Eli Stamnes and Richard Wyn Jones



In the last few years Critical Security Studies (CSS) has emerged as a new approach to the academic study of security. This article argues that its genesis is best understood as a reaction to two developments, namely ‘real world’ changes after the end of the Cold War and the far-reaching philosophical debates that have recently been taking place within the social sciences. The authors argue for a conceptualisation of CSS based on an explicit commitment to human emancipation. They then illustrate their preferred understanding of security through a discussion of Burundi. This case study not only illustrates the theoretical claims of CSS but also serves as a contribution to a more comprehensive understanding of the security issues with which this country and its inhabitants are faced. 


            This article has two inter-related aims. It seeks first to outline and explain the approach to the study of security which has developed under the rubric ‘Critical Security Studies’ (CSS). This is done initially by comparing the principles and precepts of CSS to those of the more orthodox approach to the study of security – described here as Traditional Security Studies.[1] Subsequently, the particular case of Burundi is explored from a Critical Security perspective. This serves not only to further illustrate the nature of the Critical Security approach, but also – the second aim of the chapter – to provide a conceptual framework for understanding the theory and practice of security in Burundi. 

Critical Security Studies

            CSS represents an important departure in the academic study of security, strategy and peace. The approach owes its genesis to two coincidental developments: the end of the Cold War, and major debates within the social sciences as to their nature, purpose and method. Both developments ask fundamental questions about some of the central tenets that underpinned traditional Security Studies and Peace Studies alike. The vibrant and often contentious exchanges that have resulted from this period of geopolitical transformation and intellectual uncertainty look set to continue well into the 21st century. Nevertheless, the contours of a new approach are now clearly emerging. CSS, as it has come to be known, is nothing if not ambitious. Its proponents not only offer a thoroughgoing critique of both the theory and practice of the hitherto dominant, traditional approach to security (usually called Strategic Studies or National Security Studies). They also aim to provide intellectual support and justification for alternative security practices that can emancipate humanity from the fear and insecurity which dominate the lives of people across the world (Bilgin, Booth and Wyn Jones, 1998; Wyn Jones, 1999; Booth, forthcoming). 

The Critique of Traditional Security Studies

            Traditional Security Studies was in many ways a direct product of the Cold War. Despite ritualistic references to such great military thinkers of the past such as Sun Tzu and Clausewitz, the subject was almost exclusively concerned with superpower rivalry, and in particular its nuclear manifestations. It was the perceived exigencies of Cold War competition that encouraged Security Studies to flourish in Western academia and research institutes. Furthermore, from the late 1940s onwards, such was the symbiotic nature of the relationship between security specialists in academia and defense establishments in government and industry that some posited the existence of a military­industrial-academic complex; each element had vested interest in promoting a militarised conception of relations between states.

            The ethical issues raised by this entanglement were trenchantly highlighted in the 1960s by such writers as Anatol Rapoport (1960; 1964; 1970) and Philip Green (1966; 1973), but equally as problematic for the future of human security was the state-centric conceptualisation of the subject. Making sovereign states conceived in rather simplistic unitary terms, the exclusive ‘referent object’ for the understanding of security issues had far-reaching implications. Because Cold War Security Studies was about states, the security of individuals was subsumed under the ambit of the state, and often sacrificed to the demands of realpolitik. The more secure the apartheid state of South Africa became, for example (in terms of strong borders and increased legitimacy for the regime) the less secure became the lives of the majority of the population. Security Studies, implicitly or explicitly, generated information and analysis for states, and specifically for the ruling elites within them (Wyn Jones, 1999). Security Studies within universities became an arm of statism.

            Moreover, Cold War Security Studies conceived state security in almost exclusively militarised terms. Security was equated with military security. True, during the era of détente and oil shocks from the late 1960s to the mid-1970s, a broader range of issues was allowed on to the agenda, in recognition of the growing sensitivity of economic and environmental concerns. However, this constituted more of a parenthesis rather than a break; things returned quickly to security-business-as-usual with the re-intensification of the Cold War in the late 1970s.

            Traditional Security Studies was never without its critics. As already noted, some questioned whether close involvement with government was compatible with scholarly independence. But criticism was not confined to academia. Events such as the ‘Ban the Bomb’ marches in Britain in the late 1950s, support for the ‘nuclear freeze’ in the United States in the 1980s, and growing demands in Western Europe for ‘pan-European’ perspectives, illustrated that crosscutting differences over the theory and practice of security existed even within the West (see, for example, Kaldor, 1991).

            In the early 1980s, the widespread fear and disaffection created by the Second Cold War led to two inter-related and hugely significant developments. First, the introduction of Cruise and Pershing II missiles into western Europe led a massive resurgence in peace activism that challenged some of the central nostrums of security policy. The newly invigorated peace movement also inspired and was inspired by a second development, namely the growth of ‘alternative defense’ thinking, which challenged the establishment and its experts on their own grounds.

            The crucial insight upon which ‘alternative defense’ was based was that the zero-sum notions of security prevalent during the Cold War were actually destabilising, and a major source of insecurity in international relations. Thus for alternative defense thinkers, real peace and security was not just the absence of war, but rather depended on the establishment of the conditions for social justice and mutual understanding – a view also shared by Peace Research. Alternative defense thinking concentrated on seeking means whereby the so-called ‘security dilemma’ could at least be mitigated. Consequently they developed such concepts as common security, non-offensive defense, a nuclear freeze, military confidence building, democracy and disarmament, ‘détente from below,’ and alternative security orders. Surprisingly perhaps, it was ultimately in the Eastern bloc that these ideas were to have most impact. The principles and precepts of ‘alternative defense’ directly influenced the ‘new thinking’ that emerged in the Soviet Union after 1986 under its new leader Mikhail Gorbachev (Risse-Kappen, 1994, Wyn Jones, 1999). Despite being condemned in many Western capitals for being hopelessly deluded and idealistic, alternative defense thinkers played a crucial role in the ending of the Cold War. 

Towards a Critical Security Studies

            Given the symbiotic relationship between the Cold War and Security Studies, it is not surprising that the end of the former led to a crisis in the latter. Not least among the factors that have led some to call for a fundamental rethink of the way security is conceptualised has been the inability of ‘security specialists’ to respond adequately in the face of the new – or old, but hitherto neglected – issues which have emerged since the late 1980s. These issues have forced all but the most hidebound to reconsider what we actually mean when we use the term ‘security’, and what is the appropriate ‘referent object’ for security discourse (the contours of the ensuing debate can be traced through the following collections: Lipschutz, 1995; Brown, Lynn-Jones and Miller, 1995; Brown, Coté Jr., Lynn-Jones and Miller, 1997; Lynn-Jones and Miller, 1995; Krause and Williams, 1997; Buzan, Wæver and de Wilde, 1998 [2]; Booth, forthcoming). Proponents of CSS have combined an eagerness to explore these crucial issues with a commitment also to reconsider the methodology by which security should be studied (Krause and Williams, 1997; Wyn Jones, 1999). Again, the influence of the alternative defense thinkers of the early 1980s is important here. They had been concerned to support and influence those social movements that had emerged to protest the continuing superpower arms spiral. In doing this they had broken fundamentally from traditional security specialists who had often claimed a spurious objectivity and detachment for their work while in reality their teaching and research legitimised the prevailing status quo by making it into a ‘false necessity’.

