‘The United States is committed to the world-wide elimination of torture and we are leading this fight by example’. President George W. Bush’s administration included this quote in a statement to commemorate the 2003 United Nations International Day in Support of Victims of Torture.Footnote 1 The statement also condemns governments that are ‘deliberately inflicting severe physical or mental pain or suffering on those within their custody or control’ and vows to work with like-minded states to prevent ‘other cruel and unusual punishment’. This otherwise-routine message alarmed the Central Intelligence Agency (CIA), and the agency’s top lawyer called the White House to share concerns. Three months earlier, the CIA had subjected Khalid Sheikh Mohammed (KSM), the mastermind of the September 11 attacks, to repeated applications of ‘the waterboard’, a technique in which water is poured over a cloth on a person’s face to suffocate him and simulate a drowning death. On the waterboard, the body involuntarily panics in a desperate search for air; it is excruciating.Footnote 2 The White House assured the CIA of continued support for the interrogation program. If Bush and others in the administration were leading the fight against torture in public, they were undermining it in private. Soon they would undermine it publically as well.
States sometimes resort to torture as an interrogation method, but it is hard to see why. Torture is barbaric and antithetical to what most states – especially liberal ones – claim to represent. Michael Ignatieff writes that ‘torture should remain anathema to a liberal democracy and should never be regulated, countenanced, or covertly accepted in a war on terror. For torture, when committed by a state, expresses the state’s ultimate view that human beings are expendable’.Footnote 3 Henry Shue (Reference Shue1978, 124), referring to torture, writes, ‘No other practice except slavery is so universally and unanimously condemned in law and human convention’. Yet Shue argues that ‘torture is widespread…’ The United States, Israel, Britain, and France have all used torture to gather intelligence.Footnote 4 What accounts for this behavior, especially when liberal democracies have not only denounced torture, but codified the ban against it?
The hypocrisy would not surprise most international relations scholars, especially realists and rational choice theorists that expect states to say and do things in pursuit of self-gain, even if their words and deeds are mutually inconsistent (see, for instance, Krasner Reference Krasner1999). However, there are (at least) five instrumental reasons for avoiding interrogational torture. First, torture is a clumsy method for gathering intelligence. Perhaps the most persistent critique of interrogational torture is its flimsy connection to truth: victims are likely to say anything, including false statements, to stop the suffering.Footnote 5 I defend the claim that torture is an ineffective means of gathering intelligence – a topic that has engendered considerable debate in the popular press since 2001 – in more detail in the Appendix. Second, torture can fuel the enemy’s recruitment efforts, giving life to terrorist and insurgent groups.Footnote 6 Third, torture may deter the enemy from surrendering, prolonging wars and causing more casualties on both sides.Footnote 7 Fourth, torture can produce extreme stress in the interrogator (Price Reference Price2014). Fifth, torture runs counter to the ‘hearts and minds’ campaigns that many modern counterinsurgency strategists advocate (US Army 2006). Torture does not easily lend itself to obvious, straightforward functional explanations. So why have states practiced interrogational torture? More broadly, why would states do something that is both odious and of dubious instrumental value?
This puzzle falls partly in the ‘blind spots’ of the two most prominent theoretical frameworks in international relations scholarship: rationalism and constructivism.Footnote 8 Rationalists insist that states make choices based on self-interested, cost–benefit analysis; sometimes these choices fall in line with international norms (which can impose costs for non-compliance), and sometimes they do not. Thus, rationalism can account for the norm-breaking side of the puzzle but cannot easily explain why states often do not act in their self-interest and sometimes do not update when the returns from torture are poor. Social constructivists argue that shared understandings, norms, and rules constitute states’ interests and identities, and so their choices are influenced, if not determined, by a sense of propriety, however thin. Hence, constructivists are well positioned to account for why states might not maximize their utility, but they would have more trouble explaining why states operate outside of a norm that is widely held. It is not that these frameworks cannot account for torture; I think they can. The point is that this puzzle takes rationalist and constructivist orientations out of their comfort zones.
I argue that part of the answer may lie in a more complete understanding of normative effects. I suggest two ways that the norm against torture can attract state leaders to the very act that the norm proscribes. First, the norm against torture may play a part in convincing leaders that torture works. Norms constrain states; leaders may believe that those who are unconstrained have an advantage. Second, breaking ethical norms affords leaders a heroic narrative: leaders can prove to themselves and others that they are willing to make hard choices and sacrifice their morals for a “greater good”. These accounts reveal unexpected and perverse ways in which norms shape behavior; as such, the argument extends current constructivist thinking about norms.
Although I do not test these arguments against competitors here, I draw repeatedly on the example of the use of torture in the US war on terror to illustrate the theoretical argument. The United States has previously used torture, but the interrogation program that followed the attacks of September 11 represented a dramatic shift (in terms of scale, personnel, and eventually, disclosure) in the extent to which the US government embraced torture. This case is especially puzzling since torture has long been illegal in the United States, the State Department has acted as a whistle-blower concerning torture in other countries, and arguments against the effectiveness of torture were not unknown. Furthermore, the Bush administration was interested in gathering intelligence to prevent future attacks. Although revenge and even sadism may have played a part in the use of torture, Robert Jervis (Reference Jervis2010) demonstrates through numerous examples that the Bush administration was no less interested in accurate intelligence than its predecessors. As I show, torture’s status as off-limits attracted the Bush administration, which was convinced that it needed a far-reaching approach to combat terrorists, and it furnished the administration with a narrative of moral heroism to justify its actions.
This paper makes three theoretical contributions to constructivism. First, it expands our understanding of what norms can do. Previous scholars have suggested how norms can have unexpected consequences (Ron Reference Ron1997; Tannenwald Reference Tannenwald1999, Reference Tannenwald2007; Carpenter Reference Carpenter2003); in my argument, norms lead to exactly the opposite of the standard effect. Second, the paper focuses on how norms affect political actors ‘on the ground’, and avoids the temptation to treat norm-influenced behavior as an automated response to normative scripts. The combination of wide-ranging normative effects with individual and group interests shaped in part by norms leads to a fuller understanding of norm compliance and violations.
