The pattern of conflicts developing the Greater Horn of Africa is at once depressing, and exciting. It is depressing because conflict, at whatever level, entails not only loss of life, but also severe distractions from the challenges of Africa's development as it enters the turn of the century. It is exciting because conflict also brings with it the need to reasses relationships. It signals much that is thriving - or decaying - in the body politic, and hence poses challenges about future action. More pertinently perhaps, the patterns of conflict in the Greater Horn of Africa pose serious questions about the practices of its management, and centralises the need to engage in debate about creative conflict management.

The challenge of creative conflict management in this region poses serious challenges for the analyst or manager of conflict. It requires that possible and emerging avenues of conflict within states be identified and tracked. Beyond this, those conflicts that are most likely to be diffused across borders (of all sorts) and internationalised must be identified and proper modes of their management designed. The various levels of individual conflicts must be disentangled, and the interfaces between diferent types of conflict (e.g between political and environmental conflict) identified. Within this broad approach, the conceptual basis on which analysis and practical conflict management must be embedded must be defined: for, without a sound conceptual basis any type of conflict management is bound to founder.

This paper is concerned wit the regional dimension of the conflicts in the Greater Horn of Africa region, and particularly with the modes and practices of their management. Its eventual aim is to draw out the lessons to be learnt from past and current conflict management practices in the region, and hence to suggest a programme for strengthening these as a new and more challenging era dawns. In this pursuit, it examines the conceptual bases on which regional conflict management should be founded, surveys patterns of conflict management practices in the Greater Horn, and illustrates this empirically through case studies of official and unoficial conflict management in the region. It argues on that basis that mediation has emerged as the preferred practice of conflict management in the region, and in that spirit explores, through examples, the outstanding problems of mediation. Ultimately, the lessons learnt from mediation practices in the region are drawn out, and the strategic map of future needs of regional conflict management suggested.

Until fairly recently the idea of conflict systems was strange to conflict analysts, and dead to conflict managers. Individual conflicts were suffocated within territorial borders. They erupted, lived, were responded to, and re-incaranated later as individual conflicts which had no implications for, or relationships with, regional diplomatic, political, environmental and other structures This was not just because of lack of strategic thinking, but also because conflict managers failed to appreciate conflict as an organic being whose life cycle had amoebic characteristics. For example, the conflict in Sudan was appreciated and managed in the period before 1983 as if it had no transboundary realities. Secessionist conflicts such as in Ethiopia were treated as problems which only concerned the bordered realities of Ethiopian power politics whereas, as became evident after Eritrean independence, Eritrea has shifted the diplomatic and strategic balance of the whole region.

The notion of conflict systems champion the belief that every conflict has intimate relationships regionally, and what might at first appear as individualised conflicts in fact are parts of wider pattern of conflict regionally. It rejects the idea that conflicts do not have transborder realities, and instead perceives individual conflicts as an integral part of a wider conflict system. On the ground, for example, the conflict in Zaire in 1996-7 was not locked within Zaire's territorial borders. It possessed frontiers that transcended those borders, and intimate links with the conflicts in Rwanda, Burundi and Uganda. The parties to that conflict also had interests and linkages whose hinterland went beyond Zaire. What is true of political conflicts systems is even more so for environmental conflict systems, whose ecology and realities respect no territorial borders. Conflicts must therefore be seen in terms of their reality as part of wider conflict systems. In this respect, the countries of the Greater Horn relate together through shared conflicts even more than through shared borders.

The idea of perceiving conflicts within a system has some important implications for practical (and creative) conflict management. It means, for example, acceptance of the reality that management of a particular individual conflict which does not take into account systemic (or regional) realities is unlikely to be effective, much less enduring in its outcome. It also implies, in practical terms, that conflict management efforts which do not engage other interested actors within the conflict system are unlikely to succeed. It is in the catering for all interests within the system that successful outcomes are based. If interested actors i.e. systemic actors, are not involved in management processes, this means ultimately that their interests in the system will not be catered for. Hence they will be able to sabotage any outcome of a management process that did not involve, or even comtemplate them. The settlement of the Sudan conflict in 1972 for instance, failed to conceptualise the internal conflict in Sudan within its systemic setting. Since it involved only (some) of the Sudanese parties, and none of other partie in the conflict system, the life of that peace was not only checquered, but ultimately collapsed in the ashes of its individualistic foundations.

Having identified a conflict system and its interlooking dynamics, the conflict analyst or manager who wishes to be effective must identify its nodal points, and most importantly, its epicentre - the eye of the storm. Conflict systems, like all organic things, have epicentres around which their existence revolves. Identifing accurately such an epicentre is a sine qua non for effective management. However, as is evident for example from analysis of various conflict systems, the reality is that epicentres of conflict systems keep changing. Hence conflict management processes must constantly be aware of these sometimes subtle but oten quite dramatic shifts. In the Great Lakes conflict system for example, the epicentre has sifted in the last five year from Burundi to Zaire. In the Horn of Africa conflict system, the epicentre has shifted during the same period from Somalia to Sudan. As the epicentres change, conflict management processes must similarly change emphasis to accomodate the shifting balances. In practice what has happened is to abadon one epicentre for another, or altogether as has happened in the Great Las conflict system. What should happen instead is to change emphasis, shift gears accordingly, and accomodate the emerging realities of the new centre. Thus, conflict management practices within conflict system must be alert to these realities and hence be based on a foundation that accomodates them. If they are rigid, outcomes will emerge within a time warp, outdated even before the ink is dry on the treaties marking the end of such peace processes.