            Through their rejection of simplistic notions of ‘detachment’ and ‘objectivity’, alternative defense thinkers were precursors of a much more widespread move in the social sciences to question the very nature and purpose of social research (Bernstein, 1988). While such questioning has been vigorous, there has been little consensus as to what might constitute acceptable answers. This is reflected in CSS where there is disagreement, for example, between those who follow such post­structuralist thinkers as Foucault and Derrida and those who are more persuaded by the arguments of Critical Theorists such as Horkheimer and Habermas. Whereas the former group of scholars focus largely on deconstructioning/de-essentialising claims about security, the latter group’s theorising also involves a reconstructive element. Nevertheless, whatever their differences in focus and methodology, all proponents of CSS seem to regard the ultimate aim of their work as aiding in the transformation of prevailing patterns of power and domination in favour of those who are currently disempowered and disenfranchised. In the case of the post-structuralists, this purpose is usually implicit in their work – to be read ‘between the lines’ as it were (but see the discussion in chapter 4 of Wyn Jones, 1999 for a more detailed survey). However, in the case of thinkers influenced by Critical Theory, this intent is recognised in terms of an explicit commitment to human emancipation. It is within this latter strand that this article has its basis.

            As well as introducing a new level of methodological awareness and sophistication into Security Studies, proponents of CSS have also re-worked the key concept of ‘security’ in important ways. In doing so they have provided new answers to such central questions as: What is security? Whose security should we be concerned about? And, what is the purpose of studying security? The answers to these questions provide the basis for the claim of supporters of CSS that it represents ‘the next stage’ in this new era of world politics.   

What is security?

            As we have already seen, Traditional Security Studies understood security in almost entirely military terms. This may have had a certain logic during the period of nuclear, mutual assured destruction (MAD). However, even then, this understanding did not produce a conceptualisation of security that spoke the concerns of the vast majority of the world’s population living in the Third World: for them a broader conception of security always made more sense. In the wake of the end of the Cold War, the narrowly militarised understanding of security simply became anachronistic. The vocabulary of nuclear deterrence, arms control, limited war and superpower crisis management had minimal relevance in a world in which the salient security concerns increasingly related to ethnic conflict, refugee flows, famine, peaceful settlement of disputes, the politics of identity, confidence­building, humanitarian intervention, conventional war, and so on. A narrowly militarised understanding fails to grasp the complexity of the ‘new’ issues on the security agenda, not to mention their life-and-death relevance to many people on the planet.

            Published in 1983, Barry Buzan’s seminal study People States and Fear was the first sustained attempt from within Security Studies to re-draw a broader notion of security that moved beyond a purely military focus. Although still arguing within a state-centric framework, Buzan argued that by paying disproportionate attention to the military dimension, analysts ignored other issues which were of equal importance when it came to understanding how policy-makers and populations alike perceive their security interests. In response, he proposed to broaden the conceptualisation of security to include four other ‘sectors’ in addition to that of the military: these were political, economic, societal and environmental sectors. This influential formulation has been criticised. Some traditionalists argue that adopting a broader conception of security will dilute the coherence of Security Studies (Walt, 1991). Some radical critics are wary of the danger that labelling issues such as environmental degradation as ‘security concerns’ will lead to their militarisation (Deudney, 1990). However, these arguments have been forcefully countered by those who favour the ‘broadening of security’. On the one hand they have pointed out that the parsimoniousness of the traditional approach has blinded analysts to the inter-relationship between different threats, or in Buzan’s terms, between the different sectors. Thus one of the consequences of rejecting any attempt to narrow the analysis through an arbitrary definition is to allow for a more variegated and realistic assessment of the dynamics of security. Interestingly, a broader notion of security also echoes what is increasingly becoming the explicit practice of states. Even Britain, for example, now distinguishes between ‘defence policy’, which focuses exclusively on the military dimension, and ‘security policy’, which takes a broader view incorporating economic and trade issues.

            In response to worries that ‘broadening security’ runs the risk of militarising the new issues incorporated onto the security agenda, supporters argue that broadening is itself part of a move to promulgate a different understanding of security: this understanding rejects the zero-sum notions prevalent during the Cold War (Wyn Jones, 1999).

            In addition to supporting moves to broaden notions of security, proponents of CSS have also called for its ‘deepening’. Deepening involves the recognition that security is a derivative concept (Walker, 1997; Booth, 1991, 1997). That is, our conceptions of security depend on the particular philosophical world-view we have. Traditional Security Studies was based on an often-implicit understanding of a world characterised by a continual struggle for power among sovereign states competing in an anarchical international system. According to this understanding, inter-state conflict is endemic and is destined to remain so, and this should establish the character of Security Studies. However, while this view may have achieved the status of ‘common-sense’ among Cold War security specialists, this is far from being the only possible world-view. When conceptions of security are derived from alternative perspectives – those of Feminists, Marxists, World Order thinkers and so on – very different understandings of ‘security’ emerge. Recognising the derivative nature of security, CSS has sought to expose and criticise the world-view that has underpinned traditional Security Studies and explore the understandings of security that can be built on the basis of possible alternatives. To this end CSS is concerned to explore regional security questions, including those of a military character, and promote thinking about practical policies to advance the security of people as individuals and groups (Booth and Vale, 1997).   

Whose security?

            One of the most pervasive assumptions underlying traditional Security Studies is state­centrism. State-centrism privileges the role of the state in world politics, regarding it as the sole legitimate focus for decision making and loyalty. The state became the only referent object of Cold War Security Studies. Strategic problems were analysed in terms of their impact on states. ‘National security’ was stressed as opposed to the security of individuals, groups of one sort or another, civil society, world society or common humanity. The justification for this was the doctrine that the state provides security for its own citizens. However, even a cursory examination of the empirical reality of world politics exposes the fact that this is a hopelessly idealised view (Booth, 1991; Reus-Smit, 1992; Wyn Jones, 1999). States tend to prioritise threats to their own security and often end up neglecting those posed to individuals, groups within the state, and certainly humanity as a whole. Indeed, in many cases, the state itself is one of the main threats to the security of its own population. This fact is all too well confirmed by the widespread pattern of state-sanctioned political oppression, human rights violations and torture.

