The concept of ‘safe areas’ emerged in the early 1990s as a way of responding to increasing displacement triggered by internal conflicts. As a form of protection, their record was mixed—for every success like northern Iraq in 1991, there was a failure like the collapse of the Srebrenica safe area in 1995. But why did the safe area concept itself emerge at this time? Traditionally, safe areas were akin to humanitarian spaces anchored in consent. The shift in the early 1990s was to replace consent with an international military presence, including military forces and peacekeepers. This article argues that this shift was only possible because of two critical changes which occurred within the United Nations: the recognition that civilian protection represented an international problem and the UN Security Council broadening how it interpreted the notion of ‘threats to international peace and security’ to include issues such as forced migrant flows.
Once again, safe areas are being proposed as a means for civilian protection.1 Safe areas represent a novel form of humanitarian space, which was first used in the 1990s. They refer to ‘operations undertaken by international actors which have the primary purpose of providing direct protection to civilians and internally displaced persons (IDPs) within a state’s borders in a temporary and designated geographic area. The principle underlying safe areas is that in situations of widespread conflict and atrocities, safe areas allow threatened civilian populations to remain within their state while receiving physical protection and humanitarian assistance’.2 In this sense, safe areas are a way in which the United Nations, other international actors, and states can provide civilians affected by conflict or atrocity crimes with direct and immediate protection and without bearing the larger costs of full-scale invasion or regime change.
And yet, why did the safe area concept, as a novel form of action, happen to emerge specifically in the early 1990s? Norms emerge and exist in a complex space. Individual international norms help to constitute shared understandings of appropriate behavior held by states and other actors. The emergence of the safe area concept, I argue, was driven by a process of normative contingency as three different sets of normative understandings—around safe zones in international humanitarian law, around refugee containment and the protection of internally displaced persons, and around the practices of the UN Security Council itself—were brought together during a rare crisis event—the 1990–1991 Gulf War—to recast how civilian protection was understood at the international level.
The first set of normative understandings was the idea of a particular space in war zones which could be placed hors de combat as established through International Humanitarian Law (IHL) and, in particular, the Geneva Conventions. The second, with an even longer tradition, are norms around the protection of forced migrants. While refugee status is anchored in the 1951 Refugee Convention, by the late 1980s states in the developed world were increasingly seeking to contain would-be refugees leading to growing numbers of internally displaced persons (IDPs). Hence, by 1990, there was a growing focus on the part of the international community with the need to assist these IDPs and provide them with protection. The third was the practices of the UN Security Council itself. In the late 1980s and early 1990s the Council as an institution was steadily reinterpreting the Charter requirements that it maintain international peace and security. For the first time, both human rights and the issue of displacement—specifically large-scale refugee flows—were integrated into its agenda. Thus, we see at this point in time the intersection between an established conception of safety zones in international humanitarian law and two sets of norms—around forced migrant protection and the role of the UN Security Council—which were in the process of transformation.
This reorientation, I argue, led to the creation of the first true ‘safe area’ by the United States, United Kingdom, and France in northern Iraq in April 1991, following the Gulf War and as a response to a massive exodus by the Iraqi Kurds. The northern Iraq safe area represented an evolution of the safety zones laid out in the Geneva Conventions, one which would be protected not through belligerent agreement but through force of arms sent in by other states. At the same time, the creation of the safe area allowed the US to avoid both the provision of asylum to the would-be refugees or to push its Turkish ally to reopen its borders to allow the Kurds to seek asylum there.3
More widely, it created a new precedent for approaching civilian protection spatially. What the crisis in northern Iraq did was to cause key actors to think about these norms in different ways, linking protection to the deployment of troops within another country’s territory. Thus, in the longer term the decision to create this safe area triggered lasting debates about how the United Nations should engage in civilian protection during conflicts, debates which led to the creation and collapse of the Srebrenica safe area and the introduction of the Protection of Civilians concept. In other words, normative contingency—a process of contestation set off by a crisis which led to a novel linkages of these three sets of normative understandings—created a new normative space through which safe areas were understood as a possible international civilian protection response.
I begin by exploring how normative contingency led to the linkage of three discrete sets of norms—detailing spatial forms of protection offered to civilians through international humanitarian law, around forced migrant protection, and the changing role of the UN Security Council—to create a new form of international response for problems of civilian protection. The second part of the article then explores how these three sets of norms led to a remarkable shift in practice on the part of the UN Security Council and its member states; first condemning internal repression in Iraq for the first time, then the US, UK, and France-led (and unauthorised) humanitarian intervention in northern Iraq. The final section explores the legacies of this shift both in terms of the use of the safe areas concept itself and its linkage to the containment of forced migrants, but also its wider effects on how peacekeeping was undertaken in the 1990s.
