Having concluded our chronological examination of the concepts of responsibility in relation to violence, we can now take a closer look at the concrete practices and decisions implemented in the field. The purpose of this section is to provide an insight into these, principally through reference to the three appended case studies, but also through the use of elements drawn from other situations, past and present, that have featured in the policy document review and in recent meetings and discussions at head office.

However, it is only through reading the case studies in detail that we can appreciate the range of views, the complexities of the decisions that had to be made, and the ways in which they were integrated into widely varying contexts. By attempting to marshal this variety into a set of more general remarks, we inevitably run the risk of rigidifying practices which, while deployed in relation to the framework described at length in the previous sections, were also reactions to concrete situations. Practices, in short, derived from the reference framework but which also contributed to its formation and sometimes exceeded it.

In order to reduce the risk of this mechanistic bias (which might lead to the impression that there is always a specific response to a given situation), we will not venture too far in the direction of why – our knowledge of the motives or perceptions at the root of actions will always be fragmentary. Our point of departure is rather what: ‘what is it that we do?’. It concerns the concrete action we implement in various situations, towards various groups, in link with threats that have to be identified.

BEING PRESENT

As we have seen, there was a distinct evolution in the discourse of presence between the 1980s and the present day. In the 1980’s, presence was a responsibility and an act in itself, and encapsulated the commitment of the doctor-witness – an expression of solidarity, a form of militant support, and the pre-requisite for being a troublesome witness. It ceased to incarnate this ‘manifesto’ in the 1990’s as conflicts became more accessible to foreign actors. In the middle of that decade, the events at Kibeho and Srebrenica led to a major reappraisal of the dissuasive virtues of presence. But presence was still viewed both as a means of providing populations with moral support and as a signal to the authors of violence that they were exposed to a potentially troublesome external focus. References to presence as a value in itself became rarer as the figure of the aid worker emerged; as time went on, “presence” was increasingly used as a synonym for “medical action”.

But beyond the truism that there is no concrete action without presence, the idea of “being present” nonetheless remains charged with a particular meaning. The desire to work “as closely to the violence as possible”, whether in war or in conditions of insecurity, is a 40 remarkable echo of the earlier desire to be “close to populations”. And as MSF continues to struggle to establish a medical presence in the most unstable areas of countries afflicted by war, it certainly does not believe that it is doing any harm: “We bring a little security to the people we work with. If this was not so, every justification for humanitarian aid would be shattered” (interview with R. Brauman). Moreover, it should be borne in mind that during the 1990s, the question of whether to set up programmes in North Korea, the prisons in Rwanda or the displaced persons camps in Burundi and Mozambique was in every instance answered positively – the action would go ahead. The more recent project in Burma was governed by the same logic: the question of doing more harm than good was not settled in advance but on site: the priority was to establish the presence.

On a more concrete level, the field team’s reaction to the announcement that the Nyala displaced persons camp in Darfur was to be closed in early 2004 sheds further light on the meaning attributed to presence. Having argued against the closure, the team decided to spend the night before it was to take place in the camp and to remain there throughout the process, in the hope that their physical presence would deter possible abuse. But when working in Mornay several weeks later, the same team began to worry that its presence was in fact jeopardising the inmates despite their claim that it was protecting them from the campaign of destruction raging around them. The team could not divine the intentions of the militias which plagued the area and feared that the sense of security it fostered would turn out to be a trap. MSF personnel are often told that their presence is seen as a form of protection:

“When everybody says ‘you protect us by being here,’ you can’t say, ‘no, no, that’s not true; you shouldn’t believe it’ … You’re obliged to listen to it” (interview with T. Allafort, head of the emergency unit). “In some situations, you can’t deny the impact you make just through being there, through the fact of your presence, even though the effect it has wasn’t what you were aiming for – so it creates responsibilities for you … When you saw how displaced persons in Darfur reacted, when you listened to them, they thought you were making a great contribution to their protection, that it had something to do with it, and that it was important for you to express yourself. They all thought that, from the political leader to the ordinary inhabitant – the ordinary inhabitant said it was a good thing we were there, and not just to hand out supplies; they genuinely believed that we were contributing to their security, they had been subjected to considerable violence, they were very happy to see MSF teams … and we said, ‘We won’t keep quiet, count on us’. But we also said, ‘You shouldn’t have any illusions; the presence of humanitarian organizations has never stopped massacres occurring’” (interview with J-H Bradol). “Was our presence effective? We harboured that fantasy, the population harboured that fantasy” (former head of mission, Colombia).