            Rather than continue to privilege states as the referent object for Security Studies, proponents of the Critical approach have argued that the concept of security should be extended to referents other than the state. In his influential essay on ‘Security and Emancipation’ (1991), Ken Booth insisted on the need for analysts to view states as means rather than ends in themselves. That is, while states may in some cases be providers of security, and are certainly a crucially important element in the politics of security, this does not mean that they should be privileged as referents for the theory and practice of security. Booth argues that the ultimate referent for security should be individual human beings. This formulation has not won universal acceptance even in the Critical camp. Some, for example, have suggested that ‘civil society’ is a more appropriate referent. However, given that making individuals the ‘ultimate’ referent for discussion of security concerns does not preclude – in fact, encourages – consideration of the boarder contexts in which human beings live their lives, the logic of favouring the individual referent is particularly strong (for an extended discussion of this issue see Wyn Jones, 1999). 

What sort of Security Studies?

            In addition to challenging the way in which security has been conceptualised, CSS also offers an alternative understanding of the meaning and purpose of studying security in the context of universities. As we have noted, behind a veil of spurious detachment and objectivity, traditional Security Studies has served the interests of statism and official thinking. Ultimately the subject has legitimised and supported the interests of those who benefit from the status quo, on both a local and global scale. CSS follows in the tradition of alternative defence thinking by regarding itself as thought in the service of ‘bottom-up’ progressive social change. Robert Cox (1981: 128) has famously argued that “all theory is for someone and for some purpose”. If this is the case then CSS is primarily for those made insecure by the prevailing order. In Critical Theory terms, its purpose is to aid in the process of their emancipation. 

Critical Security Studies and the future

            The traditional, Cold War approach to security has attracted many critics, though it is still resilient in many institutions. CSS is perhaps best understood as an attempt to bring together these various strands of criticism and to weld them into a coherent whole. A variety of different approaches have important contributions to make to the CSS project: the work of the Peace Researchers in widening the concept of violence and emphasising the positive conception of peace, the work of alternative defence thinkers in focusing on ‘common security’ and ways of mitigating the security dilemma, the work of critics within Security Studies in broadening the security agenda, the role of Third World specialists in emphasising the role of the structure of the international economic system in engendering insecurity in the South, the insights of feminists in underlining the relationship between the personal, the political and the international, and demonstrating the centrality of identity politics to the understanding international phenomena, the analyses of Critical Theorists in exposing traditional approaches to theory and in outlining a theoretical approach explicitly oriented towards human emancipation, and the work of World Order theorists in outlining alternative visions of attainable and sustainable world orders (Booth, 1997).

            By bringing these strands together, underpinning them with greater methodological sophistication, and providing new direction and greater coherence, CSS hopes to become an important voice informing and legitimising those political practices which promote security, community and emancipation for the whole of humanity (Booth, forthcoming). However if this extremely ambitious aim is to be achieved, the Critical Security perspective must do more than assist analysts in clarifying conceptual and meta-theoretical issues – although this is undoubtedly important – it must also prove its ability to throw new light on empirical studies. While the task of applying the conceptual insights of CSS to ‘real people in real places’ is still in its relatively early stages, prominent proponents of the approach such as Ken Booth and Peter Vale (1997: 329) are explicitly aware that it is the success of this venture that will determine “if critical security studies are to flourish and lead to a revisioning of security in world politics”.[3] The following discussion of Burundi’s security situation represents our own modest contribution to this effort.[4] It seeks to demonstrate how both ‘broadening’ and ‘extending’ the concept of security is necessary in order to understand the security issues that concern the people of this benighted central-African state. Furthermore, it will seek to demonstrate the (potential) utility of fastening upon some of the areas of unfulfilled potential that exist within the specific Burundian situation in order to edge developments in Burundi in a more emancipatory direction. 

Burundi: whose security?

            “Disputes – international : none” – so reads the section dedicated to Burundi in the 1997 CIA World Fact Book. This suggests that when viewed from the perspective of Traditional Security Studies with its stress inter-state military conflict (actual or potential), this was a country whose affairs were of little relevance to the concerns of security analysts. In the intervening period, Burundi seems to have become embroiled in the chaos in its neighbour to the west, the Democratic Republic of Congo (formerly Zaire), and may thus now fall under the remit of Traditional Security Studies. However even in 1997, it would seem to be a particularly perverse form of Security Studies that did not recognise that Burundi was beset by the most serious of security problems. As will be seen below, sudden violent death had been an ever-present threat for the Burundian population during the previous five years. Moreover, the chronic insecurity of the 5.5 million inhabitants was further compounded by their country’s extreme poverty (see for example Amnesty International, 1997; 1998; Evans, 1997).

            Burundi provides a striking empirical example of the inadequacy of the traditional security paradigm. Let us first consider the issue of the appropriate referent object for Security Studies. The traditional focus on the state was justified, in familiar contractarian terms, on the grounds that the state provides for the security of its own population (Krause and Williams, 1997). In the words of Osgood and Tucker (1967: 284), the state was regarded as the “indispensable condition of value”. The example of Burundi not only serves to refute this assumption, but totally reverse the terms of debate. For ever since Burundi gained its independence in 1962, the state has been the major source of insecurity for the greater proportion of its population. Most horrifically, on several occasions the state’s armed forces have carried out massacres during which tens of thousands have been slaughtered.

            This endemic insecurity is a result of the deep ethnic and socio-economic fissures in Burundian society. Since independence, with the exception of one short period in the 1990s, Burundi has been ruled by Tutsi dominated governments. The Tutsi form a minority (approximately 14%) of the population, and successive governments have maintained their domination over the Hutu majority (approximately 85% of the population) by ensuring their “near exclusion […] from public life, knowledge and wealth” (Reyntjens, 1993: 563). In this sense, Burundi has provided a mirror-image to neighbouring Rwanda. There the Hutu majority (the proportions of both groups are broadly similar in both countries) has dominated the Tutsi for most of the period since the revolution of 1961 and the break-up of Ruanda-Urundi (Lemarchand, 1970). But despite this difference in the balance of power between both groups, the post-independence history of both states has been equally grisly.

            Since Burundi gained independence it has been convulsed by regular bouts of violence: coups have been the ‘normal’ means of securing (intra-Tutsi) regime change, while several Hutu uprisings have attempted to challenge discrimination and oppression, in turn sparking extremely violent reactions by the armed forces. In the ensuing spirals of killings and counter-killings hundreds of thousands have been forced to flee, with each violent episode serving to further solidify and harden the intercommunal boundaries. The result is that Hutu-Tutsi divisions, while historically caste-based, have taken on the appearance of being timeless, primordial and intractable ‘ethnic’ divisions, and Hutu-Tutsi relations have come to be perceived in purely zero-sum terms (Lemarchand, 1970; Reyntjens, 1993; Evans, 1997).

            In 1972 the state’s reaction to a Hutu uprising may be properly characterised as genocidal. The uprising was a reaction to the arrest and killing of the former King of Burundi, Ntare V, by the regime that had deposed him in 1966 and established the first republic under Michel Micombero. In reaction to his killing, groups of Hutu who felt themselves to be increasingly discriminated against by the regime attacked military centres around Burundi. This triggered brutal reprisals. In the following weeks Hutu were systematically slaughtered, with members of the Hutu ‘elite’ – essentially anyone with education or in a position of leadership and responsibility – being particularly targeted. This latter group included all Hutu officers and most of the Hutu soldiers in the armed forces; most Hutu teachers, nurses, students, secondary school pupils, business men and civil servants were also killed. Priests were another target due to the Church’s role in education and political life. Thus the Hutu were effectively decapitated; their organic leadership hunted down and slaughtered as part of a sustained campaign of murder which left a death toll of between 150,000 - 250,000 and forced more than 100,000 to flee the country (Malkki, 1995; Lemarchand, 1989).

