo-tag -ci (also see Extracts (1), (14), (21), (22), and (23)). In this conversation, Kyeng is talking about a Thanksgiving Day sale at the university story, where she bought clothes for a discount price. In the immediately preceding context, Hani asked Kyeng whether books are on sale too. In line 1, Yun, Kyung’s husband responds with ci-marked RfC, formatted with pre-verbal negation, with which he prompts Kyung to confirm that books are not on sale:^[15]

Extract (12): NC Talk: TA Meeting 30:37, No. 151

Extract (13) is case where the RfC is formatted with -cianha, a form of pseudo-tag grammaticalized from -ci, conflated with the negation anh (Kawanish and Sohn 1993). In the preceding context, Yun said that he was unable to have a renowned professor in Yun’s own field serve on a committee because she is at another university. In line 1, Hijun responds with an RfC marked with the pseudo-tag -cianha, affording Yun with an option that the latter could have taken but did not (‘You can ask her to serve as an outside committee member, right?’). Yun responds with partial confirmation with an account in lines 2–3:

Extract (13): Lunch Discussion NC 42:58, No. 98

Hijun’s RfC has a strong import of offering a solution to the recipient’s problem, in such a way that the latter is assertively prodded to appreciate its relevance and respond accountably.^[16] Compared with the pseudo-tag -ci, which is geared to soliciting straightforward confirmation of facts, often ‘in collusion’ (as in Extracts (1) and (12)), the pseudo-tag -cianha indexes a more assertive stance of the questioner who orients to raising the recipient’s commitment to the confirmable at hand from a ‘momentary’ lapse (Kim 2022).

The preceding observations suggest that, if we use a radically functional approach, we may treat the pseudo-tags -ci and -cianha as ‘tags’, though they are SESs agglutinated to the verb stem. If RfCs formatted with pseudo-tags are included in the category of ‘tags’, together with ‘genuine tags’ (i.e., kuci tagged on the TCU), the frequency rate would increase to 36% (n = 71/200), with 68 pseudo-tags (34%) plus three tags whose host TCU does not involve a pseudo-tag (2%), as shown in Table 6.

This issue merits a separate analysis of types of ‘tags’ in terms of their formal and functional features, from a cross-linguistic perspective (Enfield et al. 2012).

4.7 Prosodic design

In the majority of the RfCs in the data, the confirmable was produced with rising intonation at the ending (n = 146/200, 73%), which suggests that rising intonation constitutes a crucial prosodic feature of RfCs in Korean. Table 7 shows the final intonation of the confirmable, with and without a tag.

Note that all five instances of tags were produced with rising intonation. They are all integrated into the preceding TCU prosodically, with no separate contour vis-à-vis the confirmable (Extract (11)). A subtle prosodic change, however, was observed in those cases such that the TCU-final verbal predicate is produced with continuing intonation at the ending, which flows into the first syllable of the tag (ku ‘that’), followed by the second syllable of the tag (-ci ‘committal’), and then produced in rising intonation. While the number of tokens is small (n = 5/200), this observation suggests that a ‘level-rise’ prosodic contour, where there is no prosodic ‘break’ between the host TCU and the tag, may be one of the features associated with the Korean RfCs formatted with a tag.

In the next section, we turn to resources used to build responses to RfCs.

5.1 Responsive actions

Overall, 61% of the responses to RfCs are confirmation (n = 116/190), 21% are disconfirmation (n = 39/190), and 18% are neither confirmation nor disconfirmation (n = 35/190). In the data, there are ten instances of RfCs to which no verbal response was produced.^[17] Extracts (14) and (15) show examples of confirmation and disconfirmation, illustrating typical response formats involving response tokens ney (‘yes’ (polite)) and ani-yo (‘no’ (polite)), respectively.

5.2 Response tokens

About half of the RfCs (n = 87/190, 47%) identified in the data were answered with a response token. Response tokens typically include ‘yes’ (yey/ney (polite) and ung/um/eh (plain)) and ‘no’ (ani (plain), ani-yo (informal polite), and ani-pnita (formal polite)). As Table 8 shows, I also categorized kulechi (‘certainly it is’) and kulem (‘of course it is’) as response tokens, which contain the indexical form kule (‘like that’), a highly productive indexical form used for constructing a wide range of discourse markers/connectives (Park 2008).

