Scholarly article on topic 'Path Dependence in Settlement Processes: Explaining Settlement in Northern Ireland'

Path Dependence in Settlement Processes: Explaining Settlement in Northern Ireland Academic research paper on "Political Science"

0
0
Share paper
Academic journal
Political Studies
OECD Field of science
Keywords
{""}

Academic research paper on topic "Path Dependence in Settlement Processes: Explaining Settlement in Northern Ireland"

doi: 10.1111/j.1467-9248.2007.00623.x

Path Dependence in Settlement Processes: Explaining Settlement in Northern Ireland

Joseph Ruane Jennifer Todd

University College Cork University College Dublin

The recent literature on path dependence provides a model that can be used in explanation of ethnic conflict and settlement processes. Using Northern Ireland as a case study, this article identifies path dependent patterns of conflict embedded in long-term processes of political development whose change may interrupt these patterns. It highlights the importance of long-term state trajectories in constituting and reproducing these patterns, the generation of 'endogenous' processes of change and the impact of wider geopolitical processes in strengthening these. It shows how and why factors such as power, perception, networks and institutions vary in their impact on conflict and explains when they work together to produce settlement.

Internal communal conflicts vary so widely in their form that valid generalisations are not easily found (Brown, 1996; David, 1997). Settlements are even more difficult to explain, for they vary not simply in the type of conflict that they settle, but also in their status and stability (Darby and MacGinty, 2003, pp. 1-6). Much contemporary scholarship searches inductively for recurrent proximate factors (or clusters of such factors) that explain settlement in a particular range of cases (Brown, 1996; King et al., 2005; Stedman, 2003). Other scholars proceed by applying general theoretical models drawn from international relations theory to settlement processes (David, 1997; Hauss, 2001; Lake and Rothchild, 1998). Both approaches assume that similar causes lead to similar effects, that law-like generalisations can be found, at least if we suitably delimit the range of cases, and that contemporary conditions are crucial in the causal process. But while existing work has identified a range of factors relevant to conflict and settlement (Brown, 1996, p. 14, p. 577), it has not found substantive law-like generalisations.We argue that a different research agenda may be more fruitful, one that takes a historically sensitive approach to political phenomena and places notions of'path dependence, critical junctures, sequencing, events, duration, timing and unintended consequences' (Pierson, 2004, pp. 5-6) at the heart of the analysis. Our concern is not simply with what factors are important for settlement, but when, and why at that time and not at others. As we show below, this is not a rejection of comparison or generalisation but a shift in its focus.

In what follows, we do not attempt to deal with the entire literature on settlement processes, but rather focus on four empirically promising and theoretically powerful approaches that highlight, respectively, relations of power, cognitive framing,

© 2007 The Authors. Journal compilation © 2007 Political Studies Association

«alk. Political Studies

IriHi Association

social networks and institutional opportunities.1 Using the test case of Northern Ireland, where a protracted conflict was brought to an agreed (if unstable) settlement in 1998, we show the limited explanatory power of each of these factors when taken out of its temporal context. A more adequate explanation requires that we open a space for process, temporality, positive feedback, the path-dependent 'lock-in' of patterns and the modes by which this may be interrupted (Mahoney, 2000; Pierson, 2000).

Patterns of persistence and continuity in social life may be said to be 'path-dependent'when they are 'developmental trajectories that are inherently difficult to reverse' (Hacker, in Pierson, 2004, p. 21), produced by positive feedback patterns. Ethnic conflict exemplifies many of the factors associated with such trajectories (David, 1994; Pierson, 2004, p. 24). In situations of ethnic conflict there are high set-up costs (much is invested collectively and individually in setting up institutions of ethnic dominance, in organising against them and in emphasising the ethnic aspect ofidentity); learning, coordination and anticipation effects (networks, language, symbolism, shared understandings); and these are interrelated in a whole matrix of socio-political institutions (Ruane and Todd, 2004; Wimmer, forthcoming). Indeed ethnic conflict provides an arena in which to explore questions raised but not resolved in the literature on institutional path dependence: questions of interdependence, intersection and interaction between path-dependent processes of different degrees of generality,'lock-in' and temporal duration (Mahoney, 2000, pp. 529-35; Pierson, 2004, p. 27). It invites analysis at once of the wider institutional matrix of ethnic dominance and contest which locks in propensities to conflict and of particular shorter-term patterns of intense and violent conflict. We argue below that the interrelation of the two processes can make settlement particularly difficult to achieve.

Claims that a process is path-dependent require a definition of the pattern or path that is reproduced (that which remains constant in and through other changes), identification of the 'critical juncture' when this path was taken rather than another possible path, delineation ofthe mechanisms ofreproduction and the ways in which pressures for change are deflected or absorbed. Most particularly, for the purposes of this article, they also require analysis of how path-dependent trajectories come to an end, how the mechanisms reproducing the patterns are interrupted and how radical change - in this case settlement - is possible.The path-dependence literature is sometimes criticised for supposing that change can only be produced by exogenous shocks, while in fact institutional transformation may be endogenously and incrementally produced (Cortell and Peterson, 1999; 2001; 2002; Crouch and Farrell,2002;Streeck andThelen,2005;Thelen,2003). Ethnic conflict patterns are particularly resilient to incremental change, but, as we show below, settlement involves an interrelation of endogenous and exogenous factors.

Using the Northern Ireland case we identify path-dependent patterns of conflict embedded in turn in long-term processes of political development whose change may interrupt these patterns. The case study highlights the impact of long-term

state trajectories in constituting and reproducing these patterns and the impact of wider geopolitical processes in strengthening 'endogenous' tendencies to change. This resembles a model of institutional change in which exogenous shocks open 'windows of opportunity' which may or may not be taken depending on actors' policy preferences and the distribution of power resources (Cortell and Peterson, 1999). However, the endogenous/exogenous distinction is not clear-cut in this case, for state trajectories, seemingly external to the short-term pattern of conflict, are internal to the long-term one.

Competing Explanations of Settlement Processes

The most plausible and fruitful explanations of settlement tend to prioritise one of the following four factors: power relations, cultural distinctions and cognitive frames, social networks (and their role in forming community boundaries) and institutional opportunities.

