Inequality in Action: Granting Service Requests in a Highly Resource-Constrained Context by Tam, Whitehead and Raymond

Inequality in Action: Granting Service Requests in a Highly Resource-Constrained Context

Catherine L. Tam, Kevin A. Whitehead, and Geoffrey Raymond

Published in: H. Z. Waring & N. Tadic (Eds.), (2024). Critical Conversation Analysis: Inequality and Injustice in Talk-in-Interaction (pp. 91–114). Bristol, U.K.: Multilingual Matters.

INTRODUCTION

Calling an emergency service line implies a need for an emergency service (Drew & Walker, 2010; Heritage & Clayman, 2010; Zimmerman, 1984, 1992), displays the caller's entitlement to receive the service (Curl & Drew, 2008; M. R. Whalen & Zimmerman, 1990), and positions them as a (potential) beneficiary of a service, with call-takers serving as gatekeepers to the service and thus as (potential) benefactors (Clayman & Heritage, 2014; Drew & Walker, 2010; Heritage & Clayman, 2010; Raymond & Zimmerman, 2007, 2016). Conversely, a call-taker's granting of the request for service and the caller's tacit acceptance of the granting by moving to the call's closing conveys the participants' mutual understanding that a service will be dispatched that is appropriately calibrated and timed in light of the nature and severity of the emergency at hand. In Garfinkel's (1963, 1967) terms, participants act in accordance with "the natural attitude of daily life" (also see Schütz, 1953, p. 5), displaying "trust" that the emergency service institution will function as expected as a background condition for the accomplishment of the actions of requesting and providing services (cf. Garcia & Parmer, 1999; Heritage, 1984; Watson, 2009). In this respect, a caller's willingness to end the call and await the arrival of the service reflects their trust that institutional actors will fulfill the obligations entailed by the call-taker's granting of the service. Where participants may have grounds to anticipate that the fulfillment of this "social contract" may be so substantially delayed as to call into question its status as an emergency service, they may deploy practices to prospectively manage the trouble that might otherwise ensue from such failures. These practices may be understood as being designed to shore up trust in the institution's ability to meet its obligations in the face of circumstances that may give rise to doubt or concern in this regard.

Precisely because trust is foundational to institutional realities, participants must routinely manage conduct, events, and outcomes that may undermine it. For example, callers may contend with a call-taker's skepticism regarding their claims (Garcia & Parmer, 1999) or even the project that occasioned their call (Raymond & Zimmerman, 2016). Conversely, callers doubting whether service providers will respond in a timely manner to putative emergencies may attempt to "game the system" by formulating circumstances in ways designed to secure a highest-priority dispatch (see Moskos, 2008, pp. 89–110). The import of these routine, "seen but unnoticed" (Garfinkel, 1967, p. 37) ways in which participants contend with trust that has been diminished or otherwise frayed are cast into stark relief by circumstances where communities lose trust altogether. For example, Desmond, Papachristos, and Kirk (2016) document a substantial decline in 911 calls by Milwaukee residents following a highly publicized, racially motivated beating of a Black civilian by white officers. In the aftermath of the beating and the city's failure to take action against its perpetrators, Desmond et al. (2016, p. 870) observed a "large and durable" decline in 911 calls (approximately 22,000 fewer calls in one year) from Black communities in the year following the beating, making the "cit[y] as a whole, and the Black community in particular, less safe". Relatedly, Bell (2017) uses the concept of "legal estrangement" to describe the chronic and pervasive loss of trust associated with "the intuition among many people in poor communities of color that the law operates to exclude them from society" (p. 2054).

Cases such as those documented by Desmond et al. (2016) and Bell (2017) thus show that either specific events or exclusionary practices used over long periods can be associated with communities losing trust in public institutions. Similarly to the cases described by Desmond et al. (2016) and Bell (2017), inequalities in the provision of emergency services in South Africa are directly connected to histories of racial oppression and exclusion. Specifically, the South African healthcare system consists of public and private sector service providers, with the public sector's emergency services being managed at a provincial government level. Although almost 65% of the South African population rely on public sector services, ongoing legacies of the apartheid system along with subsequent mismanagement and corruption, have contributed to public healthcare being substantially under-resourced and over-burdened relative to the private sector (see, e.g., Coovadia, Jewkes, Barron, Sanders, & McIntyre, 2009; Horwitz, 2009).

In this report we examine calls to an emergency service call center in South Africa, showing how participants' orientations to these material circumstances – and the issues of trust that may arise from the ways that emergency services are constrained by them – become evident in their conduct in the calls. Specifically, we consider some practices deployed by call-takers in the turns in which they grant service requests and examine how calls are brought to closing thereafter. Crucially, the practices we describe have not been reported as routine components of grantings and closings in previous studies of emergency service calls.

Conversation analytic research on emergency calls across a range of countries has examined how the benefactor and beneficiary positions are produced in and through the call's structure, which is built around a service request/response adjacency pair sequence (Zimmerman, 1984), along with pre- and insert expansions of this sequence yielding an overall structure consisting of five distinct phases: (i) opening/identification, (ii) request, (iii) interrogative series, (iv) response, and (v) closing (Heritage & Clayman, 2010; Kevoe-Feldman, 2019; Zimmerman, 1984, 1992). Routine grantings of service request and call closing (cf. Raymond & Zimmerman, 2016, p. 722) in previous CA literature are shown in Excerpts 1 and 2 (for further instances, see Raymond & Zimmerman, 2007, p. 37; J. Whalen, Zimmerman, & Whalen, 1988, p. 344; Zimmerman, 1984, p. 214; 1992, p. 37):

(1) [Heritage & Clayman, 2010, p. 90]

01  CT: We'll get somebody there right away.=

02   C: =o:kay thank yo[u

03  CT: [<mm bye>

(2) [Raymond & Zimmerman, 2016, p. 723]

01  CT: Okay we'll get somebody over there.

02   C: Tha::nk you.

03  CT: Mmhm b[ye.

04   C: [°Buh bye

As these cases demonstrate, the turn in which the granting of the request is produced may (or may not) include a turn-initial okay (as in Excerpt 2, line 1), which displays the call-taker's receipt and acceptance of the information the caller has given, and marks the transition from the interrogative series to the call-taker's granting of the service request (Heritage & Clayman, 2010; Raymond & Zimmerman, 2007, 2016; Zimmerman, 1984). The granting itself recurrently consists of a single component in the form of a service announcement, which in some cases includes an indication of the immediacy of the dispatch of the service (as in Excerpt 1, line 1), while in other cases leaves this unspecified (as in Excerpt 2, line 1). The caller then recurrently responds to the granting with a "service receipt" (Raymond & Zimmerman, 2016) in the form of a token such as okay (as in Excerpt 1, line 2) and, where relevant, appreciations such as thank you (as in line 2 of both excerpts). These may in turn be ratified by the call-taker's production of a receipt token such as mm (as seen in line 3 of both excerpts). Finally, the call-taker moves to bring the call to a close by producing a terminal particle such as bye (as in line 3 of both excerpts), to which the caller may respond with a reciprocal terminal particle before the call is terminated.

