The Role of Framing Mechanisms in Explaining System Wide Change: the case of the Northern Ireland Conflict and Peace Process

System-wide change is often challenging to achieve due to complex and fragmented in- stitutions, dispersed and diffused power structures, conﬁdence-sapping histories of failure and the inﬂuence of multiple and overlapping ﬁelds. This study examines how a large com- plex system-wide problem such as the Northern Ireland Conﬂict and Peace Process was paradoxically opened up and made more receptive to change by widening of the way the problem was framed. We demonstrate how and why the framing enables the mobilization of cooperation and the delivery of contextually appropriate collective action critical to the achievement of outcomes in system-wide change processes. More speciﬁcally, we examine how and why such complex and precarious processes emerge over extended timescales through four mechanisms: frame contesting, reframing, frame reproduction and frame defending. Each of these mechanisms is agentic, dynamic, purposive and politically charged. The time-series analysis of these interlinked mechanisms is a crucial and innovative feature of the study. We encourage management and organizational scholars to elevate their gaze to the system-wide changes so emblematic of contemporary society and offer an outline agenda for research.


Introduction
The paper addresses why and how success can be achievable in delivering system-wide changes. Our analytical approach involves linking together the context, mechanisms, processes and outcomes of the change. In so doing, we point to the importance of frames in mobilizing cooperation around a shared issue or cause (Campbell, 2004;Fligstein and McAdam, 2012). More specifically, we examine how and why such complex and precarious processes emerge over time through four mechanisms: frame contesting, reframing, frame reproduction and frame defending. Each of these mechanisms is agentic, dynamic, purposive and politically charged. The time-series analysis of these interlinked mechanisms is a crucial and innovative feature of the study. The study reveals how the reframing of a large complex system-wide problem such as the Northern Ireland Conflict and Peace Process (NICPP) can paradoxically be opened up and made more receptive to change by widening of the way the problem was framed. In this case, the reframing of the problem from the two tribes to the three strands created both more complexity but also more malleability and allowed other mechanisms to play out in the subsequent process.
We demonstrate how and why the mobilization of cooperation and the delivery of contextually appropriate collective action are critical to C  This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. the achievement of outcomes in system-wide change processes. Our operational definition of system-wide change processes as purposive and emergent changes that address issues of scale and straddle multiple arenas and transcend multiple field boundaries takes us into a range of practical challenges of data collection and method rarely attempted by scholars of management and organization theory. This kind of research demands access to multiple stakeholders, sometimes political and other elites, and above all time-series process data across various levels of context. The present study was substantially enabled by its location in three related studies in the same setting and by its rich complement of primary and secondary data. As such, the paper may provide some methodological guidance and encouragement for other scholars wishing to study system-wide change processes. In what follows we outline our conceptual framework and analytical arguments and then illustrate their power and significance through an analysis of one of the biggest system-wide changes of our time, the NICPP.