The traditional delimination of the Horn of Africa region consisted of Somalia, Ethiopia, Djibouti and Sudan. That delimitation was inspired by strategic concerns, particularly of the super powers during the Cold War. However, because of the interplay of wider diplomatic and strategic concerns in the region, that delimitation came to be extended loically in practice to include Kenya, and in some reckoning, Uganda. Given the fluidity of the international relation of the region, not even that expanded delimitation could be cast in stone. With its secession from Ethiopia, Eritrea clearly belongs to the strategic map of the Horn, and can not practically be excluded.

Within the Eastern and Central African region there exist two conflict systems: the Horn of Africa conflict system, and the Great Lakes conflict system. The latter comprises Burundi, Rwanda, Zaire, Uganda and Tanzania, and by some accounts, Kenya. Becaue conflict systems are in part spurred by diplomatic and economic realities, they often overlap. The Horn of Africa and Great Lakes conflict systems have a clear overlap whose sub-set includes Kenya and Uganda. Perhaps this is why these two countries are active at several levels of the two conflict systems. This interface is useful in defining either the Greater Horn of Africa or the Greater Great Lakes conflict systems. The Greater Horn of Africa is delimited by the Horn of Africa, and the interface between that conflict system and the Great Lakes one. Hence the Greater Horn of Africa conflict system comprises Uganda, Kenya, Sudan, Somalia, Ethiopia, Eritrea, Djibouti and Tanzania. At the periphery of this Greater Horn of Africa conflict system, sometimes affected by it and sometimes interacting with it, are Rwanda and Burundi.

This delimitation of the Greater Horn of Africa conflict system and those which surround, interface and interact with it has important bearings on the conflict management practices in the region. It means that any conflict within the system should be managed in terms of the wider conflict system, failing which important nuances, connections and realities might be overlooked, thus rendering such an individualistic management approach nugatory. It also opens up significantly the strategic map of actors, interests and players who must be involved in the management processes of conflict in the system. Hence a conflict system management approach expands significantly the mediation system of any conflict in the region.

The institutioal pattern of conflict management practices in the Greater Horn of Africa conflict system is still evolving, but even so, some distinct patterns are emerging. The Organisation of African Unity (OAU) provides the (historical) background against which these patterns are emerging. During the Cold War, OAU conflict management was driven in part by the interventionist imperatives of the super power rivalry. Where the interests of the super powers were involved, such as in the Horn of Africa - particularly in Ethiopia and Somalia - the OAU as an institution was not able to gain a serious entree, or even establish a firm management foothold. Also, Cold War doctrines of African international politics such as the doctrines of non-interference, uti possidetis juris, and respect for the sovereign equality of states, were the main pillars of OAU conflict management, which the organisation held dear. International legal notions such as the right to self determination provided the philosophical rationale for the existence of the OAU as a regional organisation with conflict management concerns. More seriously, the OAU was wedded to the Realist view that international and domestic politics were two distinct concerns, which in practice bore no relation to each other: hence its dichotomising between internal and inter-state conflict.

The end of the Cold War revealed serious shortcomings in OAU conflict management. It meant amongst other things that super power military intervention in African conflicts, especiall internal conflicts, came to an end, as the reluctance of the United States to become militarily involved in Rwanda and Zaire suggests. The OAU was therefore faced with a series of emerging internal conflicts which its Cold War doctrines did not permit it to handle effectively. At the same time throughout the international system, events were demonstrating that the dichotomy between domestic and international affairs could not properly exist on the ground. The OAU's practice of managing only inter-state conflicts while allowing internal conflicts to simmer was therefore put under a sharp and uncomfortable spotlight. With the settlement of the conflict in South Africa and the institutional banishment of the apartheid system, one of the pillars of the political existence of the OAU was thereby removed. The OAU was thus confronted by emerging (post-Cold War) notions about self determination and post-colonial self determination. Internal self-determination including internal self-determination is directed internally to dictatorships, while post-colonial self determination addresses the realities of ethnic groups formerly (during the Cold War) divided by territorial borders, wishing to come together again and form state units.

What all this meant was that the OAU, faced with these post Cold War challenges, needed to re-invent itself. Part of this process required ultimately that it establish mechanisms through which it could address the large problem of internal and internationalised conflicts in Africa. Logically also, it could not do this without simultaneously addressing the issue of democracy and good governance in the continent. Its 1992 review exercise suggested that it realised the urgency and need to do so. It established a Mechanism for Conflict Prevention, Management and Resolution whose short career has been chequered: it neither goes far enough in defining a new conflict management methodology, nor addresses the serious fault that its underlying principles are the same ones which for so long hampered OAU conflict management. In terms of addressing specifically internal conflict, the OAU has made some progress in that now, unlike in the past, it openly condemns military coups d'etat such as in Burundi and Sierra Leone. Indeed, in the Comoros, it was actively involved in negotiating the return of the overthrown civilian government to office.

While the OAU is grappling with these various crises of identity, there has been a significant movement towards sub-regional conflict management in Africa In this important shift, sub-regional organisations have taken the lead in trying to manage those conflicts that occur within their respective conflict systems, and particularly those which threaten the stability and security of the individual systems. Hence, IGADD (D) member states were engaged in the mediation of the conflict in Sudan, while ECOWAS engaged in peace-keeping activities in the Liberian conflict. Within the Greater Horn of Africa conflict system, some important conflict management strategies and practices have emerged, or are in the proces of emerging. These represent important shifts away from the continental strategies of the OAU. The three dominant conflict management approaches in this conflict system are institutionalised (IGADD), ad hoc (summits of the Great Lakes Heads of State), and inchoate (the East African Cooperation).