Our review of the literature revealed only one case, reported by Drew and Walker (2010, p. 109), and shown in Excerpt 3, in which the granting turn included additional components beyond those shown in Excerpts 1 and 2. As Drew and Walker (2010, pp. 108–109) note, a key feature of this case is the question of whether the caller's complaint (a nosebleed resulting from being punched "by a taxi driver") is serious enough to warrant provision of the requested service, although the call-taker does eventually grant the request and dispatch the police to the incident.

(3) [Drew & Walker, 2010, p. 109]

01  CT: =Alr:ght we'll get police back down

02       to you as soon as we can sir.=Okay?=

03   C: =Okay.

04  CT: Bye bye.

05  CT: ((makes a loud snoring sound))

In addition to the turn-initial Alr:ght and the service announcement (we'll get the police back down to you), the granting in this case includes two additional components not present in the routine cases shown in Excerpts 1–2. The first of these is an indication of contingency (as soon as we can) that implicates a possible delay, and the second is a tag question (okay?) that invites the caller's acceptance of this possibility. While Drew and Walker's (2010, pp. 107–110) analysis of this case does not focus on these details of the call-taker's granting and the caller's response, our analysis is nevertheless consistent with their observations about the high-entitlement form used in the request ("I need the police, right now"). Specifically, the two components used to expand the call taker's service announcement apparently address, and seek the caller's acceptance of, a possible discrepancy between what the caller has sought and what the call-taker can or will provide – that is, services will be dispatched, but as a matter of lower priority than the caller evidently expects. In this way, the expansion of the service announcement seeks to preemptively manage the possible mistrust that might otherwise emerge by reference to the delay the call-taker is evidently anticipating. Note then, while the caller produces the service receipt Okay (line 3), he does not produce an appreciation of the granting of the sort seen in Excerpts 1–2.

These observations are further elucidated by Raymond and Zimmerman's (2016) more recent analysis of two types of trouble that may occasion expansions of closings in emergency calls. The first of these is what Raymond and Zimmerman (2016) call "routine troubles," which involve cases in which the aligned projects of the participants (seeking and providing help) give rise to contingencies related to achieving their possible completion, such as call taker verifying information provided by the caller and caller providing additional information relevant for the dispatch of help. The expansions occasioned by such cases are thus designed to bring these articulated projects to completion and sustain the alignment of the identities of the call-taker and caller as benefactor and beneficiary respectively. In contrast, in "non-routine troubles, one or both of these features (alignment and completion) are at issue" (Raymond & Zimmerman, 2016, p. 727; emphasis in original). This includes cases where (as in Excerpt 3) callers' responses to the call-taker's indication of service provision "simply acknowledge it with 'okay' while declining to accept it as a resolution to the project they have pursued up to that point" (Raymond & Zimmerman, 2016, p. 729). Returning to Excerpt 2, we can now note that the caller's acknowledgement of the call-taker's contingent granting of the service without appreciating it treats the promise of service as departing from or falling short of what he had sought, thereby displaying that he and the call-taker have not fully accomplished alignment as help-seeker and help-provider (Raymond & Zimmerman, 2016).

A further, unequivocal, indication of the misalignment between the caller and call-taker in Excerpt 3 is then provided by the call-taker's production of a loud snoring sound after terminating the call (line 05), which, as Drew and Walker (2010, p. 109) note, displays his assessment of the call as "a waste of police time." The additional, non-routine, components of the granting in this case – as in the cases Raymond and Zimmerman (2016) examine – thus appear to be occasioned by, and reflect the call-taker's orientation to, the questionable legitimacy of the service request and therefore the non-routine nature of the call (also see Garcia & Parmer, 1999; J. Whalen et al., 1988; M. R. Whalen & Zimmerman, 1990). In contrast, in our data call-takers' routine uses of these additional components in granting the service, callers' uptake of them, and the additional complexities of subsequent moves to closing of calls together indicate that they are designed to manage issues of trust arising from participants' orientations to structural limits on the capacity of the institution to provide a timely service, rather than from call takers' assessment of the legitimacy of the caller's emergency. Our analyses of these practices thus show how participants in settings characterized by high levels of material inequality and associated resource constraints may work to manage the gap between what may be expected and what can be provided by the service institution, and thereby mitigate the potential issues of trust that may arise from this gap. We thereby demonstrate the value of a critical CA approach for investigating how social problems with origins outside of particular forms of institutional interaction can become observable in participants' orientations and conduct in the interactions.

DATA AND METHOD

We utilize a critical CA approach to examine a collection (see Schegloff, 1996) consisting of the grantings and subsequent movement to closing in 63 recorded calls to a government-operated emergency medical call center in the Western Cape province of South Africa. While the call center from which our data are drawn was reportedly relatively high-performing compared to call centers operated by other provincial governments at the time of recording, it nonetheless operated under highly resource-constrained conditions; receiving an average of approximately 1800 calls every 24 hours, while having only 65 ambulances available to dispatch for emergency responses. Consequently, wait times for arrival of an ambulance can vary widely, from as short as a few minutes to as long as several hours. Thus, while it is a routine feature of emergency dispatch worldwide that call-takers cannot reliably estimate or inform callers how long the wait for an ambulance may be, the scope of this uncertainty is of a different order of magnitude for call-takers – and, by extension, callers – in our data than it is for emergency services in better-resourced contexts.

Although South Africa has 11 official languages, the majority of residents of the region served by the call center speak Afrikaans or English as a first language, with a sizeable minority of isiXhosa speakers and small minorities of speakers of other languages also resident in the region. As is the case for other government institutions in South Africa, the official language of the call center was English, but call-takers were typically bilingual in English and Afrikaans. As fluent speakers of both English and Afrikaans, we were able to include calls conducted in both of these languages in our analysis, but we excluded a small number of calls conducted in isiXhosa due to our lack of fluency in this language.

