Coming to Terms with Separatist Insurgencies Open Access

This article seeks to extend the understanding of how certain types of civil wars, separatist insurgencies, may be brought to an end or nearer to an end through external (international) mediation. The article seeks to establish which conditions in a domestic conflict, if present, will make it more likely that an international mediation will be brought to bear on a domestic conflict and then make it more likely that a settlement or resolution will be reached and implemented. One prevalent theory about how protracted internal conflicts can be resolved through the intervention of a third party is “ripeness theory,” an approach first promulgated by I. William Zartman in the mid‐1980s and since utilized as an analytic tool by many others (Kriesberg 1987; Haass 1990; Stedman 1991; Lieberfeld 1999; Hancock 2001).

Ripeness theory posits that if a “mutually hurting stalemate” is reached in a protracted conflict, the likelihood of mediated negotiation is enhanced (Zartman 1985; Schrodt, Yilmaz, and Gerner 2003). The mutually hurting stalemate concept holds that mediation of a protracted conflict is most likely to occur when the parties are embroiled in a conflict from which they cannot achieve victory and where the “unending” nature of the conflict delivers pain to both (Lieberfeld 1999; Zartman 2001). Further, the idea proposes that a ripe moment occurs when both sides begin to see “a way out” of the conflict and that a formula can be devised that is “just and satisfactory to both parties” (Zartman 1989: 291; Pruitt 1997).

A major criticism of ripeness is that at its base the idea is tautological, that is, a mechanism that can only mount an explanation by reference to itself (e.g., A is A because it is A). A number of scholars have rendered this comment on the concept (Kleiboer 1994; Rothstein 2007; Kelly 2008). One author puts the “tautological” criticism in the following words: “The way ripeness has typically been used reduces the concept to a mere tautology, and theories that relate to it end up being unfalsifiable claims. It is hard to differentiate the concept from parties' consent to mediation and successful outcomes. Ripeness is a necessary, if not sufficient, factor for initiation and success of negotiations [citing Zartman 2000]. If mediation was successfully initiated or finalized, in other words, the moment had to be ripe” (Schrodt, Yilmaz, and Gerner 2003).

Zartman, the progenitor of the concept, has come to its defense:

Although it may be desirable for ripeness theory to maintain a stronger predictive capability, in my judgment, even in the absence of that more robust capability, the theory still has utility because it can be used to identify those elements of ripeness that either the parties to a conflict, or an external mediator would be able to recognize and act upon. Ted Gurr (2000a: 58) has noted that “disputes over self‐determination are best settled by negotiation and mutual accommodation.”

In the case of separatist conflicts, Gurr (2000a: 57) notes that it is best if the mutual accommodation occurs “after an initial show of forceful resolve rather than after prolonged warfare.” Clearly, this does not always happen, and in the two cases of protracted conflict presented here, the insurgencies realized they were “strategically overmatched,” and the governments concluded that “it is cheaper to negotiate regional and cultural autonomy and redistribute some funds than it is to fight endless insurgencies” (Gurr 2000a: 57). Gurr appears to be affirming Zartman's (2001) contention that the parties can sometimes recognize when negotiation rather than endless war offers a better alternative. The focus of this article is how external interveners may be able to recognize and then proffer a willingness to mediate or respond positively to a request by at least one of the conflict parties to take such a role.

If ripeness is indeed a condition that at some point violent conflicts may enter into, and that can be discerned by the parties to that conflict or by an external actor, the question arises as to how this recognition, or perception, occurs. Scholars have suggested that objective “elements” or “referents” of ripeness can sometimes be identified (Zartman 2001; Pruitt 2005). If this recognition by the parties or a mediator occurs, the likelihood of a suspension in hostilities would increase. If an outside observer were able to detect these referents, then it is likely the level of subjectivity imbued in conducting an analysis of a conflict, which relies on the notion of ripeness, could be reduced.

Included within this recognition of ripeness is a sense by the parties that “a way out” has presented itself that will not bring greater dishonor or incur greater material costs than that which would be incurred by remaining embroiled in the conflict (Kriesberg 1987). “Ripeness is a matter of perception, and as with any subjective perception, there are likely to be objective facts to be perceived. These can be highlighted by a mediator oran opposing party when they are not immediately recognized by the party itself . . .” (Zartman 2003).

Concluded mediated agreements directed toward ending separatist conflicts can be rejected by the constituents of one of the parties, however, particularly on the government side. Accountability to constituents can be a potentially important factor in bargaining when the negotiating team is responsible for dividing a tangible good (Druckman 1994). Determining the territorial bounds of a state would seem to qualify as this type of negotiation, requiring the participation of constituents. That participation was less than sufficient in the Philippines negotiations between the government of the Republic of the Philippines (GRP) and the Moro Islamic Liberation Front (MILF), which is one of the cases studied here.

In the Indonesian case analyzed here, negotiations conducted between the government of Indonesia (GOI) and the Free Aceh Movement (Gerakan Aceh Merdeka or GAM), constituent participation was precluded by at least two factors. First, unlike in the Philippines, there was general agreement in Indonesia regarding the territorial extent of the homeland in the Aceh region claimed by the insurgent group. The degree of self‐government or autonomy was being negotiated, not Indonesian territorial integrity. Second, the government held the belief, valid or not, that Indonesian law negated the need for consultations with constituent organizations, although there was protest (not widespread) from some opposition political party members protesting the lack of consultation (Miller 2009).

In an important early study of labor–management bargaining, Richard Walton and Robert McKersie (1965: 289) noted that negotiators representing constituents may be placed in inherently difficult situations, which creates what they called “boundary conflict.” This conflict results from the pull one negotiating team may feel in the direction of the demands made by the other side across the table as well as an opposing pull from their own constituents' expectations. Boundary conflict occurs when a negotiator believes she cannot ignore the expectations of the other side, becoming isolated from her own side and developing views distinct from those of her constituents (Walton and McKersie 1965). (The Philippines' insurgent conflict discussed below provides an example of this kind of conflict.)

Drawing largely from experimental work, Jeffrey Rubin and Bert Brown (1975) noted that constituents who were highly dependent on a representative were able to place pressure on their representatives to display commitment toward constituents' positions. Sometimes, a representative's commitment to his constituents' positions impedes his ability to perceive genuinely superior arguments made by the opposing side (Rubin and Brown 1975). Evidence from the Philippine case suggests that arguments submitted by the MILF negotiators to their Philippine government counterparts were regarded as reasonably compelling, given the protracted length of the conflict. But as will be shown, the government negotiators' constituents in Philippine civil society were not persuaded.

To explain why reaching an agreement to end certain separatist conflicts through mediation can become more likely, but implementation uncertain, I argue that if the state alters its military strategy from offense to defense against an insurgency and the insurgents revise their demand from secession to self‐determination through homeland autonomy, achieving a mediated agreement becomes more likely. But implementation remains uncertain if self‐determination is defined in territorial terms prior to adequate constituent review.