Issue framing and change
Framing has become an important theme in the organization, leadership and institutional literatures. The way ideas are presented (the 'frame') to an audience influences the choices actors make and their willingness to engage in change (Campbell, 2004). Frames create a form of conceptual scaffolding (Snow and Benford, 1988) that helps actors to understand the world (Hoffman, 1999) and provides normative guides for action (Goffman, 1972), through which actors can gain influence and alter power structures (Fligstein, 2001). Frames condition actors' choices for sensemaking (Thornton, Ocasio and Lounsbury, 2012), orient actors to work together towards a shared goal or need (Evans and Kay, 2008;Mead, 1967) and fuse a capacity to act (Stone, 1989).
Framing is a process of 'evoking meaning, in line with existing cultural categories of understanding and as a basis for mobilizing support and gaining legitimacy' (Cornelissen and Werner, 2014, p. 182). Thornton, Ocasio and Lounsbury (2012) suggest that framing is an important means by which institutional logics at the societal level get translated to the field level and micro-level meanings endogenous to a field can also influence recursively higher-order levels of meaning and activity. These institutional logics (Friedland and Alford, 1991) shape field boundaries, identities and interactions (Greenwood and Suddaby, 2006). Framing processes are imbued with history as well as present-day concerns (Berger and Luckmann, 1967;Ybema, 2014). Therefore, examining framing processes can reveal the 'communicative constitution, reproduction, and transformation of institutions' (Cornelissen and Werner, 2014, p. 14) as meanings and practices are initiated, reconstituted, diffused and translated within and across organizational fields over time (Campbell, 2004). Although a nascent literature has pointed to the importance of framing and large-scale change (e.g. Granqvist and Laurila, 2011;Gray, Purdy and Ansari, 2015;Gurses and Ozcan, 2015;Hargrave and Van de Ven, 2006;Litrico and David, 2017;McAdam, McCarthy and Zald, 1996), this literature does not fully explain how framing mechanisms deliver system-wide change over time and the influence of field and inter-field conditions.
Many issues are complex, multi-faceted and produce a wide range of frames. For example, the issue of obesity may be framed in terms of, but not limited to, diet, exercise, body image, advertising, food manufacturing, social inequalities, education, healthcare, the economy, the role of corporations, civil liberties or human rights. Highly complex issues can be simplified or reframed (Grint, 2005), and this has implications for change. For example, if obesity is framed predominantly as a food manufacturing problem, change initiatives may be limited to governments and regulators imposing restrictions on foods with high sugar or fat content. Actors, particularly those in positions of authority, can use framing to exercise power by means of drawing attention to, and generating support for, one particular understanding at the expense of others (Grint, 2005). Actors have the opportunity to frame lines of action and mobilize people in the service of these action frames (Fligstein, 2001;Jasper, 2004Jasper, , 2006Snow and Benford, 1988). Communicative acts such as stories, the selective presentation of facts or emotional appeals are used to manipulate the ways in which people process information (Fairhurst and Sarr, 1996). Grint (2005) uses the examples of the Brent Spar disaster, the Cuban Missile Crises and the 'War on Terror' in Iraq to illustrate how a persuasive reframing of a wicked problem as a tame problem (Rittel and Weber, 1973) can be used to legitimize the use of a predetermined, tried and trusted process or can be reframed as a critical problem (a crisis) in order to enforce a strong-arm course of action on followers (Grint, 2005). As such, frames are imbued with power and politics as actors use frames to reconstruct current problems as a strategy to instigate or resist change (Fairhurst and Sarr, 1996;Grint, 2005).
State and institutional actors can shape one another's micro framing (Olsen, 2017), which can build up over time (Ansari, Fiss and Zajac, 2010) and generate a 'societal level force' for change (Marsden, 2016;Tremblay et al., 2017). For example, Gray, Purdy and Ansari (2015) argue that the 'rights' frame that emerged from the civil rights movements provided a generalized cultural resource for subsequent rights movements, such as gay, animal and women's rights. Phillips, Lawrence, and Hardy (2004) reveal how reframing whales from creatures of destruction to friendly foes enabled the Canadian whale-watching industry. Similarly, Zietsma and Lawrence (2010) show how two initially powerless environmental activists and aboriginal people collectively achieved change in tree logging by framing the issue as government being 'an industry "lap dog"' (p. 205). Despite these examples, few studies examine how framing mechanisms unfold in system-wide change over extended timeframes and how these processes are shaped by field and inter-field conditions.

System-wide change
The big issues of change of our age no longer lie just within organizations or institutions, and in the fields of management and organization studies we need to prepare ourselves for this reality. The events in the European Union (EU) in the last several years, first of all grappling with the position of Greece, mass migration from the Middle East and Africa, and now with the issue of Brexit, illustrate the challenges of delivering collective action and purposeful change. We define 'systemwide change' as purposive and emergent actions that address issues of scale that straddle multiple arenas (Evans and Kay, 2008) and transcend multiple field boundaries (Fligstein, 2001). These 'systems of systems' are 'dynamic situations that consist of complex systems of changing problems that interact with each other' (Ackoff, 1979, p. 93). Each part of the system is contained in a larger system and is the product of the interactions of its parts (Ackoff, 1971).
Examples of system-wide changes are plentiful. They include climate change and poverty reduction (e.g. Ferraro, Etzon and Gehna, 2015), building resilience in post-conflict societies involving multiple independent but connected institutions (Mahmoud and Makoond, 2017) and the creation of regimes to drive urban development (Stone, 1989). We also see attempts at system-wide change focused on racial/ethnic mobilization to address discrimination and inequality (Mora, 2014a, 2014b), revolutions (McAdam, Tarrow and Tilly, 2001 and civil rights movements (Hargrave and Van De Ven, 2006); disruptive technologies requiring coalitions of actors who individually lack resources, power or legitimacy groups to produce change by themselves (Hargrave and Van De Ven, 2006); collective action to enable publicsector reform (Ferlie et al., 1996); and initiatives designed to address intractable health care issues (Pettigrew, Ferlie and McKee, 1992). These issues have often been referred to as grand challenges (Ferraro, Etzon and Gehna, 2015), wicked problems (Head and Alford, 2015;Rittel and Webber, 1973) or 'messes' (Ackoff, 1979, p. 93). Systemwide changes perennially involve what Sterman (2001) has called combinatorial and dynamic complexity, involving multiple, diverse and interconnected elements. As we shall see, other features of such system-wide change include multiple and geographically dispersed stakeholders, divergent interests and values, dispersed and diffused power systems and the perpetual challenge of delivering collective action around change issues which themselves are variously perceived and acted upon.