The Inter-Governmental Authority on Drought and Desertification (IGADD) was set up by states of the Horn of Africa in order to develop joint approaches to dealing with common problems, especially those related to life-threatening drought in the region, and the attendant threat posed by encroaching desertification. In its pursuit of these limited concerns, the organisation realised that such common solutions could not be achieved provided there were festering and violent conflicts within its member states. Such conflicts made it practically impossible to implement programmes the organisation came up with. It therefore found itself increasingly preoccupied with issues of conflict management, and particularly of the Sudan conflict, whose effects were felt system-wide. In this spirit, a committee of member states - Kenya, Uganda, Ethiopia and Eritrea - was appointed to mediate in the conflict. That mediation, which is discussed later on, brought to light a serious institutional limitation in that although IGADD member states could mediate in an internal conflict of one of its member states, its Charter did not specifically mandate it to do so. Hence the IGADD Charter was amended in April 1996 in order to give IGADD a conflict management mandate. Thus IGADD was transformed into the Inter-Governmental Authority on Development (IGAD). It was believed that the broad phraseology of `development' encapsulates wider concerns, including conflict management. As part of this institutional transformation, the IGAD secretariat was also enlarged to include a department of conflict management. That department is part of the wider division of political and humanitarian affairs. It is mandated and expected to coordinate IGAD's conflict management concerns.

The Summit of the Great Lakes Head of State is not institutionalised like IGAD is, and is more ad hoc in character. The Summit was first summoned over the conflict in Burundi. The Heads of State of the countries in the Great Lakes conflict system, and those concerned from the Greater Horn of Africa met in order to take collective measures to manage the conflict in Burundi, whose effects were being felt system-wide. This was the first time that heads of state within a conflict system had met with a specifically conflict management agenda. That process led to the agreement to impose sanctions on Burundi, the first time also that such collective action had ever occurred in the region. Although the Great Lakes summit on Burundi had earlier decided to lift partially the sanctions, pending progress on negotiations between the government and contending groups, such sanctions were reimposed in September 1997, in the face of the Burundi government's intransigence.

A Burundi syle ad hoc summit was also summoned over the conflict in Zaire, except that this time it did so under the chairmanship of President Moi of Kenya. The ad hoc Nairobi summits were however unable to take concerted measures, owing in part to competition between some members about the diplomatic centre of power in the whole process. The process thus moved first to South Africa through Ugandan diplomatic manoueverings, to the OAU where it was handed over, and finally back to South Africa under OAU auspices, but in truth finally consolidating Ugandan diplomatic gamesmanship.

The East African Cooperation is an attempt to institutionalise a forum through which cooperation between the East African states can obviate conflicts whose sources are economic. The Cooperation is less ambitious than the East Africa Community which collapsed in 1977 amid growing conflicts between its member states of Kenya, Uganda and Tanzania. The East African Cooperation is inchoate because its institutional format is still evolving, as are its areas of operational concern.

The philosophy of the Cooperation is a modern day experiment in functionalism. It is believed in its corridors that the more member states act in cooperation in various areas, the less conflict between them is likely to break out. It is also believed that, unlike the East African Community, this cooperation should be more people than state-centred. This belief was prompted by the view that the East African Community broke up because of unmediated conflicts between the heads of state. Indeed, so strong was this belief that in the original draft of the Agreement creating the East African Cooperation, heads of states of the three countries were given a minimal role in the Cooperation's programmes and activities. This was subsequently changed following the appointment of the first Executive Secretary, to enshrine an institutionalised role for the three heads of state. Because it was felt that the programmes of the Cooperation could not take off without the support of the heads of state, they are now required to meet in summit annually to give impetus to the Cooperation's evolving programmes. It is in this spirit that the summit of April 1997 had before it for approval for a Strategy Paper which charts out emerging and projected aeas of cooperation.

One of the interesting aspects of the Cooperation is that it does to provide specifically for conflict management activities. This is strange, given that others like IGAD found it necessary to institutionalise a conflict management role, following its practical conflict management experience. The thinking behind eschewing a specifically conflict management component in the Cooperation is that since it will be engaged in diverse areas of cooperation touching on all aspects of regional life - security, trade, economics, immigration, etc. - these will lead to integration, and with it avenues of conflict will be removed. This is theoretcially coherent, even elegant, but in practical terms rather quixotic. States, like individuals, will always experience conflict between them. The idea is to balance such conflict overwhelmingly with cooperation. This is best achieved through effective conflict management mechanisms, so that conflict, which is endemic in international life, does not overwhelm cooperative relationships, and like the East African Commnity, destroy all isntitutions of cooperation. Since conflict cannot be eliminated (and neither is this desirable), the practical response to it should be to manage it effectively. For this reason, the conceptual basis of East African Cooperation needs revisiting. In this, as in other respects, it must not step in the ame water as its predecessor.

