The futures of anticipatory reason: Contingency and speculation in the sting operation

Introduction

The future does not exist. This basic and obvious fact gives it a special strategic function. The future does not denote events that ‘will’ or ‘may’ happen at a later time. Rather, the future serves as an inexhaustible source of unknowns that is leveraged to smuggle in truths, beliefs, claims, that might be too speculative, too unfounded, to circulate by regular means. We argue that the increasingly prominent invocations of the future in contemporary security practices constitute exactly this strategic and speculative kind of truthmaking. Synthesizing and extending recent scholarship on preemption and anticipation, we locate the significance of future-oriented security not in its ‘accuracy’ in relation to temporally distant events, but in its ability to legitimate a wider range of truth-claims. Insofar as a rationality is defined by social norms for what is sayable and knowable, the future is a strategic technique for manipulating these margins of knowability and provability. The political and moral significance of this future lies not in the fulfilment of prediction, but in the concrete actions that are taken in the present in the name of prediction.

To describe this future at work, we use the metaphor of the epistemic black market. Here, questionable truth-claims are laundered into relatively legitimate and operationalizable forms. Drawing on theories of contingency and potentiality, this article seeks to de-emphasize the temporal and predictive imagination, where the future is said to exist as an objective array of possible events and thus can be ‘known’ through probabilistic calculation. What is posed as a way to predict the future engages instead in a process of creative production, endowing a wider range of speculative truth-claims with a degree of epistemic legitimacy and affective impact. Yet, as we will show, this ‘flexibility’ in truthmaking brings neither an equalizing of power relations nor anarchic disorganization. Instead, these techniques tend towards a conservative retrenchment of familiar patterns of control.

Our argument builds on the lessons of what might be called ‘anticipatory reason’: the various techniques of prediction, preemption and so many variations of pre- that have become a dominant attitude to knowledge-making (see, for example, Adey and Anderson, 2012; Amoore, 2011; Aradau and Van Munster, 2012; Harcourt, 2007; Massumi, 2015a; Samimian-Darash, 2013). This interest in extracting actionable insights from projected futures has been particularly prominent in practices of security – as covered extensively in this journal in recent years. The justificatory backdrop here is the sentiment that threats like global terrorism are contributing to an increasingly unpredictable world. In response, security organizes the world into an array of calculable events and aspires to an ever-greater efficiency in the ratio of what is possible to what is anticipated. This article contributes a more nuanced and strategic understanding of the future. We argue that the dominant temporal directionality of prediction functions as justificatory cover for state power – power that is enacted through unproven and unprovable judgments about crime and terror, risk and probability.

Below, we unfold this view of the future by putting recent scholarship in anticipatory reason and security together with resonant analyses of the future as contingent and strategic. The latter include sociological theories of a postwar ‘hollowed out’ or ‘presentist’ future, as well as comparable observations about futures as tradeable speculations in financialization and data analytics. To ground the discussion, we turn primarily to the ongoing expansion of ‘sting operations’ in post-9/11 US counter-terrorism practice. We focus on three cases: (1) the Newburgh Four, a 2009 sting in which four poor, black ex-cons were lured into a plot by an informant; (2) the Fort Dix Five, where a group consisting of an Egyptian American, three Albanian brothers and a Turkish-American were convicted of conspiracy to murder US military personnel in 2008; and (3) Sami Osmakac, an Albanian American, again coaxed by an informant into preparing a violent attack in 2012.

The article begins with a conceptual overview of the future, then examines three interconnected aspects of the epistemic black market: (1) The future serves as a flexible and inexhaustible source of contingency, rather than a discrete temporal entity. What is often sold as the probabilistic calculation of future events involves creative production of speculations. (2) Invocations of the future have a laundering effect, in which speculations are produced into relatively legitimate, calculable and actionable things. Such unofficial activity loosens the thresholds of proof and reason that must be met to justify actions like arrests and monitoring, allowing incumbent power relations to be exercised with greater impunity. (3) The value of this future is counted not in its ability to predict a given event with accuracy, but in its ability to produce tradeable informational goods. This value can be counted in the securitization of populations as much as direct monetary profit. It is this strategic utility that is central to the securitized future, even as it establishes itself via the insistent fantasy of predicting the temporal beyond.

What is the future?

The idea of the future has a history, one that has been recently traced in the notions of ‘conceptual history’ (Koselleck, 2002, 2004) and ‘regimes of historicity’ (Hartog, 2003). Within this body of work, there is a periodization of the future (divided into three eras) that illustrates well the spatial and temporal metaphors used to articulate the future – and their limits. The premodern epoch (from Antiquity to the 18th century) was dominated by an idealized past; the future was, in effect, a return to this past. Set within ritualistic and theological understandings of circular or eschatological time, the future is past and preordained. Notably, throughout this period, utopia was largely articulated in spatial terms, such as Thomas More’s ‘non-place’ (see Hölscher, 1996).