Thus the relatively positive impact of our presence, whether overvalued, recorded or in some cases simply imagined, is not an idea that volunteers alone have developed. This idea, most MSF personnel are certainly aware of its fragility.

OBTAINING ACCESS

For MSF, the primary issue is therefore one of access to suffering populations, as the constant and intense efforts to obtain it (access to a country or most often to a particular area) clearly demonstrateIn the 1980s, the de facto restrictions on access led to the acceptance that some areas were beyond our reach.. The persistent attempts to reach the combat zones in Sri Lanka in 2006 recalled the protracted efforts (lasting over a year) to establish a presence in the island’s northern sector in the 1990s. Exploratory missions have multiplied in every one of the countries in which we work; mobile care centres were introduced in order to reach villagers hiding in the bush in the Central African Republic (CAR), just as they had been for the Rwandan refugees hiding in the forests of Zaire. The denial of access to combat zones and areas marked by violence is a major concern of the members of MSF, who see it as encouraging the development of violence. When access was repeatedly denied, as in the second Chechen war, Liberia (2003), Sri Lanka (2006) and the CAR (2007), MSF regularly resorted to public statements. We did even more so in situations where we lost access to the populations we had been assisting, as in eastern Zaire (during the hunting down of the Rwandan refugees in 1996- 97), Darfur (the closure of Nyala in 2004) and Liberia (when populations were “trapped by the fighting” in 2003). Besides calling for access, every statement expressed fears concerning current or feared violence.

But access is never ‘all or nothing’; in many instances a choice has to be made between what has already been acquired and what is being sought, between remaining silent and speaking out. Thus in Darfur at the end of 2003 and the beginning of 2004, access was possible but precarious: there were few expatriates working in the area and visas and travel permits were hard to obtain. However, MSF refrained from commenting on the violence and impediments to access in the hope of maintaining and consolidating its presence (the access already acquired). Given the context it did not believe its ability to negotiate was sufficiently strong. From the establishment of a presence (in the sense of medical action in this instance), we turn to the concrete practices related to violence. These are sometimes visible and sometimes veiled; they may be marginal, exceptional or current; some have been institutionalised and some have not; they may be deployed for the benefit of individuals or of groups. The logic behind them varies according to the nature of the action, and thus may be analysed on a variety of levels. In an overview necessarily marked by contrasts, we shall attempt to extract some of the more consistent elements.

IMPROVING THE SAFETY OF CARE

Despite their different contexts and periods, the case studies are striking for the common theme they present – concern for the fate of people suffering violence is expressed above all through the forms the action takes (the implementation of aid), i.e. through practices which we do not usually call “protection”, but which all derive from the concern to prevent exposure to violence.

This preoccupation is expressed in widespread efforts to encourage respect for the neutrality of the care space. As the “no weapons” signs on the doors of MSF hospitals and clinics indicate, teams are often engaged in a daily struggle to prevent armed men intruding into their work spaces. The issue was pushed to extremes in Rwanda, where the preservation of the hospital’s immunity was a major achievement: the team was able to derive some satisfaction from the fact that at least “no atrocity had been committed in the hospital” (Board meeting, May 1994). Similar efforts have been made in more recent, less acute situations. In Colombia, for example, the teams attempted to ‘demilitarise’ the areas visited by their mobile clinics by persuading the militias to keep away. Through our energetic reactions, we can make such actors aware that we 42 will not tolerate violations of the care space. For example in Somalia in 1993, when MSF called for an inquiry into the bombing of a hospital in which its teams were working. Likewise, the intrusion of large numbers of armed men looking for ‘suspects’ led to formal protests (Rutshuru, DRC, 2005) and the temporary closure of a clinic (Bentiu, Sudan, 2003). However, efforts to create a ‘non-militarised’ space may obscure the fact that there is more to security than the banning of weapons: the recent discovery that women had been raped during the night in the hospital in Adre (Chad) should remind us of that. Moreover, the constant team effort to maintain a weapons-free space may simply become a reflex action which loses sight of the desired outcome (the safety of staff and patients). This issue resurfaced quite recently, when a health minister arrived at an MSF hospital to visit his seriously injured chauffeur, but was refused access because he was accompanied by bodyguards.

At different levels, each of the above examples illustrates the issue of maintaining a level of security which is not negotiable to ensure that medical action does not itself become a source of jeopardy – this is a required minimum that rules of conduct, however, cannot fully guarantee.

The same logic governs the treatment of victims of violence. For several years now, practices and ‘ways of doing things’, often linked to the requirements of quality and access, have been designed to ensure that the actual delivery of care does not add to the violence the patient has already sufferedThis intermingling of issues (quality, access, vulnerability) shows that although MSF denies having a protective role, it has most certainly not abandoned the issues surrounding people’s security, but rather approaches them at the level of the patient..

