Reducing the severity of incidents or emergency in Estonian emergency calls

1 Introduction

The topic of our article is reducing severity of incident or emergency by callers in their first turn in the calls to the Estonian Emergency Response Centre (EERC) (telephone number 112). By ‘reducing severity’, we refer to the instances where callers claim uncertainty about whether there is an incident or emergency or indicate that the incident or emergency they are reporting is minor, insignificant, or accidental. Such a reduction is untypical and unexpected in emergency calls, as one would typically expect callers to emphasize the seriousness of their situation to ensure prompt assistance.

Emergency calls are institutional talks which concern medical conditions, police-relevant incidents, and rescue operations, and are characterized by specific structure and roles of the participants (e.g., Heritage and Clayman 2010, 51–68). The caller’s purpose is to get help, while the call-taker has right and obligation to decide whether and what kind of help is needed (Raymond and Zimmerman 2007, 36, Heritage and Clayman 2010, 53–4, 56). For the call-takers, the call is a routine activity, and they follow the norms and agenda of the Emergency Centre. However, for callers, a call to 112 is typically a rare or unique event, they are often reporting a critical situation and may be stressed, nervous, or even angry (Heritage and Clayman 2010, 87–100). Most of the social actions in emergency calls are question–answer sequences where the caller is epistemically K+ and has knowledge of the problem that have occasioned the call, and the call-taker is epistemically K− and asks the questions to get information (Heritage 2012). At the same time, the call-taker has professional knowledge and experience in dealing with similar problems and situations.

Only part of the talks are real emergency calls where the caller reports an incident or emergency that requires help. Others are classified as non-emergency calls which can be divided into unintentional (misdialed numbers) and intentional non-emergency calls like offensive calls (e.g., from intoxicated persons), prank calls (e.g., from children), lonely complaint calls lacking any incident or emergency, and also irrelevant calls such as asking for the time or other information (see Tracy and Tracy 1998, Sampson 2002, Vaajala et al. 2013, Bain et al. 2016, Bolaños-Carpio 2017, EENA 2020). According to the guidelines of EERC, one should call to Centre when a situation is critical in such a way that someone’s life, health, or property is in danger (www.112.ee/en). However, the Centre also recommends people to call even when they doubt the seriousness of the incident or have reason to believe that something dangerous is potentially occurring or may occur (EERC 2020).

In the field of conversation analysis, numerous studies have explored emergency calls (see overviews Heritage and Clayman 2010, 51–100; Kevoe-Feldman 2019). We will focus here on the studies examining the first three turns of the calls.

In the first turn, call-takers respond to the summons; in the second turn, callers formulate their reason for their call (describe their problem, incident or emergency, report some activity, or make an explicit request for help), and in the third turn, call-takers initiate an interrogative series by asking questions regarding the information provided by the caller (Heritage and Clayman 2010, 57, Wakin and Zimmerman 1999, Whalen and Zimmerman 1987). Call-takers may decide that immediate assistance is required and start querying about details necessary for sending help (e.g., caller’s address). However, they also may decide that additional information is needed and begin a question–answer series to determine whether there is a real emergency (Kevoe-Feldman 2019, 233–4).

Research has demonstrated how call-takers’ responding variants to the summons can impact the continuation of the call. If call-takers respond to the summons with only an emergency number, callers usually inquire about a specific service. If their response includes a query about the incident, a description of the problem typically follows (Cromdal et al. 2012).

Estonian call-takers answer the summons with a standardized format turn, prototype variant of which is häirekeskus tere mis juhtus? ‘Emergency Response Centre hello what happened?’ (Extract 1, line 1). This is a multi-unit turn which combines self-identification of the emergency service (Häirekeskus ‘Emergency Response Centre’), a greeting (tere ‘hello’), and a query about the incident (mis juhtus ‘what happened’).^[1]

The crucial part of the turn is the question mis juhtus ‘what happened’ which occurs in 90% of the calls in our database. The verb juhtuma ‘to (unexpectedly) happen’ is related to the noun juhtum ‘(unexpected) distinctive single event/incident’. The question itself is a telling question, which projects a longer answer (Thompson et al. 2015, Rumm 2019) and assumes that callers will report the incident, describe it, or categorize the emergency in their response (see also Whalen and Zimmerman 1987).

The design of callers’ responses affects how the call continues and influences the chance of receiving help. For example, Zimmerman (1992) has shown how callers provide a maximally complete story in their opening turn to elicit the appropriate form of help. Also, it has been shown that the design of explicit request affects the continuation of the calls (Larsen 2013, Kent and Antaki 2020). If the callers design their request in a format that expresses a weak/low entitlement (e.g., I was wondering if x/Would somebody be able to x.), call-takers typically begin a question series to ask incident-related questions. Simultaneously, such request design projects more likely that the request will not be granted.