The data extracts included in the analysis that follows were selected so as to illustrate the range of components present in grantings across the data set, and the range of variations in their production by call-takers and uptake by callers. In cases where calls were partly or entirely conducted in Afrikaans, English translations are provided in italics on the lines below the corresponding Afrikaans talk. Participants' names have been replaced with pseudonyms, and all other identifying information revealed in the calls has been altered in the transcripts.

As is unavoidably the case in undertaking any sociological analysis, we necessarily used our own and others' members' knowledge in order to recognize the members' knowledge being used by the participants in our data (cf. Garfinkel 1967). In addition to providing the descriptions of the setting and data, we have worked throughout our analysis, on a case-by-case basis, to render our recognition and use of such members' knowledge as explicitly as possible, especially when analyzing details that may be opaque for readers who are unfamiliar with features of the setting and other relevant contextual matters to which participants appeared to be oriented.

ANALYSIS

We begin our analysis by examining the practices participants routinely use to manage potential service delays – and the associated issues of trust in the institution described above – in cases where the call-taker grants the caller's request prospectively, i.e., by indicating that an ambulance will be dispatched at some point in the future. We then consider how participants may manage these issues of trust even in cases where the call-taker indicates that an ambulance has already been dispatched. Finally, we examine a deviant case in which trouble arises in relation to the call-taker's use of these practices for managing trust, with the caller treating them as implicating the call-taker's assessment of the seriousness of the emergency, as opposed to the capacity of the institution to provide services in a timely manner.

Managing Trust in Cases of Pending Dispatch

In addition to the service announcement component systematically produced by call-takers in the calls reported in the literature (as shown in Excerpts 1–2 above), call-takers in both English and Afrikaans language calls in our data recurrently and routinely produced contingency and tag question components similar to those produced by the call-taker in the deviant case shown in Excerpt 3. This is illustrated by Excerpt 4, in which the caller is a 29-week pregnant woman who has reported labor pains and vaginal bleeding.

(4) Routine granting [22510543]

01  CT: Mevrou, ons >stuur die< ambulaans

        Madam we'll send the ambulance

02       >uit daarso< so gou as moontlik, ↑hoor?

        out there as soon as possible, hear?

03       (0.6)

04   C: Baie dank↑ie.

        Thank you very much.

05  CT: Plesi::er:.

        Pleasure.

06       (.)

07  CT: Buh=by:e:

08   C: By:e.

The granting turn begins with a service announcement (madam we'll send the ambulance out there; lines 01–02), followed by an indication of contingency (as soon as possible; line 02), and a tag question (hear?; line 02). Crucially, in contrast to the case shown in Excerpt 3, there is no evidence in Excerpt 4 (nor in other cases throughout our data in which these additional components are produced) of their production as conveying skepticism on the part of the call-taker with respect to the legitimacy of the request for service, and thus as reflecting a "non-routine trouble" of the type described by Raymond and Zimmerman (2016) in the call. Instead, in our data these components appear to display call-takers' orientation to, and serve as practices for managing, "routine troubles" of a different nature from those identified by Raymond and Zimmerman (2016) – namely, troubles arising routinely from the resource-related contingencies that characterize the context in which the call center operates, and the potential consequences of these contingencies for participants' trust in the institution's capacity to provide immediate service.

Specifically, the second component (the indication of contingency) is oriented to what callers in this context are likely to know, or perhaps not know, about the potential wait times for the service they have requested. That is, it effectively acknowledges that a service-seeker may experience what they could deem to be a significant wait time, while simultaneously being designed to reassure them that the wait time will be as short as possible. As such, while the contingency component in Excerpt 3 evidently arises from the individual agency or judgment of the call-taker in relation to a particular, potentially questionable, request for service, call-takers' routine production of this component in our data seems to be a practice for pre-emptively managing systematic uncertainties implicating (mis)trust in the institution's capacity to provide timely services even in cases of entirely legitimate emergencies.

This routine uncertainty, in turn, gives rise to a systematic uncertainty as to whether the (contingent) granting of the service request adequately meets the caller's service provision expectations, and thus whether (in Raymond and Zimmerman's [2016] terms) the alignment of the caller and call-taker as help-seeker and help-provider has been adequately accomplished and the interaction can be brought to a close. The third component, the tag question, invites the caller to acknowledge the contingent terms of the granting, and thereby displays the call-taker's orientation to establishing an intersubjective understanding and acceptance of this uncertainty (cf. Hepburn & Potter, 2010; Heritage, 2002; Sacks, Schegloff, & Jefferson, 1974).

Moreover, in contrast to the dissatisfaction displayed by the Excerpt 3 caller in response to the contingent granting of the service request, the caller in Excerpt 4 (and recurrently in other calls in our data, as in Excerpts 1–2) aligns with the granting: Following a brief silence (line 03), the caller displays appreciation. In this way, the caller aligns with the call-taker's treatment of the contingent service granting as arising from potential mistrust in the institution's capacity to provide a timely service, rather than (as in Excerpt 3) as casting doubt on the legitimacy of the emergency. As such, the caller and call-taker accomplish alignment as help-seeker and help-provider, and (similarly to the routine cases shown in Excerpts 1–2), the call moves thereafter to closing within a few turns (lines 05 to 08).

Further evidence for the systematic use of these practices can be seen in cases, such as Excerpt 5 below, in which the granting is produced twice as a result of a re-opening of the interrogative phase by the call-taker, with the second iteration of the granting including the same components as the first. In this case, the patient is an insulin-dependent diabetic who is reportedly confused and intermittently losing consciousness.

(5) Repeated granting [10224]

01  CT: O:kay. .h U::m: we're gonna ↑send out an

02       ambulance <as  ̄soon as possible,>  ̄hey,

03       <and he did take his medication,

04       <i[s he on ↑insulin?

05   C: [°Yes.°

06       (1.5)

07   C: Yes, he's on insulin.

08  CT: Is Thomas on insulin?

09       (0.8)

10  CT: In:su[lin: depen:dent. ((typing noises))

11   C: [Yes.

12  CT: An' he did eat, you said ↑ey?

13       (1.5)

14   C: >°Het hy geëet al?°< ((speaking off phone))

        Did he eat?

15       (2.0) ((inaudible talk by person off phone))

16   C: Yes, he ate already.

17       (0.4)

18  CT: O:kay ma'am.

19       (.)

20  CT: We're >gonna send out the ambulance<

21       as soon as we can, ↑hey?

