20 Introduction
Notes
1. Apossible third contender is the newly launched “critical security studies,”
committed to seeking alternatives to realist, statist, and positivist orthodoxies. Some
of what follows might be seen as fitting that description, but we have no prior commitment
to antistate or antirealist positions, and we are driven more by methodological
collectivism than by methodological individualism. More on critical security
studies in Chapter 2.
2. We are aware that in some other literatures the term region has a different
meaning fromours. The term was originally introduced at the subunit level. In nineteenth-century
France, a political movement formulated regionalism as an ideal for
political organization that was located in the middle of the continuum between centralized
government and political autonomy. This politicized notion of the region
lives on in separatist movements. Also, contemporary journals like Regional
Politics and Policy (published since 1990), International Regional Science Review
(since 1975), Journal ofRegional Science (since 1958), andRegional Studies (since
1967) are devoted primarily to the situation of ethnic minorities in specific subunit
regions and to issues of administration and planning at different political levels—
that is, political centralization and decentralization. Additionally, there is a Europe
of the regions: The contemporary map of the EU is subdivided not only into states
but also into thousands of smaller units (a Swissification of Europe) and also
increasingly into a variety of transnational “regions” (the Baltic Sea region, the
Alpe-Adriatic, and the like), which in our terminology would appear as subregions
and transregions, respectively. In this study, region refers to what that other literature
sometimes calls maCroregions (cf. Joeimiemi 1993, 1997).
3. The security complex is not objective in the sense of “independent of
actors.” In much traditional security analysis, region is defined “objectively” purely
in terms of geography or history (cf. current debates about whether Russia is a part
of Europe). In this sense, a region is simply an arena for security and one that is not
influenced by security policies—the analyst observes “objective” reality and tells
the actors to which region they belong. In contrast, security complexes are specifically
defined by security interactions among units. Since we argue that security is
not an objective issue but a product of the behavior of actors, security complexes
are not objective in the"traditional sense. Nor is the security complex to be seen as a
discursive construction by the actors. We are not (in this context) interested in
whether the actors define themselves as a region or whether they claim that their
true region is something larger or smaller. Security complexes do not require that
their members think in terms of the concept security complex (cf. note 6, Chapter
2). Analysts apply the term security complex (and therefore designate a region)
based upon the contingent, historically specific, and possibly changing constellation
generated by the interdependent security practices of the actors. On this basis, lines
can be drawn on a map, and the theory canbe put into operation.
CHAPTER 2
Security Analysis:
Conceptual Apparatus
What Is Security?
What quality makes something a security issue in international relations? It
is important to add the qualification “in international relations,” because
the character of security in that context is not identical to the use of the
term in everyday language. Although it shares some qualities with “social
security,” or security as applied to various civilian guard or police functions,
international security has its own distinctive, more extreme meaning.
Unlike soeial security, which has strong links to matters of entitlement and
social justice, international security is more firmly rooted in the traditions
of power politics. We are not following a rigid domestic-international distinction,
because many of our cases are not state defined. But we are claiming
that international security has a distinctive agenda.1
The answer to what makes something an international security
issue can be found in the traditional military-political understanding of
security. In this context, security is about survival. It is when an issue is
presented as posing an existential threat to a designated referent object (traditionally,
but not necessarily, the state, incorporating government, territory,
and society). The special nature of security threats justifies the use of
extraordinary measures to handle them. The invocation of security has been
the key to legitimizing the use of force, but more generally it has opened
the way for the state to mobilize, or to take special powers, to handle/existential
threats. Traditionally, by saying “security,” a state representative
declares an emérgency condition, thus claiming a right to use whatever
means are necessary to block a threatening development (Wæver 1988,
1995b).
When we consider the wider agenda, what do the terms existential
threat and emergency measures mean? How, in practice, can the analyst
draw the line between processes of politicization and processes of securitization
on this basis? Existential threat can only be understood in relation to
the particular character of the referent object in question. We are not dealing
here with a universal standard based in some sense on what threatens
individual human life. The essential quality of existence will vary greatly
22 Security Analysis
across different sectors and levels of analysis; therefore, so will the nature
of existential threats.
In the military sector, the referent object is usually the state, although it
may also be other kinds of political entities. It is also possible to imagine
circumstances in which threats to the survival of the armed forces would
elevate those forces to referent object status in their own right, perhaps
serving to justify a coup against the existing government and its policy
(whether of disarmament or of hopeless conflict). Traditional security studies
tends to see all military affairs as instances of security, but this may not
be the case. For many of the advanced democracies, defense of the state is
becoming only one, and perhaps not even the main de facto, function of the
armed forces. Their militaries may be increasingly trained and called upon
to support routine world order activities, such as peacekeeping or humanitarian
intervention, that cannot be viewed as concerning existential threats
to their states or even as emergency action in the sense of suspending normal
rules.
In the political sector, existential threats are traditionally defined in
terms of the constituting principle—sovereignty, but sometimes also ideology—of
the state. Sovereignty can be existentially threatened by anything
that questions recognition, legitimacy, or governing authority. Among the
ever more interdependent and institutionalized relations characteristic of
the West (and increasingly of the international system as a whole), a variety
of supranational referent objects are also becoming important. The
European Union (EU) can be existentially threatened by events that might
undo its integration process. International regimes, and international society
more broadly, can be existentially threatened by situations that undermine
the rules, norms, and institutions that constitute those regimes.
In the economic sector, the referent objects and existential threats are
more difficult to pin down. Firms are most commonly existentially threatened
by bankruptcy and sometimes by changes to laws that make them illegal
or unviable (as after communist revolutions). But in the market economy
firms are, with few exceptions, expected to come and go, and only
rarely do they try to securitize their own survival. National economies have
a greater claim to the right of survival, but rarely will a threat to that survival
(national bankruptcy or an inability to provide for the basic needs of
the population) actually arise apart from wider security contexts, such as
war. Unless the survival of the population is in question, the huge range of
the national economy doing better or doing worse cannot be seen as existentially
threatening. As in the political sector, supranational referent
objects from specific regimes to the global market itself can be existentially
threatened by factors that might undermine the rules, norms, and institutions
that constitute them.
In the societal sector, as we have defined it, the referent object is largescale
collective identities that can function independent of the state, such as
Security Analysis 23
nations and religions. Given the peculiar nature of this type of referent
object, it is extremely difficult to establish hard boundaries that differentiate
existential from lesser threats. Collective identities naturally evolve and
change in response to internal and external developments. Such changes
may be seen as invasive or heretical and their sources pointed to as existential
threats, or they may be accepted as part of the evolution of identity.
Given the conservative nature of “identity,” it is always possible to paint
challenges and changes as threats to identity, because “we will no longer be
us,” no longer the way we were or the way we ought to be to be true to our
“identity.” Thus, whether migrants or rival identities are securitized
depends upon whether the holders of the collective identity take a relatively
closed-minded or a relatively open-minded view of how their identity is
constituted and maintained. The abilities to maintain and reproduce a language,
a set of behavioral customs, or a conception of ethnic purity can all
be cast in terms of survival.
In the environmental sector, the range of possible referent objects is
very large, ranging from relatively concrete things, such as the survival of
individual species (tigers, whales, humankind) or types of habitat (rain
forests, lakes), to much fuzzier, larger-scale issues, such as maintenance of
the planetary climate and biosphere within the narrow band human beings
have come to consider to be normal during their few thousand years of civilization.
Underlying many of these referent objects are baseline concerns
ubout the relationship between the human species and the rest of the biosphere
and whether that relationship can be sustained without risking a collupse
of the achieved levels of civilization, a wholesale disruption of the
planet’s biological legacy, or both. The interplay among all of these factors
is immensely complicated. At either the macro or the micro extreme are
some clear cases of existential threat (the survival of species, the survival
of human civilization) that can be securitized. In between, somewhat as in
Ihe economic sector, lies a huge mass of problems that are more difficult,
although not impossible, to construct in existential terms.
Securitization
"Security” is the move that takes politics beyond the established rules of
the game and frames the issue either as a special kind of politics or as
above politics. Securitization can thus be seen as a more extreme version of
politicization. In theory, any public issue can be located on the spectrum
ranging from nonpoliticized (meaning the state does not deal with it and it
In not in any other way made an issue of public debate and decision)
through politicized (meaning the issue is part of public policy, requiring
government decision and resource allocations or, more rarely, some other
form of communal governance) to securitized (meaning the issue is pre­
24 Security Analysis
sented as an existential threat, requiring emergency measures andjustifying
actions outside the normal bounds of political procedure). In principle, the
placement of issues on this spectrum is open: Depending Upon circumstances,
any issue can end up on any part of the spectrum.2 In practice,
placement varies substantially from state to state (and also across time).