Among the 87 response tokens, 69 (79%) were used with confirmation (all ‘yes’-type or kule-form response tokens) and 18 (21%) with disconfirmation (all ‘no’-type response tokens involving the form ani). Examples of the response tokens ney (‘yes’) and ani-yo (‘no’) are provided in Extracts (14) and (15), respectively, which were introduced in Section 5.1.

Response tokens are used in 59% of confirmations (n = 69/116)^[19] and 44% of disconfirmation (n = 17/39). There is one response token used in the ‘neither’-type response (n = 1/35, 2.9%), where the negative response token ani (‘no’) prefaces an evasive answer organized as an extended account.

Interjections like um/ung/yey (‘yes’), as response tokens, are often embedded in a ‘subordinate-action’ context (Enfield et al. 2019), e.g., a parenthetically engendered sequence. In Extract (1), which is re-introduced below, Kyeng’s um (‘yes’) emerges as a response to Yun’s RfC that initiates a parenthetical sequence (Mazeland 2007, Schegloff 2007), after which the talk is allowed to move quickly back to the main narrative (i.e., Byen’s continued telling):

Extract (1): NC Talk 3 05:91, No. 57

Extract (18) illustrates a case where eh, another informal form of affirmative response token, is used:

Extract (18): Park, S.-H. Data #4 18:54, No. 194

In the data examined, there are three instances of eh produced as a response token, as in Extract (18), and they are all used in a non-minimal response, i.e., followed by further response (Section 5.5). The association of eh with non-minimal response suggests that, compared with other informal response tokens like ung/um (‘yes’), which are frequently constitutive of a minimal response, the status of eh as a response token is not as fully established.

In this respect, it should be noted that the response token eh is often formally indistinguishable from the acknowledgement/change-of-state token eh, which, just like response cries, serves as a preface framing a response. For instance, consider Extract (19), where Inho’s turn-initial eh is not being produced as a response token (comparable to um/ung (‘yes’)), but as a particle/change-of-state token (‘oh’) framing the upcoming confirmation:

Extract (19): SB Post-Service Gathering 12:43, No. 38

Inho’s prefatory use of eh marks his orientational shift, projecting a strong confirmation, as suggested by the exaggerated nodding gesture that accompanies his response, which is also marked with the ‘committal’ particle -ci. It manifests features commonly associated with the English change-of-state token oh, used in a response to a question (Heritage 1998). This conveys the sense that Inho is treating Yun’s RfC as ‘inapposite’, or at least superfluous, in that it is inquiring about something that can be readily presupposed; i.e., if Inho was a Catholic, he can be presumed to have a baptismal name.

There are also, in the data, five instances of kule-marked response tokens. Compared with ‘unmarked’ interjection-type response tokens such as ung/um/eh/yey/ney (‘yes’), which are ‘designed not to introduce any pragmatic turbulence’ as unmarked way to confirm (Enfield et al. 2019, 291), kule-marked forms constitute ‘pragmatically marked’ members of response tokens, indexing the respondent’s more assertive confirmatory stance. Consider Extract (20), where kuleh-ci is used as a response token, marked with the politeness particle -yo:

Extract (20): SB Post-Service Gathering 05:30, No. 32

At line 1, Chen makes an RfC marked by the SES -ney, serving as a modulating resource that frames the confirmable as something that has just been ‘noticed’, accompanied by the inference marker kulenikka (‘so/then’ [“(I’ve just noticed that) you’re the only son, then.”]) (Sections 4.3 and 4.4). At line 2, Min responds with kuleh-ci-yo (‘Right ( = Certainly it is).’), a form indexing his epistemic authority, which draws upon the fact that the confirmable (i.e., whether Min is the only son) is well within his own epistemic domain.

In this article, ‘short’ kule-forms such as kulehci/kuchi/kuci(-yo) (‘right/it is like that’) or kulem (‘of course it is’) are included in the category of ‘response tokens’ (cf. Betz 2015), since they manifest features of a ‘pre-fabricated’ token, as clearly demonstrated by the way kuci is used as a tag.^[20] ‘Long’ kule-forms, e.g., those that contain modal markers like -keyss or -(u)lke(s) (e.g., kule-keyss-ci-yo ‘I believe so’ or kule-lke-eyyo ‘I guess so’), are excluded, on grounds that the modally modulated process of epistemic mitigation renders their compositional structure too heavy/complex to be treated as a ‘token/interjection’.

