Previous Next

Reconciliation and Architectures of Commitment: Sequencing peace in Bougainville

1. Peacebuilding Compared and the Bougainville conflict

The story in brief

Bougainville suffered a terrible civil war for a decade from 1988 that pitted separatist forces of the Bougainville Revolutionary Army (BRA) against the national military and police of Papua New Guinea. The fighting evolved to set Bougainville factions against one another in the worst killing. This book argues that peacebuilding in Bougainville was shaped by bottom-up traditional and Christian reconciliation practices and a carefully crafted top-down political settlement. These two processes operated in symbiotic fashion, each making space for, and reinforcing, the other. There are important lessons in how each was designed and how one was connected to the other. For the student of peacebuilding, there is much to learn from the genius of this symbiosis between a top-down architecture of credible commitment and bottom-up reconciliation.

It has been a peace that has progressively become more resilient since 1998. The sequential sustaining of the peace has been patient—what Volker Boege (2006) has called a slow-food approach to peacebuilding. One wave of bottom-up reconciliation has built on previous waves, expanding the geographical reach of the peace and the breadth and depth of forgiveness across the society. The architecture of the top-down peace settlement has also been sequenced, with linkages that require one side to meet a commitment before the other side will deliver their next undertaking in an agreed sequence (Regan 2008; Wolfers 2006a). In this architecture, international peacekeepers played an exemplary role in securing the credible commitments. While peacekeepers were rarely hands-on mediators of the indigenous reconciliation, one of their greatest contributions was to initiate conversations between local enemies who were afraid of each other, allowing initial meetings to occur under the peacekeepers’ security umbrella.

The conclusion of the book is that the very top-down architecture of the peace agreement that has been such a strength is also potentially its greatest weakness. This is because it is far from clear whether there is credible commitment of the PNG Parliament and of regional powers to the final crunch of the peace deal. If Bougainville votes in a referendum for independence in the course of this second decade of the twenty-first century—as provided for in the peace deal—and Papua New Guinea refuses to honour the wishes expressed in that vote, young men will be motivated to return to arms to vindicate the blood of their fathers. The sequence of credible commitments so honourably completed in the peace process to date could tragically heighten a sense of betrayal if the will of the people in the agreed referendum is dishonoured. Political leadership is needed in Port Moresby and regional preventive diplomacy is required to grasp the nettle of that final commitment. This can be delivered alongside an honourable and open political campaign to persuade the people of Bougainville that they could be better off if they vote for autonomous provincial government integrated within the state of Papua New Guinea.

In the next chapter, we place the conflict in the context of the colonial history, the history of mining exploitation and the identity politics of Bougainville. Subsequent chapters describe the unfolding of the violence itself and then of 11 peace processes that failed. The sequence of failed agreements nevertheless laid foundations for a final deal that has stuck for more than a decade. A fascinating feature of the final deal is that it was catalysed by a desperate attempt by Papua New Guinea’s leadership to prevail militarily by contracting the private military corporation Sandline. When the international media, led by journalist Mary-Louise O’Callaghan, exposed the plan to deploy mercenaries, the plan collapsed, and ultimately Sandline collapsed.[1] As it raged, the Bougainville war seemed geopolitically obscure and more minor than it in fact was in scale. But in retrospect, many in the corridors of the United Nations see Bougainville as a success story of a sophisticated UN-backed peace architecture—a success that has reinvigorated the international norm against the use of mercenaries (Percy 2007) that was deeply endangered in the mid 1990s, and a peace that shows the potential of indigenous restorative justice in peacebuilding. The concluding two chapters of this book interpret more lessons from the conflict in comparative perspective.

The present chapter outlines the ambitions—methodological and substantive—of the Peacebuilding Compared project, of which this is the second volume. Readers who have read the first chapter of our first volume (Braithwaite et al. 2010) can skip to Chapter 2 without missing much.

Comparing conflict, comparing peacebuilding

The Peacebuilding Compared project adopts a broad conception of peacebuilding that is not distinguished from preventive diplomacy before conflict, peacemaking to end it or peacekeeping to monitor a peace agreement. Rather it is conceived here to incorporate all of these things, not wanting to separate peacebuilding after one conflict from preventive diplomacy to prevent a future one. Peacebuilding is therefore conceived here as tapping into a broad interpretation of peacebuilding as ‘creation of a new environment’ conducive to peace, in the words of former UN Secretary-General Boutros Boutros-Ghali (1995). The animating question for the project is what changes to an environment contribute most to enduring peace? We do not wish to close off by definition whether that contribution comes from this or that modality or phase of a peace process.

