'Deepening' and 'Widening':  An Analysis of Security Definitions in the 1990s
Sarah Tarry

Combined Operation:  The Appropriation of Stoney Point Reserve and the Creation of Camp Ipperwash
P. Whitney Lackenbauer

Introduction

The end of the Cold War is often regarded as marking the beginning of a fundamentally different political environment both in a real and in a theoretical sense. This has been especially pronounced within Security Studies as pressure to redefine its most important concept, that of 'security' itself, has become a preoccupation for the past decade. As a result, several academics have reconsidered what is, and what should be, included within this concept and whether a broader definition constitutes a more accurate depiction of reality.

Reaching consensus on this "essentially contested concept" (Buzan 1991) has remained elusive, however, which is largely reflective of an increasingly entrenched cleavage along the so-called traditional versus non-traditional line. While traditionalists favour the maintenance of the Cold War conception of security - defined in military and state-centric terms - the non-traditionalists have attempted to broaden and deepen the definition. These non-traditionalists argue that other issues, such as economic, environmental and social threats, endanger the lives of individuals rather than strictly the survival of states.

The outcome is that the security debate has become fundamentally normative. Instead of focussing on what constitutes a methodologically sound and analytically meaningful definition, academics are engaged in a value-oriented debate about the actors that should be included and the issues that should be examined. Agreement on the definition of security is a prerequisite, however, before this concept can be usefully employed in understanding the phenomenon it is intended to capture.

Therefore, the purpose of my paper is to examine several existing definitions in regards to their methodological soundness, in order to provide an analytical point of departure for academics, regardless of their different ideological preferences. In this pursuit, a survey of the literature will focus on six influential perspectives within the contemporary security debate.

The classical ideological cleavage in International Relations, between the two schools of so-called realism and liberalism, is analytically insufficient to capture the nature of the debate in Security Studies. While some realists, such as Stephen Walt, continue to emphasize their traditional preoccupation with military threats, others, such as Barry Buzan and Mohammed Ayoob, agree that a broader definition of security is necessary. Moreover, the realist-liberal dichotomy cannot encompass the post-modernists, who are the most prominent critics of the traditional definition of security.

The analytical distinction between traditionalists and non-traditionalists is not meant to suggest that consensus among the latter has been reached either. Indeed, there is disagreement between two sub-groups - the so-called 'wideners' and 'deepeners' about the concept (Buzan/Wæver/de Wilde 1998; Krause/Williams 1996). The wideners argue that a predominantly military definition does not acknowledge that the greatest threats to state survival may not be military, but environmental, social and economic. The deepeners, on the other hand, ask the question of whose security is being threatened and support the construction of a definition that allows for individual or structural referent objects, as opposed to the state.

The purpose of a concise definition methodologically is that "we specify precisely what we mean when we use particular terms" (Babbie 1995: 114). The word 'security,' in itself, has no meaning. Assignment of content, which enables us to distinguish it from other concepts (cf. Salmon 1995: 57-8), is the important first step before theories can be developed to explain the causes and effects of phenomena labeled as 'security.'

A definition that is excessively broad, however, has the implication of preventing the concept from being analytically useful. Thus, this paper will measure each of the definitions against three criteria, irrespective of their normative underpinning. The first criterion is that a definition must provide a clear delineation of those cases which are included and those which are excluded from being a security concern. Second, the categorization of a definition must be mutually exclusive, thereby eliminating ambiguous cases that could fit in both or neither category. Third, a definition must have explanatory capacity, such as instances in reality, which actually do fulfill the requirements of being a security threat. At the same time, the category must not be conceptualized so broadly that almost everything can be subsumed under it.

The structure of the paper is as follows. Section I examines the traditionalist position, focussing on Stephen Walt's definition of security. Section II explores the non-traditionalist 'widening' perspective. Mohammed Ayoob, Ole Wæver, as well as Michael Klare and Daniel Thomas explain conceptions of security, which, although they include a broader range of threats than the traditionalists, are still largely state-centric. Section III explores two perspectives of non-traditionalists seeking to 'deepen' the definition of security. It includes the arguments made by Ken Booth on the one hand, and Spike Peterson on the other.

Each subsection will conclude with an application of the respective definition to Québec's threat of separation from Canada. This example is not so much chosen because of its domestic significance, but because it illustrates rather nicely the practical implications of the six definitions under examination. As the paper will show, they assess the potential of Québec independence very differently.

Section I: The Traditionalist Position

Stephen Walt

The definition that currently is initiating the greatest debate within the Security Studies community is that of Stephen Walt's traditionalist perspective (1991). Firmly rooted in realism, Walt articulates a position that is state-centric and restricts the application of 'security' to threats in the military realm. He equates security with peace and the prevention of conflict through military means (deterrence policies, non-offensive defense and the like).