22       (1.3)

23   C: Thank you ma'am.

24  CT: ↑Alri:g[ht the:n, ba ↑by:e:?

25   C: [(°Okay, thank you.°)

26   C: (°Bye°.)

27       (0.8)

28   C: (°Bye°.)

In lines 01–02, the call-taker grants the request for service, with the granting including similar components to those seen in Excerpt 4 – a service announcement, an indication of contingency, and a tag question. However, before the caller responds to the granting, the call-taker quickly initiates a further series of questions about the patient, which the caller answers (lines 03–16) – with this expansion of the call being an instance of what Raymond and Zimmerman (2016), as noted above, describe as a routine trouble. After registering receipt and acceptance of the caller's responses (line 18), the call-taker re-issues the granting, which again includes the service announcement (line 20) as well as the contingency component and the tag question (line 21). As Schegloff (2004) notes, repeats of prior utterances may omit elements that the speaker thereby treats as "dispensable," and the call-taker's inclusion of all of these components in the second iterations of grantings in cases such as this thereby treats them as "indispensable" elements of the action produced through this turn. As in Excerpt 4, following an initial silence (line 22), the caller displays appreciation for the granting (line 23), and the call is brought to a close shortly thereafter.

Excerpt 6 illustrates an alternative form of the contingency component that call-takers in some cases in our data – both in English and Afrikaans calls – utilized. In this case, an ambulance has been requested on behalf of an 18-year-old full-term pregnant patient who is reportedly in labor, but whose water has not yet broken.

(6) Alternative granting [IR 10198]

01  CT: [>°Okay, ons gaan° vir julle ↑eerste<

        we're going to for you first

02       beskikbaar ambulaans >uitstuur, (net)

        available ambulance send out, (just)

03       uitkyk vir die< ambulaans, ↑hoer?

        look out for the ambulance, hear?

04       (0.5)

05   C: pt=O:kay.=

06  CT: =Dankie:,  ̄by:e[:.

        Thank you, bye.

07   C: [Okay, by::e.

As in Excerpts 4 and 5, the call-taker in this case produces a service announcement followed by a contingency component, but in this case the contingency component is produced in the form the first available ambulance (lines 01–02). While more explicitly (compared to the as soon as possible form of the contingency component used in Excerpts 4 and 5) indicating to the caller that an ambulance may not be immediately available, this form of the component also more explicitly informs the caller that the patient will be "first in line" for an ambulance once one is available, and thus that the legitimacy of the emergency is not in question. The granting in this case also includes an additional instruction to the caller to just look out for the ambulance (lines 02–03), which is evidently designed to manage an additional resource-related difficulty that ambulance drivers reported recurrently encountering in locating patients in areas with inadequate street signage and/or numbering, especially when (as in this case) callers do not have a landline or mobile phone number where they can be reached in the event of such locational difficulties.

Also, as in Excerpts 4 and 5, the granting ends with a tag question, although it appears after the additional instruction rather than after the contingency component. The tag question here nonetheless appears to serve a similar function to those produced in the prior excerpts, prompting the caller to align with the instruction the call-taker has issued and, by extension, with the contingent service the call-taker has granted. In response, the caller acknowledges the call-taker's turn (line 06) but does not produce a display of appreciation of the sort produced by the callers in Excerpts 4 and 5. There is thus evidence of misalignment of the caller and call-taker as help-seeker and help provider, which the call-taker manages by thanking the caller (line 06), thereby implementing a reversal of the default alignment of caller and call-taker as beneficiary and benefactor by treating the caller's actions as benefitting the call-taker, rather than vice versa (cf. Raymond & Zimmerman, 2016; also see Clayman & Heritage, 2014). This is facilitated in part by the caller's status as someone seeking help on behalf of a patient rather than on her own behalf (see Raymond & Zimmerman, 2016), and by the call-taker having just issued an instruction that recruits the caller to assist the institution in carrying out its work.

The call-taker then immediately moves to close the call by producing the terminal particle bye (line 06), and the caller responds (slightly in overlap with the end of the call-taker's turn) by aligning with both the reversal of the benefactor-beneficiary identities with Okay and with the move to close the call with a reciprocal terminal particle (line 07). The misalignment in this case is thus swiftly resolved without further expansion of the closing phase of the call.

Managing Trust in Cases of Immediate Dispatch

The systematic issues of trust managed using the practices described in the prior section can be further appreciated in cases where the ambulance is dispatched immediately, but the service granting is nonetheless treated as a matter of uncertainty. This can be seen in cases in our data in which call-takers indicate (in the turn ordinarily occupied by the granting) that an ambulance has already been dispatched. That is, rather than merely granting the request for service, the call-taker may report that the request has already been granted and the ambulance is en route, as in Excerpt 7, which involves a full-term pregnant patient in labor.

(7) Immediate dispatch [301023]

01  CT: Okay, ons s- (.) ambulaans is op

        Okay, we s- (.) ambulance is on

02       pad, ↑hoer?

        the way, hear?

03       (0.2)

04   C: .h Okay, baie dankie meneer.=

        Okay, thank you very much sir.

05  CT: =°Danki:e, ba ↑by[e.°

        Thank you, ba bye.

06   C: [Okay, ↑by:e:.

Here, the call-taker initially appears to be headed toward a granting similar to those seen in previous excerpts, before initiating repair (Schegloff, 2013; Schegloff, Jefferson, & Sacks, 1977) to instead formulate the service as having already been dispatched: The call-taker cuts off after saying ons s- – apparently headed toward a granting of the form ons sal... (we will...) – and informs the caller that the ambulaans is op pad (ambulance is on the way – lines 01–02). While this self-initiated repair does not reveal whether the call-taker was headed toward the production of a simple granting similar to those seen in Excerpts 1–2, or a granting including the contingency component seen in Excerpts 3–6, it does display the call-taker's orientation to the immediate dispatch of an ambulance as an alternative to the original (and thereby "default") trajectory of the granting, suggesting that call-takers can adapt their granting practices to prevailing conditions on a case-by-case basis.

Also noteworthy in this case is the call-taker's production of a tag question in the granting turn, thereby prompting the caller to align with the granting of the service even in the absence of the contingency components seen in Excerpts 4–6 (also see Schegloff, 2004). This demonstrates that even when an ambulance is available immediately, call-takers may nonetheless work to pre-emptively manage potential troubles with respect to their alignment with callers as help-provider and help-seeker, thereby displaying an orientation to pervasive uncertainty about the timely delivery of the service even though it has been announced as having already been dispatched.