Some states will politicize religion (Iran, Saudi Arabia, Burma) and some
will not (France, the United States). Some will securitize culture (the former
USSR, Iran) and some will not (the UK, the Netherlands). In the case
of issues (notably the environment) that have moved dramatically out of the
nonpoliticized category, we face the double question of whether they have
merely been politicized or have also been securitized. This link between
politicization and securitization does not imply that securitization always
goes through the state; politicization as well as securitization can be enacted
in other fora as well. As will be seen later, it is possible for other social
entities to raise an issue to the level of general consideration or even to the
status of sanctioned urgency among themselves.
In this approach, the meaning of a concept lies in its usage and is not
something we can define analytically or philosophically according to what
would be “best.” The meaning lies not in what people consciously think the
concept means but in how they implicitly use it in some ways and not others.
In the case of security, textual analysis (Wæver 1988, 1995b, 1995c)
suggests that something is designated as an international security issue
because it can be argued that this issue is more important than other issues
and should take absolute priority. This is the reason we link the issue to
what might seem a fairly demanding criterion: that the issue is presented as
an existential threat. If one can argue that something overflows the normal
political logic of weighing issues against each other, this must be the case
because it can upset the entire process of weighing as such: lfIf we do not
tackle this problem, everything else will be irrelevant (because we will not
be here or will not be free to deal with it in our own way).” Thereby, the
actor has claimed a right to handle the issue through extraordinary means,
to break the normal political rules of the game (e.g., in the form of secrecy,
levying taxes or conscription, placing limitations on otherwise inviolable
rights, or focusing society’s energy and resources on a specific task).
“Security” is thus a self-referential practice, because it is in this practice
that the issue becomes a security issue—not necessarily because a real existential
threat exists but because the issue is presented as such a threat.
Of course, places do exist where secrecy or violation of rights is the
rule and where security arguments are not needed to legitimize such acts.
The earlier illustrations were for a liberal-democratic society; in other societies
there will also be “rules,” as there are in any society, and when a securitizing
actor uses a rhetoric of existential threat and thereby takes an issue
out of what under those conditions is “normal politics,” we have a case of
Security Analysis 25
securitization. Thus, the exact definition and criteria of securitization is
constituted by the intersubjective establishment of an existential threat with
u saliency sufficient to>have substantial political effects. Securitization can
be studied directly; it does not need indicators. The way to study securitization
is to study discourse and political constellations: When does an argument
with this particular rhetorical and semiotic structure achieve sufficient
effect to make an audience tolerate violations of rules that would otherwise
have to be obeyed? If by means of an argument about the priority and
urgency of an existential threat the securitizing actor has managed to break
free of procedures or rules he or she would otherwise be bound by, we are
witnessing a case of securitization.
Even if the general logic of securitization is clear, we have to be precise
about its threshold. A discourse that takes the form of presenting something
as an existential threat to a referent object does not by itself create
securitization—this is a securitizing move, but the issue is securitized only
if and when the audience accepts it as such. (Accept does not necessarily
mean in civilized, dominance-free discussion; it only means that an order
always rests on coercion as well as on consent. Since securitization can
never only be imposed, there is some need to argue one’s case.) We do not
push the demand so high as to say that an emergency measure has to be
adopted, only that the existential threat has to be argued and just gain
enough resonance for a platform to be made from which it is possible to
legitimize emergency measures or other steps that would not have been
possible had the discourse not taken the form of existential threats, point of
no return, and necessity. If no signs of such acceptance exist, we can talk
only of a securitizing move, not of an object actually being securitized. The
distinction between a securitizing move and successful securitization is
important in the chapters that follow.
Securitization is not fulfilled only by breaking rules (which can take
many forms) nor solely by existential threats (which can lead to nothing)
but by cases of existential threats that legitimize the breaking of rules. Still,
we have a problem of size or significance. Many actions can take this form
on a small scale—for example, a family securitizing its lifestyle as dependent
on keeping a specific job (and therefore using dirty tricks in competition
at the firm) or the Pentagon designating hackers as “a catastrophic
threat” and “a serious threat to national security” (San Francisco
( 'lironicle, May 23, 1996: All), which could possibly lead to actions within
the computer field but with no cascading effects on other security issues.
<)ur concept of international security has a clear definition of what we are
interested in, but it does not tell us how we sort the important cases from
the less important ones. We do not want to sort by arbitrarily assigning
degrees of importance to referent objects and sectors, for instance, defining
Mule as more important than environment or military as more securitylike
26 Security Analysis
than identity. Doing so would undermine the logic of both widening the
security agenda and taking a securitization approach to that agenda. It
would constrain arbitrarily and a priori what we can see and thus make it
impossible to capture the extent to which the security agenda has actually
changed or been widened.
A better measure of importance is the scale of chain reactions on other
securitizations: How big an impact does the securitizing move have on
wider patterns of relations? A securitizing move can easily upset orders of
mutual accommodation among units. The security act is negotiated between
securitizer and audience^—that is, internally within the unit—but thereby
the securitizing agent can obtain permission to override rules that would
otherwise bind it. Typically, the agent will override such rules, because by
depicting a threat the securitizing agent often says someone cannot be dealt
with in the normal way. In the extreme case—war—we do not have to discuss
with the other party; we try to eliminate them. This self-based violation
of rules is the security act, and the fear that the other party will not let
us survive as a subject is the foundational motivation for that act. In a securitized
situation, a unit does not rely on the social resources of rules shared
intersubjectively among units but relies instead on its own resources,
demanding the right to govern its actions by its own priorities (Wæver
1996b). A successful securitization thus has three components (or steps);
existential threats, emergency action, and effects on interunit relations by
breaking free of rules.
The distinguishing feature of securitization is a specific rhetorical
structure (survival, priority of action “because if the problem is not handled
now it will be too late, and we will not exist to remedy our failure”). This
definition can function as a tool for finding security actors and phenomena
in sectors other than the military-political one, where it is often hard to
define when to include new issues on the security agenda. Must new issues
affect the military sector or be as “dangerous” as war (Deudney 1990)? To
circumvent these restrictive ties to traditional security, one needs a clear
idea of the essential quality of security in general.
That quality is the staging of existential issues in politics to lift them
above politics. In security discourse, an issue is dramatized and presented
as an issue of supreme priority; thus, by labeling it as security, an agent
claims a need for and a right to treat it by extraordinary means. For the
analyst to grasp this act, the task is not to assess some objective threats
that “really” endanger some object to be defended or secured; rather, it
is to understand the processes of constructing a shared understanding of
what is to be considered and collectively responded to as a threat.
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) (Wæver 1988; Austin 1975:
98ff,).
Security Analysis 27
Sectors and Institutionalization of Security
What we can study is this practice: Who can “do” or “speak” security successfully,
on what issues, under what conditions, and with what effects? It
is important to note that the security speech act is not defined by uttering
the word security. What is essential is the designation of an existential
threat requiring emergency action or special measures and the acceptance
of that designation by a significant audience. There will be instances in
which the word security appears without this logic and other cases that
operate according to that logic with only a metaphorical security reference.
As spelled out later, in some cases securitization has become institutionalized.
Constant drama does not have to be present, because it is implicitly
assumed that when we talk of this (typically, but not necessarily, defense
issues), we are by definition in the area of urgency: By saying “defense”
(or, in Holland, “dikes”)* one has also implicitly said security and priority.
We use this logic as a definition of security becauseitf has a consistency and
precision the word as such lacks. There is a concept of international security
with this specific meaning, which is implied in most usages of the word.
Our claim is that it is possible to dig into the practice connected to this
concept of security in international relations (which is distinct from other
concepts of security) and find a characteristic pattern with an inner logic. If
wc place the survival of collective units and principles—the politics of
existential threat—as the defining core of security studies, we have the
basis for applying security analysis to a variety of sectors without losing
the essential quality of the concept. This is the answer to those who hold
that security studies cannot expand its agenda beyond the traditional military-political
one without debasing the concept of security itself.