Three realist assumptions figure prominently in Walt's definition. First, because of the absence of a world government, the international system is characterized by an objectively observable, i.e., not socially constructed, anarchy. Second, while states are not the only actors in the international system, they are the most important ones. Third, generalizations about state behaviour can be developed because the primary objective of all states is survival. As war is assumed to be the most serious hazard to the survival of states, the study of war becomes the focus of realism. In his article, "The Renaissance of Security Studies" (1991) Walt argues that Security Studies, and by extension security itself:

may be defined as the study of the threat, use and control of military force. It explores the conditions that make the use of force more likely, the ways that the use of force affects individuals, states and societies, and the specific policies that states adopt in order to prepare for, prevent, or engage in war (1991: 212; italics in original).

While one may challenge realist premises and assumptions, Walt's definition methodologically satisfies the first criterion by excluding cases along two dimensions. First, it is clear that the military realm is privileged and other potential threats are excluded; thus, unless an ecological, economic or social problem intrudes into the military realm, it does not comprise part of a state's security analysis. For example, the environmental problem of water shortage would only be considered a national security concern if a state threatened, or initiated, a military response against another state, in defense of its interests. Second, since the state is the referent object, Walt excludes sources of insecurity at the individual level. Thus, even if a government posed an immediate and lethal threat to its citizens, this would not be considered a security threat within this definition.

Because this terminology is unambiguous in the issues and referent object to which it applies, it also fulfils the second criterion. Since it is generally clear when military force is being employed by one state against another, either as a threat or in reality, the concept has a tangibility that is absent in other concepts such as justice and power. This clarity may help to explain, in part, the longevity of military-centred conceptions of security.

It is questionable, however, whether this definition meets the final criterion, which is the applicability to reality. As Mohammed Ayoob (1997) illustrates, traditional definitions assume that the greatest threats to national security are external ones. However, the overwhelming frequency of military force since 1945 has been in intrastate and not interstate conflicts. This trend has continued in the post-Cold War environment where the number of civil conflicts - former Yugoslavia, Rwanda and Somalia, to name a few - far surpass the number of interstate conflicts, which essentially only consist of the Gulf War. Furthermore, J. Ann Tickner (1995: 179) argues that of the approximately 127 "significant wars" which have occurred since 1945, all but two of them were within so-called Less Developed Countries (LDCs). Therefore, she accuses the traditionalists of being ethnocentric in their conceptualizations, as they narrowly define security in terms of conflict between the great powers.

Thus, because Walt's definition does not recognize intrastate conflict as being a source of insecurity, the potential separation of Québec from Canada does not qualify as a national security threat. Even if the Government of Canada used force to suppress a secessionist movement in Québec, this internal dispute would still not satisfy Walt's conditions. To do so would first require a declaration of sovereignty, and either Québec or the Rest of Canada (ROC) would then have to threaten or mobilize military force against the other.

In spite of its soundness on methodological grounds, the traditionalist definition of security itself is widely criticized by numerous academics. The most prominent source of discontent is the failure of Walt's conceptualization to meet satisfactorily the final criterion, applicability to the 'real' world. Based on different understandings of the nature of the post-Cold War environment, other scholars have developed alternative definitions of security in response to this security problématique.

Section II: The 'Wideners'

Mohammed Ayoob

Responding to the proportionately greater number of intrastate conflicts in the Third World as compared to the First World, Mohammed Ayoob's definition stresses political-institutional underdevelopment as the predominant source of conflict. He, therefore, addresses the third definitional criterion, which Walt does not meet. Ayoob (1997: 130) argues that national security is a function of state building, which requires that a state possesses more than simply "security hardware" (control of coercive force) but also "security software" (legitimacy and integration). Ayoob advances a different definition:

[s]ecurity or insecurity is defined in relation to vulnerabilities, both internal and external, that threaten to, or have the potential to, bring down or significantly weaken state structures, both territorial and institutional, and regimes (1997: 130; emphasis added).

In fulfilling the initial requirement of a sound definition, Ayoob's "subaltern realism" excludes cases in two ways. First, there is a clearly stated threshold for security threats, which any issue must pass to be considered as such. A problem needs to become sufficiently politicized to have the potential to threaten the survival of the state, its boundaries, political institutions, or governing regimes. Unlike Walt, Ayoob recognizes that a wide variety of crises may arise with the capacity to threaten the state, many of which are not exogenous. However, widening the security agenda is not an indiscriminate undertaking; Ayoob admits that not all issues have the same influence over time and among specific countries, thereby making it analytically inappropriate to attach issues like environmental degradation or migration to a definition of security.

