Defining ‘Security’
Before we can explore the ontology of security, we must establish a preferred meaning of the word and purge it, as far as possible, of ambiguity. This preliminary task is especially important when one’s subject is a term in common currency and deployed in a wide variety of contexts for a wide variety of purposes.Footnote 1 The more familiar a word, the more likely we are to take for granted that our own particular understanding of it is widely shared and that its meaning goes without saying. This is as true of the word ‘security’ as it is for almost every other key concept in the study of world politics.
A good place to start, when attempting to narrow and refine meaning, is with a dictionary of record. The Oxford English Dictionary (OED) entry for ‘security’ is more than 9,000 words long and includes nine major senses and 20 minor ones.Footnote 2 Ignoring rare, archaic, or technical definitions, and also those that define the word with respect to a specific referent or threat,Footnote 3 we are left with the following:
1. The state or condition of being or feeling secure.
1a. Freedom from care, anxiety, or apprehension; absence of worry or anxiety; confidence in one’s safety or well-being.
2. Freedom from danger or threat.
2a. The state or condition of being protected from or not exposed to danger; safety.
2c. The condition or fact of being secure or unthreatened in a particular situation …
There are several notable features of these definitions when taken collectively. One is that they denote both objective states of affairs (e.g., ‘The state or condition of being secure’) and subjective impressions (‘feeling secure’; ‘freedom from care, anxiety or apprehension; absence of worry or anxiety; confidence in one’s safety or well-being’). A second is circularity – i.e., defining the noun in terms of its corresponding adjective. The circularity is not resolved by consulting the entry for ‘secure,’ which also includes objective and subjective definitions (e.g., ‘Protected from or not exposed to danger; certain to remain safe and unthreatened’; ‘Free from care, apprehension, or anxiety; carefree, untroubled’).Footnote 4
Our first task is to decide whether, in articulating a full ontology of security appropriate to the study of international politics, we wish to embrace both objective and subjective meanings. There are compelling reasons to shun the latter, the most obvious of which is that one can feel secure without actually being secure, and vice versa. Most of the passengers aboard the 1912 maiden voyage of RMS Titanic were convinced that the ship was unsinkable when plainly it was not.Footnote 5 Conversely, people commonly underestimate the safety of commercial aviation.Footnote 6 Since ‘feeling secure’ can be written as ‘believing rightly or wrongly that one is secure’ while ‘being [objectively] secure’ cannot similarly be rewritten in terms of psychological states, the objective condition is clearly the foundational concept. Moreover, embracing a fundamentally psychological understanding of security would justify allocating resources on the basis of irrational phobias, putting into jeopardy – if perhaps only through neglect – important referent objects.
A third notable feature of these definitions is that they invoke both ‘safety’ and ‘absence of danger or threat.’ Consulting the OED entry for ‘safety’ quickly reveals that these are synonyms (safety is ‘The state of being protected from or guarded against hurt or injury; freedom from danger’).Footnote 7 These invocations are reasonable insofar as they relate to the (preferable) objective understanding of security, but they are curiously absolute. One can only be confident of a complete ‘absence of danger or threat’ in the short term. Titanic was free of danger from icebergs on the morning of April 14, but this was no longer true just before midnight. Taking the long view, Titanic was never perfectly safe from icebergs – or from collisions, a boiler explosion, a rogue wave, being torpedoed by a submarine, or what have you. While the efforts that her designers and builders made to ‘secure’ her from threats such as these were considerable, they would have been less than perfect even had they been more conscientious. Nothing is ever absolutely safe.Footnote 8 ‘Security,’ therefore, must be thought of as a matter of degree that can vary over time. Risk analysis operationalizes this insight explicitly. No one gives away insurance policies for free.
Security, then, is best thought of as an ‘objective’ condition, but, as a matter of degree, it is always relative. Things can be more or less secure, but never absolutely secure. It is possible to imagine optimizing security – designing systems and procedures with practically unsurpassable prospects of protecting a referent against various threats – but it is not obvious that we could ever know with absolute confidence that we had done this, and in practice we generally aim for thresholds that we believe to be ‘good enough.’ Among the most popular ISO standards, for example, are the Information Security Management Systems standards in the ISO/IEC 27000 family.Footnote 9 These reflect the good-faith, well-informed judgments of technical experts on the minimal acceptable ways of protecting data. But there are even better ways. They would simply be too costly for most organizations to implement. So, the thresholds of security that we aim for in practice reflect not simply performance but cost as well.Footnote 10
Ocean liners and aircraft are tangible and observable things, as many security referents are. But, as the information security example demonstrates, others are not. We can, and do, speak of and act to promote the security of things both natural and constructed, physical and social. States can and do pass laws and devote resources, for example, to protect wetlands, endangered species, critical infrastructure, heritage languages, democracy, and the rule of law. Ontologically, these security referents run the gamut. Many cross ontological categories. Data, for example, are only ‘data’ if both stored physically and rendered intersubjectively meaningful by systems of social practices. They can be destroyed by attacking either aspect.
