Abstract
This article explores the construction of offences as ‘accidental’ in police-suspect interactions. The data comprise audio-recorded investigative interviews, which were analysed using conversation analysis. In these interviews, suspects often do not explicitly state the nature of their defence when answering police officers’ questions; instead, suspects’ defensive practices or techniques are embedded in the narrative accounts they give of what happened, thus exhibiting rather claiming their ‘innocence’. My focus here is on a particular type of defence, namely, one in which suspects portray an event as having been ‘accidental’. I show that this defence of ‘accident’ is associated with several discourse features including: building a plausible and trivial context in which the untoward incident occurred, describing the untoward action or series of actions, using impersonal or agentless constructions, and representing the disproportionality between the putative victim’s reaction and the aggressor’s untoward conduct. The accountability of these descriptions, however, does not rely on one unique feature, but rather on suspects’ ability to combine these features in such a way that each establishes the grounds for others.
Keywords
Introduction
A disturbing case was reported in the British media in 2015, concerning the murder of a 16-year-old girl from Bristol, Becky Watts. Her stepbrother, Nathan Mathews, and his girlfriend, Shauna Hoare, were arrested on suspicion of murdering Watts, after her dismembered body had been found in a house in Bristol. In his interview, Mathews explained that he had planned to kidnap the teenager in order to scare her into being more appreciative of life and of other people, particularly since she had been behaving badly towards his disabled mother, doing things such as leaving objects on the floor to trip her over. However, he claimed, his plan had gone wrong; according to him, he went to Watts’ room using a mask then panicked when the mask slipped, and she could see his face. Eventually, he admitted to having killed her, but characterized the incident as a ‘terrible accident’, stating that his girlfriend played no part in it. In court, the prosecution presented a quite different version, according to which the defendants operated in collaboration with one another in a sexually motivated crime. The police found that he had been accessing pornographic websites and watching pornographic videos, including some in which the actresses dressed like schoolgirls. The investigators had also discovered several intimate messages exchanged between Mathews and Hoare in which they discussed kidnapping a petite girl. 1 In short, the prosecution case was that the defendants acted ‘by design’ – that the death of Becky Watts was no accident, but a planned murder.
This case displays a number of important issues: first, it demonstrates that there may be alternative or competing accounts of what actually happened, each account supporting a different version of how and why the incident occurred. Second, in events that might be regarded as involving transgressive and deviant actions, people may attempt to defend themselves by contesting or deflecting allegations about their conduct; and third, portraying the incident as an accident is one such defensive strategy suspects may use to contest the attribution of intentionality to their actions.
In this paper, I will explore the use of the defence of ‘accident’ in police interviews with suspects in England, focussing on how suspects construct potential criminal offences as having arisen from accidental conduct. Before doing so, I will briefly discuss the literature on accounting and defensive strategies in police settings and then present the data and methods used for this study.
Accounting and defensive strategies
Accounts have been considered as a linguistic device mobilized by those whose actions have been submitted to evaluative inquiry (Scott and Lyman, 1968); in other words, they are verbal statements in which speakers explain to another their potentially deviant actions. These explanations could work as excuses or justifications (Austin, 1957; Scott and Lyman, 1968), the latter being an account in which the person accepts responsibility for the incident but denies any wrongdoing, arguing that it was the correct or right thing to do in the circumstances; whilst in the former, the speaker admits wrongdoing but seeks to mitigate or deny responsibility for that misconduct (e.g. having acted under duress).
Although that distinction appears to make sense on an abstract level, it falls short in describing precisely the actual use of these defensive techniques in police settings, as we shall see in the next sections. In concrete episodes, ‘to deny responsibility represents a partial claim that the act was not “really” wrong’ (Emerson, 1969: 143). For instance, when someone portrays the incident as an accident, the person is not only denying or mitigating responsibility, but also characterizing the episode as something other than wrong (malfeasance). Hence, the distinction operates best at a theoretical level, as a taxonomy, but it does not clearly capture what participants are actually doing as they answer questions in police interviews, for example. In this sense, the term ‘account’ should not be used to refer only to explicitly motivational or explanatory objects, but also to any descriptive resource that, in a particular interactional context, is used to display defensiveness and deflect blame allocation (Atkinson and Drew, 1979).