The third contribution may be articulated best by example. Arch-realist John Mearsheimer (Reference Mearsheimer2006, 162) argues, ‘In the anarchic world of international politics, it is better to be Godzilla than Bambi’. Constructivists have done excellent work unpacking what Mearsheimer and other international relations scholars mean by ‘anarchy’, and what logic – or lack thereof – the anarchy of the international system entails (e.g., Wendt Reference Wendt1992). This paper investigates what it means to be Godzilla. It might mean being a state with strength and ruthlessness. Just as norms help to define our particular culture of anarchy, norms can also tell us what ruthlessness is.
In the next section, I discuss some of the previous work on the ‘irregular’ effects of norms. Then I present the two unintended effects of norms, with accompanying examples mainly from the post-September 11 war on terror. I then turn to alternative explanations of torture, not to test them against my explanation, but to see where my explanation fits with existing accounts. After briefly exploring other possible unintended consequences of norms, I conclude with some initial thoughts on scope conditions and suggestions for future work on this topic.
Norms and ‘Irregular’ effects
Norms are standards of appropriate behavior for a given identity (Katzenstein Reference Katzenstein1996). Nina Tannenwald (Reference Tannenwald1999, Reference Tannenwald2007) identifies three effects of norms in her work on the normative prohibition against the use of nuclear weapons. First, norms impose regulations on state actors, forcing them to consider the costs and benefits of compliance; second, norms have constitutive effects, creating or defining forms of behavior, roles, and identities; and third, norms can have ‘shadow’ or permissive effects, in which a norm claims our attention, allowing other acts, which are similar and equally deplorable, but categorically different, to escape condemnation. The fuel-air explosives that the United States used in the first Gulf War exemplify this latter effect. These weapons were as powerful as small nuclear devices, but because they were not actually in the nuclear category, they eluded harsh criticism.
Tannenwald (Reference Tannenwald2007, 44) understands permissive effects to be a subcategory of constitutive effects. Permissive effects ‘are defined in relation to the categories constituted by the norm or taboo and the larger normative context in which the taboo is nested’. For instance, the norm against the use of weapons of mass destruction (WMD) defines the categorization of WMD. In doing so, it defines non-WMD weapons (like those fuel-air explosives). That said, permissive effects can arise in the shadow of regulative or constitutive effects. Regulative constraints on the use of WMD can make the use of non-WMD ‘cheaper’ than might otherwise be the case. Likewise, the norm against WMD defines this categorization of weapons as unusable; by extension, it categorizes non-WMD as usable (Tannenwald Reference Tannenwald2007, 47).
International relations scholars have identified other permissive effects and unexpected consequences of norms, including the norm against torture. James Ron (Reference Ron1997) provides a nice illustration of Foucault’s work chronicling the state-level shift from punishment to discipline. Ron explains how the horror over overt torture has forced governments to choose other forms of social control that are subtler but potentially more comprehensive. Instead of doing away with torture entirely, some states try to regulate and, by extension, legitimize torture. One result is a rise in the use of ‘stealth’ interrogation techniques that leave no scars, including forced standing and restriction of air, which Ron observes in Israel in the early 1990s.Footnote 9 The Israeli government’s shift in policy had multiple consequences. On the one hand, detainees suffered fewer broken bones and other lasting physical impairments from the ‘stealth’ techniques that were used. On the other hand, the lack of physical scars added to the mental trauma by removing an obvious excuse for any confessions (real or fabricated) made under duress.
Ron’s example shows how norms can have indirect effects. Domestic and international monitoring constrained the Israeli government (a regulative effect), but the government’s response had the unforeseen consequence of increasing psychological damage among detainees in some cases. The norm against overt physical torture may have also paved the way for government efforts to legitimize alternative forms of ‘discipline’.
Two perverse effects of the norm against torture
My argument extends previous work on the unintended effects of norms. I argue that the full effect of a norm can include violations of the very act that the norm proscribes. I detail two ways in which this can happen. First, the norm against torture attracts political leaders who are looking for an advantage in the face of challenging intelligence-gathering scenarios. Second, the norm offers political leaders a narrative of sacrifice for the greater good. Both stories connect the norm to torture, but each has its own causal pathway.
Nasty equals effective
The first causal story proceeds as follows: an executive administration faces a security challenge, such as terrorism or an insurgency, which puts a premium on intelligence. These political leaders feel pressure from the domestic population, their peers, or even from themselves, to produce results, but not necessarily explicit pressure to torture. They are looking for any advantage they can find, and they feel the need to get tough with the hardened opposition by thinking outside the box. Norms and laws – that is, the rules – define that box. The leadership decides that it cannot afford to be ‘hamstrung’ by the rules and that it needs either to push against those rules as far as it can without getting itself in trouble or to break the rules outright to obtain intelligence.
Two parts work together here. The first is a belief – or maybe just a suspicion – that in the nasty world of international politics, cheaters win. Such a belief would hardly be surprising. Observers of international politics have appreciated cheaters’ successes since Odysseus packed a wooden ‘gift’ horse with soldiers. Some of history’s most famous backstabbers such as Bismarck and Hitler saw considerable success as a direct result of their backstabbing. The set-up of the Prisoner’s Dilemma rewards the defector in a world of defectors. The belief that cheaters win in international politics is a reasonable cognitive shortcut even if it is not always true.Footnote 10
The second part of this explanation is the flipside of the usual constitutive effect of the norm. Proscriptive norms define what rule-following is. By extension, they also define what cheating is. Putting the two parts together, the off-limits nature of certain behaviors attracts the gaze of those who believe that cheaters win in international politics.