            In comparison to the horrors of 1972, events in 1988 appeared to be something of a watershed in inter-communal relations in Burundi. Then international reaction to yet another massacre in August of that year, forced the President, Major Pierre Buyoya, to change course.[5] Ethnic divisions were recognised and a reconciliation program was established. This included the appointment of a Hutu Prime Minister and the creation of a Cabinet with equal numbers of Hutu and Tutsi portfolios. A ‘Commission to Study the Question of National Unity’ was established and resulted in a charter approved by referendum in February 1991. This was followed by work on a new constitution, which also was approved by referendum eleven months later. New political parties as well as an independent press emerged, and the recruitment practices for civil servants and the admission examinations for secondary schools were changed in order to encourage Hutu participation (Lemarchand, 1989; Reyntjens, 1993; Chretien, 1996; Des Forges, 1995). This process continued despite efforts to obstruct it from both the armed forces and extremist Hutu (from the PALIPEHUTU movement). Many Hutu gained employment in the state apparatus, and the proportion of Hutu within the higher education system increased. Indeed refugees who had been living in Tanzania since the 1972 genocide began to return, such that in 1993, an arbitration programme was established to deal with the disputes over property that might arise from this process (Mallki, 1995: 284; Des Forges, 1994: 205). Undoubtedly the highpoint of this period of reconciliation were the first democratic presidential and legislative elections held in June 1993. These were won by Melchior Ndadaye and the FRODEBU party, winning 65% and 71% of the vote respectively (Reyntjens, 1993: 567-572). Further hopes were raised by the new president’s conciliatory attitude towards the former president Buyoya, the UPRONA party and the Tutsi in general. However, these hopes were violently dashed when President Ndadaye, together with his three possible successors, were all killed in a coup d’etat, after merely 100 days in office. (Des Forges, 1995: 207; Evans, 1997:27) Yet again, the perpetrators came from within the armed forces. In the following days these forces went on to kill thousands more. Yet again this instigated a wave of killing and counter-killing which left around 50,000 dead and forced almost a million to flee (Des Forges, 1994: 203; Malkki, 1995: 285-6).[6]

            Quite apart from the general challenge to people of conscience everywhere that the horrors of Burundi undoubtedly should provide, it is a history which confronts Traditional Security Studies with fundamental and thorough-going questions about its conceptual framework. For this is an example, in extremis, of that framework’s normative and empirical failings. In the Burundian context, it should be apparent that any attempt to privilege the state as the provider of security is simply grotesque. Moreover, Burundi also forces proponents of the traditional approach to confront difficult – even existential – questions over who or what constitute a state? Who does a state consist of when, as in the case of Burundi, ‘its’ inhabitants are regularly forced to flee the country in their tens and hundreds of thousands? These Barundi live abroad for years, while simultaneously refugees from the neighbouring countries settle in Burundi.[7] Are these refugees a part of the state? And if so, of which state? Given the statist assumptions of Traditional Security Studies, these questions would seem to be of crucial of importance. But whatever the theological niceties of the discussion, it is clear that any serious discussion of security in this area must give central consideration to the fate of these refugees. Refugees loom large in the Tutsi’s fear of Hutu revenge, and these ‘outside elements’ have often been blamed by the state or it’s supporters for fomenting rebellion (Malkki, 1995; Chretien, 1996). These accusations contain within them a kernel of truth in as much as it is clear that refugees from the different countries in the Great Lakes Region have been instrumental in fomenting violence in other countries (see Evans, 1997 for more details on this). All of which underlines that the traditional simplifying assumption of a basic congruence between state borders and the boundaries of various societies is not only untenable but also empirically unhelpful (see Wæver et al., 1993 who distinguish between state and society in the European context). To treat the different states as distinct, discrete units in any simple neorealist sense is in fact highly misleading.

            Given the problems with privileging the state as the referent object for any discourse on security in Burundi, what then of other candidates for the role of referent object? One obvious candidate is ethnic identity given that the situation in Burundi is often understood in terms of an ‘ethnic conflict’.

            It can hardly be denied that understanding Burundi’s catastrophic post-independence history in ethnic terms throws a certain light upon event there. As has already mentioned, the character and implications of the country’s ethnic divisions became rapidly more divisive and murderous in the immediate aftermath of independence. Three linked developments were of particular importance: the assassination of a Hutu Prime Minister in 1965, the success of Hutu candidates in subsequent legislative elections, and the coup in 1966 which replaced Burundi’s constitutional monarchy with a Tutsi-dominated, one-party regime (Lemarchand, 1970; Chretien, 1996). Historically there had been long-standing social divisions between the Hutu, Tutsi and a third category, the Twa, which represent 1% of the population. These divisions were largely caste-based – the Tutsi were predominantly pastorialists, the Hutu predominantly cultivators and the Twa predominantly hunters, potters and ironworkers – with the Ganwa (‘the princes of the blood’), a noble caste of Tutsi, standing at the top of the socio-economic pyramid. A feudal system revolved around the kettle clientship, with its intricate system of mutual rights and duties between patron and client. This was separate from and subsidiary to a decentralised political structure, in which local Ganwa chiefs shared power with the king. But it is nonetheless important to note that this intricate social pattern did allow movement from one ‘ethnic’ category to another, and that intermarriage was not unusual (Lemarchand, 1970; Des Forges, 1994; Malkki, 1995; Evans, 1997). It was in many ways the colonial powers, Germany and Belgium, who solidified and exacerbated the divisions between Hutu and Tutsi. They magnified the political importance of ‘ethnic’ distinctions during their drive to ‘tidy up’ the existing structures of governance, whose end-result was to privilege the Tutsi.