5.3 Clusters of response tokens

Out of 87 responses to RfCs that involve a response token, 12 cases (14%) involve clusters of response tokens (Table 8). Clustered response tokens collected for this study are all delivered in one intonational contour, produced in a cluster of two, e.g., yey yey, as shown in Extract (21):

Extract (21): NC Talk 3 04:51, No. 56

Kyeng’s response tokens at line 28, along with Yun’s repeat of the place term Byen used in her RfC (‘Hope Ranch’) in line 29, serves as a ‘go-ahead’ signal with which she lets Byen produce a subsequent telling about Hope Ranch (not shown in the data).

Also consider Extract (22), where we find that two clusters of yey yey (‘yes yes’) are produced, preceded by a change-of-state/realization marker (ah ‘oh’):

Extract (22): SB Post-Service Gathering 53:05, No. 48

In response to Jihi’s allusively formulated RfC, marked with the pseudo-tag -ci (line 5), Hun initially responds by initiating repair at line 7 (“You mean an international call service plan?”). In response to Jihi’s repair turn at lines 8–9 where she flatly disconfirms Hun’s candidate understanding^[21] and provides a more specific ‘clue’ (motheyl ‘motel’; cf. Kim 1999), Hun produces a change-of-state token (ah) followed by a cluster of two response tokens (yey yey), acknowledging the success of Jihi’s repair (line 11).^[22] The cluster of response tokens is then repeated after an aborted utterance (‘( ) that-’), serving as a go-ahead signal that enables the questioner to move on to her main telling; in the subsequent talk (not shown in the data), Jihi makes a request, asking Hun if he would be able to find a motel for a newly arriving visiting professor to stay temporarily before moving to his apartment.

These observations suggest that response tokens produced in cluster index the respondent’s orientation to boosting the sense of ‘go ahead’ through doubling the confirmatory force of the response token, reciprocating the questioner’s interpersonal commitment indexed by the pseudo-tag -ci (‘committal’) in the RfC (line 26 in Extract (21) and line 5 in Extract (22)). Through repeating a response token, the speaker acknowledges the epistemic supremacy attributed to him/her (Levinson 2013), while reciprocally mitigating his/her own epistemic independence. Note, in this respect, that, in Extract (21), Kyeng, as a resident of Santa Barbara, finds herself in a position where she can claim epistemic independence as she responds to Byen’s RfC which targets an area near Santa Barbara.^[23] Likewise, in Extract (22), Hun is a local who is known to be knowledgeable about accommodation in the area.

5.4 Position of response tokens within the responsive turn

In the data analyzed, the first response token was predominantly produced turn-initially (n = 83/87, 95%), with only a small number of response tokens being situated in the final position (n = 4/87, 5%). These few instances include cases where the recipient claims epistemic authority regarding the confirmable posed by the RfC, as exemplified in Extract (23). In this conversation, Byen and Ari, who are teaching Korean professionally, are sharing problems associated with teaching Korean in American universities. In response to Byen’s RfC marked with the pseudo-tag -ci, Ari produces a confirmation in line 6, with the response tokens, produced in the cluster of two, being positioned turn-finally:

Extract (23): NC Talk 3 03:26, No. 54

Ari’s response is produced as a transformative answer (Stivers and Hayashi 2010), in which Byen’s first assessment made through her RfC, pyello ep ‘there’s few’, is downgraded to manh-ci anh ‘there’re not many’. With the response tokens being relegated to turn-final position, Ari’s reformulated assessment is produced in the form of ‘non-type-conforming’ response (Raymond 2003), exhibiting her claim of epistemic independence, and conveying the sense that she is resisting the terms on which Byen’s question is based.

The use of a cluster of response tokens here seems to embody Ari’s orientation to mitigating her claim of epistemic independence, i.e., by reducing the sense of resistance by doubling the confirmatory force of the affirmative response token. With the cluster of response tokens being positioned turn-finally, she exhibits a delicately nuanced stance post hoc in resisting the terms of Byen’s ci-marked RfC, with which she pursues Ari’s straightforward confirmation. This can be partially attributed to the fact that, while Ari is a more experienced teacher of Korean than Byen and thus can make a stronger claim of epistemic rights, they, as co-participants at a conference, and who have just met for the first time, are not on close terms with each other (cf. Raymond and Heritage 2006).

5.5 Minimal and non-minimal responses

Out of 190 cases where an RfC received a verbal response, 45 were responded to minimally (i.e., with response tokens only (24%)), and 42 non-minimally (i.e., with response tokens followed by a further response (22%)). In the other 103 cases (54%; ‘other’ category), the responsive turn did not involve a response token.