A focus on changing power relations has typified rational choice and neo-realist models of ethnic conflict and settlement processes (David, 1997; Lake and Rothchild, 1998).These approaches typically take as given the self-definition of the actors and their categorisation of their aims, and focus on their resources and strategies. Power is the key resource, and if power instability gives incentive for conflict, power stalemate - where actors can prevent each other from attaining their ends - gives incentive for settlement; settlement in turn is unstable without a credible guarantee that the weakening of one party's resources by compromise (disarmament etc.) will not be exploited by the other (Lake and Rothchild, 1998; Walter, 2001; Zartman, 1989). The clarity and relevance of these arguments in highlighting factors that foster or subvert the quest for settlement are clear. But so too are their limits. In particular, actors' perceptions and categorisations - for example, the distinction between different orders of time, awareness of power stalemates, calculations about short or long-term shifts in the power balance - are insufficiently brought into this explanation (Zartman, 2003).

A focus on cultural distinctions and cognitive frames is typical of'new international relations' theories and cognitivist approaches to ethnic conflict. Analyses of micro-processes of conflict study how and when actors adopt or reject ethnic labels and link or de-link ethnic categories to personal dignity (Brubaker, 2002; Kakar, 1996; Petersen, 2002). Some argue that the central factor in moving from conflict to settlement is cognitive shift - the recognition that there is a potential win-win situation - and that this can occur at any stage of conflict (Hauss, 2001, p. 218).Yet why such re-categorisations occur when they do is seldom adequately explained. Whether the emphasis is on political entrepreneurs as key agents in the process of collective redefinition (Brubaker, 2002; McAdam et al., 2001) or on exemplary micro-interactional contexts (Arthur, 1999), neither the timing of change nor its extent is fully explained. This approach points to one set of mechanisms relevant to settlement processes but leads to further questions of why and when they become effective.

A third approach focuses on civil society and social networks. Institutionalised interactions across boundaries are shown to lessen the propensity for conflict (Varshney, 2001), while global civil society constitutes an arena where linkage politics and 'people power' can replace self-seeking ethnic and national politicians (see Albrow, 1996, pp. 176-83). Network-oriented analyses link the cognitive with the interactional so as to see how communal boundaries are constituted as 'bright' or 'blurred' in everyday interaction, and how this defines the range of choices that actors face (Alba, 2005; Wimmer, forthcoming). These approaches build variously on theories of'bridging' social capital (Putnam, 2003), the role of civil society in transitions to democracy (for discussion, see Keane, 1998) and theories of how boundaries and distinction are challenged and maintained (Lamont, 2000). However, while these approaches can explain constraints on conflict and local variations in its form, they do not explain the move to settlement (see Farrington, 2004).

A fourth approach gives particular attention to institutions - whether specific institutional forms or longer-term institutional configurations - and the opportunities they may give for compromise settlements (Coakley, 2003; Keating, 2001; McGarry and O'Leary, 2004). In general, new institutionalists assume that 'the institutional organisation of the polity and economy structures conflict so as to privilege some interests while demobilising others' (Hall and Taylor, 1996, p. 937). However, this raises the further question of the relationship of institutions to the wider societal context (including cognitive frames, power relations and informal networks) which may subvert formal institutional rules and 'convert' old institutions to new functions, or new institutions to old (Thelen, 2003, pp. 228-33). Subversion of formally democratic and intentionally egalitarian institutions for particular ethnic purposes is common in ethnic conflicts. Where institutional analysis sticks narrowly to specific institutional design, it is insufficient to grasp this; where it broadens to analyse the institutional context, it is compatible with the approach suggested here.

Each of these four sets of factors gives incentive for settlement. Yet the effect of each is not constant.Whether or not institutional changes or a new power balance give overriding incentives for settlement depends on the wider context, including the presence or absence of the other factors.This suggests the need for a historical and contextual approach concerned not only with what factors are important but when they are important, and how the former may depend upon the latter. In this article we focus on a case where these issues arise with unusual clarity: Northern Ireland, where a settlement was finally reached after a quarter-century ofconflict.

Northern Ireland: Explaining Settlement

Agreement was reached between most of the main political parties in Northern Ireland on Good Friday, 10 April 1998. What were the conditions which brought agreement and what explains its timing? How did unionists (at least a majority of

them) reach agreement with republicans (extreme nationalists who had only recently ended a campaign of violence) when six years earlier they had failed to reach agreement with moderate nationalists (see Bloomfield, 1998; 2001)? How did republicans come to accept a settlement that, formally at least, appeared to offer fewer constitutional gains than did the Sunningdale Agreement, which they had violently rejected a quarter-century earlier (Wolff, 2001)? How could actors who in the recent past had declared conflict, and even violence, to be inevitable, a rational and justified response to threat, injury and insult, now begin to speak as if it were unthinkable, irrational, a feature of a distant past?

Each of the four approaches discussed above finds echoes in the literature on the 1998 agreement. While scholars typically look to multiple factors in explanation of conflict and agreement, the quest for explanatory clarity often leads to a prioritisation of one set of factors over others. The factor most often cited in explaining the Good Friday Agreement (GFA) is power, and in particular power stalemate: a military stalemate between the Irish Republican Army (IRA) and the British army which gave impetus to peace; and a political stalemate between unionists and nationalists which gave impetus to negotiations. For example, Robert English (2003, pp. 307-13) takes republicans' inability to fulfil their goals at either the military or the political level as the main factor motivating them to call a ceasefire and to negotiate a settlement (see also Schultze, 1997). Stalemate was certainly important in providing the conditions for a settlement: if any parties had been able to achieve victory they would have grasped the opportunity. But it was far from a sufficient condition as neither the military nor the political stalemate was new. Some republicans had perceived a military stalemate as early as 1975; others did so from the mid-1980s (English, 2003, p. 307). Political stalemate was also intermittently evident from the 1970s. For most of the period, however, republicans and unionists used temporary stalemate in one field to play for victory in another. If stalemate became a key condition of settlement in the 1990s, this demands further explanation.