It should be understood that the phrase “right to self‐determination” (beyond its application to formerly colonized peoples) can apply to a range of political options, from self‐identification (mostly in cultural terms) to self‐government not including secession, to outright secession and independent sovereign state status. As Hurst Hannum (1993) has made clear, international law does not recognize a “right” to secession, although secession is not prohibited either. There are, however, numerous international legal instruments that confer upon peoples a right to self‐determination (Hannum 1993).

Hannum (1998) has suggested that a right to self‐determination should not be based primarily on a claim to territory. He argues that most current states were founded directly or indirectly on conquest at a point in history when many of the “acts of conquest” were considered legitimate (Hannum 1998). But this position begs the essential question of considering who it was that was conferring “legitimacy” on such acts and, consequently, on the present‐day legality of how certain states are territorially constituted.

Lea Brilmayer has offered the outlines of an answer to this question, arguing “that every separatist movement is built upon a claim to territory, usually based on historical grievance, and that without a normatively sound claim to territory, self‐determination arguments do not form a plausible basis for secession” (Brilmayer 1991: 192). In her view, claims to secession would only be valid in those cases where it can be convincingly shown that the constitution of certain present‐day states is illegitimate, that is to say, where a historical wrong was committed in order for particular states to be defined territorially as they currently are (1991).

Both of the insurgent movements studied here have at one time sought secession. Both the MILF and GAM based their claim on historical evidence suggesting that the conquest, respectively, of the territory in the Philippines and Indonesia that the movements now claim as their “homeland” was indeed illegitimate (Gershman 2001; Kingsbury 2007a).

But as will be shown below, both insurgent movements articulated a shift in their demands from secession to self‐determination. “Without entering into a debate of whether the right to self‐determination is jus cogens, it would seem difficult to question its status as a ‘right’ in international law” (Hannum 1990: 45). Part of the argument herein is that both the MILF and GAM have come to interpret the right to self‐determination as a “right to autonomy” and have sought the affirmation of this latter right in negotiations with the state in which they reside.

The states under study here, Indonesia and the Philippines, have both experienced multiple protracted insurgencies. The insurgencies under examination have largely been concluded through mediated negotiations, although recently a prospective settlement in the Philippines has undergone a setback (ICG 2008a). I contend that “ripe moments” occurred and were exploited by external mediators (as well as the parties) in Indonesia and the Philippines but that this exploitation was truncated in the latter case because of a failure to implement the concluded “agreement in principle.”

In lieu of regime change (with the installation of a government controlled by the insurgent group) or outright secession and the creation of a separate state, what insurgents may seek over the course of a conflict is a more equitable distribution of government revenues, a greater share in natural resource exploitation, an end to discrimination/repression, or a measure of autonomy (Lapidoth 1997; Gurr 2000a, 2000b). The demand for autonomy usually entails a significant “devolution” of authority from the national government to a lesser jurisdiction, defined in part by territory (Chattopadhyay 2002). To gain one or more of these aims, short of a military defeat of government forces, some level of discussion with the state is usually necessary. But will the parties recognize that negotiation with the other side could provide advantages greater than the advantages associated with continuing hostilities (Druckman 1994; Bercovitch, Diehl, and Coertz 1997)?

An insurgent group's decision to revise its demand for self‐determination and a government decision to alter its military objective may both exemplify the observable objective referents that have been put forward by some observers as a means of detecting ripeness in certain protracted insurgent conflicts (Pruitt 2005). Gurr has noted that the parties in certain separatist wars have recognized that the cost of accommodation is probably less than the cost of continued conflict, but when this recognition is lacking, such wars can become protracted (Gurr 2000b).

This article seeks to answer the question of what shape accommodating behaviors might take, and how can international mediators help governments negotiate with insurgencies, “providing both sides with incentives for choosing autonomy rather than secession” (Gurr 2000a: 62). But mediation success includes the implementation of concluded agreements. Negotiating teams must wield authority sufficient to gain the approval of those agreements from the respective teams' constituencies (Druckman 1994). Constituent approval may be unobtainable, however, because of the specific self‐determination demands agreed upon in negotiations.

The insurgent conflict under study here is between the GRP and the MILF, a former faction of the Moro National Liberation Front (MNLF). From the July 18, 1997 signing of the “Agreement for General Cessation of Hostilities” (AGCH) reached with the administration of Philippine President Fidel Ramos to the July 27, 2008 issuance of the “Memorandum of Agreement on Ancestral Domain” (MoA‐AD), the MILF and the GRP have reached agreements or “understandings” on a range of issues that include, but are not limited to, the geographical extent of a Moro homeland, the allocation of the designated region's resources, and, obviously, the governance of that homeland. The latter pertains to what has been termed the devolution of legal and political authority from the central government to the Moro homeland, generally understood as the island of Mindanao and certain other peripheral territories (chiefly Palawan and the Sulu island chain) in the southern portion of the larger Philippine archipelago. It is this matter of the relationship between the Philippine state on the one hand, and a proposed political entity in the south, distinct from the Philippine state — an entity where the MILF presumably would have a major role in governance — that remained heavily contested between the two parties until late 2008.

A lack of progress in negotiations with the MILF moved the government of President Joseph Estrada (who succeeded Ramos) to launch an “all‐out war” against the MILF in April 2000 (Santos 2004). This campaign, mostly intended to wrest control of certain swaths of Mindanao territory away from the MILF, resulted in the temporary halt of the GRP–MILF negotiations (Bacani 2006). The negotiations between the GRP and the MILF only resumed with the installation of President Gloria Macapagal‐Arroyo in January 2001, followed by the signing of the “Agreement on the General Framework for the Resumption of Peace Talks” in March 2001.

The latter document prepared the ground for the “Agreement on Peace between the Government of the Republic of the Philippines and the Moro Islamic Liberation Front” (Tripoli Peace Agreement of 2001) and became the “background” document establishing general principles for much of the subsequent negotiations between the parties (GRP–MILF 2001). In this phase of the GRP–MILF negotiations, President Arroyo requested that the Malaysian government intervene as a facilitator in the talks between her government and the MILF (Bacani 2006).

In this time period, Malaysia acted not as a “strong” mediator, that is, an intervening actor able to suggest proposals and counteroffers to the parties, but rather only as a facilitator, that is, providing a locale for the talks and serving as a “go‐between” (Martin and Tuminez 2008). Facilitation can be understood as helping groups find a solution to conflict without imposing an outcome (Harish 2006). Facilitation is closely associated with the “good offices” concept in which “a third party acting with the consent of the disputing states serves as a friendly intermediary in an effort to induce them to negotiate between themselves without necessarily offering the disputing states substantive suggestions of settlement” (Bledsoe and Bozcek 1987: 290).

A key principle of the Tripoli Agreement (not to be confused with the 1976 Tripoli Agreement between the GRP and the MNLF), was the explicit recognition that a “Bangsamoro homeland” could be created without sacrificing the sovereignty of the Philippine state (GRP–MILF 2001; Bacani 2006). In that agreement the parties stated their belief that “normalization” in “conflict affected areas” could be achieved (Agreement on Peace 2001: sec. A).