Second, as its focus is clearly state-centric, Ayoob's definition, like Walt's, excludes threats to individual and global security. Hence, violence against women and environmental pollution are not national security matters unless those who profess to control the state view them as such. While Ayoob does not share Walt's commitment to anarchy as the primary influence on the behaviour of states, he justifies his state-centrism arguing that the state system remains the primary mode of organization. Therefore, security analysis should not undermine the move toward statehood, especially for LDCs (1997: 131).

Despite clearly meeting the first criterion, Ayoob's definition has some ambiguity. He appears to use the term 'regime' as it is used in the public debate, where it has the negative connotation of being an authoritarian government, rather than in the political science conception of regime as any type of government. Insecurity is defined in relation to threatening the continued existence of, among other things, regimes, so one would be inclined to assume that peaceful changes of government within democratic regimes threaten the survival of the state. However, as he is focussing his analysis on the weak state building in the Third World, it is clear within the broader context of the article that he does not mean to imply that regime change in an established democracy raises the same problems as in a LDC. Hence, while readers can only be expected to judge on the basis of what an author writes, and not on what he intends, this ambiguity could be easily resolved and, therefore, does not reduce the soundness of Ayoob's definition in principle.

In fulfilling the final definitional criterion, which is its pertinence in reality, Ayoob's broader conceptualization is superior to Walt's more narrow one. The ability to include a wider variety of potential threats allows Ayoob to account for a greater number of cases of violent post-Cold War intrastate conflicts. Thus, while this definition considers the Iraqi invasion of Kuwait in 1990 as a security threat to the territorial integrity of the latter, it also explains the Croat uprising as endangering the maintenance of the political institutions (and later, the territorial integrity) of the Yugoslav state (Ullman 1996).

Despite the subaltern context of the definition, it can also be applied to issues in the developed world. With respect to the threat of Québec's separation from Canada, this would be considered a matter of national insecurity for three reasons. First, separation would reduce the existing boundaries of Canada; second, it would, at a minimum, challenge the Canadian institution of the constitution; and finally, separation would threaten the current federal government, since several cabinet ministers have constituencies in Québec (most notably, the current Prime Minister).

Ole Wæver

According to Ole Wæver, the concept of security has two meanings. First it is used in everyday language to describe a freedom from threat, and second, it is employed in Security Studies to capture the survival of the state. While conceptualizing security in terms of 'individual security' or 'global security' have become more popular in the contemporary debate, Wæver insists that "as concepts, neither individual or international security exist" (1996: 48; emphasis in original). Since national security is the sole linguistically familiar conception, of the three, he views it as the only possible referent object. Wæver acknowledges, however, that dynamics at both the individual and international level can and do play a significant role in influencing national security.

While disregarding security at the individual level, Wæver advocates a "re-conceptualization of the security field in terms of a duality of state and societal security" (1995: 67). The primary objective of both state and society is survival. For the state, it is defined in terms of sovereignty, whereas for the society, it is defined in terms of identity. According to Wæver, societal insecurity is the result of an increasing incapacity of states to protect the interests of their citizens resulting from the progressive dispersion of political power (1995: 67).

In regards to both state and society, the threshold of any issue to be considered a security threat is if it is "securitized." The so-called securitization of political, economic or ecological problems occurs when the elite (either of the respective state or society) claims the problem to be such. In other words, Wæver argues that security is a speech act wherein only after an issue is defined by the elite as constituting a 'security threat' can it be considered as such. An issue becomes securitized once "a state-representative moves a particular development into a specific area, and thereby claims a special right to use whatever means are necessary to block it" (1995: 55). While the government has the responsibility of speaking 'security' on behalf of the state, Wæver also argues that the government is usually the orator of security for the 'society' as well. However, he admits that the relative success of societal elites who speak 'security' can only be determined in hindsight, as one considers whether the speaker had a significant amount of popular support among the society.

Wæver's two-dimensional conceptualization can only partially fulfill the criteria necessary to be considered a sound definition in methodological terms. On the one hand, like the previous definitions, this one excludes potential threats at the individual level and so avoids encompassing an infinite number of cases. As well, identifying the 'speakers' of security is easily operationalized in those cases where the use of the term by the elite is explicit, making it clear when specific cases are included and when they are not. This, of course, means that governments could invoke 'security' as a means to take unlimited measures against its own or neighbouring populations, when the very survival of the state was not threatened. However, according to Wæver, the 'objective reality' of the threat is less important than how states react to the perceived threat.