A particularly telling orientation of this nature on the part of a caller can be seen in Excerpt 8. Here, a non-contingent granting is followed by an expansion of the closing sequence in which the caller asks about the likely wait time for the ambulance, with the call-taker then indicating that it has already been dispatched. In this case, the patient is a 60-year-old man who is experiencing chest pains and has a history of hypertension.

(8) Non-contingent granting [10646]

01  CT: .h >Okay, ↑thank you ma'am we're

02       gonna send a <ambulance out, ↑hey:?

03       You must just [( )

04   C: [Okay, th↑a:nk you my dear.

05       <To twentix Queen Mary Street in

06       Ruy:terwa:cht?

07  CT: Yes, [we've got it [( )

08   C: [.hhh [Okay- Um:: they w-

09       they w↑on't be too long,  ̄he:y?

10  CT: No, dey on their way now.

11   C: O:kay thank >yo[u my< dear:, ↑bye[::

12  CT: [Okay. [Bye.

The granting the call-taker initially produces in this case (lines 01–02) resembles those shown in Excerpts 1–2 by lacking the contingency component seen in Excerpts 3–6, and thereby implying the immediate dispatch of the ambulance without explicitly (as in Excerpt 7) indicating that it has already been dispatched. However, it does include the tag question (hey?) similar to those observed in Excerpts 3–7, and the call-taker also begins to formulate an additional instruction component (You must just), as seen in Excerpt 6, before aborting it as the caller begins to respond in overlap (lines 03–04). While the caller initially responds by aligning with and appreciating the granting (line 04; cf. Excerpts 1, 2, 4, 5, and 7), she then initiates an expansion of the granting sequence, requesting confirmation that the call-taker has the address at which the patient is located (lines 05–06). Following the call-taker's confirmation (line 07), the caller produces an aligning Okay (line 08) before cutting off to initiate a further sequence expansion with a query regarding the potential wait time for the ambulance (lines 08–09). This query serves as evidence that in the absence of either the type of assurance provided by the contingency component (as in Excerpts 4–6), or the production of an explicit indication by the call-taker that the ambulance is en route already (as in Excerpt 7), callers may treat as inadequate the type of service announcement that, in calls such as those shown in Excerpts 1–2, is treated as indicative of immediate dispatch. That is, if a granting includes a service announcement in the prospective we're gonna... form without also including a contingency component, callers may treat the contingency component as relevantly absent from the granting, thereby treating the precise nature of the granting as uncertain.

It is noteworthy, however, that the polarity of the caller's query projects a no answer (Raymond, 2003), and thus treats the granting the call-taker has provided as most likely, albeit not unequivocally, indicating the immediate or imminent dispatch of an ambulance. The call-taker aligns with this producing (as projected by the question) a turn-initial No before reporting, similarly to the call-taker in Excerpt 7, that the ambulance team is on their way now (line 10). The caller and call-taker thus collaboratively treat the contingency-free service announcement the call-taker initially produces as effectively equivalent (or at least near-equivalent) to the "en route," form of the granting she subsequently provides. The caller then produces alignment and appreciation tokens, and the call quickly moves to closing. The expansion of the closing phase of the call in this case thus involves a non-routine trouble that apparently arises from systematic uncertainty as to whether a form of granting recurrently treated in well-resourced contexts as indicating immediate dispatch of a service can be treated as such in this resource-constrained context. This uncertainty is then resolved – and realignment of the caller and call-taker as help-seeker and help-provider is accomplished – by the call-taker's confirmation that this is indeed so.

A Deviant Case

In the cases examined above, and consistently across our data set, the practices used in the granting in addition to the service announcement are systematically treated as arising from the contextual conditions in which the institution is operating. As such, even in cases where there is evidence of misalignment of the caller and call-taker as help-seeker and help-provider, the troubles are treated as arising from the potential uncertainty of the institution's capacity to adequately fulfill service requests that have been granted, rather than from the call-taker's judgments as to the legitimacy of the medical emergency for which service has been requested. In contrast, Excerpt 9a (in which the patient is reportedly experiencing paralysis and severe pain in her legs) includes an extended expansion of the closing phase of the call following a caller's display of dissatisfaction with a granting, which she explicitly links to the contingency component of the granting. Subsequently, the caller explicitly treats the contingent granting as arising from the call-taker's judgment of the (lack of) legitimacy of the emergency, despite the call-taker (as in the cases above) giving no indication of skepticism in this regard.

(9a) Deviant case [10293]

01  CT: .hhhh >Okay ↑ma'am, I'm 'onna send an

02       ambulance out to you< as soon as

03       possible, ↑hey?

04   C: Okay, how soon is "soon as possible"?

05  CT: Um:: ↑ma'am I cannot tell you, but as soon

06       as an ambulance is available, they'll

07       definitely send one through.

08   C: Is it in an ↑hour, two ↑hours?

09       (1.3)

10  CT: I cannot tell you but we hope so ma'am:.

11       (.)

12  CT: [I'll try and do my bes:t for you.

13   C: [(What-)

14       Okay, 'cause she's l↑aying there >in a lot of

15       pain, she needs< to get to the hospital.

16  CT: No problem.

17       (0.4)

18   C: Alright, ↑thank [you.

19  CT: [Okay, >thank you so

20       mu[ch for calling.<

21       [((C hangs up))

The granting in this case is much like those in the excerpts shown above, including a service announcement, indication of contingency, and a tag question (lines 01–03). As in the previous excerpts, the caller's response begins with the service receipt Okay (line 04), but the caller initiates an expansion of the granting sequence by requesting a more granular specification of the meaning of soon in the contingency component of the granting, thereby explicitly linking her question to this component. In doing so, the caller treats the contingency component as hedging rather than as reassuring, thus treating the granting as being of questionable adequacy and displaying misalignment as a help-seeker in response to the help the call-taker has promised will be provided. Also noteworthy is the low-contingency, high-entitlement form of the caller's question (how soon is "soon as possible"?) which orients to a high degree of entitlement to the information she is requesting – compared, for example, to a form such as Is it possible to tell me how soon it will be? (see Curl & Drew, 2008; Drew & Walker, 2010). This further underscores her dissatisfaction with the contingent granting the call-taker has produced, and the attendant misalignment between herself and the call-taker.

In response to the caller's question, the call-taker states that she cannot tell the caller how soon it will be (line 05), with the emphasized word cannot possibly serving as an appeal to institutional authority or policy over which she personally has no control. She then produces a different form of the contingency component of the granting and indicates that they'll definitely send an ambulance (lines 06–07), again grounding her response in institutional constraints while displaying that the responsibility for sending the ambulance rests with an institutional they beyond her purview, and hence is a matter of resource (i.e., ambulance) availability rather than her judgment of the legitimacy of the emergency.