The question then became how the idea of an ancestral domain, that is, homeland, could be realized, when the parties maintained clearly differing views on what the idea of “self‐determination” would mean for the Moro population of the Mindanao region. Self‐determination was first offered to the MILF by the GRP in March 2007 and at the time was regarded as a breakthrough by the MILF because it would allow self‐governance (although not independence) for Moros within their ancestral domain territory (Gopalakrishnan 2007). The MILF and the GRP conducted several rounds of negotiations between the June 2001 Tripoli Agreement and the release of the GRP's “Draft Final Peace Agreement” in February 2003. When this draft did not receive a warm reception from the MILF, the Arroyo government suspended negotiations and launched the Buliok offensive in February 2003 (IISS 2005).

Negotiations did not recommence until the Armed Forces of the Philippines (AFP) withdrew from the Buliok camps (relatively remote areas where the MILF had earlier established military control) in early 2004 and the MILF renounced ties to the terrorist groups Abu Sayaf and Jemaah Islamiyah shortly thereafter, a precondition required by the GRP. The Malaysian government continued to maintain its role as facilitator into 2004 (Santos 2004; Bacani 2006). The MILF continued to insist on its ultimate goal of independence in some form, although the group had decided to hold that demand in “abeyance” in order to pursue talks on the rehabilitation of the areas in Mindanao most adversely affected by the fighting (Gershman 2001).

For its part, the GRP insisted that any agreement reached on the geographical expanse and political structure of a Moro ancestral domain must not harm or threaten harm to the sovereignty and territorial integrity of the Philippine state. But as one observer has put it, both parties seemed to share a tacit recognition that a Bangsamoro homeland was not necessarily incompatible with Philippine territorial integrity and sovereignty (Bacani 2006).

Importantly, in its withdrawal from the Buliok camps, the GRP had determined that in the best interest of promoting success in the negotiations, the army would shift its overall strategy from seeking to defeat the MILF to maintaining a defense against the insurgent group. In a press release President Arroyo stated: “Today the peace panels of our government and the MILF have agreed on a mutual cessation of hostilities. I am directing the Armed Forces to downgrade our operational status from punitive operations to active defense” (Office of the Press Secretary 2003, emphasis added). The fact that this introduction of a restraint in military strategy was linked to a reciprocal ceasefire notwithstanding, such a strategy alteration is notable.

The parties seemed to agree that the Bangsamoro were a people defined in large part by religious belief and cultural history, distinct from the majority population of the Philippines, with the former residing mostly in the geographically defined territory of Mindanao (GRP–MILF 2001). (The term Bangsamoro is generally interchangeable with Moro and is derived from the Malay Bangsa, meaning nation or people, and the Spanish Moro, referring to Arabs or Muslims.) The negotiating parties also agreed on the need to expand in jurisdictional, and not just geographical terms, the bounds of a Moro homeland or ancestral domain. In the new MoA‐AD, the parties agreed that the territorial expanse of ancestral domain was to include not only the ARMM (Autonomous Region of Muslim Mindanao) created in 1989 by the Organic Act for the ARMM (Republic Act 6734), but also 712 additional villages, many of which contained populations who were not majority Moro (BBC News 2008).

This understanding was articulated in the MoA‐AD under the title “Governance.” The MoA‐AD refers to a comprehensive compact (yet to be written) that would “flesh out” and finalize the former and also refers to a future Bangsamoro Juridicial Entity (BJE), a term that suggests that Moro self‐determination would lead to a self‐governing polity, with a defined territory, distinct from the Philippine state (GRP–MILF 2008). Paragraph four under the “Governance” title of the MoA‐AD states that “a period of transition shall be established in a Comprehensive Compact specifying the relationship between the Central Government and the BJE” (GRP–MILF 2008).

It is the specification of the relationship between the Philippine government and a future BJE (the MILF currently serves as a surrogate for the latter) that has severely disrupted the conflict resolution process in Mindanao. The government made clear that throughout the discussions leading to the promulgation of the MoA‐AD, it never intended to grant independence (in the sense that Philippine territorial integrity would be diminished) to a Moro homeland. In a public address, a presidential advisor in the peace process noted that the “granting of independence is not and never was the intent of the talks between the government and the Moro Islamic Liberation Front in [pursuing] the peace process in Mindanao” (Esperon 2008).

The insurgents had earlier in the negotiations rejected the possibility of autonomy, particularly if such a prospect was to be based on the model provided by the ARMM, although the MNLF (which preceded the MILF) had previously recognized this as a prospectively viable entity when it signed a final peace agreement with the GRP in 1996 (Tuminez 2007). An offer by the Arroyo government in 2003 for “enhanced autonomy” (first put forward by the Estrada administration) was also rebuffed by the MILF (Minda News 2010). Prior to a fresh attempt to end the conflict by meeting with the Philippine government in Kuala Lumpur in 2004, the MILF reiterated its aversion to an autonomy solution for the conflict (Philippine Daily Inquirer 2004). In the MoA‐AD's final formulation (August 2008), however, both the GRP and the Moro rebels agreed that “the core of the BJE [Bangsamoro Juridicial Entity] shall constitute the present geographic area of the ARMM” (GRP–MILF 2008).

The MoA‐AD's formulation generated certain expectations on the part of the MILF and the Moro population, not the least of which was that the opposition of certain members of the Philippine Congress and non‐Muslim segments of the general population opposed to further expansion of the autonomous region could be circumvented. “In the new framework, a GRP–MILF peace would govern the enabling law for the Moro homeland, thus preventing Congress from emasculating Moro gains from negotiations. ARMM enlargement and the creation of a genuine Moro autonomy could theoretically happen without being held hostage to the obstreperous and obstinate opposition of Congress or local anti‐Moro groups” (Tuminez 2007).

And, under the governance section the MoA‐AD includes the following statement on territory: “The ultimate objective of entrenching the Bangsamoro homeland as a territorial space is to secure their identity and posterity, to protect their property rights, and resources as well as to establish a system of governance suitable and acceptable to them as a distinct dominant people” (GRP–MILF 2008). But this agreement between the negotiating parties that self‐determination understood as self‐governance required a territorial “set‐aside” of a certain expanse failed to find acceptance among particular congressional members or certain segments of civil society.

In recent years, the role of civil society in bringing insurgent conflict to an end has come under increasing scrutiny. Catherine Barnes (2006: 19) has defined the term “civil society” as: “[t]he web of social relations that exist in the space between the state, the market (activities with the aim of extracting profit) and the private life of families and individuals. Interlinked with the concept of ‘civil society’ is the idea of social capital: the values, traditions and networks that enable coordination and cooperation between people.”Anthony Wanis‐St. John (2008: 15) defines it as: “the vast array of public oriented associations that are not formal parts of the governing institutions of the state: everything from community associations to religious institutions, trade unions, nongovernmental organizations (such as human rights groups, relief organizations, development organizations, and conflict resolution groups), business associations, and professional associations such as the Bar or accountants' associations.”