We focus on the talks where callers describe the problem in their opening turn. In some cases, they describe the incident or emergency simply as a fact like in Extract 1 (meil ‘plahvatas üks=eeee (.) ‘auto ‘klaav ‘one autoclave exploded’). In some cases, they upgrade the severity of the emergency to increase the probability getting help, e.g., using extreme case formulations, intensifiers, or providing extensive details to highlight the seriousness of the problem (e.g., Tracy and Anderson 1999, Rääbis and Rumm 2022). In Extract 2, the caller begins the turn with a neutral utterance mul=ee ja- ‘vasakus jalas on siuke (.) ‘valu ‘I have got such pain in my left leg’ (line 4). Then he upgrades the problem using the intensifiers nii (‘so’) and üldse (‘at all’) (lines 5, 7) and the adverb palju (‘many, much, a lot of’) (line 6).

However, there are also cases where callers use mitigated formulations that express uncertainty or lack of knowledge regarding the existence of an incident or emergency or use constructions, indicating that the incident or emergency is minor or insignificant. In Extract 3, the caller informs the Emergency Response Centre of a traffic accident he was involved in and reduces the severity of the accident by using the adjective väike (‘small, little’) (line 3).

In our article, we analyze talks where callers reduce the severity of an emergency or incident in their first turn. We have the following research questions. Which verbal means do callers use to reduce the severity of incidents or emergencies, what functions do these reducing markers serve in the talk and how to explain the reducing? How does reducing markers which diminish callers’ epistemic status and express low entitlement affect the progression of the talk, and the dispatch of assistance?

As far as we know, these issues have not been previously studied, and our aim is to fill this gap. The answers to these questions will provide us a better insight into the call-takers’ ways of thinking and help improve the quality of the service.

2 Data and method

Our data come from the Emergency Calls sub-corpus of the Corpus of Spoken Estonian of the University of Tartu (Hennoste et al. 2009). The corpus comprises 1,000 emergency calls of which 230 calls have been transcribed in detail. For reasons of confidentiality, all data enabling the identification of the speakers (names, addresses, social security codes, car numbers, etc.) have been changed. In the Extracts of the article, personal and place names and other references to specific persons or places have been replaced by the common name in backslashes (e.g., I am calling from the /city/) (see also Upser 2022, 426).^[2]

We follow the principles of interactional linguistics (Couper-Kuhlen and Selting 2018, 18–26) and use conversation analysis as a basic method (Sidnell and Stivers 2013, 77–99). From two starting points in interactional linguistics, we chose the one which starts from linguistic resources in certain sequential contexts and asks how the structures of language in this context are used in social interaction (Couper-Kuhlen and Selting 2018, 15–6).

Our starting point was the surprising observation that callers sometimes reduce their problems in their first turn. Next, we systematically went through callers’ first turns, identifying all cases where they expressed uncertainty about the possible existence of an incident or emergency, or used constructions suggesting that the incident or emergency was minor, insignificant, or accidental. After setting aside non-emergency calls, we identified 39 calls that constitute our collection for analysis. We divided our findings into words and syntactic constructions and analyzed the functions of different reducing means in the talk. Thereafter, we sought to explain the use of reducing markers and analyzed how the diminishing of epistemic status and expression of low entitlement by callers affected the subsequent actions of call-takers, and how diminishing influenced the provision of assistance.

3 Analysis

In the following sections, we first describe the use of lexical means and then the syntactic constructions and their functions. In Section 3.1, we focus on lexical means and their functions. We highlight various lexical means and demonstrate how they are used to reduce the severity of the incident and to indicate the speaker’s uncertainty about the information. In Section 3.2, we present general assessments manifested by negative utterances at the beginning of callers’ turns. With these utterances, the caller claims that there is no emergency, points out a potential emergency or incident, or downgrades the certainty of the upcoming description.

4 Discussion and conclusion

We analyzed the first three turns in calls to the Estonian Emergency Centre, focusing on callers’ first turns which are responses to the call-takers question mis juhtus ‘what happened’. Our analysis revealed that in 17% of cases in our data, callers reduce the severity of the incident or emergency, either by expressing uncertainty about whether there is an incident or emergency or by indicating that the incident or emergency they report is minor, insignificant, or accidental. We analyzed which means are used by callers for reducing and examined the functions of the means.

The first group of means employed are lexical items (Extracts 4 and 5), most of which are pragmatic/discourse particles (hedges, softeners). These particles are not specific to emergency calls, most of them are among the most frequent pragmatic particles in spoken Estonian (Hennoste et al. 2021, 2022).

The second group of means consists of negative utterances at the beginning of the turn. These utterances are syntactically and semantically (but not prosodically) completed clauses followed by a second part of the clause construction containing specific information about the caller’s problem. Functionally, these utterances are assessments that fall into three groups based on the information they project.

Some assessments project uncertain information, indicating that callers are uncertain about whether there is an incident or emergency. These assessments include the expression of uncertainty about the information like m:a=i=ole ‘kindel ‘I am not sure’ (Extract 5), and turn initial epistemic marker ma=i=tea ‘I don’t know’ (Extract 9). The marker I don’t know is used in the same function in various contexts, and its use has been revealed across several languages (see overview by Lindström et al. 2016). As far as we know, this usage has not been observed in emergency calls.

The second group of assessments project information about an incident that the caller does not qualify as an emergency, like mul ei ole ‘häireolukord ‘I don’t have an emergency situation’ (Extract 6).

