The (de)legitimation of torture: rhetoric, shaming and narrative contestation in two British cases

Existing studies on democracies’ involvement in torture emphasise how governments have been able to circumvent the international anti-torture norm and shape public discourse on the issue through powerful rhetorical strategies of denial and exception. Less attention has been paid, however, to the rhetoric of opponents of torture and how it impacts on governments and security agencies. This article proposes a typology of four common arguments against torture, which make use variously of ethical, utilitarian and ‘shaming’ rhetoric. These arguments often take a narrative form and are extensively contested by governments. Drawing on the literature on rhetorical coercion, I argue that anti-torture narratives can play an important role in constraining democratic states and significantly reducing their perpetration of torture. Yet the multiplicity of narratives at play opens up opportunities for governments to accept some messages against torture while simultaneously contesting others in a way which enables them to continue their involvement in torture. I develop this argument through a comparative analysis of the role of torture in two British counterterrorism campaigns – against Irish republican terrorism in the 1970s and against jihadist violence after 9/11. Differences in the content and salience of the narratives advanced by critics of the government during the two time periods explain much about why the British government contested some arguments against torture, but accepted others. This variation helps to explain in turn why British security agencies carried out coercive interrogations on a wide scale during the 1970s, while their perpetration of torture was significantly reduced in the post-9/11 case.

Torture is a site of intense contestation at the intersection of human rights and security practice. To torture is to engage in one of the most egregious violations of human dignity. Yet torture has also long been represented as a useful means of gaining not only intelligence but even 'truth' (Wisnewski, 2010). In the face of these contradictions, democratic governments do not respond consistently. This article seeks to shed light on their varied approaches to the question of torture through a comparative analysis of two British cases. Facing Irish terrorism during the 1970s, the UK's security agencies carried out coercive interrogations on a wide scale. In their campaign against jihadist violence after 9/11, by comparison, there was a significant reduction in the British agencies' perpetration of coercive interrogation, although they did collude in torture by foreign intelligence services. Why does the extent of British involvement in torture vary between the two campaigns?
In addressing this question, I develop a theoretical analysis of anti-torture narratives and government efforts to contest them -a rhetorical struggle referred to here as the '(de) legitimation' of torture. The article introduces insights from the literatures on rhetoric and 'shaming' in International Relations (IR) to understand the dynamics of torture discourse. A central focus is on the role of 'rhetorical coercion' in this area; in other words, the strategic use of language for the purpose of denying an opponent sufficient 'rhetorical materials. . . to craft a socially sustainable rebuttal' (Krebs and Jackson, 2007: 42). One type of rhetorical coercion is 'shaming' which involves condemning a target actor for violating its own international commitments with a view to motivating a change in the target's behaviour (Friman, 2015;Schimmelfennig, 2001). This article proposes a typology of four common anti-torture narratives, which make use variously of ethical, utilitarian and shaming rhetoric. It also analyses the effects of each narrative on rhetorical struggles over coercive interrogation. I argue that under certain conditions, anti-torture narratives can play an important role in constraining democratic states and significantly reducing their perpetration of torture. Yet the multiplicity of narratives at play opens opportunities for target governments to engage in strategic rhetorical manoeuvres -to accept some messages against torture while simultaneously contesting others in a way which enables them to continue their involvement in torture.
Applying this framework to the British cases, the article examines the efforts of various actors to shame and rhetorically coerce the UK government for its involvement in torture during the 1970s and 2000s. It outlines how these actors deployed four key antitorture narratives in the two time periods and accounts for why their efforts had little impact in the 1970s but proved more influential in the post-9/11 case. The authority of the accuser and the clarity of the standard of legitimacy invoked both contributed to this variation. But, more importantly, it was differences in the content and salience of the narratives advanced by critics of torture, which led to a different configuration of prevailing narratives within the government in the two cases. In the second (post 9/11) case, both 'shaming' messages and narratives depicting the strategic ineffectiveness of torture proved difficult for British officials to dismiss and indeed were embraced by many within the government in a way that imposed significant constraints on interrogation practices. Yet officials proved more resistant to two other narratives, which focused on the ethical impermissibility and tactical ineffectiveness of torture. Attitudes within the government on these fundamental issues essentially did not change and, in this context, officials reacted strategically to public narratives against torture by covertly colluding with foreign services' torture programmes after 2001. An account of the particular configurations of narratives prevalent during the two time periods is key for understanding why British security agencies carried out coercive interrogations on a wide scale during the 1970s, while the nature and degree of their involvement in torture was significantly different in the post-9/11 case.
This article speaks to an ongoing scholarly debate on the extent to which democratic states face constraints on their involvement in torture. A central strand of this literature holds that the international human rights regime often has only superficial effects on democracies, for example, prompting them to make greater use of 'stealth torture', which causes horrific pain but leaves few physical marks on its victims (Rejali, 2007). Governments have also been able to circumvent the international antitorture norm and shape public discourse on the issue through powerful rhetorical strategies of denial, exception or 'plausible legality' (Blakeley and Raphael, 2017;Neal, 2010;Sanders, 2011). A second strand of the literature emphasises, by contrast, the constraints on democracies' involvement in torture. It has shown how non-governmental organisations (NGOs) and human rights lawyers worked through the structure of domestic and international law to stymie America's post-9/11 torture programme (Blakeley, 2013). Moreover, Del Rosso (2015) demonstrates how certain types of evidence, and the timing of their release, led mainstream US politics to make a transition from denial to acknowledgement and criticism of the government's perpetration of torture. Keating (2014) shows how the United States' defection from the international prohibition against torture after 9/11 failed to secure the support of international society.
This article makes two contributions to the academic debate on torture. First, as outlined above, it develops a theoretical analysis of the rhetorical constraints on democracies' involvement in torture, which is currently not present in the literature. While scholars have shown great interest in governments' rhetorical strategies, they have paid less attention to the rhetoric of opponents of torture. Second, this article offers the first qualitative analysis of the effects of shaming on a state's degree of involvement in torture, thus addressing an important lacuna in the literatures on torture and shaming. Research into the effects of 'naming and shaming' on torture has been dominated by quantitative analyses -but these studies have produced ambiguous results (Conrad and DeMeritt, 2014: 176;Conrad and Moore, 2010: 468;Hafner-Burton, 2008: 703). A further limitation of these large-N analyses is that they do not capture well how shaming interacts with other types of rhetorical coercion. This study sheds light on this interaction, showing how shaming can be buttressed by other narratives -for example, about effectiveness -and thus have a greater influence on government officials than would be likely through shaming alone.
It is also worth noting that much of the empirical literature on torture has focused on single case studies, often of post-9/11 America. 1 This article further expands our knowledge base beyond the US case through a historical comparative analysis of two British cases. The comparative method also offers additional analytical leverage, enabling one to better identify the precise conditions that led to different outcomes across two or more cases (Lightcap, 2011: 32-33).
In its empirical analysis, this article concentrates on the agencies with primary responsibility for the investigation and repression of terrorism against the United Kingdom. In the 1970s, these were the British Army (which had the lead role in Northern Ireland operations between 1971 and 1976) and the police (which assumed primacy in Northern Ireland in 1976-1977 and had the lead role in Great Britain throughout). In the post-9/11 fight against jihadist terrorism, the key actors were the counterterrorist police and the UK's intelligence agencies. 2 The main concentration, therefore, is on how British security agencies have treated those suspected of posing a direct threat to the national territory of the UK, 3 from Irish republicans in the 1970s to British and foreign jihadists in the 2001-2010 period, including those who were detained by foreign intelligence agencies. I follow the United Nations (UN) definition of torture as 'any act by which severe pain or suffering, whether physical or mental, is intentionally inflicted on a person for such purposes as obtaining from him. . . information or a confession, punishing him, or intimidating or coercing him. . .' 4 The terms 'torture' and 'coercive interrogation' are used interchangeably.
The article proceeds in three steps. The first section establishes that there was variation in the degree to which the UK was involved in torture in the two cases under study. The next section sets out the theoretical expectations which guide my analysis of rhetorical coercion, shaming and narrative contestation. The third section begins by proposing a typology of four key anti-torture narratives. Having outlined the methods underpinning my empirical analysis, this section traces how opponents of torture deployed these narratives during the 1970s and 2000s in an effort to shame and rhetorically coerce the UK government. I outline the government's response to these narratives and show why they had varying effects on the rhetoric and behaviour of British officials over time. The conclusion lays out the implications of my findings for scholarship on shaming, rhetorical coercion and international norms.