Although, Mindanao has a large, active, and diverse civil society sphere, predominantly Muslim organizations are underrepresented and predominately Christian organizations are overrepresented (Rood 2005; Bacani 2006). Several Mindanao civil society groups have shown a sustained interest in supporting the effort to end the conflict. Business associations, religiously affiliated organizations, and activist nongovernmental organizations (NGOs) have participated in this work and they have often joined to form networks (Rood 2005).

But because of the reasonably well‐documented discrimination and marginalization of the Muslim community in Mindanao — a condition that helped spawn the insurgency in the first place — Muslim civil society efforts (because of the lower density of Muslim organizations) to end the conflict have been weaker than those of predominately Christian civil society groups (Rood 2005; ICG 2008a). This state of affairs has important implications for the idea that “front channel” negotiators would do well to make serious efforts to consult with civil society groups, if not give them an actual “seat at the table” as the peace talks go forward (Wanis‐St. John 2008). Front channel or Track I negotiators, if they were so inclined, would have to pay close attention to the composition of civil society groups brought into the peace process in Mindanao to be certain that both the Christian and Muslim communities were well represented (Pfaffenholz, Kew, and Wanis‐St. John 2006)

In July 2004, President Arroyo asked the Malaysian government to expand its role as an intervener in the conflict from facilitation to mediation. This moved Malaysia from the referee role to a role in which its representatives were empowered to suggest alternative solutions (Santos 2004; Camilleri 2007). In the talks mediated by the Malaysian government, which led to the creation of the MoA‐AD, the following interpretation of what is meant by self‐governance was inserted into the document (paragraph 8 under governance) on ancestral domain:

But it has been those details that have proven, as of this writing, to be insurmountable for members of the Philippine Congress and particular groups in civil society. In August 2008, the Arroyo government stated that the MoA‐AD had been cancelled, partly because of skirmishes that had occurred in several provinces but also because of strong objection from legislative members, as well as civil society groups, to several of its provisions (Asia News 2008). Opposition to signing the MoA‐AD was sufficiently cohesive enough to lead to a petition to the Supreme Court asking for an injunction against the signing, which had been slated for August 5, 2008 (ICG 2008a). The injunction was granted and the court declared the agreement unconstitutional in October 2008 (ICG 2008a; Zamboanga.com.Editorials 2008).

The events leading to the unconstitutional determination by the court, produced in large part through the effort mounted by the constituents of the GRP front‐channel negotiating team, could have resulted from those constituents' belief that the costs of “no‐agreement” were lower than the costs that would result from a ratification of the MoU‐AD (Putnam 1988). That is, those constituents decided they would prefer to retain the status quo in Mindanao (with the potential for continuing hostilities) rather than abide by an agreement that extended unacceptable concessions to the MILF. “No‐agreement often represents the status quo, although in some cases no agreement may in fact lead to a worsening situation” (Putnam 1988: 442). Whether the situation in the southern Philippines has in fact worsened remains uncertain.

The Malaysian government continues to seek a resolution to the conflict (Lingaa 2006). The Malaysian ambassador to the Philippines has stated that his country “will continue to play a key role in the peace process” (Sy, Katigbak, and Unson 2008). Although the Organization of Islamic Conference (OIC) has indicated that “the talks should be re‐instituted in the context of Philippine territorial and sovereign integrity” (Sy, Katigbak, and Unson 2008), for the moment the talks remain in suspension.

Among several insurgencies in Indonesia, I focus here on the conflict in the province of Aceh, where the insurgency was waged by the Free Aceh Movement (Gerakan Aceh Merdeka [GAM]) between 1976 and 2005. The dispute between Aceh and the state was initially a central government versus “peripheral areas” dispute over natural resources and state revenues (Miller 2008). At the outset of the conflict, GAM sought independence from the government in Jakarta, greater control of natural resources residing in Aceh, and a greater share of state revenues, particularly those generated by Aceh's natural gas deposits (Sims 2000).

The Acehnese population believed it had achieved a degree of territorial control, if not outright sovereignty, in the precolonial period over what is today the northernmost province of Indonesia on the island of Sumatra (Bacani 2006; Kingsbury 2007a, 2007b; Barnes 2009). While petitions for autonomy to the governments of President Sukarno and President Suharto were initially given sober consideration, neither government followed through with concrete positive steps in the direction of autonomy (Schulze 2007a). Although the Acehnese had mounted a guerilla resistance early in the Dutch occupation that eventually contributed to the Dutch withdrawal from the region, after independence in 1949 the Indonesian state incorporated Aceh (Siregar 2008; Barnes 2009).

After the rebellion began, the Indonesian government instituted an exceptionally harsh counterinsurgency policy, which itself generated further demands for independence (Schulze 2007c). But probably the most important motivation for the insurgency was the discovery of immense deposits of oil and natural gas in the region and the failure of the central authorities to return to the province an equitable share of the revenues from the exploitation of these natural resources (Schulze 2007a; Barnes 2009). The Aceh provincial government had pressed in the years following independence to be allowed to implement Shari'ah (Islamic law) more widely, and the Aceh local government was allowed to do so after a plebiscite in 2000. But GAM itself is not an Islamic movement, and some evidence indicates that the central authorities in Indonesia allowed the wider application of Shari'ah as a counter to the movement's influence (Kingsbury 2007a; Siregar 2008). GAM has attempted to mobilize the population based on Acehnese nationalism rather than fidelity to Islam (Brown 2004; Tan 2007).

The Indonesian government declared Aceh to be Derah Operasi Militer, a special military zone, in 1989, a designation that endured until 1998 when President Suharto was forced to leave office (Eye on Aceh 2005). During this period the Indonesian military, Tentara Nasional Indonesia (TNI) directed high levels of violence against GAM cadre and suspected GAM members, which resulted in pervasive human rights violations (Large and Aguswandi 2008).

The end of the Suharto regime brought the ascension of Abdurrahman Wahid to the presidency in October 1999. The insurgents and the government reached a short‐lived cease‐fire in May 2000, but after two extensions were appended to the original three‐month time frame (Eye on Aceh 2005;Schulze 2007a), it disintegrated by January 2001 under President Megawati Sukarnoputri.

GAM entered into a new agreement in December 2002 (Cessation of Hostilities Agreement [CoHA]) with the new national leadership (Aspinall and Crouch 2003). This agreement between the parties was assisted by mediators from the Henry Dunant Centre of Geneva (now known as the Centre for Humanitarian Dialogue [HDC]), a Geneva‐based nongovernmental organization (Aspinall and Crouch 2003; Huber 2008).

The purpose of the CoHA was to provide an opportunity to negotiate a resolution of the conflict, but the attempt failed, at least in part because the parties were unable to reconcile two juxtaposed positions. Article 1(b) of the CoHA stated that the purpose of the agreement was “to continue the confidence building process . . . to enable . . . the peace process to proceed to the next phases” (GOI–GAM 2002). The government offered “special autonomy” as a status for Aceh, while at this stage GAM continued to insist on independence, which seemed to require creation of a separate sovereign entity (Aspinall and Crouch 2003). But during the negotiations in the CoHA interim, the government insisted on GAM disarmament (Kingsbury 2007b; Schulze 2007c), a step that the group refused to take. Consequently, the talks ended, and hostilities began anew in May 2003 (BBC News 2003; Huber 2004).