The last group projects potential incident or emergency. Here, the negative utterance ei juhtund midagi ‘nothing happened’ is used, which formally expresses the absence of an incident (Extract 7), and variants of the same utterance containing adverbs or adjectives (like suur ‘big’, kiire ‘urgent’, traagiline ‘tragic’, etc.) which in negative constructions downgrade the reference of the utterance and therefore seriousness of an incident (Extract 8). These assessments seem to be specific to emergency calls, as they do not occur in a similar function outside of emergency calls in the Corpus of Spoken Estonian of the University of Tartu (approximately 2 million transcribed words).

Assessments are typically followed by the contrastive–concessive conjunction aga ‘but’. In our examples, it is generally a substantive contrast between mitigating and the following part of the turn. If the assessment states that there is no emergency or incident, the following part contains information that says that something might be an issue (Extracts 6 and 7). The conjunction could also indicate that the following part of the construction contains unexpected information compared to what was previously said (Extract 8) and could also serve as a point-making device which indicates that the speakers are moving to the important part of the answer (Extracts 6, 7, and 8; Schiffrin 1987, 164–77).

In some cases, callers use a single reducing tool (Extracts 6 and 7), but often multiple different tools are employed in the same turn (Extracts 4, 5, 8, and 9).

The caller’s purpose is to get help, and reducing the severity of the situation is untypical and unexpected in emergency calls. This raises the question of why callers reduce the severity of the incident or emergency. There could be several explanations as to why callers use reducing means.

First, spoken dialogue employs a lot of softeners/hedges (Extracts 4 and 5). This is driven by the need to respond quickly to the partner’s turn, which does not always allow for finding the exact expression (Biber and Conrad 2009, 236–41, Hennoste et al. 2021). In emergency calls, another aspect that explains the use of softeners/hedges also becomes evident: the caller is often stressed or nervous. In addition, the fact that the call-taker’s answer to the summons lacks a conventional opening sequence (greeting exchange), and in the very first turn, there is a question might also have an impact (Heritage and Clayman 2010, 53–4, 56, Rääbis 2009).

Second, sometimes the caller is closely related to the person involved in the incident and reducing the incident/emergency may serve to reduce potential blame (Extract 4, also Tracy and Anderson 1999).

Third, the callers are expected to have knowledge about the incident or emergency they report. Knowledge is a scalar phenomenon, ranging from complete knowledge to very little knowledge. Callers in our data express uncertainty about the information they possess, indicating that they do not take full epistemic responsibility for it. The marker I don’t know (Extract 9) signals the smallest possible epistemic responsibility for the information (Lindström et al. 2016).

Fourth, although the Emergency Response Centre recommends calling even for minor issues, by assessment, there is no emergency the callers indicate that their information does not meet their criteria for an emergency and contradicts the caller’s understanding that the call to the Emergency Centre implies an emergency (Extract 6).

Fifth, by using the utterance nothing has happened, callers align their response with the question which implies an incident (Extract 7). At the same time, it highlights that the answer is unexpected from the perspective of the call-taker’s question what happened assuming the incident.

Sixth, if the caller’s knowledge is incomplete or the problem does not meet the emergency/incident criteria for them, then their response is a dispreferred answer and therefore general assessments at the beginning of the turn are used also to delay the dispreferred information (Schegloff 2007, 58–96).

The most interesting case is the utterance that nothing has happened with adjectives/adverbials which implies that nothing serious or tragic has happened but projects information about a potential emergency (Extract 8). These cases need further analysis to explain the motivations of the callers.

The caller’s descriptions of the emergency can also be interpreted as implicit requests for help. Using reducing markers, callers express a low entitlement. Studies on callers’ explicit requests have revealed that when callers use the formats that express a low entitlement, call-takers initiate the question series, and such formulations often project a lower probability of sending assistance (Larsen 2013, Kent and Antaki 2020). This suggests that reducing the severity of the problem can lead to a similar outcome. However, our data and analysis do not support this conclusion.

First, there are cases where the call-taker initiates the question series (like in Extract 6), but there are also cases where the call-taker immediately decides on the need for assistance. This is evidenced by the fact that she asks immediately for information necessary for dispatching assistance (the address, Extract 9).

Second, reducing the severity of the incident does not mean that the call-taker accepts the caller’s assessment and decreases the likelihood of sending help. On the contrary, among the 39 analyzed cases, assistance was not dispatched in only seven cases; two cases did not qualify as incidents, and in one case, the potential incident resolved itself. In addition, there is no difference between reducing the severity of the incident and expressing uncertain knowledge about its existence.

In summary, our analysis suggests that call-takers do not rely on callers’ assessments of the probability, smallness, or existence of the incident or emergency. Instead, they use the acquired specific information and rely on their agenda and experience to determine what information to consider as a feature of an emergency.

Given that reducing the severity of incidents or emergencies by callers in their first turn is a common practice in Estonian emergency calls. It is important to note that despite callers’ attempts to mitigate the severity of the situation, it is often a serious matter. It is essential for call-takers to efficiently recognize and address potential emergencies based on the information provided. Therefore, it is crucial to address this behavior in the training of emergency call-takers to improve the quality of service.