Formulations in French Emergency Calls Transferred to Physicians

ABSTRACT

This article investigates a collection of 64 formulations delivered by physicians in French emergency calls. When a call is transferred from assistant emergency medical dispatcher (EMD) to physician EMD, the latter typically formulates a version of prior talk for the caller to confirm (e.g. alors votre mari euh il tremble c’est ça “so your husband uh he’s shaking, is that correct”). Formulations allow physicians to connect the current interaction to the previous one and deflect callers from launching into potentially extended new beginnings. Formulations allow physicians to move to key questions and instructions faster, as a formulation opens a restricted sequential slot for callers to either confirm or disconfirm. However, because of their selective nature, callers may use their responsive turn to alter physicians’ formulations—repairing, expanding, or pushing back against the terms of the formulation. Data are in French, with English translation.

When calling the French medical emergency number (15), the first person to take the call is an assistant de régulation médicale (comparable to emergency medical dispatcher or EMD), who identifies and prioritizes the medical situation and, if deemed relevant, transfers the call to an emergency physician (médecin régulateur) for more specialized medical assessment and instructions to the caller. I refer to the first role as “assistant EMD” and the latter as “physician EMD.” When a call is transferred to a physician EMD, the assistant EMD typically puts the caller on hold, contacts a physician EMD on the telephone, tells them which file (or “case card”) to select on the computer screen, and shares key information orally about the case before disconnecting from the call. Either assistant EMD or physician EMD can dispatch a response team, such as an ambulance crew. From the point of view of the caller being transferred, a unique call occurs, even though from an institutional perspective, three distinct interactions are recorded: (1) interaction between the caller and assistant EMD, (2) handover interaction between assistant and physician EMD, and (3) interaction between caller and physician EMD.

This study centers on what I term “physician formulation,” a type of turn that physician EMDs routinely deliver early in their interaction with callers. An example can be seen in lines 8–9 in Extract 1. With his formulation at line 8, the physician (PHY) requests confirmation on the social relationship between caller (CAL) and patient (vot` maman, “your mom”) and a version of the current situation (qu` a du mal à respirer, “who has trouble breathing”).

Extract 1. acr085_D

This article focuses on the interactional function of physician formulations, that is, the interactional job this practice accomplishes in this specific environment in which a caller has been transferred from assistant EMD to physician EMD. Based on the analysis of what physicians formulate, how they position their formulations, and what happens in the absence of a formulation, I argue that formulations are a connective and deflective practice. Analyzing the sequential development of the interaction, this study underlines the selective and necessarily partial character of formulations, and I argue that this makes them a potential locus of interactional work by callers to alter the terms of formulations.

The next section provides theoretical background on emergency calls and formulations in talk-in-interaction. The data and methods are presented in the third section. In Section 4, I analyze formulations as a connective and deflective practice, and in Section 5 I focus on callers’ responses to formulations. I discuss the findings of this research and further implications in Section 6, and conclude in Section 7.

Background

Emergency calls

The long-standing interest of conversation analysis (CA) for emergency calls can be traced back to the 1980s with pioneering research conducted on 911 calls in the United States, with a more specific focus on calls to the police (Whalen & Zimmerman, 1990; Zimmerman, 1984). Whalen and Zimmerman (1987) then Wakin and Zimmerman (2010) described the reduction and specialization of sequential structure compared to ordinary telephone interactions. Heritage and Clayman (2010) stressed that emergency calls are a vivid example of institutional interaction, as participants endorse the roles of caller and call-taker to engage in a task-orientated interaction centered around a single order of business. Zimmerman (1984, 1992) proposed that emergency calls are centered around an “interrogative series,” an extended series of question-answer sequences by which the call-taker assesses the situation and decides on institutional response. For a fuller overview of conversation analytic research on emergency calls, see Kevoe-Feldman (2019), and Riou (2024) for medical emergencies specifically.

Many emergency medical services (EMS) worldwide use systematic dispatch protocols, with a more or less scripted algorithm indicating which question or instruction must be delivered to callers in what sequence. France does not currently use a systematic dispatch protocol. Instead, EMDs follow general institutional guidelines on how to handle emergency calls and adjust response, but they are not bound by a specific set of questions or criteria during their interaction with callers. The French system also differs from many other EMS in that it is two-tiered: Callers first talk to an assistant EMD and then to a physician EMD. Callers may be unaware of the existence and respective roles of assistant and physician EMDs (Robert, 2022), and they do not necessarily know or guess that the two institutional representatives communicate together and share access to the digital case card—that is, the document in which dispatch information is entered automatically and manually (e.g., date and time of call, location and contact details, crews dispatched). Goffman’s (1974) concept of frames has been mobilized to analyze this type of interactional difficulties in emergency calls, first by Tracy (1997), who identified conflicting interactional frames as a key characteristic of emergency calls. She showed that whereas 911 call-takers may operate with a public service frame, callers may assume a customer service frame. In her study of emergency calls in Canadian French, Laforest (2011) focused on callers’ imagined representations of their role vs. the role in which EMDs cast them, and she underlined that emergency calls are for lay callers a rare—and sometimes unique—experience of this specific type of institutional interaction.