The dissolution of the CoHA was followed by a presidential decision to impose martial law in the province (ICG 2003). This period was marked by a virulent TNI campaign to gain “victory” in the Aceh conflict. The military measured progress in part by the casualties inflicted on insurgent fighters and their alleged civilian supporters by the at least 35,000 military troops put into the field by the TNI (Huber 2008).

Strong evidence indicates that the intensified military campaign was effective in reducing GAM's military reach (Schulze 2007b; Huber 2008). Many GAM fighters did surrender to TNI forces, and most of the rest retreated to remote forest and mountain regions (Eye on Aceh 2005). One observer suggested that by the middle of 2004, GAM was in danger of military defeat (Schulze 2007b). But whether GAM and its supporters in the Aceh civilian population were close to capitulating politically remains debatable (ICG 2003; Miller 2009).

In any case, in early 2004 soon‐to‐be Vice‐President Jusuf Kalla (through the national elections of April 2004) made overtures to the GAM leadership in the field rather than to the Swedish‐based leadership in exile. An Acehnese senior field officer, Muzakkir Manaf, gave permission to start discussions in Malaysia with representatives of the Indonesian government, presumably with instructions from Kalla (Huber 2008). In October 2004, preliminary agreement was reached on the implementation of the special autonomy law for Aceh that had been offered during the CoHA time frame (Aspinall and Crouch 2003).

On December 26, 2004, a massive earthquake in the Indian Ocean generated a devastating tsunami that swept over Banda Aceh, killing more than one hundred thousand residents and leaving hundreds of thousands homeless (Center of Excellence in Disaster Management and Humanitarian Assistance 2005; Matsui and Sato 2007). Up to this point, the hostilities had been ongoing since the breakdown of the CoHA talks. The humanitarian disaster persuaded both parties to prioritize the delivery of disaster relief to Aceh over the military struggle and political campaign conducted by each (Barnes 2009). The need to provide international aid to the tsunami victims likely provided the critical impetus for an external mediation of the conflict (Aglionby 2005).

Factors other than the tsunami do also seem to have held a prominent role in prompting the mediation (Schulze 2007b). Particularly noteworthy were the pretsunami contacts between members of the GAM leadership‐in‐exile in Stockholm and high‐ranking personnel in the Indonesian government in mid and late 2004 (Sillanpää 2006; Huber 2008). But some evidence indicates that the humanitarian crisis generated by the tsunami was critical. On January 28, 2005, prenegotiations, or “talks about talks,” between the GOI and GAM were initiated in Helsinki (CMI 2006; Djuli and Rahman 2008). The prenegotiations concerned the scope of the human cost inflicted on Aceh by the tsunami and the responsibility of the parties to the conflict in the region to respond to the humanitarian disaster. Thus, the first two days (prenegotiations) of what came to be the first of five sets of negotiations between the parties were devoted to humanitarian needs in the aftermath of the tsunami (Center of Excellence in Disaster Management and Humanitarian Assistance 2005).

Just prior to the prenegotiations, the head of the Indonesian military announced the suspension of attacks against GAM forces in Aceh province (Center of Excellence in Disaster Management and Humanitarian Assistance 2005; Matsui and Sato 2007). Some have suggested that the government determined the humanitarian crisis to be the opportune moment to act because it knew already that GAM could not be defeated decisively and thus decided to employ a containment strategy (Schulze 2007b). But whether the government decision was made for strategic or humanitarian reasons, the importance of implementing the shift in military strategy from offense to defense was clear: “the administration adopted the strategy of not crushing GAM but rather co‐opting it into the establishment and rendering it harmless” (Matsui and Sato 2007).

The decision to refrain from “active pursuit” of GAM military units in the field appears to have been driven in large part by President Susilo Bambang Yudhoyono's personal beliefs (and presumably also Vice President Kalla's) regarding effective solutions to the Aceh insurgency — beliefs that had been formed well before Yudhoyono's inauguration (Harish 2004; Shie 2005; Morfit 2007). One observer commented that Yudhoyono “appears open to new approaches in resolving the Acehnese conflict” (Shie 2005). After twenty‐seven years in the Indonesian military, Yudhoyono had become a strong supporter of nonmilitary solutions to separatist conflicts generally, Aceh in particular, and was prepared to offer real concessions to GAM, even before becoming president (Harish 2004; Aspinall and Crouch 2006).

After being elected in the first democratic national presidential elections in Indonesia's history, Yudhoyono took steps to weed out high‐ranking potential spoilers in the military who were believed to be working against a negotiated peace agreement (Aspinall and Crouch 2006; Morfit 2007; Huber 2008). The TNI chief of staff, Ryamizard Ryacudu, had been nominated by outgoing President Megawati to become commander‐in‐chief.

Because Yudhoyono regarded Ryacudu as an opponent to the Aceh peace process and a potential spoiler, he blocked his appointment, retaining instead General Endriartono Sutarto, seen as a peace process supporter, in the post of commander‐in‐chief (Morfit 2007). Yudhoyono seems to have made a political decision to “stand down” against the TNI in the early fall 2004, in the interests of providing credibility to the overtures made to GAM field commanders by Kalla in that period. At least some of the meetings held by government representatives with the field commanders utilized Malaysian good offices (Huber 2008).

GAM had stood fast with its demand for undiluted independence since the inception of the insurgency in 1976. Then in January 2005, twenty months after the previous formal talks with the government and four to five weeks after the December 26 tsunami, a statement was issued from the Kalla‐orchestrated talks that a discussion of the special autonomy concept had begun, although the term preferred by GAM was self‐government (Joyo Indonesia News 2005; Djuli and Rahman 2008). This declaration by GAM had been preceded by the work of Damien Kingsbury, an Australian scholar and political advisor to the GAM leadership and Vice President Kalla, in the fall of 2004.

With the facilitation of the Crisis Management Initiative (CMI), a Finnish‐based NGO, Kalla had made preliminary contact with GAM prior to the December 26 tsunami (Huber 2008). After the tsunami, he asked GAM advisor Kingsbury whether GAM had softened its position regarding independence (Matsui and Sato 2007). Kingsbury wrote that he asked GAM's political leadership “whether it was prepared to accept anything less than full independence. This was predicated upon the first principle question of what independence was intended to achieve. Once having established what purpose independence was intended to serve, it was then possible to ask whether that purpose could be achieved by means other than independence” (Kingsbury 2007b: 102).

In this discussion, which took place in October 2004, GAM let it be known that the group would consider a solution to the conflict other than independence. This softening of its position was made known to Vice President Kalla after the latter made contact with Kingsbury, and informally with senior GAM officials in Malaysia in the aftermath of the tsunami disaster (Matsui and Sato 2007; Schulze 2007c). At this point Kalla asked CMI to formally mediate the negotiations between GAM and the GOI (Huber 2008).