The Role of Normative Contingency
How international norms emerge and become implemented within states and other actors remains debated within international relations scholarship. For Finnemore and Sikkink, in their norm life cycle model,4 the process occurs as norm entrepreneurs seek to place new issues on the international agenda. The new norm may then be adopted by states and, following a new norm’s adoption by a critical mass of states, internalised, which means conformance with the norm becomes almost automatic. However, such a view neglects the ‘impermanence of internalization’,5 the idea that a norm may not assume a fixed definition accepted by all actors through such a process. Instead, recent work has focused on a conception of norms as works in progress which, in Krook and True’s words, ‘tend to be vague, enabling their content to be filled in many ways and thereby to be appropriated for a variety of different purposes’.6 Thus, Wiener argues that a norm’s meaning is ‘constituted through an interactive process. Interpretation is thus derived from the social practice of enacting meaning that is used in a specific context’.7
It is through such a process that normative contingency can play a critical role in developing not only new interpretations of specific norms, but also in linking together disparate norms in new ways. Wiener views the contingency of normative meaning to indicate a ‘change of constitutive social practices both cultural and organizational, and hence normative meaning over time’. Such a change, she argues, is likely to enhance the contestation of meaning over a particular norm, as are the changes of social practices which change the social environment and reference frame provided by a social institution and a situation of crisis which raises the stakes for shared understanding as time for deliberation is more limited.8 Thus, ‘points of contestation are triggered by contingency, out of context application and in situations of crisis’.9
While Wiener sees normative contingency as helping to explain the likelihood of contestation occurring around a single norm, I argue that this can also play a role in how states and other actors at the international level understand and link together a number of different norms, including both sets of norms which form a regime but also norms which are otherwise discrete and separate.10 Such linkages between norms are not unique; Finnemore notes that many of the norms associated historically with intervention are interwoven and interdependent: ‘Humanitarianism—its influence and definition—is bound up in other normative changes, particularly sovereignty norms and human rights norms’.11 In the case of safe areas, three key sets of norms played a role, relating to international humanitarian law; the protection of refugees and internally displaced persons; and the practices of the UN Security Council. I now turn to each of these.
Safe Areas and International Law
The protection of civilians in wartime is a key legal norm developed within international humanitarian law. But such forms of protection have long had a spatial element associated with them. The first legal prohibitions emerged with the 1907 Hague Convention.12 Its Article 25 noted that ‘the attack or bombardment, by whatever means, of towns, villages, dwellings, or buildings which are undefended is prohibited’. It also required, in Article 27, that all necessary steps be taken to spare ‘buildings dedicated to religion, art, science, or charitable purposes, historic monuments, hospitals, and places where the sick and wounded are collected, provided they are not being used at the time for military purposes’.
However, the 1929 Geneva Convention was silent on these questions, which other organisations viewed as problematic.13 This led in 1934 to a draft Monaco Convention. The Monaco Draft envisioned a new treaty to offer protection to both military noncombatants (principally the wounded) and to the civilian population. It did this by requiring belligerents to guarantee protection to ‘sanitary towns’, used for the requirements of medical services only and with any other military utilisation excluded, provided formal notification was provided. In addition, undefended towns ‘containing no military objectives’ could be given a similar status to sanitary towns provided notification was also given.14
The Monaco Draft presaged the creation of a number of safety zones in the late 1930s. During the Spanish Civil War in November 1936, General Franciso Franco designated portions of Madrid as a neutral zone following days of aerial bombardment of the city by his forces, which had caused the government to flee the city.15 The zone, Sandoz notes, ‘was by and large respected’ even though ‘it had not been created by the party in whose territory it was situated, nor was it subject to supervisory procedures. Moreover, no formal agreement had been reached on establishing the zone which was, in effect, the result of a unilateral military decision…’16 While they respected the zone, Franco’s troops laid siege to Madrid, which fell on 27 March 1939.