Different degrees of involvement in torture: UK 1970s vs UK post-9/11
The nature and extent of British involvement in torture differed significantly in the country's two most important counterterrorism campaigns of modern times -against the Irish Republican Army (IRA) after 1969 and jihadist terrorism after 9/11. This section compares these two cases and evaluates some possible explanations for the variation observed.

Interrogation and torture in 1970s Britain
The British Army and the Royal Ulster Constabulary (RUC) police force used coercive interrogation against many detainees in Northern Ireland during the 1970s, while some police forces in mainland Great Britain were also accused of abusing terrorist suspects in their custody. In 1971, the British Army introduced five interrogation techniques for use in Northern Ireland, which it had developed in previous colonial counter-insurgency campaigns. These were: (a) hooding, (b) standing for hours at a wall with legs apart and arms high, (c) subjection to noise and sensory isolation, (d) a diet of bread and water and (e) sleep deprivation. Fourteen detainees were selected for exposure to the five techniques in 1971 (European Court of Human Rights, 1978: para. 96). It was usually police officers who carried out these coercive interrogations at centres operated jointly by the army and the RUC (ECtHR, 1978: paras. 92-115). A report by Amnesty International (1972: 41) concluded what the British government itself later stated in private -that the use of these techniques against detainees had amounted to torture (Carver, 2016: 110).
Although the government declared in 1972 that the five techniques had been discontinued, coercive interrogations and abuse of detainees continued in practice. Between 1971 and 1974 alone, approximately 2000 complaints of ill-treatment or assault were made against the police and the army in Northern Ireland. In total, 218 members of the security forces were prosecuted for assault between 1972 and 1977 and 155 were convicted (ECtHR, 1978: para. 140). The report of a detailed investigation by Amnesty International (1978) concluded that RUC mistreatment of suspects took place on a frequent basis. A public inquiry followed, which substantiated the allegations of mistreatment in a number of cases and recommended reforms (Bennett, 1979). A number of former RUC officers later admitted that coercive interrogation by the police had been widespread in Northern Ireland during the 1970s. They outlined the techniques used, which included beatings, punches to the centre of the stomach and forcing detainees' wrists or elbows into painful positions, sometimes for long periods of time (Cobain, 2012: 182-183).
In Great Britain, meanwhile, there was substantial evidence that some English police forces coerced confessions out of terrorist suspects during the 1970s, particularly in notorious miscarriages of justice cases. Some of the Birmingham Six -wrongly convicted and later acquitted of carrying out deadly bombings in 1974 -alleged that they had been beaten while in custody. Doctors verified that the men had sustained extensive injuries, some of which were visible when they appeared in court. The judge at their trial also acknowledged that they had been 'quite outrageously assaulted' while in custody (Faul and Murray, 1976). Other victims of miscarriages of justice -the Guildford Four -also made credible allegations that members of an English police force had beaten and forced false confessions out of them (Ewing, 2006: 47-54). While not on the same scale as the abuse in Northern Ireland, such cases indicated that some police forces in Great Britain also carried out coercive interrogation of terrorist suspects during the 1970s.
Britain's record on torture in the post-9/11 era In contrast to the 1970s, by all accounts there have been no coercive interrogations of detainees within the UK in the period since 9/11. Jihadist terrorist suspects have not faced the intimidating and violent cells that many Irish detainees endured in the 1970s. The 'vast majority' of terrorist suspects arrested in Britain during the 2000s 'maintain[ed] their right to silence' according to Peter Clarke, a former head of counterterrorism at the Metropolitan Police. 'The walls of Paddington Green police station rarely reverberate with outpourings of guilt and contrition', he wrote (Clarke, 2008). However, there was no question of the police resorting to coercive interrogation in order to get suspects to talk.
Beyond the domestic sphere, Britain did become involved in torture after 9/11 through its collusion with the prisoner abuse of many foreign intelligence services, including those of the United States, Pakistan, Libya and Egypt. A parliamentary committee inquiry into the period 2001-2010 found evidence that in 232 cases the UK's intelligence agencies continued to supply questions or intelligence to foreign services after they knew or suspected that a detainee was being mistreated. In 13 cases, British officials witnessed a prisoner being mistreated by others (ISC (Intelligence and Security Committee), 2018: 2-3). Indeed, UK officers allegedly interviewed some detainees after -or in betweenperiods of torture by a foreign security service. Britain was also complicit in other states' rendition operations -the transfer of prisoners, outside of any legal process, to countries where the risk of torture was high. The parliamentary inquiry found evidence that in 28 cases UK intelligence 'suggested, planned or agreed to rendition operations proposed by others'. In a further 22 cases, the UK 'provided intelligence to enable a rendition operation to take place' (ISC, 2018: 3). In 2004, for example, Britain's foreign intelligence service, MI6, provided intelligence which enabled US agents to abduct at least two Libyan dissidents and fly them into the hands of the Gaddafi regime, which subsequently tortured them. A senior MI6 official, Mark Allen, wrote to his Libyan counterpart to 'congratulate' him on one of these operations and request access to the resulting intelligence (quoted in Blakeley and Raphael, 2017: 255). Overall, as the High Court later put it, the UK had 'facilitated' abusive interrogations by foreign security services through providing questions, intelligence and other cooperation (UK High Court, 2009: para. 88). Britain colluded in these coercive interrogations with a view to gaining the intelligence produced by them.