Sectors are “views of the international system through a lens that highlights
one particular aspect of the relationship and interaction among all of
its constituent units” (Buzan, Jones, and Little 1993: 31). Given that the
analytical purpose of sectors is to differentiate types of interaction (military,
political, economic, societal, and environmental), it seems reasonable
to expect (1) that one will find units and values that are characteristic of,
and rooted in, particular sectors (although, like the state, they may also
appear in other sectors); and (2) that the nature of survival and threat will
differ across different sectors and types of unit. M other words, security is a
generic term that has a distinct meaning but varies in form. Security means
Niirvival in the face of existential threats, but what constitutes an existential
IIneat is not the same across different sectors. One purpose of the following
chapters is to unfold this sectoral logic of security more fully.
Securitization can be either ad hoc or institutionalized. If a given type
of threat is persistent or recurrent, it is no surprise to find that the response
and sense of urgency become institutionalized. This situation is most visible
in the military sector, where states have long endured threats of armed
coercion or invasion and in response have built up standing bureaucracies)
28 Security Analysis
procedures., and military establishments to deal with tkose threats. Although
such a procedure may seem to reduce security to a species of normal politics,
it does not do-s©> The. need for drama in establishing securitization
falls apway, because it is implicitly assujned that when we talk of this*j$sue
we are by definition area of urgency. As is the ^fse for defense,MSues
in most countries and for the dikes in the Netherlands, urgency has been
(ef^lished,; by the previous^use of the security move. There fg| no further
need to fgell oust that this issue has to take precedence, that£|Ét#;security
issue—by saying “defense” or “dikes,” one has also implicitly said ‘(security"’
an(^“priority.” This can be shown by trying to inquire about the rationale
Jør decisions in thege areas. Behind the first layers of ordinary bureaucratic
arguments* onejgdll ultimately find 4?f4probably irritated—repe^^in
of a security argument so wgljhpstablished that it is taken for granted.
Sojme security practices aresnot legitimised in public b^ sec^i|f disCQj|fi|fe
because they are not out in the public at all (e.g.,.jtih;e “black programs”
in the United States, which are not presented in the budget), but this
is actually a very clear case of the security logic.|§j|ja democracy, at some
point it must be argued in the public sphere why a situation constitutes
security and therefore can legitimately be handled differently. One could
not take something out budget without giving a reason for the use of
sfph ^> effaor4iPW procedure. When this procedure has been legitimized
through security rhetoricB ttbecomes institutionalized as a package legiS
imization, and i^j|s thfpi possible to have black security boxes in theipolitical
proccss. The speech act reduces public influence on this- issue,*but in
democracies one must legitimize in public why from now on the details
\\jd|ljgiot be.presented publicly (because of the danger of giving useful informgÉjm
to the enemy and the like). In all cases, the establishment of secret
services some element of this.logical sequence. Not every act is presented
with the drama of urgency and priority, because it has been estaty-,,
listed in a general sense that this k an entire field that has been moved to a
form of treatment legitimate only because this area has been defined as
security.
In well-developed stft|es*v-armed forces and intelligence services are;
carefullyfteparated from normal political life, and their use isi4|bject t®a
elaborate procedures of authorization. Where such separation is||$)$'»|
place, as in many weak states (.Nigeria under Abacha, the USSR under
Stalin) or Mutates mob||jzed for totalgwar, much of normal p itie s is i
pyshed into tj|e security realm.^The prominence of institutionalize||iøjMd
tary security underpins not only the claim of those who want to confine |
security studies to the military sector but also the de facto primacy of the
state in security affairs. But nothing is necessary about this particular construction;
it comes out of a certain history and has formidable institutional
momentum but is not fixed for all time. Where the threat profiles warrant
them, one can see other types of institutionalized security structures, such
as those concerned with flood control in the Netherlands. One of the diffiSecurity
Analysis 29
cullies facing those attempting to securitize environmental issues is that the
threats are both new (or newly discovered) and controversial regarding
their existential urgency.;;iConsequen||^„fhey do nofelyet) have institutions*
and they find themselves operating in a-political context dominated by
necurity institutions designed for other types of threat. Although
(in one-sense securitization is a further intensification of
politicization (thus usually making an even stronger role for the state), in
miother sense it is opposed to politicizgtion. Politicization means to make
tin issue appear to be open, a matter of choicer/something that is decided
upon and that therefore entails responsibility, in contrast to issues that
either could not be different »(laws of nature) or should not be put under
political control (e.g.* a free economy* the private sphere, and matteSffor
expert decision)||By contrast* securitization on therinternational level
(although often not on the domestic one) means to present an issue as
urgent and existential, as so important that it should not be exposed to the
noimal haggling of politics but should be dealt with decisively by top feadet'Nprior
to other issues.
National security should not be idealized. It works to silence oppositimi
and has given power holders many opportunities to exploit “threats”
for domestic purposes, to claim a right to handle something with less demo-
11illic control and constraint. Our belief*itherefore,?i$ not ‘fthe more security
the better.” Basically, security should be seen as negative, as a failure to
ili iil with issues as normal politics. Ideally, politics should be able to unfold
according to routine procedures without this extraordinary elevation of spei
Ilk "threats” to a prepolitical immediacy. In some cases securitization of
lynnes is unavoidable, as when states are faced with an implacable or barI*«ii
i i i i i aggressor. Because of its prioritizing imperativ*©* securitization also
11its tactical attractions-—for example, as a way to obtain sufficient attention
fin environmental problems. But desecuritization is the optimal long-range
option, since it means not to have issues phrased as “threats against which
Wt have countermeasures” but to move them out of this threat-defense
Nii|uence and into the ordinary public sphere (Wæver 1995b).
When considering securitizing moves such as “environmental securityp
in m"war on crime,” one has to weigh the always problematic side effects of
applying a mind-set of security against the possible advantages of focus,
•mention, and mobilization. Thus, although in the abstract desecuritization is
ilie ideal, in specific situations one can choose securitization—only one
should not believe this is an innocent reflection of the issue being a. security
tliM'Hl; it is always a political choice to securitize or to accept a securitization.
Subjective, Objective, and Intersubjective Security
I shading the essential quality of international security lakes one some
wav toward pinning down a general but nonetheless still fairly confined
30 Security Analysis
meaning of the concept that can operate both within and beyond the traditional
m ilitary-political understanding of that concept. But this does not
solve all of the problems. Commentators on security at least as far back as
Arnold Wolfers (1962: 151) have noted that security can be approached
both objectively (there is a real threat) and subjectively (there is a perceived
threat) and that nothing ensures that these two approaches will line
up. This distinction turns out to be crucial in formulating an international
security concept for a multisectoral agenda.
Our argument is that securitization, like politicization, has to be understood
as an essentially intersubjective process. Even if one wanted to take a
more objectivist approach, it is unclear how this could be done except in
cases in which the threat is unambiguous and immediate. (An example
would be hostile tanks crossing the border; even here, “hostile” is an
attribute not of the vehicle but of the socially constituted relationship. A
foreign tank could be part of a peacekeeping force.) It is not easy to judge
the securitization of an issue against some measure of whether that issue is
“really” a threat; doing so would demand an objective measure of security
that no security theory has yet provided. Even if one could solve the measurement
problem, it is not clear that the objectivist approach would be particularly
helpful. Different states and nations have different thresholds for
defining a threat: Finns are concerned about immigration at a level of 0.3
percent foreigners, whereas Switzerland functions with a level of 14.7 percent
(Romero 1990).4
Regardless of whether an analyst finds that an actor’s disposition
toward high or low thresholds leads to correct assessments, this disposition
has real effects. And other actors need to grasp the logic this unit follows.
When states or nations securitize an issue—“correctly” or not—it is a political
fact that has consequences, because this securitization will cause the
actor to operate in a different mode than he or she would have otherwise.
This is the classical diplomatic (and classical realist) lesson, which holds
that good statesmanship has to understand the threshold at which other
actors will feel threatened and therefore more generally to understand how
the world looks to those actors, even if one disagrees (Carr 1939; Kissinger
1957; Wæver 1995d).