The result is a kind of ‘upside-down world’, where good is bad and bad is good. For instance, in the early 1930s, Japanese political leaders began to develop biological weapons, ‘in part because Japanese officials were impressed that germ warfare had been banned by the Geneva Convention of 1925. If it was so awful that it had to be banned under international law, the officers reasoned, it must make a great weapon’ (Kristof Reference Kristof1995; Kier and Mercer Reference Kier and Mercer1996). For these officials looking for an advantage in future wars, ‘awful’ was ‘great’.
The war on terror provides more illustrations of the nasty-equals-effective proposition. The story starts with a need for intelligence: the attacks on 11 September 2001, put a premium on intelligence. The attacks themselves were an intelligence failure, and the public was likely to judge the failure to anticipate subsequent attacks much more harshly. Moreover, counterterrorism relies heavily on intelligence because the enemy does not wear a uniform; separating terrorists from civilians is a meticulous act of intelligence (Gross Reference Gross2010). The war on terror was ‘a war in which intelligence [was] everything’.Footnote 11
Almost immediately, the Bush administration imagined a war in which normative (as well as financial) constraints would not bind those charged with winning it. The day of the attacks, President Bush reportedly said as much:
‘I want you all to understand that we are at war and we will stay at war until this is done. Nothing else matters. Everything is available for the pursuit of this war. Any barriers in your way, they’re gone…’ When, later in the discussion, Secretary Rumsfeld noted that international law allowed the use of force only to prevent future attacks and not for retribution, Bush nearly bit his head off. ‘No’, the President yelled in the narrow conference room, ‘I don’t care what the international lawyers say, we are going to kick some ass’ (Clarke Reference Clarke2004, 24).
With the war on terror just a few hours old, Bush was already sharing his belief that to really ‘kick ass’ in wartime, one had to ignore the constraints of ‘international lawyers’.
A few days later, in a well-known interview with Tim Russert on Meet the Press, Vice President Dick Cheney laid out his philosophy for how to fight the war on terror:
VICE PRES. CHENEY: I’m going to be careful here, Tim, because I–clearly it would be inappropriate for me to talk about operational matters, specific options or the kinds of activities we might undertake going forward … We…have to work, though, sort of the dark side, if you will. We’ve got to spend time in the shadows in the intelligence world. A lot of what needs to be done here will have to be done quietly, without any discussion, using sources and methods that are available to our intelligence agencies, if we’re going to be successful. That’s the world these folks operate in, and so it’s going to be vital for us to use any means at our disposal, basically, to achieve our objective.
MR. RUSSERT: There have been restrictions placed on the United States intelligence gathering, reluctance to use unsavory characters, those who violated human rights, to assist in intelligence gathering. Will we lift some of those restrictions?
VICE PRES. CHENEY: Oh, I think so. I think the – one of the by-products, if you will, of this tragic set of circumstances is that we’ll see a very thorough sort of reassessment of how we operate and the kinds of people we deal with. There’s – if you’re going to deal only with sort of officially approved, certified good guys, you’re not going to find out what the bad guys are doing. You need to be able to penetrate these organizations. You need to have on the payroll some very unsavory characters if, in fact, you’re going to be able to learn all that needs to be learned in order to forestall these kinds of activities. It is a mean, nasty, dangerous dirty business out there, and we have to operate in that arena. I’m convinced we can do it; we can do it successfully. But we need to make certain that we have not tied the hands, if you will, of our intelligence communities in terms of accomplishing their mission.
MR. RUSSERT: These terrorists play by a whole set of different rules. It’s going to force us, in your words, to get mean, dirty and nasty in order to take them on, right? And they should realize there will be more than simply a pinprick bombing.
VICE PRES. CHENEY: Yeah… (White House 2001; emphasis added).
Cheney’s assertion that the United States must fight ‘dirty’ as the terrorists do exemplifies the nasty-equals-effective argument. Interestingly, Russert uses the phrase ‘whole set of different rules’. His phrasing and Cheney’s response of affirmation imply that for Cheney, terrorists’ lawlessness gives them an advantage, and so in order to get the upper hand, the United States would have to ‘get mean’.
Some of Cheney’s ‘unsavory characters’ were part of the intelligence agency in Egypt, where the Bush administration, via the CIA, sent one of its earliest (allegedly) high-value detainees. Officials believed that Ibn al-Shaykh al-Libi, whom Pakistani authorities captured and turned over to the United States in the first few days of 2002, was the most senior Al Qaeda official captured to date (Schmitt Reference Schmitt2002). At first, the FBI handled much of the questioning, and according to some reports, al-Libi was talking (Frontline 2005). But the CIA then forcibly removed al-Libi from Bagram Air Base and the FBI’s jurisdiction and transferred him to Egypt through extraordinary rendition.
The United States has had an extraordinary rendition program – the extrajudicial transfer of persons from one state to another – since at least the Reagan administration (Garcia Reference Garcia2010). After September 11, the Bush administration expanded the program. Extraordinary rendition is widely understood to be a way to circumvent torture laws. States that are parties to the United Nations Convention Against Torture can render their prisoners to states with looser codes as long as there are no ‘substantial grounds for believing’ that detainees will be tortured (United Nations 1984). The subjective nature of that phrase gives leaders ‘plausible deniability’Footnote 12 and the potential to claim, as Secretary of State Condoleezza Rice did in a speech in 2005, that ‘the United States has not transported anyone, and will not transport anyone, to a country when we believe he will be tortured’ (Gilmore Reference Gilmore2005).
Yet as former CIA officer Vinnie Castrovince explains, ‘[Al-Libi is] carried off to Egypt, who torture him [sic]. And we know that he’s going to be tortured. Anyone who’s worked on Egypt, has worked on other countries in the Middle East, knows that’ (Frontline 2006). Egypt’s primary – perhaps sole – draw was that it did not follow ethical and legal norms of interrogation. In the words of former CIA agent Robert Baer, ‘If you want a serious interrogation, you send a prisoner to Jordan. If you want them to be tortured, you send them to Syria. If you want someone to disappear – never to see them again – you send them to Egypt’ (ACLU 2005).