These sub-regional practices of conflict management are encouraging, and are also largely a fresh experience for Africa generally and the Greater Horn of Africa conflict system in particular. As expected, being almost experimental, they are experiencing teething problems. They need strengthening in terms of focussing more directly on isues of peace management, rather than purely conflict management. IGAD has this sort of structure in place. Its focus on developmet, supported by isntitutionalised confict management, bears the roots of a framework for peace management. The problem IGAD is likely to face is that of being - even aspiring to be - a mini OAU, or even worse, being embraced in the bear hug of the OAU. The East African Cooperation as it is currently structured has in place some of the processes of peace management, but none on conflict management. The Great Lakes of state summits on the other hand, seem not to have contemplated that conflict management can only be consolidated by an equally heavy dose of peace management.

There have been many instances of conflict management in the Geater Horn of Africa conflict system. This section will concern itself overwhelmingly with post-Cold War practices of conflict management in this system. The rationale for this is that during the Cold War, conflict management in the region was almost exclusively the concern of the OAU. OAU conflict management dichotomised between internal and inter-state conflict. In directing its management practices almost exclusively on the latter, it failed to make an impact on the management of internal conflicts which not only continued to fester, but were also contagious within the system, and hence became diffused conflict-system wide. The conflict in Sudan is a prime example of this. Other 'internal' conflicts such as those in Uganda and Somalia, while not diffusing throughout the system, did so into significant portions of it. By the time the idea dawned that sub-regional approaches could effectively step where the OAU had feared to tread, the conflicts had become firmly entrenched in the landscape of the Greater Horn of Africa conflict system. This foundation of conflict within the international relations of the system required a different management architecture, which this section will examine.

Two conceptual bases need to be touched in any investigation of conflict management practices in the Greater Horn conflict system. The first is the relationship between track one and track two diplomacy, and the second is the distinction between settlement and resolution of conflict.

Track one conflict management refers to official, state approaches to conflict management. These are nested on offficial diplomatic concerns, their foreign policy limitations, and are often trapped within the paraphernalia of diplomatic protocol, and the intellectual bricolage of official (sometimes officious) positions. While track one confict mangement operates within the walls of official diplomacy, track two is not shackled by these concerns. It is non-official, and while it might nod at official foreign and diplomatic policy, it is not tied to them. Practitioners of track two diplomacy have no official relationship with states; while they are sometimes former officials, they can also be scholas, churches etc.

Track one and track two diplomacy have historically had an uneasy relationship. Three broad schools of thought about this relationship have emerged. The first (not surprisingly champioed by official diplomats) is that track one diplomacy is the dominant track, and should exclude track two from its processes. The second (whose exponents are track two players) is that track two diplomacy is emerging as the dominant management track and should eventually supercede track one. The third school of thought, which is the more progressive, maintains that the two tracks have different merits, and that they ought to cooperate in conflict management if it is to be effective and enduring.

Related to this debate is the distinction between settlement and resolution of conflict. The distinction is fundamental to the whole conflict management project, because, if third parties have no idea whether they want to settle or to resolve conflict, the output of their efforts at management is likely to be unmitigated confusion. Settlement of conflict entails the parties to a negotiation going through essentially bargaining processes. Such processes are founded on power relationships between the parties. Because power is the motivating credo, any bargain that is struck is likely to endure only as long as the power relationships that obtained during the negotiation subsist. Settlement processes are more concerned with power rather than with the causes of the conflict. Their outcome is thus imposed, and zero-sum in nature.

Resolution on the other hand rejects power as the basis of conflict management processes, and believe in legitimised outcomes. In resolution processes, the third party does not play a dominant role. Instead the parties are sovereign to the extent that they mutually engage in the search for a solution to their conflict. They examine the causes of their conflict and work out the basis of their post-conflict relationship. Because this process is mutual, the outcome is acceptable to both, hence legitimised. unlike in settlement, resolution processes adddress the perceptual and psychological basis of conflict in the belief that in this way, a win-win solution can be reached.

The argument is that track one conflict management is not inherently inclined towards painstaking resolution processes. And also, because itnernational relationships between states are power-based, track one processes incline towards power, and hence settlement. Track two on the other hand believes that international relationships are dominated not by power but by a whole set of other valued relationships. In track two conflict management, the threats to these relationships are explored through a painstaking resolution process. In truth, some issues are quite amenable to settlement, namely disputes for example about physical aspects like borders; about percentages of representation for each party within a post-conflict parliament or army, as was the case in the Uganda conflict mediated in 1985. But some issues are only amenable to resolution, namely conflicts. Such issues underlie all conflicts. They include issues such as the right to participate in political, social and decision-making processes of the community; the right to have one's dignity reespected and reflected in political, social and other life; and the right to self-determination (at all levels). Conflict management, to be effective and enduring must address both diputes and conflicts, and this requires both settlement and resolution processes to co-exist. It therefore calls for a cooperative relationship between track one and track two diplomacy. Without such cooperation, conflict management processes will be nugatory.

There have been several official conflict management processes in the Greater Horn of Africa conflict system. This section examines these, and particularly tries to draw out the extent to which individual instances of conflict management have taken cognisance of the systemic bases of the conflicts in question.

The strategies of response to the current conflict in Burundi are interesting in the context of systemic conflict management. They are, in effect, a combination of the responses to the Uganda and Sudan conflicts outlined above. A two-pronged approach to the Burundi conflict is clearly discernible: individual mediation, which is being undertaken by former President Nyerere, and the Great Lakes Heads of State summit which, through its own inputs, has tried to support the Nyerere framework. This is a fresh approach to conflict management within the Greater Horn of Africa conflict system.