Any ontology of security that will satisfy both our theoretical and practical requirements, then, must fit under a large tent that treats all of these classes of ‘objects’ as real.Footnote 11 This is required by the lowest-common-denominator/common-sense metaphysics requirement that I specified in the Introduction. Whatever else people believe about the world, they live their lives in such a way as to demonstrate an epistemic commitment to all of them. No one tries to walk through unopened doors. No one denies the existence of language or social norms. No one can fully describe the world as they experience it or as they wish it to be without reference to ideas and ideals. They may hold additional beliefs that are not widely shared, or that are perhaps even idiosyncratic, requiring additional ontological commitments – for example, beliefs in God, gods, angels, magic, dragons, unicorns, or imaginary friends – but unless these beliefs manifest in politically relevant systems of social meaning whose ontological status as a class we must perforce embrace, we can remain comfortably agnostic about them.
My commitment here to the necessity of this common-sense ontology, I submit and confess, smacks of circularity, as I have essentially been presuming the reality of the classes of things for which I am attempting to make the case. The paradox is reminiscent of that bedevilling a common interpretation of René Descartes’ famous Cogito Ergo Sum (‘I think, therefore I am’) whose consequent is meant to prove the existence of an antecedent already presumed.Footnote 12 There is no fully satisfactory philosophical resolution here. But there is some solace, I believe, in the thought that denying the reality of these things requires a commitment to one or more propositions the truth of which would at worst render my project moot but the falsity of which could be catastrophic if acted upon (I am thinking here, for example, of nihilism, radical skepticism, and solipsism).Footnote 13 This is my ontological analogue to Pascal‘s wager.Footnote 14 In any case, we know from ancient Greek philosophy that no inquiry can begin without an axiom (an unquestioned assumption) of some kind, and the intuition that we can attempt to ‘secure’ things that are tangible, intangible, physical, social, natural, constructed, ideational, or ontologically hybrid seems to be about as unobjectionable a place to start as one can imagine, given the (apparent) reality that this is what in fact we do.Footnote 15
Now, some readers may be inclined to think my operative ontology old-fashioned. In effect, I am embracing a Newtonian view that presumes both a meaningful difference between observer and observed and the existence of a mind-independent world. At least two lines of criticism are available to those who find this unsatisfactory. One is that a Newtonian view overstates the discreteness of things. The other is that we live in a quantum universe, not a Newtonian one.
A good example of the first line of criticism would assert the superiority of ‘relational ontology,’ or the view that things exist only by virtue of their relationship to other things.Footnote 16 This is not a view that I would necessarily reject, but I would qualify my endorsement thus: knowing something or fully describing something at any given time certainly does very often, at least, require relating it to something else.Footnote 17 There is no such thing as a mother without a child, for example, or a state without a population. But the mother is also a person, as is the child, and these two facts do not depend upon the existence of the relationship. And while a state would certainly not in any meaningful sense be a state if it had no population, those who give a state reality would not necessarily cease to be if the state did. Relationality, in other words, varies by context when we find it, and can operate asymmetrically. But in some cases it is simply not relevant. We have no reason to believe that the sun exists only because we exist to observe it, because our planet orbits it, or because it stands in relation to anything else whatsoever. We have good reason to believe, in fact, that the sun existed before us or any of its planets and that it will continue to exist, if in a dramatically different form (i.e., as a white dwarf), long after we are gone.Footnote 18 My working ontology, in other words, does not reject relationalism; it subsumes it.
The second objection is that Newton was wrong.Footnote 19 The universe, fundamentally, is a field of probabilistic entities whose building blocks violate Aristotle’s laws of logicFootnote 20 and whose properties cannot be fully known or described.Footnote 21 The very attempt to know or describe them alters them. So, at any rate, says quantum physics. This may well be right, although two of the founding fathers of quantum physics – Niels Bohr and Werner Heisenberg – never agreed on how to interpret experimental findings, and debate rages on.Footnote 22 The equations certainly work and have paved the way for otherwise unimaginable technologies such as quantum computing; so there is no reason to doubt that there is a ‘there’ there, whatever ‘there’ may be. But the key point is that the overwhelming majority of people live their lives as if Newton were right, and do so successfully. We wouldn’t even know how to live our lives in a quantum way. No one builds a bridge using quantum mechanics; no one calculates braking distance using quantum equations. The Newtonian worldview works perfectly well for any purpose that does not involve extremely small spatial or temporal scales. Very possibly this is because quantum effects wash out at meso scales. This is where life operates.