But the Bush administration would not be able to render all of their detainees to places like Egypt, and may not have wanted to outsource so much intelligence gathering even if they could. So they developed an in-house program. In late 2001, the defense department asked the Office of Legal Counsel (OLC) whether detainees in the war on terror were protected by the Geneva Conventions III on the Treatment of Prisoners of War (GPW). John Yoo and Robert Delahaunty of the OLC, in a memo to the general counsel at the Defense Department, argued in early January 2002 that the GPW need not apply to Al-Qaeda or Taliban detainees, but remained at least nominally agnostic about whether waving Geneva guidelines as a matter of policy was a sound idea. On January 18, President Bush told members of his administration that GPW would not apply, and Rumsfeld relayed this news to his subordinates the next day. The only additional instructions for detainee treatment contained in Rumsfeld’s (Reference Rumsfeld2002) memo on January 19 sounded notes of restraint, but left a sizeable loophole: ‘The Combatant Commanders shall, in detaining Al Qaida and Taliban individuals under the control of the Department of Defense, treat them humanely and, to the extent appropriate and consistent with military necessity, in a manner consistent with the Geneva Conventions of 1949’ (emphasis added).
The State Department and some members of the military dissented, and an internal debate ensued. In late January, a memo to President Bush bearing White House counsel Alberto Gonzales’ signature, but likely penned in large part by Cheney’s lawyer David Addington (Gellman and Becker Reference Gellman and Becker2007a), summed up the policy argument against applying Geneva:
As you have said, the war against terrorism is a new kind of war. It is not the traditional clash between nations adhering the laws of war that formed the backdrop for GPW. The nature of the new war places a high premium on other factors, such as the ability to quickly obtain information from captured terrorists and their sponsors in order to avoid further atrocities against American civilians … [T]his new paradigm renders obsolete Geneva’s strict limitations on questioning of enemy prisoners… (Gonzales Reference Gonzales2002).
The heightened need for intelligence meant that American interrogators could not be fenced in by rules, the memo argues. The decision to forego Geneva, which President Bush again confirmed 2 weeks later (White House 2002), was instrumental in leading to abuses, especially those in the military (Phillips Reference Phillips2012; Hutchinson et al. Reference Hutchinson, Jones, D’Alemberte, Gushee, al-Hibri, Irvine, Kennedy, Pickering, Sessions and Thomson2013).
Meanwhile, the CIA began to set up a separate interrogation program. Since the CIA had no recent interrogation experience, it sought the advice of others. The CIA ended up developing a program based on the work of psychologists in the Survival, Evasion, Resistance, and Escape (SERE) training endured by military members most endanger of capture. One would be hard pressed to find a more extreme set of officially sanctioned tactics in the United States. The CIA did partner with experienced interrogators from the FBI. But the FBI interrogators were subordinate to the program developed primarily by hired psychologists from SERE training who had never conducted a live interrogation (Soufan Reference Soufan2011). The CIA may have valued experience, but it valued harsh measures even more.
When Abu Zubaydah was captured in March 2002, he took al-Libi’s place as the United States’ most important detainee to date in the war on terror. As with al-Libi, FBI interrogators at first reported success in getting Zubaydah to talk using rapport-building methods (Soufan Reference Soufan2011). But they soon came in conflict with the CIA contractors. The CIA had primary authority over interrogation, and the coercive measures began. The CIA was simultaneously petitioning the OLC for legal clearance regarding a set of techniques to use on Zubaydah. The OLC, in a memo bearing the name of Jay Bybee but mostly written by John Yoo, followed up on a few months of verbal assurances with written affirmation of almost all of the requested techniques, including sleep deprivation and waterboarding (Bybee Reference Bybee2002).
Yoo explains that the CIA wanted to know ‘what the legal limits of interrogation [were]’ (Gellman and Becker Reference Gellman and Becker2007b). The CIA did go beyond what was sanctioned (CIA Inspector General 2004; Blakeley Reference Blakeley2011), but it can hardly be described as a rogue organization in the post-9/11 context. When the CIA was first pursuing the assignment of interrogation for high-value detainees in the days following September 11, the organization’s willingness to push boundaries was a key factor in why Bush chose them. ‘The truth’, said one CIA official, ‘is that the President wanted it. So everyone else wanted to be the most aggressive. A lot of ambitious people played on Cheney and the President’s fascination with this’ (Mayer 2008, 53). Neither much of the CIA nor much of the Bush administration needed extensive convincing about harsher methods. The turning point was September 11 itself. In a statement before the US Congress in 2002, the CIA’s lead counterterrorism official Cofer Black cryptically reported, ‘All I want to say is that there was ‘before’ 9/11 and ‘after’ 9/11. After 9/11 the gloves come off’. For Black as well as many others, following norms and rules after 9/11 would have been like wearing gloves in a bare-knuckled world. This is why Black concluded his statement by advocating the continuation of a policy of ‘no limits aggression’.
Defenders of harsh interrogation, including Bush and Cheney, continued their advocacy of such practices through Bush’s two terms and beyond.Footnote 13In a chapter on the administration’s use of torture, Holmes (Reference Holmes2007, 279) writes:
One reason the Administration’s defenders remain unmoved by the numerous powerful arguments against the utility of coercive interrogation might be that they are not really focused on the information extracted by these methods… They may prize the cruel, inhumane, and degrading treatment of prisoners precisely because it violates international norms and the rules of war. Psychologically persuasive evidence that a medicine is effective is that it tastes unbearably foul (emphasis in the original).
Holmes goes a bit too far here; former administration officials and CIA officers use examples to argue their side, even if the evidence does not always add up. However, their initial intuitions and reading of the evidence about the efficacy of torture may well have been (and continue to be) informed by torture’s status as an off-limits practice.