Accounting practices are a pervasive feature of most legal settings, including police interviews and interrogations, in which, when questioned by police officers (hereafter POs) about a potential unlawful conduct, suspects will often respond by describing the event in such a way as to deflect blame from themselves. In this sense, police-suspect interactions are a rich context in which to identify and analyse the language resources used by people to account for their untoward behaviour. Nevertheless, the literature on interactions in police interviews and interrogations has focussed more on POs’ questioning (Ferraz de Almeida and Drew, 2020; Johnson, 2020; Komter, 2003; Stokoe, 2009), setting aside investigating suspects’ construction of defensive accounts. Among the exceptions is Watson’s (1997) research on the methods offenders use to describe their victims in US murder police interrogations, explaining their motives and attempting to blame the victims for the incident. Similarly, Stokoe (2010) examined the categorial and sequential properties of denials by men accused of assaulting women in police-suspects interviews (England). Whilst blaming the victim and simple denials are certainly two forms of defences employed by suspects, they are not the only ones available to them. As I have pointed earlier, in order to defend against an accusation, one does not have to deny or blame someone else for it. One can also portray the event as an accident. By analysing instances in which this particular defence technique is used in police interviews in England, I hope to contribute to the understanding of the language resources used by people to describe their conduct and for recognizing these descriptions as a particular form of account.
Data and methods
This paper draws upon 27 audio-recorded police interviews with suspects in a police station in England. These recordings were produced as a standard police procedure and collected by other researchers as part of a project on neighbour disputes (see Edwards and Stokoe, 2011; Stokoe and Edwards, 2007). The extracts presented here were selected from three interviews in which the suspect, whilst building their accounts, portrayed the incident as an accident. All names and other identifying information were pseudonymized at source. These extracts were transcribed according to the conventions widely used in conversation analysis (henceforth CA) (Jefferson, 2004). CA was used to analyse the data in order ‘to discover and explicate the practices through which interactants produce and understand talk’, and thereby make visible the accountability of descriptions and social actions (Drew, 2005: 75).
Portraying an event as an accident
Building a defence in which an incident is characterized as an accident (i.e. having been accidental in character, or having resulted accidentally from some circumstances that perhaps might not have been anticipated) is not a privilege of high-profile cases such as the murder of Becky Watts. Even in more mundane cases, suspects may construct accounts in which their conduct is represented as accidental.
The following extract is from an interview between two POs and an 83-year-old man suspected of having exposed himself to a young girl.
In this extract, having seemed at first not to have heard the question (line 7), the suspect replied positively to the repeated question, confirming that he had exposed himself to a young girl (‘Oh yes (.) I have (.) yes’, line 11). The police officer then formulated the suspect’s answer, expressing some surprise, conveyed by the intonation at the end of the utterance (‘=You have’, line 12). The suspect subsequently confirmed again that he had exposed himself, but added that he had done so accidentally, not intentionally (line 13).
This brief example shows that, as in the case of Becky Watts, one defensive technique available to the suspect is to claim that the incident was an accident, hence a misfortune rather than a deliberate transgression (Austin, 1966). The suspect made this claim straightforwardly, having admitted doing the action about which the police inquired, using a formula ‘I did it, but it was an accident’. However, this is not the only way in which a suspect may indicate that an apparent transgression happened accidentally. As I will go on to show, there are cases in which the suspect does not say explicitly that the event was or has arisen from an accident, but rather leaves it to the police officer to infer the accidental nature of the event or incident.
In this paper, I focus on the dimensions that underlie the understandability of particular accounts through which suspects represent apparently untoward conduct as having been an innocent misfortune. What are the properties of talk that enable the POs, as competent members, to hear a particular description as a portrayal of an event/incident as an accident? In the following sections, I will present and discuss each of the analytical properties or dimensions, identified through an analysis of three interviews from my data set.
The description of the action
The following extract is from an interview in which two police officers (PO1 and PO2) questioned a teenage girl (S) who had been arrested on suspicion of having assaulted 2 another girl on the way home from school.