On the other hand, the concept of 'society' introduced by Wæver in his analysis is not as readily operationalized because it raises as many questions as it answers. What is the relationship between a society, a nation, and a state? Can more than one society exist within a state? Can societies overlap? What constitutes membership within a particular society? Against this backdrop, Immanuel Wallerstein (1984: 1-2) argues that the concept of 'society' cannot be empirically determined:

endless, passionate debate about which 'entities' constitute which concrete 'societies' throws fundamental doubt, it seems to me, on the utility of the concept of 'societies' as a starting point for analysis. States are at least visible, functioning organisms…but where can we find 'societies' other than in the minds of the analysts, or of the orators? Social science would, in my view, make a great leap forward if it dispensed entirely with the term.

Wæver contends that a national security threat can be viewed as "developments that threaten the sovereignty or independence of a state in a particularly rapid or dramatic fashion, and deprive it of the capacity to manage itself" (1995: 54; emphasis added). As a realist conception of interstate conflict, this latter definition neglects domestic challenges to state survival. The idea of state sovereignty as its benchmark seems to contradict security as a speech act, since the elite could speak economic, environmental or social issues as national security threats.

Wæver faces a final conceptual problem of categorization due to his association of society with identity. As Håkan Wiberg illustrates with the case of Macedonia:

to the extent identity is anchored in language, Bulgaria is the main threat: it regards Macedonian as a Bulgarian dialect…To the extent it is anchored in religion, the Serbs are the main threat: the Macedonian church [is] still under the Serb patriarchate…To the extent that it is anchored in statehood, the Albanian minority will not accept Macedonians defining themselves as the state carrying people. When it is defined by territory and history, the Greeks object strongly (quoted in Krause/Williams 1996: 244; emphasis in original).

Hence, identity remains variable, changing to reflect different circumstances or perceptions of threat. While Wæver may intend this in order to reflect the complexity of identity, it makes the exclusion of cases virtually impossible. States may consist of an infinite number of societies with potentially transnational and/or overlapping identities. Furthermore, the boundaries and constituents of a society could be subject to rapid changes in response to political, economic and other conditions.

Ambiguity, i.e., the second criterion, is a problem this definition cannot overcome. Wæver's article fails to make clear whether 'security' has to be explicitly spoken or whether the invocation of threats to 'sovereignty' and 'identity' are sufficient. This is especially relevant within the Canadian context, where the potential separation of Québec has not been 'spoken' as a national security threat by the federal government, but where the protection of sovereignty and identity have been frequently declared.

The employment of Wæver's two-dimensional concept of state and societal security to the situation in Québec reveals the complexity of the situation and the inherent contradictions between the interests of the state in relation to society. There are, broadly speaking, four dimensions in the application of this conception. First, for the federal government, Québec sovereignty may eventually pose a security threat to the maintenance of existing Canadian boundaries. Second, the survival of a 'Canadian society' based on a bilingual culture and tradition would also be threatened by the separation of Québec. However, from the perspective of Québec separatists, the third dimension of security in Wæverian terms, is that the Canadian state continues to stifle, and thereby threaten, the development of Québecois identity. Fourth, in Wæver's terminology, Québec consists of at least two different societies with diametrically opposed identities, i.e., one nationalist and the other separatist; however, the current premier claims to 'speak' identity for the entire province. Whether evolving political conditions in Québec cause a shift in the identification of its populace will, as the author admits, only be clear in hindsight, perhaps only after the next provincial election.

Michael Klare and Daniel Thomas

In contrast to Ayoob and Wæver, Klare and Thomas (1994: 3) contend that the concept of security needs to be expanded because of a declining significance of geographical boundaries. They perceive state actors to be less able to respond to global problems like environmental degradation and financial currency crises. Instead of focussing on domestic threats to state survival, they advocate a "world security" concept that accounts for the global nature of contemporary problems. To them, world security is

distinguished by the belief that security involves more than protection against military attack…ecological, economic and demographic trends pose serious challenges to [developed] countries. And even in the less-developed "South," where the threat of armed attack remains constant, nonmilitary trends pose equal or greater threats to people's security (1994: 4; emphasis added).

Through the inclusion of such varied global problems as human rights abuse, economic crises and ecological threats, it becomes obvious that Klare and Thomas view global security as equitable to human security. Their assumption is that all actors are influenced equally by global threats and that they are motivated to respond cooperatively to them. Such a conceptualization, however, ignores the contradictions of an all-encompassing blanket of threats. For example, banning deforestation and over-harvesting of farmland as a protective measure for the environment, would, at least in the short-term, reduce the economic viability of those people who depend on the land resources for food production.

In regards to fulfilling the first criterion of a 'good' definition, the possibility to exclude cases, Klare and Thomas' conceptualization of security falls short in three respects. First, the attachment to particular global problems prevents the definition from distinguishing between issues that are security threats and those that are not. As there is no specified threshold for operationalizing when an issue constitutes a threat, world security becomes an all-inclusive concept allowing the analyst excessive flexibility to include whatever s/he deems a threat to international security.