Despite the rich conversation analytical research on emergency calls, not much attention has been devoted to interactional practices surrounding call transfer. Two notable exceptions are Robert (2022) and Kevoe-Feldman and Pomerantz (2018). Robert (2022) analyzed how disaffiliation and conflict travels from one interaction to the next when calls are transferred to multiple institutional representatives in French emergency calls. Kevoe-Feldman and Pomerantz (2018) focused on how 911 call-takers from a state police dispatch center transfer callers to a local dispatch center in the United States. They identified a practice by which call-takers prepare callers for call transfer, and argued that there is an ideal interactional slot for them to do so. In the absence of such a practice, callers may not understand they are about to be transferred to another call-taker, and may move to disengage from the call. What happens after call transfer remains largely undocumented. Because many lay callers do not necessarily expect to be transferred to a new interlocutor (the physician EMD), French emergency calls provide an interesting observation ground for interactional practices deployed by physicians and callers to deal with this discontinuity.

Formulation in talk-in-interaction

Formulating is “the practice of proposing a version of events which (apparently) follows directly from the other person’s own account, but introduces a transformation” (Antaki, 2008, p. 26). Formulations do not necessarily repeat or recycle a specific segment of prior talk, for example, in the form of reported speech. Indeed, Bolden (2010) identified a practice by which participants do not formulate a version of prior talk, but of something that was left unsaid by co-participants.

Deppermann (2011a) identified that conversation-analytic research into “formulation” focused on three related but distinct phenomena. In the sense of Garfinkel and Sacks (1970), formulation refers to metacommunicative practices by which participants explicitly say what it is they are doing in the interaction. Fele (2023) focused on this first meaning of formulation for his study of Italian emergency calls in which callers use their first turn to say they are not calling for an emergency. Another conception of formulation was introduced by Schegloff (1972), who focused on the referential work of participants—that is, how they name or describe a person, place, or activity. Finally, a third use of the term formulation stems from Heritage and Watson (1979) and focuses on “formulations which are produced as responsive actions and which make confirmation by the producer of the original version relevant” Deppermann (2011a, p. 117). Deppermann (2011a) termed this “the narrow use” of formulation, and this is the approach taken here.

Research on formulation has been tightly connected to institutional interactions. As Drew (2003) noted, all the cases of formulations analyzed in the seminal papers by Heritage and Watson (1979, 1980) came from institutional interaction. Persson (2013) observed that Steensig and Larsen’s (2008) analysis of “you say X”-questions in Danish can be connected to research on formulation, even though the authors did not use the terms “formulation” or “formulating.” Most of the occurrences analyzed by Steensig and Larsen (2008) come from institutional talk-in-interaction, with only two cases in 60 hours of ordinary conversations. Interestingly, 31 of their 53 cases come from just two hours of emergency calls.

Drew (2003) argued that formulations can have different interactional functions, shaped by context. Formulations are used in professional meetings as a closing and topic-management device (Barnes, 2007). Antaki (2008) observed that formulations in psychotherapy can be used by therapists as an interpreting tool, for diagnosis and history-taking, and to manage interactional progress. In child counseling, formulations are a key practice to achieve active listening (Hutchby, 2005). Beach and Dixson (2001) observed that, during yearly checkups at a preventive medicine center, formulations simultaneously maintained mutual understanding and selected for attention one aspect of the patient’s talk, while disattending others. In mediation about neighbor disputes, Stokoe and Sikveland (2016) found that mediators formulated clients’ complaints to create an environment in which they could then deliver a “solution-focused question,” by which they guided clients into finding solutions to their conflict.

In his discussion of formulations in various institutional contexts, Drew (2003) observed that formulations are rare in ordinary conversation. However, he looked for formulations “resembling in any way ‘(so) what you’re saying is …’” (Drew, 2003, p. 297). Persson (2013) made the case that most research on formulating has focused on turns containing an explicit or lexical marker of inference, such as so (Bolden, 2008) or and (Bolden, 2010). Persson (2013) found 134 formulations in a 25-hour corpus of French talk-in-interaction. This does not make formulations a frequent phenomenon, but it highlights that their design can vary. Rather than focusing on formulations exhibiting a specific format, this article focuses on formulations occurring in a particular interactional environment—that is, the opening of emergency calls transferred to physicians.

Data and methods

The data used for this article are emergency calls in French handled in 2018 and 2019 by SAMU Rhône, the emergency medical dispatch center based in Lyon, France. All calls were for patients in cardiac arrest. This means that a patient’s heart had stopped when the response team arrived on scene; however, cardiac arrest was not necessarily identified by EMDs at the time of the call. A patient in cardiac arrest was unconscious and not breathing, therefore someone other than the patient called the emergency telephone number. Callers typically knew the patient as a relative, friend, colleague, or neighbor, but occasionally, the caller was a bystander unknown to the caller.