The caller then pursues her request in a more specific format that includes candidate estimates of wait times that the caller thereby treats as likely and/or reasonable – an ↑hour, two ↑hours? (line 08) – while maintaining her use of a high-entitlement, low-contingency form of the question. The 1.3-second silence (line 09) that follows suggests the call-taker's difficulty in responding to this request and foreshadows a further dispreferred response by the call-taker (Heritage, 1984; Pomerantz, 1984), which she then produces at line 10. In this response, the call-taker maintains her orientation to her (institution-based) inability to answer the question, as she again emphasizes the word cannot, before expressing hope that the service will be provided within the timeframe proposed by the caller. While the call-taker's use of the collective we here does mark a shift (following her prior use of they) to include her in the institution on behalf of which she is expressing this hope, it nonetheless continues to show that her personal evaluation of the legitimacy of the emergency is aligned with that of the caller, and thus that any delay in service is solely due to an institutional constraint.

Following a further brief silence (line 11) at a place prepared for the caller to align with the call-taker's assurances, and thus possibly again foreshadowing continued non-alignment by the caller, the call-taker pursues the caller's alignment by shifting from collective to personal pronouns (also see Lerner & Kitzinger, 2007) in producing the further assurance, I'll try and do my bes:t for you (line 12). The caller's formulation of this assurance in the first person – and emphasized by the extreme case formulation (Pomerantz, 1986) my best – proposes that a potential long wait should be understood as an unavoidable outcome of institutional capacity rather than as resulting from a lack of recognition of the legitimacy of the emergency on the part of the call-taker. Nonetheless, the caller responds by producing (following a second service receipt, Okay, in line 14) an account of the seriousness of the patient's condition as in support of a further claim of the urgency of the service request (lines 14–15). She thus appeals to the individual judgment of the call-taker with respect to the legitimacy and seriousness of the emergency as a basis for a less contingent service than the call-taker has thus far promised, thereby treating the contingent granting as a product of skepticism in this regard on the call-taker's part rather than institutional constraints beyond the call-taker's control.

After the call-taker aligns with this appeal by the caller (line 16), the caller produces a third service receipt followed by her first display of appreciation (line 18) – although the brief silence that precedes these tokens (line 17) may display an orientation to her alignment with the call-taker as reluctantly arrived-at. Further evidence of the caller's continued dissatisfaction is observable in her termination of the call prior to the call-taker's completion of a display of appreciation (lines 19–21). While this abrupt termination of the call results in it not being responded to, the call-taker's display of appreciation serves (as in Excerpt 6 above) to reverse default alignment of the caller and call-taker as beneficiary and benefactor. It thereby marks the call-taker's ongoing orientation to the misalignment between herself and the caller throughout this part of the call by thanking the caller for having accepted the diminished service the call-taker could provide.

While the evidence available in the call and dispatch package do not provide for a conclusive account for the basis of the caller's orientation in this case, earlier parts of the call (shown in Excerpt 9b below) do offer some evidence in this regard.

(9b) Deviant case opening [10293]

01  CT: Emergency medical service, Zian ↑speaking?

02   C: .hhhh W:=what do I do if I need to get an

03       ambulance to come out and fetch somebody,

04       they can't move, their legs are lame, and

05       and their back is also- they can't move at all.

06       (.)

07  CT: tch ↑Okay, you've  ̄got the right number,

08       ma'am. .hhh I'm speaking ↑to?

09       (.)

10   C: Nicole.

11  CT: .hh Nicole, tell me, do you have a contact

12       detail for me? <Cell phone number for me,

13       ↑please?

14   C: .hh Um: (.) okay, the person that i- that it's

15       for, can I give you their details?

16  CT: Okay, please.

17       ((1:47 omitted, CT gathers further details))

18  CT: Ma'am tell me, the patient is it a male or

19       ↑female?

20   C: Female.

21  CT: How old is she?

22   C: She::'s about forty nine I think.

Following the institutional and personal identifications produced by the call-taker in the call opening (line 01), the caller requests information about how to request an ambulance (lines 02–03) rather than (as is typical in other calls in the data set, and as reported in the literature discussed above) simply issuing an immediate request for an ambulance. In this way, the caller overtly displays a lack of knowledge (cf. Heritage, 2012) relating to the process of requesting an ambulance from a service-provision institution such as this one. The caller then refers to the patient using the "generic" person reference form (Whitehead & Lerner, 2020) somebody (line 03), rather than using a more informative reference form – for example, a kinship term or another categorical reference form – that would provide an account for her calling on the patient's behalf (Kitzinger, 2005). Moreover, the caller subsequently maintains this distancing from the patient (cf. Jackson, 2013) by referring to her as the person that i- that it's for (lines 14–15) and by offering the patient's phone number rather than her own (line 15) in response to the call-taker's request for a contact number (lines 11–13). Moreover, the caller displays her limited familiarity with the patient by providing a hedged estimate of her age through the turn-final I think (line 22), rather than offering an exact number.

Although the available evidence is tacit and therefore inconclusive, this consistent distancing of herself from the patient by the caller may be evidence that she is oriented to asymmetries in their respective social statuses and/or categories, such that she is seeking to secure an emergency service for a patient who is reliant on an under-resourced (public) institution, in contrast to her own taken-for-granted access to well-resourced (private) services. Similarly, the caller's displayed unfamiliarity with the institution, the high-entitlement, low-contingency form of her requests for information, and the dissatisfaction she displays with the call-taker's contingent granting of the service, may be evidence for her orientation to occupying a position in South Africa's stratified social order associated with eligibility for more immediate emergency services than a public institution in this context could promise to provide. In short, the misalignment between the caller and call-taker in this case may indicate the caller's tacit orientation to a set of class and/or race-based privileges that other callers to this institution typically do not take for granted (cf. Dominguez-Whitehead & Whitehead, 2014; Whitehead, 2020).