Following Austin (1957), the event as a whole could be described using a term such as ‘assault’, which seems to characterise what happened during a spate of time as a single activity. However, that apparent single piece of conduct is actually composed by a multitude of constituents or bits of actions. For instance, as can be seen from the extract above, the potential transgression is comprised of at least four main components: (i) running to catch a bus, (ii) spilling something on someone – an action which might have initiated the incident, (iii) someone turning and pushing, and (iv) retaliation (‘slapped or punched her first’) in anticipation. Each of these components consists of a number of other components. One can then deconstruct or dismantle the ‘machinery of the act, and describe (and excuse) separately the intelligence, the appreciation, the planning, the decision, the execution and so forth’ (Austin, 1957: 27). By doing so, one can analyse, step-by-step, how suspects construct their versions, especially in terms of the descriptive resources used to portray their innocence.
Thus, the first property to observe is related to the description of the action selected by the suspect to depict each part of the episode. As is evident in extract 2 above, by putting together that she was running and [thereby] spilt some liquid on someone, the suspect did not explicitly portray her conduct in the scene as having been accidental. Until that moment, it was up to the POs to draw that inference from the suspect’s description of the series of actions that might have initiated the incident. However, that description enabled the PO herself to infer that the suspect might have tripped whilst running, then spilt the liquid, as is evident in extract 3 below. This is implied through the description ‘and I was runni:ng’ (line 12), which is not used simply as an action-in-a-sequence (as it would be in, e.g. ‘I was walking, then I was running, then I was walking again’). The description ‘as I was running’ works in an adverbial manner, that is, to characterize and project the next action, spilling (‘an:d I spilt some. . .’, line 14). The claimed accidental character of the event, although not stated explicitly, is nevertheless evident in the suspect’s account; in other words, it is implied or embedded in the description of a careful sequence of events (Auburn et al., 1995) constructed by the suspect, rather than being stated explicitly as happened in extract 1. As the questioning continues, the PO approached the potentially accidental character of the suspect’s action by asking what had precipitated her trip (‘what caused you to trip?’, line 1).
The suspect did not say she had tripped. When in line 1 the PO asked about what caused the suspect to trip, she displays an inference, connecting ‘running’ and ‘spilling’, the two actions explicitly described by the suspect in her previous answers. The inference appears as an embedded presupposition in PO’s question: that the suspect was building a version in which the incident was precipitated by an accident. In making this inference, the PO exposes – and most likely challenges afterwards (see extract 4 below) the suspects’ account. The suspect becomes accountable for her version and should not only describe the actions of running to catch the bus and spilling soft drink on someone, but also establish a causal relation between them.
Looking again at extract 2, after building a description of a series of pieces of conduct which form the potential transgression, the suspect characterized the upshot of her conduct as ‘fighting’ (‘we just started fighti:ng’, line 22), implying the incident was an exchange rather than a unilateral attack. This account is supported by the use of the pronoun ‘we’ as the actor-agent and by the subsequent descriptions she built about the details of the fight, where the supposed victim was actually ‘grabbing on to her’, whilst she was ‘just hitting on her’ (lines 24–25). Her action description thereby minimized her responsibility by subsuming her actions within a different form of activity, that is, a fight; and in a fight, responsibility may be distributed between ‘the fighters’.
The following extract occurred halfway through the same interview, and begins with the first question asked in this interview by the second police officer (PO2).
PO2 began by building a version of the episode that contrasted with and challenged the suspect’s description – and implicit explanations – of how the incident had begun. He introduced that version by formulating (Ferraz de Almeida and Drew, 2020) the suspect’s prior account (‘=You’re saying. . .’, lines 1–2). The contrast is achieved by restating some particular pieces or items from the suspect’s version (‘the incident started because you tripped and spilled some pop’, lines 1–2) and introducing some first-hand knowledge about the consequences of the event (‘We’ve seen the top that Maureen was wearing’, line 4), highlighting an inconsistency between the suspect’s claim and some physical evidence (‘And I would say that that hasn’t been caused by someone tripping’, line 6). The description here is one that relies on shared commonsense knowledge. In terms of identity, PO2 does not seem to be talking as a policeman with specialized technical knowledge about fibres or cloths; he is just looking at someone’s clothes as ‘anyone’ would and reaching a conclusion which relies on mundane reasoning practices (Pollner, 1987). Building on Hart and Honoré (1985), it is the ordinary person’s notions of causation (and not the expert’s or the scientist’s) with which the police and the law are concerned.