Alternative explanations
While this article emphasises the role of anti-torture narratives, Blakeley and Raphael's extensive analysis of UK involvement in torture after 9/11 offers an alternative explanation, which depicts Britain as a 'petty sovereign operating. . . under the aegis of US exceptionalism'. There was 'no need' for the UK to carry out torture itself because the United States had established an extensive torture infrastructure across the world (Blakeley and Raphael, 2017: 243, 251, 261). Britain could simply outsource torture to others and receive intelligence from them. This, it may be argued, reflects the nature of contemporary international terrorism. One could also extend this argument to conclude that, in contrast, when facing the Irish terrorism of the 1970s, Britain carried out torture because it could not outsource its dirty work on this domestic threat to any other country.
While this argument tells an important part of the story, it does not provide a full understanding of the British case because contemporary jihadist terrorism which targets the UK is as much a domestic phenomenon as it is international. Attack planning has been carried out mostly by British nationals based in the UK (Foley, 2013: 29-31). US or other countries' intelligence is very useful against such threats -but it is far from sufficient. The need for the British government to generate its own intelligence on suspects' activities within the UK far outweighs its need for information from foreign partners. During the 1970s, it used torture as a means of gaining such intelligence. In the 2000s, it did not. On 21 July 2005 for example, soon after the 7/7 London bombings, a second plot came very close to causing explosions and loss of life in London once again. The five men responsible escaped and a desperate manhunt for them ensued. When the first of these men were captured, British police officers did not use coercive interrogation to find out the locations of their accomplices. A focus on outsourcing does not help us to understand why.
A second explanation would focus on specific changes to domestic law. Undoubtedly, this did play a role in the evolution of the UK's approach to interrogation. The Police and Criminal Evidence Act 1984 (PACE) introduced the audio tape recording of detainee interviews and strengthened suspects' right to have a lawyer present with them in police stations. Over time, PACE created a context in which English lawyers were able to defend detainees' rights effectively in the initial detention period after arrest and provided practical protections against coercive interrogation. (Carver, 2016: 118-119). Notwithstanding its importance, domestic legislation can be changed and it does not offer an adequate explanation of the UK's record on interrogation after 11 September 2001. After all, the government pushed through legislation in 2000 and again after 9/11 that departed from ordinary criminal justice procedures in some respects, such as extending the length of maximum pre-charge detention in terrorism cases (Foley, 2013: 183). There was no legal impediment to the government extending this special regime so as to also deprive terrorist suspects of the PACE protections covering the interrogation process. Indeed, countries such as Spain do precisely this in terrorism cases. I argue below that to understand why the UK did not take this step, we need to consider the influence of certain anti-torture narratives on the British government during the 2000s. Let us now turn to consider the theoretical expectations which underpin this argument.

From rhetorical coercion to narrative contestation
Legitimation -understood here as 'the public justification of policy' -plays a central role in world politics. There is a growing body of IR scholarship, which indicates that the public rhetoric of legitimation has significant and independent effects on how governments formulate and execute foreign policy (Goddard and Krebs, 2015: 5). In many cases, if an action cannot be justified to relevant audiences, it may not be taken. In political struggles, rival actors put considerable effort into justifying their positions before domestic or international audiences. As Krebs and Jackson have shown, such actors often engage in rhetorical coercion: they seek to use language strategically with a view to denying their opponent sufficient 'rhetorical materials. . . to craft a socially sustainable rebuttal'. If one side 'wins', it is because 'the audience deems certain rhetorical deployments acceptable and others impermissible'. As part of this kind of rhetorical coercion, one party may seek to manoeuvre another onto 'more favourable rhetorical terrain' (Krebs and Jackson, 2007: 42, 44, 47). As we shall see below, opponents of torture do not confine themselves to talking only about the short-term tactical effects of torture but rather seek to shift governments into publicly considering other effects of torture.

Narrative contestation
While rhetorical coercion can have significant effects, it is often deeply contested as target states make extensive efforts to resist such attempts to influence them (Friman, 2015). 5 In this process, accusers' messages and state counter-messages often take a narrative form. Scholars in the field of IR are increasingly bringing this concept into their research on a wide range of topics (Krebs, 2015). In essence, a narrative is a framework that allows humans to connect various phenomena around some kind of causal chain. Causality and causal claims are central to any narrative. The end point of the causal chain -the claimed outcome -is crucial for giving meaning to the previous links in the chain (Miskimmon et al., 2013: 5). A desirable outcome casts a positive light on the decisions, actions and events that are said to have led to it. A negative outcome can delegitimate many if not all of the actions that preceded it. Narratives are also 'necessarily selective', as Krebs (2015: 11) puts it. To sustain its thread, a narrative must highlight some elements and leave others out of the story. This opens an opportunity for actors who seek to deploy narratives to achieve a political effect. The state and its opponents thus deploy rival narratives in an attempt to sway relevant audiences. This narrative contestation involves contrasting interpretations of what the relevant setting is, which actors should be included in the story, how their purposes should be characterised, how their actions and means should be depicted and what the causal effects of their actions are (Krebs, 2015: 12).
What conditions shape the outcome of narrative contestation? Which factors have the most important influence on whether rhetorical coercion is effective in producing the desired outcome? (i.e. a situation where the target refrains from or significantly reduces its involvement in the 'illegitimate' behaviour concerned). Two factors, emphasised in the literature, are the authority and credibility of the accuser(s) and the relative clarity of the standard of legitimacy used by them. I outline these factors below and show how they are illustrated well by the two British cases examined here. The IR literature on rhetorical effectiveness also emphasises that we need to pay attention to the content of the rhetoric being transmitted in order to understand the outcomes of political struggles (Goddard and Krebs, 2015: 29). When analysing rhetorical coercion on the issue of torture, it is useful to pay attention to the multiplicity of narrative content.