This study followed the methodology of conversation analysis, and ethnographic observation at the emergency medical dispatch center preceded data collection. Extracts were transcribed using standard conversation analytic notation, adapted for French by Groupe ICOR (2013). The list of symbols can be found in the Appendix.

I used a corpus of 100 emergency calls, with data including the audio recordings as well as the digital case cards in which EMDs entered and shared information during calls. I focused on the subset of 65 calls that were transferred from assistant EMD to physician EMD. To identify formulations, I used the criteria formalized by Persson (2013) for his analysis of formulations in ordinary French talk-in-interaction:

1. “tacitly displays, but overtly explicates, an understanding which is analyzably inferable from previous talk”;

2. “makes relevant as a next turn a confirmation or disconfirmation by the addressee”;

3. “concerns the addressee’s domain of knowledge and experience, circumstances to which the addressee has a more direct epistemic access”;

4. “does not explicitly assess the circumstances that the formulation concerns, neither positively nor negatively”; and

5. “is not syntactically interrogative, that is, both syntactically declarative utterances and utterances lacking a finite verb may function as formulations” (Persson, 2013, p. 21).

This study was approved by the Comité éthique et scientifique pour les recherches, les études et les évaluations dans le domaine de la santé (CESREES), which is the national ethics committee for research on health data (approval no. 11339421 Bis), and by the Commission Nationale de l’Informatique et des Libertés (CNIL), which is the French administrative body regulating personal data (approval no. DR-2023-213). The CESREES and CNIL waived the need to obtain consent for the collection, analysis, and publication of the retrospectively obtained and anonymized data for this study.

Physicians’ formulations: Connect and deflect

Connecting current interaction to previous interaction

Physician formulations targeted some aspect of their prior interaction with the assistant EMD. This clinical handover between assistant and physician EMD contained oral and written components, as it combined a brief telephone conversation and shared use of the digital case card. Although it is not the focus of the present study, it is worth mentioning that assistant EMDs typically delivered formulations themselves during handover, targeting some aspect of their prior talk with callers. In this sense, physician formulations can be seen as formulations of formulations. Physician formulations targeted a variety of informational items: patient name, age, gender, location, social relationship between patient and caller, current symptom(s), circumstances leading to the emergency call, and medical history. Each of these items can be sourced by the physician from their interaction with the assistant EMD and/or the digital case card pulled up on the computer screen.

In most calls (32/46), only one formulation was delivered, but some calls contained two (n = 10) or even three (n = 4) separate formulation turns, for a total of 64 physician formulations. When a physician delivered more than one formulation, the first one tended to contain nonmedical material, and aimed simply to check patient and/or caller identity. I refer to this type of formulation as “biographical.” An example can be seen in Extract 2. After a turn-constructional unit devoted to greeting and categorical self-identification (line 5), the physician delivers a first formulation with latching (line 6). This first formulation line 6 recapitulates the patient’s name, age, and the social relationship between caller and patient (vous êtes le fils de, “you are the son of”). Without pause but following a long in-breath, the physician launches into a second formulation (line 8). The second formulation is broken up into three components, each with rising final intonation: et-prefaced (“and”) relative clause characterizing the patient’s condition (un malaise, “fainted/collapsed”), a preposition phrase expressing location (dans ses toilettes, “in his toilet”), and explicit request for confirmation (c’est ça, “is that correct”). I refer to this type of formulation as “situational.” The caller orients to each formulation turn lines 6 and 8 as response-relevant, confirming both (lines 7 and 11). The caller’s second responsive turn (lines 11–12) does more than mere confirmation, which is discussed in the following section.

Extract 2. acr041_C

Physician formulations occurred early in the caller/physician interactions, after sequences commonly observed in the openings of emergency calls—namely, summons-answer, greeting, and identification. In Extract 2, the physician’s turn at line 5 interlocks greeting and self-identification (médecin à la régulation du samu, “physician dispatcher from samu”). In a latched turn, she then delivers a formulation at line 6. The summons-answer sequence occurs in all calls, while greeting and identification are common but not systematic. Physicians typically provide a categorical identification indicating their social role as a physician, and occasionally give their name as in this case. The caller usually simply receipts (cf. oui “yes” line 7) instead of responding with recognition or self-identification—but for a counter-example, see Extract 6.

In lines 4–6 of Extract 3, the physician delivers in the course of one turn a greeting (bonjour madame, “hello ma’am”), self-identification (je suis le docteur NOM, “I am doctor NAME”), formulation (c’est pour votre maman, “this is for your mom”), and dispatch announcement (les secours sont déjà en route, “help is already on the way”). With this turn including a formulation, the physician signals to the caller that this new interaction is a continuation of the previous one with the assistant EMD. The formulation in particular displays that caller and physician share knowledge about who the call is for (the mother) and who is requesting help on her behalf (the caller).