DISCUSSION AND CONCLUSION

Our analysis has demonstrated call-takers' use of components of grantings – namely indications of contingency and tag questions – that are routinely present in our data but rarely reported in previous studies. In contrast to the potential for such additional components to be treated as indicative of the call-taker's assessment of the legitimacy of a request for service as questionable (as in Excerpts 3 and 9), their routine uses and uptake suggest that they are "artful practices" (Garfinkel, 1967, p. 32) designed to manage the issues of trust associated with granting requests for service in a highly resource-constrained context. Our findings thus demonstrate the additional interactional burdens that participants (but most especially call-takers) may bear in managing systematic uncertainties with respect to the institution's capacity to immediately provide a service that has been granted – including their management of additional interactional troubles and associated call expansions that may arise from their deployment of these practices. In this way, it demonstrates one way in which matters understood to be "public issues of social structure" nevertheless come to be experienced as "personal troubles" (Mills, 1959, p. 8), through participants' efforts to manage the interactional contingencies these issues give rise to.

In revealing some of the pervasive interactional burdens associated with global and national-level inequalities, our findings serve to distinguish the type of "critical CA" we have conducted in this research with what has been called "applied CA" (Antaki, 2011) – or what we could call "interventionist CA." The latter type of work offers critical findings on the workings of talk-in-interaction that can serve as bases for interventions designed to alter the interactional conduct of participants as a means of bringing about more favorable outcomes. In contrast, a more favorable set of outcomes in relation to the issues of inequality revealed by our analysis is unlikely to result from changes in the interactional conduct of the participants, since the problems it addresses do not arise in the first instance from the interaction in the call, or from the call-taker's evaluation of the legitimacy or seriousness of the emergency at hand. Instead, call-takers can only seek to address the threats to trust associated with potentially systematically delayed services by inviting callers to anticipate and acknowledge or accept them. In this respect, they are by all appearances, doing the best they can to address these difficult circumstances one caller at a time, thereby temporarily papering over problems that, to be adequately resolved, require broader structural changes to improve the resource capacity of the institution – interventions that would require forms of expertise that extend beyond what CA can offer. We have thereby shown how critical CA can identify structural problems that are visible in the interaction, but are not of the interaction – though they must nevertheless be managed there.

ACKNOWLEDGEMENTS

We are indebted to the staff of the Health Communication Research Unit (School of Human and Community Development, University of the Witwatersrand), and especially its Director, Claire Penn, for their facilitation of the collection of the data used for this research, and for their support and encouragement. We would also like to thank Dexter Timm for his assistance in providing the data, and for offering valuable insights into the call center's processes. A report of this study was presented to the 5th International Conference on Conversation Analysis (ICCA-18) in Loughborough, United Kingdom, July 2018. We are grateful to Sandy Thompson for her assistance in developing the analysis, and to Hansun Zhang Waring, Nadja Tadic, and three anonymous reviewers for their helpful suggestions on earlier drafts of the chapter.

NOTES

¹ In a case reported by Garcia and Parmer (1999, pp. 301–302), after making a service announcement similar to those shown in Excerpts 1–2 (we'll send someone there sir), the call-taker instructs the caller to stay on thuh phone with me, okay? Since the tag question okay in that case is appended to an additional instruction not associated with provision of the service, but instead contributing to the project of keeping the caller on the line after the service request has been granted, it is designed to manage a different contingency than those on which we focus.

² It is important to note that we are not claiming that the use of these components is unique to the national context from which our data are drawn, nor that they are necessarily more common in this context (or others like it) than in those represented in the literature we have reviewed here. Distributional claims of this nature are beyond the scope of our analysis, which instead focuses on how these practices are used by the participants in our data. Future research could consider whether and/or how frequently they occur in particular contexts, and/or whether their deployment in other contexts is designed to manage the contingencies that they evidently manage in our data as opposed to being used in the ways evident in Excerpt 3, or in other ways altogether.

³ A reviewer suggested that authors include a "positionality statement" in the Data and Method section. The analytic viability of a pre-positioned, generalized statement focused on our positioning in relation to this research is not clear to us, since it would not consequentially inform readers about the bases for analyses that we have developed by reference to the specifics of the cases and practices at hand. Such a statement would instead invite readers to substitute common-sense reasoning about these analyses based on the unavoidably selective set of considerations we might include in the statement for evaluations of the analyses on the basis of their empirical merits (cf. Schegloff, 1997, 2005; Whitehead, 2020). If our analyses are found wanting in some way, then readers should attend to those shortcomings, which only then might raise questions about their possible bases, including whether they may relate to some feature of our positionality – that is, beyond our membership in the only category of generalized relevance for the work we have done here, namely conversation analyst.

⁴ The dispatch packages in our data set indicate that in some cases callers/patients make use of alternative, private, forms of transport after waiting for some time for an ambulance to arrive, resulting in the patient no longer being at the location to which the ambulance was dispatched when it arrives. The use of tag questions may thus also be a method for prompting callers to tacitly commit to waiting for the ambulance to arrive rather than seeking alternative transport after terminating the call.

⁵ Silences such as this, at places prepared for alignment by the caller, may be evidence for incipient trouble in relation to the alignment of the caller and call-taker as help-seeker and help-provider (cf., e.g., Heritage, 1984; Pomerantz, 1984). However, there is no explicit indication of disalignment by the caller, either in this case or in numerous others in the data in which silences are present at similar places (see Excerpts 5 and 6), and even in calls where the granting does not include a contingency component (see Excerpt 7). Thus, even if such silences do indicate a degree of dissatisfaction on the part of callers, the callers in these cases recurrently opt not to "go on record" as such.

⁶ In this case the address has been provided by a call-taker from a general emergency service line, 117, that the caller contacted to report the emergency prior to being connected to this call center. In such cases, the 117 call-taker provides the EMS call-taker with the relevant information that they have received from the caller, before providing the EMS call-taker with the opportunity to speak directly to the caller. In this case the EMS call-taker has taken up this opportunity in order to gather further details of the patient's medical condition from the caller, before issuing the granting shown in the excerpt. The caller's request for confirmation that the EMS call-taker has the address thus displays uncertainty as to whether this information was shared (accurately) by the 117 call-taker.

⁷ These features of the call may strike those familiar with South Africa – including ourselves – as characteristic of a middle-class white person calling on behalf of her Black domestic worker. However, the tacit nature of the evidence is such that a claim of this degree of specificity is necessarily speculative.

REFERENCES

Antaki, C. (2011). Applied conversation analysis: Intervention and change in institutional talk. London: Palgrave Macmillan.

Bell, M. C. (2017). Police reform and the dismantling of legal estrangement. The Yale Law Journal, 126(7), 2054–2150.

Clayman, S., & Heritage, J. (2014). Benefactors and beneficiaries: Benefactive status and stance in the management of offers and requests. In P. Drew & E. Couper-Kuhlen (Eds.), Requesting in social interaction (pp. 55–86). Amsterdam, The Netherlands: Benjamins.