Another approach to explaining the 1998 agreement puts ideological, discursive and cognitive change at the heart of the analysis. By the 1990s virtually all the parties had adopted new political discourses, incorporating current international concepts of pluralism, rights, parity of esteem, recognition and equality (Bourke, 2003; Coakley, 2002).Was settlement then a product of shifting ideological frames that allowed actors to recognise the possibility of compromise? There was a change in the dominant language of politics and for some actors - in particular republicans - it was very radical (Ruane, 2004).Yet two issues arise. First, since the new language had been available since the 1970s (articulated by John Hume and by the Alliance party), the timing ofthe shift for republicans and unionists remains to be explained. Second, its precise significance and its causal import require clarification. Did shifting ideological frames allow successful negotiations, or did they follow from the success ofnegotiations and remain contingent on the success of the settlement? The new political language does not itself answer the question

since it was, on occasion, used not as an indication of intent but as a strategic resource in the pursuit of conflicting interests (Ruane and Todd, 1996, pp. 99-108; 2003).

Another explanation focuses on the build-up of social networks and civil society organisations that blurred communal boundaries and enabled the public to see beyond ethnic divisions. It was the build-up of social capital and 'people power' that ultimately impelled politicians to negotiate a settlement (Guelke, 2003). In the same vein, some explain the crises of implementation of the Good Friday Agreement by the marginalisation of civil society and the prioritisation of nationalist and unionist politics in the institutions of the Agreement (Wilson and Wilford, 2003). While there is evidence that cross-community civil society organisations can restrain conflict and even provoke identity shift in settlement-sustaining directions (Darby, 1986; Hargie and Dickson, 2003; Todd, 2006), their importance in producing and underpinning settlement in this case is open to question. First, there was a relatively small increase in institutionalised contexts of cross-community interaction in the decade preceding settlement.2 Second, voting behaviour since the Agreement, far from pressing the parties to maintain the Agreement, has rewarded those parties least likely to do so.3

Finally, a major strand in the literature focuses on the institutions of the Good Friday Agreement: consociational, egalitarian and neo-federalist, with 'double protection' for minorities in either British or Irish jurisdictions (McGarry and O'Leary, 2004). Without doubt, the institutions are an advance on any previous initiative and have allowed opposed political parties to cooperate in government. Yet the new forms of democracy are not themselves a sufficient explanation of agreement. Where the Good Friday Agreement of 1998 differs most radically from the failed Sunningdale Agreement of 1973 is less in the form of representation or in the arrangements for North-South cooperation but in the context -the increased electoral and demographic strength of nationalism, the integration of a now electorally strong Sinn Féin into government and the shared British-Irish strategy of conflict management (McGarry and O'Leary, 2004, pp. 260-99; Ruane andTodd, 1999; Wolff, 2001).The institutional provisions were effective in producing agreement because of the context in which they were embedded.

Each of the factors discussed above played a role in the emergence of a settlement and all serious interpretations of the Agreement acknowledge them. Our argument is that we can better understand the precise role that they played when we adopt an approach sensitive to history and process in which such concepts as path dependence, critical junctures, duration and timing are highlighted.

Path Dependence and the Conditions of Settlement in Northern Ireland

How did settlement come about? What happened in the 1990s to make it possible? Why did factors which had been present for some time without leading

to settlement now work together to make settlement appear ever more rational and probable? The answer lies in a change in the context in which they operated. That context consisted of a long-term path-dependent pattern of relationships within which a particularly intense and violent pattern was locked in for three decades in Northern Ireland (1968-98). Change came through the long build-up of endogenous change, intensified by geopolitical change. To use the concepts of Andrew Cortell and Susan Peterson (1999), exogenous change produced windows of opportunity in terms of which actors - whose new power resources had been the product of a slow-moving process of change - could begin to break the patterns. In this case, the exogenous geopolitical changes also affected the preferences of the actors, and speeded up slower processes of state change already under way. In what follows, we trace the long-term pattern and we show how the recent phase of intense and violent conflict fitted within this and how the complex path-dependent patterns were disrupted.

We have argued elsewhere that the Irish/Northern Irish conflict is driven by a system of relationships that emerged at a particular historical conjuncture and then reproduced itself over time (Ruane and Todd, 1996). The parallels between this interpretation and a path-dependence one are clear: in both cases there is a moment of crystallisation and then oflock-in.The moment of emergence was the century-and-a-half of political, religious and settler-native conflict that accompanied the integration of Ireland into the English state in the sixteenth and seventeenth centuries.The moment of crystallisation - of no return - was the end of the seventeenth century which saw the definitive victory of the British Protestant minority in Ireland. It had political and military control over the country, its position was underwritten by the English state in return for its loyalty and the majority Catholic population of largely indigenous stock was economically and politically subordinate and culturally marginalised.

The resulting stark power disparity and inequality locked in and partially fused a complex set of oppositions (religious, ethnic, cultural and colonial) and created a situation where rational self-interest (for security or economic livelihood or influence) led individuals to band together as Protestants or as Catholics. Those with the power to drive change had every reason to maintain the status quo: once the Crown had committed itselfto this manner ofrule, any shift to a different one would be costly with no guarantee of a successful outcome. The pattern of relationships had properties of increasing returns, since it formed political and cultural identities as oppositional and interests as conflicting. The exclusion of Catholics was justified by their disaffection, and exclusion reinforced disaffection. Once formed into communal blocs, individuals' security and well-being came to involve communal solidarity. Interests, emotions and values converged: socialisation patterns locked 'human capital' into the pattern, engaging emotions and values, mobilising religious fears, producing a tendency to 'essentialise' differences and mobilise different groups into totalising communities (Farrell, 2000; Hirst, 2002; Wright, 1996). What had crystallised was a system of relationships based on

cultural difference, power relations and communal belonging with each of these reinforcing the others and being reinforced by it in return. There were strong incentives for individuals to reproduce the pattern, and strong disincentives to step outside it.