Extract 3. acr094_D

Packaging the formulation into the same turn as the greeting and dispatch announcement means that no interactional space is created for the caller to respond individually to the greeting or formulation. Therefore, the caller’s response (d’accord, “all right” line 7) can be interpreted as confirmation to any or all the actions accomplished by the physician’s turn. Alhough the physician’s turn is response mobilizing by virtue of its final rising intonation, he does not wait for the caller to respond. In overlap, the caller responds (line 7) and the physician delivers a second formulation at line 8 (↓e::lle e::st tombée de sa hau↑teu:rə¹ “did she fall from standing height”), with turn-initial lengthening suggesting that he is possibly searching for information on the case card while speaking. This turn functions as a pivot, as it formulates a version of events (the patient had a fall) and introduces a follow-up question (de sa hauteur, “from standing height”) to assess the severity of the fall. Thus, formulating allows a physician to connect the new interaction with the caller’s previous interaction with the assistant EMD, before moving on to the main business of the caller/physician interaction.

This research resonates with the existing conversation analytic literature on healthcare encounters involving a series of different interactions, in which participants seek to establish early on what is shared knowledge, and how the new interaction is connected to the previous one. For example, Robinson (2006) identified different question formats used by physicians to solicit patients’ presenting concerns, orienting to their concerns as new, follow-up, or routine. White et al. (2014) showed that surgeon–patient interactions in New Zealand are opened with a discussion of the referral letter, in which surgeons convey the knowledge they get from the referral letter, which patients can confirm or disconfirm. White et al. (2014) argued that these “referral recognition sequences” address patients’ reluctance to tell physicians what they might already know, and ultimately fosters alignment. This finding particularly resonates with our analysis of physician formulations, as they are instructing callers regarding the proper management of information between the assistant and physician EMDs.

Deflecting new beginnings

In addition to the connective use of formulations, physicians use them as a deflective practice. This interactional function of physician formulation is best evidenced by cases in which physicians do not formulate. The trajectory of such calls highlights what physicians seek to bypass or deflect with a formulation.

In 19 calls out of 65, physicians did not deliver any formulation during call opening. In five of these cases, calls had been transferred to physicians so that they could take over first aid instruction to callers. In such cases, assistant EMDs had already categorized the medical situation (e.g., cardiac arrest, choking) and initiated the project of performing first aid. A formulation sequence was obviously not expected, and physicians moved on to instructions right after, or instead of, greetings and categorical identification, as in Extract 4, in which the physician takes over instructions for ongoing chest compressions from line 4 onward.

Extract 4. acr035_C

This leaves 14 cases in which a physician formulation could have been relevant, as patient assessment had not been completed at the time of call transfer, and had to be taken up by the physician EMD based on the information already collected by the assistant EMD. A case in point is presented in Extract 5, in which two bystanders (CAL and BYS) are calling from a public location for an unconscious man they found lying on the ground. Right after the greeting and identification sequences, the caller launches into an extended telling from line 6 onward.

Extract 5. acr075_C

In the beginning of her telling, the caller used definite article le to refer to the patient (le monsieur, “the man”), as part of a left-dislocation that reiterates the subject as a pronoun (il, “he”), topicalizing the patient. Definite article le signals that the referent of “le monsieur” has already been established or is unambiguously identifiable in context. However, the next time the caller refers to the patient, she switches to the indefinite article (un monsieur, “a man”), which signals the opposite—namely, that the referent is not preestablished, for example, because it is mentioned for the first time. This is reinforced by the caller’s use of an existential structure (y’a, “there is”), typically used to introduce a new topic. This self-repair from definite to indefinite article, combined to existential structure, suggests that from the caller’s perspective, her referential work started with the assistant EMD does not carry over to her interaction with the physician.

Overall, the caller treats her interaction with the physician as disconnected from her previous interaction with the assistant EMD. Her extended turn at lines line 6–15 functions as a reason-for-the-call. With a reason-for-the-call, emergency callers express their main concern about the situation they are experiencing and for which they are (implicitly or explicitly) requesting help. Depending on local specificities, the reason-for-the-call sequence can be opened by the caller or the call-taker, for example, with a turn such as “tell me exactly what happened” if the Medical Priority Dispatch System is used (Clawson et al. 2015). Callers can also preempt their reason-for-the-call in other sequences early in the call (Riou et al., 2018). In any case, reason-for-the-call occurs early in the call, and in France, it occurs in the first interaction between caller and assistant EMD. If not prompted by assistant EMDs, callers deliver their reason-for-the-call at the first sequential opportunity (Rollet, 2015). That the caller in Extract 5 launches into a reason-for-the-call sequence again with the physician suggests that she treats the call transfer as an interactional reset, discontinuing her previous interaction with the assistant EMD. Because the physician’s next turn does not occur until line 16, the caller gradually discovers at each transitional-relevant place that she can continue her extended telling.

A second case with no physician formulation is shown in Extract 6, in which the caller has been recruited by his neighbor to get help for her husband, who is experiencing chest pain. The caller volunteers again his name (line 8) and address (line 9) right after the greeting sequence, in partial overlap with the physician’s categorical identification (line 7). In overlap with the physician’s receipt (d’accord, “alright” line 11), the caller provides his reason-for-the-call in a multi-turn telling (line 10 onward).