In one of his few public appearances since the end of his term, Bush described his thought process after capturing 9/11 mastermind KSM: ‘The first thing you do is ask, what’s legal? What do the lawyers say is possible’ (Hamby Reference Hamby2009)? In other words, how far would interrogators be able to go in terms of severity?Footnote 14 Interestingly, Bush does not say that he first asked, ‘Who are the most experienced interrogators? Who has the best record of success?’ Degree of severity likely informed estimates of the probability of successful interrogation.
The Bush administration pushed the normative envelope in a number of areas during the war on terror, including indefinite detention, closed trials, preventive war, and targeted killing, in addition to detainee abuse. This suggests that torture was not an act specifically plucked from a pile of war tactics for its stellar record as a means of producing intelligence; rather, torture emerged from a broader set of beliefs about the efficacy of policies that some have considered too harsh or unjust to be accepted practice.
Heroic narrative
The torture norm also gives political leaders a heroic narrative of making hard choices to keep a country safe, especially when ticking time bomb scenarios are invoked. Holmes (Reference Holmes2007, 274) makes the point directly:
The ticking time bomb fable… suggests the quiet heroism of those who, defying moral norms and legal conventions, choose torture. They sacrifice their scruples for the greater good. They come close to being martyrs for their country, sacrificing moral conscience on the altar of political responsibility.Footnote 15
This tale of sacrificial heroism would make no sense if torture were not a deplorable, norm-breaking act. Traditional interrogation methods of rapport-building would not entail the same kind of sacrifice.
Norms again play a constitutive role here: for certain actors, proscriptive norms contribute to the definition (or redefinition) of ‘heroism’. As with the first perverse effect, those ‘certain actors’ and their prior beliefs are crucial. Norms affect different actors differently, and although it is difficult to know at the outset who will find norm-breaking heroic, we might hypothesize that democratic leaders are especially susceptible. Democratic governments pride themselves on the attention they pay to human rights. They believe that they have the scruples to sacrifice. This makes the decision to torture ‘brave’ in a perverse sense. Political leaders must rise to the challenge and show that democratic rule-following has not softened them.
Scholars have noted themes of heroism in the narratives that the Bush administration constructed in the war on terror. Some of the heroism is about being on the ‘good’ side of a good vs. evil struggle, and the more horrible the opponent, the more heroic the defender (Jackson Reference Jackson2007; Devetak Reference Devetak2005). Securing the legitimate right to torture could be a way of reinforcing these identities (Blakeley Reference Blakeley2007). My argument adds another element: part of the ‘heroism’ can include doing the moral labor of lowering oneself, at least temporarily, to the level of one’s enemies. Building up a narrative of being ‘good’ not only accentuates the sacrifice; it also implies that a supposedly peace-loving country can come away innocent and unscathed thanks to the heroes in office who don the torturer’s clothes and intervene on the citizens’ behalf.
Some political leaders in the war on terror told tales of heroism to justify the authorization of torture.Footnote 16 After his terms, Bush said of the harsh interrogation program that ‘the techniques…were necessary and are necessary to be used on a rare occasion to get information to protect the American people’ (Fox News 2009). If torture is used only rarely, then the decision to torture has the potential to be a difficult one. Bush repeatedly characterized the decision to use enhanced interrogation as a tough choice. ‘You have to make the tough decisions’, he told an audience in Michigan in 2009. ‘They’ve captured a guy who murdered 3000 citizens … that affected me … They come in and say he may have more information … and we had an anthrax attack … and they say he may have more information. What do you do?’ (Gardner Reference Gardner2009). Bush wanted to portray himself as a president who rose to the challenge when his country needed him: ‘I made the decision, within the law, to get information so I can say to myself, ‘I’ve done what it takes to do my duty to protect the American people’. I can tell you that the information we got saved lives’ (Hamby Reference Hamby2009). There is some evidence that Bush used the ‘what(ever) it takes’ line – a line that conjures images of TV’s fictional counter-terrorist torture hero Jack Bauer – in private as well (Mayer 2008, 31). The ‘I’ve done what it takes’ language suggests a sense of sacrifice and heroically tough decision making. It would be hard to imagine Bush saying he had ‘done what it takes’ if he had approved only traditional, rapport-building methods. No proscriptive norm, no sacrifice.
Sometimes the target audience for political leaders selling heroic narratives may be themselves. In the quote above, Bush says, ‘I can say to myself, “I’ve done what it takes to do my duty…”’ Though it is not clear whether Bush really had this conversation with himself, it is certainly possible. Another plausible if unconfirmed example comes from one of Cofer Black’s private exchanges. Regarding the CIA’s ambitious assassination, detention, and interrogation plans, Black’s colleague Tyler Drumheller reports that Black said to him, ‘We’ll probably all be prosecuted’. Drumheller apparently ‘thought that Black practically relished the possibility, casting himself as a tough but noble hero, forced to sacrifice himself for his country. ‘He was giving his Jack Nicholson, A Few Good Men speech’, Drumheller recalled’ (Mayer 2008, 41).
One final note about causation here: the claim that the norm against torture can attract violators does not mean that the absence of the norm would result in less torture. Rather, torture would occur by a different route. I aim to account for torture in the world in which we live. A counterfactual world would require a different story even if it ended up in the same place.
Perverse effects and other explanations
Rather than test the perverse effects argument against competing hypotheses, this section explores how the perverse effects propositions can work with other explanations to provide a fuller analysis.
Leaders may draw on intuition to inform their belief that torture works. Rejali (Reference Rejali2007a, 449) calls this ‘folklore about pain’: the belief that ‘all people avoid pain and seek pleasure, more injury produces more pain, and so it is simply a matter of calibrating the quantity of pain for each individual’ in order to get the individual to divulge valuable information. Ruth Blakeley (Reference Blakeley2007, 376) echoes Rejali’s argument and takes a step toward my own when she explains that people believe torture is effective because of its ‘horribleness’. If the need for intelligence is strong enough, leaders need only believe that torture might work.