Two zones were also established in China in 1937 following the Japanese invasion. In Shanghai, Robert Jacquinot de Besange, a French Jesuit, set up a private de-militarised zone which was administered by a committee reflecting the international communities of the area and which protected some 300,000 people. Similar safety zones were established in Nanjing, Wuhan, Canton, and other Chinese cites.17 The ICRC also played a direct role in negotiating and then administering a neutralised zone in Jerusalem in 1948 to protect its 150,000 inhabitants following the anticipated withdrawal of the United Kingdom.18
The Monaco Draft did not enter into force, but its ideas were coupled with the ICRC’s own experience with safe areas to include the idea of protected hospital zones and localities within the 1949 Geneva Conventions and, subsequently, the 1977 Additional Protocols. Geneva Convention IV provides for a range of different spatial areas to protect civilians. First, Article 14 of that Convention, on ‘Hospital and Safety Zones and Localities’, allows belligerents to establish, through mutual agreement, hospital and safety zones. However, the people who can be protected from the effects of war within them are limited to the ‘wounded, sick and aged persons, children under fifteen, expectant mothers and mothers of children under seven’.19 By contrast, Article 15 allows for the negotiated creation of neutralised zones ‘in the regions where fighting is taking place’ in order to shelter two groups: wounded and sick combatants and non-combatants; and civilian persons who take no part in hostilities.20 These neutralised zones are more encompassing than hospital and safety zones by allowing for the inclusion of all civilians.
Additional Protocol I of 1977 further extends these obligations.21 In particular, Article 59 provides for the creation of non-defended localities covering the entire civilian population.22 These localities must have all combatants and mobile weapons and military equipment removed, and neither hostile acts nor support for military operations can be undertaken. Critically, if these requirements are met, the locality can be declared unilaterally without the consent of other parties. Once established, there is a blanket prohibition against attack: ‘it is prohibited for the Parties to the conflict to attack, by any means whatsoever, non-defended localities’.23 The Protocol also allows for the creation of demilitarised zones, but these need to be established through express agreement.24 Each of these Conventions is widely ratified, with the Geneva Conventions having 196 state parties, and the Additional Protocol I having 174.25
There is no similar explicit provision for non-international armed conflicts. But, as Clapham, Gaeta, and Sassoli note, ‘Common Article 3 [ of the Conventions] states that the parties to the conflict may bring into force between them all or part of the provisions relating to international armed conflicts … and thus also those relating to safety zones and localities’.26 Further, it has been suggested that these provisions have become customary international law as well.27
These prohibitions have been brought into international criminal law. The Statute of the International Criminal Tribunal for the former Yugoslavia established that the Tribunal had the power to prosecute violations including ‘attack, or bombardment, by whatever means, of undefended towns, villages, dwellings, or buildings…’28 The Rome Statue of the International Criminal Court includes as war crimes a range of violations, including ‘intentionally directing attacks against the civilian population…’ and ‘attacking or bombarding, by whatever means, towns, villages, dwellings or building which are undefended and which are not military objectives’.29
This may reflect issues in how signatories understand the obligations associated with such zones, as they are defined differently between Geneva Convention IV and Additional Protocol I, with Additional Protocol I allowing the creation of non-defended localities without negotiation or mutual agreement. Thus, while such zones are defined in law, how they should work in practice remains an open question.
Formalization of the system of protected zones in the Geneva Conventions has not led to any proliferation of such zones during conflicts that have arisen since the Conventions were adopted. Moreover, in the few instances where protected zones have since been created, they have been established in dire emergency and fall only very roughly into the categories envisaged in the Geneva Conventions, if at all.31
A Shift to Containment
The second important normative shift began in the 1980s, as states, particularly within the developed world, began reinterpreting their obligations towards refugees and other forced migrants. As with the Geneva Conventions, these standards had been clearly defined in international law, with the 1951 Refugee Convention both defining refugee states and establishing that all signatory states have a requirement to not ‘expel or return (“refouler”) a refugee in any manner whatsoever to the frontiers of territories where his life or freedom would be threatened on account of his race, religion, nationality, membership of a particular social group or political opinion’.32
But the Refugee Convention was designed primarily as an instrument to respond to the persecution of individuals by the state. As the incidence of internal conflicts rose beginning in the 1970s, this resulted in significant numbers of new refugees.33 Yet these refugees did not easily fall within this definition, as they were fleeing from situations of generalised violence or violence by non-state actors.34 This also meant that by the late 1980s, states in the developed world were facing growing numbers of asylum seekers, even as the end of the Cold War removed a clear interest for states to accept refugees from Communist regimes.35 In response, Western states began to create a range of deterrent and extraterritorial measures designed to contain would be refugees to their regions and even countries of origin.36
