A discursive analysis of compliance, resistance and escalation to threats in sexually exploitative interactions between offenders and male children

There is a notable scarcity of empirical studies focusing on online sexual grooming using real world, naturally occurring data. Limited research with real victims (as opposed to decoys) has indicated that more overt forceful threats are employed by offenders in such interaction; however, they tell us little about how these threats are built up and managed by both parties. Furthermore, the majority of research focuses on female victims, with limited attention paid to male victims. The current study presents a discursive psychology analysis of chat logs between one offender (posing as a teenage girl) and ﬁve male victims under the age of 16 years, in order to explore how victims attempt to resist such manoeuvres in situ , and how offenders manage such resistance. The sexualized nature evidenced in our data contrasts with other ﬁndings which suggest that boys are not sexually solicited and that interactions with boys are less aggressive and forceful. Our ﬁndings demonstrate for the ﬁrst time how an offender escalated his issuing of threats following victims’ resistance and non-compliance to requests. Turning points that appeared odd in the online interactions suggest that they may be used to encourage children to be more reﬂective about any further engagement.

Our current understanding of online sexual grooming is predominantly based upon research which explores retrospective narrative accounts by offenders and/or victims, or analyses online interactions between offenders and decoys.This body of research has usefully illustrated various strategies and modes of manipulation; however, it falls short of examining how victims attempt to resist such manoeuvres in situ, and how offenders manage such resistance.In particular, sexual extortion, including threats to publish sexual images of the victim, is less extensively researched (Chiang & Grant, 2018).Moreover, there is a scarcity of studies exploring online interactions between offenders and victims using real-world conversational data between an adult and a child under the age of 18 years (i.e., a child not posed by an adult decoy).
The few studies that do exist demonstrate that more overt forceful threats are employed by offenders in such interactions; however, they tell us little about how these threats are built up and managed by both parties.Furthermore, the majority of research focuses on female victims, with limited attention paid to male victims, despite claims that interactions involving female and male children are qualitatively different (Grosskopf, 2010;van Gijn-Grosvenor & Lamb, 2016).The current study aims to fill this gap by examining in detail how making deals becomes a resource through which sexualized behaviour is operationalized in interactions between a male offender (posing as a teenage girl) and his five male victims.Our analysis pays attention to how deontic orientations are raised and made relevant, particularly in relation to the escalation to threats.As we are interested in the dyadic nature of manipulation and sexualization, discursive psychology is employed in order to pay more attention to how threats are initiated and carried out.
Sexual grooming is defined as 'a process by which a person prepares a child, significant adults and the environment for theabuse of this child.Specificgoals include gaining access to the child, gaining the child's compliance and maintaining the child's secrecy to avoid disclosure.This process serves to strengthenthe offender's abusive pattern, as it may be used as a means of justifying or denying their actions' (Craven, Brown, & Gilchrist, 2006, p. 297).Although this definition is based on sexual grooming in the physical world, its features can be applied to the online environment.Therefore, 'online sexual grooming' refers to the process offenders may engage in, whereby they prepare a child for abuse via Internet technologies.
Offenders who approach and attempt to engage children online are not a homogenous group (Ospina, Hartstall, & Dennett, 2010), with modes and strategies of manipulation varying from one offender to another (Whittle, Hamilton-Giachritsis, Beech, & Collings, 2013).Despite these differences, researchers have attempted to outline the various stages of online sexual grooming (Whittle, Hamilton-Giachritsis, & Beech, 2014).O'Connell (2003) proposed a model of 'cybersexploitation' based on a longitudinal observational study, identifying five distinct stages: (1) initiating a friendship; (2) relationship-forming; (3) risk assessment; (4) exclusivity; and (5) the sexual stage.However, this supposedly sequential order of 'grooming' has been challenged in light of offenders' varying styles of manipulation, as well as their adopting and recycling of different strategies where needed (Black, Wollis, Woodworth, & Hancock, 2015;Whittle et al., 2014;Williams, Elliott, & Beech, 2013).Consequently, victims have been described to become 'caught in a web' (Quayle, Jonsson, & L€ o€ of, 2012;Whittle et al., 2014).In the following sections, the present paper therefore aims to provide an overview of the modes and strategies of manipulation offenders may employ in order to offer insight into how they pursue and possibly achieve their goals.

Frequency and intensity
One form of manipulation relates to the frequency and intensity of online interactions which function to create a level of intimacy.For example, Whittle et al. (2014) demonstrated that frequent and intense contact involving sexual talk, compliments and flattery, as well as the sharing of sexual images, increased the dependency of their eight (mostly female) victims.Victims reported feeling 'addicted' to the conversations which often continued throughout the night and across a variety of social media platforms.Some victims equated this intensity to being in a 'loving relationship'.However, Whittle et al. (2014) noted that the two male victims in their study reported a lower level of frequency and argued that this may account for a lesser level of involvement/entrapment with the offender.Winters, Kaylor, and Jeglic (2017) examined conversational strategies of 100 offenders and decoys (of which 95% posed as girls) by using data published on the 'Perverted Justice' website (a United States non-profit foundation which uses decoys, that is, adult volunteers who pose as children, in order to expose 'sexual offenders').They found that sexual content was introduced quickly in the interaction; this was also observed in Marcum's (2007) study of three separate cases (involving adult males and decoys posing as female children) which had been published on the Perverted Justice website.Marcum (2007) described how the adults in the three cases used the following strategies to manipulate victims: (1) presenting sex as a choice; (2) normalizing sex with minors; and (3) raising the possibility that others may view sex with the offender as illegal (as a way of attempting to get the 'victim' to view their own role differently).