Anti-torture narratives: ethical, utilitarian and shaming rhetoric
Opponents of torture in democracies tend to rely a great deal on four types of narratives. Though they have varying effects, all four are considered here as potential instances of rhetorical coercion. First, ethical claims have been central to the case against torture. The argument that torture is morally wrong is often developed through narratives which focus on the pain and suffering caused by coercive interrogation (Del Rosso, 2015: 140-142;Wisnewski, 2010).
A second prominent type of anti-torture narrative has been one which uses utilitarian rhetoric to depict coercive interrogation as a tactically ineffective method of gaining intelligence. Detainees may say whatever they believe will make the torture stop, including falsehoods, and in any case standard interrogation techniques offer a more effective way of gaining accurate information (Del Rosso, 2015: 138-144;Wisnewski, 2010: 153-157). The third type of narrative also takes a utilitarian approach, this time to depict torture as strategically ineffective and counterproductive. It highlights the broader effects of torture on a counterterrorism campaign, suggesting that if the state is known to torture, that will alienate key sections of the population, increase support for the terrorist group and help it to recruit more people into violence (Del Rosso, 2015: 145-148;Ralph, 2013: 118-130). Claims to effectiveness form an important part of governments' efforts to legitimate their policies. Consequently, narratives which cast doubt on effectiveness need to be engaged with and are an important contestation ground for rhetorical struggles on countless issues, including torture.
The fourth type is a narrative on how torture contradicts states' international commitments and damages their reputation. By drawing attention to violations of international legal commitments, this narrative is related to the practice of shaming, which is the subject of an extensive literature covering torture and many other issue areas. Shaming is commonly understood as the public exposure and condemnation of a target state for illegitimate actions which violate its own international commitments, done with a view to damaging its reputation and thus motivating a change in behaviour (Friman, 2015: 143-144;Schimmelfennig, 2001: 64). Scholarship on the politics of shame relies on a number of premises. On the one hand, it is assumed that international norms or standards of legitimacy have some influence on states, which have publicly committed themselves to upholding these standards. As members of regional and international communities, political actors care about their reputation and the perceived legitimacy of their behaviour. On the other hand, however, actors are only 'weakly socialised' into the community. International norms are not necessarily internalised by actors; norms do not directly motivate their behaviour and state officials may regard such norms as mere external constraints (Schimmelfennig, 2001: 48, 58, 62-63, 65). This is where shaming comes in. It is through being shamed that the targeted state may be brought to focus on, and perhaps come closer to fulfilling, its own professed standards of conduct. The organisations or governments, which drive this shaming effort, direct their messages to several audiences. They focus on those who are responsible for the illegitimate behaviour, such as political leaders and senior government and security officials of the target state. At the same time, they also direct their messages to other elements of the political class, media or civil society in the targeted country, along with key constituencies in allied countries and broader international audiences (Vennesson, 2014: 31). The resulting process of shaming can work in at least two ways. Some of the state officials directly responsible for the objectionable deeds may feel ashamed following this public exposure and condemnation, leading them to push for changes in their government's behaviour. Yet feelings of shame are not necessary for shaming to work. Even officials who see little or nothing wrong in their government's actions may still be concerned about the negative effect of this public shaming on their reputation among the various domestic and international audiences where the message is resonating (Schimmelfennig, 2001: 64). That, by itself, can be enough to stimulate changes in behaviour.
As noted in the introduction, this article provides the first qualitative analysis of the effects of shaming on a state's degree of involvement in torture. The analysis, offered below, indicates that the literature needs to pay more attention to the interaction between shaming and other types of rhetorical coercion. Busby and Greenhill (2015: 106) examine how shaming can be combined with material sanctions to influence target states. They suggest that 'shaming by itself is. . . a weak tool'. If this is true, it is equally important to analyse how shaming can be combined with other kinds of rhetoric to have an impact on target governments.

Dynamic responses to rhetorical coercion
One contribution of this article is to develop a theoretical analysis of the rhetorical constraints on democracies' involvement in torture. To understand these constraints fully, we need to focus attention on the target state. After all, research on shaming and socialisation has been criticised for treating these phenomena as a one-way process in which the accuser impacts on the target state (Adler-Nissen, 2014: 150-151). Building on this critique, this article speaks to the literature which analyses shaming effectiveness primarily from the perspective of the target (Adler-Nissen, 2014;Vennesson, 2014). This focus gives rise to two more specific points on the dynamic, 'socio-strategic' response (Mantilla, 2018: 319) of the target state to rhetorical coercion and constraint. First, I argue that greater attention needs to be paid to the multiplicity of narrative content and the resultant complexity of contestation. The previous section illustrated how accusers propagate multiple narratives simultaneously. However, this opens opportunities for rhetorical manoeuvre by the target government -to accept some critical narratives, while simultaneously contesting others in an attempt to circumvent the full implications of the overall norm at stake. Many different narrative combinations could result from this multifaceted contest. However, as some critical or shaming messages fail and others prove effective, the result is a specific configuration of prevailing narratives, which enable and constrain target governments in particular ways. The specific configuration of narratives accepted by the British government during the 2000s, for example, is crucial for understanding the particular way in which it sought to both avoid torture and collude in torture.
The second point is to show how target states react dynamically to shaming and rhetorical coercion -even when they do not resist them. Petrova (2016) analyses NGOs' positive altercasting of a target state, showing how a government may in part be praised into meeting NGO demands. Her analysis indicates how social influence rarely works through pure shaming or disapproval alone. The analysis developed in this article explores another way in which a 'positive' narrative can be developed to legitimate government action. It shows how a target government can creatively recontextualise the critical or shaming narratives of its accusers. Here, the target accepts some of its opponents' criticism -but recasts these points in the context of a positive overall narrative for itself. 6 While this manoeuvre allows the target government to 'look good' to relevant audiences, it may simultaneously reinforce its own rhetorical entrapment and constrain its subsequent behaviour, as will be illustrated below.

Four narratives and their effects in the UK: 1970s vs post-9/11
As noted above, the literature on torture has focused largely on governments' rhetorical strategies. While valuable, this research focus nevertheless gives the impression that the government had the rhetorical field almost to itself. The following sections show, by contrast, that there was deep narrative contestation between the UK government and opponents of torture, which had a significant impact on the rhetoric and behaviour of British officials.

Typology and methods
The preceding discussion drew on secondary literature to identify four key anti-torture narratives. In the empirical research for this article, I examined the rhetoric of opponents of British torture, observing the presence or absence of these four narratives across a range of sources from the 1970s and 2000s, including reports by NGOs and other activists, court proceedings and public inquiries. Thus, a theoretically informed but empirically grounded typology of anti-torture narratives was constructed. Though they usually take a narrative form in public discourse, the four parts of the typology can be stated briefly as follows (with each one then classified according to the type of rhetoric used): The following sections show how these four narratives have been deployed by various actors for the purposes of influencing the British government's position on torture. I analyse the effects of this rhetorical coercion on government discourse, tracing the varying salience of the four anti-torture messages in UK government narratives across the two cases. The result was a different configuration of prevailing narratives in the two time periods, which I argue is key for understanding the British government's varying degrees of involvement in torture in the 1970s and after 9/11. The methods used for this analysis were twofold. First, a qualitative analysis of government documents was carried out. Relevant sources were selected according to three criteria that they contain 'clear articulations' of the position of the British government, have the 'formal authority' to define this position and are 'widely read and attended to' (Hansen, 2006: 76). I identified patterns in the ways in which involvement in torture was legitimated and delegitimated in government sources, including speeches and interventions in Parliament, strategy documents, official reviews of operations and internal communications. 7 On this basis, I develop an account of government narratives (I do not seek to demonstrate here that these are broader elite or societal narratives). While this first method was a qualitative analysis of heterogeneous documents, the second method used was a content analysis of a homogeneous data source. A research assistant coded speeches and interventions in Parliament by government ministers during 10 key debates on torture from the 1970s and 2000s in order to observe how many times the 4 key narratives appeared in each time period. This resulted in 52 discursive observations, allowing us to compare the salience of the narratives across the 2 cases (see online supplemental file for further detail on both methods).