So, yes, my operative ontology may be old-fashioned. It may also in certain respects be inaccurate. But it works, and I believe it can command wider assent than more esoteric, avant-garde alternatives. It is, I submit, commonsensical precisely because it does a better job of helping us make sense of the world than any available substitute.
Securing
There are different ways of securing something, and the best way typically depends upon exactly what threatens it. Suppose, for example, that you wish to secure sensitive data. These will no doubt be stored in some form (paper, magnetic, electronic, etc.) and in some place (typically in a building, in a vault, in file cabinets, or on computers). Also typically, you will want to protect these data from at least three different kinds of harm: destruction, alteration, or theft.Footnote 23 Destruction might be accomplished in various ways, both natural (earthquake, fire, flood) and intentional (attack, sabotage, arson, or – if stored electronically – hacking). Alteration and theft can also often be accomplished by hacking, but under certain circumstances, by covert action or burglary as well. Depending upon the specific threat–referent pair, you may be well advised to invest most heavily in prevention or deterrence (trying to keep the threat from arising in the first place),Footnote 24 interdiction (keeping the threat at bay in case it manifests), defence (protecting the referent from immediate harm when the threat cannot be kept at bay), resilience (ruggedizing or inoculating the referent when defence is impossible), or recovery (restoring the referent following harm). If you are unsure what particular threat poses the greatest risk to your data, you are likely to invest in as many of these as possible. For example, you will store your data in buildings built to an appropriate earthquake code, above a flood plain, and/or with an adequate firebreak if in an area prone to forest fires; you will hire guards and cooperate with law enforcement to stymie thieves or terrorists; you will invest in good-quality strongboxes and alarms; you will put your computers behind firewalls, or perhaps even air gaps; you will encrypt data, put in place strict user access protocols, and screen users for reliability and trustworthiness; you will store redundant copies of your data offsite; and you will seek to anticipate and design around potential common-mode failures.
Because there are different ways of attempting to provide security, we must beware of conflating any one of them with security as such. In interwar Europe, for example, when a traditional understanding of security as protection of the state from armed attack prevailed, security was more or less synonymous with defence when arguably it would have been much better served by efforts to prevent Nazism from arising.Footnote 25 During the Cold War, with a traditional understanding of security still firmly in place and defence against nuclear attack virtually impossible, deterrence came to the fore. With the benefit of hindsight, we can see today that nuclear deterrence was often more problem than solution – a major security threat in and of itself.Footnote 26
We are no longer conceptually hobbled by a traditional understanding of security as protection of the state from armed attack. Today’s security agenda is positively kaleidoscopic – complex, rich, variegated, and changeable. This has given rise in some quarters to a curious fetishization of resilience.Footnote 27 Building resilience may well be the most effective mode of providing security in many cases, but the gusto with which the concept has been taken up and the range of problems to which it has been applied surely give rise to questions about whether enthusiasm for the concept has outstripped the care and rigour with which we apply it.Footnote 28
A second temptation to resist is confusing the concept of security with any of its necessary, sufficient, or permissive conditions. The first two categories, I submit, are empty sets. I cannot think of an earthly referent object that faces no possible harm. There may well be some things that are never threatened as a matter of fact, but this would be a matter of purest luck.
With respect to the third category, a great many things can conduce to security under various circumstances. Among the most powerful of these, perhaps – if we are particularly concerned with human beings and their quality of life as referents – are justice, empowerment, and ‘emancipation.’ While justice, empowerment, and emancipation can most certainly contribute to (human) security, it would be a mistake to confuse any of them with security as such. This is a temptation to which even one of the field’s most preeminent scholars succumbed:
“Security” means the absence of threats. [Technically, this is incorrect, as we have just seen; but let us ignore this for the moment.] Emancipation is the freeing of people (as individuals and groups) from those physical and human constraints which stop them carrying out what they would freely choose to do. War and the threat of war is one of those constraints, together with poverty, poor education, political oppression[,] and so on. Security and emancipation are two sides of the same coin. Emancipation, not power or order, produces true security. Emancipation, theoretically, is security.Footnote 29
While emancipation may certainly ‘produce’ security in certain circumstances, effects are not their own causes. Freedom and security are two different things.
Security, Securitization, and Threat Perception
Taking all of the above into account, we are left with the following working definition: Security is an objective condition of relative safety from harm. Crucially, the definition itself leaves entirely open the questions of which referents are worth securing and how best to secure them. Much of the rest of this book is concerned with the former; the latter – a question of policy rather than theory – I largely leave to others. But before we delve into the question of how to know how to invest wisely in security, it would be helpful to clarify the relationship of the concept both to ‘securitization’ and to threat perception.