            Since the most recent attempt to transcend ethnic divisions by fostering a sense of ‘national unity’ was brutally crushed in October 1993, ethnic divisions have taken on an even greater importance.[8] At least 150,000 have been killed in the violent conflict motivated overwhelmingly by ethnic differences (Evans, 1997: 9). Reyntjens (1993: 582) claims that “ethnicity has re-emerged as the single most important factor of political life” and the United Nations commission who investigated the assassination of President Ndadaye concluded that an ‘overwhelming majority’ of the Barundi identified themselves with either the Hutu or the Tutsi category (Evans, 1997: 22). In this context arguments that the Hutu-Tutsi distinctions are not ‘really’ ethnic, and that the ethnic ideology is based on “a confusion between concepts of race and class” seem to miss the point (Chretien, 1996: 206; Makinda, 1997; Des Forges, 1994). For even if these distinctions have little ‘objective’ basis, conflict based on the perceived differences between Hutu and Tutsi continue to have deadly consequences (Lemarchand, 1994).[9]

            Does the undoubted importance of ethnic identities in the ongoing conflict in Burundi justify focusing on these identities as the referent object in any security analysis of the country? In our view the answer is no, and this is for at least two reasons. First of all to do so, as  Bilgin et al. (1998) has pointed out, would be to logically confuse agents and referents. While ethnic identity is clearly a very important element of the security problematique in Burundi, to privilege it as the referent object for security is merely to reintroduce all the problems identified with the privileging of the state at another level (Krause and Williams, 1997; Shaw, 1994). Moreover, and relatedly, to privilege ethnic identity, underpinned as it is in this case by a particularly pathological ‘self-other’ distinction, would be to privilege exactly that which needs to be removed or at least ameliorated for progress towards a more sustainable and secure environment to be made. Secondly, to understand Burundi purely in terms of Hutu-Tutsi conflict is effectively to silence other aspects to the conflict: these include the socio­economic aspects already alluded too, intra-group conflict, and the situation of the Twa. The latter group is almost always ignored by analyses of the conflict with much of the literature suggesting that they are not interested or active in politics and that they can thus be safely ignored. It is however far from certain that the Twa’s security interests would be served by ‘adding’ them to a discussion which focuses on ethnic identities. This is because the ethnic characterisations (stereotyping) of the Twa tend to be extremely negative, focusing as they do on the Twa’s ‘pygmoid features’ and alleged intellectual deficiencies  – the Twa are regularly seen as “half-monkeys” (Malkki, 1995: 21, 29-30). In this context, fastening upon and further privileging ethnic identities seems likely merely to reinforce these stereotypes.

            For these particular reasons, as well as the more general reasons alluded to in the first part of this article, the ultimate referent object in the following discussion will be the Burundi individual – the Murundi.[10] It is interesting (and indeed heartening) to note in this context that the Cairo Declaration which resulted from the 1995 African Heads of State summit declares that “the security of individuals was a fundamental problem in Burundi” (Evans, 1997: 34). 

Security for the Murundi

            Focusing on individual security as the ultimate referent object in any attempt to understand the theory and practice of security inevitably opens up a ‘broader’ conception of security as well, for it is not only issues related to military security which impinge upon any individual’s security situation (Wyn Jones, 1999). Therefore other security concerns will also be discussed in the remainder of this article which will review the contemporary security situation as it affects the Murundi before suggesting some possible ways forward. However, given the catastrophic effects of the on-going uncivil war, the section will begin by focusing on the seemingly endless spiral of ethnically based violence.

            Even though the attempted coup in 1993, during which President Ndadaye was murdered, did not succeed as the junior officers who carried it out were disavowed by the Chief of Staff, it did generate a debilitating vacuum in Burundian politics. Surviving members of the government gradually re-emerged from foreign embassies where they had sought refuge from the coup and subsequent massacres, but almost three months were to elapse before a new president could be sworn in and a government established. Moreover, despite the consensual character of the new government based as it was on power-sharing between a Hutu President and Tutsi Prime Minister, the mutual mistrust between both groups grew as ethnically motivated violence continued (Des Forges, 1994; Chretien, 1996: Evans, 1997). Like his predecessor, new President Ntaryamira was destined to enjoy only a brief period in office as he was killed when the plane carrying him and Rwandan President Habyarimana was shot down outside Kigali on the 6 April 1994. This action was the signal for the start of the appalling genocide in Rwanda. In contrast Burundi initially remained relatively calm. However, the situation was extremely tense. Yet another Hutu, Sylvestre Ntibantunganya, was elected President by the National Assembly, while the country’s Constitution was suspended and replaced by a Convention of Government. However, this could not halt Burundi’s slide towards civil war. Indeed some analysts have argued that the suspension of the constitution and the establishment of the National Council for the Defence of Democracy (CNDD) along with its armed branch the Front for the Defence of Democracy (FDD) were partly responsible for the deterioration in the situation (Chretien, 1996: 210; Evans, 1997).

            Since then fighting between the armed forces and various groups of Hutu militia, massacres of (predominantly Hutu) civilians, arbitrary detention and torture have all become commonplace. By the summer of 1996 Burundi’s internal situation deteriorated to the extent that African state leaders decided to give security assistance to the country. This in turn was the prelude to yet another military coup. President Ntibantunganya was forced to flee and the perpetrators of the coup offered the Presidency to former President Buyoya. Since then, although Buyoya has gradually introduced various conciliatory measures at the political level under considerable external pressure, the violence and distrust between both sides of the ethnic divide has continued (Evans, 1997: 36-7; Amnesty International, 1996). A cease-fire between all groups involved in the fighting agreed on the 20 June 1998 has had little or no effect (Burundi-Bureau, 1998b; Radio Burundi, 22 June 1998). As a result, the life of almost every Murundi is currently characterised by extreme insecurity.

            That said, it is important to understand that this insecurity is not only a function of the fear generated by living in a country roamed by marauding bands of killers – uniformed or otherwise. Rather this insecurity is manifested in areas of life not normally considered to fall within the purview of security specialists. One good example is housing. Housing is usually regarded as a welfare issue, but in Burundi regroupment camps (regroupés) are central to the security concerns of individuals and their families.

            Since early 1996, civilians from rural areas have been forced to leave their homes and settle in so-called regroupment camps. The government has claimed that this has been for the civilian’s own protection because of the ongoing fighting (Amnesty International, 1997: Evans, 1997: 39). However, people have been kept in the camps long after the fighting in their areas has subsided, fuelling the suspicion that the regroupment policy is based more on a desire to isolate Hutu rebel groups from potential and actual supporters than to protect the civilian population. There is ample evidence to support such suspicion. Population movement seems to have been effected almost entirely through coercion in operations often characterised by arbitrary arrests, rapes and massacres. Subsequently many have ‘disappeared’ from the camps and Amnesty International (1997; 1998) have concluded that these have been victims of extrajudicial executions. Moreover, the camps have proved to be anything but secure for their inhabitants. They have been subject to countless attacks in which hundreds have been killed and raped, and the hygiene and health conditions inside are almost uniformly appalling. Finally, several camps have been connected to systematic attempts at political indoctrination in which residents have been “re-educated” or “de-toxicated” (Amnesty International, 1997; Burundi-Bureau, 1998a).

            The government argues that both Hutu and Tutsi are equally affected by the regroupment system, but in fact most Tutsi live in so-called displaced peoples camps (déplacés). Although the conditions within these camps are equally unhealthy and unhygienic, there are crucial differences between them and the regroupment camps, most notably in terms of freedom of movement. Whereas the residents in the regroupés are not allowed to leave their camp or have had restrictions placed on their movement, the inhabitants of the déplacés can move freely, and indeed, move from the camps if they so wish. Hutu fears that the regroupment policy is intended to effect a permanent shift in settlement patterns in Burundi are heightened by the fact that many of the residents’ properties were burned and destroyed shortly after they were taken away to the camps. Furthermore, the few who have been allowed to leave the camps have been resettled along roads and not allowed to go home. Several hundred thousand people in Burundi live with insecurity generated by this situation (Burundi-Bureau, 1998a; Amnesty International, 1997).