In projects involving the treatment of victims of sexual violence (VSV), confidentiality gradually came to be seen as a vital factor in gaining access to victims. These people were not sick – they were the victims of a singular, invisible and stigmatising form of violence, and any publicity of the prejudice they had endured was likely to expose them to further danger. This was a new challenge for field workers, who had long been accustomed to working in hospitals and refugee camps where ‘quantitative’ treatment often threatened the confidentiality of interviews. A number of examples, some of them recent, illustrate the problems arising from a twofold requirement – the publicizing of the offer of care and the confidentiality of the care itself. One team, for example, finding that few VSV were turning up for consultations, asked the “elders” to draw up lists of rape victims. The need for confidentiality is now drummed into field teams and has gradually gained acceptance as a vital component of such activities, but it is still very often experienced as an operational headacheThe problem was acknowledged with the creation of a post specifically dedicated to victims of sexual violence in 2007, long after other sections had taken this step. 50. Certification arose from a context of sexual violence but is just as relevant for the wounded, although the issue had not arisen in 30 years of treating such patients.. Having addressed this problem, MSF encountered further difficulties arising from interaction with “protection” agencies which wanted to produce reports on the incidence of rape: should it contribute to the documentation of rape and (like many agencies) ‘bear witness’, or should it focus on preserving confidentiality? The case studies show that whenever these two demands were perceived as incompatible, we opted for the second course, challenging the comparison with sanctioned “protection” activities and disputing in point of fact the claim that they improved personal safety. This position is clearly connected to the personalisation of the carer-patient relationship arising from the treatment of VSV.

The problem of confidentiality extended to the certificates made available to victims of violence.50 Head office has made great efforts to convince doctors that this medico-legal practice formed part of their responsibilities, yet the certificate is also a potentially dangerous document for the patient, given the information it contains. It was thus stressed that victims should be free to choose whether to accept it when it was offered, accept it at a later date, or simply refuse the offer. Once again, we see a paradox arising from the fact that certification is currently associated with “protection”, even at MSF – indeed it is to my knowledge the only legitimate use of the term within the organization at present. Here we find ourselves at the heart of the difficulties connected to the term’s polysemous nature: the certificate is certainly a form of protection in the strict legal sense, i.e. the document should enable the victim to exercise her rights at a later date by filing a formal complaint. In this sense, there is precisely no specific reason at present why it should contribute to protection in its more common sense, i. e. physical security (the insistence on the possible dangers linked to the possession of such a document probably reduces the likelihood that teams will regard it as a ‘protective’ measure). Moreover, a document linked to an individual’s legal protection does not in itself exhaust the actions MSF may deem possible or desirable at a wider, more collective level. The internal debates sometimes reflect this ambiguity, with the reference to certification forming the basis for the idea that MSF ‘does what it has to do’ in terms of “protection” (see the North Kivu case study).

REDUCING EXPOSURE TO RISK

From the pressing need to secure the confines of the hospital (as a space) and of care (as a moment), we turn to the reduction of exposure to risk, which is in most cases a matter of adapting the methods of aid implementation whenever specific threats are identified.

As Rwandan refugees were being hunted down in 1996-97, two different ‘explo’ teams came to the same conclusion: their efforts to reach the refugees in the forest did not appear to “improve … their chances of survival”, and indeed increased the risk by enabling their killers to locate them. Both teams adapted their approach: they stopped the search for refugees, restricted the offer of care to sites near major roads and stopped collecting information from villagers living in the depths of the forest, having realised that this could also expose people to danger. A similar adjustment concerning the location of aid delivery took place when the Mornay project in Darfur opened in 2004, after the team had noticed that the edges of the wadi, some distance from the camp itself, were dangerous areas for displaced persons. The decision to install a water distribution point closer to the camp was directly linked to the identification of this particular threatSee the Darfur case study for a comparison of water provision and the gathering of fodder.. There are many instances of such adaptations, all made after examining the possibility that the action itself might increase exposure to the violence that had been identified. In many cases they involve the rearrangement of or abstention from certain activities. For example, in the health centres currently provisioned by MSF in the DRC, distributions are halted when the risk of pillage seems especially high. This was also the case in the displaced persons camps in Liberia in 2002: “Our strategy is to minimise the exposure and targeting of displaced persons as much as possible by concentrating on medical treatment and water supply (we avoid aid distributions, which might trigger systematic pillaging)” (Board meeting, February 2002). In 2001, the organization decided against providing assistance in the refugee camps on the Guinea-Liberia border because they were too close to the combat zone (Board meeting, 2 March 2001)MSF pleaded for the refugees to be relocated to a site some distance from the border. It is the refugees who refused, preferring to stay in a danger zone which was nonetheless closer to home.. Such practices therefore fall within the province of aid action.