Moreover, by using an impersonal or passive construction (Pomerantz, 1978), ‘it’s been thrown at somebody’ (line 9), PO2 characterized the scene in a way that contrasted with that offered by the suspect, thereby avoiding directly accusing her of (deliberately) throwing the liquid at the claimed victim. Whereas ‘spilling’, especially if preceded by ‘tripping’, implies an accident, ‘throwing’ conveys some degree of intentionality, which, as noted previously, turns out to be one of the points the police have to prove to charge the suspect with an assault.
The suspect’s version that the incident began through an accidental spillage and subsequently turned into a fight was challenged by the police officer later in the same interview:
PO2 opened with the characterization ‘accident’ (line 1), then substituting or adding ‘fight’, ‘or’ working either to add an item to a list or as a repair marker, substituting an item for a prior item (line 1). He then set those characterizations aside (‘those’ being ‘accident’ and ‘fights happen’) with the suggestion that what happened fitted another pattern, an ‘ongoing bullying problem’ (line 4), and re-categorized the event as an ‘assault’ (line 8), thereby successively upgrading the seriousness of the incident and, by implication, the blameworthiness of the suspect’s conduct that led to the incident. Moreover, the ways he portrayed the incident have different practical implications, especially in terms of legal relevance, since whether within a broader bullying context or not, an assault represents a criminal offence.
This recharacterization work relied also on the contrast that PO2 drew between the injuries sustained by the suspect (none), and those that PO2 later went onto detail that were sustained by the alleged victim, as seen in the extract below:
The fact that the suspect did not sustain any injuries from the incident, whereas the putative victim had to be kept 3 in the hospital overnight to have her damaged kidneys monitored is used to recast the event. Instead of being characterized as a fight that was the outcome of an accidental spillage of soft drink, the event is now portrayed, accountably, as an assault, committed deliberately, which had serious health implications for one of the parties involved.
In extract 5, the PO did not directly contest the suspect’s account, but rather showed how the same event can be reinterpreted if it is put into a different context, in this case, a context of ongoing bullying. If one takes those same facts and put them into that picture, there will be a different gestalt (Garfinkel, 2002). That provides a different context in which to interpret the conduct of both girls – if the context is a pattern of bullying, then the incident can be taken to have a more ‘blameable’ cause, that is, that the suspect intended to cause harm to the victim (mens rea element 4 ).
A similar device was used by an adult male suspect during an interview in which he was being questioned about an incident involving an alleged assault and criminal damage that occurred in his ex-girlfriend’s neighbourhood.
Here too, the suspect did not explicitly claim that the incident was an accident. Instead, his defence of ‘accident’ is embedded in a description in which he happened to have his hand in plaster (‘>as you can see I’ve got my hand in a plaster cast<’, lines 40–41), so that by knocking on the little window he inadvertently (accidentally) broke it. Although the accidental character of the incident is conveyed through several features of the description, including the context in which the action took place and the impersonal construction (see pp.18–19), it is important to highlight also the verb the suspect selected to describe the physical contact between his hands and the door. By describing his action as ‘knocking’, instead of ‘banging’ or ‘punching’, for example, he characterized his conduct as perfectly innocent, as something that anyone would do for calling someone inside the house. He then continued by describing a scene in which someone he ‘vaguely remembered’ lived there, and who responded to the breaking of the window by running and hurling towards him (lines 44–45), when they started to ‘wrestle’ in the little porch (line 47), from which point the suspect begins to cast himself as victim rather than perpetrator. However, the key point is that through his descriptions of the scene and his conduct the suspect’s descriptions imply that the consequences of his actions, breaking the door, the ‘fight’ and so on, were not intended, but were rather unfortunate accidents.