In some cases, however, it does matter how others judge the reasonableness
of a securitization, because this influences how other actors in the
system will respond to a security claim. What may seem a legitimate securitization
within a given political community may appear paranoid to those
outside it (e.g., Western perceptions of Soviet concerns about pop music
and jeans). Conversely, outsiders may perceive that a political community
undersecuritizes a “real” threat and thus endangers itself or free rides (e.g.,
U.S. perceptions of Danish defense policy during the Cold War). The way
the securitization processes of one actor fit with the perceptions of others
about what constitutes a “real” threat matters in shaping the interplay of
Security Analysis 31
Nccurities within the international system. Both within and between actors,
Ihe extent of shared intersubjective understandings of security is one key to
understanding behavior.
In any case, 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. Such rationalist universalism will easily
he “right” on its own terms, but it will be of very little help in political
analysts. It is more relevant to grasp the processes and dynamics of securilizution,
because if one knows who can “do” security on what issue and
under what conditions, it will sometimes be possible to maneuver the internetion
among actors and thereby curb security dilemmas.
The distinction between subjective and objective is useful for highlighting
the fact that we want to avoid a view of security that is given
objectively and emphasize that security is determined by actors and in this
respect is subjective. The label subjective, however, is not fully adequate.
Whether an issue is a security issue is not something individuals decide
nlone. Securitization is intersubjective and socially constructed: Does a referent
object hold general legitimacy as something that should survive,
which entails that actors can make reference to it, point to something as a
threat, and thereby get others to follow or at least tolerate actions not otherwise
legitimate? This quality is not held in subjective and isolated minds; it
i n a social quality, a part of a discursive, socially constituted, intersubjecllve
realm. For individuals or groups to speak security does not guarantee
Niiccess (cf. Derrida 1977a; Wæver 1995b). Successful securitization is not
decided by the securitizer but by the audience of the security speech act:
Does the audience accept that something is an existential threat to a shared
vnlue? Thus, security (as with all politics) ultimately rests neither with the
objects nor with the subjects but among the subjects (cf. Arendt 1958,
1959; Wæver 1990; Huysmans 1996).
Social Power and Facilitating Conditions
This relationship among subjects is not equal or symmetrical, and the possibility
for successful securitization will vary dramatically with the position
held by the actor. Security is thus very much a structured field in which
dome actors are placed in positions of power by virtue of being generally
HOcepted voices of security, by having the power to define security (Bigo
IW4f 1996, forthcoming). This power, however, is never absolute: No one
Inguaranteed the ability to make people accept a claim for necessary secuilly
action (as even the Communist elites of Eastern Europe learned; see
Wæver 1995b), nor is anyone excluded from attempts to articulate alternative
interpretations of security. The field is structured or biased, but no one
conclusively “holds" the power of securitization.5Therefore, it is our view
32 Security Analysis
(contra Bigo) that one can not make the actors of securitization the fixed
point of analysis—the practice of securitization is the center of analysis. In
concrete analysis, however, it is important to be specific about who is more
or less privileged in articulating security. To study securitization is to study
the power politics of a concept.
Based on a clear idea of the nature of security, securitization studies
aims to gain an increasingly precise understanding of who securitizes, on
what issues (threats), for whom (referent objects), why, with what results,
and, not least, under what conditions (i.e., what explains when securitization
is successful).
The impossibility of applying objective standards of securityness
relates to a trivial but rarely noticed feature of security arguments: They are
about the future, about alternative futures—always hypothetical—and
about counterfactuals. A security argument always involves two predictions:
What will happen if we do not take “security action” (the threat), and
what will happen if we do (How is the submitted security policy supposed
to work?). A security theory that could tell politicians and citizens what
actually constitute security problems and what do not would demand that
such predictions should be possible to make on a scientific basis, which
means society would have to be a closed, mechanical, and deterministic
system. Even this condition, however, would not be enough, because a second
complication is that securityness is not only a matter of degree—“how
threatening”—but is also a qualitative question: Do we choose to attach the
security label with its ensuing effects? Actors can choose to handle a major
challenge in other ways and thus not securitize it. The use of a specific conceptualization
is always a choice—it is politics, it is not possible to decide
by investigating the threat scientifically.
An objective measure for security can never replace the study of securitization,
because the security quality is supplied by politics, but this does
not mean a study of the features of the threat itself is irrelevant. On the contrary,
these features rank high among the “facilitating conditions” of the
security speech act. Facilitating conditions are the conditions under which
the speech act works, in contrast to cases in which the act misfires or is
abused (Austin 1975 [1962]). Conditions for a successful speech act fall
into two categories: (1) the internal, linguistic-grammatical—to follow the
rules of the act (or, as Austin argues, accepted conventional procedures
must exist, and the act has to be executed according to these procedures),
and (2) the external, contextual and social—to hold a position from which
the act can be made (“The particular persons and circumstances in a given
case must be appropriate for the invocation of the particular procedure
invoked” [Austin 1975 (1962): 34]).
A successful speech act is a combination of language and society, of
both intrinsic features of speech and the group that authorizes and recognizes
that speech (Bourdieu 1991 [1982]; Butler 1996a, b). Among the
Security Analysis 33
internal conditions of a speech act, the most important is to follow the security
form, the grammar of security, and construct a plot that includes existential
threat, point of no return, and a possible way out—the general grammar
of security as such plus the particular dialects of the different sectors,
such as talk identity in the societal sector, recognition and sovereignty in
the political sector, sustainability in the environmental sector, and so on (cf.
Wæver 1996b). The external aspect of a speech act has two main conditions.
One is the social capital of the enunciator, the securitizing actor, who
must be in a position of authority, although this should not be defined as
official authority. The other external condition has to do with threat. It is
more likely that one can conjure a security threat if certain objects can be
referred to that are generally held to be threatening—be they tanks, hostile
Nentiments, or polluted waters. In themselves, these objects never make for
necessary securitization, but they are definitely facilitating conditions.
After thus subdividing the social, external speech-act conditions into
nctor authority and threat related, we can sum up the facilitating conditions
as follows: (1) the demand internal to the speech act of following the grammar
of security, (2) the social conditions regarding the position of authority
for the securitizing actor—that is, the relationship between speaker and
audience and thereby the likelihood of the audience accepting the claims
made in a securitizing attempt, and (3) features of the alleged threats that
cither facilitate or impede securitization.
Actor and Analyst in Securitization Studies
Approaching security from a speech-act perspective raises questions about
the relationship between actors and analysts in defining and understanding
the security agenda. As analysts, we define security as we have done here
because it is the only way that makes coherent sense of what actors do. We
have identified a particular sociopolitical logic that is characteristic of
security, and that logic is what we study. Although analysts unavoidably
play a role in the construction (or deconstruction) of security issues (viz.,
the long argument between peace research and strategic studies or the U.S.
debate about the securityness of the Vietnam War), it is not their primary
task to determine whether some threat represents a “real” security problem.
Objective security assessment is beyond our means of analysis; the
main point is that actors and their audiences securitize certain issues as a
specific form of political act. Actors who securitize do not necessarily say
'Security,” nor does their use of the term security necessarily always constitute
a security act. We use our criteria to see if they take the form of “politics
of existential threats,” with the argument that an issue takes priority
over everything else and therefore allows for a breaking of the rules. As a
Ilist step, the designation of what constitutes a security issue comes from
34 Security Analysis
political actors, not analysts, but analystsinterpret political actors’ actions
and sort, out when these actions fulfillghe security criteria. It is, further, the
analyst who judges whether the, actor is effective in mobilizing support
around this security reference (i.e., the attempted securitizers are “judged”
first by other soeifg|actors;and citizen% and the degree of their following is
the® interpreted and measured by u^j Finally, to assess the significance of
an instance of sepuritization, analysts study its effects on other units. The
actor commands at only one very crucial step: the performance of a political
act in a sfeuritpmode.
■.Thus, it |f the actor, not th,e analyst who decides whether something is
to be handled as an existential threapThis does not make analysts hostage
to the self-understanding of actors for £he duration of the analysis. In all
subsequent questions of cause-effect relationships—what are the effects of
these security acts, who influenced decisions, and so on—we do not intend
to give actors any defining role. Thus, a concept such as security complex jg,
defined not by whet]ggr actqtf labelj|h£mse|ves a complex (they do nottø.
but by analystpinterpretation of who is actually interconnected in terms of
securitp interactionA(Security complex is basically an analytical term;
security is a political practic%Jliat we have distilled into a specific, more
precise category on the basis of the way the concept is used.) The speechact
approach says only tha$||4s the actor who by securitizing an issue—and,
the audience by- acccpting the claim—makes it a security issue. At that
level, the analyst cannot and should not replace the actor.