Similarly, the realisation that letting people leave the hospital after dark might actually increase their exposure to physical harm led to practical adjustments, as at Rutshuru (DRC). However, it was not immediately apparent to everyone that additional measures were required. On several occasions, rape victims who had come in for a consultation had been allowed to leave the hospital at nightfall; the increased danger on the roads at this hour was not spontaneously linked to MSF action on the issue of exposure. The comments of outside observers and those with a greater awareness of the problem probably sensitised the team to the dangers inherent in this element of the action, which could in fact be ‘turned’ against the victim. Once the coordinators had become aware of the risk, they established a rule that anyone arriving for a consultation should be kept in hospital overnight if dusk was approaching. Once again, the personalisation of care and the urgent need to examine the impact of our own actions combined to make the possibility of ‘damage’, of the patient being exposed to further violence, less and less acceptable. The role MSF itself might play in exposing people to danger was thus brought into sharper focus.

However, the possibility that aid itself can become a source of danger does not imply that there is such a thing as ‘pure’ aid or ‘turning against people’ aid. Such terms are employed simply to indicate how growing awareness of the issue leads to the risks, benefits and adverse effects of a mission being gauged differently by those at headquarters and those working in the field.

This observation also applies to the extreme situation in which aid is used by armed actors to commit acts of violence, i.e. when aid becomes an active rather than a passive accomplice. As we have seen, MSF (and particularly MSF-France) considers this a quantum leap since the Ethiopian episode, given that its responsibility is directly at stake here. When it became apparent that aid was “serving the executioners” (Ethiopia in 1985 and Zaire in 1997), programme leaders at MSF-France called for radical action, believing that the need to “stop it” and denounce it was so urgent that it overrode all other considerations. Nevertheless, an examination of the hunting down of Rwandan refugees in eastern Zaire in 1996-97 leaves us in no doubt that the image of aid as bait was constructed over time, as information came in and awareness grew; it was not immediately apparent to the teams in the field or to those in the offices of the various sections involved. Retrospective references to “bait” or “hunting down” should not disguise the fact that at the time, the successive ‘explo’ teams were primarily concerned with situations of violence and emergencies. They slowly built up a picture of what was happening – the use of aid to locate and then kill refugees – as they attempted to collect information and confirm suspicions; their observations combined to create a significant canvas. But even then, their observations were not enough to trigger an automatic withdrawal or denunciation, for at that point the organization had an exceptionally severe operational emergency on its hands. Moreover, the head office concerned was reluctant to denounce incidents that were still perceived as marginal. Several months later, they were integrated into the broader picture of an elimination strategy, and formally denounced by MSF-France, against the advice of the other sections.

We are therefore presented with a range of elements – the mutual exclusion between operational intensity and a tight focus on violence; the link in the chain represented by the head office and the role it plays in formalising a situation; the role played by description before the action begins; and the ‘operational culture’ peculiar to each section. All these elements determined the emergence or non-emergence of the protection issue in this episode. In short, while the use of aid as bait may have seemed particularly clear to MSF in terms of what was at stake operationally – it was an “intolerable” situation, i.e. the limit for an organization whose priority is to avoid harm – it was certainly not apparent as such from the outset. As always, the response given to a situation cannot be dissociated from the way the reality is collectively depicted.

Our examination of the practices designed to reduce the possibility of exposing staff and patients to violence reveals the existence of a “do no harm” policy which is directly related to the responsibilities of what we do. There is a legitimate field of action, but its limits undergo constant revision – as we continue with the ‘traditional’ practices, we introduce and systematise new practices. We shall now turn from this solid foundation and investigate the practices implemented in the more fluid space that lies beyond the realm of “avoiding harm”.

SHIELDING PEOPLE FROM VIOLENCE

When the genocide in Rwanda was at its height (April-June 1994), the teams working in Kigali hospital decided not to discharge patients at the conclusion of their treatment – these people had been admitted for legitimate medical reasons and were retained with the specific aim of physically protecting them. The hospital became a structure from which no one emerged, and expanded day by day as MSF commandeered neighbouring buildings to turn them into additional wards and increase the number of beds. In such drastic conditions, the teams were gambling that the medical argument would counterbalance the determination of the Hutu militias to exterminate the Tutsi populationWe can assume that the teams did not consider shielding people who were not patients, for this would have negated the medical argument. . In this instance, as mentioned earlier, “no atrocity was committed in the hospital”.

