A review of classic anthropological ethnographies provides ample evidence that people who are recognized as qualified mediators usually possess a number of different responsibilities and abilities at once. He or she embodies the offices of asylum granter, mediator, and ritual expert. This specific register of knowledge, skills, competencies and responsibilities enables this person to mediate even in cases of extreme social transgression and deviance. Mediators of this type must be able to effect three processes: they must be able to suspend power differentials between the parties involved during the mediation process, guarantee security and protection for the parties involved, and provide for social/ritual reintegration of the protagonists once a settlement is reached. The mediators’ capacities find expression in larger-scale institutional arrangements and conflict settlement forums of which mediation is just one component.

A careful analysis of complex conflict scenarios shows that some of the supposedly basic conditions for mediation processes – such as the presumption that conflicting parties agree to engage in a mediation process and also agree on the choice of mediator – may be pre-empted by the extended institutional agency of the mediator. Sometimes agents actively impose themselves as mediators in ongoing disputes, especially if the dispute has the potential to involve violence; sometimes they are forced to accept the role of a mediator, even against their own will, at the instigation of one party (usually the accused). Thus some of the basic conditions and certainties that are usually associated with the concept of mediation, such as the voluntary nature of the mediation and consensus-oriented settlement, can become complicated. 

Mediation processes are informed by three interconnected normative logics involving three distinct legal institutions. The first of these three normative logics finds expression in the concept of retaliation, which is a universal principle of human interaction. The second normative logic refers to the temporary protection of the weaker party in the conflict constellation (usually the perpetrator, but sometimes the victim) and often takes the form of the legal institution of asylum. Such institutional arrangements go far beyond the scope of what is regulated in formal legal systems, from the UNDHR to domestic asylum law. The third normative logic is that of post-conflict social reintegration, which the expert/mediator effectuates through ritual forms of purification and social rebalancing.

One may define retaliation as a human disposition to reactive balancing of conflict or other types of situations perceived as unjust. The basic principle of retaliation, however, is more than an automatic reaction to a violent offence. It refers broadly to the full range of reactions to circumstances that are perceived to be deviant or socially transgressive. Such a constellation presupposes two opposed, nominally equal, parties. Understood in this sense, retaliation occurs at all levels of socio-political organization, from individual face-to-face interaction to nation-states and transnational organizations.

As an expression of the principle of reciprocity, retaliatory logics may inform the whole gamut of conflict resolution procedures, from consensual settlement to compensation to violent reprisal and escalation. The question is, how can the great variety of retaliatory reactions be explained? The most fundamental common denominator seems to be the equalizing and balancing quality that is inherent in the principle of reciprocity. Thus retaliatory reaction exhibits two properties. On the one hand it entails a preventive principle, as the threat of retaliation can prevent somebody from committing an inappropriate act directed against another. On the other hand, it implies the right to react in a way that allows an offended party to rebalance a perceived injustice. If the violence prevention mechanism fails and an initial violent act takes place, then reactive violence is acknowledged as legitimate; otherwise, the implicit threat of retaliation would lose its deterrent function. However, the legitimate right to violent retaliation does not necessarily lead to its immediate execution; it can be substituted by compensation. Compensation itself can be subject to negotiations that, under favorable conditions, may take on the form of a mediation process. Institutionalized protection, the second of the three components, comes into play here as a necessary interface that helps create a situation favorable to mediation.

The exercise of legitimate retaliatory violence can impede the initiation of a mediation process and thus the transformation of claims to violence into claims to compensation. Under such circumstances, protecting a perpetrator from the threat of violent retaliation may help initiate mediation. This is where the legal institution of asylum comes into the picture. 

The objects of investigation here are institutionalized and socially recognized forms of protection for individuals irrespective of the reason for asylum. Such institutions are geared toward people who are guilty of an offence. So why institutionalize the protection of perpetrators who are liable to legitimate violence? How can this be justified? In the majority of those social relationships that are an appropriate subject of mediation, the protection of the perpetrator can constitute a decisive precondition. In the absolute majority of documented ethnographic cases, asylum is provided by the mediator; only under very specific circumstances can it be refused. It is a temporary state and is, therefore, associated with a whole variety of possible further actions and developments. The ethnographic record shows that asylum is institutionally interconnected to mediation and the processing of the consequences of the act that led a person to seek asylum. 