The conflict-generating pattern long outlasted the colonial period; it was embedded in British state institutions in Ireland, it was embodied in the habitus of both communities and it became the accepted mode of territorial management whereby the British state relied on the local loyal community for administration and security. Recurrent attempts to undo the system - either by reform from above (Union in 1801, land reform in the late nineteenth century) or rebellion from below (1798,1848,1916,1920-1) - changed elements but left the relations intact (Ruane andTodd, 1996).The pattern continued after partition in Northern Ireland where the broad frame of governance - British rule with its implications for unionist power, cultural capital and security - directly affected communal interests more than did any particular policies or institutions. Constitutional conflict was therefore highlighted at the expense ofpolicy change or institutional compromise. This system produced an endogenous source of change: Catholics. Each rise in Catholics' power resources - as political and cultural modernisation redistributed power downwards - led them to challenge the existing institutional order (Hirst, 2002; Ruane and Todd, 1996; Wright, 1987, pp. 1-20; 1996, pp. 1-22). The mode of that challenge - power struggle within the cultural and communal terms of the seventeenth-century settlement - intensified conflict.

A new phase of conflict began with the civil rights movement in Northern Ireland in the 1960s. It fits the general pattern of challenge and reaction outlined above (Ruane andTodd, 1996, pp. 124-8). However the intense and violent form of conflict that ensued - a radical politicisation and militarisation of republican working-class communities and of loyalist areas, with polarising effects on the entire population - was not inevitable. It was the product of a series of contingent events in the early years: unionist miscalculations, British misjudgements (not least internment) and the absence of a coherent nationalist political strategy.

Once set in motion in 1969, a process of intensification and polarisation sharpened identities, closed communal boundaries and made every sector ofthe society an arena of power struggle. The extremes used each other's behaviour to justify their own, increasing the numbers of their active supporters, keeping the middle ground of their community ambivalent and making impossible any form of compromise settlement. A range of ritualised interactions - public rioting, symbolic violence, intimidation and murder - hugely reinforced the unwillingness to compromise. But it was not simply that: each side was aware that bigger issues were at stake, that long-term individual interests and self-respect were interrelated with collective power and state control. That longer-term pattern to Irish history was perceived (from opposite perspectives and with different emphases) by republicans and loyalists, nationalists and unionists, and it gave historic depth and meaning to the concrete experiences, injustices and violence that they experi-

enced in a society at war. Republicans did not want immediate gains; they were determined to make this the final phase of the long historic struggle. Unionists and loyalists were unwilling to make the concessions that were demanded even by the most moderate nationalists for fear that if they lost ground now, it would allow nationalist advance in the future. This was not mythical thinking, but situated rational self-interest informed by a long-term perspective. It worked with, and reproduced, the cultural and communal categories formed over the long term. As we have seen, it subverted all movements towards settlement. If there was to be a settlement, not only would institutional opportunities for safe compromise have to be provided in the short term, but the longer-term pattern of relationships would have to be breached decisively.

What had changed by the 1990s? A long, incremental, slow-moving process of augmentation of Catholic power resources had reached a threshold where neither repression nor concession could bring stability, and where both republicans and unionists began to have incentive for settlement. This intensified the process of British repositioning that had been ongoing, although very uneven, since 1969, which was facilitated and further intensified in a context of exogenous geopolitical change. The shift in the role of the state(s) broke key aspects of the long-term pattern and opened a 'window of opportunity' that the parties were able and willing to use to reach an agreement.

From the start of the current round of'troubles', nationalist pressure had provoked an uneven process of British reform, while the slowness of the process and the extent of repression had hardened nationalist (and particularly republican) resolve. Only by the 1980s did the British government recognise the extent of nationalist alienation and embark on a path to resolve it, breaking with the old pattern of territorial management, taking the Irish state, rather than the unionist community, as their local partner and intensifying a programme of reform. The process strengthened nationalists institutionally, culturally and economically, with a knock-on demographic impact. By the 1990s, republicans were aware of the new power resources potentially at their community's disposal and were contemplating an unarmed strategy, while unionists wished to negotiate a stronger position for themselves within the union. But there was no plausible form of agreement that did not weaken one or other party in the longer term. Meanwhile both Irish and British governments saw that more radical change was needed to bring stability in the short term, and still more in the long term when demographic change might provoke further crises.

The process of state repositioning that began with the Anglo-Irish Agreement of 1985 had potentially radical effects. By the 1990s, the British government explicitly committed itself to a political reconstitution of Northern Ireland, in partnership with the Irish government.4 Northern Ireland would be reconstituted as a distinctive kind of European region, one that comprised two peoples pledging allegiance to two different states, and whose conflicting aspirations could be accommodated by cross-border arrangements and overlapping institutions and an

input from outside mediators (the US and the EU), with egalitarian and bi-national policies and with sovereignty decreasingly important in the daily run of affairs.5 The change in the British trajectory was given impetus by the wider geopolitical context. Its trigger was the imperative to reconfigure state institutions to succeed in the new global economy (Krieger, 1999) and, more immediately again in the 1990s, to restore the special relationship with the US. Within Britain this entailed a restructuring of the British polity (devolution-all-around, multi-culturalist reconfigurations of'official' British identity) that also impacted directly on the public culture in Northern Ireland and on the interactional patterns deemed acceptable in institutional settings: traditional unionist notions of the sovereignty of the 'Crown in Parliament' and the hope of full integration into the United Kingdom were becoming politically irrelevant.

There was a corresponding Irish dimension. The Irish government's post-1950s strategy of openness first to economic and later to cultural and political impacts had developed, by the 1980s, into a distinctive self-conscious project of adapting to the new global environment by playing off powerful British, US and European forces against one another to national advantage. This led to major socioeconomic changes: the 'Celtic tiger' economy (O Riain, 2004), while secularisation weakened the entanglement ofreligious and cultural differences and interests with national categories, encouraging some unionists (most prominently the business class) to reconsider their relationship to the Irish state. There was also a direct impact on strategy towards Northern Ireland: sovereignty was no longer of core importance, borders were permeable and economic interests paramount; Irish unity came to mean ever greater island-wide integration while keeping an openness to Britain in both parts of the island (Hayward, 2004). Through the 1970s and 1980s, the Irish state poured major diplomatic resources and energy into promoting this as the way to settle the Northern Ireland conflict and their rhetoric of European regionalism became the language of the peace process in Britain, Europe and the US (FitzGerald, 1991).