The protracted conflict in Burundi, in all its dimensions predates the end of the Cold War, and has from time to time degenerated into genocide. This conflict is intricately connected to the other conflicts in that region, particularly in Rwanda and Zaire. Clearly therefore, the only management option that has any chance of success is one that addresses the cross-system basis of the conflict. However, peculiarly Burundian dimensions of the conflict must also be addressed. While the Great Lakes heads of state summit addresses the systemic dimension of the conflict, the Nyerere engagement addresses the Burundian internal dimension. There is therefore in place a fairly complex management approach to this conflict.

Nyerere's mediation of the Burundi conflict differs from Moi's mediation of the Uganda conflict because firstly, Nyerere is not a sitting head of state. Thus, although he might also champion Tanzania's interests, this is not within an official context. Secondly, Nyerere has the official blessings of the OAU, and indeed his mediation was undertaken at the OAU' request, although on the understanding that his independence must not be compromised. The mediation itself is being conducted within a power framework, and any leverage that Nyerere can muster is being brought to bear on the process. Nyerere appears more to be power broking than helping the parties search for a mutually negotiated solution. This explains partly why the Burundi government has attacked Nyerere's mediation, and called for a more 'neutral' mediator in the conflict.

The Great Lakes heads of state summit bolstered the Nyerere process considerably. The heads of state took the view that the most effective approach to the conflict is to act in concert, and in this respect they imposed economic sanctions on Burundi in an effort to prompt the government to negotiate with its opponents. The heads of state were also concerned with making a broader statement about conflicts and their causes in the system. They argued against governments which take power extra-constitutionally, and signalled that such processes are unacceptable withn the system. The sanctions imposed on Burundi were a fresh approach to sub-regional conflict management. The sanctions were accompanied by stiff conditions, most important being that the Buyoya government negotiate with its opponents and make significant progress towards establishing and institionalising democratic government in Burundi.

The sanctions have had some effect at least from the perspective of the Burundi government. However, even as they were partially lifted in late April 1997 there were doubts about whether all the states involved had an equal commitment to their working. But this does not reduce in any way the validity of the experiment, and the fact that sanctions were a viable option and response by the conflict system generally to an internal conflict. Meanwhile Nyerere's mediation does not appear to have made much headway. This may be explained in part by the fact that the conflict underwent a fairly sharp transition soon after the mediation started, with Buyoya's coup d'etat. The effect was to change the parties to the conflict and thus the basis of the mediation, and hence to change the issues. While some of the issues remain the same, nevertheles the conflict was considerably transformed, and this has serious repercusions for the chances of a successful outcome to the mediation.

Alongside these efforts at official conflict management, there have been several attempts at unofficial conflict management in the Greater Horn of Africa. These have centred on churches and church organisations, and most of them have been concerned with the Sudan conflict.

The East African Community was an exercise in functionalism which for a time worked, and even looked like it might endure. Through the Community the three East African states cooperated in various economic areas, and undertook common services. The political integration that underlay hopes of eventual political commmunity never worked out. In its operations however, the East African community hid many serious structural conflicts, largely because the other two member states felt that Kenya was getting the better deal out of the arrangement. These structural conflicts became exacerbated by deteriorating personal relationships between the three Heads of State, especially after the unseating of President Obote in 1971. These deteriorating relations brought the structural conflicts to the fore, and the East African Community relationship formally ended in 1977.

Although the community broke up in this way, the conflict nevertheless remained, because the issue of the distribution of the assets of the community still needed addressing. The three states agreed to negotiate a solution to that conflict with the help of a mediator, who spent some years engaged in this mediation. The bulk of the issues at hand in this exercise was the distribution of the assets of the defunct ommunity. This essentially was a bargaining exercise in which the parties horse-traded about who should retain or acquire which assets. Because the issues were thus clearly delimited, bargaining was an approach that was maybe the most approprite in this case. Indeed, the exercise more resembled an arbitration than it did a mediation, especially given that the political conflicts underlying the economic ones were outside the remit-and indeed competence - of the mediator.

Some salutory lessons from the East African Community experience appear to have been learnt. Some effort has been made to remove the structural conflict elements of economic cooperation by for example insisting on the principle of equal rather than weighted contributions to the budget. Decision-making has also been less concentrated on the heads of state, on the basis that cooperation should not be held hostage to the changing personal relations between them. The main problem with the current East African Cooperation arrangements is that there is no provision for a mechanism through which conflicts arising from cooperation can be managed. The argument being floated within the cooperation is that the more areas of cooperation there are, the less there will be conflicts. This is a theoretically coherent argument. However, in practice, and pending the attainment of full integration, there will be many conflicts and especially structural conflicts that will arise. Some mechanism for their management ought to be emplaced, otherwise the whole cooperation structure could collapse yet again.

The period since the end of the Cold War has witnessed the growth of political parties within the East African states. This is certainly a healthy development, and a useful aspect of the whole democratisation project. However, the proliferation of political parties has not been matched by mechanisms through which conflicts within and between these political parties can be managed. By their very nature, the operation of political parties generates conflicts. These conflicts are both structural and violent. On the other hand, the continued refusal by governments in power - this is clearly the case on Kenya - to envisage cooperative relationships with political parties can lead to the sabotage of regional cooperation mechanisms, because political parties can withdraw their legislative support for regional cooperative programmes that need approval of parliament. On a regional level, the East African Cooperation has made some institutional headway by putting in place a committee of parlimentarians from the three countries, who are constantly appraised about the work of the commission. The idea behind this is that eventually, when cooperation measures require ratification in the various parliaments, they will find a ready audience, and one which understands the issues.