Second, the authors' referent object in the so-called developed "North" is the state, whereas in the underdeveloped "South," it is the people, thereby encompassing two different levels of analysis. This not only comprises a methodological problem in itself, but it also raises practical difficulties in the categorization of specific cases. This differentiation prevents the classification of certain regions or states, such as the 'Asian Tigers' and other Newly Industrialized Countries (NICs) from being clear-cut. To illustrate the point, Singapore and South Korea achieved their status as NICs under authoritarian and military governments respectively, on the basis that state security, i.e., the economic reforms initiated by the governments, could be threatened by its own citizens. Hence, the North-South dichotomy is insufficient because in both cases it is unclear whether the security concerns of the people or the state should prevail. In other words, apart from the methodological concerns mentioned above, where is the threshold for the shift in the focus from individual to state security?

Third, this conceptualization does not explore state interdependence, despite being heavily reliant on it as the ontology of world security. Interdependence, within this context is assumed to lead to positive outcomes wherein all actors benefit by the resolution of a global threat. In the normative sense of being a 'good thing,' interdependence assumes that states are symmetrically interdependent with the outcome of mutual solutions being equally beneficial for all actors. This argument does not take into account the possibility of conflict arising from interdependence and the actions taken by states to mitigate their perceived vulnerabilities.

Klare and Thomas' definition cannot meet the requirements of the second criterion, the freedom from ambiguity, because it is unclear who or what retains the responsibility for resolving global security threats. Does the obligation lie with individuals, the state, or collections of states in addressing these problems? Since the concept of state sovereignty obligates a commitment to non-intervention in domestic affairs, which role can international organizations play in resolving the security concerns identified by the authors? Klare and Thomas also fail to make clear what the relationship between national and world security is, and subsequently, whose security interests take precedence in the event of a contradiction between them.

Within the context of Klare and Thomas' definition, as Canada is a country of the developed "North," the possibility of Québec's separation constitutes a security concern to the federation. However, since the authors do not recognize political-institutional issues, separation only becomes a security problem if it is regarded as an economic threat.

In concluding this section, despite varying justifications, these first three non-traditionalists share an understanding of security that rests largely with the state. As well, the wideners agree that the traditional conception of security is too narrow, and as a result, they attempt to incorporate other sources of threat. However, only Ayoob is successful in maintaining a methodologically sound definition; both Wæver, and Klare and Thomas are unable to overcome categorization difficulties and ambiguities. Moreover, Ayoob's definition is superior to Walt's militaristic notion of security in terms of the third criterion, because it can encompass more than simply one case since 1990.

Section III: The Deepeners

Ken Booth

Ken Booth's advocacy for a deepened conception of security contrasts ironically with his earlier writings, when he was an influential realist. Now, as a self-proclaimed "fallen" Realist (1997), the contemporary Booth argues that war is actually a cultural phenomenon, which implies that it has a potential for change. As culture is dynamic, Booth advocates a theoretical approach which he refers to as "utopian realism" (1991). The essence of this conception is an amalgamation of normative and empirical theories:

The normative element is made up of universal appeal, based on reason, to various world order principles…The empirical element seeks to make the world of politics more intelligible by seeking to go beyond realism to a set of ideas which offer a fuller understanding of the forces shaping 'Who gets what, when and how'…(Booth 1991: 534).

While acknowledging that "utopianism" has been widely criticized in academia, Booth defends it through his expansion of a famous metaphor introduced by Michael Oakeshott in the 1950s (1991: 536). Oakeshott suggested that realism provided the survival tools necessary to keep (political) ships afloat in a "boundless and bottomless sea." In Booth's expansion, he agrees with Oakeshott, but contends that a captain requires more than seamanship; he also needs a utopian vision to guide the ship to its destination.

Applied to the current discussion, the destination Ken Booth has in mind is security, and the route to it is the emancipation of humanity:

Emancipation means freeing people from those constraints that stop them carrying out what freely they would choose to do, of which war, poverty, oppression, and poor education are a few. Security and emancipation are in fact two sides of the same coin. It is emancipation, not power and order, in both theory and practice, that leads to stable security (1991: 539).

By locating security and insecurity at the individual level, Booth cannot satisfy the first criterion of a methodologically sound definition. Such an inclusive conceptualization prevents it from being analytically meaningful, as the individual preferences of all people could not possibly be taken into account. Instead, it appears that Booth intends to promote normative (and arguably ethnocentric) values of the superiority of freedom and liberty over order.

Moreover, how can the inevitable situation of one person's freedom contradicting another's be resolved without further increasing the 'insecurity' of both? In Hobbesian terms, emancipation would precisely lead to the state of nature, where freedom was limited only by the power possessed by individual actors. It is only if we assume that individual rationality valued the same outcomes - the prevention of war, for example - that conflict between individuals would be reduced rather than exacerbated by emancipation. Hence, due to this ambiguity, Booth's definition also fails to meet the second criterion.