In this framework, the modern future begins with the French Revolution. Indeed, ‘the decade from 1789 to 1799 was experienced by the participants as the start of a future that had never before existed’ (Koselleck, 2004: 59). Koselleck retraces this transition through the etymological shift of the concept of ‘revolution’. From signifying rotation and return, it came to mark ‘a point of departure’, one of upheaval and change; thus, a once predestined horizon was opened to human manipulation (see also Burke, 2010). This is the future implicit in the modern idea of progress as a universal and inevitable trajectory towards civilizational improvement, itself an idea that only comes together fully over the 18th and 19th centuries (Bury, 1920). This liberated future is unmoored from space and set free into the realm of temporality (Carey and Quirk, 2009: 151). The notion of utopia is also temporalized (Hölscher, 1996), placed into a glowing future that is asserted as both fact and imperative (Allen, 2016).

From the mid-20th century onward, however, that future would be ‘lost’ (see Burke, 2010; Hartog, 2003). Faith in the ‘liberated future’ in the form of modernity’s progress took major blows in the aftermath of World War I (Hölscher, 2013) and the genocidal and atomic ash of World War II. If, for modernity, the horizon was ‘that line behind which a new space of experience [would] open’ (Koselleck, 2004: 260–261), after World War II futurism ‘sank below the horizon and presentism has taken its place. We cannot see beyond it’ (Hartog, 2003: 113). The future is felt no longer as promise, but as threat: ‘the future is a time of disasters, and ones we have, moreover, brought upon ourselves’ (Hartog, 2003: xviii).

We find a resonant argument in risk studies. The idea goes that, from global terrorism to ecological and epidemiological threats, we are exposed to more diffuse and uncertain dangers than ever. While there is a split between realist assertions that the world really has become more dangerous in an objective sense ¹ and more constructivist theories that what socially ‘counts as’ risk has changed (see, for example, Dean, 1998; Ewald, 1993), both strands locate the future as a source of emergent dangers that are defined by their corrosion of existing social order (e.g. Beck, 2009). In its ostensible aim to safeguard society from what comes from beyond the horizon, security in a risk society thus contributes to a sense of an ‘extended present’ (Nowotny et al., 2001) that is ‘permanent, elusive, and almost immobile’ (Hartog, 2003: 17).

To be sure, these narratives of presentism, or even ‘lost’ futures, are partially a result of the limitations of the spatial and temporal metaphors that used to make sense of it. We do not contend that the future was lost in any absolute sense, as if what was once a natural horizon has been replaced by an artificial one. Even in Hartog’s (2003: 113) world, one ‘so enslaved to the present’, pasts and futures required for the present are constantly ‘fabricated’. This aligns partially with the Whiteheadian sense of ‘fabulation’ used by Massumi (2015a) to analyse contemporary security, and Bergsonian ‘fabrication and retrojection’, which Ayache (2015) has applied to financialization. The future does not exist in an objective sense, but is instead created in and for the now. Even as the temporal imagination endures, these approaches emphasize the increasing prominence of the future as uncertainty, and as a locus of strategy.

This emphasis on strategy is, of course, latent in recent scholarship into anticipatory reason. For instance, Brian Massumi focuses on the future’s proto- or non-cognitive epistemology, which moves us through preconscious prehensions that take the form of affective ‘loomings’ (e.g. Massumi, 2015a: 202–203). Here, the informational content is typically a ‘diversionary’ set of indefinite material, whose effect is counted in leveraging affect out of potential. Complementing this analysis, we approach truth-claims about the future as neither predictions nor shade-throwing, but a more ambivalent move. A certain degree of collective commitment to the idea of real prediction, real future knowledge, is central to the engineering of affects and projected outcomes. Consider the ‘next terrorist attack’ (NTA) (Aradau and Van Munster, 2012), a paradigmatic figuration of the evolving dangers of 21st-century terrorism and the injunction to predict embraced by security institutions:

Moments after five men planted two bombs – inert and provided by an agent provocateur – in the empty parking lots of two Riverdale, NY synagogues, a mass of FBI agents descended onto the scene; made for TV, the entourage included a superfluous bomb squad. In the aftermath, Republican Peter King told reporters, ‘This [the Newburgh Four] was a very serious threat that could have cost many lives, many lives if it had gone through. It would have been a horrible damaging tragedy’. (Baker and Hernández, 2009)

The idea of the next terrorist attack shapes the Congressman’s subjunctive warning. It is a strategy for making visible, making present, amplifying the affective force of something that remains amorphous and unknown. The next attack is a prominent example of anticipatory reason. In the case of security and surveillance, this involves an umbrella of related concepts and phenomena, from computer simulations and physically enacted ‘scenarios’ as dress rehearsals of the future (Adey and Anderson, 2012) to the dissection of bodies into data-driven correlations ‘indifferent to the actual occurrence’ of the threat (Amoore, 2011). Running throughout is a common imperative: that a threat be identified – or, rather, produced – before it arrives into the present.