Introduction of sexual content
Similarly, through isolating offenders' speech, Black et al. (2015) found that offenders asked decoys (posing as male and female children) about their knowledge of the risk of online interactions, admitting that their own behaviour was inappropriate.A large number of offenders raised sexual matters within the first 20% of the interactions.Gupta, Kumaraguru, and Sureka (2012) reported that across 502 interactions on the Perverted Justice website, relationship-forming was a more prominent stage than the sexual stage.O'Connell (2003) further suggested that the creation and sharing of sexual images can be used to lower a child's inhibitions, as well as to blackmail them in order to engage in sexualized behaviours during the sexual stage.In the Winters et al. (2017) study, pictures were typically exchanged within the first day of contact; however, the authors were unable to verify whether these pictures were in fact of a sexual nature, and urged that further research examine whether the use of sexual images desensitizes victims, and/or leads to an escalation in sexualized behaviours.Olson, Daggs, Ellevold, and Rogers (2007) proposed a theory of luring communication whereby victims become entrapped through the offender's use of sexually suggestive and coercive strategies that are employed to gain their compliance.Intimacy can therefore be achieved by means of specific linguistic strategies that are used to increase rapport with victims.For example, Williams et al. (2013) thematically analysed the first 2 hr of interactions with decoys (posing as female children) and revealed how offenders synchronized their behaviours with the child's communication style.They reported that offenders adopted a 'text' style, thereby attempting to diminish the age gap through presenting matching interests, and often employing a 'guardianship' role (Williams et al., 2013).

Linguistic and communicative strategies
Lorenzo-Dus and Izura (2017) found that in Perverted Justice chats, compliments (primarily oriented toward appearance) were employed strategically, were highly formulaic, and achieved solidarity between the interactants.Other communicative strategies employed were evidenced by Baryshevtsev and McGalone (2018) through a comparison of pronoun usage between 'contact-driven' offenders, decoys and romantic couples.A unique pattern in their data was that more 'you' pronouns were employed by 'contact-driven' offenders, suggesting a more other-oriented focus in their conversations.Drouin, Boyd, Hancock, and James (2017) found that 98% of the offenders in their study used sexual words in their conversations with undercover police officers.Additionally, offenders were found to 'chat' less with male decoys compared to female decoys.Chiang and Grant (2017) considered the rhetorical moves and linguistic features of the conversational goals during online sexual grooming in seven different transcripts taken from the Perverted Justice website.Like Baryshevtsev and McGalone (2018), the authors identified the use of 'you', but suggested that its function was to place responsibility away from the offender, and shift the onus of creating sexual content onto the 'victim' (thereby arriving at a different interpretation to Baryshevtsev & McGalone, 2018).A further strategy employed was the use of the 'lol' (laughing out loud) acronym, which arguably functions to downgrade the seriousness of preceding contributions and wards off rejection, as well as feasibly making victims more comfortable.O'Connell (2003) suggested that, over time, offenders may employ increasingly more directive, aggressive and coercive strategies in order to achieve a victim's compliance.This was supported by Williams et al. (2013) who also found that once offenders had introduced sexual content into the conversation, they maintained and escalated that focus.In Whittle et al.'s (2014) study, the majority of victims felt increasingly pressured and frightened by sexual conversations.Sexual interactions were particularly prevalent with female victims (and less so with male victims).Furthermore, victims noted the controlling (e.g., possessive/jealous) and erratic behaviours of offenders, which confused them and made them feel out of control.When offenders were abusive towards them, the anxiety experienced by victims led them to try to regain the previous state of 'tenderness'.Whittle et al. (2014) found that some offenders normalized young people's practice of sharing images, resulting in some of the female victims feeling obliged to share them for fear of the offender's interest and 'love' being withdrawn.

Escalation
In the few studies that examined real-world interactions, an emerging picture of more forceful strategies is presented.For example, in Kloess et al.'s (2015) sample, the most commonly used strategies by offenders included flattery, compliments and affection, as well as persistence and manipulation.However, it was noted that throughout interactions, offenders may modify their initially more gentle tactics, subsequently becoming more directive.Offenders employing a direct approach made use of strategies that were of an aggressive, persistent, non-compromising and pressurizing nature in order to incite victims and achieve their compliance.Kloess, Hamilton-Giachritsis, and Beech (2019) further stated that offenders used either indirect or direct approaches relatively consistently.Chiang andGrant (2017, 2018) compared interactions between offenders and adult decoys to those between offenders and child victims.Their case analysis of the interactions between one offender and 20 female victims illustrated that overt persuasion and extortion were commonly used; however, this was not observed in the interactions between offenders and adult decoys.During overt persuasion, offenders pushed victims to comply (rather than using more subtle techniques such as rapport-building).In response to offenders' requests, victims used strategies, such as justifying non-compliance and bargaining, in order to defend themselves.Chiang and Grant's (2018) finding that more overt persuasion and extortion were present in naturally occurring data was further tested in a larger number of Perverted Justice transcripts of chat logs (with decoys; Schneevogt, Chiang, & Grant, 2018).Whilst a small number of instances of overt persuasion was revealed, these were rare.Furthermore, no examples of explicit sexual extortion were found.This difference between the two data sets may partly be explained by child victims displaying some resistance to sexual advances (Chiang & Grant, 2018), and the possibility of sexual images being shared (Schneevogt et al., 2018), which is not permitted by decoys on the Perverted Justice website.