At the same time, internally displaced persons were increasingly being recognised as a distinct problem. Arguments made in favour of their inclusion within the 1951 Refugee Convention had gone nowhere.37 From the 1970s onwards, UNHCR did provide assistance to IDPs, but this was only when requested to do so by either the UN General Assembly or by the Secretary-General. This began to change in the 1980s. The 1986 Report of the Independent Commission on International Humanitarian Issues, noted that ‘the international community’s response to the plight of displaced people has been unsystematic. No single UN agency is specifically mandated to protect and assist internally displaced people’.38 Two major international conferences in 1988 and 1989 also recognised the need not only for legal instruments but also for mechanisms within the UN to respond to IDP assistance needs. These conferences served to raise the profile of IDPs among states in the United Nations, who began to push for more coordinated action. This culminated in 1990 to ECOSOC requesting the Secretary-General to initiate a review process.39 But debates over the response to IDPs remained divided over the issue of sovereignty. The resulting General Assembly resolution, 46/182,40 while it ushered in significant humanitarian reform within the UN system,41 did not specifically note the IDP issue. This marked ongoing concerns by states to ensure that the resolution included ‘a clear expression that governments had the sovereign authority to determine whether or not outside assistance would be provided to their citizenry’.42
Hence, by the late 1980s and early 1990s, the way in which forced migrants were viewed was in a state of flux at the international level. Previously welcoming states in the developed world were increasingly focusing on restricting refugee access to their own asylum systems. At the same time, IDPs were being recognised as a distinct phenomenon, but how they should be provided with assistance and protected given they remained within their own country remained unclear.
The Changing Role of the Council
Finally, the third major normative change was in how the United Nations Security Council approached its role. The UN Charter grants the Security Council the power to ‘determine the existence of any threat to the peace, breach of the peace, or act of aggression and shall make recommendations, or decide what measures shall be taken in accordance with Articles 41 and 42, to maintain or restore international peace and security’.43 And yet, how the Council interprets its role with respect to maintaining or restoring international peace and security changed markedly with the end of the cold war to include a focus on the ‘well-being of civilian populations…’44
However, the Gulf War was transformative in terms of how the Council approached this mandate. As American President George H.W. Bush and his National Security Adviser, Brent Scowcroft, would later note, in terms of the Gulf War, ‘we tried to establish a model for the use of force…Building an international response led us immediately to the United Nations, which could provide a cloak of acceptability to our efforts and mobilize world opinion behind the principles we wished to project’.45
The Council had only authorised force previously in the cases of Korea46 and Rhodesia.47 But even by the late 1980s, the Council was beginning to shift its practices towards civilian protection. This included successful diplomatic efforts to end the Iran-Iraq War,48 and a significant increase in peacekeeping operations: whilst the Council authorised a total of thirteen missions between 1948 and 1987, in the following six years alone thirteen new operations were authorised.49 And the nature of peacekeeping was changing. Operations deployed to Namibia and Nicaragua in 1989 marked the first efforts by the UN to engage in peacebuilding while the northern Iraq safe area presaged a move by the UN into peace enforcement missions in ongoing conflicts in Somalia and Bosnia.50 As Gray argues, with the Gulf War there was a newfound consensus ‘that it is for the Security Council to authorize Member States to take enforcement action, even if the precise legal basis for this in the Charter is not clear’.51
Equally important was that the Gulf War shifted how the Council approached questions of human rights. Walling notes that prior to Iraq’s invasion of Kuwait ‘human rights discourse was considered inappropriate for council discussion’.52 The Gulf War, she suggests, was a watershed for the Council because it ‘created a context in which it was possible for Security Council members to consider the relationship between human rights and international security’.53
The First Safe Area: Northern Iraq 1991
These three normative issues were brought together following the Gulf War and the liberation of Kuwait. The end of the war triggered two large uprisings in Iraq in March 1991, including the predominantly Kurdish population in the north of the country, and the predominantly Shiite populations in the south. Both uprisings were quickly crushed by the Iraqi military. In the north, this triggered a massive refugee flow as 400,000 mostly Kurdish Iraqis fled into Turkey and 1.3 million fled into Iran. Turkey responded to these flows by closing its border on the second of April because ‘it viewed the Kurdish refugee population as an ethno-national threat and feared its “Palestinianization”’.54
This left one million internally displaced Kurds within Iraq.55 The result was a massive humanitarian crisis in northern Iraq with a death rate of between 400 and 1,000 per day among the IDPs due to ‘hypothermia, exposure, exhaustion, and bacteria-ridden drinking water, which led to pneumonia, diarrhea, and cholera’.56 The crisis led to two different international responses. Interestingly, little explicit reference was made to international humanitarian law and to either safety zones or to non-defended localities. Instead, these responses were framed by changing conceptions of how the Security Council could respond to forced migrant flows within its Charter mandate, particularly when they created transboundary threats to other states. The first response, within the Security Council, was the passage of Resolution 688. Subsequently, the governments of the United States, United Kingdom, and France used that resolution to justify a military intervention into northern Iraq and the creation of the safe area on humanitarian grounds.