An offer by CMI supported (or underwritten) by the European Union (EU) was accepted by both parties and led to what became known as the Helsinki Peace Process (Herrberg 2008). The first of the five rounds of negotiations, mediated by CMI, began on January 28, 2005. A final draft of an agreement was reached on July 17, 2005, and a Memorandum of Understanding between the Government of Indonesia and the Free Aceh Movement (MoU), or Aceh Peace Accord, was signed on August 15, 2005, formally ending the conflict (GOI–GAM 2005). The agreement mandated that the Indonesian government withdrawal all “nonorganic” military and police forces (not part of a normal military presence to maintain the national defense) from Aceh and that the GAM decommission its armaments and de‐mobilize its troops, with these obligations to be carried out in a parallel fashion.

The EU took responsibility for the formation of the Aceh Monitoring Mission to observe and help manage the agreement's implementation (CMI 2006). The peace that ensued has held, and GAM‐affiliated candidates (in a formal sense GAM has disbanded, but the infrastructure remains, thus the term “affiliated”) fared quite well at the district level in Aceh in 2006 (Aspinall and Crouch 2008). In the April 2009 parliamentary elections, the GAM‐affiliated party, Partai Aceh, won thirty‐three out of sixty‐nine seats in the provincial legislature, the DPRA, the provincial parliament of Aceh (Simanjuntak 2009; Center for Peace and Conflict Resolution Studies 2009).

The insurgency in Aceh has effectively ended (Schulze 2007a; Barnes 2009). The MoU between the Indonesian government and GAM stresses (in Article 1(2)1) the need for “Aceh‐based political parties” to be incorporated into the political process in a timely manner. The EU‐led monitoring mission took responsibility for the decommissioning of weapons, the demobilization of forces, and monitoring the level of human rights violations for an extended period after the peace accord was consummated (Herrberg 2008). Substantial movement toward a resolution of those issues that first gave rise to the insurgency, with particular emphasis on the political participation of the Free Aceh Movement, appears to have occurred in Indonesia.

Of obviously critical importance, however, is that the MoU negotiated by GAM and the Indonesian government and mediated by CMI has not been rejected by the national legislature or Indonesian civil society. In contrast to the actions taken by the Philippine Congress, the Indonesian Parliament in 2006 supported the peace accord and enacted the Law on Governing Aceh (Law no. 11). This act of parliament specified that “Aceh constitutes a special provincial administration,” that is “granted a special authority to manage and administer its local governance” (UNORC 2006). (In large part this language appears to reflect that the province of Aceh had already been well defined in jurisdictional terms since the enactment of Law no. 24/1956 in 1956. That law had reformed and granted to Aceh the status of “special territory,” separating it from the province of North Sumatra [Siregar 2008].) Another critical difference between the Indonesian and the Philippine cases is that the Acehnese form the clear majority of the province's population and there has been no significant in‐migration of non‐Acehnese into the province, whereas there has been massive in‐migration of non‐Muslims into Mindanao (Bacani 2006; Martin and Tuminez 2008).

Clearly, with the summer 2008 onset of the impasse between the Arroyo administration and the Philippines Congress, a disconnect emerged between the two branches regarding the meaning of the term self‐determination. Particularly after legal challenges were mounted against the MoA‐AD, the administration made clear that its understanding of the term did not connote separateness or independence. The administration was prepared to accept a devolution of powers to a degree but insisted that Philippine sovereignty over the full territorial expanse of the state should remain intact.

When the government initially offered the concession of self‐determination in spring 2007, the MILF seemed to see it as a significant step forward (PIA 2007). MILF Chairman Al Haj Murad commented, “It can be a breakthrough. We appreciate this development; we feel it is an advancement in the search for peace” (Gopalakrishnan 2007). But the government apparently saw the concept of self‐determination as more limited, not inclusive of regional control of external defense, foreign affairs, or the monetary system, among other things. Most importantly, partly in response to a court challenge to the agreement initiated by non‐Muslim legislative members and other non‐Muslim elected office holders, the government decided to retain the idea (from the 1976 Tripoli Agreement) that any application of self‐determination would be contingent on “constitutional processes” (ICG 2008a).

Those constitutional processes involved Article 10 of the Philippine Constitution, which stipulates that “no province, city, municipality, or barangay [village] may be created, divided, merged, abolished or its boundary substantially altered, except in accordance with the criteria established in the local government code and subject to approval by a majority of the votes cast in a plebiscite in the political units directly affected” (Bacani 2006). On the basis of this language, the Supreme Court first granted an injunction to the petitioners and subsequently declared the MoA‐AD unconstitutional (Philippine Supreme Court 2008).

Without a ratified MoA‐AD, movement toward a comprehensive compact could not take place. As provided in the agreement, “any provisions of the MoA‐AD requiring amendments to the existing legal framework shall come into force upon signing of a Comprehensive Compact and upon effecting the necessary changes to the legal framework with due regard to non‐derogation of prior agreements . . .” (GRP–MILF 2008). The petitioners argued that a plebiscite (offering 737 villages not already part of the autonomous region inclusion in the Bangsamoro entity) should occur before any ratification of the MoA‐AD, not “within twelve months following the MoA‐AD's signing” (ICG 2008a).

In any case, after the injunction was issued, the Arroyo government stated it was abandoning the agreement and would cease negotiating. The GRP decision to “walk away” from the MoA‐AD almost certainly had as much to do with domestic electoral politics as it did with the court's injunction (ICG 2008a). The MILF said it considered the agreement to be binding, that the deal was completed and they would not renegotiate. Although not approaching full‐scale combat, limited hostilities between government forces and the insurgents had resumed after the injunction was issued; this fighting was apparently instigated by renegade MILF commanders in the field (ICG 2008a). The GRP then announced that it would pursue these units with vigor and would not join new negotiations until the renegade leadership had been eliminated (Philippine Insitute for Political Violence and Terrorism Research 2008b) and would cancel the MoA‐AD (Asia News 2008).

The Malaysian Ambassador to the Philippines, Dato Ahmad Rashidi Bin Hazizi, stated that his government would continue its role as mediator between the MILF and the GRP as long as the parties continued to request assistance from Malaysia and the OIC (Philippine Insitute for Political Violence and Terrorism Research 2008a). With no resumption of talks between the parties, however, the Malaysian contingent of the International Monitoring Team quit the country on November 30, 2008 (GMA News 2008ab; ICG 2008c).

Talks will almost certainly resume at some future point, but when is unclear, given the government's current stance. On its website the MILF has issued the following statement: “For the MILF, there is no way to settle the Moro problem except through the MoA‐AD, which is the framework agreement for the discussion of the solution to this problem and the armed conflict. It determines war and peace in Mindanao . . . It is a prescription for genuine Moro self‐rule without severing ties with Manila” (Luwaran 2008). But after seven years of mediated negotiations between the parties it is the nature of those “ties to Manila” that remains unsettled.

The mix of GRP constituents, that is, members of opposition political parties, print and electronic media, the courts, business associations, religious organizations, and others (Rood 2005), opposed to the ratification of the MoU‐AD could be labeled as spoilers — but with some qualification. These constituents did not have direct access to the means of violence, an attribute that Stephen Stedman has held necessary for what he designates as “inside” and “outside” spoilers (Stedman 1997). Although in Stedman's (1997) typology, inside spoilers utilize stealth to a greater degree than outside spoilers, all spoilers have the ability to use violence in their effort to undermine a peace process.