            The Burundian judicial system represents another major source of insecurity for individuals in Burundi. As UN Secretary-General Kofi Annan pointed out during his visit to Burundi in May 1998, this is a system in desperate need of reform (see UN Press Release, 1998). Far from conforming to the tenets of impartiality and fairness, the judicial system is virtually totally dominated by Tutsi. According to Albert Mbonerane (1997: 3) they use the judiciary as “an instrument of repression in the service of the government” instead of as “the guarantor of impartial application of the law”. This abuse became even more commonplace after July 1996 when Buyoya took power. The already suspended constitution was abolished and the judiciary and legal system even further politicised. This has resulted in great and manifest injustice for the Hutu. Hutu are not allowed legal representation, and it has been widely claimed that many are tortured in order to make them confess to crimes they have not committed (Amnesty International, 1998). There are great inequalities before the law. Tutsi are often set free after having committed the same crimes for which Hutu are executed. Court cases which result in execution sentences often do not take more than a few hours (Mbonerane, 1997; Amnesty International, 1998). The courts are arenas where any meaningful concern for justice has been replaced by a desire to reinforce ethnic supremacy (Mbonerane, 1997; UN Press Release, 1998). All of this is compounded by the fact that many judges are inexperienced and without proper qualifications. This, together with the fact that legal texts are written in French – a language that only 10 % of the population understand – adds to the arbitrary character of the system (Mbonerane, 1997; Amnesty International, 1998). This ethnically-biased judicial apparatus adds yet another layer of insecurity to the daily life of the Murundi.

            Other factors impact in equally profound ways on individual security in  Burundi; in this case factors, namely poverty and starvation, tend not to respect the boundaries between ethnic groups. Again these are issues which Traditional Security Studies has tended to regard as lying beyond it’s intellectual remit, but when individual security is placed central stage it is clear that they create extreme insecurity for very many individuals.

            As Burundi is not blessed with many mineral resources, its economy is predominantly agricultural. Coffee is its main export product and represents 80% of the country’s foreign exchange earnings. This makes the Burundian economy very vulnerable to changes in the world market. Otherwise, 9 out of 10 inhabitants depend on subsistence agriculture, which means that the security to plant, tend and harvest crops takes on especially crucial importance (Des Forges, 1994: 203; Malkki, 1995: 20). An economic reform programme was introduced in 1991 but has had no effect as the civil war has effectively ruined the economy with crops destroyed, population displaced and fertile land rendered unsafe or inaccessible by the conflict. Malnutrition and the diseases that thrive among a famine-weakened population have become commonplace, especially so in the camps (Des Forges, 1994: 206; Malkki, 1995: 287; Evans, 1997: 20; UN Press Release, 1998; Burundi-Bureau, 1998a). Quite apart from the direct effects on the subsistence economy, the insecurity in Burundi has slashed coffee revenues further damaging an already debt-burdened economy. All these problems have been further compounded by the external reaction to the 1996 coup when that many aid donors withdrew their support and neighbouring countries imposed economic sanctions (The Economist, 1996; Makinda, 1997: 382; Evans, 1997:20, 36). However, Burundi’s economic problems and the resulting insecurities should be seen in a broader context than that of civil war and sparse resources. Even without these debilitating specific problems, Burundi would still be massively disadvantaged by the iniquities and inequalities of the world economic system. 

Possible pathways towards emancipation

            CSS rejects the ‘scientific objectivist’ conception of knowledge on which Traditional Security Studies is based and argues instead that “both the subject and the object of scientific activity are socially constituted” (Dubiel, 1985: 104; Wyn Jones, 1999). One implication of this is that all theories whatever their claims to the contrary, prioritise and privilege certain actors and organising principles: or in Robert Cox’s (1981:128) well worn phrase, “all theory is for someone and for some purpose”. In the Burundian context, the whole aim of any CSS understanding of the situation will be to generate knowledge which can assist in some way in emancipating the men and women of that country from the insecurity and fear in which they find themselves. Within the given space-constraints, we are not in a position to develop a full analysis here. What we will attempt to do in this final section is therefore to suggest the methodology through we would suggest that emanicpatory possibilities may be identified and then provide some empirical examples.

            Methodologically, the critical approach underpinning CSS holds that there is no simple causal explanation of the cause or causes of conflict; neither is there a solution (or solutions) which can simply be applied universally. Neither does it posit emancipation as an end-point – a destination which can be arrived at and at which history will somehow come to an end. Rather a Critical Security approach, with its stress on immanence, focuses on the development of unfulfilled potentials within the actual conflict situation as it stands, as part of an emancipatory process (Wyn Jones, 1999).

            When it comes to potential pathways to a more emancipated Burundi, a few routes suggest themselves which could form part of such an emancipatory process. The ongoing negotiation process in Arusha, Tanzania is plainly one. The most recent round of talks took place on 13-22 October 1998 and was the third such meeting in a process which started on 15-21 June 1998 and continued on 20-29 July 1998. They included representatives of the Burundian Government, the political parties (both the internal and external branches) and the armed groups involved in the fighting. The negotiations, chaired by the former Tanzanian President Mwalimu Julius Nyerere, were organised around the following themes: “the nature of the Burundian conflict”, “democracy and good governance”, “peace and security”, and “development and rebuilding” (Burundi-Bureau, 1998c). From a Critical Security perspective it is interesting to view these negotiations in the light of Habermas’ notion of unfulfilled potential, namely the potential for reaching mutual understanding (and potentially at least, accommodation) through undistorted communication inherent in speech (Wyn Jones, 1999; Jones, 2000). To what degree are the parties to the discussion engaged in a genuine and inclusive attempt to promote dialogue with the aim of developing mutual understanding? Or are some or even all merely engaged in a cosmetic exercise in order to placate various constituencies either within or external to Burundi while hoping for (an ultimately unobtainable) violent solution? Obviously the answers to these questions are difficult to gauge without direct access to the talks and the participants. However there are at least some positive signs that there is increasing consensus both about the negotiations themselves and their contents. The government, the National Assembly and the FRODEBU party signed a partnership agreement in the summer of 1998 agreeing to co-operate in peace efforts, and Buyoya was sworn in as President by the National Assembly. This latter development is potentially significant as the National Assembly is the only ‘survivor’ from the 1993 political reforms and was part of the structure overthrown by the 1996 coup (Pitman, 1998). Another potentially significant straw in the wind is that elements of Burundian civil society have shown interest in participating in the negotiation process, and  this question was put on the agenda for the October 1998 round of talks with several civil society groups travelling to Arusha to make representations to the delegates. Among these were women’s groups hoping to promote the inclusion of more women into the peace efforts (Agence France Press, 1998; Burundi-Bureau, 1998b; 1998c). However despite these positive signs, there is also much cause for caution before jumping to any overly-optimistic conclusions. There remains dissent about the talks within the various parties to them and fighting still persists.