Building a context for the accident
The second property associated with the defence of ‘accident’ concerns the context constructed by suspects in order to describe each of the potential transgressive pieces of conduct.
In this extract, the suspect framed her account by depicting the unfolding scene as an ordinary, unexceptional one (lines 1–4). Her description consisted of a series of unremarkable features of her being an ordinary student, doing things every student does, in a particular order that makes sense for everyone who hears it. Each particular part of these descriptions sets up and accounts for the subsequent part and gives credibility to the version in which the incident has arisen from an accident – for example, it is feasible to spill some liquid on someone accidentally when not walking at a normal pace; one should have an excuse to run, which is provided by an attempt to catch up with a friend who went ahead to take a bus, etc. The context is one in which a student left school at the end of the day and ran to catch up with a schoolfriend who went ahead, in order to take the bus into town together. By highlighting the scene’s mundane, unremarkable and non-problematic features, the suspect is ‘doing being ordinary’ (Sacks, 1984: 413–429). She described the incident using a first-person narrative, which established the ordinariness of the scene paving the way to the incident, making available the juxtaposition of the ordinary with the extraordinary actions and events that subsequently occurred. In other words, this narrative was designed to portray the suspect as someone who was not looking for trouble, but rather the trouble came to her, accidentally (Kidwell, 2009).
In the extract below, the same device is used by the suspect, who built a narrative in which the descriptions of the actions that preceded the potential criminal offences are put into a context of a mundane scene.
The context portrayed by the suspect is one in which he asked a taxi driver to take him to the address of his ex-girlfriend’s neighbour. The normality or ordinariness of the scene, and therefore the suspect’s ‘innocence’ is conveyed particularly by his having requested the taxi driver, when they arrived at the address, to wait for him for only a short while (can you turn round in the close.hhh I’ll be two minutes’, lines 30–31) then drive him to the police station afterwards (‘and then I want to go to Potter Street Police Station’ lines 31–32). By doing so, he relied on some commonsense knowledge conveying the mundane, unproblematic and innocent nature of his conduct. Through his description, he implied that he would not have asked the taxi driver to wait for him nor told him that he wanted him to drive to a police station if he had been going to have a fight with someone (by design).
In each case, the suspect builds a particular context in which the incident occurred and the main action was committed. Each constructs a plausible or reasonable account or an account of legitimate circumstances for someone doing what they were doing – for example, running to catch a friend after school or taking a taxi and asking the driver to wait for him until gets back from a talk with his ex-girlfriend neighbour; and these accounts seem even more legitimate or plausible once one notices a contrast between the triviality or ordinariness of the context in which the incident takes place and the extraordinary nature of the incident itself. They are rather careful to put those descriptions together and without saying that it was unintentional; both suspects only imply that sense of non-deliberateness. In sum, the suspects designed their turns to depict the episodes as accidents – incidents that can be explained in terms of perfectly understandable circumstances at the time in which they occurred.
Agentless or impersonal (passive) constructions
The third device for describing an incident as having arisen ‘accidentally’ is the use of agentless or impersonal constructions by the suspect, as one can observe in the following extracts.
By contrast with extract 2 (line 8) in which the suspect had said it was she who spilt the liquid (‘I spilt some pop over Maureen’), here in extract 3 (line 9), she selected an agentless construction, that is, ‘it’, which de-agentified the conduct. By doing so, the suspect seemed to be oriented to sustain the accidental character of the episode and undermine any sort of claim that it was an intentional action. In this interview, selecting the verb ‘to spill’ allows the suspect to choose between using it transitively or intransitively (ergative verbs), as it is the case in extracts 2 and 3, respectively.
The same device of agentlessness is also used multiple times by the suspect in the interview about the fight that ensued after the suspect broke a small pane of glass in the front door; in extracts 7, 8 and 9, he defended himself against the allegation of criminal damage by saying that ‘the little window went in’ (line 42); ‘my plaster cast has gone through the little pane of glass (.) in the front door’ (lines 8–9) and ‘my plaster cast went through the little window’ (line 2), respectively. In these extracts, the suspects selected delexicalized verbs, that is, verbs that are devoid of lexical content compared to other options, for example, ‘broke’ or ‘smashed’ (Johnson, 2020).