Securitization
Investing in security involves making choices, and when we are speaking of international security, at least, the key decisions are made by political actors – namely, leaders of states. Security policy, in other words, is the output of a political process. Among the great contributions of the ‘Copenhagen School’ of International Relations to our understanding of this process is the concept of securitization, a term popularized by Barry Buzan, Ole Wæver, and Jaap de Wilde in their ground-breaking study, Security: A New Framework for Analysis.Footnote 30
Buzan and his colleagues argued that when we attempt to label something a ‘security’ problem, we mean to say that we regard it as particularly acute or important; we mean to say that it should take priority over other sorts of issues;Footnote 31 and we mean to say that it is appropriate to make extraordinary efforts and devote extraordinary resources to address it, perhaps even loosening or suspending the normal rules of politics to do so.Footnote 32 Securitization requires one or more ‘securitizing moves’ by one or more ‘securitizing actors.’ Securitizing moves can succeed or fail. Only if they succeed can an issue be said to have been ‘securitized.’
A good example of a securitizing move by a securitizing actor was the first largely (perhaps one should say ‘eventually’) successful attempt to promote a broader understanding of security as the Cold War wound down – Jessica Tuchman Mathews’s influential 1989 Foreign Affairs piece, ‘Redefining Security.’Footnote 33 ‘Global developments,’ Mathews argued, ‘now suggest the need for broadening [the] definition of national security to include resource, environmental[,] and demographic issues.’Footnote 34 Interestingly, Mathews argued that we should understand these as national security issues when she might have characterized them as new security concepts rather than as new conceptions of an existing one.Footnote 35 This was both clever and strategic, as was her choice of outlet. National security had an unimpeachable pedigree as an issue regularly warranting vast investments of time, talent, and money, and Foreign Affairs was virtually required reading for anyone involved in national security policy making. For her argument to catch on, the right audience had to hear it, and they were more likely to be persuaded if she pitched it as an elaboration of their primary problématique rather than as an alternative to it.
In crucial respects, Mathews’s argument was not new. Much of the alarm she sought to raise others had sought to raise earlier – for example, Paul and Anne Ehrlich in their 1968 bestseller, The Population Bomb.Footnote 36 But the Ehrlichs lacked Mathews’s credibility as a national security establishment insider, published in a form and forum easily ignored, attempted to change the channel rather than stream to it, and – perhaps most importantly – wrote at a time when those in a position to take action were preoccupied by traditional security issues (the Cold War in general, and Vietnam in particular).
Mathews’s was but the first step down a long and difficult road with fellow travelers as diverse as the Intergovernmental Panel on Climate Change, Al Gore, and Greta Thunberg.Footnote 37 Today, the effort to securitize environmental issues involves a broad range of public and private actors, political and scientific elites, civil society organizations, and grassroots activists. It has been largely, though not entirely, successful, encountering predictable resistance from vested interests, opportunists, and know-nothings.Footnote 38 Occasionally, the effort has encountered heavy weather even from respectable quarters.Footnote 39
Sometimes, securitizing moves are successful but inappropriate. A classic example would be the George W. Bush administration’s full-court press to make the case for war against Iraq in 2003. The effort rested on false or misleading information about Iraqi weapons of mass destruction (WMD) presented in a tendentious way – although it would appear in retrospect that at least some of the key decision-makers were utterly sincere in their statements of what they believed and in their certainty about the justice or necessity of their cause.Footnote 40 Conversely, securitizing moves sometimes fail when they ought to have succeeded. Winston Churchill’s prescient warnings about Hitler are a case in point.Footnote 41
Despite the evident utility of the securitization framework for helping us understand the processes by which perceived problems come to be seen and treated as ‘security’ problems, it is not without its flaws and limitations.Footnote 42 For example, a prominent strand of the Copenhagen School is almost certainly wrong to characterize securitization as a ‘speech act.’ As Buzan and his colleagues put it,
The process of securitization is what in language theory is called a speech act. It is not interesting as a sign referring to something more real; it is the utterance itself that is the act. By saying the words, something is done (like betting, giving a promise, naming a ship).Footnote 43
This statement is correct in characterizing a ‘speech act’ as a performative utterance.Footnote 44 By definition, a speech act has illocutionary force.Footnote 45 The conscious statement ‘I promise’ fully and sufficiently accomplishes something: namely, it establishes a moral obligation to carry out an undertaking. Similarly, when a judge utters the words, ‘I sentence you to life in prison without parole,’ ipso facto you are sentenced, and when your commanding officer says, ‘I order you to retreat,’ ipso facto you are so ordered. Not so with securitization. Characterizing something as a security problem does not ipso facto make it one. After all, as the Bush administration demonstrated, one might be wrong.