A crucial point is that the protection-granting individual inevitably gets involved in the fugitive’s conflict as the mediator. The expert negotiates with the parties involved and helps them reach a compromise, which often involves the transfer of compensation in a face-saving manner in order to balance out the injustice. 

Once an agreement between the parties is achieved and compensation negotiated, the situation prior to the conflict should be reestablished to the extent possible. This entails the social reintegration of the protagonists, especially of the perpetrator/asylum seeker, whose actions may have led to his or her social exclusion. In such cases it becomes necessary to formally terminate the exclusion period. Ritual forms of purification can open the door to such reintegration, although the prior state of affairs may never be completely restored. The deviant act may remain on the offender’s ‘record’, but a return to normality and social routine should be reached. What the ethnographic data show is that the ritual discharge from asylum and social reintegration are among the conflict expert’s duties. While there is a great variety of ritual expressions and performances of exclusion and inclusion, such as the symbolic death of the offender and his resurrection, often in combination with a formal apology, all of these fall to one and the same person. The conflict expert is at the interface of these three institutions. 

The special properties that make an individual an expert in the arts of conflict intervention are often associated with the social ascription of political and economic non-power and marginality to the very same actors. The relevant literature shows that such conflict experts often practice endogamy, which means that they are not involved in inheritance disputes; in agrarian or rural societies they do not possess landed property; they have no access to official positions of power such as political leadership; and they have no means to achieve economic predominance. In many cases, they are bound to an occupation that nobody else is allowed to pursue and have special professional skills, such as being able to communicate with the spirit world or knowing how to work with certain substances (e.g., metals, medicinal plants). Thus they hold an ambivalent position in the literal sense of the word – they are feared and admired; they are excluded and accepted; they are socially marginalized yet are an integral part of society. 

In fact, this type of mediator has been identified in a wide variety of settings, from traditional, acephalous societies to complex state organizations. In a way they are experts because their own legal capacity and their own legal rights are restricted. This obviously self-contradictory social construction seems to ensure the credibility and objectivity of the conflict expert, even if he or she does not necessarily behave in an unbiased manner. He or she can openly express disapproval of the perpetrators’ actions and even reproach them, but there is the expectation that he or she will have no vested interest in the outcome. There is ample evidence that these professional interveners are not necessarily enthusiastic about their role. They confront a dilemma in that they are expected to remain neutral even when they do not sympathize with a perpetrator they are bound to protect. They are expected to invest all they can in the mediation process – including their reputations – even when it would be in their own financial interests to resolve the conflict as quickly as possible and get rid of the uninvited guest. To resolve the situation the mediator may strategically shift solidarity from one party to the other to support a certain argument and make the parties inclined to accept compromises. The mediator’s principal service to the parties is to find an acceptable interpretation of the disputed event so that each side may save face. The fact that they embody these various qualities – asylum provider, mediator, ritual expert – as well as their ambivalent social status all seem to support them in their production of acceptable narratives. 

A crucial challenge to the mediator is to create circumstances that enable such negotiations. One prerequisite is the balancing of power differentials between the parties. Asylum can prevent the exercise of such power differentials. For example, a stronger party may feel that it is in a position to deal directly with the perpetrator and mete out a fitting punishment. However, once the target has escaped and found safe shelter, the stronger party’s power is neutralized and it can be forced to negotiate. 

In these conventional configurations, the problem of inequality in bargaining power, which is so often put in the spotlight of criticism in contemporary ADR constellations, is circumnavigated. There is no equivalent to this non-state model of asylum in modern contexts of asymmetrical power relations between disputants, be they multinational enterprises, war parties or individual actors. Shelter-granting institutions have also not attracted the attention of the designers of contemporary post-conflict settlements who advocate the reinvention of so-called customary regulations if they consider them appropriate for all sorts of transitional justice procedures. What can be observed in such situations is a partial revitalization of institutional components deemed useful, but the necessary back-up measures are neglected or intentionally avoided. State institutions and transnational agents such as the UN engaged in conflict settlement initiatives selectively activate local knowledge regimes in order to exercise control over opaque local dispute resolution processes and achieve compatibility of the outcomes with national or transnational legal standards. 

In this outline I have argued that an examination of the conjunction of certain properties allows us to identify a specific type of legal actor and expert who is best suited to deal with conflict. While mediation as a legal institution is inextricably linked to an extended web of legal procedures and institutions, exclusion and ambivalence characterize the conflict expert’s position within the society. Interpretations based on an isolated analysis of legal institutions in conflict processing may, therefore, be misleading.