Some intuitive belief in torture’s efficacy may be necessary in order for the torture norm to have the perverse effects suggested above. Psychological reactance theory, which seeks to explain why people are attracted to forbidden fruit, holds that actors must see utility in the proscribed behavior in order to find it attractive. According to reactance theory, people seek to (re)establish behavioral freedoms that have been taken away from them. Expected utility provides a baseline interest, and the forbidden nature of a given act makes it even more attractive (Brehm Reference Brehm1966).
Examples from the war on terror show that the intuitive argument and the unintended effects accounts are complementary. The great need for intelligence is what prompted the Bush administration to look for outside-the-box solutions. Part of the Bush administration’s intuitive understanding of interrogation was that harsher meant more effective. The heroic narrative also dovetails the intuitive justification. Heroes (including false ones) are known for saying, ‘I was just doing what anyone else (who loves our country and has the courage to protect it) would do’, an appeal that Bush and Rice have aired.Footnote 17
Alexander Downes’ (Reference Downes2011) argument about how states target civilians in war when they become desperate might be applied to torture as well: states turn to torture when they become sufficiently desperate for intelligence. As with the intuitive argument above, the desperation argument has rational underpinnings. The main argument about norms presented here can be thought of as an extension of the desperation hypothesis. It gives the desperation argument more direction; by itself, the desperation argument is unclear as to what desperation will actually produce. For instance, we can imagine a desperate leader in an intelligence crisis calling around to different agencies looking for the interrogators most experienced in the non-coercive techniques that Matthew Alexander used to track down Abu Musab al-Zarqawi. The perverse effects argument shows why desperation might also get leaders to ‘think the unthinkable’. It may be hard to believe that desperate times call for the vigorous pursuit of traditional measures.
States also may use torture if coercive interrogation is engrained in an organization’s culture. Closed organizations such as intelligence agencies and military organizations develop their own set of ideas, beliefs, and norms, which frame their approach to certain tasks (Kier Reference Kier1997). Their singularity of focus gives them the image of possessing expertise in their fields, but it also limits their ability to see their actions in a larger context (Sagan Reference Sagan1994). Over time, the norms of closed institutions may contradict those of the international system, including norms that the institutions’ countries have internalized. The secretive nature of these organizations hampers international norm penetration, insulates the organizations from the benefits of the marketplace of ideas, and makes state coordination difficult. The result is ‘decisions’ (or rather, ‘output’) that may not reflect international sensibilities or work in the state’s overall best interest. Torture may be a part of that output. Despite damning evidence, agencies may press on with torture as a matter of tradition and a lack of organizational learning.Footnote 18
The perverse effects thesis can complement organizational culture theories in at least four ways. First, members of an organization may be attracted to the wrongness of torture just like other political actors, potentially shaping the organization’s culture over time. The example of Cofer Black speaks to this possibility. Second, administrations that value coercive techniques may foster competition between agencies and award the most aggressive among them, as the Bush administration did with the CIA at the beginning of the war on terror.
Third, executives lacking an organization that sufficiently values lawlessness may create a new agency whose principle purpose is to be aggressive in the face of proscriptive norms and laws. Donald Rumsfeld created the Special Access Program (SAP) to conduct secret operations and harsh interrogations because he thought the military leadership was too cautious; members of SAP were behind some of the abuses at Abu Ghraib prison in Iraq (Hersh Reference Hersh2004). Given these beginnings, it is hard to imagine a culture developing within SAP that did not condone torture.
Fourth, an administration looking to shape an organization’s culture toward more coercion can promote (or demote) individual agents that comply (or do not comply) with their overseers’ wishes. General Miller, who presided over harsh interrogations at Guantanamo Bay, was transferred to Iraq to ‘Guitmo-ize’ the Abu Ghraib prison (Priest and Stephens Reference Priest and Stephens2004) – that is, to bring torture techniques from Guantanamo to Abu Ghraib. Like the Bush administration, Miller derived the effectiveness of techniques from their proximity to the laws, as suggested by the first perverse effects proposition. Regarding the use of extreme temperatures for interrogation, Miller once declared, ‘If the Torture Statute says 80 degrees is bad, we will set the thermometer at 79.9 degrees’ (Mayer 2008). For Miller, ‘bad’ was good. Evidently, the Bush administration liked this approach.
Rejali (Reference Rejali2007b) argues that modern men may choose to torture in an attempt to establish or recover a sense of manliness. Civil society has made men weak, so the story goes; men use torture to overcome this shortcoming. His argument resembles the perverse effects explanations in two main ways. First, the fear of weakness stemming from societal rules drives the turn to torture in both explanations. Second, the recovery of manhood can be an implicit part of the heroic story that torturers tell themselves and others. In Cheney’s appearance on Meet the Press and Cofer Black’s quote about his own prosecution, bravado and machismo overlap considerably.
Rejali’s concern with idealized forms of masculinity echoes work by feminist international relations scholar J. Ann Tickner. Tickner (Reference Tickner1992) argues that the study and practice of international politics privileges viewpoints based on constructed ideals of masculinity. Chief among the celebrated attributes of leaders and states are autonomy and aggression. These characteristics connect feminist IR work to the perverse effects of the norm against torture in at least three ways. First, autonomy suggests separation from agreed-upon standards. Norms are inherently social, but autonomy is inherently asocial. If autonomy is perceived to be advantageous, part of the advantage might come from the freedom to act unconstrained by socially constructed rules. Second, the ‘promise’ of aggression features prominently in the nasty-equals-effective argument. Third, constructed ideals of masculinity contribute to how some define ‘heroism’ just as the norm against torture does.
More perverse effects
Norms may contribute to the behavior that they proscribe in other ways as well; I briefly explore three possibilities in this section. Although I stick with the example of the norm against torture, these effects may also apply to other norms.
The norm against torture may make it easier for torture apologists to dismiss sources of contradictory evidence. The norm gives advocates of torture a convenient reason to reject the arguments of those who say that torture does not work: torture’s detractors, the apologists might say to themselves, are too squeamish. The critics are convinced that torture is wrong, so they are claiming that torture also does not work so that they can avoid tough choices and wrenching trade-offs. This may be linked to the heroic narrative: the apologist convinces himself that, unlike the weak critics, he has the courage to do the difficult, ‘right’ thing.