Peacebuilding Compared hopes over more than 20 years to code 670 variables in relation to the major armed conflicts that have raged across the world since 1990. The first large volume covered six different Indonesian armed conflicts (Braithwaite et al. 2010). It is hoped the third and fourth volumes will appear in quick succession to cover the Solomon Islands and then Timor-Leste. The project started with the region around the home country of the senior authors simply because it was easier to learn how to do it in the region with which the research team was most familiar.[2] As it happens, this region experienced a great deal of armed conflict during the 1990s. It was popularly referred to as ‘the arc of instability’ around Australia. As we enter the second decade of the twenty-first century, this arc is a much more stable, though vulnerable, region.

Peacebuilding Compared started in 2005. During the first five years of the project, the senior author managed to do some serious fieldwork across each of the sites in the four nations where these first 11 conflicts occurred.[3] In some cases, he was joined by co-authors for that case with far greater knowledge of that site and its languages. Joint is better, more reflexive and reliable than solitary fieldwork, but often is not logistically possible. Thankfully in the Bougainville case, most of the authors were able to spend many weeks together in the field with co-authors. We encourage a participatory approach to the research and invite readers to check out the Peacebuilding Compared web site at <>, where more information can be found. Please feel encouraged to post ideas and information to at any time throughout the 20-year life of the project.

For the project in general so far, we have been surprised by the level of access won to key players such as presidents, state and insurgent military commanders, foreign ministers, peace agreement negotiators and peacekeeping commanders. Yet, as is clear in the appendix to this volume, in comparison with the appendices in our first volume that summarised the types of players in the conflict who were interviewed, there was always uneven coverage in the types of stakeholders accessed. In every case, there were regional specialists in the study of this conflict who had secured broader access to the key players and who had talked many times to decision makers we did not mange to tap. This means it is always more important to attend to the published fruits of the fieldwork of others than to one’s own fieldwork notes.

Yet this raises the question of what added value there could be in research of inferior coverage led by researchers with an inferior background in the regions of conflict. One added value is that sometimes inferior researchers whose fieldwork engagement is thin are nevertheless lucky enough to get superior access to some significant bits of information. So there is some value from our research in adding a little to the superior body of data and insights accumulated by the very best experts in these conflicts. Yet this is not the main contribution of comparative research. Its main added value is in the comparison and in the different ways of seeing that a comparative lens opens up. In each case study of Peacebuilding Compared, there tends to be a few scholars who have done the most insightful or thorough research on that case. The frequent citation of the work of these scholars makes it clear who they are. We are deeply grateful to them. Their work remains the scholarship to read on that case; but we do hope that by standing on their books, we might be able to peer over their shoulders to begin to see more clearly a comparative landscape of patterns of conflict across the globe.

Peacebuilding Compared offers a different kind of comparativism than the dominant kind that is based on quantitative analysis of statistical information from databases maintained by organisations such as the World Bank, the United Nations Development Programme (UNDP), the US Central Intelligence Agency and national statistics bureaus. Peacebuilding Compared uses these databases as well to code one-third of its 670 variables in relation to each conflict. Most codes, however, are of things not available in these databases, such as whether insurgents received training from a foreign power or whether significant numbers of the combatants were female, based on our interviews (and published fieldwork of others). Good examples of the kind of variable never coded in the leading quantitative research are the dynamics and shape of reconciliation processes post-conflict. This is a particularly important neglect according to some of the theoretical frameworks we address in this volume.

We also attempt to deal with two fundamental problems in the quantitative literature. One is that statisticians are often interested only in data coded at the national level. The study of ‘civil wars’ dominated by the disciplines of political science and international relations is often, moreover, interested only in armed conflicts in which one of the combatants is a state.[4] Peacebuilding Compared seeks to maximise coding at the local or provincial level. Hence the way a variable is coded for the separatist conflict in Aceh might be quite different from how it is coded for the separatist conflict in Papua at the other end of Indonesia. Another difference is that Peacebuilding Compared is content to code conflicts that are many things at once. For example, Peacebuilding Compared codes Aceh and Papua as both separatist and ethnic conflicts. This is different from the approach in the quantitative literature, which tends to force conflicts into one category or another. Third, as is clear from the summary in Table 9.1, we also enter certain codes as ‘consensus’ codes among scholars and other expert commentators on the case, and others as ‘contested but credible’.