Securitization takes place as a result of successful securitizing moves that certainly involve utterances, but while these utterances try to accomplish something, they do not accomplish anything automatically. Put another way: securitizing moves are perlocutions, not illocutions.Footnote 46
I say that it is ‘almost certainly’ wrong to describe securitization as a speech act for two reasons. The first is that John Searle, one of the most prominent speech act theorists, included ‘assertives’ amongst his taxonomy of illocutions:
The point of assertive speech acts is to commit the hearer to the truth of the proposition. It is to present the proposition as representing a state of the world. Some examples are statements, descriptions, classifications, and explanations. All assertives have the word-to-world direction of fit, and the sincerity condition of assertives is always belief. Every assertion is an expression of a belief. The simplest test for identifying assertives is to ask whether the utterance can be literally true or false. Because the assertives have the word-to-world direction of fit, they can be true or false.Footnote 47
Securitizing moves are undoubtedly attempts to persuade an audience of the truth of a proposition, and successful securitizing moves certainly do so. They would therefore appear to qualify as assertives. John Austin, however – the founder of speech act theory – ultimately decided that assertives did not, in fact, qualify as speech acts, on the ground that with an assertive, nothing is actually done.Footnote 48 Uttering the words ‘It is raining’ – an assertive – does not make it rain. Searle as much as admitted that Austin was correct by saying that assertions ‘can be literally true or false.’ One can check an assertion for facticity. One cannot – nor does one need to – check any of Searle’s other four types of speech acts (directives, commissives, expressives, or declarations) for facticity.Footnote 49 One can only check them for sincerity, legitimacy, or consistency. But an insincere promise is still a promise, and an inconsistent verdict is still a verdict – at least until it is overturned on appeal. Only a lack of legitimacy nullifies a performative. Humphrey Bogart and Katherine Hepburn did not actually become husband and wife on the set of the 1951 film African Queen when Peter Bull said, ‘By the authority vested in me by Kaiser Wilhelm II, I now pronounce you man and wife. Proceed with the execution.’
Why Searle insisted that assertives had illocutionary force is something of a mystery. Nevertheless, Buzan and his colleagues make a move that one might be tempted to construe as rescuing the notion of securitization as a speech act: namely, denying the utility of distinguishing between ‘objective’ (real) and ‘subjective’ (perceived) security threats:
… [It] is neither politically nor analytically helpful to try to define “real security” outside of the world of politics and to teach the actors to understand the term correctly. [N. B. This is exactly what I am doing in this book.] Such rationalist universalism will easily be “right” on its own terms, but it will be of very little help in political analysis. It is more relevant to grasp the processes and dynamics of securitization, because if one knows who can “do” security on what issue and under what conditions, it will sometimes be possible to maneuver the interaction among actors and thereby curb security dilemmas.Footnote 50
There are two ways to read this statement. One is very narrowly, such that the scope conditions of securitization theory are limited to ‘political analysis.’ On this reading, ‘objective’ security threats are uninteresting unless and until they manifest as securitized subjectively perceived security threats as the result of successful securitization moves. Since the set of unsecuritized subjective security problems would necessarily be empty, it would be true that successful securitization did, in fact, accomplish something – namely, elevate something to the status of a subjectively perceived security problem as a matter of politics (only). The problem with this reading is that it sets aside virtually all of the interesting challenges having to do with protecting things of value, which is what security policy is all about, and it exogenizes debate about appropriate and inappropriate securitization. This reading, in other words, reduces securitization theory to a theory of labeling and prevents it from being a theory of securing.Footnote 51
This reading has four further debilitating weaknesses. First, it brackets securitizing moves from successful securitizations. Only the latter accomplish something, but only the former involve speech. Second, it renders securitization theory arbitrary, in the sense that it defines away as uninteresting any aspect of security studies that did not involve politicization. Third, it renders securitization theory paradoxical – for how could one possibly make the case that we ought to treat something as a security problem without reference to facts beyond politics? Finally, it renders unintelligible the claim that ‘if one knows who can “do” security on what issue and under what conditions, it will sometimes be possible to maneuver the interaction among actors and thereby curb security dilemmas.’ Are security dilemmas real, or only conceptual artifacts of conflicting securitizations? If the former, securitization theory has nothing to say about them; if the latter, understanding the processes by which conflicting securitizations came about may expose them, but cannot solve them.