            In addition to this ‘grand-plan approach’ to attaining a more just and secure environment for the Barundi, there are also smaller scale initiatives which might well make a positive contribution. One is a civil society-based initiative which is attempting to ameliorate and even transcend ethnic boundaries by establishing a multi-ethnic or non-ethnic radio station. Radio is a very powerful medium in a country where the transport infrastructure is poor, where most political activity takes place in the hills among a dispersed rural population, and where other means of communications are scarce (Malkki, 1995: 20). It played an important role in the genocide in 1972 when it was used to encourage Tutsi to kill their Hutu neighbours. ‘Hate radio’ has been used by both sides for the same purpose during the civil war over the last five years. However, groups within civil society have started broadcasting programmes aimed at building trust and securing reconciliation between the groups. An American Non-Governmental Organisation (NGO), Search for Common Ground, has initiated a project where Hutu and Tutsi journalists work together. The hope is that through co-operating in a work environment these journalists will begin to view each other as individuals and fellow professionals rather than simply representatives of rival ethnic groups, and that this will result in a spirit of tolerance which will then be communicated to the listeners (Des Forges, 1994: 204; Evans, 1997: 40). This and other similar radio initiatives may help to prepare the ground for a more general process of communication and reconciliation within Burundi.

            Other potentially important ‘grass-roots’ initiatives have concentrated on involving new actors in the process. One example is the Center for Women, established in January 1996, again by Search for Common Ground. This has worked to encourage and facilitate the participation of women in the work for peace. The centre’s efforts range from providing administrative infrastructure and meeting facilities to training women in peaceful conflict resolution techniques. Women from the different regions and ethnic communities have been encouraged to discuss and promote reconciliation within their families and communities. In this way, individuals and groups who are not traditionally considered as ‘political’ actors are included in the communication process, something that may further reinforce and dynamise the negotiations taking place at both national and provincial levels (Office for the Coordination of Humanitarian Affairs, 1998). 

Concluding Remarks

            This article has attempted two tasks. First it attempted to outline and explain in general terms the approach to the study of security that has become known as CSS. In particular, it stressed how CSS eschews the traditional approach’s concentration of the military dimension of state security and concentrates instead on a more holistic understanding of individual security concerns – that is on how CSS both ‘broadens’ and ‘extends’ the orthodox understanding of security. It then illustrated these arguments by way of a discussion of Burundi. This served not only to provide an empirical illustration of the necessity of reconceptualising security, but also to provide a more coherent conceptual framework for understanding that benighted country’s problems.

            It may well be objected that many analyses of Burundi already adopt this expanded conception of security as their starting point. This is an argument that we would certainly accept but would make two points in response. First, while it has long been the case that area specialists often utilise such a conception, much conceptual work in International Relations and Security Studies is still mired in idealised models of international politics, which are neither ideal nor representative of reality. While this situation persists there remains the need to emphasise points that some may well regard as obvious – even ‘common sensical’.  Secondly, the major advantage of the CSS approach is the consistency and self-reflexivity of its theoretical framework. Much important empirical work lacks this and as a result suffers from logical lacunae or contradictions. By outlining a broad theoretical framework within which security can be properly understood, CSS can give empirical work greater direction and hopefully maximise its positive impact. For above all else, the example of Burundi illustrates the need to develop alternative strategies for attaining human security.



Agence France Press. 1998. “Burundi-Talks. Delegates at Burundi Talks Call for Transitional Government,” 16 October.

Amnesty International. 1998. Annual Report on Burundi. New York: Amnesty International Publications.

----1997. “Forced Relocation: New Patterns of Human Rights Abuses.” AI Index: AFR 16/19/97, 15 July.

----1996. “Burundi: Leaders Change, but Human Rights Abuses Continue Unabated,” 5 August.

Bernstein, Richard J. 1988. Beyond Objectivism and Relativism: Science, Hermeneutics, and Praxis. Philadelphia: University of Pennsylvania Press.

Bilgin, Pinar, Ken Booth and Richard Wyn Jones. 1998. “Security Studies: The Next Stage?” Nação e Defensa, No. 84, pp. 131-157.

Booth, Ken, ed. Forthcoming. Security, Community and Emancipation: An Introduction to Critical Security Studies. Boulder, CO: Lynne Rienner Publications.

----1997. “Security and Self: Reflections of a Fallen Realist.” In Keith Krause and Michael C. Williams, eds., Critical Security Studies: Concepts and Cases. London: UCL Press.

----1991. “Security and Emancipation.” Review of International Studies, Vol. 17, No. 4, pp. 313-26.

Booth, Ken and Peter Vale. 1997. “Critical Security Studies and Regional Insecurity: The Case of Southern Africa.” In  Keith Krause and Michael C. Williams, eds., Critical Security Studies: Concepts and Cases. London: UCL Press.

Brown, Michael E., Owen R. Coté Jr., Sean M. Lynn-Jones and Steven E. Miller, eds. 1997. Nationalism and Ethnic Conflict. Cambridge, MA: The MIT Press.

Brown, Michael E., Sean M. Lynn Jones and Steven E. Miller, eds. 1995. The Perils of Anarchy: Contemporary Realism and International Security. Cambridge, MA: The MIT Press.

Burundi-Bureau. 1998a. “The Regrouping Camps Remain in Effect.” Press Release No. 184, 29 September.

----1998b. “Burundi-Tanzania: Third Round of the Negotioations of Arusha.” News Brief  No. 162, 11 October.

----1998c. “Burundi-Tanzania: The Negotioations Begin Again by Intense Consulations.” News Brief No. 163, 13 October.

Buzan, Barry. 1991. People, Sates and Fear: An Agenda for International Security Studies [Second Edition, first edition published in 1983]. London: Harvester Wheatsheaf.

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

Cox, Robert W. 1981. “Social Forces, States and World Orders: Beyond International Relations Theory.” Millennium: Journal of International Relations, Vol. 10, No. 2, pp. 126-55.

Chretien, Jean-Pierre. 1996. “Burundi: The Obsession with Genocide.” Current History, May, pp. 206-210.

Des Forges, Allison. 1994. “Burundi: Failed Coup or Creeping Coup?” Current History, May, pp. 203-207.

Deudney, Daniel. 1990. “The Case Against Linking Environmental Degradation and National Security.” Millennium: Journal of International Studies, Vol. 19, No. 3, pp. 461-76.

Dubiel, Helmut. 1985. Theory and Politics: Studies in the Development of Critical Theory, translated by Benjamin Gregg. Cambridge, MA: MIT Press.

Evans, Glynne. 1997. “Responding to Crises in the African Great Lakes”. Adelphi Paper, No. 311. Oxford: Oxford University Press.