RfCs engendering a minimal response have the character of a ‘casual fact-checking’ question, as can be observed in Extracts (1), (3), (4), (10), (20), (21), and (22). Sequences they organize often manifest features of a pre-sequence (Extracts (10), (21), and (22)) or a parenthetical sequence (Extract (1)).

Examples of a non-minimal response are found in Extracts (5), (9), (14), (15), (18), and (23), where the response token is followed (or prefaced) by an utterance further elaborative of the responsive action.

Examples of the ‘other (than minimal or non-minimal)’ responses, where the response to an RfC does not involve a response token, are illustrated by Extracts (2), (6), (7), (8), (11), (12), (13), (16), (17), (19), (24), and (25).

Concerning their relationship with (dis)confirmation, a large majority of minimal responses are confirmation (n = 41/46, 89%), with only five instances being disconfirmation (11%). The majority of non-minimal responses were also confirmation (n = 30/43, 70%), with 12 non-minimal responses being disconfirmation (28%), and only one ‘neither (confirmation nor disconfirmation) type’ (2%). As for the ‘other (than minimal or non-minimal response)’ type responses, 45 out of 101 instances are confirmation (44%), 22 disconfirmation (22%), and 34 ‘neither (confirmation nor disconfirmation)’-type responses (34%). These distributional tendencies suggest that, compared with confirmation, which predominantly takes the form of a minimal response, disconfirmation is more likely to take the form of a non-minimal response, and that the ‘neither (confirmation nor disconfirmation)’-type responses are more likely to take the form of ‘other (than minimal or non-minimal response)’.

5.6 Full repeats

Out of 190 instances of verbal response, there are 58 repetitional responses, where an element of the RfC is repeated in the response turn (n = 58/190, 31%). Among these are 11 cases of full repeats (n = 11/190, 6%), with 47 cases being ‘partial repeat’ (n = 47/190, 25%).^[24] An example of a full repeat response is shown in Extract (24), which is produced in the form of ‘confirming allusion’ (Schegloff 1996):

Extract (24): NC Talk: Breakfast 20:55, No. 135

In line 23, Hani makes an RfC formatted with pre-verbal negation (an), which is built to Kyeng’s preceding turn through the turn-final clausal connective -kwu (‘and’) (Kim 2015a) (‘And coffee is not included?’). In line 25, Kyeng responds with a full repeat in confirmation. Here, Hani’s RfC pre-verbal negation RfC is oriented to as a ‘formulation’ (Heritage and Watson 1979) of an aspect of Kyeng’s prior talk to be confirmed, in terms of both its content and allusive/inexplicit conveyance (i.e., listing of food items contained in the breakfast combo plate where ‘coffee’ is missing; Kim 2021, 2023a).

Extract (25) is another case in which a confirmatory response takes the form of a full repeat. In the preceding context, Inho used the term swuyanghoy (‘retreat’) in describing the church summer camp activities he enjoyed in his childhood (not shown in the data). In line 1, Chen makes an RfC where he proposes the alternative term yelum sengkyeng hakkyo (‘the summer Bible school’[“Isn’t it (called) the summer Bible school these days?”]) as one that is in current use. Inho responds with confirmation, repeating the term used by Chen:

Extract (25): SB Post-Service Gathering 08:24, No. 36

Inho’s repetitional response in line 2, with the addition of the ‘committal’ SES -ci and a reformulated adverbial expression (yocum ‘these days’ → icey-nun ‘now’), strongly embodies his epistemic claim of ‘ownership’ over the confirmable posted by Chen’s RfC (Raymond and Heritage 2006; also see Extract (19), a case of partial repeat where the verb stem iss (‘exist’), used in Yun’s RfC is repeated in Inho’s response, marked with the ‘committal’ SES -ci). It conveys the sense that Inho is resisting and ‘pushing back against’ Chen’s implicit claim of epistemic rights, embodied in his RfC (Enfield et al. 2019, 286).

Repetitional responses tend to be used with response tokens less frequently than in the case of non-repetitional responses; response tokens are used with 22% of repetitional responses (n = 13/58) and 56% of non-repetitional responses (n = 74/132, 56%; Note 19). In relation to polarity, responses to negatively formatted RfCs are more likely to take the form of a repetitional answer (n = 23/58, 40%) than those to positively formatted RfCs (n = 35/132, 27%). In terms of types of responsive action, repetitional responses were produced slightly more frequently in disconfirmation (n = 14/39, 36%) than in confirmation (n = 34/116, 29%) or in neither-type response (n = 10/35, 29%).