This is a useful mechanism. However, it requires to be supported by mechanisms through which political parties can manage inter an intra-party conflicts. The currently preferred methodology of conflict management within political parties has been litigation, which is not the most effective approach. At best, it only settles the conflicts and does not go to their sources as the experiences of FORD-K, Ford Asili and the DP have shown. Arbitration has also been attempted, but the problem with arbitration in this context is that its effects are eventually similar to those of litigation. The parties in Kenya have talked about mediation, but this is yet to be put in place. The mediation option is one which should be supported and encouraged, especially in as far as it might lead to resolution of the inter-party conflicts.

What emerges clearly from the foregoing accounts of conflict management practices in the Greater Horn of Africa is that mediation is the most widely preferred - indeed dominant - mode of conflict management. This is not surprising, and is in keeping with conflict management choice making internationally. But although mediation as a conflict management practice has been preferred in this way, most of its practices in the Greater Horn of Africa region have faltered. These will be examined below. At this point, it should be noted that there has been a confusion between mediation proper, and mediation-like activities. This has meant that where such activities have been attempted, the practice has not been properly conceptualised, and this has in turn militated against successful outcomes. Surrounding this is the related problem of what actually constitutes a successful outcome to mediation. Currently the 'success indexes' in vogue such as that by Bercovitch and Kriesberg's modification suggest that almost any mediation effort will result in some form of success, either full, or partial. This is unsatisfactory, because for proper judgements to be made about mediation there needs to be a fairly clear cut criteria of success or failure.

The major problem with conflict management practices generally in the Greater Horn, and mediation in particular, has been the inability to see the process in the context of a wider conflict system. This has especially been the case with those mediations undertaken individually, such as Kenya's mediation of the Uganda conflict in 1985. Perceiving the management process within the context of the conflict system enables wider parties and interests in the conflict to be accommodated, and hence to reach a solution that is mutually supported throughout the conflict system. But it also presents its own conceptual and practical problems of management. The first problem concerns the identity of the third parties involved, and how these relate to the conflict, and the mediation. Traditionally, two types of third parties are recognised: exogenous and endognous. While the former comes from outside the conflict, the latter comes from within it. This traditional categorisation of third party is not satisfactory when seen in terms of a conflict system. Within a conflict system a third category of mediator, namely a heterogenous one is discernible. Such mediator possesses both endogenic and exogenic characteristics. In terms of practical mediation this means that such a mediator is psychologically unable to "keep a distance" from the conflict because s(he) is also an integral part of it. This certainly poses serious challenges to the basis for impartiality in mediation as articulated most forcefully by Touval.

This kind of problem was evident in Kenya's mediation of the Uganda conflict where President Moi was unable, because of his heterogenic characteristics, to distance himself from the Ugandan parties and issues, thus adversely affecting the outcome of the mediation. Similarly the IGAD(D) mediation of the Sudan conflict has also been affected by this triple identity of the mediators, and this has affected not only inter-mediator relationships, but also those between the mediators individually, and the Sudanese parties. The peculiar problems that heterogenous mediators face is one of the outstanding problems of conflict management in the Horn Africa conflict system, and one which calls for careful and detailed study.

A second problem relates to intra-mediator conflicts, and there are two levels to this problem. The first concerns conflicts between the party/parties in conflict, and the mediator(s). The second concerns conflicts between the mediators themselves. Where the mediators' relations with one or other of the conflicting parties are tense, and even hostile, this reduces the prospects for a successful outcome. Similarly where the relations between the mediators themselves are hostile they will be unable to deliver the concentration on the issues required to generate successful outcomes. This has clearly been he case with the multilateral IGAD(D) mediation of the Sudan conflcit, where deteriorating relations between some mediators like uganda and Eritrea and Sudan have not promised much for the mediation. On the other hand, hostile relations between some mediators like Kenya and Uganda have adversely affected their ability to work in concert as mediators. Clearly, a supporting mechanism by which these intra and inter-mediator conflicts can be managed needs to be designed and put in place if the kind of multilateral conflict management effected by IGAD(D) is to have any chance of success. A similar problem dogged the heads of state management of the Zaire conflict, at least in as far as the Nairobi initiative was concerned. That initiative was put in place by Kenya which summoned the first and subsequent summits in Nairobi. It soon became clear however, that there were fundamental fissures within the process, which became more pronounced with the expansion of the Heads of state group to include especially the Southern African countries of Zambia, Zimbabwe and South Africa. In the third summit some heads of state who should have been invited were not, and some who were invited declined to attend, and this effectively put an end to that initiative. The 'handing over' of the process to the OAU in Lome was diplomatic-speak for admitting the failure of the process.

A third major problem with conflict management in the Greater Horn of Africa conflict system - and of mediation generally - is that the institutional mechanisms for conflict management and mediation are very weak. This is especially the case with the OAU itself, and with sub-regional institutions such as IGAD. While the OAU has tried to address this problem through the creation of its Mechanism for Conflict Prevention, Management and Resolution, that arrangement has several shortcomings. Its philosophical foundations mean that it cannot, conceptually, reach beyond the Charter; its hiring mechanisms are (perhaps understanably) subject to the OAU's bureaucratic policies and politics; and it is unable to conceptualise the content of some functions it sees as primary to its existence, such as designing an early warning system. The problem with this latter is that because conceptually early warning is still undefined, the mechanism may be asking the wrong questions, hence getting the wrong answers, and therefore not getting the business done. The notion of early warning goes beyond having a 'new' control room, and requires also a serious analytical input.