Attributing responsibility to someone often involves discursively tagging a blameworthy action to an actor-agent who performed it (Pomerantz, 1978). Thus, by using agentless or impersonal constructions, the suspects above are building descriptions that are ‘shaped as reports of “events that happened” rather than of actions performed by actor-agents’ (Pomerantz, 1978: 117). Suspects use that device in order to deny or mitigate their responsibility regarding the crime for which they are investigated.
Disproportionality of the victim’s reaction
The fifth and last feature associated with characterizing an incident as having been accidental is the claimed disproportionality between the suspect’s conduct and the victim’s reaction, a disproportionality that, in the suspect’s account, is conveyed through contrasting the suspect’s reasonableness with the victim’s unreasonableness.
The central issues in this extract concern temporality and intentionality. In other words, which action has initiated the sequence of events. On one hand, the suspect’s description conveys that the alleged victim hit her first, which means that the suspect hitting the victim is depicted as a reaction, not a first action. On the other hand, the suspect acknowledges that the victim’s action was itself a reaction, namely a reaction to the suspect spilling drinking on the victim. However, this first action, that is, spilling, was not deliberate. In sum, the suspect’s account consists of:
a non-intentional first action by the suspect (‘I was runni:ng (. . .) and I spilt some pop over Maureen’:, lines 6–8);
an intentional reaction by the victim (‘and then she turned around and pushed me (. . .) and went to punch me’, lines 8–11);
an intentional defensive reaction by the suspect (‘And then (. . .) I slapped her (0.2) or punched her first’, lines 11–14).
By portraying her first action as accidental, the suspect provided the grounds for characterizing the victim’s reaction as the first action to have been done deliberately. The sense of deliberateness is conveyed by a hendyadic construction (‘she turned around and pushed me’, line 9), used commonly to complain about someone’s conduct (Drew et al., 2021). However, the suspect still had to account for the fact that both she and the alleged victim had acted deliberately and reactively. By way of displaying her innocence, the suspect distinguished between her own and the victim’s reactions, by conveying the sense of disproportionality of the victim’s reaction in comparison to what she (the suspect) had previously done. The suspect showed how unexpected the event was by exhibiting the reasonableness of her conduct, which was entirely innocent though it had an unfortunate consequence (her spilling drink on the other girl), but which resulted in the putative victim’s unreasonable reaction (turning around, pushing and preparing to punch her). In lines 8 and 9, the victim is portrayed as the one who first acted deliberately by turning around and pushing the suspect. After a pause (line 10) she added that the victim was moving towards her to punch her, an additional element which also serves to establish the contrast between the suspect’s reasonableness and the victim’s unreasonableness, through the ‘aggravated’ character of her response. – a response that was unexpected and unwarranted, transforming, in her account, not only the spilling but the whole event as an accident by recasting the fight as the unfortunate consequence of an accidental act.
A similar pattern is visible in a previous extract.
The suspect constructed a defence that involved describing background to the incident in such a way as to depict the reasonableness, harmless and transparently innocent character of his actions leading up to the incident. His account of having taken a taxi to where the incident happened included having told the taxi driver that he would be back in 2 minutes (line 31), after which he wanted to be driven to the police station (lines 31–32), evidently implying that nothing problematic would take place. That background was coupled with the non-deliberate nature (‘unfortunately’, ‘knocking on the door’, ‘the little window went in’) of the first transgressive piece of conduct (lines 40–42). Whereas the suspect constructed his conduct as having been reasonable and the incident as having been an accident, he described the victim’s reaction as disproportionate, conveyed through the deliberate and the exaggerated nature of the victim’s actions (‘a bloke (. . .) came running out (.) and hurled himself towards me’, lines 44–45). This contrast between the suspect’s and the victim’s actions is also evident when the former said he was ‘trying to tell him not to touch the kids next door any more’, whereas the victim, aided by his wife, starts beating him around the head (lines 49–50).