The push for an international response began from the Turkish government. In a letter to the Security Council on 2 April 1991, the Turkish government argued the actions of the Iraqi government ‘violate all norms of behaviour towards civilian populations and constitute a threat to the region’s peace and security’.57 But in a telephone call with US President George H.W. Bush, Turkish President Turgut Ozal had suggested broader international action, including an ‘appeal to the UN for a UN presence in northern Iraq to stop the killing’ while also accusing the Iraqis of ‘pushing’ refugees to the Turkish border.58 Two days later, the Iranian government also sent a letter, noting that while they would keep their border open ‘the prolongation of the situation … will have consequences that in fact threaten regional peace and security…’59
At the same time, the French government was also playing an active role in highlighting the crisis within the UN. They had raised the Kurdish situation directly in the Council on April 3rd 1991, arguing that ‘the Security Council has a duty to say something about this situation’.60 Roland Dumas, the French Foreign Minister, noted this situation ‘argued for the recognition of an international “duty to intervene,” to prevent gross violations of human rights and the rights of minorities’.61 French President Francois Mitterrand also suggested that the political and moral authority of the Council would be seriously affected if it failed to take action.62
These concerns led the Council to pass Resolution 688 on April 5th. While recalling Article 2(7) of the Charter and not explicitly invoking Chapter VII, the resolution not only condemned the repression ‘of the Iraqi civilian population in many parts of Iraq…’, and that ‘the consequences of which threaten international peace and security in the region’ but demanded that Iraq ‘immediately end this repression’.63 This resolution was a significant change in the Council’s practice. As Walling notes, the resolution ‘challenged the traditional meaning of sovereignty by interfering in the internal affairs of the Iraqi state and by linking minimal standards of human rights protection to the meaning of legitimate sovereign authority’.64
And it is important to note that this was a limited step in two ways. First, the Council demanded no further action from Iraq other than ending the repression and allowing immediate access for humanitarian organisations.65 The French government tempered its language in the Council, and rather than calling for a duty of intervention noted only that the actions against the Kurds were assuming ‘the dimension of a crime against humanity’ and that the demands made in the resolution ‘are the minimum that the members of the international community must make…’66
The second was that the majority of states involved in the deliberations over the resolution framed the issue exclusively in terms of the refugee flows being a threat to international peace and security. The Turkish representative noted that ‘there is no way in which what is going on in northern Iraq can be justified as an internal affair of that country … No country can cope with such a massive influx of destitute people fleeing for their lives’.67 The Iranian representative argued that flow of civilians added an international dimension to the crisis and threatened the security of neighbouring countries.68
States which opposed the resolution did so by arguing not that the problem did not exist, but that it was not within the Council’s mandate to respond. Thus, the Yemeni representative noted that it was a problem that did not threaten international peace and security;69 the Zimbabwean representative that it was a ‘domestic political conflict’;70 and the Cuban representative noted that while there were legitimate humanitarian concerns the Council had ‘no right to violate the principle of non-intervention’.71 The Iraqi government itself argued simply that the resolution was ‘a flagrant, illegitimate intervention in Iraq’s internal affairs and a violation of Article 2…’72
This limited language led to the USSR supporting the resolution, while China abstained, allowing it to pass (India, serving as an elected member, also abstained).73 The Chinese delegate noted that the refugee flows were concerning, but that a state’s internal affairs were involved in the question.74 David Hannay, then the United Kingdom’s Permanent Representative to the United Nations, noted that while the resolution contained no enforcement provisions, ‘the pattern of voting on the resolution made it clear that, if we had reverted to the Council with a request for increased authority we would have been very unlikely to have got it’.75
This was recognised as a significant shift in how the Council viewed its mandate to respond to forced migrant situations and to some instances of internal repression. Participants at a June 1991 conference on human rights protection for internally displaced persons noted ‘the resolution had made clear that civilians trapped in conflict situations should be helped no matter what the position of their government’.76 Stromseth similarly suggested it was precedential both in the Security Council being willing to ‘condemn internal repression as a threat to international and security’ and to spur ‘institutional reforms to make UN humanitarian responses more prompt and effective’.77 By significantly reinterpreting its mandate within the Charter, for the first time the Council had passed a resolution in which a country’s internal situation and the ensuing forced migrant flows led the Council to demand that the country alter its behaviour.