Gender and sexual grooming Our current understanding of male children who are victims of online child sexual exploitation and abuse is limited, given that the majority of research has focused on female victims (Grosskopf, 2010;van Gijn-Grosvenor & Lamb, 2016;Winters et al., 2017).Across ten interviews and fifteen transcripts of chat logs, Grosskopf (2010) explored Australian police's experiences of posing as a child (aged between 13 and 14 years) as part of online child abuse investigations.The author found that in contrast to the findings of existing research on online interactions involving decoy 'girls', online interactions involving decoy 'boys' were less sexually explicit, aggressive and coercive.Whilst their sexuality and sexual experiences were established and discussed early on (in order to gauge their receptiveness), conversations around sexual topics were predominantly used to cultivate a 'friendship' (rather than a 'relationship', which appears to be more prevalent in online interactions involving 'girls').Interactions did therefore not feature the stages of relationship-forming or exclusivity (as outlined by O'Connell, 2003).
Furthermore, the interactions revealed that gay pornography was shared with a view to desensitize 'boys', with offenders arguably exploiting their sexual curiosity (a point also made by Wolak et al., 2004).In a study by van Gijn-Grosvenor and Lamb (2016), which involved the examination of online interactions between 101 offenders and decoys (with 52 posing as female children and 49 posing as male children) that were published on the Perverted Justice website, offenders appeared to pursue and focus more on immediate sexual gratification in online interactions with 'boys'whilst 'boys' were asked about the appearance of their genitals, 'girls' were asked about their general physical appearance.
Overall, the above studies have paved the way in providing insight into the range of manipulative strategies offenders may employ.However, the majority of them predominantly rely on self-report or analyses of online interactions between offenders and decoys.
A key criticism of the use of decoys posing as children is that the data derived from such interactions do not reflect the true nature of those that are occurring naturally and take place in the real world (Black et al., 2015;Chiang & Grant, 2017;De Santisteban et al., 2018;Lorenzo-Dus & Izura, 2017;Williams et al., 2013).There is an assumption that offenders behave in a similar way irrespective of whether they are interacting with a decoy or a real child (van Gijn-Grosvenor & Lamb, 2016).However, Winters et al. (2017) critically point out that decoys may initiate or welcome sexual content in order to obtain proof of devious intent.Williams et al. (2013) further argue that a decoy is likely to continue to engage with the offender even if the content is disturbing, whereas a true victim may choose to disengage (Kloess, Hamilton-Giachritsis, & Beech, 2017).Finally, Lorenzo-Dus and Izura, (2017) highlight how the majority of research studies are limited to an analysis at the content level, or merely focus on the offender's contributions, and therefore 'miss the structures, functions and inter-relationship aspects' surrounding these contributions (p.80).
Our study aims to build upon the evidence base of real world, naturally occurring online interactions by presenting a case study of transcripts of chat logs between an offender (who is posing as a teenage girl) and five male children.The nature of the data is particularly novel and complex in light of the offender posing as a teenage girl and allows us to examine whether some of the suggested gender differences for male children will be supported by our data.In particular, the analysis focuses on how threats escalate through the social practice of constructing deals, which we argue become a flexible resource employed by the offender and his victims.We will therefore be able to respond to a call by Winter et al. (2007) for future research to explore whether the exchange of sexual images leads to an escalation in sexualized behaviour.As part of our analysis, we employ discursive psychology, which is suited to highlighting how the dyadic nature of manipulation and sexualization operates in situ.Discursive psychology 'treats discourse as doing things in interaction and examines the ways in which psychological concepts are produced and made consequential in interaction' (Wiggins, 2017, p. 4).In contrast to other approaches, discursive psychology grounds analysis in how co-interactants display their understanding of a previous utterance within their talk/text rather than based on researcher interpretations (Wiggins, 2017).

Context
The data used in the present study consisted of transcripts of chat logs that were provided by a UK law enforcement agency.All case material was anonymized by a designated officer at the law enforcement agency prior to it being made available to the researchers.Originally, these data formed part of a larger data set that was analysed as part of a doctoral research project.The data presented here were not analysed as part of the doctoral research project due to a delay in the data being made available and resulting time constraints.
The project was granted full ethical approval by by the Science Technology Engineering and Mathematics Ethics Committee at the University of Birmingham.It is important to note that approval to work on this very sensitive data was based on the explicit understanding that the data would be accessed and worked on via an encrypted memory stick.The data were derived from closed cases that involved investigations of child sexual offences.As such, the male children in our study are vulnerable individuals who have often been victimized on multiple occasions.Whilst our analysis does not attempt to hide the coconstructed nature of these types of interactions, it prioritizes focus on the offender's manipulation strategies and seeks to balance the representation of both parties' contributions in a way that does not distort the analysis, whilst acknowledging that such representations involve children who have been victimized and sexually exploited.

Sample and data
Two cases were identified and selected by a designated officer at the law enforcement agency through their intelligence system, based on meeting the research project's inclusion criteria of the individual having committed (1) an offence of sexual grooming under Section 15 of the Sexual Offences Act 2003 (Home Office, 2003), or (2) any other offence under the Sexual Offences Act 2003 that involved sexual grooming.1Furthermore, individuals had to have been convicted of these offences, and their victims identified as children (i.e., below the age of 18 years), in order to be included in the sample.The transcripts of chat logs therefore represent case material which the law enforcement agency was able to retrieve from the offender's computer.
Unfortunately, only one of the cases could be used for the purpose of analysis due to the brevity of conversation/interaction in the other case.The case presented here consisted of five transcripts of chat logs between an adult male (posing as a teenage girl) and five male children under the age of 16 years (aged between 13 and 15 years), which took place between April 2011 and January 2013. 2 In such interactions, posts are necessarily short due to the nature of the mode of communication.The number of available transcripts was five, with a total length of 95 pages (single-spaced format, font size 10).