Extract 6. acr072_C

This case is slightly unusual in that the caller does not start directly at the reason-for-the-call but also reenacts the address sequence. This makes it a stronger example of call trajectory in the absence of physician formulation: Callers then launch into new beginnings. After a long in-breath, the caller delivers lines 10, 12, and 14 at a rather fast pace and without pause. The caller then gradually introduces new elements to his reason-for-the-call in lines 16, 19, and 20, as the physician indicates receipt and delivers continuers (d’accord, “alright” lines 11, 18; oui, “yes” line 13) but does not claim the floor despite silences at lines 15 and 17.

Extract 7 comes from the same caller’s previous interaction—that is, with the assistant EMD, and corroborates the argument that callers launch into new beginnings in the absence of a physician formulation. The caller’s reason-for-the-call conveyed to the assistant EMD in Extract 7 echoes what he then says to the physician in Extract 6. Note in particular how j` vous appelle (“I’m calling you” line 4), j’ai ma voisine là qui vient de sonner (“I have my neighbor here who’s just rung” line 4), and elle vient d` sonner là affolée (“she’s just rung now distraught uh” line 13) in Extract 7 with the assistant EMD are recycled in Extract 6, with the physician as j` vous appelle (“I’m calling you” line 10) and sa femme est venue affolée sonner (“his wife came and rang distraught” line 12). The echo is similarly striking between il a une oppression à la poitrine (line 8 in Extract 7) and il a des oppressions à la poitrine (line 16 in Extract 6), although the word for “pressure” is upgraded from singular (une oppression) to plural (des oppressions). Comparing Extracts 6 and 7 gives a measure of how much callers can recycle from their interaction with the assistant EMD to the physician. In sum, when callers launch into a new reason-for-the-call sequence, they treat their interaction with physicians as a new beginning rather than a continuation of the previous interaction with the assistant EMD. I argue that formulations are used by physicians to deflect such an interactional reset: By recapitulating some aspect of prior talk, formulations signal to callers that information has been shared between assistant and physician EMD, and therefore that a reason-for-the-call has already been established, at least partly.

Extract 7. acr072_A

The examples discussed do not show that callers orient to the absence of physician formulations. Such a claim would propose that callers, who may be calling the emergency number for the first and only time in their lives, have the specific normative expectation that physicians would provide a formulation. However, what the data suggest is that when present, formulations deflect what callers typically do in this sequential environment. Timing is critical in both examples shown above. In Extract 5, the caller launches into her reason-for-the-call immediately after her greeting, despite the physician’s long in-breath indicating self-selection for the upcoming turn. In Extract6, the caller’s reason-for-the-call at line 10 is delivered right after his address turn, without waiting for the physician to respond, and without a beat of silence, but prefaced with a long in-breath audibly indicating his not yielding the floor. Thus, in the opening of caller/physician interactions, callers may deliver a reason-for-the-call at the very first opportunity (as in caller/assistant EMD interactions, cf., Rollet, 2015), or create that opportunity with competitive overlap with physicians.

When physicians do not deliver a formulation at the earliest opportunity, this leads to a substantially different call trajectory. Callers’ new reasons-for-the-call vary in content and length but, as illustrated in Extracts 5 and 6, they tend to display characteristics of storytelling (Mandelbaum, 2013) delivered over a multi-turn project, for example, background material such as location, a temporal sequence of events with complicating action, and a story ending. Additionally, they can recycle information already conveyed to the assistant EMD, which is therefore potentially already known or available to the physician. The physician is consequently cast into the role of story recipient, as can be seen with the physician’s continuers in Extract 6 (d’accord “alright” 11, 18, 22; oui “yes” 13) and closing receipt token (d’accord 15) in Extract 5. Neither physician in Extract 5 or 6 competes for the floor during the caller’s telling. Even though there is no interactional evidence that physicians orient to caller’s new beginnings as problematic when it does happen, it is still a call trajectory that most physicians actively avoid, precisely by delivering a formulation. From the point of view of the institutional agenda of emergency medical dispatch, callers launching into new reasons-for-the-call can be considered a risky interactional trajectory regarding time management. There is a strong institutional incentive for physician EMDs to make every turn-at-talk count, and to maintain control of the interactional initiative. In that light, an interactional reset in which callers provide their reason-for-the-call again means that the latter launch into a potentially extended telling, casting the physician into the role of story recipient for an unknown amount of time, and with no obvious opportunity to regain the floor before the caller’s project is completed.

In this section, I argued that formulations reach back to previous talk and, as such, function as a signal to callers that their interaction with the physician is not a new beginning but a continuation of their interaction with the assistant EMD. In addition to this connective feature, formulations are also a deflective practice by which physicians block an interactional reset where callers would launch again into their reason-for-call.