The constituents drawn from Philippine civil society opposed to the MoU‐AD took a legal path in their opposition. They took a position against what was in their view an agreement that had been adopted because of a “peace at any price” negotiating posture toward the Mindanao conflict on the part of the government. They thus preferred a possible return to warfare, rather than accept an agreement that conceded “too much” (Wanis‐St. John 2006: 138). But the warfare, of course, would have to be conducted by the Philippine military, an outcome that could only be initiated by the GRP, not the latter's constituents, at least not directly.

The GRP constituents in the GRP–MILF negotiations can perhaps be better described as constituents in civil society who “provide the rationale and moral justification for violence . . . and that shape perceptions of what is going on and advocate war as the answer” (Barnes 2006: 20) rather than as spoilers who use violence themselves. This characterization seems even more apt if one agrees that chances of a return to hostilities in Mindanao are significantly higher, now that the outcome of the MoU‐AD ratification process has been rejection or “no‐agreement” (as of April 2010).

Constituents who — while not explicitly advocating war — do not rule out war as an alternative to a flawed peace agreement have been characterized by Thania Pfaffenholz, Darren Kew, and Anthony Wanis‐St. John (2006: 69) as “potential spoilers . . . civil society groups that could have a significant role in undermining an agreement or fomenting instability in the peace building phase.” This definition can help us understand what occurred following the GRP–MILF negotiation. These kinds of spoiler groups have themselves no need to possess a military capability. But as occurred in the MoU‐AD ratification process, such groups are able to potentially deny approval of a peace agreement by the use of legitimate legal and political mechanisms.

If successful in their efforts, these government constituents can create a “no‐agreement” condition and become more than just potential spoilers. Such a condition need not necessarily bring renewed hostilities between the parties, but is not strongly averse to them either. These groups find engaging in renewed hostilities to be preferable to accepting what they perceive to be a deeply flawed peace agreement.

Negotiating the construction of a new agreement remains a third alternative. But when rejection of the signed agreement brings substantial risk of renewed hostilities, the spoilers response may be “so be it”— demonstrating “risk‐acceptant” behavior (Levy 1997; Lounsbery and Pearson 2009: 36). The likelihood of renewed hostilities in Mindanao as a consequence of MoU‐AD rejection is unclear, but certainly greater than zero.

In the GAM insurgency, the prerequisites for obtaining an offer to mediate and for a high likelihood of mediation success were in place before CMI intervened. The Indonesian government, prompted in part by the tsunami humanitarian disaster, made the strategic decision to forego defeat of the GAM insurgents in favor of containment. Prior to, and during, prenegotiations, GAM made clear its willingness to engage in talks with the Indonesian government within the framework of “special autonomy,” that is, self‐government. Kalla then contacted GAM privately through CMI intermediaries. After becoming certain that GAM's stance had softened (at least partly as a consequence of the tsunami), with the aid of GAM advisor Kingsbury, the vice president began full‐scale negotiations (Matsui and Sato 2007).

The Indonesian government made the approach to CMI in a manner that was similar to Philippine President Arroyo's overture to the Malaysian government, although the Indonesian approach seems to have been conducted much more privately. But a key distinction between the two third‐party interventions occurred in the period just prior to the beginning of external mediation. In the Indonesian case, GAM became open to a form of self‐governance that was less than full independence, and the Indonesian military ceased offensive operation against the GAM. In the Philippine case, the Philippine military had adopted a defensive posture toward the MILF. Although the MILF had indeed ceased insisting on a need for a separate state, it had not abandoned a particular variant of self‐determination as a demand, that is, “self‐governance,” which for the MILF meant a “homeland” with a territory delimited by cultural, demographic, and historic factors (Bacani 2006; Tuminez 2007; Martin and Tuminez 2008).

Clearly, for a separatist insurgency, territory is a critical issue. But for GAM, the issue of delimiting the territory over which the movement would exercise self‐determination had already been established back in 1956 when the Indonesian Parliament enacted Law no. 24/1956, which reconstituted Aceh as a distinct entity apart from North Sumatra. (Aceh had been amalgamated with North Sumatra in 1951, following the country's transformation in 1950 from a federal to a unitary state [Kingsbury 2007a; Siregar 2008].) Law no. 24/1956 granted Aceh the status of “special territory,” and in 1959 the national government had designated Aceh as having a “special administrative” status (Kingsbury 2007a).

In 2002, the Indonesian Parliament passed Law no. 18/2001 on Special Autonomy for Aceh. But the law contained a major flaw: members of insurgent groups faced restrictions in their ability to participate in elections (McGibbon 2004). A key to gaining agreement on the MoU was that it lifted this restriction. As a consequence, GAM members won significant victories in the 2006 Aceh provincial elections, including the governor's office (Aspinall and Crouch 2008).

The larger point is that the insurgent conflict in Aceh was not a struggle over competing claims to contested territory. The Aceh conflict was concerned with the degree of sovereignty, or authority over an agreed‐upon fixed territory, which the national government would be required to relinquish in order to gain an end to the hostilities and for the Acehnese people to gain self‐determination. Thus, prior review of the MoU did not occur because as Vice President Kalla stated, the Parliament did not “need to know”— the conflict was internal, not external, and, according to Kalla, Parliamentary approval was not necessary under Indonesian law (Kalla 2008).

This is not to say that groups within Indonesia's government outside of the executive were entirely subservient to the vice president's understanding of Parliamentary prerogatives. Several parties (the Indonesian Democratic Party‐Struggle, Sukarnopatri's party, and the National Awakening Party of Abdurraman Wahid) complained that the government had failed to consult them regarding the MoU (Miller 2009). The party members argued that the house branch of the national legislature should have been appraised of MoU specifics before the document was signed. The accuracy of the vice president's views notwithstanding, prior review by the GOI's constituency was not imperative in the government's view, because of a preexisting wide agreement regarding the territorial expanse of an Acehnese homeland.

The Mindanao conflict in the Philippines concerned a devolution of authority as well. But for the MILF, self‐determination for the Moro population could only be achieved through a recognition by the state of territorial claims synonymous with a homeland shaped by its historic, cultural, and religious experience. For the MILF, gaining a homeland required the acquisition of an expanse of territory in Mindanao that would bring recognition of, and promote, Moro distinctiveness, but not necessarily separateness (Gershman 2001). The notion of homeland inherently connotes not just geographical division, but differences in political and cultural behavior by the population therein, apart from the population outside the homeland. This behavior has been labeled “territoriality,” which “describes not [a] simple designation of space, but the behavior patterns which emerge from the political use (the occupation) of territory. Territoriality thus requires territory to be definitively bounded rather than loosely so, and thereby creates the conditions for the physical exclusivity of people . . . [Territoriality] requires the control of territory as the very basis of legitimacy” (Bishai 2004: 62).