Analytical procedure
The study employed discursive psychology (Edwards & Potter, 1992;Hepburn & Wiggins, 2007;Wiggins, 2017) to analyse the data.Discursive psychology takes into account that discourse is situated in a particular interactional context, within which it is produced (here, online chat): (1) within a rhetorical framework as part of which discourse constructs one version of events, and thereby undermines alternative versions; and (2) within the turn-taking sequence, drawing upon conversation analytic understanding of grounding analyses in how utterances are understood by co-interactants as displayed in the next turn of talk/text (Wiggins, 2017).
Analytic stages followed those outlined by Wiggins (2017) and included the first author reading the data in depth, as well as coding the transcripts by examining the interaction in terms of social actions and psychological constructs.We noticed that the social action of deals was prevalent in the data, and a corpus of instances was collected, with the analysis focusing on how psychological matters are invoked as part of these actions, that is, how component parts are put together to do particular work, and we considered how these were oriented to.In total, there were 63 examples of deals being constructed, with 43 initiated by the offender and 20 initiated by the victim.There were an additional 34 examples where deals were made reference to (e.g., 'what would I have to do to get that video').Once we had a collection of bargaining sequences, we began to build up a collection of instances of resistance.During sequences of resistance, we considered how deontic authority was obliquely referenced by offenders, and how this became consequential as threats escalated.The analysis focuses on instances of these patterns to illustrate the phenomena.
Our discursive analysis draws upon the sequential placement and turn-design features associated with requests (Curl & Drew, 2008), and Craven and Potter's (2010) research on directives, specifically an extension of Curl and Drew's (2008) focus on entitlement and contingency, by outlining how directives are structurally designed to project compliance and deontic authority (Stevanovic, 2011;Stevanovic & Per€ akyl€ a, 2012), as well as how utterances are built as a threat (Hepburn & Potter, 2011).

Results
Across the data, sexualized interactions developed quickly.Relationship-building was not really evident except in a minimal way (e.g., praising and flattering victims).We identified one key social action which can be glossed (and is oriented to by participants) as 'making deals'.Deals typically involved written communication that was centred on the exchange of photographs, webcam displays of bodies or sexualized behaviour, and the exchange of videos between the offender and victims, and were ultimately key to eliciting compliance to requests and directives about sexual activities.Due to the predominance of the initiation by the offender, the analytic focus is restricted to examples which are arguably one-sided; however, they do incorporate victim orientations to this practice.First, we will outline how deals are often initiated and worked up as equitable.Next, we will focus on offender and victim orientations to issues of parity as a way to resist or gain compliance.We will then consider how deals escalate to threats.Finally, we will examine how subsequent deals invoke the deontic status (participants' rights to determine future courses of action) of the offender.

Constructing deals as 'equitable'
Deal-making between the offender and the victims occurred as a commonplace activity across the data.These bargaining sequences are initially worked up as a reciprocal practice, involving the sharing of photographs or videos (and often live webcam in the case of the victim).Some bargains were relatively straightforward question/answer sequences as in Extract 1 above.The way the question is designed with the use of 'wanna' constructs this as a 'free' choice.Framing the offer as making a 'deal' also works up the notion of a 'fair' exchange.This offer is minimally accepted with 'k'.In line 3, we see the proposal of the terms couched with 'how about', which again works up the notion of the proposed terms of a deal in a way that constructs the offer as an 'equitable', 'joint' venture.The offender makes gender relevant (Stokoe & Smithson, 2001) through the reference to 'my pussy'.Only the offender has epistemic access to his 'true' identity as an adult male (Heritage, 2012).
One particular interactional practice in designing deals to build up a sense of equity in the transaction is the use of particular formulations, such as 'I will. .., if you. ..', or variations thereof, such as 'if you. .., I will. ..' and 'if I. .., will you. ..'. Formulations such as these constitute actions that are routine and expectable, with Edwards (1994) discussing them as scripted, and arguing that they perform interactional business where they occur.
Directly before the extract below, the offender and victim had both expressed 'liking' each other's pictures, and the offender had also intimated that he had naked pictures and a video of 'herself', which the victims had requested to see.
Resistance and escalation to threats 995 In the extract above, the offender sets up the parameters of the exchange of photographs, including the quantity and format (lines 1-2).The offender constructs the deal as a proposal using an 'if you. .., I will. ..' formulation ('and if u will show urs then I will show more xx'), working up the deal as conditional, yet reciprocal.The addition to 'and I don't show my video for free :P' arguably raises a question about the value of the video format.The bargaining in combination with the cheeky/playful emoticon implies the worth of this offer as higher than a 'pic' (Pinch & Clark, 1986).
We see evidence that this interpretation is valid in the next turn from the victim (line 3), as he seeks clarification of what would be required to gain the video.In line 4, the offender again orients to the idea of parity in the exchanges.This orientation to parity is further developed with the conditional use of the device 'as im gonna. .., u should. ..' (line 5).The design of the construction works up the entitlement of the offender as an appeal for parity in terms of the proposed exchange with the victim (Curl & Drew, 2008).The victim accepts the deal, adopting a similar formulation (line 6), demonstrating a synchronicity of style at this point in the negotiating process.Both parties orient to an equal footing in terms of agency.
In Extract 3, the offender's request to see the victim's sister is refused, with the victim providing a justification of her being unavailable (lines 1 & 2).The offender pursues his goal through the strategy of orienting to a conditional exchange with the 'I will. .., if you . ..' formulation.The framing of the transaction is couched in an 'equitable' manner, whereby justifications for non-compliance are accepted.Edwards (1994) argued that such formulations can be used to make them appear routine and needing no account.
As the offender is posing as a teenage girl, his request to see the victim's sister might appear strange; however, there is no evidence in the subsequent response that would support the victim treating the request as dubious (other than his refusal to comply).In Extract 4, the request to 'get ur mate with u', and conditional offer to 'get my sister with me', is responded to with an evasive 'probably'.Combined with the winking emoticon, this is potentially worked up as something that the victim is agreeable to, which is supported by the extreme case formulation (Pomerantz, 1986) of 'anything you want us too'.Again, there is a potential indication that the sexualized activity here is problematic; however, it is not treated as such by the victim.As can be seen in the extracts below, the orientation to a deal is also constructed by the victims.