Callers’ responses: Confirm and alter

The “power of the formulation to delete, select, and transform” (Antaki, 2008, p. 31) means that the choice of what to formulate and offer for confirmation is necessarily selective and partial. In addition, physicians do not have direct access to callers’ prior talk: Their formulations target prior talk from a different interaction, in which the caller was not a participant. Instead, physicians formulate something about their interaction with the assistant EMD, an interaction that in itself formulates a prior interaction between caller and assistant EMD. In this specific institutional context, we can expect the filtering effect of formulation (Antaki, 2008) to be present and result in formulations that do not necessarily encapsulate the caller’s main concerns in their own terms, leading in turn to potential elaboration, disalignment, and push back in callers’ responsive turns.

Confirming formulations

Out of the 64 physician formulations, 49 received a confirming answer. However, only half of those (26/49) were mere confirmation, defined here as an answer whose format is an interjection or repetition, two basic options to answer polar questions (Enfield et al., 2019). The distinction between “mere confirmation” and “more than confirmation” is borrowed from Persson’s (2013) prosodic analysis of formulations in French talk-in-interaction. Extract 8 showcases both interjection and repetition confirmations. The caller’s confirmations at lines 3 and 7 feature interjections, with polarity matching that of the formulation, either positive (oui, “yes,” line 3) or negative (nan nan, “nah nah,” line 7). In addition, the answer at line 7 also contains partial repetition (plus du tout, “not at all anymore”) of the physician’s question at line 5.

Extract 8. acr057_D

Confirming answers have been consistently identified as the preferred response to polar questions (Schegloff, 2007). Interjection answers in particular far outweigh any other type of confirmation in English talk-in-interaction, representing 88% of all confirming answers in Stivers’s (2022) data. Although there is no equivalent research quantifying confirming answers in French, it still raises questions that, in the data presented here, more than half (38/64) of caller responses to physician formulations are not mere confirmations. This is because they do more than confirmation, are non-answer responses, or receive no response at all. Table 1 summarizes caller responses to physician formulations.

Table 1. Caller response to physician formulation.

Caller response	Biographical formulation	Situational formulation	Total
Confirming answer	18	33	51
Mere confirmation	15	11	26
Expanded response	1	9	10
Narrative expansion	2	13	15
Non-confirming response	3	10	13
Disconfirming answer	2	7	9
Non-answer response or no response	1	3	4
Total	21	43	64

Interestingly, 15 out of the 26 mere confirmations were given after formulations that targeted only biographical details about the caller and/or patient. Two cases discussed so far illustrate this: line 7 in Extract 2 and line 3 in Extract 8. “Biographical” formulations are more factual in nature, are thus arguably less affected by the filtering effect that formulations have on prior talk, and most of them (18/21) received confirmation. By contrast, “situational” formulations, which integrate material about the situation (e.g., symptom, patient’s position), are more vulnerable to the filtering effect of formulation. While 33 out of 43 situational formulations received confirmation, only 11 of those answers were mere confirmations. Callers therefore seemed to orient to most situational formulations as requiring elaboration or disconfirmation, two trajectories analyzed in the following section.

Altering formulations

Extract 2 was used above to illustrate a physician delivering two formulation turns in one call. Turning back to this example (reproduced here in part), we can see that, despite the caller’s confirming answer to the second (situational) formulation, he pushed back with an expanded response altering the terms of the formulation.

Extract 9. acr041_C (reproduced from Extract 2)

The physician’s second formulation at line 8 described the patient as having had “un malaise,” which can refer to fainting or a general feeling of discomfort. I translated it as “fainted/collapsed” to maintain ambiguity as to whether the patient lost consciousness or not, however, contrarily to “collapsed,” “malaise” does not necessarily imply that the patient fell down. The physician opted for the past conditional (aurait fait), which can be analyzed as a marker of reportative evidentiality (Dendale, 2018). I translated it with epistemic marker “seems to have” in the transcripts for the sake of idiomaticity. However, this use of the past conditional is closer to “allegedly” in journalistic or legal English, as the speaker signals that they are reporting a proposition coming from another source, without personally confirming or disconfirming it. With “aurait fait un malaise” at line 8 in Extract 9, the physician therefore signals that she is reporting information obtained from the assistant EMD, but that she is suspending its accuracy until the caller confirms or disconfirms. This creates an opportunity for the caller to do more than confirmation, which he seizes in his next turn. The caller’s response at line 11 initially confirms (oui, “yes”), then downgrades this confirmation with epistemic disclaimer je sais pas (“I don’t know”; see Pekarek Doehler, 2016), gently pushes back against the term “malaise” (une sorte de malaise, “sort of fainted/collapsed”) and repairs it with an alternative term prefaced with ou (“or”): ou en état de choc (“or in a state of shock”). The caller thus ostensibly provides the confirmation projected by the physician’s formulation, but uses this sequential opportunity to alter the terms of the formulation and advance his own concerns about the patient.

Extract 10 is a case in which the caller disconfirms after physician formulation. The formulation at line 6 requests confirmation as to whether the caller’s husband is “shaking” (“tremble”). This characterization does not originate from the caller. During handover (transcript not shown), the assistant EMD hypothesized to the physician that the patient was having a seizure (je pense qu’il convulse, “I think he’s convulsing”), based on the caller’s report of strange noises and movements of the mouth.