In sum, by portraying the victim’s reaction to the suspect’s untoward though accidental action as disproportionate, the suspect not only mitigated his responsibility regarding the incident, but also partially attributes the blame to the other person(s) involved, portraying the victim as having played a significant role in escalating the incident into what has come to be regarded as an assault, causing criminal damage, etc. The disproportionate character is established by describing the conduct of both suspect and victim; the unreasonableness of the latter’s actions not only derives from the accidental character of the event, but also enhances this character. These descriptions convey the inadvertent and non-deliberate character of the harm. In all these cases, it is undisputed that a harm has been done (e.g. breaking of a window, the injuries sustained by the girl), but the actions were not part of any sort of plan. It was not their intention to break someone’s window or to have a fight with and injure someone; the alleged victims overreacted, their responses were not commensurate to the previous (accidental) actions performed by the suspects.
Conclusion
The analysis above reveals some important issues concerning how different versions of the same episode are produced within police interviews. First, although POs conceive of police interviews as a search for information, it turns out to be that they involve collaborative work whereby information is interactionally produced. Suspects build their versions of incidents and episodes according to the questions they are asked and the information POs bring to the table. In other words, they will move across different defensive strategies throughout the interview, using multiple descriptive resources – movements that can be discerned in the deconstruction of the event into multiple actions or constituent elements, instead of looking at the incident as a single unit.
When Nathan Matthews was questioned about the death of Becky Watts, he described it as a ‘terrible accident’; the man questioned on suspicion of exposing himself to a teenage neighbour likewise described the incident as an ‘accident’. In each case, they labelled or categorized the event as accidental. However, in most cases, as I have shown throughout this paper, suspects do not say explicitly that the event was an accident, but rather describe a series of actions and circumstances through which a conclusion might be reached that the suspect’s conduct during the incident in question was not committed by design. As a descriptive strategy, this sort of defensiveness can thus be produced by a suspect even without using the word ‘accident’, conveying that implicitly and leaving the police interviewer to make that inference. Put another way, by using particular descriptive devices and resources, suspects manage to exhibit rather than claim that their actions were not done deliberately (Drew, 1992; Sacks, 1992).
But which then are the dimensions that underlie the understandability of this account as one in which a suspect is representing him/herself as being innocent? What are the properties of talk that enable the POs to hear a particular description as a portrayal of an event as having been [arisen of] an accident? As I showed in the previous sections, the accountability – hearability and recognisability – of these descriptions as being a particular defence strategy – portraying an event as an accident – is achieved through some main analytical properties, features or practices which include: (a) building a plausible and trivial context in which the untoward incident occurred; (b) the description of the untoward action; (c) the use of impersonal or agentless constructions; (d) representing the disproportionality between the victim’s reaction and the aggressor’s untoward conduct – or making the victim the assailant.
These features are evidently intertwined and their separation into different sections has been done for analytical purposes. The accountability of the descriptions I have examined here does not rely on one unique feature, but rather on combining (some of) them in a way that each one establishes the grounds to the other one. This form of arranging their narratives, and the co-selection of items and action descriptions, can constitute for these actions and for their agents, that is, suspects, a particular moral profile. The examples I have analysed exhibit the agents’ actions as having been accidental/non-deliberate both in terms of their content and their sequential organization (Jayyusi, 1993). The understandability, indeed the accountability (Robinson, 2016) of the suspects’ responses is achieved by deciding what actions are attributed to who, how these actions are described and the order in which these actions were produced.
Footnotes
Acknowledgements
In revising the paper for publication, I have benefitted from the insightful comments and suggestions made by Paul Drew and the anonymous reviewers. This paper is based on one of the chapters of my PhD thesis, completed at Loughborough University. I am grateful to Alison May, Martha Komter and Christian Greiffenhagen for comments on earlier drafts of this work. I am also grateful to Elizabeth Stokoe and Derek Edwards for making these data available to me, and to the police authority concerned for permission to use these data. The project for which they were originally collected by Stokoe and Edwards is ESRC grant number RES-148-25-0010 ‘Identities in neighbour discourse: community, conflict and exclusion’.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded in part by the CAPES Foundation, Ministry of Education of Brazil, grant number 0667-2014-5.