In Extract 5, the victim pursues photographs (previously mentioned by the offender) after some sexualized behaviour has taken place.In line 2, the offender responds using a conditional 'if', but employs the vague 'something nice of u xxxxx', arguably opening up the floor to sexualized interaction (which does indeed follow).Note here also how the offender uses 'x' as kisses, of which there are five, in contrast to the victim's two kisses in line 1.Kisses appear to be used to build rapport/relationship status, a point we return to later.
Extract 6 occurs after some sexualized behaviour has taken place via webcam.In line 1, the victim asks a polar question, setting the terms of how a response should be constructed in that they embody a preference for agreement (Heritage & Raymond, 2012).The victim also uses the modal format 'will you' to request a picture or video which displays an entitlement to a preferred response (Curl & Drew, 2008).The request is not initially responded to, as the offender initiates a new sequence (he continues to bluntly direct the victim to perform sexual acts via webcam), showing disaffiliation.However, in line 4, a conditional acceptance is delivered, which 'pushes back' from the constraints of the question (Lee, 2013), and operates as a transformative response, demonstrating a level of resistance (Stivers & Hayashi, 2010).
Thus far, we have demonstrated that deals are a social practice oriented to by both the offender and the victim.The formulaic nature of deals, using conditional framings such as 'if I. .., will you. ..' (and variations thereof) are routinely deployed within these bargaining sequences, and appear to work up the notion of deals as 'equitable', with an orientation to equal agency or symmetry between the co-interactants.However, in Extract 6, we see the offender resisting through a transformative response.At this point, both parties appear to have agency and be equally invested in the transactional nature of bargaining for sexualized material.Once on this trajectory of bargaining, it might be harder to resist becoming entangled at a later stage; whilst this rhetorical move is not automatically successful, it can be rebuked or contested (Edwards, 1994).

Breaches in the formulation of deals
The social practice of orienting to sexualized transactions and bargaining as deals has been evidenced above; however, such formulations raise issues of accountability (Edwards, 1994).In considering the data extracts presented below, it is useful to consider the distribution of deontic rights (i.e., determining how the world ought to be) when talk is oriented to future actions (Stevanovic & Per€ akyl€ a, 2012).Whilst the reader has knowledge of the epistemic status (Heritage, 2012) of the offender, the victim does not.Such issues become relevant at certain points during sexualized interactions, when the notion of equity and fairness is raised by victims and is also taken up by offenders.Notably, throughout the data, the offender orients to videos he has made as having 'more value' than photographs or sexual acts given or performed by the victim via webcam.
Prior to the exchange above, the offender had been issuing directives to the victim to perform sexual acts via webcam, to which the victim acquiesced.The request for a photograph initiated above is delivered with the modal verb 'I wanna', which orients to an entitlement for the request to be fulfilled (Curl & Drew, 2008).In the next turn, the discourse marker 'but' (line 2) therefore marks the victim's response as dispreferred (Clift, 2016), resisted on the grounds that he has complied with the offender's previous request (thereby challenging the offender's deontic right, and attempting to work up an equal deontic status).
In line 3, the offender minimally agrees, but employs one strategy prevalent across the data setthat the webcam picture is not clear, and the implication is therefore that the 'deal' is not 'fair'.This interpretation is validated in the victim's response, which raises the possibility that the promised video (in return for his sexual acts) is now potentially being withheld.The offender also orients to this notion of parity in his response, which constructs the inclusion of two people as more valuable than one.In line 6, a further attempt at appealing to a sense of justice is invoked by the victim as an extreme case formulation (Pomerantz, 1986) of 'but ive showed you everything', designed to defend the challenge and orient to a lack of parity in the sexualized exchanges.As with examples above, the construction of 'with my sister doin stuff' might be treated as dubious by the victim; however, this is not evidenced in the victim's turns.
One might typically expect further bargaining to occur at this point; instead, the offender does not pursue the issue but initiates a new question (line 7).Not responding to a plea or request, and thereby breaking the socially expected pattern (Sacks, 1992), is a further mode of manipulating the trajectory of the interaction.This strategy of nonengagement is employed frequently across the data and tends to work, as it does in this instance with the response 'yeah'.Below, we re-join the interaction a bit later on.
In line 1, the offender's request for action is structured in an interrogative format, which conveys a mitigated stance on deontic rights (Stevanovic, 2011).The use of flattery ('sexy best ass') is arguably designed to ingratiate, in order to gain compliance with the request, as has been noted in previous research (Black et al., 2015;Marcum, 2007;Williams et al., 2013;Winters et al., 2017).However, the victim provides a dispreferred response to the request for a (further) photograph, using the conditional 'if', thereby displaying resistance, which is also designed in line with a bargaining trajectory.The insertion of kisses by the offender, and the sequential placement of these within this extract, is a subtle way to confer favour, as are the numbers of kisses (i.e., xxxx) given, which are increased or reduced strategically at key points in the interaction.Notably above, there is only one kiss from the offender after the victim resisted complying with the offender's request (line 3).Excuses for the delay in sending the video include the length of time it will take to send (lines 4 & 8), bedtime (line 9), and parental presence and punishment (lines 14 & 15).
The offender also invokes his 'epistemic status' as a 'child' through providing further justifications for not complying with the victim's request.In line 6, he constructs a version of reality where he is restricted by bedtimes and parental control of laptop accessibility (in lines 14-15).These invocations are deployed as strategies to resist compliance with the victim's requests, and work up his identity as a teenage girl.The following extract (below) is again subsequent to a series of sequences where another victim obeyed the offender's instructions and performed a series of sexual acts via webcam.