1970s: narrative contestation
As allegations of torture by British security forces in Northern Ireland emerged during 1971, those who expressed condemnation or criticism included not only civil rights and Irish nationalist activists in Northern Ireland but also the Irish government in Dublin, elements of the British media, some opposition Members of the UK Parliament and Amnesty International (Newbery, 2015: 86-90). The British government established a three-person committee to conduct a review of UK interrogation policy -the Parker Inquiry. The committee's majority report, authored by Lord Parker and John Boyd Carpenter, recommended permitting the security forces to use their 'five techniques' of coercive interrogation. However, the third member of the committee -Lord Gardinercame out strongly against the five techniques. He and the other critics mentioned above deployed four main narratives in their efforts to shame or influence the UK government to move away from coercive interrogation.
Torture is morally wrong. Lord Gardiner's minority report in part focused attention on the interrogation room, making it a key setting in his narrative. Referring to medical research, he sought to show how the five techniques caused suffering and severe psychological damage in detainees. He concluded that the techniques were 'illegal' and 'not morally justifiable' (Gardiner, 1972: 11-18, 20-22). This ethics narrative does not appear to have elicited any moral qualms inside the British government. In our content analysis, we did not find any unequivocal statements from ministers that torture was morally unacceptable in the parliamentary debates on torture analysed from the 1970s (see online supplemental file). Instead, they sought to muddy the ethical waters. As the Home Secretary Reginald Maudling told Parliament: 'torture is not acceptable, but merely asking [detainees] if they would be good enough to help in the investigation is equally not acceptable' (Hansard, 17 Nov 71, column 444). Officials thus avoided the label of torture and instead referred to their techniques as 'interrogation in depth'. The government's central narrative was that the five techniques had produced intelligence on the IRA which had saved the 'lives of innocent citizens', as Prime Minister Edward Heath put it (Hansard, 02 Mar 72, column 743). The majority report of the Parker Inquiry (1972: 7-9) similarly defended the five techniques, concluding that there was 'no reason' to rule them out 'on moral grounds'.
Torture is tactically ineffective. Lord Gardiner's minority report argued that coercive interrogation was not an effective tactic for gaining intelligence. His narrative referred back to the Second World War, outlining how Britain had used standard interrogation techniques, which had elicited a great deal of 'vital information' from German prisoners. Reviewing the information gained from the five techniques, he believed that 'substantially as much information' could have been gained through 'our well tried and effective war-time methods' (Gardiner, 1972: 18-19). British government officials constructed a very different historical narrative, emphasising how the 5 techniques had been developed and perfected over a 20-year period in various colonial counter-insurgency campaigns from Malaya to Aden (Newbery, 2015: 14-17;Parker, 1972: 3). Ministers repeatedly stated that coercive interrogation gained vital intelligence (twice per debate on average during the parliamentary debates on torture analysed from the 1970s). As Reginald Maudling put it, 'the interrogation yielded information of great value that would not otherwise have been available' (Hansard, 17 Nov 71, column 443). His ministerial colleague, Lord Balniel, drove the point home. 'Polite and leisurely questioning of suspected terrorists will not provide this information', he said (Hansard, 17 Nov 71, column 496).
Torture is strategically ineffective. While he had devoted considerable attention to the tactical inefficiency of torture, Lord Gardiner devoted only one (somewhat muted) paragraph of his report to its counterproductive strategic effects, such as alienating populations and increasing support for terrorist groups (Gardiner, 1972: 19). Furthermore, a review of key statements and reports by human rights NGOs, such as Amnesty International, and other activists indicates that negative strategic effects were not a central argument of the opponents of British torture during the 1970s. 8 Unsurprisingly, we found that no minister made this argument either in the parliamentary debates on torture analysed from that decade. It is significant that the low salience of the 'strategy' anti-torture narrative during the 1970s is associated with the continuation of coercive interrogation in the UK throughout that decade. As we shall see, when the narrative of strategic ineffectiveness became more salient later on, it proved influential in helping to shift the government's position on torture.

Torture contradicts your international commitments and damages your reputation.
Shaming was a key feature of Lord Gardiner's rhetoric. He constructed a narrative around Britain's contradiction of its international commitments in respect of torture and the effects of this on its international reputation. He wrote: 'If. . . we now depart from world standards which we have helped to create, I believe that we should both gravely damage our own reputation and deal a severe blow to the whole world movement to improve Human Rights'. The Republic of Ireland also sought to shame Britain for violating its international commitments. In December 1971, the Irish government initiated Ireland v UK before the European Commission of Human Rights in Strasbourg alleging that Britain had breached the European Convention on Human Rights (ECHR), notably Article 3 which prohibits torture and inhuman treatment. Dublin ultimately submitted written evidence in respect of 228 cases of alleged abuse between 1971and 1974(ECtHR, 1978. In 1976, the Strasbourg Commission ruled that Britain had breached the ECHR and that the interrogation techniques used in Northern Ireland were torture. The British government had strenuously denied throughout that it had breached the ECHR both in public and before the Commission. It appealed the judgment to the European Court of Human Rights, which ruled in 1978 that the techniques did not amount to torture. However, it confirmed that Britain had violated the Convention since the techniques did constitute 'inhuman or degrading treatment', prohibited under Article 3. The government was to some extent shamed by Lord Gardiner. In mid-1972, it accepted the conclusion of his minority report (rather than the majority report of the Parker Inquiry) and issued a directive which prohibited the future use of the five techniques in Britain. The ban on the five techniques was also issued under the shadow of Ireland v UK. There is evidence which suggests that London's decision may have been influenced in part by the hope that it would persuade the Irish government to drop its case (Newbery, 2015: 103).
Overall, however, shaming had a limited impact on the British government during the 1970s. While both Gardiner and Dublin sought to shame Britain for contradicting its international commitments, this issue was not to the fore in ministers' public statements on interrogation policy. We found no references to the UK's international legal commitments regarding torture in ministerial statements during the key debates analysed from the 1970s. Those at the top of government took a wry view of their decision to ban the five techniques. The Foreign Secretary sent a telegram to British ambassadors in over 30 countries, which stated that it had been right to use the techniques in Northern Ireland. 'Nevertheless', he continued, 'in the British way, HMG [Her Majesty's Government] have thought it right to err on the side of caution and not to permit the techniques to be used in future. . .' (quoted in Newbery, 2015: 103).
European Court rulings, including Ireland v UK (1978), had important effects over the long term, as will be outlined below. During the 1970s itself, however, the British government's response to this case, and to the 'contradiction narrative' underpinning it, was to contest it in court and in public. Foreshadowing the approach of the United States after 9/11, British officials in the 1970s did not openly contest the global norm against torture. They claimed rather that their techniques of 'interrogation in depth' (a euphemism to rival that of the Bush administration) fell short of the definition of torture. This section has outlined how British officials dynamically contested the messages of their accusers, constructing rival narratives in which their coercive interrogation techniques were effective, based on long experience, morally justified and not a violation of the ECHR. Thus, shaming and rhetorical coercion had little impact during the 1970s. It was against the background of this deep narrative contestation that -as a previous section outlined -security forces in the UK continued to inflict coercive interrogation on Irish terrorist suspects throughout the 1970s.