The norm against torture might also cause political leaders to overestimate support for torture. Politicians may figure that members of the public are too timid to disclose their support for such a grisly and forbidden act, and thus conclude that poll numbers are artificially deflated. This is a reasonable conclusion from Gronke et al.’s (Reference Worth, Mazzetti and Shane2010) survey of public opinion polls on torture taken over the previous 8 years. Gronke et al. found that the public approved of torture and harsh interrogations drops when specific techniques are mentioned. Rather than being a more accurate measure of support (as Gronke et al. claim), poll readers in positions of power might conclude that a substantial portion of the population wants their government to use torture against detainees, but the public shies away from dirtying its hands in the details of techniques.Footnote 19 A politician could decide, ‘That’s why they hired me’.Footnote 20
Similarly, soldiers and bureaucrats may overestimate their superiors’ desire to use torture. Those in charge of interrogation may assume that their superiors cannot simply say, ‘Go torture them’, because of the laws and norms against torture. The result may be that the interrogator discounts evidence that his bosses want him to restrain himself, assuming that they want to maintain some plausible deniability. The expectation of rules-bound language leads interrogators to ‘read between the lines’.
The effects suggested in this section and elsewhere in this paper do not constitute an exhaustive list. They demonstrate the next steps – but certainly not the last – in exploring the full effects of a proscriptive norm like the one against torture.
Conclusion: limits and generalizability
Constructivists argue that norms shape and define interests, identities, and categories, but they have yet to fully explore the various ways in which this is so. The standard account suggests that norms have an impact when groups and individuals follow the norms’ script. Yet norms may motivate a whole range of unanticipated behavior, and the same norm may indicate different things for different actors. I have argued here that proscriptive norms can attract actors to the very acts that the norm prohibits, but there could be many more roads from norms to the acts they proscribe. The example of torture, while possessing the analytically useful attribute of not directly benefiting the political leaders who approve it, may be only one instance of many. Analyzing the ‘irregular’ effects of other norms could be a fruitful area of further research.
One limitation of the explanations presented here comes right out of the example I’ve used, the US war on terror. Most detainee abuse ended after 2005 (Hutchinson et al. Reference Hutchinson, Jones, D’Alemberte, Gushee, al-Hibri, Irvine, Kennedy, Pickering, Sessions and Thomson2013, 7). Yet beliefs about the utility or importance of off-limits practices, including torture, probably did not change. Bush and Cheney continued to trumpet the efficacy of ‘enhanced interrogation techniques’ after the administration was out of office. This may be the nature of explaining events and behavior that have ‘so many causes’ (Jervis Reference Jervis1988, 675): one thing causes a series of events (in this case, torture) to occur, and something else ends it. Still, the inability to explain the variation in the war on terror shows the limits of my argument’s explanatory power.
Another limitation of these perverse effects explanations is that they currently do not specify at the outset which individuals will hold the beliefs about cheating and heroism that help determine how the norm will affect perceptions and interests. The fact that the same individual might change beliefs based on certain circumstances complicates matters further. It should still be possible to hypothesize which kinds of individuals and what circumstances make certain perverse effects of a given norm more likely. It might even be possible to test these hypotheses, though comprehensive testing could prove difficult specifically with data-scarce cases of torture.
In addition to individual-level variation, future work on irregular effects of norms could tackle the effect of different norms. Which norms are more likely to have the perverse effects outlined here? While this paper leaves a fuller investigation to a future project, I offer a few points of departure. First, and perhaps most obviously, norms will not lead to their own violation if the norm is so robust and deeply internalized that thoughts of violation do not even occur to the actor. For instance, the norm against cannibalism is so robust in the United States that US politicians and generals would probably never even think of violation of the norm as a policy, no matter how intimidating to enemies the policy might prove. But if a person can ‘think the unthinkable’ – as with torture – then it is not really unthinkable.
Second, norms that prohibit harsh acts whose efficacy is debated may open the door for alternative means of determining effectiveness. The norm against targeted killings is one possible example.Footnote 21 Because the effectiveness of targeted killing is not well established,Footnote 22 some may use the harshness as an indicator. They could also frame norm violations as heroic and brave, necessary under the circumstances; this would be a much more difficult story to sell if almost no one thought the behavior was effective. Importantly, targeted killing may actually be an effective tool of war; the perverse effect of the norm does not simply explain blunders, and hunches about the effectiveness of cheating may in fact be correct.
Third, ethical norms preventing acts that are almost guaranteed to harm someone physically,Footnote 23 emotionally, psychologically, or in other ways may be more likely to be susceptible to the perverse effects I describe. One possible example is civilian victimization. Civilian victimization clearly causes harm, and some may expect such pain to yield gain in difficult circumstances (such as those faced by Churchill below). Likewise, the choice to target civilians requires a perverse ‘heroism’: that actors contain within themselves the tension between what is decent and what ‘must be done’.Footnote 24 By contrast, consider the norm of creating states with similar features (see Meyer et al. Reference Meyer, Boli, Thomas and Ramirez1997). It is more difficult to imagine the founders of a new nation-state believing that those who cheat on the standard form of state-building have an advantage, and it would be odd if an actor tried to sell the move as heroic.
Under which circumstances might these perverse effects occur? I offer here another non-exhaustive list. First, moments that citizens, civilian leaders, or the military determine to be critical to the life of the state might push key individuals to see norms as a fence dividing them from more effective means. Walzer called these ‘supreme emergencies’, a phrase used by Winston Churchill in 1939. Walzer (Reference Walzer2000, 251) argued that the phrase ‘contains an argument: that there is a fear beyond the ordinary fearfulness…of war, and a danger to which that fear corresponds, and that this fear and danger may well require exactly those measures which the war convention bars’. Having the courage (or ‘courage’) to reach for extraordinary means at extraordinary times may be framed as heroic.