Crucially, the potential dangers of the next attack are exemplified, but never exhausted, by a specific case such as the Newburgh Four. The very claim to have preempted a specific ‘next attack’ supplies the rationale for ever-more-expansive and costly surveillance systems to catch the innumerable other next attacks that must be out there. Each specific attack is considered to capture important aspects of this broader pattern, but that pattern always reserves further surprises for its next victims. Speaking of the anthropocene, Timothy Morton (2013) offers the term hyperobject: something that lies beyond the horizon of the present and human knowability, but is glimpsed through myriad local instantiations. The polar bear may emblematize the problem of species extinction, even as it is understood that the broader object of climate change exceeds every such representation. In the same way, every near miss, preemptive arrest and even actualized attack is understood as a non-exhaustive instantiation of the future that remains a repository for indefinite uncertainties. In a probabilistic and temporal framework, one might expect that this hyperobjective horizon (NTA1), upon investigation, is revealed to consist of a set of specific threats (NTA2) that can then be preempted out of existence. But the problem is that no finite set of threats exhausts the general possibility of the next attack. It is the indefinite availability of NTA1 that is reaffirmed through periodic identifications of NTA2, fuelling the equally indefinite expansion of surveillance apparatuses. This is our departure point for a strategic understanding of the future in anticipatory reason: the atemporal future as a productive source of contingency.

From probability to contingency

Applications of anticipatory reason in security are diverse. If more traditional approaches focused on the institutionalized production of ‘actuarial reason’ (Harcourt, 2007), anticipatory reason casts its net across a wider range of more amorphous futures. We find its popular lexicon in Donald Rumsfeld’s infamous typology of ‘known unknowns’ and ‘unknown unknowns’ (Daase and Kessler, 2007; Hannah, 2010), as well as Philip K. Dick’s mid-century coinage of pre-crime (see, for example, Burris, 2016). Anticipatory reason hits a widely shared nerve across the various efforts to manage future outcomes, within and beyond the domain of security.

Across these diverse contexts, the emphasis on anticipation often fuels a kind of arms race: the time of decision shrinks, even as the time of possibility widens. An ever more expansive array of not-yet-realized threats are claimed for securitization, extending to potentialities that are not only temporally but causally remote. ² The capture of ‘possible uncertainties’, which are expected to resemble past events, extends now to ‘potential uncertainties’ whose form remains unclear even as they are acted against (Samimian-Darash, 2013). These diffuse and amorphous threats are argued to require ever-faster and earlier intervention. In war zones, drones – iconized by their ‘unmanned’ status, but in reality requiring a complex and distributed network of analysts, pilots, sensor operators and more (Elish, 2017) – operate under a ‘rapid shrinking of the space and time for deliberation’ (Andrejevic, 2017: 879; see also Gregory, 2011: 196). There, speed is not a luxurious technological achievement, but a moral and natural necessity demanded by reality (Andrejevic, 2017: 881). Control must be established faster and earlier, even as the target of that control expands to an ever-broader array of future events.

The sting operation in many ways exemplifies this logic. As in the Newburgh Plot above, the public is told that the next terror attack would have happened if not for the work of law enforcement in ferreting out threats lurking within the population. While questions arise about how ‘real’ a threat was, authorities respond that they cannot afford to wait and see whether the threat materializes (Massumi, 2015a). This kind of anticipatory strategy has become a central part of post-9/11 US counter-terrorism. One report estimates that around 30% of counter-terrorism convictions between 2002 and 2011 were fabricated through stings (Human Rights Watch and Columbia Law School Human Rights Institute, 2014). While the full dataset remains publicly unavailable, The Intercept has compiled a partial list of some 300 individuals charged with terrorism-related offences on the basis of such covert operations since 11 September 2001 (Aaronson and Williams, n.d.). Insider officials such as Andrew Liepman, director of the National Counterterrorism Center from 2005 to 2012, have described an emergent cultural and moral imperative in the post-9/11 years: ‘zero failure, zero attack threshold’ (Barker, 2016; also see Aradau and Van Munster, 2007; Cooper, 2006). This ‘zero tolerance’ mantra requires security to continuously refine its ability to calculate and predict, to the point that it is expected to calculate the incalculable. Louise Amoore (2014: 424) cites the L’Aquila earthquake of 2009, where scientists and public officials were convicted and sentenced for ‘failing to predict’ a natural disaster that the prosecution themselves admitted was ‘incalculable’. Although most of the convictions were eventually overturned by the Italian Supreme Court, the episode remains a telling marker of shifting expectations around the work of anticipation and upon the experts tasked with it.