Post-9/11: narrative reconfiguration
While the British authorities had contested all efforts to shame or criticise them during the 1970s, their response to these narratives evolved in the following decades. When the debate about torture once again intensified in the years after the September 11 attacks on America, important parts of the UK government and security establishment continued to disregard two of their accusers' main narratives, regarding the ethics and tactical effectiveness of torture. However, the other two narratives, concerning international reputation and strategic effectiveness, were essentially accepted by that time across most parts of the British government and its security agencies.
Torture is morally wrong. In the post-9/11 era, UK-based human rights NGOs and newspapers denounced the United States' use of torture as morally unacceptable and criticised British collusion in torture (Amnesty International, 2004: 88-89;Cobain, 2009). British government ministers agreed that torture was not morally acceptable (stating this about once per parliamentary debate on average). However, they also raised doubts on the ethical issue (also about once per debate on average) when they spoke of the moral 'dilemma' facing the government in cases where the need for intelligence might involve the UK colluding in torture or mistreatment carried out by foreign security services. One Foreign Office minister, Ivan Lewis, told Parliament that the government regularly faced 'some incredibly difficult judgement calls' on this issue (Hansard, 07 Jul 2009, column 946). As Home Secretary, Theresa May spoke in similar terms. 'It's a difficult judgement', she said, 'it's a difficult balance' (ISC, 2018a: 75).
Torture is tactically ineffective. During the 2000s, British-based human rights' NGOs and researchers repeatedly emphasised that torture as an interrogation technique produces false or unreliable intelligence (Amnesty International, 2004: 87-88;Sands, 2009: 9-11;Walton, 2008). However, government ministers never signalled their acceptance of this point in the parliamentary debates analysed. Instead, they stated in just over half of the debates that information gained through coercive interrogation could be important for protecting the security of the British people. The UK's counterterrorism strategy, CON-TEST, also indicated that intelligence obtained through torture 'may still be used to investigate and stop terrorist attacks' (UK Government, 2009: 78). Indeed, the government and intelligence agencies' official policy on cooperation with foreign secret services assumed that torture could be tactically effective in obtaining valuable intelligence. In operation between 2004 and 2010, this confidential policy required the intelligence agencies to 'balance the risk of mistreatment. . . against the need for the proposed action'. If there was an 'operational imperative for [cooperating with coercive interrogation by a foreign service] such as. . . obtaining life-saving intelligence', the importance of that would be weighed against 'the level of mistreatment anticipated' (UKSIA (UK Security and Intelligence Agencies), 2006: para. 38).
Torture contradicts your international commitments and damages your reputation. While efforts to influence the British government's attitude to torture on ethical and tactical grounds continued to have minimal impact, two other narratives proved over time to be more effective. As outlined above, Ireland v UK was a concern for the British government during the 1970s. In the following decades, the government was brought back before the Strasbourg Commission and the European Court of Human Rights on many occasions for its actions in Northern Ireland, such as its repeated killings of terrorist suspects, and it suffered some further adverse rulings. The government's reaction to such rulings alternated between defiant dismissal of the court (Dickson, 2010: 257) and a concern for its reputation on human rights. The latter impulse began to gain an upper hand by the mid-to-late 1990s, however.
During that decade, criticism of the UK's failure to uphold its international obligations on human rights increased in salience. Both the UN Committee against Torture and the European Committee for the Prevention of Torture (CPT) made their first visits to Britain in the early 1990s, documenting a number of embarrassing allegations of illtreatment in Northern Ireland detention centres (Carver, 2016: 112). Furthermore, as Dickson (2010: 255-258) outlines, the European Court became more assertive during the 1990s, symbolised in its decision that the UK had violated the ECHR right to life in one key ruling in 1995. It was clear by this point that Britain could be repeatedly exposed both in court and by the increasingly active scrutiny committees on torture if it violated its international commitments.
The government's reaction to this indicated a concern for its reputation on human rights. By the mid-1990s, the CPT reported that the UK authorities' cooperation with their inquiries was 'excellent' and the UN committee noted that allegations of mistreatment had substantially reduced since its first visit (Carver, 2016: 112;CPT, 2000: 8). At the same time, advocacy and discussion of human rights in Britain was more intense than it had been in previous decades, culminating in the government supporting the passing of the Human Rights Act 1998, which incorporated the rights contained in the ECHR into UK law (Feldman, 2002: 70-79). In the context of this discussion, the government was by this time showing a greater awareness of its international obligations on human rights and of how publicly failing to uphold them damaged Britain's reputation. Noticeably absent from government statements on torture in Parliament during the 1970s, by the 2000s, ministers were referring to the importance of Britain upholding its international obligations an average of four times per debate. The Foreign Secretary David Miliband, for example, emphasised 'our obligations under the UN convention against torture', before going on to state that 'the highest values of human rights. . . need to be adhered to. . . and it is very important that they are seen to be adhered to'. If they were not, he said, it would be an 'embarrassment' to the government (Hansard, 21 Feb 2008, column 555).
Three findings from the IR literatures on shaming and rhetoric help us to understand why the government was vulnerable to shaming and reputational concerns. First, as Schimmelfennig (2001: 65) has shown, if the standard of legitimacy used in shaming is clear and unambiguous, 'it becomes difficult to rhetorically circumvent its practical implications'. The legal prohibition of torture and inhuman and degrading treatment, contained in the ECHR, has more clarity than that which applies in other contexts including the United States (Mayerfeld, 2011). This enabled the European Court to rule against the UK in the Irish case. A second finding of the literature concerns the authority and credibility of the speaker or accuser (Goddard and Krebs, 2015: 27). The Irish government -as a member state (and peer of the UK) in the Council of Europe -had the authority to bring an interstate case against Britain before Europe's judicial authorities. The UK, like other European nations, has put itself under the auspices of a relatively strong European Human Rights regime, which means that it is subject to the judgements of external courts (Mayerfeld, 2011: 127-128). Thus, in addition to the Irish government, in 1978 the European Court of Human Rights effectively joined the effort to shame the British government for its use of coercive interrogation. The court had not only credibility but also the institutional authority to issue significant rulings against Britain in this and other cases. The UN and European committees on torture had a lower profile, but they too had credibility and were taken seriously by the government.
The UK case also supports a third finding of the literature, which emphasises that one of the most effective types of rhetorical coercion is shaming -showing that the target state's actual conduct is in contradiction with its declared international commitment to normative standards of behaviour (Schimmelfennig, 2001;Vennesson, 2014: 30-31). When the European Court ruled in 1978 that the UK had indeed violated its commitments under the ECHR in Northern Ireland, the government was embarrassed and it felt that its international reputation had been tarnished (Newbery, 2015: 124-125). As outlined above, this concern for reputation increased following further European Court rulings, particularly from the 1990s, and it coloured officials' recollection of the Irish case. A government official, who played a leading role in UK counterterrorism after 9/11, spoke of the reputational damage caused by this ECHR case: 'It was a disaster on stilts', he said, 'Very little intelligence of value was gained from these interrogations basically because they interrogated the wrong people. . . And then we got taken to Strasbourg. It was not a happy outcome'. He noted that 'it was impossible to win the PR battle, especially in America. . . and human rights organisations were up in arms'. And yet, the result could have been worse, he pointed out: 'In the end, we got away with cruel and inhuman treatment' (Interview, 2016). The government had 'got away' with it in some sense because the court did not rule that it had committed torture. Yet these words also indicated the embarrassment of officials at the Irish case's tarnishing of the UK's reputation and the defensive position that they found themselves in following European Court rulings.
The government had essentially accepted the implications of the shaming or 'reputation' narrative, which is that given that the UK has publicly committed itself to refraining from torture, flagrant contradictions of this position cause damage to its international reputation. Expressed in Parliament by Ministers, as noted above, this narrative also extended down to the level of senior intelligence and security officials and influenced some of their messages to staff in the first months after 9/11. For example, in January 2002, the intelligence agencies in London received reports that US agents were mistreating detainees in Afghanistan. On account of this, the agencies sent a communication to their officers in the field, which stated: Her Majesty's Government's stated commitment to human rights makes it important that the Americans understand that we cannot be party to such ill-treatment nor can we be seen to condone it. In no case should [the detainees] be coerced during or in conjunction with a SIS [MI6] interview of them. (UKSIA, 2002) A key internal policy document, issued in 2004, also emphasised to officers that if British intelligence was known to be involved in the mistreatment of detainees, this 'could result in damage to the reputation of the [intelligence] Agencies' (UKSIA, 2006: para. 9). Thus, while 'the Americans' were known to be abusing detainees, a strong signal was sent to UK intelligence officers that they by contrast could not perpetrate or directly cooperate with coercive interrogations.
Torture is strategically ineffective. Between the 1980s and the 2000s, Human Rights NGOs and some journalists and politicians in the UK and Ireland increasingly advanced a narrative which held that torture and other repressive British operations had alienated sections of the population in Northern Ireland and increased support for the IRA. They pointed to a good deal of information that supported their narrative, was publicly available and thus could not be easily dismissed by the UK government (Coogan, 1997: 149-152;English, 2003: 140-143;Liberty, 2004: 15, 18). Though such criticisms had been discussed for many years, they crystallised and began to be accepted by British officials in the mid-to-late 1990s as IRA violence reduced and the UK and Ireland entered a peace process. The Irish government and Northern Irish nationalists argued strongly that facing up to past mistakes was an important part of the reconciliation process, and the British publicly signalled their acceptance in a way that they had not done before. Having long resisted it, in 1998 the government established an independent inquiry into Bloody Sunday, showing a willingness reconsider this infamous 1972 operation in which the British Army shot dead 13 civilians. Prime Minister Tony Blair (2010: 165) accepted this demand, he said, 'to assuage Nationalist opinion and under pressure from the Irish'. Blair also apologised for the injustice suffered by the Guildford Four (Ewing, 2006: 54). In talks with UK government officials in 1997, the republican leader Gerry Adams spoke about how repressive British operations in the early 1970s had increased support for the IRA. Blair's Chief of Staff Jonathan Powell (2008: 46) was struck by Adams's criticism and he came to accept that the 'brutality' of Britain's five techniques of interrogation in Northern Ireland had 'alienated a generation of Catholics'. At the same time, British Army figures like Michael Dewar (1996: 55-60) were admitting that the five techniques had to be discontinued in light of 'international opinion' and that Bloody Sunday was 'catastrophic' in how it alienated the Catholic population.
In the post-9/11 era, the British government repeatedly showed its acceptance of the narrative that torture is counterproductive for any counterterrorism strategy. In contrast to the 1970s when it was completely absent from the debates, during the 2000s government representatives did make this point in just over half of the ministerial statements to Parliament analysed. UK officials' response to the 'strategy' narrative of their accusers was a little more creative than a simple acceptance of the criticism. As indicated above, social influence rarely works through a simple mea culpa alone. UK officials were able to accept this critical narrative by recontextualising it in important ways. They developed a public rhetoric of 'lesson learning', in which they depicted themselves as actors capable of self-awareness and reasoned change, based on a hard-nosed analysis of policy effectiveness. In 2006, for example, the Head of the British Army introduced an official review of operations in Northern Ireland by stating that it was 'important to learn from mistakes, where they were made, and to ensure that they are not repeated' (CGS (Chief of the General Staff), 2006: i). The Army's review admitted that with the introduction of internment in 1971 and 'the use of deep interrogation techniques, . . .the information operations opportunity handed to the republican movement was enormous'. It stated that these repressive methods 'had a major impact on popular opinion across Ireland, in Europe and the US. Put simply, on balance and with the benefit of hindsight, it was a major mistake' (CGS, 2006: para. 220).
This rhetoric of lesson learning found its way into Anglo-American intelligence discussions in the days after the 9/11 attacks. Listening to the US plans to hit back hard against Al Qaeda, one senior UK intelligence officer who had previously overseen operations in Northern Ireland was said to have told his American counterpart: 'You need to learn from our history' (Mayer, 2009: 29-30). The lesson that he and many other British officials spoke of after 9/11 concerned the importance of a 'proportionate' response to terrorism (Foley, 2013: 261-262). The government and security establishment thus recast its accusers' critical points in the context of a positive narrative for itself -one which says that Britain has learned from its past experience to craft a 'proportionate' and 'strategic' approach to counterterrorism. This was expressed notably by Sir David Omand, a high-ranking government official after 9/11 and the lead author of the UK's counterterrorism strategy, CONTEST. Omand recalled the 'dire consequences' of the British security forces' use of coercive interrogation in 1971. 'The fall-out from that episode polarized attitudes in Northern Ireland and overseas for many years', he wrote. The decision of the UK government not to use torture after 9/11, in his view, 'reflect[ed] a more strategic view of the long term impact of short-term security actions' (Omand, 2010: 270-273).
Thus, a prevailing narrative across the British government and security establishment during the 2000s was that if torture is known to be used, it will alienate key constituencies (such as Irish or Muslim communities) in a way that is counterproductive for a counterterrorism strategy over the long term. As the government and intelligence agencies' key internal policy document on torture noted, it 'could result in further radicalisation, leading to an increase in the threat from terrorism' (UKSIA, 2006: para. 9). Initially developed by critics of the British government, the salience of this narrative had increased greatly since the 1970s and it was now largely accepted by the target government.