This point does not suggest that we feel obliged to agree with this
securitizing act. Onefiaf the purposes of this approach should be that
becomes possible to evaluate whether one finds|| good or bad to securitize
a certain issue. One rarely manages to counter a securitizing attempt by
spyingåt-ananalvgtBBfou are not really threatened, you only think so.” But
|Jj§S possible t®|ji|g|$gith some force;whether it is a good idea to make this
ilfitl a security issue—to transfer it Jg the agenda of panic politics—or
whether it is better handled within normal politics. As witnessed in the discussion
about environmental security, even environmentalists have had
strong second thoughts aboff the effect®*of putting the environmental agenda
in security terms. The securitization;approach serves to underline the
responsibility of talking security, the responsibility of actors as well as of
analysts who choose to frame an issue as a security jifsue. They cannot hide
behind the claim that anything in iitsei®constitutes ^ security issue.
The relationship of analyst to actor is one area in whigfc pur approach
differs from that taken by many scholars with whom we share some theoretical
premises. An emerging school of “critical secwlty studies”,(CSS)
wants to challenge conventional security studies by applying postpositivist
perspectives, such as critical theory and poststructuralism (Krause and
Williams 1996, 1997). Much of its work, like ours, deals with the social
construction of security (cl. also Klein 1994; Campbell 1993), but CSS
mostly has the intent (known from poststructuralism as well as from conSecurity
Analysis 35
structivism in international relation^ of showing that change is possible
because things are socially constituted.
We, in contrast, believe even the socially constituted is often sedimented
as structure and becomes so relatively stable as practice that one must do
analysis also on the basis that it continues,,igsing one’s understanding of the
social construction of security not only to criticize this fact but also to
understand the dynamics of security and thereby maneuver them. This
leads us to a stronger emphasis on collectivities and on understanding
thresholds that trigger securitization in order to avoid them. With our securitization
perspective, we abstain from attempts to talk about what “real
security” would be for people, what are “actual”;security problems larger
than those propagated by elites, and the like. To be able to talk about these
issues, one has to make basically different ontological choices than ours
and must define some emancipatory ideal. Such an approach is therefore
complementary to ours; it can do what we voluntarily abstain from, and we
can do what it is unable to: understand the mechanisms of securitization
while keeping a distance from security—that is, not assuming that security
is a good to be spread to ever more sectors.
There are other differences between the two approaches (much of CSS
takes the individual as the true reference for security—human security—
and thus in its individualism differs from our methodological collectivism
and focus on collectivities; cf. Chapter 9), but the political attitude and its
corresponding view of constructivism and structuralism is probably the
most consistent one. The analyst in critical security studies takes on a larger
burden than the analyst in our approach; he or she can brush away existing
security construction disclosed as arbitrary and point to some other issues
that are more important security problems. Our approach links itself more
closely to existing actors, tries to understandsjtheir modus operandi, and
assumes that future management of security will have to include handling
these actors—-as, for instance, in strategies aimed at mitigating security
dilemmas and fostering mutual awareness in security complexes. Although
our philosophical position is in some sense more radically constructivist in
holding security to always be a political construction and not something the
analyst can describe as it “really’lls, in our purposes we are closer to traditional
security studies, which at its best attempted to grasp security constellations
and thereby steer them into benign interactions. This stands in contrast
to the “critical” purposes of CSS, which point toward a more
wholesale refutation of current power wielders.
The Units of Security Analysis:
Actors and Referent Objects
The speech-act approach to security requires a distinction among three
types of units involved in security analysis.
36 Security Analysis
1. Referent objects: things that are seen to be existentially threatened
and that have a legitimate claim to survival.
2. Securitizing actors: actors who securitize issues by declaring something—|f
referent object—existentially threatened.
3. Functional actors: actors who affect the dynamics of a sector.
Without being the referent object or the actor calling for security on
behalf of the referent object, this is an actor who significantly influences
decisions in the field of security. A polluting company, for
example, can be a central actor in the environmental sector—it is
not a referent object and is not trying to securitize environmental
issues (quite the contrary
The most important and difficult distinction is that between referent
objects and securitizing actors, and this distinction requires some discussion.
We deal with functional actors in the sector chapters.
The referentobjectTor security has traditionally been the state and, in a
more hidden wap the n^|on.#or a State, survival is about sovereignty, and
for a nation itBNMxrat identity (Wæver et al. 1993, chapter 2). But if one
follows the securitization approach outlined earlier, a much more open
spectrum of possibilities has to be allowed. In principle, securitizing actors
can attempt to construct anything as a referent object. In practice, however,
the constraints of facilitating conditions mean actors are much more likely
to be successful with some types of referent objects than with others.
Security action is usually taken on behalf of, and with reference to, a collectivity.
The referent object ffjthat to which one can point and say; “It has
to survive* therefore it is necessary to . . . ”
Size or scale seems to be one crucial variable^ determining what Constitutes
a successful referent object of security. At the micro end of the
spectrum, individuals or small groups can seldom establish a wider security
legitimacy*-j®&thei®fown right. They may speak about security to and of
themselves, but few will listen. At the system end of the scale, problems
also exist in establishing security legitimacy. For example, attempts have
been made to construct ill of humankind as a security referent-pmost:
notably in terms of shared fears of nuclear annihilation during the Cold War
but also in the context of environmental fears. Another system-scale
attempt was the failed move by socialists in 1914 to mobilize in the name
of the international working class. Thus far, however, the system level has
rarely been able to compete with the middle scale, although this does not
mean it will not become more attractive in the future as international circumstances
change.
In practice, the middle scale of limited collectivities has proved the
most amenable to securitization as durable referent objects. One explanation
for this success is that such limited collectivities (states, nations, and,
as anticipated by Huntington, civilizations) engage in self-reinforcing rivalSecurity
Analysis 37
ries with®ther limitefiieoiieolHvities, and suchtinteraeiion strengthens their
“wé” feeling. Becausf&they involve aÆeférence to aHwe,” they are social
constructs operative in the interaction among people. A main criterion of
this type of referent is that it forms an interpretative community—it is the
context in wlÉiK principles of legitimacy and valuationl^rculate and within
which the individual constructs an interpretation of events. The referent rs a
social context with the dignity of a^“site .of judgment” (Foucault 1979). If
rivalry is a facilitating condition; for successful securitizatioriljlmiddle-level
collectivities1will! always have an advantage over the systemfjj|velilli this
respect. Somehow, the system-levéfecandidates are SKUtoo sublléiand indirect
to trigger the levels of. mass identity necessary for secutlfization.
Lacking the dynamic underpinning of rivalry, their.attempt at universalist
political allegiance confronts the middle-level collectivities and loseSt>%f|
The apparent primacy of the middlc-Hevel, limited collectivities opens
the way for an attack on our approach from traditional state-centric security,
analysts (and perhaps also from certain types of liberals). Their argument
goes like this: Security, by definition, is and should be about the state, and
the state is and should be about security, with*the emphasis on military and
political.security. A hard-line liberal might say the state has no legitimate
functions other than security; When security is expanded beyond the state,
we. have problematic securitizations such as environmental security; when
th©state expands beyond security, we have problems such as the conflation
of economic security with prOtectionismaltt is?j§assible to take the statesecurity
position and argue poSAiiiea'My against all attem ptto “do” security
with reference to other referent objects On the ground tMt only through the
state can the process of securitization be controlled democratically and kept
in check.
We acknowledge that there is some analytical truth, as well as a legitftf
mate political position, in this tight link between state and security. But the
logic of our approach forces us to reject the use of such a narrow and selfclosing
definitional move. We have constructed a wider conceptual net
within which the state-centric poltiOn is a possible but not a predetermined
outcome. In using tbiSfécheme, one majåip&fin&fhit the state is the most
importanifecurity referent; if so, Ibis-finding would carry much more force
than if it were made true by definition and would also remain open to
change. We do not say security is only about the state (although there is
much truth to the argument that the state is the ideal security actor) nor that
security is equally available to all-fjfetates and other, social motement$IP
Security is an area of competing actors, but it is å biased one in whicfotthe
state is still generally privileged as the actor historically endowed with
security tasks and most adequately structured for the purpose. This explanation
acknowledges the difference between a state-centric approach and a
state-dominated field.