Extract 10. acr089_C

Following the physician’s formulation, the caller disconfirms at line 8 with the negative interjection “non” and a repetition reversing the polarity of the formulation (“il tremble pas”), then launches into an extended turn in which she recycles much of what she previously told the assistant EMD (transcript not shown). The assistant EMD’s notionalization (Deppermann, 2011b) of the case to the physician as a “convulsing” situation was rephrased by the physician to the caller as a “shaking” (“tremble”) situation. With her response at line 8, the caller reaches back into interactional history (Deppermann, 2018)² for her original version of events, thus correcting the semantic transformation displayed in the physician’s formulation.

Physician formulations occur in a sequential environment that is particularly difficult for participants to navigate.

Extract 11. acr051_C

In Extract 11, the formulation at line 3 describes the caller as the patient’s wife (“la femme de monsieur”). The caller confirms but immediately transforms it as “his partner” (“sa compagne”), which she repeats after a short repair sequence when the line breaks. This suggests that callers are highly attuned to physician’s formulations and how they describe the medical situation as well as social relationships. Callers treat as repairables such items offered for confirmation in physicians’ formulation, and attend to them even when they otherwise express a sense of urgency at the situation they are facing. During ethnographic observation, some EMDs expressed annoyance at callers who devote interactional space to what physicians perceive as medically irrelevant details and therefore a waste of precious time. It is possible however that callers orient to such details in physician’s formulations as possible signs that their situation was not adequately understood by the assistant EMD and/or not adequately shared with the physician EMD.

Extract 12 is an extreme case in which the physician’s formulation precedes open conflict between participants and early termination of the call.

Extract 12. acr074_C

The initial caller told the assistant EMD he was calling for his father (c’est pour mon papa, “it’s for my dad” transcript not shown). The physician integrates this social relationship to his formulation at line 4 (vot` papa, “your dad”). However, while assistant and physician EMDs were attending to handover, and unbeknownst to them, the patient’s son gave the telephone to someone else. This new caller’s first turn (oui, “yes” line 2), in the sequence reopening the communication channel, is delivered with a rising-falling pitch contour but otherwise calm voice. In response to the formulation, the caller takes issue with being categorized as the patient’s father (écoutez c’est pas mon papa c’est un ami, “listen he’s not my dad he’s a friend” line 7), switching to a chanting, wailing voice situated rather high in his pitch register. The caller expresses great distress in this call (j` vous en supplie, “I’m begging you” line 8; on sait plus quoi faire nous, “we don’t know what to do anymore,” line 11) and his understanding that the patient is dying (lines 14 and 20). Conflict between caller and physician EMD quickly develops and its audible starting point is the caller’s response to the formulation. Categorizing the patient as the caller’s father is likely not the main reason why the call escalates, yet disaffiliation starts in the formulation sequence. The caller devotes his second turn-constructional unit to repair it (line 7), even before expressing his concern at the urgency of the situation. The caller’s responsive turn after physician formulation can therefore be used as an anchor to do more than confirmation or disconfirmation, and to push back against more than simply the version of events presented by the physician.

Overall, few caller responses simply confirmed physician formulation. When callers did not outright disconfirm, many of their confirming answers did more than confirmation. This indicates that callers used their responsive turn as an interactional opportunity beyond the connective function of physician formulations. Such findings echo Stivers and Heritage’s (2001) observations on doctor–patient interaction, in which patients’ expanded responses during history-taking “exploit the local environment to raise a matter that is apparently ‘on the patient’s mind”’ (p. 165), making them “resources for more extensive departures through which the patient can introduce her own agenda of concerns” (p. 155).

Discussion

In the absence of a physician formulation, or before it can be delivered, callers provide their reason-for-the-call, treating call transfer as an interactional reset, which can take the form of extended tellings casting the physician into the role of story recipient for an unknown amount of time. As formulations project confirmation, they restrict interactional space for callers’ next turn, with a clear opportunity for physicians to regain the floor shortly. Some callers provide mere confirmation, and the physician can move on to other questions and instructions. When callers do not confirm the formulation, or do so with elaboration, transformation, or push back, their response still comes in a shape that is more restricted than a reason-for-the-call would be. The selection and transformation of formulations is what callers target in their responses, yet this is also what propels the call forward, instead of returning a step behind to reason-for-the-call as if there had not been any prior interaction with the assistant EMD. Despite the fact that formulations are selective and partial—and therefore prey to caller elaboration, alteration, or push back—they lead to call trajectories in which physicians can advance their institutional agenda. Formulating an upshot of the case very early after call transfer participates to “shaping prior talk so as to become a suitable basis for next actions to be performed” (Deppermann, 2011a, p. 125).

This article furthers our understanding of formulations in institutional interaction. Studying physician formulations in emergency calls brings to light an unusual situation in which formulating is a recurring practice in call opening. Even though formulations have been mostly observed and studied in institutional interactions, previous research had not identified a specific sequential slot in which a version of prior talk routinely occurs. In addition, the formulations analyzed so far in the existing research have targeted prior talk by the co-participant, which is typically very close temporarily and sequentially to the formulation, and not the case here.