In addition, Mohammed Ayoob criticizes Booth on the grounds that his definition "obfuscates" the meaning of both 'security' and 'emancipation' (1997: 127). Ayoob discusses how the emancipation of Kurds from Northern Iraq illustrates that 'emancipation' in the sense of the right to self-determination may increase rather than decrease the insecurity of individuals and groups. Transcending Hobbes' state of nature, which is based on domestic disorder, to the international level, he argues that among the subalterns, where strong state apparatus' have yet to develop, the emancipation of ethnic groups would almost certainly lead to chaos and to the dissolution of state borders. Further, Ayoob reminds us of the fluidity of the concept of ethnicity, which is "subject to change depending on the context in which it operates at any given time" (1997: 127).

Within the context of Canada, Booth's definition would identify the federal government as the primary security threat for Québec sovereignists who would freely choose to separate from the federation. Since Booth (1991: 539) characterizes states as being "means and not ends," he would consider the Canadian state to be a threat to the distinct Québecois identity. This emancipation of the sovereignists would, however, be in fierce opposition to Québec nationalists and the ROC who would freely choose to have Québec remain within the federation. As both choices are equally legitimate among equal individuals, it is difficult to imagine how this contradiction could be resolved in such a manner as to not make all Canadians more insecure.

Thus, while Booth's equation of emancipation and security has intrinsic and emotional appeal, it seems to fall short as a definition of security in two respects. First, as an excessively broad conceptualization that encompasses the perceptions of all people, the definition becomes analytically meaningless as a tool for understanding the phenomena it is intended to capture. Second, in practice, the emancipation of humanity could be used to create a condition of anarchy, where violence is legitimized, and existing divisions between people are exacerbated.

Spike Peterson

The trademark of feminism in Security Studies is its objection to the traditionalist military-based and state-centric notion of security, arguing that this narrow conceptualization does more to decrease the security of women than to increase it. This is not only because the military establishment and state institutions are patriarchal, hierarchical and have traditionally excluded feminist contributions, but also because financial, social and political resources directed toward the military cannot be administered elsewhere in improving the health and social well-being of society.

Peterson contends that the preeminence of state sovereignty is a primary source of insecurity for women, because it limits the " construction of political community " (1992: 32) to masculine and patriarchal institutions and policies. Thus, she advocates a move toward perceiving security in global terms. While it is not explicitly defined by Peterson, her 'world security' concept would address global problems such as nuclear proliferation, ecological concerns, human rights abuses and widespread poverty on the one hand, and the systemic oppression of women, on the other (1992: 31). Furthermore, it is intended to transcend the ideologies and identities that produce systemic (global) insecurity and legitimate violence (both direct and indirect) against women.

As a result of its association with structural violence, this conceptualization of security employs the macro-level as its unit of analysis. Peterson asserts "…the structural violence of gender (and class) hierarchy - i.e. women's systemic insecurity - is revealed as an internal as well as external dimension of state systems" (1992: 32; italics in original). However, this definition is prevented from satisfying the first criterion for two reasons. First, this conception in itself, is a circular argument because the term insecurity is defined by another term 'structural violence' and the latter term is not defined. For the definition of structural violence, a second source needed to be consulted. J. Ann Tickner argues that it "extends the meaning of violence beyond its association with physical violence to the indirect violence done to individuals when unjust economic and political structures reduce their life expectancy through lack of access to basic material needs" (1995: 187; italics added). Because violence against women is seen as a systemic phenomenon, the insecurity of women is universal. Therefore, Peterson's conceptualization broadly generalizes that all women are insecure and all men are not. It implies that to understand security, we must understand the specific and potentially unique sources of insecurity of more than 50 percent of the world's population.

Second, the inclusion of class further complicates and obscures the definition for two reasons. On the one hand, as class is not based on gender but on socio-economic status, some men would also be considered insecure, which contradicts the universal distinction along gender lines. On the other hand, given the global distribution of wealth, even if we accept that both women and lower classes are insecure, a definition of security would then have to incorporate the individual sources of insecurity for most of the world's population simultaneously. With the number of cases and the subsequent variety of insecurities this implies, security becomes an unmanageable and incomprehensible concept.

Furthermore, as Peterson's 'world security' concept is not clearly defined, it is also ambiguous and, therefore, cannot fulfill the second definitional criterion. As was true for Klare and Thomas' 'world security' concept, this perspective raises numerous questions. What does global security look like? How are individual sources of insecurity accommodated with structural sources? How can we move from the idea of widespread insecurity to the practice of 'disrupting' the system?