These changes gave a rationale and a possibility for an agreement that was already in the strategic interests of both republicans (who stood to gain influence and build more power resources) and unionists (who needed to negotiate a safe place for themselves within the new United Kingdom).The new state strategies and the 'peace process' itself were justified in terms of emerging international norms which no longer held borders sacrosanct (Guelke, 2002).There is debate on the direct effects of international influence on party strategies in Northern Ireland.6 Indirectly, however, the wider geopolitical context was crucial in legitimating and guaranteeing the changes in state strategies for the actors in Northern Ireland: that British (and Irish) state repositioning was a rational response to a new international order, and seen as such by major actors like the US, showed that the change was set to deepen and broaden for the foreseeable future, indeed to be the future. The political actors in Northern Ireland intuitively thought in terms of long-term patterns of conflict: these patterns were now visibly changing. Classic

tenets of republicanism (that the British government was an imperial presence in Ireland which could not be negotiated with; that only violence would persuade it to withdraw) were put in the balance, jettisoned or overturned (Ruane, 2004). Unionists - at least leading groups within the main Ulster Unionist party -recognised that to secure the Union they had to accept that it would inevitably be looser, with a changed form and cultural substance and a different concept of Britishness, and that an Irish dimension was inevitable (Aughey, 1999; Farrington, 2006; Patterson, 2004). This also changed the functions and significance of institutions and ideas that most of the parties had long rejected: consociationalism took on a new appeal for enough sections of each community to permit agreement to be reached.

The loosening of the long-term pattern did not disrupt all the mechanisms reproducing conflict and violence. A multiplicity of such mechanisms continued to function at local and regional levels - not least the marching controversies (O'Neill, 2000) - but they were disaggregated and violent opposition was marginalised. Why then did Protestant opposition to the Agreement increase so rapidly after 1998 (Ruane and Todd, 2001)? The difficulty of implementation lay - and still lies - in precisely the ambiguities in the new state trajectories that helped broker agreement. The two governments may have a 'deconstructed sovereignty' vision of what is now in place, but they have not foreclosed on the possibility that this is a transitional phase rather than a permanent state. The GFA is open to the interpretation that it is yet another step in the diminution ofBritish sovereignty over Ireland, signalling a further step in the long advance by nationalists and a base from which a new challenge will be made in the future. This is the loyalist fear and the republican hope: it constitutes the rationale ofthe unionist opposition to the Agreement (Farrington, 2006).

What role did power stalemate, cognitive and ideological change, the building of cross-community linkages and the nature of the new institutions play in producing settlement? Each took on a new significance in the context of changing state trajectories. Stalemate was important. If republicans could have forced the British out within a realistic time period, they would not have abandoned the armed struggle. If the British government could have definitively defeated the IRA it is unlikely that it would have gone through the tortuous negotiations of the 1990s. However stalemate was as consistent with long-term struggle for position as it was with negotiations for settlement. It was only when the parameters of the long-term struggle changed that stalemate (and the recurrent fears that it might be broken to the enemy's advantage) became an incentive for negotiations and agreement. Ideological change also played a role, although its precise significance varied with the evolving situation. The new language and ideology were consistent with old strategies and aims, and if they eventually signalled a change in cognitive perspectives, they did not begin in this way. In a situation of uncertainty and ambiguity, the discourse of pluralism became a political lingua franca for those in opposed entrenched positions which indicated a willingness to move if, and

only if, realistic opportunities should emerge. Cross-community linkages were less important in the settlement process. We have argued that the Agreement is an intervention in a path-dependent process, an unpicking of its mechanisms of reproduction and lock-in. As this proceeds, it becomes possible to build up and harness cross-community networks and relationships, but it is a misconception of the nature of the conflict to assume that all that is required for settlement is to find ways ofinstitutionally mobilising an already existent peaceful civil society. Finally, we have argued that the context of changing British strategy and changing power resources gave a new significance to the prospect of consociational democracy that had been rejected by republicans and most unionists in 1973-4. The 1998 institutions were better designed and more inclusive than those of the previous period, but this is not the most important difference; it is rather that the long-term opportunities and dangers held out by the institutions were radically changed. Sunningdale was an attempt at a 'quick-fix' political solution to a problem that was structural and systemic in nature. The Good Friday Agreement became possible not because 'slow learners' had eventually grasped the need to cooperate in representative institutions, but because there had been a major shift at once in the communal power balance, in the trajectories of the two states and in the guarantees and support offered internationally.

Our argument, in short, is that geopolitical shifts (international and state) gave the window of opportunity for change. Changed communal power resources gave an incentive and means, and allowed a reassessment of the parties' short-term preferences. At that stage, rational self-interest began to point to settlement rather than to conflict and allowed agreement to be reached. Ambiguities in state trajectories continue to destabilise the Agreement. However the change in the pattern of relationships is highly unlikely to be reversed even should the conso-ciational institutions fail to be reinstated.

Settlement as the Breach of Path-Dependent Patterns

We have argued here for a particular strategy for explaining settlement processes. Instead of searching for covering laws that identify particular factors (or even clusters of factors) likely to lead to settlement, we suggest that it may be more fruitful to search for underlying path-dependent patterns which regulate how these factors function. This is not a rejection of comparison or generalisation but a shift in its focus. The defining characteristic of a path-dependent pattern is precisely its uniqueness, stemming from the particular contingent events that set a particular combination of elements and mechanisms in place (Mahoney, 2000, pp. 507-8). However, the presence of path-dependent patterns in protracted conflicts is likely to be general, just as are the types of factors involved and the types of interventions which break their patterning. This approach suggests the need to reframe the competing explanations with which we began. Rather than identifying sufficient or even necessary conditions of settlement, they are better seen as describing social mechanisms, causally relevant to settlement but not

necessary or sufficient to produce it.7 The functioning and outcomes of these mechanisms are dependent not just on their sequencing and interrelations (see McAdam et al., 2001) but also on their relation to underlying path-dependent patterns. The following points of comparative relevance follow from the discussion above.