In all this, the relationship between prediction and futures is typically described in terms of probabilistic identification: the future may not be a predestined sequence of events, but it exists as an array of variously weighted possibilities, which can at least partially be calculated and managed through increasingly sophisticated apparatuses of security. But there is a certain incongruence between the urge to decide and the horizon of possibility; that is, between (1) the idea that security can and must know the future in a probabilistic and objective way, and (2) the idea that the future dangers we must securitize against are increasingly unknown and distributed. This tension demands that we think beyond discrete and linear temporalities. The ‘next terrorist attack’, as an amorphous and inexhaustible source of dangers, justifies and requires an equally diverse and ever-expanding suite of security practices based upon mechanisms that seek to define that future by extrapolating past instances. In the sting operation, this often takes the form of extracting weak indicators. One example here is the ‘Indicators of Mobilization to Violence’, a scoring system used by the US Federal Bureau of Investigation (FBI) since 2015 to identify individuals likely to go on to carry out a next attack (Currier, 2017). The survey searches for the earliest proto-emergence of a propensity for violence or religious extremism (which themselves are mere proxies for an actual terrorist attack), even down to the growing of beards or playing paintball. Although we do not know how extensively Indicators of Mobilization to Violence itself was used, such weak indicators often feature in known accounts of sting operations. For the Fort Dix Five, it was their recreational hobby of visiting gun ranges that led authorities to them. The men had dropped off a video of a vacation to be copied to DVD at a local Circuit City; in it, an employee saw men shooting guns and speaking in a foreign language, and thus called the authorities. For Sami Osmakac, it was his verbal tirades against democracy and Christianity that made him a readymade target for an informant-led sting (see Aaronson, 2015). For James Cromitie of the Newburgh Four, it was his supposed ‘Arabic accent’ that attracted the attention of the informant who had recently began phishing at the local mosque Cromitie sparsely attended (US v Cromitie et al., 2010: 677–678). These could be considered ‘weak indicators’ in that both activities – guns and speech – are, at least in the US context, constitutionally protected. What is made visible here is that any act of prediction is predicated on a presumed relationship between proxies: certain hints, indicators, are made to stand in for far broader sets of claims and possibilities, in order to close the gap between contingency and action.

Such creative and proactive assembling of futures (that, by definition, are never intended to be realized) is found at every stage of the sting operation. Once identified for intervention, the informant encourages the target(s) to expand on their controversial views – on Jews (Newburgh), US intervention in foreign affairs (Fort Dix) and democracy (Osmakac) – and captures them on tape; these tapes are spliced together at trial, often eliding the hundreds of hours of non-radical mundane babble. Second, the informant goads the target(s) into preparatory or conspiratorial action – such as downloading jihadist propaganda, doing reconnaissance, attaining maps and purchasing weapons – often by promising money (Newburgh) or by exploiting the personal or financial relationships they develop with their targets (Newburgh, Fort Dix, Osmakac). Third, the weapons are often bought from or through the informant. Moreover, the informant suggests particular types of weapons that have a cultural resonance with terror groups (particularly, the AK-47 and rocket-propelled grenades) – the weapons of terrorists in TV and film (see Szpunar, 2018).

One might argue that these extrapolations are a fundamentally accurate way to know the future, and the only question is the variance in the quality of the prediction. But the incongruence between the future as an open horizon of unknowns (which necessitates surveillance) and the future as a set of objective predictions remains. In particular, the premediatory work of the informant (Szpunar, 2017) suggests that we might instead think of these predictive actions as creative action. As we have mentioned, Brian Massumi (2015a: 207) deploys Whitehead to argue that both past and future are fabulated via long-range ‘forecasts and back-casts’, and our own position in doing so is always the lived, fleeting, momentary now of what Foucault calls ‘effective history’. The active fabrication of a future puts into motion yet another kind of reality. In the case of the sting operation, that other reality constitutes the prosecution and imprisonment of individuals that might once have been considered unremarkable.

Tasked with preventing a future’s actualization, the predictions created remain unverifiable. Agents in the Osmakac case spoke of him as a ‘retarded fool’ who needed the agency’s support to turn his ‘pipe-dream scenario’ into any semblance of a real plot/threat. They were able to capture Osmakac red-handed at the scene of the attack – an ideal outcome for the purposes of prosecution. But this ‘Hollywood ending’ (as the agents described it internally) was only possible because the agents paid for his cab there (Aaronson, 2015). The presiding judge of the Newburgh case was certain that ‘only the government could have made a terrorist out of Mr. Cromitie, a man whose buffoonery is positively Shakespearean in its scope’ (Harris, 2011). The informant in the Fort Dix case stated that four of the five accused were not in on, or aware of, a conspiratorial plot (Hussain and Ghalayini, 2015; see also Szpunar, 2018). The FBI’s material and psychological encouragement was essential for the targets to become ‘dangerous enough’ to be legally and operationally eligible for arrest. The fake bombs, the ghastly mugshots, the specificity of the targeted location, the subjunctive rhetoric – ‘it would have been a horrible damaging tragedy’ – stitch together a certain genre of speculated futures (also see Hong, 2015). Here, prediction is not fundamentally anchored to the actualization of the predicted event at a certain time; the relation is primarily strategic. Prediction happens in, and happens for, the ‘now’ of predicting.

Brian Massumi (2015a: 240) presents a formula for the conditional logic of preemption: ‘could have, would have, just as good as was’. The formula holds true here, while taking on a different epistemological inflection. The invocations of the future in these security practices entail a fabrication of contingency. The future stands not as an objective thing to be progressively unveiled and stabilized through probabilistic methods, but as a font of contingency that allows for flexible and dynamic utilizations. Consider Elie Ayache’s (2015) The Medium of Contingency, which contrasts probability and contingency in stark terms, overlapping partly with the ongoing philosophical interest in contingency ³ and with a particular interest in financial markets. Ayache firmly rejects the idea that the world might be governed through a structural calculus of relative possibilities unfolding on a linear temporal track. Neither stochastic variations nor neatly modelled patterns exist in any objective way; the market, so thoroughly saturated with such calculations, consists in fact of absolutely contingent and disruptive events – the event of naming a price for the purposes of trading (Ayache, 2015: 3, 25–26, 35–38). If one predicts rain tomorrow, all that exists is the world in the present and the actions taken vis-a-vis that prediction: no rain, only umbrellas (Ayache, 2015: 85). No terror attack, only FBI agents descending on the scene and men in handcuffs.