A second (stronger) reading of the statement is to deny that there is any meaningful distinction between objective and subjectively perceived security threats regardless of the politics of security. This reading rescues securitization theory from charges of arbitrariness and paradox, but at the cost of plausibility and utility. No matter what story we tell about the securitization of Iraqi WMD in 2002 and 2003, the fact is that there simply weren’t any. We cannot meaningfully gauge the ‘threat’ Iraq posed if we deny ourselves the right to say that certain things perceived as threats were not. Once we admit that there can be objective security threats about which we may be correct or incorrect, neither a securitizing move nor a successful securitization can be a speech act.
A final difficulty with this particular strand of the Copenhagen School’s approach to security is its insistence that securitization involves presenting something as an ‘existential’ threat.Footnote 52 It is plain that we do, in fact, treat things as security threats that do not existentially threaten a referent object. Islamist terrorism has very obviously been securitized in American politics since 2001, and yet it is impossible to tell a plausible story about al-Qaeda, the so-called Islamic State, or any other radicalized group or lone wolf posing an existential threat to the United States. The closest one might come is to say that in the immediate aftermath of 9/11 it seemed briefly plausible to imagine that al-Qaeda might provoke a hysterical overreaction of the kind that would existentially threaten American liberal democracy.Footnote 53 But arguably, what should be securitized in this case would be the hysterical overreaction to Islamist terrorism, not Islamist terrorism itself.Footnote 54 Own goals can hurt more than defensive lapses.
It is not clear exactly what level of harm something must seem capable of inflicting before it warrants treatment as a ‘security’ threat – perhaps this is best seen as intersubjectively constructed – but it seems clear enough that we do not ordinarily require something to pose an existential threat before we securitize it.
It is difficult to avoid the conclusion that attempting to connect securitization to speech act theory was an unnecessary mistake. Juha Vuori writes, ‘Both social and linguistic analysis [are] necessary to understand the performative of securitization.’Footnote 55 I would say social and rhetorical analysis are necessary to understand securitization, which is not performative. Securitization theory can stand on its own without speech act theory. There is, after all, much in the Copenhagen School’s conceptual armamentarium that is certainly useful for understanding how and why things come to be treated as security problems.
But the more interesting question is: how well do we do this? To answer this question requires that we be sensitive to security as a matter of degree; that we embrace the distinction between objective (real) and subjective (perceived) security threats; that we attend to the challenges of threat perception; and – most importantly – that we understand what makes something worth securing in the first place.
Threat Perception
Securitizing appropriately requires accurate threat assessment. It goes without saying that we will only devote scarce resources to problems that we perceive, so it is imperative to narrow as far as possible (and preferably eliminate) the gap between objective and subjective threats.
Alas, threat perception is fraught. In some respects, we are very good at it; in other respects, we are dreadful.
Take, for example, the problem of climate change, now widely recognized as perhaps the most serious known threat to a wide range of security referents. Scientists had discovered the phenomenon of climate variability as early as the eighteenth century, and by the early twentieth a tolerably clear picture was emerging that human activity might be affecting the climate directly. ‘The atmosphere contains altogether 1,500,000,000,000 tons of carbon dioxide,’ wrote one prescient science journalist in 1912; ‘Consequently the combustion of coal at the present rate will double it in about 200 years, unless it is removed by some means in enormous quantities … [I]t may well be that the enormous present-day combustion of coal is producing carbon dioxide so fast that it will have important climatic effects.’Footnote 56 And yet it would take almost another 100 years for climate change to be securitized – long after annual atmospheric greenhouse gas emissions had reached levels much too high to roll back in time to avoid dramatic, and possibly catastrophic, long-term effects.Footnote 57
A key explanation for this ‘securitization lag’ is, quite simply, the way humans are wired to perceive threats, as psychologist Daniel Gilbert explained in an entertaining 2006 opinion piece titled, ‘If Only Gay Sex Caused Global Warming’:
No one seems to care about the upcoming attack on the World Trade Center site. Why? Because it won’t involve villains with box cutters. Instead, it will involve melting ice sheets that swell the oceans and turn that particular block of lower Manhattan into an aquarium.
The odds of this happening in the next few decades are better than the odds that a disgruntled Saudi will sneak onto an airplane and detonate a shoe bomb. And yet our government will spend billions of dollars this year to prevent global terrorism and … well, essentially nothing to prevent global warming.
Why are we less worried about the more likely disaster? Because the human brain evolved to respond to threats that have four features – features that terrorism has and that global warming lacks.
First, global warming lacks a mustache. No, really. We are social mammals whose brains are highly specialized for thinking about others. Understanding what others are up to – what they know and want, what they are doing and planning – has been so crucial to the survival of our species that our brains have developed an obsession with all things human. We think about people and their intentions; talk about them; look for and remember them.