First, some conflicts are so locked in, with such a multiplicity of convergent reasons for conflict that no single change - however good the institutions or balanced the power relations - is enough to bring settlement. Only a change sufficient to breach the path-dependent pattern will allow these factors to have effect. Since actors themselves intuitively recognise the long-term patterns, expect them to continue and devise their strategies accordingly, change has to be sufficiently radical and long-term to convince them that the patterns are broken.

Second, this allows settlements to be seen as products of 'critical junctures', disruptions of paths or intersections between different paths (Mahoney, 2000; Thelen, 2003). These critical junctures are rare, not because they are a product simply of unpredictable exogenous shocks but because even endogenous processes of change have to build to threshold levels to breach the patterns of conflict.

Third, one way that such change can occur is through change in the geopolitical context which at once intensifies ongoing slow processes of change, magnifies their significance and 'guarantees' to the actors in the region that these changes will not easily be reversed. One might argue that the basic difference between the relatively successful Northern Irish settlement and the unsuccessful Oslo peace process lies in the way changes in state trajectories in Northern Ireland were intensified in the new global environment (Ben-Porat, 2005).

Fourth, radical, long-term changes in state trajectory facilitate settlement in part by provoking radical re-categorisations among the parties to the conflict: this allows them to see new long-term opportunities in institutional provisions which before they would have rejected as insufficient or as radically unfair.This does not have to involve change in goals - among both unionists and republicans it was rather change in constitutive assumptions. Such change does not mean that previous beliefs were shallow, but rather that they were intimately connected to existing patterns of power and opportunity.

Fifth, among the many reasons why peace processes and settlements break down (see Darby and MacGinty, 2003) is that the actors come to believe that their re-categorisations were mistaken. Since 1998, increasing numbers of unionists have come to believe that the GFA did not secure their position but instead benefited nationalists.8The misjudgement may also be in another direction.Those republicans who believed that the changes were moving with the 'grain of history' to nationalist victory may find that instead they have become locked in a new path where victory is even farther away than before (McIntyre, 2001). One

of the difficulties for pro-settlement leaders in contemporary Northern Ireland is in judging just how far state trajectories and patterns of conflict have in fact shifted (see Ruane, 1999).

The explanation of settlement suggested here builds on the insights of new approaches to conflict, focusing on sequences and patterns of factors, lock-in and path dependence. This approach has not, to our knowledge, before been used in analysis of settlements. It is so difficult to find adequate explanations of settlement, either in general or - as we have seen here - in hard cases, that it is important to open up another approach. This form of explanation allows us to identify the difficulties of achieving settlement and the problems with incremental approaches to settlement, without being deterministic. Possibilities of radical change exist, if critical junctures are created and grasped.

(Accepted: 1 June 2006) About the Authors

Jennifer Todd, School of Politics and International Relations, University College Dublin, Belfield, Dublin 4, Ireland; email: jennifer.todd@ucd.ie

Joseph Ruane, Department of Sociology, University College Cork, Cork, Ireland, email: j.ruane@ucc.ie

The authors acknowledge helpful comments and criticisms from Christopher Farrington and anonymous reviewers and the Editor of Political Studies. Jennifer Todd acknowledges the Contentions and Transitions research programme at the Geary Institute, University College Dublin.

1 We make no claims here to cover the entire literature, and in particular a detailed discussion of inductive approaches is beyond the scope of this article. However our conclusion applies equally to them.

2 There was a new large paid professional cross-community civil society sector of employees, but at the grass roots less changed; for example, the percentage of children at integrated schools increased from 1 percent to 4 percent between the early and late 1990s, a significant increase but not one that affected a significant proportion of the population (Gallagher and Dunn, 1991; Stephen, 2000, p. 167).

3 In the 2003 regional elections, the two 'extreme' parties, Sinn Fein and the Democratic Unionist party, were voted into majority positions within their respective blocs. For the results, see Irish Political Studies, Data Yearbook 2004.

4 See for example the joint British and Irish government declaration (the 'Downing Street Declaration') of 15 December 1993, paras 1-4, para 9, reprinted in McGarry and O'Leary (1995, appendix A).

5 See Owen (2002) for a discussion of the radical nature of the change.

6 Cox, 1997; English, 2003, p. 307; Guelke, 2002; Laffan, 2005.

7 For a strong statement on the contingency of outcomes of social mechanisms, see Elster (1998). The seeming contingency of outcome of the mechanisms identified here, however, is explicable in terms of their context.

8 See Life and Times surveys (political attitudes module, GOODFRI) at www.ark.ac.uk/nilt.

References

Alba, R. (2005) 'Bright vs. Blurred Boundaries: Second Generation Assimilation and Exclusion in France, Germany and the United States', Ethnic and Racial Studies, 28 (1), 20-49.

Albrow, M. (1996) The Global Age. Cambridge: Polity.

Arthur, P. (1999)' "Quiet Diplomacy and Personal Conversation"TrackTwo Diplomacy and the Peace Process in Northern Ireland', in J. Ruane and J. Todd (eds), After the Good Friday Agreement: Analysing Political Change in Northern Ireland. Dublin: University College Dublin Press, pp. 71-95.

Aughey, A. (1999) 'A New Beginning:The Prospects for a Politics of Civility in Northern Ireland', in J. Ruane and J. Todd (eds), After the Good Friday Agreement: Understanding Political Change in Northern Ireland. Dublin: University College Dublin Press, pp. 122—44.

Ben-Porat, G. (2005) 'A New Middle East? Globalisation, Peace and the "Double Movement" ', International Relations, 19 (1), 30-62.

Bloomfield, D. (1998) Political Dialogue in Northern Ireland: The Brooke Initiative, 1989—92. Houndsmills: Macmillan.

Bloomfield, D. (2001) Developing Dialogue in Northern Ireland: The Mayhew Talks, 1992. Houndsmills: Palgrave Macmillan.

Bourke, R. (2003) Peace in Ireland: The War of Ideas. London: Pimlico.

Brown, M. E. (ed.) (1996) The International Dimensions of Internal Conflict. Cambridge MA: MIT Press.

Brubaker, R. (2002) 'Ethnicity without Groups', Archives Européennes de Sociologie, 43 (2), 163-89.