Coovadia, H., Jewkes, R., Barron, P., Sanders, D., & McIntyre, D. (2009). The health and health system of South Africa: Historical roots of current public health challenges. The Lancet, 374(9692), 817–834.

Curl, T. S., & Drew, P. (2008). Contingency and action: A comparison of two forms of requesting. Research on Language and Social Interaction, 41(2), 129–153.

Desmond, M., Papachristos, A. V., & Kirk, D. S. (2016). Police violence and citizen crime reporting in the black community. American Sociological Review, 81(5), 857–876.

Dominguez-Whitehead, Y., & Whitehead, K. A. (2014). Food talk: A window into inequality among university students. Text & Talk, 34(1), 49–68.

Drew, P., & Walker, T. (2010). Citizens' emergency calls: Requesting assistance in calls to the police. In M. Coulthard & A. Johnson (Eds.), The Routledge handbook of forensic linguistics (pp. 95–110). Abingdon, UK: Routledge.

Garcia, A. C., & Parmer, P. A. (1999). Misplaced mistrust: The collaborative construction of doubt in 911 emergency calls. Symbolic Interaction, 22(4), 297–324.

Garfinkel, H. (1963). A conception of, and experiments with, 'trust' as a condition of stable concerted actions. In O. J. Harvey (Ed.), Motivation and social interaction (pp. 187–238). New York, NY: Ronald Press.

Garfinkel, H. (1967). Studies in ethnomethodology. Engelwood Cliffs, NJ: Prentice-Hall.

Hepburn, A., & Potter, J. (2010). Interrogating tears: Some uses of 'tag questions' in a child protection helpline. In A. F. Freed & S. Ehrlich (Eds.), Why do you ask? The function of questions in institutional discourse (pp. 69–86). Oxford, UK: Oxford University Press.

Heritage, J. (1984). Garfinkel and ethnomethodology. Cambridge, UK: Polity.

Heritage, J. (2002). The limits of questioning: Negative interrogatives and hostile question content. Journal of Pragmatics, 34, 1427–1446.

Heritage, J. (2012). Epistemics in action: Action formation and territories of knowledge. Research on Language & Social Interaction, 45(1), 1–29.

Heritage, J., & Clayman, S. (2010). Talk in action: Interactions, identities, and institutions. Oxford: Wiley-Blackwell.

Horwitz, S. (2009). Health and health care under apartheid. Johannesburg, South Africa: Adler Museum of Medicine, University of the Witwatersrand.

Jackson, C. (2013). 'Why do these people's opinions matter?' positioning known referents as unnameable others. Discourse Studies, 15(3), 299–317.

Kevoe-Feldman, H. (2019). Inside the emergency service call-center: Reviewing thirty years of language and social interaction research. Research on Language and Social Interaction, 52(3), 227–240.

Kitzinger, C. (2005). Heteronormativity in action: Reproducing the heterosexual nuclear family in after-hours medical calls. Social Problems, 52(4), 477–498.

Lerner, G. H., & Kitzinger, C. (2007). Extraction and aggregation in the repair of individual and collective self-reference. Discourse Studies, 9(4), 526–557.

Mills, C. W. (1959). The sociological imagination. Oxford, UK: Oxford University Press.

Moskos, P. (2008). Cop in the hood: My year policing Baltimore's Eastern District. Princeton, NJ: Princeton University Press.

Pomerantz, A. (1984). Agreeing and disagreeing with assessments: Some features of preferred/dispreferred turn shapes. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action: Studies in conversation analysis (pp. 57–101). Cambridge, UK: Cambridge University Press.

Pomerantz, A. (1986). Extreme case formulations: A way of legitimizing claims. Human Studies, 9, 219–229.

Raymond, G. (2003). Grammar and social organization: Yes/no interrogatives and the structure of responding. American Sociological Review, 68(6), 939–967.

Raymond, G., & Zimmerman, D. H. (2007). Rights and responsibilities in calls for help: The case of the mountain glade fire. Research on Language and Social Interaction, 40(1), 33–61.

Raymond, G., & Zimmerman, D. H. (2016). Closing matters: Alignment and misalignment in sequence and call closings in institutional interaction. Discourse Studies, 18(6), 716–736.

Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the organization of turn taking in conversation. Language, 50, 696–735.

Schegloff, E. A. (1996). Confirming allusions: Toward an empirical account of action. American Journal of Sociology, 102(1), 161–216.

Schegloff, E. A. (1997). Whose text? Whose context? Discourse & Society, 8(2), 165–187.

Schegloff, E. A. (2004). On dispensability. Research on Language and Social Interaction, 37(2), 95–149.

Schegloff, E. A. (2005). On integrity in inquiry...of the investigated, not the investigator. Discourse Studies, 7(4–5), 455–480.

Schegloff, E. A. (2013). Ten operations in self-initiated, same-turn repair. In M. Hayashi, G. Raymond, & J. Sidnell (Eds.), Conversational repair and human understanding (pp. 41–70). Cambridge: Cambridge University Press.

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.

Schütz, A. (1953). Common-sense and scientific interpretation of human action. Philosophy and Phenomenological Research, 14, 1–38.

Watson, R. (2009). Constitutive practices and Garfinkel's notion of trust: Revisited. Journal of Classical Sociology, 9(4), 475–499.

Whalen, J., Zimmerman, D. H., & Whalen, M. R. (1988). When words fail: A single case analysis. Social Problems, 35(4), 335–362.

Whalen, M. R., & Zimmerman, D. H. (1990). Describing trouble: Practical epistemology in citizen calls to the police. Language in Society, 19(4), 465–492.

Whitehead, K. A. (2020). The problem of context in the analysis of talk-in-interaction: The case of implicit whiteness in post-apartheid South Africa. Social Psychology Quarterly, 83(3), 294–313.

Whitehead, K. A., & Lerner, G. H. (2020). Referring to somebody: Generic person reference as an interactional resource. Journal of Pragmatics, 161, 46–56.

Zimmerman, D. H. (1984). Talk and its occasion: The case of calling the police. In D. Schiffrin (Ed.), Meaning, form and use in context: Linguistic applications (pp. 210–228). Washington, DC: Georgetown University Roundtable on Language and Linguistics.

Zimmerman, D. H. (1992). The interactional organization of calls for emergency assistance. In P. Drew & J. Heritage (Eds.), Talk at work: Interaction in institutional settings (Vol. Cambridge University Press, pp. 418–469). Cambridge, UK: Cambridge University Press.