That’s why we worry more about anthrax (with an annual death toll of roughly zero) than influenza (with an annual death toll of a quarter-million to a half-million people). Influenza is a natural accident, anthrax is an intentional action, and the smallest action captures our attention in a way that the largest accident doesn’t. If two airplanes had been hit by lightning and crashed into a New York skyscraper, few of us would be able to name the date on which it happened.
Global warming isn’t trying to kill us, and that’s a shame. If climate change had been visited on us by a brutal dictator or an evil empire, the war on warming would be this nation’s top priority.
The second reason why global warming doesn’t put our brains on orange alert is that it doesn’t violate our moral sensibilities. It doesn’t cause our blood to boil (at least not figuratively) because it doesn’t force us to entertain thoughts that we find indecent, impious[,] or repulsive. When people feel insulted or disgusted, they generally do something about it, such as whacking each other over the head, or voting. Moral emotions are the brain’s call to action.
Although all human societies have moral rules about food and sex, none has a moral rule about atmospheric chemistry. And so we are outraged about every breach of protocol except Kyoto. Yes, global warming is bad, but it doesn’t make us feel nauseated or angry or disgraced, and thus we don’t feel compelled to rail against it as we do against other momentous threats to our species, such as flag burning. The fact is that if climate change were caused by gay sex, or by the practice of eating kittens, millions of protesters would be massing in the streets.
The third reason why global warming doesn’t trigger our concern is that we see it as a threat to our futures – not our afternoons. Like all animals, people are quick to respond to clear and present danger, which is why it takes us just a few milliseconds to duck when a wayward baseball comes speeding toward our eyes …
There is a fourth reason why we just can’t seem to get worked up about global warming. The human brain is exquisitely sensitive to changes in light, sound, temperature, pressure, size, weight and just about everything else. But if the rate of change is slow enough, the change will go undetected. If the low hum of a refrigerator were to increase in pitch over the course of several weeks, the appliance could be singing soprano by the end of the month and no one would be the wiser …
The human brain is a remarkable device that was designed to rise to special occasions. We are the progeny of people who hunted and gathered, whose lives were brief and whose greatest threat was a man with a stick. When terrorists attack, we respond with crushing force and firm resolve, just as our ancestors would have. Global warming is a deadly threat precisely because it fails to trip the brain’s alarm, leaving us soundly asleep in a burning bed.Footnote 58
There is no doubt that evolution has selected for certain threat-perception skills and not for others, but threat perception is an enormously complex process involving both shared and idiosyncratic elements – which makes the politics of securitization as fraught as threat perception itself.
Among the shared features are the basic neurological structures and processes involved in monitoring our environment and processing threat cues. In the human brain, the amygdala plays a crucial role. Subjects with basolateral amygdala damage experience great difficulty perceiving threats.Footnote 59 But a wide variety of other structures are involved as well. MRI tests show that areas of the temporal, parietal, and occipital lobes all light up when a subject sees a menacing-looking person approach, indicating that those parts of the brain tasked with sight, spatial awareness, and social categorization all interact in a coordinated way to assess degree of threat.Footnote 60 Again, damage or deficits in one or more of these areas will affect one’s response.
Physiology is not the sole determinant of this capacity, of course; various dispositional and contextual considerations bear as well. There is strong evidence that emotion can have a powerful effect on threat perception, for example. General (trait) anxiety would appear to be a particularly important variable; dispositionally anxious people are quicker to perceive threat, have lower threat perception thresholds, and are more responsive to threat priming.Footnote 61 Moustaches are not necessarily required to elicit threat perceptions, but they are more likely to do so if someone had been assaulted by a man with a moustache in the past. Similarly, someone who fears snakes may become alarmed simply by seeing something with a curvilinear shape.Footnote 62
Threat perception is further complicated by perfectly normal decision-making heuristics and the biases they induce.Footnote 63 It is rare that we are able to make decisions on the basis of full and accurate information about every possible option, so we often must rely on shortcuts or rules of thumb to help us cope with uncertainty, limited time, and limited cognitive resources. Usually, these cause us no difficulty. People make reasonable decisions every day.Footnote 64 But heuristics and biases will occasionally lead us wildly astray.