The blunt deontic authority evidenced in the delivery of the imperative in line 1 is resisted by the victim with the extended directive ('noo').The victim provides a justification based on the concept of equivalent sharing ('noo send me the video first ive done enough').This treats the imperative as problematic in terms of deontic rights.The offender disputes this orientation to parity by citing the length and vague content of the video, and claiming that the victim has not complied with a request to show his 'ass'.In this way, the offender is able to eventually secure a compliant response from the victim.
The concept of fairness is a resource deployed by both the offender and the victim.However, in our data, it is the offender who typically secures compliance rather than the victim.Throughout the data, the tactic of the offender withholding material as a means of securing compliance is prevalent.Tactics used by the offender also include constructing problems with technology, the time associated in sending long videos, and the presence of a parent.Using these resources enables the offender to delay requests by the victim, whilst maintaining the notion of equity in transactions.
The deontic rights of both parties appear to be challenged via such exchanges and orientations to fairness.From the victim's perspective, the offender is claiming a higher deontic status than 'she' should.Elements of resistance are grounded in the mistaken understanding of an equal deontic status and are challenged.The social practice of deals could be equated to what other researchers refer to as being 'caught in a web' (Quayle et al., 2012;Whittle et al., 2014).

Maintaining control and escalation to threats
In the last section, we saw how the victim's resistance begins to challenge the deontic authority of the offender.When the offender pushes the victim to engage in everincreasing, sexualized interactions, the victim must make judgements about their own deontic rights relative to the offender.
Above we see the offender pursuing a previously attempted goal of getting the victim to include one of his friends.This is managed in the form of a directive, which projects compliance with little acknowledgement of contingencies that would thwart such compliance (Craven & Potter, 2010), albeit here, a 'please' is added, which softens the force of the command.The addition of the offer to 'do wat ever u want' also somewhat mitigates the force of the directive with a return to the familiar bargaining.'I swear' is also added here which perhaps orients to the non-receipt of a video from previous deals.The victim gives a dispreferred response (line 4) without offering an excuse for non-compliance, exerting a deontic right.The offender pursues compliance by raising issues of contingency ('I know u can but u don't want'), which the victim aligns himself to (line 6).Issues of agency are raised by the offender in line 7, and followed by a further plea in line 9 ('please').Again, the victim provides a further dispreferred response without any expansion or account displaying a deontic right to resist (line 10).
According to Stevanovic (2011), 'participants need to be aware of their deontic rights relative to one another in order to understand how certain turns at talk are to be interpreted as social actions' (p.5).The first display that all is not right comes with the candidate question raised by the offender ('why u scared'), which is rejected by the victim.This rejection is again met with a candidate reason for not wanting to participate in the suggested activity, at which point, the victim provides a justification of 'I cba4 to get them' to close down the sequence.The next utterance from the offender ('I can force you', line 15) appears to invoke some deontic status that the victim is unaware of.This utterance works as a threat, albeit softened with 'but I don't wanna do it'.The victim's use of a question mark at this point displays that the interaction has somehow gone awry.At this point, the offender does not engage with the questioning of his deontic rights or reveal his epistemic status.Indeed, the revelation of his identity as a male adult remains undisclosed across the data set.Instead of pursuing the threat, the offender resumes his previous strategy of dealmaking which achieves compliance from the victim.
Lines 1-4 above show an orientation to 'unfairness', worked up by the offender as justification for not complying with the victim's request to send his video, which is countered with the victim's rebuttal orienting to a notion of parity ('Ive showed my hole for 1 hour. ..').The directive in line 5 to 'show again' is resisted by the victim with the dispreferred 'no', and a recycling of his conditional request for the video.The victim thereby displays agency and independence in these negotiations.
Once deal-making is an established pattern, it becomes available as a resource for both parties (though, as indicated earlier, more so by the offender) in an attempt to achieve compliance.Resistance from the victim is typically countered by the notion of reciprocal sharing, and also by the offender; here, in the form of a directive ('WE GONNA DO IT BOTH').The use of capital letters stresses the velocity of the command, and arguably displays anger or shouting (lines 8 & 9).Hepburn and Potter (2011) discuss how modified versions of the same directive, with elevated volume, potentially display anger.In lines 9-10, the offender subverts the 'I will. .., if you. ..' formulation to the more forceful 'I WILL SEND WHEN U FUCKING SHOW NAKED', and issues a warning ('do not fucking piss me off').This warning implies a noxious upshot contingent on the victim's response (Hepburn & Potter, 2011); however, the victim resists by reinstating a symmetrical stance, 'I'm already pissed off'.After a further failed directive (line 12), the interaction appears to enter a closing sequence.
The insertion of the assessment 'I'm not losing anything' (line 15), however, re-opens the sequence obliquely through working up the notion of the offender having the 'upper hand'.This is followed by a warning ('you gonna lose and u losing anyways').With these assessments, the offender projects what Heritage and Raymond (2012) have referred to as having more information (a K+ knowledgeable epistemic position) than the victim (who has a KÀ lower knowledge position).The victim's response of 'how' registers their Kepistemic position.
Sikveland, Kevoe-Feldman, and Stokoe ( 2019) consider what kind of communicative practices lay the ground work for turning points in negotiations between suicidal persons and police negotiators.In their data, they found that suicidal persons display independence and agency, and that negotiators can use these as instances of reasoning to align with the suicidal person and secure successful outcomes.In our data, turning points occur when the offender is thwarted by resistance or challenges from victims; arguably victims display agency but this is based on (mis-)judgements of their own deontic status relative to the offender's (Stevanovic, 2011).