In contrast to the other definitions that have been reviewed, feminism is not simply a theoretical discourse, but has activists who attempt to implement feminist beliefs in practice. Thus, this case stands alone in allowing us to compare theory and reality within the context of the threat of Québec's separation from the ROC.

According to Peterson's 'world security' concept, this threat would constitute a source of insecurity because of its basis in nationalism. As "perhaps the most powerful ideology of collective authority and political identity in the modern era" nationalism has tremendous repercussions for women because it "obscures social hierarchies" (Peterson 1992: 47). In other words, 'effective' nationalism regards everyone in the respective national community as sharing the same fundamental characteristics. Thus, Québecois identity excludes other possible sources of identity, such as gender, class, and ethnicity, while reinforcing language and culture as the primary criteria for differentiating people.

There appears to be, however, an inconsistency between what some feminists, such as Peterson, claim theoretically, and how other feminists act in reality. According to Micheline Dumont, many Québec feminists have pragmatically chosen to support the sovereignist movement, as the famous slogan "No women's liberation without Québec liberation. No Québec liberation without women's liberation" indicates (1992: 76). Therefore, instead of opposing nationalism, many feminists in Québec have been very supportive of it. Dumont contends that because the Parti Québecois was the first political party to adopt a specific policy on women (1992: 88), separation is perceived as being a positive step for all women in that province.

To complete this section, in contrast to the wideners examined in Section II, the deepeners focus on different referent objects in their definitions of security. The methodological implication is that security becomes all things to all people; in other words, security as a concept loses its analytical value because it cannot exclude any cases. As we have seen, the criticism of analytical incoherence seems to be true for both Booth's emancipation definition, and for Peterson's (re)visions of security.

Conclusion

The purpose of this paper was to conduct a survey of six definitions of security in post-Cold War Security Studies literature. Many scholars, realists and liberals alike, have capitalized on this apparently changed political climate to re-visualize security in broader terms. With a variety of justifications, the five non-traditionalist perspectives presented above challenge the traditional militaristic notion of security that had previously prevailed. As this paper illustrates, because many of these emerging definitions are fundamentally normative, they are so general as to be analytically meaningless.

The conceptual difficulty in constructing an analytically meaningful definition with practical utility is not to be underestimated. As Klare and Thomas (1994: 2) assert, a concept is not a theory and so cannot be disproved; rather, it can only be accepted or rejected based on its analytical usefulness and "its consistency with one's own normative or theoretical inclinations." Because a battle of values can never be decisively won, this paper has emphasized that analytical utility arises from a methodologically sound definition, instead of by judging the validity of the values held by the authors. The first criterion of a 'good' definition is that it is not all-inclusive, or in other words, that it excludes cases. The second criterion is that the categorization be unambiguous so that it is always clear which cases are included and which are not. The final criterion is an ability to transcend a purely analytical focus and to apply to situations in reality.

The six security definitions were organized and examined along a traditional versus non-traditional cleavage. The fundamental difference between these two perspectives is the limited association with military security on the one hand, and with a much broader range of threats on the other. The non-traditionalist position was sub-divided into the 'wideners' and the 'deepeners,' where the essential differences were between the security issues they highlighted and the referent objects they chose.

In attempting to fulfill the first criterion, only two of the six definitions could clearly exclude cases. Walt's traditionalist perspective is the most easily applicable analytically, since it excludes all cases except those in which military force is employed in response to perceived national security threats. Ayoob also meets this criterion because he identifies a political threshold for when any issue can be considered a security threat as well as excluding potential non-state security issues.

The remaining four definitions, however, are unable to satisfy the first criterion. In Wæver's two-dimensional notion of state and societal security, the latter concept does not appear to be operationalizeable, preventing analysts from constructing definitive categorizations. Klare and Thomas had the immediate methodological problem of encompassing two levels of analysis simultaneously as well as permanently affixing specific threats to their definition and thereby ignoring the varying importance of those issues over time and space. Booth's emancipation definition, with the individual as the referent object, cannot exclude cases since any threat to the freedom of an individual must be taken into account. Peterson's macro-level definition had a similar difficulty because of its incorporation of structural violence and the universality of particular threats.

In terms of meeting the second criterion, the absence of ambiguity, there was a similar result. Walt's definition is unambiguous about the use or threat of military means as constituting a threat. The inclusion of the undefined term 'regime' in Ayoob's definition causes the categorization of normal political change in democratic 'regimes' to be unclear. However, the author could easily resolve this problem simply by explicitly defining the concepts he employs.

The transition from analytical meaningfulness to practical applicability is the purpose of the final criterion. It is this third methodological test that isolates Ayoob's definition as the sole analytically coherent definition that can also be applied to several cases in the post-Cold War era. By excluding intrastate conflict as a source of national insecurity, Walt's definition is relevant only in the recently rare circumstance of interstate war.