A difference from the ethnographic/qualitative literature is that Peacebuilding Compared is much less engaged with adjudicating the most contested debates about the case. We just code them as contested interpretations and we report the nature of the contestation in our narrative. What we are interested in doing is ruling out non-credible interpretations. Conflict zones are teeming with them: wild unsubstantiated rumours, ridiculous theories propagated by people who spread lies to protect their culpability, clever pieces of misinformation planted by double agents, imagined histories concocted by supposed combatants with grandiose visions of their self-importance to saving their nation. A significant level of fieldwork on the ground and in the capitals of combatant and peacekeeping states (or at UN headquarters) is needed. The intent is not to get the research team to the point where it can settle the most contested debates among the experts, but to the point where it can rule out most (hopefully all) the myriad non-credible interpretations.

A distinctive comparativism

This renders Peacebuilding Compared a distinctive form of comparativism. The approach was motivated by reading most of the best research as falling into one of two camps. The first includes a large number of wonderful studies of particular conflicts—or comparing a couple—written by scholars who have deep knowledge and long experience of that region. The second is the more recent quantitative tradition led by outstanding comparativists such as Ted Gurr, Jack Goldstone, Paul Collier, Anke Hoeffler, Virginia Page Fortna, James Fearon, David Laitin, Michael Doyle and Nicholas Sambanis, among others cited in the references. In choosing a method that aspires to significant fieldwork engagement that is inferior to the best ethnographic work, and is on a smaller number of cases than the best quantitative work, we are simply filling a methodological niche that has been under-exploited in the literature. We do not have the view that it is necessarily a superior method to the dominant two.[5] One of its demands is that it requires one person to read very extensively on each case and to be in the room or under the tree for most of the fieldwork. Otherwise it would be impossible to code the 670 variables consistently across cases. Otherwise the thematic unity of narrative volumes such as this might offer no advance on an edited collection of haphazard comparisons—insightful though such casual comparativism can be.

By 2030, we hope that some sort of cluster analysis or fuzzy set analysis to the best quantitative standards of that time will reveal something new about types of conflicts. We would also hope to define which might be the most important of probably a long list of risk factors that conduce to the persistence of armed conflict—and which are the most important protective factors for preserving peace. Narrative and analytical books such as this lay an important foundation for this future quantitative work. They discover new variables that are worth coding for all cases and new complexities in the dynamics among these variables that might ultimately account for why certain quantitative models will not explain much and why others might do so.

A final part of the method was to invite the people who seemed to be producing some of the best insights and writing the best books on the case to be members of an advisory panel. Our thanks to Sinclair Dinnen for this idea. We asked the advisory panel to suggest important people to interview, to read our first draft, comment on erroneous insights within it and on research and lines of inquiry that needed to be pursued before the next draft. Many were internationally distinguished scholars. Others were PhD students, including a number from The Australian National University, who had the luxury of recently spending long periods in the field, which senior scholars sometimes cannot manage.

Our ethical obligations under The Australian National University’s Research Ethics Committee approval were explained to all participants. These included an obligation to report quotes and insights from each informant without identifying them unless they specifically indicated that they wanted to be quoted as the source of an insight. Wherever a quote appears without a citation to some other source in the literature, it is an anonymous quote from an informant interviewed for Peacebuilding Compared.

[1] Peace journalism is of neglected importance in the peacebuilding literature. In our first volume, we found peace journalism important in ending the major Indonesian conflicts of the past decade, and rumour-mongering, sensationalist journalism to be a contributor to the onset of some of those armed conflicts (Braithwaite et al. 2010). Mary-Louise O’Callaghan deserves to be honoured for breaking a story that revived the anti-mercenary norm internationally and connected Sandline in Bougainville to a wider politics of exposing the ‘blood diamonds’ trade in Africa and a nefarious mining company politics of interference in civil war through proxies such as Sandline. O’Callaghan was also an important player in our next volume on the Solomon Islands peace. A key Australian government strategic policymaker, Hugh White, tells the story of his old friend, Mary-Louise, calling him in 2000 to urge him to support a peacekeeping intervention in the Solomons. White recounts with some embarrassment, because he thinks it a pity in retrospect that O’Callaghan’s advice was not followed at the time—his reply being that such a military intervention ‘would not fit our paradigm’ in the Pacific. O’Callaghan’s retort, abruptly ending the conversation, was ‘Well change your ******* paradigm’. Australia did change its paradigm, leading a regional peacekeeping intervention in the Solomon Islands in 2003 that so far has restored peace, with Hugh White a critical catalyst of the policy shift. During the intervening years, O’Callaghan had campaigned for that intervention in her columns in The Australian newspaper and other media.