This study also provides further illustration of formulations’ filtering effect, as previously identified in the literature (Antaki, 2008; Drew, 2003). In the data presented here, the filtering effect of formulation is doubled. This “double filter” results from the specific institutional setting: Assistant EMDs formulate to physicians a version of callers’ talk, and then physicians formulate a version of that back to callers. Thus, physician formulations target prior talk from a different interaction, in which the formulation recipient (the caller) was not co-present. Stokoe and Sikveland (2016, p. 103) highlighted that a series of formulations can “work iteratively across long encounters to progress to a resolution of the dispute.” The data analyzed here highlight that formulations can have an even larger scope than previously identified, and connect different interactions despite change in participants.

This study resonates with previous research investigating the tensions between progressivity and intersubjectivity in emergency calls, summarized by Fele (2023, p. 44) as the “two conflicting demands of an emergency response: on the one hand, the response should be fast, but based on accurate information; on the other, obtaining accurate information takes time and impinges on the progressivity of the call.” Determining the exact location of the incident is an interactional project during which progressivity is frequently hampered (Fele, 2023; Nattrass et al., 2017; Paoletti, 2012). Perera et al. (2022) documented call-takers’ balancing act when asking if a defibrillator is available. That sequence is particularly challenging for lay callers, and leads to frequent repair initiation. Call-takers consequently face the dilemma of either prioritizing intersubjectivity (ensuring caller comprehension about defibrillator) or progressivity (moving on to chest compressions instead). The present study adds to this focus of inquiry by highlighting a deflecting practice used by EMDs to promote progressivity, at the cost of callers potentially taking issue with the way their concerns are formulated back to them.

Conclusion

In French emergency calls, physicians routinely deliver a formulation when callers are transferred to them by assistant EMDs. Their formulations recapitulate biographical details (such as patient age or name) and/or situational details (such as symptoms or patient position) for the caller to confirm. Physician formulation in the opening of transferred emergency calls is a powerful practice to connect the new interaction to the previous one, and therefore deflect callers from launching into new beginnings with a reason-for-the-call. New reasons-for-the-call can be at odds with physicians’ institutional agenda when they take the form of extended storytelling and cast the physician into the role of story-recipient. Instead, formulating allows physicians to shape and restrict upcoming caller turns into confirmation, yet still provides callers with an opportunity to elaborate or disconfirm, thereby advancing their concerns in their own terms. The power of formulating comes at a cost, however. Because formulations are necessarily selective and partial, callers often respond with elaborations and/or disconfirmation. Yet, even though callers can alter or push back against formulations, in a time-sensitive situation, formulating is still a more powerful practice to propel the call forward than not providing any formulation at all.

Acknowledgments

The author acknowledges Hospices Civiles de Lyon and, more particularly, Christian Di Filippo, Pierre-Yves Gueugniaud, David Schiavo, and Karim Tazarourte for facilitating access to data, as well as Delphine Hugenschmitt and Aurélie Henry for help with data extraction at the emergency medical dispatch center of SAMU Rhône. Many thanks are due to three anonymous reviewers for their thorough engagement with the manuscript and key observations which helped improve analysis.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ The symbol ə is used here to represent a nonstandard pronunciation known as “schwa prépausal” (Hansen & Hansen, 2003). At the end of the turn line 8 in Extract 3, an extra syllable is created through the insertion of reduced vowel [ə] at the end of the word hauteur (“height”). This phenomenon is transcribed here in order to represent overlap with line 9 accurately. Other occurrences in the data are not transcribed.

² This use of “interactional history” does not exactly correspond to the original definition, namely, “previously shared interactional experiences of participants in interaction” (Deppermann, 2018, p. 297), since the interaction between caller and physician EMD is their first interaction together. However, I argue that the concept of “interactional history” can usefully encapsulate the succession of interactions between the caller and two institutional representatives (assistant EMD and physician EMD) who are in contact with each other and maintain a digital record of all exchanges.

Appendix.

Transcription conventions

Table

⇒	target line where a formulation begins
→	other target line in the transcript
⌈non⌉	overlap with next turn
⌊non⌋	overlap with prior turn
.tsk	alveolar click
.h, .hh	in-breath
h, hh	out-breath
non-	cutoff
no::n	stretching of the sound preceding colon(s)
°non°	lower volume or whisper
NON	louder volume or yelling
non	contrastive stress
’	standard elision (e.g., j’ai “I have”)
`	nonstandard elision (e.g., p`tit “small”)
.	falling final pitch contour
,	slightly rising final pitch contour
?	rising final pitch contour
↑	pitch upstep
↓	pitch downstep
(.)	silence < 0.2s
(0.5)	pause or silence measured in seconds
xx	unintelligible, each “x” represents a syllable
inaud.	inaudible
(())	paralinguistic sounds and transcriber’s comments