When the Philippine government granted the MILF additional territory beyond the autonomous region with greater autonomy, in the view of some segments of the general population, it risked the creation of “a state within a state” or what has been referred to as an “insurgent state” (McColl 1969). Such an entity claims or demands partial sovereignty or a devolution of authority over its autonomous area of governance. Self‐governance for the MILF included claiming a certain expanse of territory in Mindanao beyond the existing autonomous region with a significant Muslim population. In the MoU‐AD, the Arroyo administration conceded the expansion of the autonomous region into this territory to become part of the “Bangsamoro homeland.”

Such a concession may challenge the results of laboratory experiments that suggest that negotiators held accountable to constituents are no more likely to compromise than negotiators not so accountable (Breaugh and Klimoski 1977). But it is unclear in this instance whether the GRP negotiating team fully understood the degree to which it was accountable to a national constituency, although it has been noted that the Philippine negotiators were not oblivious to potential constitutional challenges to provisions in the MoU‐AD granting an expanded Moro ancestral domain or homeland (Tuminez 2007). Ultimately, the Philippine Supreme Court determined this was a concession the Arroyo government did not have the authority to grant under the constitution.

The impasse in the Mindanao conflict remains, while the Aceh conflict has been reduced to political argument conducted within electoral processes (ICG 2008b). It should be noted that GAM candidates fared quite well in the December 2006 municipal and district elections (Aspinall and Crouch 2008). The mediators conducting the Indonesian conflict were able to seize an opportunity created when the GOI altered its military strategy while GAM revised its primary political demand as part of a peace process ongoing both prior to, and in the aftermath of, natural disaster.

In the Philippine conflict, the government also amended its military strategy. But the MILF's alteration of its primary demand was conditional. With the aid of the Malaysian mediation, the MILF revised its central demand from secession to self‐determination, understood as self‐governance. The GRP and the MILF were able to agree on this view of self‐determination in the form of the MoU‐AD. But a significant (non‐Muslim) portion of the electorate and members of the Philippine Congress raised objections to that agreement and petitioned the Supreme Court successfully to effectively void the agreement by ruling it unconstitutional. An opportunity for a more thorough advance review of the MoU‐AD by the GRP negotiating team's constituency may have avoided this outcome.

Constituent review was not completely absent in the Philippines peace process (Santos 2004; Rood 2005). But as Steven Rood (2005: 19) has pointed out, “Christian (or settler communities) opposed to the accommodation of Muslim aspirations are not well represented in [Philippine] civil society.” What Rood (2005: 19) is suggesting is that although there is significant opposition to concessions in peace agreements on the part of “average citizens,” in organized (Christian NGOs) civil society, there is support for the peace process. Soliman Santos (2004: 75) has argued for the “further development of Mindanao/Bangsamoro civil society participation” in the peace process.

Opposition political parties, holding legislative seats or not, could also be considered part of civil society. In her formulation of civil society composition, Catherine Barnes (2006: 20) appears to accept this view: “[i]n democracies, civil society interfaces with the state through parliamentary institutions (with parliamentarians often seen as serving a bridging role as the elected representatives of civil society).” Even when the loyal opposition is not part of the “government,” that is, its representatives are not present at the negotiating table, they must nonetheless be counted as part of the government's constituents and would certainly qualify as interested citizens. The point is that opposition legislators have the ability to act as spoilers and negotiators must consider their perspectives and reactions when negotiating with insurgents.

In fact, it was representatives in the Philippine Congress from the Sarangani, Cotabato, and South Cotabato provinces in Mindanao — each with sizable non‐Muslim populations — who were among the most active voices mobilizing against the MoU‐AD. The participation of communities from those in the Track Two (less formal, back‐channel) component of the peace process (Putnam 1988; Wanis‐St. John 2006) could have increased the likelihood of MoU‐AD participation.

The back‐channel “alternate negotiation path” offers an opportunity for greater civil society and constituent participation in a peace process. In the Philippine process, some of the government's constituents seemed to be not overly concerned about the costs of no‐agreement with MILF. If that is true, then the “win‐set” of the GRP negotiating team has been effectively reduced. Robert Putnam (1988) has defined the win‐set of a Level Two constituency as “the set of all possible Level One agreements that would ‘win’— that is, gain the necessary majority among the constituents when simply voted up or down.” Stated differently, this means the number and “magnitude” of the concessions the GRP negotiating team could offer to the MILF was smaller than the GRP team had originally estimated.

In Indonesia, parties concurred that compounding the Aceh population's suffering with additional casualties from combat should be avoided (Aglionby 2005). Thus, GAM was prepared to discuss something less than complete independence as an outcome to the conflict. The recognition by the adversaries of the humanitarian disaster's scope and the mutual recognition by the parties of the way the disaster could damage their own goals led to the objective referents of an altered Indonesian military strategy and a revised primary political demand from the Aceh insurgents. Neither party wished to be viewed as having taking actions that at best prevented the most effective disaster response and at worst even added to the magnitude of human suffering. Each party had been vying for the support of the population, with GAM espousing nationalist motivations for the rebellion and the GOI advocating a more fundamentalist Islamic practice for the population (Kingsbury 2007a).

Either party's continuation of hostilities in the face of the humanitarian disaster would risk losing substantial popular support. In the perception of both adversaries, the aftermath of the tsunami was a moment affording “a way out” (Zartman 1995). Martti Ahtisaari, the lead mediator in the GOI–GAM peace process and the 2008 Nobel Peace Prize winner, has said as much. “Certainly the conflict was more ‘ripe’ in 2005 than in 2000 . . . With all the skill in the world, I cannot make a mediation process ‘pregnant’ in the sense of bearing the fruit of peace. The parties themselves must do this” (Ahtisaari 2008a, 2008b).

The cases examined in this article provide only limited support for my investigative thesis. To reiterate, if the state alters its military strategy against an insurgency from offense (pursuit) to defense (of civilian populations), and the insurgents revise their demand from secession to self‐determination acquired through homeland autonomy, achieving a mediated agreement becomes more likely. But the implementation of a negotiated agreement is uncertain if self‐determination is defined in territorial terms prior to adequate constituent review.

In order to gain an agreement, in each case it was critical for the government to make a strategic decision to abandon offensive operations against the insurgents and for the insurgents to adopt self‐determination as their primary goal, instead of independence. But for separatist movements, self‐determination has a territorial component. To make implementation of a negotiated agreement more likely, national constituents should achieve an adequate prior appreciation of the territorial component's extent.

A lesson learned from this study would appear to be that democracies engaged, or about to be engaged, in negotiations with an insurgent movement should develop a degree of prescience or foresight with regard to the likelihood of civil society objection to a proposed agreement to end hostilities. How to achieve this foresight would not be a simple matter, but it likely would include establishing “links” between the government (the executive) and organized civil society groups.

A legislative opposition could possibly serve as such a linkage, contingent on whether the organized civil society groups and opposition legislators had roughly the same base, for example, region of origin, religious affiliation, ethnicity, or socioeconomic rank. Recognition and use of these bonds — which may fuse into coalitions — by the executive could be helpful. Bringing about the adequate prior review by civil society of a proposed agreement, particularly regarding the territorial expanse of a proposed homeland, might then become more attainable.