This view extends the recent scholarship in anticipatory reason and security to its logical conclusion: that we don’t just move in degrees from relatively known to relatively unknown future threats, but that the future was never quite ‘knowable’ to begin with, and that it has always been a byproduct of truthmaking mechanisms that convert this contingency into tradeable forms. In the constant reinvention of governmental and security practices, risk calculus and future conjectures are both part of the same wider toolset used not for predicting, but for creative manufacture of truth-claims that can then be traded and leveraged for the kinds of control that exceed immediate needs and justifications.

This focus on contingency takes us out of the linear temporal disposition and forces us to consider both a future and present as kinds of real, material actions that are always contingent. Ayache stresses that contingent actions in the market are not ‘potential’ or ‘probable’, actions whose validity and effect are to arrive at some future point; rather, the very act of making a prediction is itself a real action, whose strategic consequences play out immediately in the ensuing negotiations and recalculations (as evident in the unfolding and prosecution of the sting operation). The market price is thus a ‘present settler of future contingency’ (Ayache, 2015: 167). In this sense, the future of anticipatory reason operates outside temporality. The future does not exist, in the sense that the event that prediction describes exists neither as a temporally displaced event nor as something that is actualized. What does exist is the act of prediction in the present as creative action, something that takes advantage of contingency (i.e. the non-existence of the future as actualized events) in order to make a difference in the now. Importantly, the former view continues to be utilized for justificatory cover. The future as something around the corner, and as boundless possibility, provides unfalsifiable and inexhaustible source of justifications for securitization. The following sections build on this contingent and atemporal notion of the future, addressing how the future tense of contingency is used to launder a variety of marginal and speculative truth-claims in the service of security.

Selective speculations

The growing utilization of the future for creative production of speculation is consistent with the historical expansion in the remit of security. As Michael Dillon (2015) has shown, security as a technology of modern government entails the indefinite work of liberal governmentality: the constant rediscovery, adjustment and recapture of the most effective means for managing the productive freedom of biopolitical subjects. ⁴ As with theories on the ‘presentism’ of the future, there is a historical shift towards a sense of the future as dominated by finitude. Dillon describes Foucaultian governmentality as the closure of a teleological vision in which all things proceed inexorably towards a final moment, in favour of a ‘factical finitude’ in which government toils as ‘the indefinite management of many finite things’ (see Dillon, 2015: 7, 18–19, 34–35). In this sense, securitization involves the constant identification of new areas that might be charted out and prepared for social sorting. Crucially, the sorting work involved is never neutral (in the sense of the fantasy of objectivity) and always selective. If the theory of liberal governmentality can be read as a reactive strategy to capture and recapture the movements of ‘many finite things’ (Dillon, 2015: 117), this work of knowing also entails rearranging the world to make humans and things more knowable for the way the state knows – as exemplified in James Scott’s (1998) Seeing Like a State.

Back in the sting operation, the creative production of futures involves precisely this work of selective speculation. Informants were often asked to speculate on the stand in lieu of surveillance tapes, which they at times shut off or were instructed to shut off by their FBI handlers. In the Fort Dix case, the informant Mahmoud Omar was asked if he had heard the defendants discussing targets or plans among themselves. (There are plenty of recordings in which the primary defendant, Mohamed Shnewer, and Omar discuss plans, but this does not in itself constitute a conspiracy.) Omar would repeatedly respond, ‘Something happened, but it didn’t happen in front of me’ (US v Shnewer et al., 2008: 4296). Perhaps most astonishing are the speculations of the government expert witness, Evan Kohlmann. Kohlmann, who holds a law degree and independently compiled a database of jihadist propaganda, has become a mainstay in government prosecutions of this kind. In describing videos downloaded by Shnewer (at Omar’s behest) as some of the ‘most important films al-Qaeda has ever made’ (US v Shnewer et al., 2008: 5853), Kohlmann mused about what effect they might have had on the defendants. Speaking about a video entitled ‘The State of the Ummah’, Kohlmann claimed that ‘this video had a tremendous impact when it was first shown. The raw cuts were sent to al-Qaeda trainees at camps in Afghanistan. The trainees went crazy. They started screaming and yelling, people started crying’ (US v Shnewer et al., 2008: 5844–5845). Later, he speculates that despite the lack of evidence that Shnewer had any contact with any actual jihadists, one could not rule out such a possibility since the (now defunct) peer-to-peer file service from which Shnewer downloaded videos had a chat function. In the absence of any concrete evidence, Kohlmann does not lean towards the negative with regard to Shnewer’s potential contact with jihadists; instead, he utters a suggestive ‘I don’t know’ (US v Shnewer et al., 2008: 5909). Such claims were made to substantiate the potential future threat of the defendants – threats that become all the more important to prove now that they have been neutralized. These kinds of truth-claims operate not in a binary of truth and lies, information and mis-/disinformation, but in the flexible and ambiguous margin of the projected future. They can neither be proven nor disproven, and therein lies the utility.