With the turn 'I still keeping ur naked pics', the offender works up some leverage against the victim and potentially acts as a warning and/or threat.Indeed, the victim's response of having photographs of the offender (line 21) displays that he treats this as a threat in which he has an equal deontic footing, and the potential to undermine the offender's threat.However, this construction of parity is quickly challenged by the offender, noting that his photographs do not include his face, thereby effectively minimizing any opportunity for them to be used against him.A simple 'okay' displays the victim's acknowledgement that this undermines any equality.
The addition, 'I bet u don't wanna ur parents know about them' similarly works as a threat through a candidate noxious upshot.It is returned with the preferred agreement 'ye' from the victim.The 'if you. .., I will. ..' formulation is reworked as a threat through the conditional grammar structure 'so don't make me piss off and I wont', which places the victim as in control of the unwanted threat action (Hepburn & Potter, 2011).The victim orients to a further potential noxious upshot with his question about the possibility of the offender posting his photographs on Facebook.The victim's resistance has effectively been undermined, resulting in a response of 'thank you', and later 'okay' (line 29).Indeed, in the last line of the transcript we see the offender return to a new sequence, where a new directive is issued and accepted.
One of the features of threats is that the speaker has the power to affect a negative upshot (Hepburn & Potter, 2011).'The fundamental point of the threat is to change the current action of the recipient rather than align with it' (Hepburn & Potter, 2011, p. 10).The analysis above has begun to describe how the offender builds up displays of their previously hidden deontic status through orienting to negative upshots.
A loss of bargaining power Once threats are operationalized, they are invoked frequently in order to secure compliance from the victim.The deontic status of the offender becomes increasingly evident to the victims; however, they still question this in some instances.
Extract 12 occurs after the offender has threatened to share the victim's photographs with others on a network site.Fear is made relevant in the offender's question 'r u scared?' (Edwards, 1998).The victim agrees with the assessment implicit in the question, and it is upgraded with 'I really am'.His pleas for the threats to stop are delivered with a crying emoticon.Hepburn and Potter (2011) urged that future research into threats should examine the sequential placement of emotional displays.In our data, there was evidence to suggest that displays of emotion were frequently employed by the victims as a means of appealing to the offender to desist with threats and were also attended to by the offender (Edwards, 1998).
The offender uses the formulation 'will u show me ur naked if I stop?', which is treated as a false deal, evidenced in the victim's response (line 5).In lines 6-7, the victim makes a conditional dealseeking proof of the 'stuff of me' before 'showing me you'.The offender manages this resistance by working up his contingency through the threat displayed in his assessment ('u can just see them on internet').This threat is met with an upgraded plea from the victim (lines 9 & 10).At this point, the offender returns to offering a further deal, and employing the formulation 'if you. ..I will'.The offender designs his offer in a way that might appease the victim through the deletion of material and not telling others, as well as the offer of friendship.The offender seems to work up the deal in a way that might return things to the status quo.
Despite receiving threats, victims often tried to regain some control by attempting to reinstate the social practice of deals through bargaining.
The extract above occurs after a sequence in which the offender threatened the victim with posting his photographs online, after failed attempts to convince the victim to include his younger sister via webcam.The victim makes two attempts to reinstate a deal his initial attempt is conditional ('maybe if you. ..', line 3).The offender's response is designed to placate the victim (line 4), and adding the victim's name potentially personalizes their 'relationship' (line 4).However, with the discourse marker 'so', the offender then works up a conditional reprieve (lines 4-6).The threat 'if you didnt do it then I will have to post it ok?' effectively removes any agency on the offender's behalf.The victim's second attempt at a deal (line 7) is rejected by the offender.The offender then reverts to the formulation ('if u get her with u and do stuff then I will delete it and leave u alone'), working up pressure via this conditional offer to comply with his request.The victim upgrades his offer with a request in the form of a deal (lines 7-8).This revised offer and promise is problematized by the offender with 'I don't need ur promise cause ur lier', and any return to a deal is clearly dismissed with the offender's 'you listen to what I say not I listen to wat u say' (line 16).Any pretence at an equal deontic status is gone.
Overall, victims could typically be seen to make attempts to revert to the social practice of deal-making after threats have been issued; however, contingencies have shifted and, whilst the offender also reverts to offering deals, he also displays his deontic status and power over the victim through negative upshots.

Discussion
We have shown how, in our data, the social practice of deals, featuring the exchange of sexual images and/or videos, was used as a resource to entrap victims in order to exhibit control over them.In contrast to previous research, we found that boys were sexually solicited and that such interactions were aggressive and forceful.Our key contribution is that by employing a discursive approach we have been able to highlight how the offender escalated his issuing of threats following victims' resistance and non-compliance to requests, and how manipulation, in the form of overt persuasion and extortion, were live issues in naturally occurring data (in contrast to transcripts of chat logs derived from the Perverted Justice website).
In our data, the social practice of deal-making was employed to normalize the practice of sharing images and 'hooking' the victim.We found that threats to publish sexual images of the victim were mobilized through this social practice of deal-making.We illustrated how bargaining sequences were initially reciprocal, and how both parties appeared to treat each other as having an equal deontic status.However, we also demonstrated how the deontic status of both parties was challenged when issues of parity were raised.It is here that we saw some resistance from victims to exchange images or engage in certain sexual acts via webcam.In resisting, victims constructed a disparity in the transactions.This led to the offender working up counter-claims of parity, such as problems with material sent by the victim, and justifying their own disparity.The concept of fairness was therefore a resource used by both the offender and his victims to elicit a 'fair' exchange; however, in our case study, it is the offender (rather than the victim) who typically secured compliance.