But whereas the middle level in general, and the state in particular,
38 Security Analysis
might enjoy primacy in the selection of referent objects, that is not the end
of the story. Being a middle-level, limited collectivity is insufficient for
achieving status as a referent object. This is probably best illustrated in the
case of economic security, where one would think firms are the natural limited
collectivity units. But by their very nature, firms rarely have a strong
claim to a right of survival. If the survival of a firm is threatened, the firm
will not be able to: legitimize action beyond the normal, legal rules of the
game. We rarely see middle-fcvel security policy in this field except when
economic arguments: can be linked to what in economic terms is the secondary
unit—the state—which can claim a natural right to survive, to
defend its existence, and to take extraordinary measures (protectionism and
the like) on a national issue (such as maintaining the capability for military
mobilization) if deemed necessary.
Nor do system-level' referent objects always lose out. Thus far they
have done so in the military and political sectors, where the security of
humankind has generally had less appeal than that of the state. But the
story is different in other sectors. The environment is becoming an interesting
case, because groups are using a securitizing logic that exactly follows
the format prescribed in the previous section: The environment has to survive;
therefore, this issue should take priority over all others, because if the
environment is degraded to the point of no return all other issues will lose
their meaning. If the normal system (politics according to the rules as they
exist) is not able to handle this situation, we (Greenpeace and especially the
more extremist ecoterrorists) will have to take extraordinary measures to
save the environment. Sustainability might be the environmentalists equivalent
of the state’s sovereignty: and the nation’s identity; it is the essential
constitutive principle that has to be protected. If this idea catches on, the
environment itself may be on the way to becoming a referent object—an
object by reference to which security action can be taken in a socially significant
way. We discuss this more fully in Chapter 4.
Once this door is opened, one can see other plausible candidates for
security referent objects at the system level. Humankind as a whole
achieved some status as a referent object in relation to nuclear weapons and
could do so again—perhaps more successfully in relation to environmental
disasters, such as new ice ages or collisions between -the earth and one
or more of the many large rocks that occupy near-earth space. The level of
human civilization could also become the referent object in relation to environmental
threats. In the economic sector, system-level referents may be
more effective vehicles for security discourse than limited collectivities,
such as the firm and the state. Already, systems of rules or sets of principles,
such as “the liberal world economy” and “free trade,” have some status
as referent objects in the economic sector. A similar practice could grow
in the political sector around international society or democracy (the latter
as an extension of the democracy = peace hypothesis). Our position is that
Security Analysis 39
no principled, logical exclusion of referent objects should take place at the
system level; therefore, we investigate the issue in each of the sector chap-
ters.
Also, the individual is again>a factor in security debate. As argued by
Ken Booth ^199®1994j||995), muc^iofpscurity analysis blanks outfsÉhe
effects on actual human beings of the issues discussed; thus, his argument
is an attempt to securitize‘concrete individuals in their competition ^ith
aggregate categories. Emma Rothschildffl995) has argued that historically,
a major part of liberal thought had the individual as the referent ofSecurity;
thus, there is a respectable philosophical tradition to build on. In the 1980sJ
with projects like the Brandt and Palme CommisSions|gseo#|fp?thought
drifted back toward the individual, and Rothschild argues convincingly that
regardless bf whether it is intellectually coherent or ethically idealj?securitization
of the individual is a real political practice of' our times;, (In this
book, the lidividual wall reappear primarily;in the pfeiBcål-sector chapter,
because it is usually. a question of establishing the principle of, for example,
human rights rather than of specific individuals appearing one by one
asssecuritized referent objects.8)
To concludc, one can study security discourse to learn what referent
objects are appealed to and can study outcomes to see which hold security
legitimacy So an appeal to their necessary survival is able to mobilize support.!Traditionally,
the middle level has been the most fruitful generator of
referent objects, but lately more has been heard about system- and microlevel
possibilities (Rothschild 1995). Referent objects must establish secffe
rity legitimacy in terms of a claim to survival. Bureaucracies, political
regimes, and firms seldom hold this Sense of guaranteed survival and thus
are not usually classed as referent objccts. Logically, they could try to
establish a claim to Survival and thus to security legitimacy, but empirically
this is not usually possible. In practice, security is not totally subjective.
There are socially defined'limits to what can and cannot be securitized,
although those limits can be changed. This means security analysis is interested
mainly in successful Instances of securitization—the cases in which
other people follow the secufitteing lead, creating a social, intersubjective
constitution of a referent object on a mass scale. Unsuccessful or partially
successful attempts at seci#å#zatioø are interesting primarily for the
insights they offer into the stability of social attitudes toward security legits
imacy, the process by which those attitudes are maintained or changed, and
the possible future direction of security politics. In these larger patterns,
desecuritization is at least as interesting as securitization, but .theffrccesppl
acts of securitization take a central place because they constitute the currently
valid specific meaning of security.
Critics will undoubtedly protest our abdication of the critical use of
objective security measures as a way to question dominant definitions (cf.
McSweeney 1996). When a threat is not securitized, should one not be able
40 Security Analysis
to show that this is a threat? Yes, the securitization perspective, which basically
removes the objective ground from the dominant discourse, opens the
possibility of problematizing both actual securitization and the absence of
securitization, but it cannot do so by proving that something “is” a security
problem—at least not without shifting from the role of analyst to securitizing
actor. Thus, it is not advisable to add to our basic securitization perspective
that there are also objective security problems (to hold against
false securitizations and the lack thereof). Doing so would introduce an
incompatible ontology that would ultimately undermine the basic idea of
security as a specific social category that arises out of, and is constituted in,
political practice.
What one can add are arguments about the likely effects.9 One can try
to show the effects of either excessive securitization—security dilemmas—
or of not securitizing—the inability to handle an issue effectively unless it
is securitized. Only within society and by one’s own participation in political
practice can one contribute to securitization or desecuritization, which
is a different matter from the threat “being” a security problem. Things can
be facilitators of securitization—it is made easier if one can point to matters
associated with threats, but the ultimate locus of securityness is social
rather than technical, and it is between a securitizing actor and its audience
in reference to something they value.
A securitizing actor is someone, or a group, who performs the security
speech act. Common players in this role are political leaders, bureaucracies,
governments, lobbyists, and pressure groups. These actors are not usually
the referent objects for security, because only rarely can they speak
security through reference to the need to defend their own survival. Their
argument will normally be that it is necessary to defend the security of the
state, nation, civilization, or some other larger community, principle, or
system. Only occasionally will actors such as governments or firms be able
to speak successfully of security on their own behalf.
The notion of an “actor” is in itself problematic. To say precisely who
or what acts is always tricky, because one can disaggregate any collective
into subunits and on down to individuals and say, “It is not really ‘the state’
that acts but some particular department—or in the last instance individuals.”
But to disaggregate everything into individuals is not very helpful,
because much of social life is understandable only when collectivities are
seen as more than the sum of their “members” and are treated as social realities
(methodological collectivism).
Identifying actors is thus more complicated than identifying referent
objects. The former involves a level-of-analysis problem: The same event
can be attributed to different levels (individual, bureaucracy, or state, for
instance). Unlike the case with the referent object, a speech act is often not
self-defining in terms of who or what speaks, and the designation “actor” is
thus in some sense arbitrary. Ultimately, individuals can always be said to
Security Analysis 41
be the actors, but if they are locked into strong roles it is usually more relevant
to see as the “speaker” the collectivities for which individuals are designated
authoritative representatives (e.g., parties, states, or pressure
groups)—for example, France-materialized-as-de Gaulle rather than the
person de Gaulle. If one wants to downgrade the role of the analyst in
defining actors, one option is to let other actors settle the matter. Other
states treated de Gaulle as acting on behalf of France and held France
responsible for his acts; thus, in the world of “diplomatics” France was
constituted as the actor (Manning 1962; Wæver forthcoming-c). How to
identify the securitizing actor is in the last instance less a question of who
performs the speech than of what logic shapes the action. Is it an action
according to individual logic or organizational logic, and is the individual
or the organization generally held responsible by other actors? Focusing on
the organizational logic of the speech act is probably the best way to identify
who or what is the securitizing actor.