Among the more common heuristics is ‘availability,’ or the tendency to associate frequency or risk with ease of recall. The availability heuristic largely explains why people overestimate the dangers of flying: we can all vividly recall ghastly examples of past aviation disasters.Footnote 65 Another is the ‘representativeness heuristic,’ or the tendency to make judgments of likelihood based on something’s apparent similarity to a class.Footnote 66 This is why someone speaking Arabic on a cellphone might trigger panic on a Southwest Airlines flight in the post-9/11 world.Footnote 67 Another is ‘anchoring and adjustment,’ or the tendency to assign estimates close to a seeded value.Footnote 68 People’s beliefs about the likelihood or severity of global warming, for example, can be manipulated by experiences of, or stories about, unusually high temperatures.Footnote 69 A fourth is the ‘affect heuristic,’ or the tendency (inter alia) to overestimate the risks and costs of things that induce fear or dread and underestimate the risks – and overestimate the benefits – of things that induce positive emotions.Footnote 70 These are just a few examples. There are other heuristics and biases as well, and evidence that they often interact to compound errors in a fascinating variety of ways.Footnote 71
Scaling up from the individual to the group, there is ample evidence that social, cultural, and racial factors can also affect threat perception. Some of the most interesting studies exploring these dynamics do so with reference to attitudes toward immigration – certainly a highly salient issue today, and one that has been widely securitized. Findings on this score are an interesting mix of unsurprising and counterintuitive. With respect to the former, much of the relevant literature confirms the basic insights of Social Identity Theory, which posits ‘othering’ out-groups as a mechanism of promoting in-group self-esteem.Footnote 72 Out-groups are easier to ‘other,’ moreover, in a context of ethnic or religious difference, or when people are rendered strange by lack of direct personal contact. In contrast, contact and familiarity can breed comfort and promote welcoming attitudes.Footnote 73 Similarly, racial difference inflates threat perception and racial similarity deflates it.Footnote 74 These insights would not astonish a lay observer. On the other hand, one might expect national attachment (patriotism or chauvinism) to correlate more strongly with anti-immigrant sentiment and threat perception in an ethno-nationalist state than in a republican or multicultural one, but this does not appear to be the case – nor does anti-immigrant threat perception correlate with the restrictiveness of national migration policies.Footnote 75 To explain this requires referring to the specific content, rather than the generic form, of national political cultures. Finally, no doubt many would expect immigration to trigger threat perceptions most strongly when people see immigrants as competition for scarce resources or opportunities, but the evidence suggests otherwise. ‘Realistic’ (material) threat is far less powerful than ‘symbolic’ threat.Footnote 76 An International Relations theorist might not find this particularly surprising, of course, as it would comport with what we know about ontological security.Footnote 77 But an economist would find it mystifying.
All of this helps explain why we systematically overestimate the threat of terrorism, underestimate the threat of climate change, and respond to both inappropriately. The 9/11 attacks provided a shockingly graphic availability hook. One survey shortly afterward showed that the average American felt that they had a greater than one-in-five chance of being hurt in a terrorist attack within the next year,Footnote 78 something that would have required nearly 30,000 perfectly distributed 9/11-scale attacks, or roughly 80 per day. The surge of Islamophobia that followed the attacks can no doubt be attributed in large part to representativeness: after all, the world’s 1.5 billion Muslims all resemble the nineteen 9/11 hijackers on the dimension of religious affiliation. A good-quality base-rate analysis would have flagged young males of any religious affiliation as a much greater source of terrorist violence than Muslims in general, but would have assigned members of both groups extremely low likelihood of committing terrorist acts.Footnote 79 The global supply of transnational terrorists is much smaller than people think. The affect heuristic helps us understand why people assigned unrealistically high estimates of the likelihood of terrorist attacks post-9/11. Terrorism is something that inspires a particular sense of horror and dread. And standard in-group/out-group dynamics can help explain why xenophobia and racism might amplify Islamophobia among those not inoculated against it by personal and social relationships with Muslims. Climate change triggers none of these things. Small wonder that resource allocations to both threats are hugely disproportionate and responses largely inappropriate.
The fixes for threat misperception begin, of course, with knowledge of its causes and attentiveness to its perils and pitfalls. The fact that we can know that we have overreacted to the threat of terrorism and underreacted to the threat of climate change testifies to this. Accurate threat perception requires good-quality inputs, awareness of the vital importance of base rate information, and dispassionate analysis – sometimes simple and sometimes sophisticated, depending on the problem at hand. These are all vital for appropriate securitization.
But they are not sufficient. Appropriate securitization still requires securitizing moves by securitizing actors with the requisite access, voice, and skills. Getting the right resources to the right problems is, as the Copenhagen School teaches us, an inherently political task in which narrative construction, persuasiveness, and charisma all play a vital role.Footnote 80 But it also requires something yet more: knowledge of what is worth securing, and why. This is something no quantity of data and no political or rhetorical process can supply. For this we need a deep dive into the theory of value and a tour of potentially important security referents. These, respectively, are the subjects of the next and subsequent chapters.