Coakley,J. (ed.) (2002) Changing Shades of Orange and Green: Redefining the Union and the Nation in Contemporary Ireland. Dublin: University College Dublin Press.

Coakley, J. (ed.) (2003) The Territorial Management of Ethnic Conflict. London: Frank Cass.

Cortell, A. P. and Peterson, S. (1999) 'Altered States: Explaining Domestic Institutional Change', British Journal of Political Science, 29 (1), 177-200.

Cortell, A. P. and Peterson, S. (2001) 'Limiting the Unintended Consequences of Institutional Change',

Comparative Political Studies, 34 (7), 768-99.

Cortell, A. P. and Peterson, S. (eds) (2002) Altered States: International Relations, Domestic Politics and Institutional Change. Lanham MD: Lexington Books.

Cox, M. (1997) 'Bringing in the "International":The IRA Ceasefire and the End of the Cold War', International Affairs, 73 (4), 671-93.

Crouch, C. and Farrell, H. (2002) 'Breaking the Path of Institutional Development? Alternatives to the New Determinism', EUI Working Paper Series, SPS no. 2002/4.

Darby, J. (1986) Intimidation and the Control of Conflict in Northern Ireland. Dublin: Gill and Macmillan.

Darby, J. and MacGinty, R. (eds.) (2003) Contemporary Peacemaking: Conflict, Violence and Peace Processes. Houndsmills: Palgrave Macmillan.

David, P. A. (1994) 'Why are Institutions the "Carriers of History"?: Path Dependence and the Evolution of Conventions, Organizations and Institutions', Structural Change and Economic Dynamics, 5 (2), 205-20.

David, S. R. (1997) 'Internal War: Causes and Cures', World Politics, 49 (4), 552-76.

Elster, J. (1998) 'A Plea for Mechanisms', in P. Hedstrom and R. Swedborg (eds), Social Mechanisms:An Analytic Approach to Social Theory. Cambridge: Cambridge University Press, pp. 45-73.

English, R. (2003) Armed Struggle. A History of the IRA. London: Macmillan.

Farrell, S. (2000) Rituals and Riots: Sectarian Violence and Political Culture in Ulster. Lexington KY: University of Kentucky Press, pp. 1784-886.

Farrington, C. (2004) 'Models of Civil Society and their Implications for the Northern Ireland Peace Process', IBIS Working Paper no. 43 (Dublin: Institute for British Irish Studies).

Farrington, C. (2006) Ulster Unionism and the Peace Process in Northern Ireland. Houndsmills: Palgrave Macmillan.

FitzGerald, G. (1991) All in a Life. Dublin: Gill and Macmillan.

Gallagher, A. M. and Dunn, S. (1991) 'Community Relations in Northern Ireland', in P. Stringer and G. Robinson (eds), Social Attitudes in Northern Ireland. Belfast: Blackstaff, pp. 7-22.

Guelke, A. (2002) 'Northern Ireland and the International System', in W. Crotty and D. Schmitt (eds), Ireland on the World Stage. Harlow: Pearson Education, pp. 127-39.

Guelke, A. (2003) 'Civil Society and the Northern Ireland Peace Process', Voluntas: International Journal of Voluntary and Non-profit Organisations, 14 (1), 61-78.

Hall, P. A. and Taylor, R. C. R. (1996) 'Political Science and the Three New Institutionalisms', Political Studies, 44 (4), 936-57.

Hargie, O. and Dickson, D. (eds) (2003) Researching the Troubles: Social Science Perspectives on the Northern Ireland Conflict. Edinburgh: Mainstream Publishing.

Hauss, C. (2001) International Conflict Resolution: International Relations for the 21st Century. London: Continuum.

Hayward, K. (2004) 'The Politics of Nuance: Irish Official Discourse on Northern Ireland', Irish Political Studies, 19 (1), 18-38.

Hirst, C. (2002) Religion, Politics and Violence in Nineteenth Century Belfast:The Pound and Sandy Row. Dublin: Four Courts Press.

Kakar, S. (1996) The Colors of Violence, Cultural Identities, Religion and Conflict. Chicago IL: Chicago University Press.

Keane, J. (1998) Civil Society: Old Visions, New Images. Cambridge: Polity.

Keating, M. (2001) Plurinational Democracy: Stateless Nations in a Post-Sovereignty Era. Oxford: Oxford University Press.

King, C., Lust-Okar, E. and Shapiro, I. (2005) 'Problems and Prospects for Democratic Settlements: South Africa as a Model for the Middle East and Northern Ireland', Politics and Society, 33 (2), 277-326.

Krieger, J. (1999) British Politics in the Global Age: Can Social Democracy Survive? Cambridge: Polity.

Laffan, B. (2005) 'The European Context: A New Political Dimension in Ireland, North and South', in J. Coakley, B. Laffan and J. Todd (eds), Renovation or Revolution: New Territorial Politics in Ireland and the United Kingdom. Dublin: University College Dublin Press, pp. 166-84.

Lake, D. A. and Rothchild, D. (eds) (1998) The International Spread ofEthnic Conflict: Fear, Diffusion and Escalation. Princeton NJ: Princeton University Press.

Lamont, M. (2000) The Dignity of Working Men: Morality and the Boundaries of Race, Class, and Immigration. Cambridge MA: Harvard University Press.

McAdam, D., Tarrow, S. and Tilly, C. (2001) Dynamics of Contention. Cambridge Studies in Contentious Politics. Cambridge: Cambridge University Press.

McGarry, J. and O'Leary, B. (1995) Explaining Northern Ireland: Broken Images. Oxford: Blackwell.

McGarry, J. and O'Leary, B. (2004) The Northern Ireland Conflict: Consociational Engagements. Oxford: Oxford University Press.

McIntyre, A. (2001) 'Modern Irish Republicanism and the Belfast Agreement: Chickens Coming Home to Roost or Turkeys Celebrating Christmas?', in R. Wilford (ed.), Aspects of the Belfast Agreement. Oxford: Oxford University Press, pp. 202-22.

Mahoney, J. (2000) 'Path Dependence in Historical Sociology', Theory and Society, 29 (4), 507-48.