Rhetorical manoeuvres: avoiding torture, colluding in torture
The British case illustrates well the dynamics of rhetorical manoeuvre in a context of multiple narratives. As outlined above, opponents of torture did succeed in manoeuvring the government onto more favourable rhetorical terrain for them. Thus, the broad acceptance within government of two anti-torture narratives -concerning strategy and reputation -explains much about why UK intelligence and security officials generally refrained from perpetrating torture after 9/11. Unlike their American counterparts, there were no circumstances in which British officials could publicly defend an 'enhanced interrogation' programme of their own. They had accepted that the public effects of such a move would be 'dire', as Omand put it.
However, British officials resisted the narratives which emphasised the ethical impermissibility and tactical ineffectiveness of torture. Their rhetorical manoeuvre was to accept two anti-torture narratives while simultaneously contesting two others in an attempt to circumvent the full implications of the international norm against torture. As the shaming literature would expect, this norm was not internalised by UK officials. Instead, they varied their behaviour according to their perception of the likelihood of exposure. They did not perpetrate torture themselves -but they did collude in the violations of other states.
Collusion flowed directly from the particular configuration of narratives accepted by the UK government. They told themselves that torture could be morally acceptable and tactically beneficial in gaining intelligence. These two narratives were concerned with the private realm of torture -the ethics and 'secret' tactical 'benefits' of what happens inside the interrogation room. In the narrative world of British officials, the downsides of torture were only in the public realm (if it was exposed to the public) -its negative strategic effects on certain constituencies and how it would damage the UK government's international reputation. The question for officials therefore was which course of action was more likely to lead to public exposure? If the state perpetrates torture, it places itself very close to the 'scene of the crime'. And the UK's coercive interrogations of the 1970s did lead to the government being condemned in court. In this context, collusion was an attractive option for UK officials because it allowed them to increase the distance between themselves and the act of torture, while still being able to reap some of its perceived tactical benefits. They believed that it would be possible to maintain a covert collusion in torture, which was carried out not by them but by other governments in distant lands, without this involvement becoming public knowledge (Aldrich, 2011: 162-163;Cobain, 2009). If the fact of collusion never became publicly known, the British government would never have to suffer the negative strategic or reputational effects of its involvement in torture.
In the event, however, British collusion did not remain covert. It was uncovered by investigative journalists and activists who sought to shame the UK government for its links to torture by foreign security services. It has been the subject of parliamentary committee inquiries, police investigations, and an aborted judge-led inquiry (Cobain, 2012: 264-271;ISC, 2018: 8-11). In the context of this shaming and controversy, in 2010 the government issued a new, more restrictive policy governing the intelligence agencies' liaison with foreign services. Since 2010, there have been far fewer allegations that British officials have colluded in torture, compared with the 2001-2010 period (ISC, 2018a: 1, 25).