Thus, it appears that for Klare and Thomas, Booth and Peterson, normative inclinations assumed priority over the importance of a useful concept of security. While their value preferences captured in these security definitions are not problematic per se, it seems that they must still fulfill the three methodological criteria in order to avoid 'intellectual incoherence.'

By transcending the normative commitments of several authors within the contemporary debate, the paper highlights the importance of evaluating definitions for their analytical applicability rather than for their intrinsic appeal to human emotions. Perhaps, as value preferences continue to divide academics in the subdiscipline of Security Studies, it is only on methodological grounds that this apparent gap between perspectives can be bridged and the widespread disagreements resolved in respect to this "essentially contested concept."

Ayoob, Mohammed. 1997. Defining Security: A Subaltern Realist Perspective. In Critical Security Studies, edited by Keith Krause and Michael C. Williams, 121-146. Minneapolis, MN: University of Minnesota Press.

Babbie, Earl. 1995. The Practice of Social Research. 7th ed. Belmont CF: Wadsworth Publishing.

Booth, Ken. 1991. Security in Anarchy: Utopian Realism in Theory and Practice. International Affairs 67(3): 527-45.

__________. 1997. Security and Self: Reflections of a Fallen Realist. In Critical Security Studies, edited by Keith Krause and Michael C. Williams, 83-120. Minneapolis, MN: University of Minnesota Press.

Buzan, Barry. 1991. People, States and Fear: An Agenda for International Security Studies in the Post-Cold War Era. 2nd ed. Boulder, CO: Lynne Rienner.

Buzan, Barry, Ole Wæver, and Jaap de Wilde. 1998. Security: A New Framework for Analysis. Boulder, CO: Lynne Rienner.

Dalby, Simon. 1997. Contesting an Essential Concept: Reading the Dilemmas in Contemporary Security Discourse. In Critical Security Studies, edited by Keith Krause and Michael C. Williams, 3-32. Minneapolis, MN: University of Minnesota Press.

Dumont, Micheline. 1992. The Origins of the Women's Movement in Quebec. In Challenging Times: The Women's Movement in Canada and the United States, edited by Constance Backhouse and David H. Flaherty, 72-89. Montreal & Kingston: McGill-Queen's University Press.

Grieco, Joseph M. 1993. Understanding the Problem of International Cooperation: The Limits of Neoliberal Institutionalism and the Future of Realist Theory. In Neorealism and Neoliberalism: The Contemporary Debate, edited by David A. Baldwin, 301-38. New York: Columbia University Press.

Klare, Michael T., and Daniel C. Thomas, eds. 1994. World Security: Challenges for a New Century. New York: St. Martin's Press.

Krasner, Stephen D. 1983. International Regimes. Ithaca, NY: Cornell University Press.

Krause, Keith, and Michael C. Williams. 1996. Broadening the Agenda of Security Studies: Politics and Methods. Mershon International Studies Review 40: 229-54.

__________. 1997. From Strategy to Security: Foundations of Critical Security Studies. In Critical Security Studies, edited by Keith Krause and Michael C. Williams, 33-60. Minneapolis, MN: University of Minnesota Press.

Lipschutz, Ronnie D., ed. 1995. On Security. New York: Columbia University Press.

Mearsheimer, John. 1990. Back to the Future: Instability in Europe after the Cold War. International Security 15(1): 5-57.

Peterson, V. Spike, ed. 1992. Feminist (Re)Visions of International Relations Theory. Boulder, CO, Lynne Rienner.

Salmon, Merrillee H. 1995. Introduction to Logic and Critical Thinking. 2nd ed. Fort Worth, TX: Harcourt Brace & Company.

Sylvester, Christine. 1994. Feminist Homesteadings of Security and Cooperation. In Feminist Theory and International Relations in a Postmodern Era. Cambridge: UK: Cambridge University Press.

Tickner, J. Ann. 1995. Re-visioning Security. In International Relations Theory Today, edited by Ken Booth and Steve Smith, 175-97. University Park, PA: Pennsylvania State University Press.

Ullman, Richard. 1991. Securing Europe. Princeton, NJ: Princeton University Press.

__________. 1996. The World and Yugoslavia's Wars. New York: Council on Foreign Relations.

Wæver, Ole. 1995. Securitization and Desecuritization. In On Security, edited by Ronnie D. Lipschutz. New York: Columbia University Press.

Wallerstein, Immanuel. 1984. The Politics of the World-Economy: The States, the Movements, and the Civilizations. Cambridge, UK: Cambridge University Press.

Walt, Stephen M. 1991. The Renaissance of Security Studies. Mershon International Studies Review 41: 211-39.