Field teams are currently taking similar initiatives in less extreme circumstances, usually in order to shield a patient from the threats he or she faces as an individual. These practices have a low profile and seldom appear in activity reports unless it is to highlight pressing issues – security problems which jeopardize the safety of MSF personnel, or problems arising from case management, for example. On several occasions, the team in Port-au-Prince (Haiti), protected injured individuals who would have been at risk had they left the hospital. Whether a child who had suffered violence from family members or a youth hunted by other youths, they were kept – indeed hidden – in the building pending a solution to the predicament. These “protection cases” then became management problems in that they were blocking an increasing number of beds (in the same way as “social” cases – paraplegics, people who have become dependent and cannot be looked after by their families, etc.). The practical difficulties arising from bed blocking played a large part in the creation of the petit comité violence, an initiative which institutionalised the search for concrete solutions – procedures of removal or transfer to an appropriate environmentThe petit comité violence (PCV) is the name given to the sub-group (head of mission, social assistant, psychologist, medical coordinator, …) which meets to discuss complex cases that require a collective input in the search for solutions. The caseload is thus composed of protection cases and social cases which raise similar issues (the identification of an appropriate treatment structure, the provision of financial support, etc.). However, the inclusion of the word violence in the group’s title is significant, for it is a direct reference to the legitimate ‘core’ of the programme, although the cases are usually distinguished by an inextricable mesh of economic, psychological, legal and security problems.. The “social” activities at Matare in Kenya also involve working with “protection cases” such as victims of domestic violence. In the DRC, field workers refer to the assistance provided to individual cases as ad hoc intervention, but in fact a budget is now envisaged to provide support for people who would face further violence if they returned to their homes (if the victim knows of a safer but more distant location, for example, the money will facilitate the move). Data collection now includes specific questions about persistent threats the patient may face (“has she a safe place to return to?”). These measures illustrate the degree of concern felt at head office over the possibility of returning a patient to a place of danger. The development may be seen as a consequence of the personalisation of care, which entails the extension of responsibilities towards the patient. It has led MSF to consider social and protection issues, to try and intervene one step higher in the chain of causality; for besides treating the consequences of violence, we are also trying to prevent further exposure to itIt is interesting to note that at a time when many “violence and social exclusion” projects have been closed, the ‘highquality’ treatment of victims of violence is in fact forcing MSF to return to the issues involved in such programmes – the complexity and entanglement of the dynamics at work, the chronic nature of the problems (poverty, the fragmentation of social and familial links, the many forms of violence, impunity, etc) – and is testing its ability to arrive at appropriate operational responses..

MSF also has some experience of attempting to shield a large group of people from the threat of violence. We recall that the Srebrenica team took patients with it when it evacuated the hospital as the enclave fell. The MSF-Holland team working in the Kivus did the same for children at its feeding centre when the crisis in eastern Zaire broke out in late 1996. The Dutch team had done something similar a few months earlier in the same region. In that instance the individuals concerned were not patients but Tutsi civilians receiving assistance in one of the villages. After some of them were massacred and UNHCR failed to respond to the team’s warning, head office supported the decision to begin evacuating a group of survivors in trucks.

To our knowledge, only one attempt has been made to evacuate persons under threat in recent years. These were North Korean refugees who were either sick or “of whom it was reasonable to assume that [they] faced death”. MSF helped them to “cross the border illegally … to reach South Korea and seek asylum” (AR, 2002-2003). These “unusual” and delicate decisions (some operations failed and the refugees were intercepted) were later described as deriving from the confrontation with concrete situations, with an entreaty“We did not choose to take the initiative in these operations … we chose not to say no when individuals, families or groups appealed to us for help” (AR 2002-03). . It should also be noted that once again the people concerned were from the project’s target group (refugees); in other words, they were linked to us by the fact of assistance. Moreover, the initiatives were probably encouraged by the atypical nature of the project: MSF was operating outside the ‘traditional’ framework from the outset. Collective evacuations of this type have always been regarded as exceptional measures.