The difference between actor and referent object in any specific case
will also usually mean there is a separate category of “audience,” those the
securitizing act attempts to convince to accept exceptional procedures
because of the specific security nature of some issue. One danger of the
phrases securitization and speech act is that too much focus can be placed
on the acting side, thus privileging the powerful while marginalizing ithose
who are the audience and judge of the act (Huysmans 1996).
One use of the distinction between actors and referent objects is to
avoid reifying some security units—for example, nations. When we say in
the chapter on societal security (and in Wæver et al. 1993) that societal
security is often about nations and their survival, we do not want to say that
“a nation acts to defend itself,” which would represent reifying and anthropomorphic
terminology. Someone—some group, movement, party, or
elite—acts with reference to the nation and claims to speak or act on behalf
of the nation.
The distinction between securitizing actor and referent object is less of
a problem in the context of the state and therefore has not previously been
clearly noted. The state (usually) has explicit rules regarding who can
speak on its behalf, so when a government says “we have to defend our
national security,” it has the right to act on behalf of the state. The government
is the state in this respect. No such formal rules of representation exist
for nations or the environment; consequently, the problem of legitimacy is
larger in these areas than in the case of the state. When someone acts in the
name of a nation, certain discursive rules are imposed on the actor, because
he or she has to speak in terms of identity, in terms that follow the logic of
“nation,” and these terms shape the discourse and action in a way that differs
from that appropriate to other referent objects. But only in the weakest
sense does this mean the nation is “acting.” The rules for what one can do
in the name of a nation are less rigid than those for a state; therefore, it will
42 Security Analysis
be easier to talk of the state acting than of the nation doing so. This is a
matter of degree rather than necessarily a qualitative difference. Consequently,
the analyst who writes about a fringe neo-Nazi group that tries to
mobilize people to defend “our national survival” against the threat posed
by immigrants will feel uncomfortable phrasing this as “the nation acting.”
It feels more correct to make the distinction between who actually does the
acting and what those actors are referring to as that which should survive
and then see how successful they are in asserting a claim to speak for that
higher entity.
These arguments show why it is important to distinguish between securitizing
actors and referent objects. But the distinctions are contextual
rather than intrinsic to specific units: In many cases, the securitizing actors
will be different from the referent object, but in others—most notably the
state—the referent object will in a sense speak for itself through its authorized
representatives. In all cases, however, the analyst is obliged to question
the success or failure of the securitizing speech act. Even governments
can fail at securitization, as happened to Britain over the Suez, the United
States in Vietnam, and the European Communist regimes domestically in
the late 1980s.
In applying the distinction among referent objects, securitizing actors,
and functional actors to the five sector chapters that follow, it is important
first to clarify the referent object(s) in each sector. In some cases, this will
constitute most of the exercise. To map societal security around the world,
it is probably more interesting—and at least logically primary—to know
where people are mobilized in the name of nations, civilizations, religions,
or tribes than to know where mobilization is effected by political parties,
where by state elites, where by social movements, where by churches, and
where by intellectuals. In the military sector, the referent object may almost
always be the state, and the securitizing actor may in some sense also be
“the state,” but a number of functional actors may also influence decisions.
If so, one would need to spend more space tracking down these functional
actors. Thus, the sector chapters will vary in terms of the weight of analysis
given to the three types of security unit. In an ideal situation—perhaps in
more complete future case studies based on this approach—-all three types
will be covered fully, in particular the articulation of referent objects and
securitizing actors.
Regions and Other Constellations of Securitization
In the part of this work aimed at tracing security complexes, the approach
is to look at the pattern of security connectedness. The investigation proceeds
in three steps: (1) Is the issue securitized successfully by any actors?
Security Analysis 43
(2) If yes, track the links and interactions from this instance—how does the
security action in this case impinge on the security of others, and where
does this then echo significantly? (3) These chains can then be collected as
a cluster of interconnected security concerns. When this case along with the
patterns from all of the other cases (of the sector in the case of homogeneous
sector-specific analysis or across sectors in the case of heterogeneous
security complex analysis; cf. Chapters 1 and 8) are aggregated, we
can see the level on which the processes of securitization and the patterns
of interaction are concentrated.
Our general assumption, and one of the key motivations for this project,
is that the post-Cold War world will exhibit substantially higher levels
of regionalization and lower levels of globalization than was the case during
the Cold War. One of qur purposes is to adapt security complex theory
to deal with this more complicated world. In the sector chapters that follow,
however,,we keep this question open. It may be that the security logic of
some sectors inherently inclines toward regionalization, whereas in other
sectors it does not. This is what we need to investigate in these chapters.
And we do so in basically the same way as is done in classical security
complex theory: by combining the concerns of major actors into a constellation,
a knot of mutual security relations.
One final problem in thinking about security regions is how to tie such
thinking into the discussion of actors and referent objects in the previous
section. Is a security complex defined by actors or referent objects? As just
argued, the security complex is actually a constellation of security concerns;
the different instances of securitization as such form the nodes
among which the lines can be drawn and the complex mapped. Because referent
objects are the more basic, enduring, and salient features on the security
landscape; the answer to our earlier question is the referent objects.
Some might object that according to our scheme referent objects do not act
and therefore cannot be the units in subsystems that are defined by interactions.
This is an illusion. Security actors speak and act in the name of referent
objects, and they generally see threats as emanating from other referent
objects. There is thus a real sense in which India and Pakistan, Turkey and
the Kurds, or Chile and ITT interact.
Since referent objects are the socially constituted units, they are often
actors for each other, even if some analytical theories point to other links in
the chains as the actors. For instance, states are to some extent real as states
and they act as states even if the literal acting is done by statesmen, because
states ascribe intentions and responsibility to each other as states (M anning
1962; Wæver forthcoming-c). This reflection is structured by the motivation
of security complex analysis, which is to reach a dynamic analysis of
security situations. We want to be able to grasp the connections between the
security of A and that of B, the security dilemmas as well as mutually rein­
44 Security Analysis
forcing security loops. Therefore, it is essential that we organize the regional
analysis around nodes that are simultaneously that which is (claimed to
be) threatened and that which is (depicted as) the source of threat.
In classical security complex theory (CSCT), the definition was
phrased in terms of primary security concerns; in the current framework, it
must be instances of securitization that connect and form the complex. In
both cases, the core is obviously the articulation of threats by the major
actors. Unfortunately, there is little conceptual literature on threats. In discussions
of the concept of security, some participants claim an actor-based
threat is a precondition for something to be a security problem (Deudney
1990). It is difficult to see what justifies this as a logical step, although it
could be an empirical connection, a structural proclivity making threats
attributed to actors more easy to securitize. We do not, however, want to
define security problems such that actors have to be the problem. Probably,
they usually are.
It follows from our general securitization perspective that what interests
us is the attribution of security problems to specific sources rather than
the actual origins of what appear as security problems. As argued by attribution
theory, there is a general psychological tendency to overestimate the
degree of choice for alter while emphasizing necessity as to ego (Hart
1978; Jervis 1976). One will therefore generally tend to “actorize” the other
side—that is, fashion the other as a willful chooser rather than a chain in a
series of events. In most cases, the fact that the other is a strategic actor
with several choices is an amplifying factor in any threat perception and
therefore assists in pushing an issue across the security threshold. Because
the other is an actor, not just a wheel in a machine, it has the potential of
outwitting us, of having intentions, or of bending or suppressing our will to
replace it with its own (cf. Clausewitz 1983 [1832]; Wæver 1995b).
This focus on actors could seem to point to securitizing actors rather
than to referent objects. This deduction, however, is probably false. What
the attribution argument implies is not that we should focus on those units
we see as actors but rather that whatever is presented as the cause of security
problems is most likely also actorized. If securitizing actor “a” on behalf
of community “A” claims A is threatened by B, he or she will present B as
an actor, as responsible for the threat, as an agent who had a choice.
Therefore, we do not have to define security complexes in terms of what
we have labeled actors in our analytical framework: The actors might operate
with other actors and thereby point to the bigger, more abstract categories—the
referent objects. On the other hand, threats do not need to be
attributed to the same categories as those the other side acted with reference
to. Actual events are likely to be varied and complex, requiring a pragmatic
approach that allows us to find the specific units of the case.
For instance, Churchill as a securitizing actor could have securitized
Nazism as a threat. This does not necessarily mean a countersecuritization
Security Analysis 45
is performed either by Nazism as actor or with Nazism as referent object.