This relationship of speculative truth-claims and the projected future characterizes the epistemic black market: a means to subvert and bypass the boundaries between actual and potential, proven and speculative. In speaking of a black market, we we aren’t necessarily referring to the familiar trope of flagrantly criminal trades, but something closer to what Foucault (2015) in The Punitive Society calls illegalisms: the extralegal, informal, tacit agreements by which participants profit from or cope with the demands of regulated and official relations. Crucially, these illegalisms are not simply a ghetto for the pernicious abuse of the legal; rather, they often handle functions that this particular legal system is not able to cover. The future, we suggest, is a vehicle for tolerated illegality in the production of truths – truths that then take the form of predictions, anticipations, visions, promises, and thereby enter collective processes for decision, action and sentiment. These production lines entail systematic, normalized relations between illegalisms and official activity, resulting in not so much transgression as resistance but a conservative retrenchment of state control.

For one, the future as leverage for illegalisms does not imply an equitable or anarchic kind of openness. At least in the case of security, it points to a more deregulated and flexible space for the state to enforce its ongoing interests. Consider a ‘stereotypical’ black market, as a physical site for the sale of all manner of irregular goods. Where regular markets cannot perform adequately, the various gaps in efficiency are often usefully filled by black markets. Thus underground economies in North Korea grew in times of economic stress, such as the widespread famines in the 1990s dubbed the Arduous March. The Kim regime would tacitly support this development, reserving the right to clamp down on it when the tides turn for the better. ⁵ This loosening and tightening of the rules of trade – whether in illicit foodstuffs or truth-claims – thus occurs in intimate relation to the state’s strategic needs.

More specifically for truth-claims, we might consider leaks as a paradigmatic example. As David Pozen (2014) has shown, the US government has long ‘leaked like a sieve’. Laws such as the Espionage Act exist that forbid such leaking, but enforcement is minimal. This does not mean any kind of leaks by anyone will do. Few leaks are genuinely transgressive, in which case the state makes extraordinary effort to create a chilling effect; consider the violence exerted on Chelsea Manning and Reality Winner, and attempted in relation to Edward Snowden. More common are neither leaks nor plants but ‘pleaks’: tacitly tolerated leakage that supplies journalists with information and gives insiders a path to air discontent, and the state itself to test and monitor public opinion. Pleaks are not mere errors in the system, but serve important and routine functions. Such illegalistically produced information is ambiguous; it carries a sense of veracity, but is also deniable and doubtable. It is this duality that makes pleaks tolerable and even useful for the state. What the future as a black market does is to filter information through this process of rendering it ambiguous, able to be bent into more hard-nosed calculations and speculative forays as suits the wielder.

In sting operations, one key source of illegalistic manoeuvring is the informant. In both the Newburgh and Fort Dix stings, the key informants had entered the US illegally, and in the latter case had committed crimes during their work with the FBI (US v Shnewer et al., 2008: 2892). Officials continue to defend these preemptive sting ops, arguing that it is difficult if not counterproductive to find ‘clean’ informants. Such illegalistic (not necessarily illegal) practices illustrate the importance of speculative truths and activities for the law. Consider that most defendants who have not yet carried out – and had no chance of carrying out – an attack are charged with conspiracy to do so. Conspiracy law in the United States is complex, but the boundaries are left purposefully loose. There are no definitive limits on the duration or time of a conspiracy, with the result that the attack need not be imminent, merely potentially emergent/contingent. The geographical scope and even the target of a plot can also be left relatively open. Moreover, tacit agreement suffices to convict, and so participation is determined by inferences drawn from the facts of the case (Siesseger, 2004; see also Chesney, 2009). It is precisely because conspiracy is a future-oriented charge that such inferences are possible. Finally, an entrapment defence only further opens the door to evidence of predisposition. Asserting that one’s actions were induced by a government agent permits the prosecution to introduce evidence of predisposition (i.e. character, conjectural past, etc.) that would otherwise be considered inadmissible (see Gantar, 2015; Laguardia, 2013; Norris, 2015; Roth, 2014; Sherman, 2009). Here we find a highly selective mobilization of ambiguity vis-a-vis the future – a set of ‘illegalistic’ strategies surrounding the application of the law that provides greater flexibility for the state at the expense of the accused.

Such selectivity in future-oriented truthmaking is often concealed via the guise of extrapolation. Here we find a parallel to the contemporaneous rise of data analytics, which shares the interest in more aggressive and early predictions (and, indeed, often participates in the revamping of security practice in areas like signals intelligence and predictive policing). The most common approaches, such as those built on Bayesian methods, constantly feed past data to extrapolate future events and fold each result (however correct or incorrect) into the next prediction. In this way, a particular imagination of probabilistic futures and objective predictions is constructed for use. Yet, as critical data studies has repeatedly shown, such calculations are often driven by a practical set of interests (rather than, say, malicious and purposeful bias) that introduce structural selectivity into the process (e.g. Cheney-Lippold, 2017; Hong, 2020; Mackenzie, 2015; Rieder, 2016).