Operation Provide Comfort
Equally important, however, were the events that occurred in the wake of the Resolution which neither invoked Chapter VII nor authorised troop deployments into northern Iraq. And yet, within eleven days, the US, UK, and France announced an armed humanitarian intervention into northern Iraq. They argued that Resolution 688 provided a justification for it. In addition, they couched this effort within humanitarian language, focusing on an internal response to forced migrant flows which traditionally would have been responded to through asylum. No leader explicitly linked this back to the provisions of the Geneva Conventions or Additional Protocols, but, as McQueen notes, ‘community norms enabled the adoption of a safe haven approach because they opened the door to third party action in response to breaches of humanitarian law’.78
Some military operations were already occurring in Iraq—the same day that Resolution 688 passed, the American government announced that it would undertake airdrops of relief supplies into northern Iraq, but these were overflights only.79 At the same time, the US government was considering how to protect refugees along the southern border of Iraq as US troops withdrew. This led to internal debates on whether there needed to be ‘a specific “zone of protection” for refugees’ and whether Security Council Resolution 687, which set the comprehensive terms with Iraq ending the Gulf War, provided for ‘coalition military action if Iraqi forces enter[ed this] demilitarized zone’.80
On April 8th, however, in a speech to a European Community, British Prime Minister John Major proposed creating ‘safe enclaves’ in northern Iraq to the European Community Summit. Major argued that this would not constitute interference in Iraq’s internal affairs because it was based on ‘protection of a population from persecution’.81 Following a suggestion from David Hannay, Major later changed the language to safe havens to avoid suggesting the creation of an autonomous Kurdish area.82 On April 16th, Major argued in the House of Commons that Resolution 688 supported his plan and that under the resolution ‘it is clearly the responsibility of the United Nations to protect both helpers and helped. If necessary, the United Nations would have to act on that responsibility and seek from its members whatever assistance, including military assistance, it might need’.83 British Secretary of State for Foreign and Commonwealth Affairs Douglas Hurd similarly argued to the House the next day that the safe havens ‘are to help provide emergency aid, as authorised by Security Council resolution 688’ and that Resolution 688 ‘insists—that’s the word it uses—that the UN operation should be carried out in Iraq. So there’s no violation of Iraqi sovereignty’.84
The American government initially was cool to the idea, with President Bush arguing that ‘when you have a refugee problem of this enormous consequence, then that comes under the heading of United Nations business’ while Secretary of Defence Richard Cheney suggested ‘the US might want to go back to the UN for “some kind of mechanism or process that would provide an area where they (the Kurds) would be safe”’.85 However, the US position quickly shifted. This was driven by three factors. The first was the significant media coverage of the Kurdish exodus and the fact that the ‘media pinned on Bush, Major, and other Western leaders much of the responsibility for this situation, in view of their encouragement to the Iraqi people to overthrow the regime’.86 One report suggested that Bush’s vacillation on the issue had caused his approval rating to fall by 12 points.87 The second was that upon his return from a trip to the border region, US Secretary of State James Baker argued directly to the President that ‘the international community has to respond quickly and effectively’.88 In his memoirs, Baker would write that ‘the desperation and deprivation’ of what he had seen made him ‘determined to see us do all we could do to prevent this becoming even more of a humanitarian catastrophe than it already was’ and galvanised him into pressing for the safe areas.89 Finally, Bush also faced pressure from the US Congress, with the Senate passing a resolution on 11 April 1991 which called on him to immediately press the Security Council ‘to adopt effective measures to assist Iraqi refugees as set forth in Resolution 688 and to enforce, pursuant to Chapter VII of the United Nations Charter, the demand in Resolution 688 that Iraq end its repression of the Iraqi civilian population’.90
Following this shift, Bush was quick to argue both publicly and privately that he felt Resolution 688 provided sufficient justification to create the safe area. In a phone call with French President Francois Mitterrand on April 11th, President Bush noted not only that the US government felt they ‘had the authority to do so under UNSC Resolution 688’ but that he had also ‘talked to the Secretary General on this, and I think he agrees with me’.91 Based on this, Bush argued to Mitterrand that ‘it is our feeling that we should accept them de facto and not try to formalize them further. My worry about formalizing them is that we will have difficulties with some members of the Security Council, like China or the Soviet Union. I would hate to go and get vetoed’.92 For his part, President Mitterrand continued to echo the French position that the international community had both a moral duty and a political requirement ‘to protect these people against this man who was militarily defeated. I think the safe haven concept is a good one. We are not detracting from the sovereignty of the state involved by removing territories. Very simply, we are acting because we must forestall genocide’.93