The second set of circumstances bears some similarities to the first. The perverse effects of norms I describe become more likely if events persuade actors that the world is a more dangerous place than they previously believed. The September 11 attacks were just such an event. The Bush administration justified ‘new thinking’ regarding detainee treatment by arguing that the war on terror was ‘new’ and the dangers faced were ‘unprecedented’ (Jackson Reference Jackson2007, 356–57). This squares with work in political psychology about how surprising events can amplify emotional responses and cause preference shifts and reversals (Mercer Reference Mercer2013).
Third, if an enemy uses ‘unscrupulous’ means in war, a political actor might be convinced that his or her country is now engaged in a ‘dirty war’ in which such means are necessary. This is especially true if the enemy lands a punch or two while operating outside of war’s laws and norms, causing the newly bruised to consider the possibility that rule-following would put them at a disadvantage. Having an enemy that does not ‘fight fair’ also weakens the argument that one’s side must stick with the laws and norms of war in order to engender the same treatment from the other side.
While I do not attempt to specify the generalizability of my argument to certain individual actor profiles, this article makes possible a broader comment about the significance of norms. Norm violators constitute a hard case for the impact of norms.Footnote 25 If norms can influence even their violators, then the scope of normative impact may be much wider than is commonly understood. For most actors, we would do well to ask not if norms influence their behavior, but how.
Acknowledgments
The author thanks Elizabeth Kier, Jamie Mayerfeld, Jonathan Mercer, the IT editors, and the reviewers for their close reading and extensive comments. Christopher Adolph, Daniel Berliner, Kendra Dupuy, and Stephan Hamberg provided helpful insights on earlier versions of this paper and theoretically similar ones.
Appendix: torture’s bad record
I give here a brief defense of the claim that torture does not compare well to other interrogation techniques.Footnote 26 I do not think my main theory strictly depends on this argument. Even if torture reliably yields valuable information, this does not imply that it was chosen for functionalist or rational reasons. Discovering how leaders arrive at the decision to torture remains important. Still, the assertion that torture does not work makes the puzzle more ‘puzzling’, and may help some readers set aside the ‘torture works’ ghosts that could haunt their evaluation of my claims. I offer three reasons to believe that torture is usually an ineffective interrogation technique, and certainly worse than other methods used by professional interrogators.
First, to my knowledge, there are no social scientific studies that document evidence of torture’s efficacy.Footnote 27 Researchers showing how torture does not work appear as boxers fighting phantoms, with no academic schools of thought on which they can land their punches. Darius Rejali’s account of whether torture works is the most thorough, yet even in his chapter titled ‘What the Apologists Say’, he cites no social scientists extolling torture’s benefits. Alan Dershowitz (Reference Dershowitz2002) and Bruce Hoffman (Reference Hoffman2002) come closest to being social-scientific defenders of torture, but their examples of torture’s successes are isolated, disputed, and not systematically compared with the effectiveness of other techniques.
Second, arguments for torture lack details about causal mechanisms that would plausibly link torture to the consistent extraction of accurate information. When data are fairly scarce, as with torture, evaluators must rely more heavily on the theory behind the act, especially the causal mechanisms that purportedly link cause and effect. We have no reason, however, to believe in the absence of evidence that torture evokes honesty in its victims; no theory has outlined the transmission belt connecting pain to the truthful disclosure of secrets. The intuitive argument – that prisoners will talk to end their torment – does not spell out why the talk will be truthful.
More sophisticated torture theories also lack the link to truth-telling. For instance, Martin Seligman, who developed of the concept of ‘learned helplessness’ that others adopted and used as a torture technique in the US war on terror, said: ‘I think helplessness would make someone more dependent, less defiant and more compliant, but I do not think it would lead reliably to more truth-telling’ (Shane and Mazzetti Reference Shane and Mazzetti2009). Even torture manuals admit that torture is likely to yield a lot of false information.Footnote 28 One might argue that non-coercive methods do not guarantee true information, and this is correct. But methods such as rapport-building do not make detainees desperate, and do not force innocent people to talk. Moreover, scientists have linked pain and torture with memory loss (O’Mara Reference O’Mara2009), feelings of guilt among the guiltless (Gray and Wegner Reference Gray and Wegner2010), and corruption of agencies and societies (Rejali Reference Rejali2007a). We can expect comparatively lower rates of these adverse effects under non-coercive methods. Thus, torture almost always compares poorly with other methods, and is often worse than doing nothing at all.
Third, considering the amount of torture practiced in history, the lack of evidence challenges torture’s apologists more than its critics. This is because torturers hold the evidence; they are more likely to hide the results of failure than success. Of course, there are many reasons to hide torture from the world regardless of outcome, including issues of legality and the desire to refrain from giving away secrets to the enemy. Even when knowledge of the occurrence of torture becomes public, however, the evidence of efficacy still remains hidden, raising suspicion that governments and agencies have embarrassing findings to hide. If the evidence that is publicly available is the most favorable to torturers, then the defenders of torture have no argument.Footnote 29 The results of the CIA’s Phoenix Program in South Vietnam, for instance, have been disclosed, and they do not speak well of the promise of torture for producing accurate intelligence (Rejali Reference Rejali2007a; Kalyvas and Kocher Reference Kalyvas and Kocher2009).
The point here is not that torture never produces intelligence; rather, all signs suggest that it is at best a clumsy method that carries great risks with little upside. For the example of the US War on Terror that this article has frequently invoked, arguments against torture were known in policy circles as well as academia. A Pentagon adviser says of the Bush administration, ‘They did it the wrong way, and took a heavy-handed approach based on coercion, instead of persuasion – which actually has a much better track record. [Torture] is evil, but it’s also stupid’ (Hersh Reference Hersh2004). The same argument can be found among journalists, professional interrogators in the FBI and CIA, and other sources.Footnote 30 The Bush administration was not moved to use torture for intelligence because ‘it’s hard to argue with success’.