Green, Philip. 1973. “Strategy, Politics, and Social Scientists.” In Morton A. Kaplan, ed., Strategic Thinking and Its Moral Implications. Chicago: University of Chicago.

----1966. Deadly Logic: The Theory of Nuclear Deterrence. Columbus, OH: Ohio State University Press.

Jones, Deiniol Lloyd. 2000. “International Mediation as Communicative Action: The Creation of Cosmopolitan Power.” In Richard Wyn Jones, ed., Critical Theory and World Politics. Boulder, CO: Lynne Rienner Publications.

Kaldor, Mary, ed. 1991. Europe from Below: An East-West Dialogue. London: Verso.

Krause, Keith and Michael C. Williams, eds. 1997. Critical Security Studies: Concepts and Cases. London: UCL Press.

Lemarchand, René. 1994. Burundi: Ethnocide as Discourse and Practice. Cambridge: Cambridge University Press.

----1989. “The Report of the National Commission to Study the Question of National Unity in Burundi: a Critical Comment.” The Journal of Modern African Studies, Vol. 27, No. 4, pp. 685-690.

----1970. Rwanda and Burundi. London: Pall Mall Press.

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

Lynn-Jones, Sean M. and Steven E. Miller. 1995. Global Dangers: Changing Dimensions of International Security. Cambridge, MA: M.I.T Press.

Makinda, Samuel M. 1997. “Ethnic Claims Versus Political Imperatives in Rwanda and Burundi.” Security Dialogue, Vol. 28, No. 3, pp. 382-4.

Malkki, Liisa H. 1995. Purity and Exile: Violence, Memory, and National Cosmology Among Hutu Refugees in Tanzania. Chicago: University of Chicago Press.

Mbonerane, Albert 1997. “Burundi: The Judiciary - A System in Need of Reform.” ANB-BIA Supplement, No. 337, pp. 1-4.

Office for the Coordination of Humanitarian Affairs. 1998. “Internal Discussions on Peace Process Continue.” Information Bulletin. Bujumbura, 7-13 September.

Osgood, Robert E. and Robert W. Tucker. 1967. Force, Order, and Justice. Baltimore: The Johns Hopkins Press.

Pitman, Todd. 1998. “FOCUS-Burundi Military Ruler Sworn in as President.” Reuters, 11 June.

Radio Burundi. 1998. “Burundi Government, All Groups Attending Talks Agree to Cease-Fire.” BBC Monitoring (23 June), 22 June.

Rapoport, Anatol. 1970. “Critique of Strategic Thinking.” In Naomi Rosenbaum, ed., Readings on the International Political System. Englewood Cliffs, New Jersey: Prentice-Hall.

----1964. Strategy and Conscience. New York: Harper and Row.

----1960. Fights, Games and Debates. Ann Arbor: University of Michigan Press.

Reus-Smit, Christian. 1992. “Realist and Resistance Utopias: Community, Security and Political Action in the New Europe.” Millennium: Journal of International Studies, Vol. 21, No. 1, pp. 1-28.

Reyntjens, Filip. 1993. “The Proof of the Pudding is in the Eating: the June 1993 Elections in Burundi.” The Journal of Modern African Studies, Vol. 31, No. 4, pp. 563-583.

Risse-Kappen, Thomas. 1994. “Ideas Do Not Float Freely: Transnational Coalitions, Domestic Structures, and the End of the Cold War.” International Organisation, Vol. 48, No. 2, pp. 185-214.

Shaw, Martin.1994. Global Society and International Relations. Cambridge: Polity Press.

The Economist . 1996. “The Other War in Central Africa”, 14 December, pp. 67-68.

United Nations Press Release. 1998. SG/SM/6554 AFR /58, 12 May.

Walker, R.B.J. 1997. “The Subject of Security.” In Keith Krause and Michael C. Williams, eds., Critical Security Studies: Concepts and Cases. London: UCL Press.

Walt, Stephen M. 1991. “The Renaissance of Security Studies.” International Studies Quarterly, Vol. 35, No. 2, pp. 211-39.

Wyn Jones, Richard. 1999. Strategy, Security and Critical Theory. Boulder, CO: Lynne Rienner Publications.

Wyn Jones, Richard and Ken Booth. 2000. “Critical Security Studies.” Entry in World Encyclopedia of Peace. Oxford: Pergamon.

Wæver, Ole, Barry Buzan, Morten Kelstrup and Pierre Lemaitre. 1993. Identity, Migration and the New Security Agenda in Europe. London: Pinter.

[1] A previous draft of this section has appeared in an entry on “Critical Security Studies” prepared by Richard Wyn Jones and Ken Booth for The World Encyclopedia of Peace (2000).

[2] It should be noted that the portrayal of CSS offered by  Buzan , Wœver and de Wilde is to say the least contentious, and in our view, descends into caricature, but to pursue this issue fully is beyond the scope of this paper.

[3] We should note that proponents of Critical Security Studies realise that the tasks of conceptualisation and the application of concepts to concrete empirical examples are dialectically inter-related. Their conceptual framework is not regarded as immutable but is rather subject to transformation in the light of practice (Wyn Jones, 1999).

[4] Please note that this article deals with the situation in Burundi up until the autumn of 1998.

[5] The  memories of the 1972 genocide are important in explaining the motivation for the killings in the communes of Ntega and Marangara in 1988 (Lemarchand, 1994; Malkki, 1995). The arrest of several Hutu and unannounced army manoeuvres were interpreted as an attempt to repeat the 1972 atrocities, and as a result a number of Tutsi were killed in reaction (Lemarchand, 1989; Evans, 1997). The armed forces reacted in their now all-too­familiar manner, causing 50,000 to flee. Total numbers of deaths is estimated at 1,500 (Malkki, 1995; Chretien,1996).

[6] The estimates of the death toll in these killings vary greatly - from 30,000 (Des Forges, 1994) to 100,000 (Malkki, 1995; Chretien, 1996). There are also significant variation in the literature as to which group suffered most. Malkki (1995) claims that most of the killed were Hutu; Des Forges (1994) suggests the losses for the Hutu and Tutsi may have been equal; while Chretien (1996) concentrates on the killing of Tutsi. There is also some variation concerning numbers of refugees, but it is clear that a majority of them fled to neighbouring countries, while a proportion fled to other parts of Burundi.

[7] Barundi is the plural for the Burundi people, while Murundi is the singular term. Kirundi is the language spoken in Burundi (Lemarchand, 1970: xiii).

[8] It has been claimed that political discourse in Burundi had already become ethnicised before October 1993 when, towards the end of that year’s election campaign, ethnic distinctions were stressed in order to galvanise political support (Des Forges, 1994: 205; Chretien, 1996: 208; Evans, 1997: 25-26).

[9] This is not to imply that work to ‘denaturalise’ the Hutu-Tutsi division is of merely esoteric concern. Rather it can play a part in a long-term strategy which aims to transcend the divide. In this context, work like Lemarchand’s Burundi: Ethnocide as Discourse and Practice (1994) is potentially very useful.

[10] See note eight.