Bush made such arguments publicly when he announced that the US military would lead an operation to establish encampments in northern Iraq. He noted that this was consistent with resolution 688 and that: ‘I want to underscore that all that we are doing is motivated by humanitarian concerns. We continue to expect the Government of Iraq not to interfere in any way with this latest relief effort’. In response to a question, he went on to note that ‘some might argue that this is an intervention into the internal affairs of Iraq. But I think the humanitarian concern, the refugee concern is so overwhelming that there will be a lot of understanding about this’.94
The governments of the Soviet Union and China also expressed reservations to the plan, with Chinese Ambassador Li Daoyu stating the move ‘went beyond the council’s mandate in the Gulf Crisis…’96 However, they took no further action, and the general view is that the safe area ‘was generally condoned if not explicitly authorized’.97
whether a Western military presence could be established under UN authority without Iraqi consent, he replied ‘No. No. No. We have to be in touch first of all with the Iraqis’. At the same time, though, he seemed prepared to look the other way if Western forces chose to act alone, so long as the UN was not asked to use force: ‘… if the countries involved do not require the United Nations flag, then that is quite different’.95
Operation Provide Comfort thus began with three missions: to provide emergency relief, create a security zone in northern Iraq, and refugee resettlement. It was designed if necessary to fight the Iraqi military, and as such included some 23,242 troops from the United States and twelve other countries, while 30 states contributed financial to the mission.98 As the American commanding officer, Lieutenant General John Shalikashvili, later noted, ‘I had no idea what [the Iraqi government] reaction would be to our insistence that we move coalition forces into Zakho [Iraq] and into the valley … And when we asked [the Iraqi forces] to withdraw to an arc of some 30 kilometres around Zakho, I was very encouraged when they did’.99 Both the American and British governments made clear that their goals were limited—forces would be mobilised ‘only insofar as they would contribute to humanitarian efforts … [the haven] would only have to be large enough to provide a temporary shelter until the displaced felt secure enough to go home’.100 The deployment itself lasted less than two months, and while the military forces had repeated confrontations with Iraqi military personnel and secret police, none escalated to violence.101
In addition, the announcement of the operation led the Iraqi government to sign a Memorandum of Understanding with the UN on 18 April which permitted UN humanitarian agencies to operate throughout the country, with Saddam Hussein claiming ‘that his acceptance of these UN terms obviated the need for military enforcement of 688’.102 On June 7th, the force turned responsibility for relief operations over to UNHCR with a security presence provided by the UN’s own guards.103 They were supported by a residual US military force in Turkey and by the maintenance of a no fly zone over northern Iraq, dubbed Operation Provide Comfort II, until 1996.104 At the same time, Kurdish refugees took the allied security presence as a positive indicator with 600,000 refugees returning by early June.105
The Legacies of the Northern Iraq Safe Area
The safe area was widely seen as a success. But it created four problematic legacies which would have lasting consequences. The first was simply that the northern Iraq case was quite unique. As Contran notes, it succeeded due to a range of factors, including the weakness of the Iraqi military, a significant coalition force remaining in the region, and strong cooperation from the Kurdish population.106 Not only did the northern Iraq safe area have a credible international military presence to secure it, but the coalition spent $2.7 billion to support the 100 day operation.107
Second, northern Iraq was not approved by the UN Security Council, though as we have seen the key proponents of the safe area argued that Security Council Resolution 688 provided them with the authority to do so. The following year, the Security Council explicitly authorised that the Bosnian town of Srebrenica be established as ‘a safe area which should be free from any armed attack or any other hostile action’,108 a concept then extended to Sarajevo, Tuzla, and four other Bosnian towns.109 However, the UN Protection Force (UNPROFOR) in Bosnian had neither a clear protection mandate nor the necessary troops to police these seven zones.110 The result, Posen suggests, was that within the safe areas, there were 5,000 troops in Sarajevo, 3,000 in Tuzla, and only 500 each in the other four safe areas.111 Rather than following the northern Iraq model with a significant military presence, consequently, the Srebrenica safe area relied on the uncertain legitimacy of the UN Security Council. This resulted in its fall in July 1995 with the genocidal deaths of 8,372 Bosnian Muslim men and boys.