It is important to note that this is not a model of rigid stepsrather, deal-making is a flexible resource employed by both parties at various points.Our finding supports the claim that sexualization, which includes the exchange of images, normalizes requests for engagement in and performance of sexual acts, with the potential to lead to blackmail and threats (O'Connell, 2003;Ospina et al., 2010;Whittle et al., 2013).However, our discursive approach allowed us to explore how threats to publish sexual images of the victim are oriented to by the victims in situ.Our study also adds support to the claim that offenders are more directive, persistent, non-compromising and pressurizing towards real victims in order to induce victims' compliance as part of sexually exploitative interactions, which contrasts with findings that are based on decoys (Chiang & Grant, 2018;Kloess et al., 2015;Schneevogt et al., 2018).
In instances where threats were escalated by the offender, the deontic rights of both parties were challenged.For example, we found that threats were not always clearutterances made by the offender in response to the victim's resistance obliquely referenced a threat (e.g., 'I can force you'), followed by the victim displaying incomprehension (e.g., '?').The victims are unaware of the offender's epistemic status and are therefore lower on the epistemic gradient (Heritage & Raymond, 2012).Once such threats were invoked, they became operationalized more frequently, and compliance was secured through the threat of negative upshots.As threats escalate, the deontic status of the offender becomes more obvious (e.g., 'you listen to what I say not I listen to wat u say').
Our study also provides support for, and extends, the groundwork conducted by Hepburn and Potter (2011) on the explication of the basic elements that constitute threats.In Hepburn and Potter's (2011) study, the noxious upshot was commonly withholding something enjoyable, whereas in our data, this tended to represent a negative aspect.Additionally, agency for the negative upshot in our data is routinely claimed by the offender, rather than being worked up to soften the 'appearance of autocratic control', as evidenced in their analysis of family mealtimes (p.13).
We found that sexualized interactions developed very quickly, and noted limited evidence of relationship-forming.The sexualized nature evidenced in our data contrasts with other research that has argued that interactions with boys are less sexually explicit, aggressive and coercive (Grosskopf, 2010).In fact, in our data, the interactions were sexualized, aggressive and forceful.This may, in part, be due to the difference between how decoys and actual children interact, particularly in relation to the sharing of images (which is prohibited with decoys).
To our knowledge, this is the first analysis of online interactions between an offender (pretending to be a teenage girl) and five male victims.Based upon our data, there appear to be times when the victims missed early signs of deception, such as the request from the offender in Extract 3 to see the victim's sister.Yet, we see no evidence of the victim treating such requests as strange, other than his refusal to comply.Turning points that appeared odd in the online interactions suggest that they may be used to encourage children to be more reflective about any further engagement.The data set merely revealed one instance in which this online identity was questioned.Stevanovic (2011) argues that 'it is entirely up to the recipient to sort out the deontic implications the speaker's utterance have on the recipient's own future actions' (p.4).Drawing upon deontic authority (Stevanovic, 2011;Stevanovic & Per€ akyl€ a, 2012), we noted how this was very apt in relation to the victims in our data.Researchers have noted that gendered discourses have positioned 'girls as potentially sexy and boys as inevitably wanting' (Rasmussen & Sɵndergaard, 2020), and this may have played some part in the interactions.Young people may transgress the normative boundaries whilst sharing images online, treating them as both harmless and violating at the same time (Rasmussen & Sɵndergaard, 2020).
Our findings highlight that offenders may escalate their threats following resistance and non-compliance by victims, demonstrating the extent of manipulative strategies offenders will employ in order to achieve their goal.Many of the children and young people who are victimized may present with a range of vulnerability factors that place them at greater risk of experiencing sexual exploitation and abuse (Whittle et al., 2014).In a study by Kloess et al. (2017), most victims appeared to be motivated by curiosity and sexual exploration/experimentation, and whilst they did engage in some risk-taking behaviours, they were equally aware of personal boundaries and assertively refused to comply with more extreme requests by offenders, resulting in successful rebuttals thereof.It is important to recognize that sexually exploitative and abusive interactions online represent dynamics that are unique to the parties involved therein.Offenders may either adapt their manipulative strategies, or use them relatively consistently, within their interactions with one particular victim, or across a number of different victims.Some victims have been found to successfully depart and remove themselves from online interactions with offenders, whilst others suffered serious offences of sexual abuse (Kloess et al., 2017).
In terms of prevention, a number of school-based sexual abuse prevention programmes (SSAPPs) have been developed in order to raise children's and young people's awareness and understanding of child sexual exploitation and abuse.According to Whittle et al. (2014), young people in their study reported engaging in safer online behaviours and Internet use post-abusive experiences, finding safety education helpful and necessary.This was further supported by May, Kloess, Davies, and Hamilton-Giachritsis (2020); however, the authors argue that prevention should not be made the sole responsibility of children and young people.The move to situate preventative measures and strategies within SSAPPs has been criticized by some for the inherent assumption that children and young people have the ability and power to avoid being exploited and abused (Williams, 2019).It goes without saying that the responsibility for the sexual exploitation and abuse of children always lies with the offender.
Whilst the number of transcripts the case comprised of was appropriate for the methodology used, our study is limited in focus on the practice of one offender.Future research would benefit from a larger data set that examines whether the social practice of deal-making operates in a similar way across offender and victim genders.Finally, the newly identified modes of manipulation, such as not responding and the use of kisses, require further investigation.Overall, in line with Chiang andGrant (2018) andSchneevogt et al. (2018), our study supports their claim that overt persuasion and extortion are present in real world, naturally occurring interactions between offenders and their victims, which is in contrast to the type of manipulation evidenced in analyses of Perverted Justice data.As such, we agree with their concern about the validity of treating such data as an imitation of what happens in real-world interactions of this nature.