Conclusion
This article has argued that there are significant rhetorical constraints on democracies' involvement in torture. It demonstrated this argument by showing that rhetorical coercion and shaming have had a substantial influence on the UK government's record on torture. The analysis drew attention to the multiplicity of anti-torture narratives in two British cases and the resultant complexity of the contestation between accusers and the UK government. The two most effective anti-torture narratives were those that laid out the negative effects of torture on counterterrorism strategy and on the UK's international reputation. The former gained much of its power from a compelling narrative about the counterproductive effects of British torture in Northern Ireland, while the latter was given weight by European Court rulings against the UK and negative findings by international scrutiny committees on torture. These two narratives were considerably more salient by the 2000s than they had been during the 1970s. The British authorities' acceptance of two of their accusers' key narratives is crucial for understanding why they were significantly less involved in torture in the post-9/11 era than they were in the 1970s. Yet given the multiplicity of narratives, target governments will have opportunities for rhetorical manoeuvre -to accept some of their accusers' messages while contesting others. The result is a specific configuration of prevailing narratives, which enable and constrain governments in particular ways. This helps to explain why the UK's security agencies avoided coercive interrogation themselves but still colluded in torture by foreign governments after 9/11. These findings have implications for scholarship on shaming, rhetorical coercion and international norms. On the one hand, it is clear that shaming efforts, which focused on the UK's reputation and international commitments on torture, had a significant impact on the government and its security agencies. We might expect this given the status of the prohibition of torture as a peremptory international norm, arguably the most fundamental human right, and one that cannot be derogated from. If shaming is important anywhere, it should be important here. Yet even when it concerns torture -this most ostensibly shameful of violations -the limits of shaming are also apparent. The UK experience indicates that, in order to have a significant impact on governments, the shaming narrative on torture and international reputation needs to be combined with a utilitarian narrative, which emphasises the strategic ineffectiveness of torture for countering terrorism. This analysis indicates that the literature on shaming needs to pay more attention to the interaction between shaming and other types of rhetorical coercion. Finally, by analysing the contestation of the anti-torture norm, this article also contributes to recent efforts to explain the influence of international norms not as the product of unidirectional diffusion and persuasion but as a contentious process in which manoeuvring and interplay between local actors shapes the direction of normative change (Wiener, 2018).
The four-part typology put forward in this article can facilitate theoretically informed empirical research not only on further cases of torture but also on narrative contestation in other domains of security practice. Researchers might usefully examine how narratives concerning ethics, tactics, strategy and reputation have been constructed to oppose practices such as targeted killings, preventive detention or surveillance -and how governments have contested such critiques.
In terms of implications for practice, the British case indicates that governments and security agencies in liberal democracies do not stop torturing because they think it is ethically unacceptable or tactically ineffective. It may be difficult to influence them on either of those points. However, they may reduce or even stop their involvement in torture because they accept that it is counterproductive for their counterterrorism strategy or has a damaging effect on their international reputation. These findings suggest that human rights campaigns should not overemphasise any purely ethical objections to torture, nor should they stake everything on an effort to show that torture is a tactically ineffective means of securing intelligence. The narratives that may have the most impact in swaying liberal democratic governments and security officials are rather those that focus on the negative effects of torture on counterterrorism strategy and international reputation. Pushing these two narratives may be the best way to help ensure that torture becomes a socially unsustainable option for security agencies.