Perhaps not surprisingly, following these failures, the UN Security Council reverted to a more traditional model of peacekeeping and did not establish another peacekeeping mission under Chapter VII until the UN Mission in Kosovo and the UN Transitional Administration in East Timor, both in 1999.114 The next year, the Brahimi Report on Peace Operations solved the quandary of how peacekeeping operations should seek to protect civilians by arguing that peacekeepers ‘who witness violence against civilians should be presumed to be authorized to stop it, within their means, in support of basic United Nations principles’.115 This alternative, framed as the Protection of Civilians (PoC) agenda, was endorsed by the Security Council in Resolution 1296.116
…existing peace-keeping operations were given additional mandates that required the use of force and therefore could not be combined with existing mandates requiring the consent of the parties, impartiality and the non-use of force. It was also not possible for them to be executed without much stronger military capabilities than had been made available, as is the case in the former Yugoslavia. In reality, nothing is more dangerous for a peace-keeping operation than to ask it to use force when its existing composition, armament, logistic support and deployment deny it the capacity to do so.113
The final problematic legacy was that the northern Iraq safe area—in providing an ad hoc response to the IDP problem created by Turkey’s closure of its borders—effectively closed off a more traditional form of response: providing asylum to the Kurds. In a telephone call between President Bush and German Chancellor Helmut Kohl on 16th April 1991, Kohl wondered if ‘we might launch a joint initiative to help them find homes’. Bush replied directly that ‘we agree Helmut. We want them to go home’,117 dismissing either deliberately or accidently Kohl’s proposal for resettlement. The creation of the safe area reflected the unwillingness of the US and other states to challenge Turkey’s border closures, or to consider other alternatives. As then UN High Commission for Refugees Sadako Ogata noted in a 1996 speech: ‘it was the converging interests of the Coalition Forces … They recognized Turkey’s security concerns not to allow the inflow of Kurdish refugees’.118 UNHCR, Long has argued, had little ability to influence these events with the Office facing ‘a choice between continuing to resist the plans already determined upon by its key donor states—who were also the key military actors—in the interests of sticking to its mandate principles, and participating reluctantly in a major humanitarian operation…’119 Alongside the protection it offered the concept of a safe area also reinforced the containment agenda being driven by developed states by establishing a logic through which would-be refugees in effect lost the option to seek asylum.120
And yet, very soon thereafter, this was the logic followed both by states and by UNHCR itself. In 1992, then US Secretary of State Lawrence Eagleburger argued that with respect to the Bosnian War, ‘[w]e must also funnel humanitarian assistance to hundreds of thousands more who are besieged inside Bosnia, so that they do not become the next wave of refugees’.122 In 1993, as UNHCR became more involved in providing assistance into that conflict, the Office argued that ‘the objective of prevention is not to obstruct escape from danger or from an intolerable situation, but to make flight unnecessary by removing or alleviating the conditions that force people to flee. Defending the right to remain does not in any way negate the right to seek and to enjoy asylum’.123 It was this logic which lead to the creation, and subsequent collapse, of the Srebrenica safe area.
In-country protection, e.g., through the establishment of internationally guaranteed safe zones, however, needs to be weighed against the rights of individuals to leave their own country, to seek and enjoy asylum or return on a voluntary basis, and not be compelled to remain in a territory where life, liberty, or physical integrity is threatened.121
This is a story of normative contingency in two ways. First, without these broader normative trends, the northern Iraq safe area likely would not have occurred. After all, even with a form of safe area established in international humanitarian law, they have been seldom used in practice. But this safe area also reflects both the potential and the problems associated with the concept. In creating the safe area, the states responsible for it linked together civilian protection, human rights, and sovereignty in a novel way. This linkage did reflect the deliberations of the Security Council around Resolution 688, even if that resolution stopped well short of authorising military intervention, and also reflected the Council’s wrestling with how to respond to the transboundary effects and refugee flows created by internal repression.
But, second, re-examining the northern Iraq safe area also demonstrates the lasting effects it has had on how we understand civilian protection. There is a clear trajectory from the safe area concept to the creation of the Protection of Civilians agenda in UN peace operations.124 And, in the debates around northern Iraq we also see the start of deliberations which would lead, ten years later, to the first developments of the Responsibility to Protect doctrine. The creation of the safe area may have reflected the confluence of three sets of normative understandings with a unique set of circumstances surrounding the Gulf War; yet it has clearly had a lasting impact both for good and ill.