IGAD on the other hand now has a conflict management mandate, and a department specifically concerned with this. But its relations with the OAU need to be sharply defined, if only to avoid duplication and buck-passing at crucial moments, like the UN and OAU are wont to do on (serious) occassion. with IGAD too, there is a need for serious recruitment policies, and not civil service exercises under the guise of hiring projects. For both the OAU and IGAD, conflict management is not something to be improvised about: just as people with some knowledge of biology do not get hired as doctors, so too should people not be hired as conflict analysts just because they have a nodding acquaintance with conflict issues. This problem is also reflected within government departments and foreign ministries. Because they do not employ conflict specialists, this tells on some of the policies and approaches adopted on conflict management within the system.

One of the most problematic aspects of conflict management in the Greater Horn system is the undefined role of exogenous third parties such as the United Nations. This may stem in part from the redefinition of its role which the end of the Cold War occassioned. But apart from this general problem, there is also the issue that the United Nations and other exogenous actors such as the European states seem to be locked in their own conceptual prisions when it comes to issues of conflict management. In the Greater Horn for example, the character and tenor of discourse of some of the conflicts and their actors has changed considerably after the Cold War, and these need to be approached in quite different ways fromt he approaches in vogue during the Cold War. The conflicts in Zaire and Rwanda for example are protracted, and are based on serious structural problems which, while they could be swept away under the carpet during the Cold War, cannot be treated in that way now. Further, the change in discourse about conflict and development after the Cold War means that these conflicts require more than just quick fixes to manage them. Exogenous third parties can contribute more effectively by supporting conflict management efforts in the system rather than taking over. A ripe moment will always present itself for their involvement as it currently has in the Sudan conflict, where the exogenous actors (like friends of IGAD) should now play a more direct role in nudging the peace process forward. This is the trend the OAU should have followed in Zaire conflict. Instead of having kept a low profile until the Lome meeting, the OAU seemed to want to rush to have a mediation - any mediation -perhaps in the desperation to record some success in post-Cold War continental conflict management.

One of the emerging trends in conflict management in the Greater Horn that is interesting in the rise of what might be termed semi-exogenous third parties. These are African third parties who belong to a different but not necessarily overlapping conflict system. They are not completely exogenous because they are from within Africa and are members of the OAU. South Africa in its involvement in the Zaire conflict is such a third party. Perhaps because it is not from the conflict system, the parties in conflict found South Africa acceptable as a mediator. Perhaps this is also so because South Africa was not perceived to have the same sort (usually negative and selfish) of interests in the conflict, or even the leverage exogenous and especially western ones have in the conflict. This is a significant development and one which should be explored more carefully to examine if it holds any promise for African conflict management. If this is the beginning of a trend, then it means that conflict managers within one system, and particularly mediators, will be searched for from other conflict systems in Africa. This would be a very significant development. It is also one which calls for analysis in terms of sub-regional peacekeeping of the type ECOWAS has been engaged in. Although it would bring about (not insurmountable) problems of logistics, it might suggest that peacekeepers in one conflict system should be drawn from a different one. The implications of this especially for the coordinating role of the OAU are vast and important.

Surrounding all these issues, is the question of the OAU Charter and its ability to respond with flexibility to conflicts, for example the current ones in the Greater Horn of Africa conflict system. Although OAU heads of state and government e-evaluated the role of the organisation after the Cold War, they said nothing about the Charter and the need for its amendment to reflect the changed structures of international relations after the Cold War. Clearly, some of the doctrines underlying the OAU Charter, while they might have been valid in a Cold War context, have been overtaken by events, and even outlived their unsefulness. The doctrine on non-interference as enshrined in article 3 (2) of the Charter is one such. The doctrine of uti possidetis is another. Developments and conflicts especially within the Greater Horn have demonstrated that the strict non-interference doctrine has at best a very truncated role in contemporary international relations and diplomacy. The doctrine of uti possidetis juris too, strictly interpreted, ignores the fact that the wish to alter territorial borders cannot be checked by the Charter against the will of citizens. This was the case in Eritrean succession which ipso facto entailed a re-ordering the territorial borders of Ethiopia. It is also the case in Sudan, where promises of autonomy and secession have led some Southern factions to sign a peace treaty with the Sudanese government. The OAU Charter needs to be amended in order to reflect international political and diplomatic reality. Current heads of state of the OAU are unlikely to accede to this necessity, but this should not preclude pressure in that direction.

One of the issues surrounding conflict management and the role of third parties is the problem of ripe moments for third party intervention. The literature on ripe moments holds that third party intervention which takes place before the moment is ripe (i.e. for the conflictants and the conflict) is unlikely to succeed. This was for instance what ailed Jimmy Carter's attempt to mediate in the Eritrean conflict in 1990. This issue of ripe moments is especially pertinent for the Greater Horn of Africa In the Zaire conflict for example, much noise was made about the difficulties and reluctance of Mobutu and Kabila to agree to negotiations. In their obsession with having a negotiation take place the UN, OAU and the United States failed to read the pattern of the conflict. Clearly, neither Mobutu nor Kabila perceived the moment to be right for negotiation. There did not exist a hurting stalemate, and both parties considered that they could settle the conflict on the military track. In the absence of such a stalemate - a plateau - they were unlikely to want to negotiate seriously; and if such negotiations were coerced, they would have been unlikely to make much headway. But, locked in their conceptual prisons, neither the UN nor the OAU seemed able to appreciate this.