Instead, Hitler could securitize England (the referent object of Churchill, so
far so good) as the threat in the name of Germany, all Germans, and the
Aryan race. What constitutes the threat for one is not necessarily the referent
object for the other. This procedure was much easier in CSCT where
security was conducted for and by India, which was also the (perceived)
threat to Pakistan and vice versa. The argument from attribution theory
gives us reason to believe that most threats will be linked to actors and that
what we analyze as referent objects will often be constructed by other
actors as actors. If, however, one draws the map too finely, a number of
actors will be securitizing slightly different referent objects (the German
race, the German people, Germany, Aryans)—differences that are important
when one is trying to look into the politics of securitizing moves—
whereas we in security complex analysis need to find the main patterns of
interaction and therefore need to bundle together the various versions of
securitizing “Germany” as one node.
When generating the security complex, the best way to define the
points between which the security arrows go might be to point to conglomerates
of a referent object and the corresponding securitizing actor. In the
extreme case, this means we have referent objects with stable spokespersons.
A stable combination of referent object and “voice” points to the classical
concept of the state as a clear instance. But even the state and sovereignty
as referent object is appealed to by other than the one official voice.
There are several actual securitizing actors, and the state as well as the
other actors occasionally securitize other referents, such as the nation, the
European Union (EU), or some principles of international society. In the
case of France, Japan, and Sudan, the name makes a relatively clear reference
to a dense network of correlated referent objects and securitizing
actors. The different securitizing actors are connected by competing for the
representation of the same referent object; the different referent objects are
unified by their mutual substitutability for each other. There is more a chain
of family resemblances than a clear-cut criterion or one primary unit. In
each case, a conglomerate of actors and referent objects is unified by the
density of overlapping security discourse and usually also nominally by a
name: the security of “France” (which can mean several different referent
objects and a large number of possible actors), of Europe and the EU, and
of “the environment.” (See the further discussion on pp. 171-175.)
The key question in security analysis is, who can “do” security in the
name of what? For a time, experts could get away with analyzing only
“states,” and the system was then the sum of the states. Regional security
meant the sum of national securities or rather a particular constellation of
security interdependence among a group of states. The approach developed
here offers more types of units to choose from, but the basic idea of security
complexes can be carried over into a world of multiple units.
46 Security Analysis
Notes
1. The history of the word security is complex (KaUfmann 1970; Der Derian
1993; Delumeau 1986; Corze 1984), but in the 1940s it was established in international
affairs with a fairly distinct meaning (Rosenberg 1993). Much of this meaning
was so easily installed because it rested on an old argument that had used the
word security much less systematically—an argument about “necessity” previously
contained primarily in the concept of raison d’état (Butterfield 1975). Especially
from the mid-nineteenth century, when the state enters ajuridical self-limitation and
self-control, this “is balanced by the designation of a range of ‘governmental acts’
which are immune to legal challenge. This juridical reserve area of executive power
is . . . the qualification which . .. calculations of security impose as a condition for
the political feasibility of a liberal democracy” (Gordon 1991: 33; cf. Foucault 1991
[1978]). The classical argument, which holds that in extreme cases the government
can use all means necessary, becomes concentrated as a specific, exceptional case
(Wæver 1988, 1995b). This meaning of security evolved separately from the use of
security in various domestic contexts (although connections definitely exist; see
Kaufman 1970). This international type of security starts to spread to new referents
and new actors; therefore, we want to retain ,a focus on international security
because it has a distinct meaning, but we do not exclude the possibility that we will
meet this kind of security increasingly in domestic contexts.
2. This argument does not imply that private issues could not in some sense be
political, an argument made forcefully by feminists. To claim such is a politicizing
move.
3. The concept of strong and weak states is elaborated and defined in Buzan
(1991: 96-107) and rests on the degree of sociopolitical cohesion within the state,
which is high for strong states and low for weak ones. The concept should not be
confused with the distinction between strong and weak powers, which is about their
capabilities vis-a-vis other powers.
4. Baldwin (1997) is the most sophisticated and consistent attempt to define
security and to structure security studies according to the idea that the purpose and
task is to assist decisionmakers in correctly assessing the relative attention to devote
to different threats.
5. The importance of “cultural capital” to the ability to perform a speech act
has been argued by Pierre Bourdieu (1991 [1982]). A speech act is not only linguistic;
it is also social and is dependent on the social position of the enunciator and
thus in a wider sense is inscribed in a social field. However, Bourdieu made this
argument to counter a tendency of some poststructuralists and philosophers of
everyday language to make the purely linguistic, internal features of a speech act
completely determining (Bourdieu 1996). He has accepted the critique by Judith
Butler (1996a, b) that since the speech act needs to include an idea of—with his
own phrase—the “social magic” whereby some are accepted as holding authority
and others are not, it has to be indeterminate, open for surprises. This is not purely a
question of a formal position of authority (Austin’s example in which “I declare you
man and wife” is an effective speech act only when performed by a properly authorized
authority; 1975 [1962]: 8-15). There is a performative force to the speech act;
to use Bourdieu’s own concepts, it has a magical efficiency, it makes What it says. A
speech act is interesting exactly because it holds the insurrecting potential to break
the ordinary, to establish meaning that is not already within the context—it reworks
or produces a context by the performative success of the act. Although it is important
to study the sociul conditions of successful speech acts, it is necessary always
to keep open the possibility that an act that had previously succeeded and for which
the formal resources and position are in place may fail and, conversely, that new
Security Analysis 47
actors can perform a speech act they had previously not been expected to perform
(Butler 1996a, b; Derrida 1977a [1972], 1977b, 1988). Therefore, the issues of
“who can do security” and “was this a case of securitization” can ultimately be
judged only in hindsight (Wæver et al. 1993: 188). They cannot be closed off by
finite criteria for success.
6. This stands in contrast to some other studies of regions where one is interested
in the construction of regions by actors (Neumann 1994; Joenniemi and
Wæver 1992; Joenniemi 1997). Both approaches to regions are relevant, but for different
purposes.
7. For those interested in pinpointing our position within the field of international
relations theory, this is probably the passage to pick. We do not take the state
or sovereignty as representing fixed limits, but we are skeptical of individualism as
the traditional alternative to state centrism. We therefore form a picture of a world
of multiple units, which might be called postsovereign realism. The units can be
overlapping (in contrast to the exclusivity of sovereign territorial states), but this
does not necessarily lead to any benign transnationalism in which the focus is on
the multiple identities of individuals relativizing all units and collectivities.
Although each individual in a world of overlapping units is a “member” of several
units, instead of focusing on any such softening effects produced by overlap, we
study how the units can continue to conduct power politics; think, for example, of
the work of Susan Strange (state-firm diplomacy; 1994) and Robert Kaplan (a very
anarchic anarchy after sovereignty; 1994). Each unit has a possibility of becoming
the reference for security action, but since the different units overlap and are placed
at different levels, there is no fixed line between domestic and international—what
is internal to one unit can be interunit when one thinks of other units. More importantly
a distinction exists between individual and collective security. This argument
is important for the present project, because if domestic and international were
fixed, there would be a risk of generating a cozy Western view of politics: Domestic
politics is normal and without security, whereas the extreme is relegated to the
international space. In other parts of the world, domestic is not cozy. This fact can
be grasped by focusing on those units and collectivities that are mobilized in such
contexts: These domestic security relations are interunit because in these places the
most powerful referent objects are smaller than the state.
8. One can contemplate cases in which concern seems to focus on a particular
individual: one girl in Sarajevo or Salman Rushdie. To a large extent, these individuals
are given such prominence and more resources are spent on them than on most
others because they are taken to represent principles. Action for some specific individual
always depends on a construction of that person as representing some category,
as deserving protection because he or she belongs to a particular social category—for
example, leader, representative, free intellectual, or revealing test case.
9. The analyst can also intervene to countersay actors in relation to the use of
the word Security. Sloppy talk of “economic security” or “environmental security”
can be questioned by arguing that the security act has not really been performed and
that the securitizing actor has not managed to establish a case for treating the threat
as existential. Whether the threat really is or is not existential in relation to the referent
object is impossible to decide from the outside, but we can study the discourse
and see if the issue has been securitized in this sense. This is primarily an intervention
into the debate among observers over the appropriateness of the use of the
security label. When intervening in direct policy debates over a securitization, the
mode of argumentation will typically be in terms of comparing the likely effects of
having the issue securitized or desecuritized.