The responsibilities linking MSF to civilians in general (those who are not patients or in receipt of assistance) are certainly more tenuous. Nonetheless, personnel have willingly attempted to shield people from violence when a situation has reached crisis point. Moreover, when people seek refuge in MSF clinics in order to escape the violence raging outside, teams are even more likely to act spontaneously and give them shelter. This happened when fighting had intensified in Rutshuru in 2005 (the expatriate team had at one point evacuated the area), and in Bentiu, when militia fighters were combing the town for recruits. In the Bentiu case, the ‘passive’ provision of shelter led to tense negotiations with the militia leader, who demanded that the civilians be handed over. The team steadfastly refused to do so, but armed men eventually forced their way into the hospital. The team made no attempt to stop them and could only protest by pointing to the violation of the care space (see above). In this case as in the North Korean civilians’, the protection they received was pretty fragile. But given the choice between sending people back into greater danger and holding on to them as long as possible, the team unhesitatingly gambled on the second option – although the wager did not exclude a rough and ready assessment of the risksA calculation which, on the one hand, includes the risks for the people concerned (the nature of the threat – certain death, ill treatment, forced enlistment – posed by the combatants), and, on the other, the risks to team members should they place themselves between combatants and the people they are looking for – how determined are the combatants, to what extent do they respect humanitarian workers, how sensitive are they about their public image, etc?.

SPEAKING PUBLICLY AND IN-HOUSE

In North Kivu, as in every unstable area in which MSF operates, the field coordinator holds regular meetings with the various armed leaders, the aim being to ensure that MSF teams can travel in safety. He also personally delivers the quarterly activity report. Several coordinators reported that they often took advantage of such visits to indicate to a leader that many victims of violence were originally from the area under his control. In effect, they were telling him that they knew what was going on, that others were watching him. It was also a way to avoid becoming “auxiliaries” of the violence by maintaining a silence which could be taken for acquiescence. In a context such as North Kivu, this finely modulated form of communication is as widespread as it is discreet; indeed there is no reason why it should be discussed or even referred to in the messages between field and head office. The room for manoeuvre here is narrow: it is necessary to “set a certain tension” when dealing with the authors of violence, but also to avoid compromising fragile interaction by taking the high moral ground, for example.

To turn to another time, place and context, tension with the Sudanese government over Darfur was the last thing required in the early months of 2004. Head office decided to launch a (non public) lobbying strategy; it briefed journalists and sent a coordinator who had recently returned from the area on a tour of the United States, in the hope of pushing the international community into “adopting a firm political stance at the earliest opportunity”, including on “violence against civilians.”Briefing quoted in C. Danet, S. Delaunay, E. Depoortere, F. Weissman, A Critique of MSF-France Operations in Darfur (Sudan) – October 2003-October 2004, MSF / Cahiers du CRASH, 2007, p.111. This type of indirect language, addressed to third parties rather than to the authors of the violence themselves, is as common today as it was in former year. The practice is specifically linked to the goal of “protection” in the broader sense: it strives to make an impact on a situation of massive and largely ignored violence (and on the massive needs thereby created) by publicity on the one hand and appeals to influential political actors on the other.

The use of public statements to achieve the same goal is more difficult to analyse. As our chronological approach has shown, changes in the content of the press releases clearly reflect the gradual decoupling of témoignage from the goal of “protection”, as exemplified by the withdrawal from calls for armed intervention and the “shouldering of responsibilities”, the decreasing reliance on references to IHL and respect for civilians, and the foregrounding of medical data and epidemiological realities. Press releases now usually focus on aspects of the aid we provide, and may thus cover topics such as obstacles, access, insecurity and diversion.

However, this new register does not mean that MSF has abandoned references to the fact of violence: as we saw earlier, it still features in the numerous press releases covering impediment to our action. On rare occasions it constitutes the core message, notably when it is exceptionally severe (the bombing or massacre of civilians, for example). The scale and/or intentional nature of violence may be taken as one of the criteria governing our perception of its gravity.

Nor does it mean that MSF actors have lost the desire to discuss the violence a population may be suffering, to bring it to the attention of the wider world; it is one motive among others when advocating a public position at a given moment. But, firstly, the desire to speak out has been reformulated in terms of the aid worker. “Access to care” in Somalia is a useful example: whereas twenty years ago we would have “testified to” the disastrous consequences of the violence and chaos, we now express concern over the “population’s lack of access to care” resulting from the insecurity (see above). And our current calls for the parties to the conflict to “enable the sick to access health facilities” echo the former emphasis on “respecting civilians”. The vocabulary may have changed, but our aim is still to draw attention to the plight of an entire population rather than issue a call for something specific which may contain a passing reference to the sick.