The ability of actors within the system to approach individual conflicts as an integral part of a conflict system is one of the most encouraging developments in the Greater Horn of African conflict system. That approach was evident in the mediatory responses to the Sudan confict, with respect to the Burundi conflict, and in the conflict in Zaire. Although response to the latter through the Nairobi peace process was not ultimately sucessful, nevertheless the fact that leaders within the system appreciated the interlocking interests and actors concerned with that conflict, was a particularly significant development.

The Nairobi peace process showed quite clearly, as did the IGAD(D) mediation in the Sudan conflict and Kenya's mediation of the Uganda conflict, that third parties are always susceptible to diploamtic, political, bureaucratic and other influences, and that these can adversely affect the ability of the third party to play a constructive role. Yet it seems that third parties are prone within this region, to jumping into conflicts from the deep end, without taking these influences into account. In other words, while the dynamics of conflcit systems appear to be appreciated and taking root, those of mediation systems are not. This means ultimately, that the earlier stages of systemic conflict management will always promise more than they deliver in this conflict system. And clearly this is what ailed the Nairobi peace process and what still ails the IGAD mediation of the Sudan conflict.

Although as was noted in the case studies there have been some track two conflict management efforts in the Greater Horn of Africa conflict system, these have been few, have been inadequately - or even wrongly - done, and have not been related to, nor cooperated with, track one approaches. None of the track one approaches have engaged unofficial tracks in their management processes, and this has been reciprocated in full measure by track two conflict management initiatives. There is quite evidently a lack of a structrue for cooperation between these two tracks, and frameworks for such cooperation need to be conceptualised and eventually operationalised. In order to make a practical difference to conflict management within the conflict system, such cooperative relationships need to be institutionalised.

An issue related to this is that there is a problem of lack of knowledge within track one and track two approaches. The sources of this problem - apart from the clear need for training - is that track one actors treat unofficial actors with feelings bordering on contempt. Track two players on the other hand view track one as incompetent. This is clearly a case of diplomacies in conflict, which needs to be resolved urgently if conflict management within the system is to prosper. In the Greater Horn of Africa conflict system for example, there is an abundance of both track one and track two players on the ground, each professing to be taking action and initiatives which would help settle or resolve the various conflicts. But these two viewpoints never merge, and inputs from one track are never shared or strengthened by those from the other. This is a sorry state of affairs indeed, and it denies conflict management efforts in this region important sources of cooperation and creativity.

Within both track one and track two conflict management exercises in the Greater Horn - as the case studies show quite graphically - there is an unfortunate pre-occupation with bargaining and its associated structures. Bargaining as a method of mediation is an approach deeply embedded inthe psyche of official, state actors. These use their roles and conflict management functions entirely within a power political framework. Eventually also, it must be admitted that such structures are necessary especially when finally consolidating the outcome of peace processes. But alone, bargaining structures are not sufficient to creat an enduring outcome. Hence track two mediation methodologies and philosophies with their preoccupation with perceptual, psychological and legitimacy components of conflict management, need to be involved within conflict management activities in the system. This perspective again harkens to the need for institutionalised cooperation between official and unofficial conflict management activities. There is really no way out if enduring outcomes - and resolution are to be contemplated.

Finally, no peace process should end at the negotiating table. Conflict, if followed by peace, requires also that the peace itself be nurtured. Hence there is a pressing need for peace management approaches to be considered as an integral component of conflict management. In the Greater Horn of Africa region, peace management has not been treated with the same urgency as conflict management has. Perhaps this is because of the high profile that conflict management involves. However, it needs to be emphasised that the relatively boring activities that peace management entails - systems of continuous negotiation at the grassroots level; emplacing peace structures in local communities; training citizens right from the grassroots on conflict, its management and processes; designing confidence building measures that can be used to bridge perceptual and psychological gaps between communities in conflict, and between those whose relations are based on long running structural conflict - are a crucial component of a working peace system. The Greater Horn of Africa system especially requires such structures of peace management. Eventually, regional and other conflict systems need to be transformed into peace systems, and this is the ultimate challenge for conflict managers and analysts.

This paper has attempted to draw up an anatomy of conflict management in the Greater Horn of Africa conflict system. In this quest, it has described and analysed the predominant patterns, practices and strategies of official and unofficial conflict management within this conflict system. The paper argued at the outset that the beginning point in approaching this task is to conceptualise the various conflicts in the Greater Horn as belonging to the same conflict system.

It was noted that although there are diverse actors and strategies of conflict management in the Greater Horn, the dominant mode and practice of conflict management is that of mediation. Some of the mediation activities that have taken place in different conflicts within the system were examined, and the problems these have given rise to enumerated. The paper has also underlined some of the outstanding problems of conflict management, which have arisen from not learning properly the lessons of systemic conflict management.

The Greater Horn of Africa conflict system has been marked by severe protracted conflicts since independence. Some of these run their course before they are transformed and begin again. But there are many more which are latent, and which are fuelled by underlying structures which generate structural conflict, and frequently erupt into violent conflict. Conflict management if it is creative, can provide mechanisms for the management of these. However, the concern of conflict managers, and others engaged in the discourse about conflict, should be with the conflict generating structures and the structural violence they give birth to. This essentially, is a project of peace management. After conflict managers have laid down their tools, it is structures and processes of peace management which will drag the Horn of Africa to the twenty first century. Conflict management, if it is creative and effective enough, must lead ineluctably to defining the project of peace management.

* * *