The 21st century ‘war on terror’ presses these futures in the service of highly differentiated strategies for the distribution of bodies and threats. This selectivity is evident in the racialization of the notion of terrorism (Volpp, 2002). It is not by circumstance that an informant chose to phish at a mosque located in a poor city (Newburgh) frequented by African Americans. Thus Stephen Paddock, responsible for the most deadly mass shooting in modern US history in Las Vegas, was not even officially designated a terrorist, despite conforming almost exactly to the legal definition of the term in the state of Nevada. In addition to racial differentiation, the ‘war on terror’ has also entailed a systematic production of a broken perverted subject, a monster that requires a collective response in the form of ‘aggressive heterosexual patriotism’ (Puar and Rai, 2002). In this context, the increasing numbers of sting or sting-like operations, and their focus on a particular subset of the population (brown, male, low socio-economic status, and so on), must be understood as a highly selective ordering of future possibilities for different parts of the population. Here we return to the productive tension of the ‘next terrorist attack’: what is ostensibly an objective effort to capture a changing and indefinite future ends up reprising the past in a ritual repetition of the last terrorist attack.

In all this, the body of the incarcerated ‘would-be’ terrorist serves as the material vessel for futures both realized and foreshortened, for the asymmetric power relations enacted through the future as a margin. The coloured, incarcerated, monstrous subject, prevented from actually becoming a terrorist, is a subject who is made to exist as nothing other than the bodily proof of prevention. In these bodies, some of which are literally placed in the legally uncertain twilight zone that is Guantánamo (Lemov, 2015), we find shades of not only Giorgio Agamben’s (1985) notion of bare life, but also Elaine Scarry’s (1985) body in pain: a body whose suffering is leveraged as powerful and (seemingly) inimitable evidence of the truth. ‘Zero tolerance’ functions as a back door for the powerful to imprint their truths upon the weak, an additional route by which populations may be further encoded, contained and incarcerated.

Tradeable informational goods

These themes of conservatism, state control and the making of monsters bring us to the strategic value of speculative futures. ‘Value’ here goes well beyond direct monetary profit and describes how security practices benefit from the invocation of the future. Nevertheless, it is precisely in the context of financialization that we find a relatively early and robust development of the future in this sense. One place where this is explicitly expressed, of course, is the stock market. The trading of futures began in the 17th century as a standardized contract that would fix the future price at which a material good, such as grain, would be purchased. These futures stabilized the uncertainties of agricultural realities, providing a certain reliability for human judgment and action – even as the conditions and gaps in these networks of promises were to be exploited for speculative profiteering. This basic mechanism would be further expanded through formalized futures contracts in the late 19th century and the launch of an international financial futures exchange in the 1970s (for an overview, see LiPuma and Lee, 2004; MacKenzie, 2017).

The historical expansion of futures trading involved a progressive unmooring of futures from material goods, enabling ever-more-sophisticated fragmentation and recombination of futures as tradeable objects. Broadly speaking, derivatives trading writ large has been analysed as ‘virtual’ assets or goods in a Deleuzian sense (Arnoldi, 2004), a process where the actual occurrence of a projected future event is no longer so relevant and what matters is the act of making that projection in the moment-to-moment fencing between traders and their algorithms (also see Hayles, 2017: 144). Here, futures are increasingly unmoored from not only the material goods that originally anchored their value but also the imagined horizon of ‘real’ events. Ayache (2015) points out that every time one takes an action in the market, one might be acting in anticipation of future events under a probabilistic model, but that action accordingly changes the field of play for everything that happens afterwards. The eventuation (or not) of whatever was predicted at some future time hardly matters (see also Massumi, 2015a: 190); at that point, other contingent actions will already have transformed the field of play (Ayache, 2015: 41–52, 148, 159–171). A similar feedback loop is found operating at frenetic timescales in high-frequency trading (see, for example, Borch et al., 2015; Pasquale, 2015). The use of the future consists not of predictions that seek to actually match the real sequence of unfolding events from a mathematical array of probabilistic states, but rather of tradeable informational goods. A bet on the future in the form of a price and bid is not, primarily, a claim about what will happen to the price later. Nor is it the information itself, a prophecy about the future, that is on sale. Primarily, the act of pricing is a performative act that itself makes and loses money.

In essence, the ‘value’ of futures trading depends on algorithms and the ability of traders to manipulate the future margins of the knowable, produce volatilities, uncertainties and asymmetries, and derive profit from this manipulation. This principle is most obviously at work in financial derivatives, which carry through Frank Knight’s (1964) canonical lesson that uncertainty (defined in his case as asymmetries between actors’ knowledge, which is always imperfect) is the basic source of all profit. Uncertainties regarding the future are operationalized as amorphous and flexible margins through which asymmetries of information and calculation may be engineered among the traders and profit may be gained (also see Zaloom, 2003: 261–262). Here, too, success is predicated on the ability to preempt possible futures and to use such predictions to nudge and manipulate subjects in the present – an obsession not just with speed, but with futures as useful fictions, which is most explicit in high-frequency trading (see also MacKenzie, 2017; Martin, 2002: 46–49; Zaloom, 2009). The future opens up margins of tolerated ambiguity, which may be exploited to maximize profit.