Secondly, the question ‘what for’ is asked more systematically to those who express a wish to communicate on a specific topic as is clear from the emergence and abandonment (perhaps provisional) of the idea of a public position on Iraq in late summer 2007See the Board discussions (31 August 2007), the operations meeting (4 September 2007) and the email exchanges between president, desk, communications department and the Foundation (11-13 September 2007).. The debates this stimulated at head office are notable for the ways in which the conceptions and expectations concerning a public statement diverged. Some participants felt the emphasis should be on the impact of violence on civilians, using our patients as an example. Others wanted to raise the broader issue of access to care for wounded Iraqis through an explicit call for a commitment on the part of the international community. One group took the view that given the chaos and the vast amount of media coverage, the obstacles to action constituted the only legitimate basis for an MSF statement. In short, the arguments reflected the range of views concerning the legitimacy of (and opportunity for) a public position on protection issues – exceeding the narrow confines of our own action – in the specific context of IraqIn this respect, it is interesting to note that among other arguments, some came up with a medical comparison (MSF’s challenges to the international community over AIDS and the response to the nutritional crisis in Niger) and suggested that given our work on the margins of Iraq, it might be a ‘solution’ if fully exploited. Others questioned the validity of the comparison, since in Iraq we could not point to specific aims (such as the use of a particular medicine to combat malnutrition or AIDS). The use of medical examples to sustain the idea of a public stance indicates the extent to which such issues have gradually replaced situations of violence as legitimate topics for appealing to or challenging the international community.. Eventually, access for our patients (and therefore the principal issue linked to our action) having improved, the plan to issue a statement was deferred.

The tension between the elements at the heart of such debates continues to influence the decision-making process whenever the possibility of a public statement on the vulnerable populations we assist is considered. These elements include the gravity of the situation (the nature and gravity of the violence and its consequences); the degree of operational ability (are we able, are we allowed to deal with the consequences?); the amount of media coverage; the degree to which our field of action matches the issues we have identified; the quality of the factual and other information at our disposal; the nature of external power relations and of those between sections, etc. Of equal importance (but absent from the Iraqi example, since we work at some distance from the theatre of violence) is the assessment of risks affecting our operational ability and security. The weight accorded to each element during the decision-making process has probably changed, resulting in the evolution of the public statement outlined above.

The elements of the debates remain the same – as remains the uncertainty as to when and how we should publicly express our views on the situations we are dealing with. The arguments should remind us that despite the developments discussed earlier, speaking out is a topic which, at both institutional and individual levels, exemplifies the tensions within MSF. These tensions highlight the ambivalence surrounding the issues of raising public awareness and applying pressure – which are sometimes decried as aims exceeding our legitimate framework, or simply as actions which are ill-timed given the context. But at the same time they are defended on the grounds that they prevent us from becoming the “auxiliaries” of the political policies whose consequences we deal with. This ambivalence emerged in an interview I conducted with the president. I asked him to clarify the reasoning behind the public statement on the “intensity of the repression” in Darfur in June 2004:As we have seen, the statement was subsequently described as an attempt to increase the amount of aid (AR 2004-05).

A: “… we were doing ‘advocacy’, as it’s called.”

Q: “What was the aim of this ‘advocacy’? To raise awareness? Generate a debate? Raise the alarm?”

A: “I don’t see it like that. I think that in many situations, it’s not a given when you are trying to work in the field. You have to create a relationship of power so that humanitarian action is respected, so that it is not co-opted. Your approach to your working environment has to be political, and [speaking out] is one of the elements … So it’s all a game in which power relations must be managed, and part of that is talking about what you see. I think you do it to prevent your own action from being co-opted … It’s a consequence of your operational deployment, and should not be avoided.”

Avoid “being co-opted”; do not become “auxiliaries”; let them know that you know; create a “tension” so that that the action will not “lose its meaning”The expression was used in the debate over the release of a film about displaced persons in Darfur in autumn 2007.. These examples of the language used internally all point towards one of the major bases for a statement (whether public or otherwise) on violence: MSF may deal with consequences rather than causes, but it seeks to ensure that the causes are not blotted out.

“Do people who have been subjected to violence have the right to turn to you for care? It is through questions like these that you raise issues of security/protection … The work of an auxiliary begins when there is a major outbreak and you don’t mention it … an outbreak of rape, for example … When you have achieved a proper grasp of your own responsibility and exercise it in a fairly political way, it clearly also has an impact on the protection of populations … Your VSV treatment centre can also serve to stifle the question its existence poses for broader society, if you are not active in your own public positioning on it. Does your work represent a form of compassion which thus enables society to ignore the problems? Or are you prepared, if confronted with a large number of cases, to depict the medical histories – [in wich case] you immediately contribute to a debate in society” (interview with J-H Bradol).