[2] The senior author had dabbled at the beginning of the decade in some writing on peacebuilding in Indonesia after several trips there in the 1990s before and after the fall of President Suharto (Braithwaite 2002:Ch. 6), and spent time as an anthropology student living in a village in Bougainville at the end of the 1960s.

[3] John Braithwaite was present for about 90 per cent of these interviews and he typed up the fieldwork notes or used voice-recognition software to record almost 90 per cent of them. The most common reason for not creating an electronic copy of fieldwork notes was that culpability for war crimes was discussed in the interview or other information was provided that might conceivably put someone in danger. The second most common reason was that there seemed so little that was truthful or valuable in them! Handwritten notes taken during such interviews were still kept, in case a changed view of their truthfulness and value emerged later. No interviews were taped. Co-researchers had often done extensive fieldwork of their own for quite separate research projects—for example, associated with Peter Reddy’s PhD thesis. The latter fieldwork is not included in the interview statistics summarised in the appendix at the end of this book.

[4] Peacebuilding Compared studies armed conflicts in which one armed group with a command structure—even if its organisational auspices were episodic or non-institutionalised—engaged in group attacks with weapons on another armed group with a command structure. This means a clash of two warlord armies or two armed gangs can count as an armed conflict for Peacebuilding Compared if it passes certain other threshold conditions. For the moment, these are that two of the following three conditions are met: that at least 200 people were killed in the fighting within three years, at least 30 000 people were driven from their homes by the fighting and an internationally sanctioned peacekeeping mission was sent to make peace in the war-torn region. Including the last condition prevents us from excluding from consideration serious armed conflicts that started but were prevented from escalating into mass slaughter by peacekeepers (for example, the arrival of UN peacekeepers in Timor-Leste in 2006). This, however, is just a starting definition for our armed conflicts that could change as new wars occur. It sets a threshold that excludes a lot of conflicts that one might want to include. Bougainville is a civil war that clearly exceeds all three provisional thresholds we have set for inclusion.

[5] One battleground between large-n quantitative methods and single case studies arises from the qualitative critique that quantitative methods freeze (into one code) dynamic phenomena that are one thing at one point in time, another thing at another point in an unfolding conflict. This means that case studies of single conflicts do not in fact have an n of 1. Rather, they are studies of many separate episodes of violence, some of which might be more ethnic, others more religious or involving attacks by different ethnic groups than the first episode. Hence, combining the results of X qualitative analyses of protracted conflicts is more like a qualitative meta-analysis than it is like combining X cases each with an n of 1. What we are attempting in Peacebuilding Compared is a unique kind of meta-analytical hybrid. John Braithwaite deploys his knowledge of the narratives of the set of episodes of violence that makes up a particular case to code most variables as ‘high’, ‘average’ or ‘low’ on that variable. If there is some doubt about how to code (a common occurrence), it is coded ‘average’. So, ‘average’ is given the broad meaning of ‘the range on this variable where most cases of armed conflict in Peacebuilding Compared lie’. If there is both doubt and thinness of data that make it very hard to code, it is also coded as ‘hard to code’. Imagine coding two variables on the extent to which greed and grievance are motivations for fighting. The first point to make is that they can both be high or both low, or they can have different values. The second is that if greed is highly prominent in some episodes, moderately present in most and totally absent in some, the greed variable will be coded ‘average’. So these three-point codes are in fact crude summaries from a sometimes large number of data points within the single case. For some variables, such as the number of combatants on various sides and the number of refugees, we code a real number (or estimate a midpoint of a best-guess range). But we code both a maximum number (the high-water mark of the number of combatants or refugees across all episodes of the conflict) and a separate variable, which is an estimated average number across the various episodes of the conflict. All this is perhaps only slightly less crude than a purported single quantitative estimate for a single conflict (as in the extant quantitative literature). However crude, it is an attempt to quantitatively summarise from qualitative cases that are more than narratives of an n of 1. This is exemplified by the discussion in the last two pages of Chapter 9 of why there are 12 data points in the Bougainville history relevant to confirming Barbara Walter’s (2002) hypothesis that reciprocity in step-by-step demobilisation is needed for peace. Moreover, this approach to aggregating from a multiple-n sensibility for each conflict is combined with actually writing an episodic, dynamic narrative for that conflict. This is what we are doing in this book. The hope is that new kinds of insights will ultimately come from the interplay between multiple case-study narratives and quantitative analysis of the codes with this multiple-n sensibility.

Previous Next