The effectiveness of the speculations circulated through the black market – in other words, their ‘profit’ – is counted not so much in the objective accuracy vis-a-vis what ‘ends up happening’, but in the degree to which those speculative claims can be legitimated and pressed into action. Louise Amoore (2011) in particular argues that data-driven security analytics does not and cannot hinge on its ‘external’ accuracy to actual events, but on the internal ‘precision’ of how it combines data towards elegant and consistent results. Massumi and Ayache also describe a feedback loop in the work of projecting futures. Recycled pasts are traded into contingent future-tensed truth-claims, which may subsequently be leveraged to – in the case of the sting operation – achieve legitimacy and judicial confirmation. The propaganda videos Shnewer downloaded at the behest of Omar (and copied with an FBI-purchased DVD burner) come to stand in for the intent, emotion and fatwa of a conspiratorial group (US v Shnewer et al., 2008: 5835). Recreational paintball becomes weapons training. Balloons become stand-ins for soldiers’ heads. As a reward for the work of creating these valuable objects, informants trade their illegal status for a green card. Subjunctive, ‘what-if’ refrains have now become a familiar fixture in public justifications of predictive security practices: the expensive new surveillance system in New York’s subways may not have caught any terrorists red-handed, but it would have deterred certain terrorists; it would have prevented concrete attacks if there had been any; and it would have been worse without the surveillance (Molotch, 2012). What might otherwise have stalled as insufficient evidence, unfounded conjecture, unproven anxieties – the angry temper and controversial political views of a young Arab-American male, the absence of terrorist attacks on a New York subway – go through the future as a clearing-house for contingency and come out as something with a clear affective presence and calculable, actionable coordinates.

In this view, risks represent a historically settled technique for rationalizing futures. Actuarial reason converts future contingencies into comparable and calculable forms. When the risks are produced, what is made and purchased – whether as derivative, as insurance – is a certain kind of exposure to contingency, with a price put on it. O’Malley (2008: 57–58) points out that Donzelot, Ewald, Castel and other prominent risk theorists throughout the 1980s and 1990s saw risk as a probabilistic production, whose abstract nature allows many different kinds of interventions to be legitimized. It is precisely this process of rationalization that allows, rather than eliminates, the profitable work of speculation (Zaloom, 2004). In this sense, anticipatory reason represents no fundamental departure from risk as a technology of government, but rather a markedly capitalist expansion of it. Analogous to the evolution of derivatives trading over the latter half of the 20th century, anticipatory reason in security practice is part of a progressive expansion of the existing market for packaged, sellable ‘futures’.

In short, the invocation of the future in sting operations demonstrates a strategic and therefore highly interested and selective practice. If financial speculation recombines the future for monetary profit, security speculation does so for the production of valuable facts – the kind of facts that are actionable for purposes of arrest, litigation and public justification. (Of course, some forms of security speculation also have monetary profit in mind, given the thriving industry of private firms like Palantir that bid for highly lucrative contracts with the police, intelligence agencies and other arms of government.) As we have seen, each act of counter-terrorist prediction utilizes the future as an amorphous and indefinitely broad horizon of possibilities, but also narrows that future into a specific range – for instance, of stereotypical jihadist motives and bearded, anti-social, Arab-descent suspects. In turn, such narrowing of the future shapes the horizon of the possible in which politics and ethics takes place. The flexibility of the future does not imply an open clearing for individual freedoms, but, more often, a space for powerful institutions to impose their interests on the social production of evidence and judgment.

Conclusion

In ‘Security and the Incalculable’, Louise Amoore (2014: 425–426) describes a meeting between Alan Turing and Ludwig Wittgenstein, the former attending a class taught by the latter. When Turing raises the predictive powers of mathematics, Wittgenstein disagrees. Numbers do no such thing as ‘predicting’ on their own, he says: it is how we use those numbers, the grammar, the relation of use. In the post-9/11 expansion of ‘anticipatory’ security, we find an emergent grammar for the processing of illegalistic material – future speculations, past data and behavioural indicators – towards more calculable and actionable kinds of output. The future, typically invoked in this context as a temporal zone of impending events that are to be objectively known through probabilistic predictions, is better understood as the utilization of contingency towards present strategic objectives. This function is achieved by what we describe in terms of an epistemic black market: the translation of relatively speculative truth-claims into more concrete forms in order to influence decisionmaking and enable anticipatory intervention. From the rhetorical figure of the ‘next terrorist attack’ to prosecution tactics against would-be terrorists preempted by sting operations, contemporary security practices seek to render an ever more expansive bandwidth of bodies, behaviours, situations, eligible for control. To call this ‘strategic’ acknowledges the latent asymmetries and power differentials in the process. The future in anticipatory security facilitates speculation not as an open field of possibility, but as the retrenchment of state control and familiar archetypes of the ‘enemy’ by other means. The epistemic and normative validity of anticipatory reason must be tethered not to the useful fiction of ‘accuracy’ against the future as an unfalsifiable and inexhaustible horizon, but to the political and moral consequences of those predictive actions in the present: the incarceration of bodies, the execution of drone strikes, the steady growth of surveillance apparatuses.