Speaking out is the product of complex decisions and intentions which vary according to time and context. The practice, a subject of permanent controversy, remains deployed in response to situations of violence – whether it denounces (as in the past) or rather describes situations (like today) ; whether this violence is already committed, ongoing or perceived as imminent; whether aimed at groups or individuals. But it is far from certain that whenever we raise this ‘voice’ – the voice of the “expert witness” who describes, who foregrounds “motivated facts” – it is always in the strict defence of the conditions surrounding our provision of aid.According to P. Redfield, MSF is an “expert witness”; it foregrounds “motivated facts … [which] all seek to establish the facts of suffering and, thereby, make a moral claim”. See P. Redfield, ‘A Less Modest Witness. Collective Advocacy and Motivated Truth in a Medical Humanitarian Movement’, The American Ethnologist, vol. 33 no. 1, February 2006At the conclusion of this lengthy itinerary, which has taken us from institutional discourse on the role of MSF and the responsibilities of other actors when faced with violence to the concrete practices deployed by individuals or systematised by the institution in specific situations, we are forced to acknowledge that it would be vain to attempt a “portrait of protection” at MSF today.

Beyond the internal and external developments that have led to the current reluctance to invoke human rights and protection, to the caution with regard to calls for military intervention in the name of protection and to the disinclination to prescribe solutions, call for the assumption of responsibilities or denounce a situation, violence – and therefore concern for the safety in particular of personnel, patients and of those we assist – is still a central feature within our operational framework. To be sure, the wealth of experience MSF has acquired has also burdened it with a permanent sense of anxiety and doubt as to the effects of its action. Our awareness of the “fragility of the goodness”The expression is from Tzvetan Todorov (The fragility of Goodness. Why Bulagria’s Jews Survived The Holocaust, Trans. Arthur Denner. Princeton: Princeton UP, 2001). we may do means that in any given situation our ambitions will be couched in cautious and modest terms; of greatest importance here is the Hippocratic injunction to “do no harm”. The endless series of ‘non-acts’ – refusals, abstentions, withdrawals and strategies to avoid exacerbating existing violence – should be understood in this light. Thus the notion of complicity remains valid, at least in the active sense of blind or involuntary participation in the implementation of disastrous projects. Therefore if this itinerary has demonstrated anything at all, it is the error of thinking in terms of care on the one hand and “protection” on the other, for the true basis of our action is concern for the fate of those we treat. In every case, addressing this concern requires clearsightedness, attention to the context, the ability to interpret complex situations of violence and the ability to identify the “priority victims”, those most at risk.

But as we have seen, views and actions evolved. The refocus on the aid worker’s role involves both refining “protection” practices aimed at individuals, and moving away from practices with a more tenuous connection to aid action. Borrowing an idea from one of the case studies, we might view responsibility in terms of two axes perpendicular to one another. On the vertical axis, representing the ethical dimension (our accountability for our own actions), our responsibilities are rising. Internally, the lower limit is the “intolerable” possibility of active complicity; on this foundation, we have extended the tasks we allocate ourselves. The emphasis on quality led to provide more and more comprehensive care in our domain and includes consideration of the physical threats a patient may face. What is the upper limit here, in a dimension that sometimes appears as one of absolute sovereignty? On the horizontal axis, representing the ‘political’ dimension (our responsibilities in the light of the general situation and our interaction with other bodies), the scope of our role is retracting. There have been many efforts to define the limits of our legitimate field, ranging from revising our role with regard to international legal procedures to banning calls for military intervention and reappraising the practice of speaking out.

However, these efforts do not imply that MSF is no longer concerned with matters beyond its own sphere of action. Even if aid issues were not involved, everyone would believe that wholesale violence would demand a response of some kind. At some stage, the vertical must cross the horizontal: no field in which MSF operates is free from dilemmas, because our field is inevitably interacting with others. While it is necessary, given the variable results of our attempts to slow down or stop violence, or to shield people from it, to focus on doing our own job as well as we can, it would be disastrous if this “acceptance of our limits” – another way of defining the field – became an opportunity to congratulate ourselves on the quality of our work within that perimeter.

This also applies to the emphasis on the treatment of victims of violence, especially in contexts that are increasingly hard to describe as wars. The expansion of responsibility regarding our own action inevitably opens the door to a host of difficulties and searching questions, often aggravated by the commonly used categories of violence – whether “domestic”, “urban”, “social”, or “other”. Questions such as “what should we do?” sometimes reflect the confusion that field teams experience. We cannot escape the pull of the real; there is no ‘right place’ in the chain of causality that would enable us to settle these questions. The long evolution that has established medical care as the core commitment at MSF is undeniably beneficial, but in no way does it solve the questions concerning the interplay between the various levels of causality, the practice of speaking out, and the possibility of action that extends beyond the injunction to ‘do no harm’.