Owen, D. (2002) 'The Resolution of Armed Conflict: Internationalization and its Lessons, Particularly in Northern Ireland', in M. Elliott (ed.), The Long Road to Peace in Northern Ireland. Liverpool: Liverpool University Press, pp. 22-40.

O' Riain, S. (2004) State Development and the Global Economy: Developmental Network States in the Global Economy. Cambridge: Cambridge University Press.

O'Neill, S. (2000) 'Liberty, Equality and the Rights of Cultures:The Marching Controversy at Drumcree', British Journal of Politics & International Relations, 2 (1), 26-45.

Patterson, H. (2004) 'Missing the Point: Academic Analyses of the "New Unionism" '. Paper presented to the Identity, Diversity and Citizenship conference, Institute for the Study of Social Change, University College Dublin, 6 October.

Petersen, R. D. (2002) Understanding Ethnic Violence: Fear Hatred and Resentment in Twentieth Century Eastern Europe. Cambridge: Cambridge University Press.

Pierson, P. (2000) 'Increasing Returns, Path Dependence and the Study of Politics', American Political Science Review, 94 (2), 251-67.

Pierson, P. (2004) Politics in Time: History, Institutions and Social Analysis. Princeton NJ: Princeton University Press.

Putnam, R. D. (2003) Better Together: Restoring the American Community. New York: Simon & Schuster.

Ruane, J. (1999) 'The End of (Irish) History: Three Readings of the Current Conjuncture', in J. Ruane and J. Todd (eds), After the Good Friday Agreement: Analysing Political Change in Northern Ireland. Dublin: University College Dublin Press, pp. 145-70.

Ruane, J. (2004) 'Contemporary Republicanism and the Strategy of Armed Struggle', in M. J. Bric and J. Coakley (eds), From Political Violence to Negotiated Settlement. Dublin: University College Dublin Press, pp. 115-32.

Ruane, J. and Todd, J. (1996) The Dynamics of Conflict in Northern Ireland: Power, Conflict, and Emancipation. Cambridge: Cambridge University Press.

Ruane, J. and Todd, J. (1999) 'The Belfast Agreement: Context, Content, Consequences', in J. Ruane and J. Todd (eds), After the Good Friday Agreement: Explaining Political Change in Northern Ireland. Dublin: University College Dublin Press, pp. 1-29.

Ruane, J. and Todd, J. (2001) 'The Politics of Transition: Explaining the Crises in the Implementation of the Belfast Agreement', Political Studies, 49 (5), 923-40.

Ruane, J. and Todd, J. (2003) 'A Changed Irish Nationalism? The Significance of the Belfast Agreement of 1998', in J. Ruane, J. Todd and A. Mandeville (eds), Europe's Old States in the New World Order: The Politics of Transition in Britain, France and Spain. Dublin: University College Dublin Press, pp. 121-45.

Ruane, J. and Todd, J. (2004) 'The Roots of Intense Ethnic Conflict may notThemselves be Ethnic: Categories, Communities and Path Dependence', Archives Européennes de Sociologie, 45 (2), 209-32.

Schultze, K. E. (1997) 'The Northern Ireland Political Process: A Viable Approach to Conflict Resolution?', Irish Political Studies, 12, 92-110.

Stedman, S.J. (2003) 'Peace Processes and the Challenge ofViolence', in J. Darby and R. MacGinty (eds), Contemporary Peacemaking: Conflict, Violence and Peace Processes. Houndsmills: Palgrave Macmillan, pp. 103-13.

Stephen, F. (2000) 'Promoting a Culture ofTolerance: Education in Northern Ireland', in M. Cox, A. Guelke and F Stephen (eds), A Farewell to Arms: From 'Long War' to Long Peace in Northern Ireland. Manchester: Manchester University Press, pp. 165-78.

Streeck, W and Thelen, K. (eds) (2005) Beyond Continuity: Institutional Change in Advanced Political Economies. Oxford: Oxford University Press.

Thelen, K. (2003) 'How Institutions Evolve: Insights from Comparative Historical Analysis', in J. Mahoney and D. Rueschemeyer (eds), Comparative Historical Analysis in the Social Sciences. Cambridge: Cambridge University Press, pp. 206-40.

Todd, J. (2006) 'Identity, Identity Change and Group Boundaries in Northern Ireland', Geary Working Paper, 2006/6.

Varshney, A. (2001) 'Ethnic Conflict and Civil Society: India and Beyond', World Politics, 53 (3), 362-98.

Walter, B. F. (2001) 'DesigningTransitions from Civil War: Demobilization, Democratization, and Commitments to Peace', in M. Brown, R. Owen, R. Coté Jr, S.M. Lynn-Jones and S. E. Miller (eds), Nationalism and Ethnic Conflict, revised edition. Cambridge MA: MIT Press, pp. 415-43.

Wilford, R. and Wilson, R. (2006) The Trouble with Northern Ireland: The Belfast Agreement and Democratic Governance. Dublin: tasc at New Island Press. Available at http://www.tascnet.ie (Accessed 20.11.06).

Wimmer, A. (forthcoming) The Making and Unmaking of Ethnic Boundaries: Toward a Comparative Theory. Unpublished Manuscript, Department of Sociology, University of California, Los Angeles.

Wolff, S. (2001) 'Context and Content: Sunningdale and Belfast Compared', in R. Wilford (ed.), Aspects of the Belfast Agreement. Oxford: Oxford University Press, pp. 11-27.

Wright, F. (1987) Northern Ireland: A Comparative Analysis. Dublin: Gill and Macmillan.

Wright, F. (1996) Two Lands, One Soil. Dublin: Gill and Macmillan.

Zartman, I. W (1989) Ripe for Resolution: Conflict and Intervention in Africa, second edition. Oxford: Oxford University Press.

Zartman, I.W (2003) 'The Timing of Peace Initiatives: Hurting Stalemates and Ripe Moments', in J. Darby and R. MacGinty (eds), Contemporary Peacemaking: Conflict,Violence and Settlement Processes. Houndsmills: Palgrave Macmillan, pp. 19-29.