A system of fields
The institutional literature has typically addressed these macro-level change issues by means of the field concept -a bounded area within which actors interact, forming 'a recognized area of institutional life' (DiMaggio and Powell, 1983, p. 148) and comprising 'a collection of diverse, interdependent organizations that participate in a common meaning system' (Scott, 2014, p. 106), where 'actors (who can be individual or collective) are attuned to and interact with one another on the basis of shared (which is not to say consensual) understandings about the purposes of the field, relationships to others in the field (including who has power and why), and the rules governing legitimate action in the field' (Fligstein and McAdam, 2012, p. 9). Yet the field concept has been criticized for being typically 'rather narrow when compared to the wicked problems facing global society' (Zietsma et al., 2017, p. 72). Even relatively 'mundane' issues cannot be settled within single fields (Zietsma et al., 2017, p. 73).
Some authors have attempted to elevate the field concept. Bourdieu and Wacquant 1992 or example, conceive of fields as social spaces that represent the structures of different components of social life through which power is constituted, contested and reproduced as actors pursue common interests. Historical institutionalists have also customarily worked on bigger issues and on a grander scale than management and organization theorists (Campbell, 2004;Pierson, 2004;Skocpol and Pierson, 2002). A small number of pioneering studies have also begun to examine the relationship between institutional fields (Campbell, 2004;Evans and Kay, 2008;Fligstein and McAdam, 2012;Mora, 2014aMora, , 2014bSchneiberg, 2007). These pioneering studies provide an 'architecture' for field overlap (Mora, 2014a(Mora, , 2014b and the idea of nested or embedded nature of multiple fields. Recently the writing of Fligstein (2001) and Fligstein andMcAdam (2011, 2012) has markedly taken forward thinking about the possibility of collective action in what they call strategic action fields. Fligstein and McAdam (2012) provide several examples of strategic action fields from a collection of universities in a given country, to markets (such as the US housing mortgage market), to policy fields (such as the US 'field of racial politics' during contention over civil rights and race in the USA). Fields can have proximate ties with other fields whose actions routinely impact the field in question or distal ties that have little influence; ties that reflect the existence of formal hierarchical relations that are mutually dependent (vertical or horizontal) and completely independent fields (Fligstein and McAdam, 2012). Strategic action fields are embedded in a system or 'web' of other fields, 'much like a Russian doll' (Fligstein and McAdam, 2012, p. 8). These lines of inquiry and interest highlight the role of ideas and framing mechanisms in delivering change outcomes and the importance of time, history, context, configuration, power and process in shaping the development and resolution of big societal issues of the day.

Methods
We studied framing mechanisms as they unfold over time in the NICPP. Initial studies of systemwide changes will inevitably rely on 'unique' or 'outlier' cases. Pettigrew (1990Pettigrew ( , 1997 argues that case selection based on atypical examples offers more potential for learning. We have drawn on the NICPP as a generative case, basing our selection of this case on four criteria that help to illuminate the processes we are seeking to explain. First, many secondary sources on the NICPP (see Table 1) highlight the central role played by framing and reframing in opening up possibilities of change where stasis had been a strong legacy feature. The NICPP also allows an examination of the conditions, which can shape the possibility of individual and collective action. Cooperation and collective action were difficult to achieve due to complex and fragmented structures, dispersed and diffused power structures, confidence-sapping histories of failure and multiple and overlapping strategic action fields.
Second, the NICPP is an example of a strategic action field (as defined by Fligstein and McAdam, 2012, p. 9) that is embedded in a system of fields. The Northern Ireland conflict and peace field involves multiple actors and institutions seeking to further their political objectives within the historic violent conflict in Northern Ireland, often referred to as the 'Troubles'. This includes political, social and institutional actors within Northern Ireland, security services such as police forces within Northern Ireland, GB and the Republic of Ireland, paramilitaries (republican and loyalist) operating within and beyond the boundaries of Northern Ireland, a spectrum of political parties inside and outside Northern Ireland and the constituent parts of the British and Irish Governments. The NICPP is affected by other fields such as British, Irish, US and EU political and economic fields, policing and security fields in multiple jurisdictions, and cultural and social fields within Northern Ireland itself and border counties in the Republic of Ireland. As such, the NICPP field permeates and intrudes on every other field in Northern Ireland.
Third, the NICPP enables system-wide change to be exposed as a natural history of development over time. System-wide changes are historical processes and products. Crucially for any processual analysis, the 'interchange between agents and contexts over time is cumulative. The legacy of

Repositories of historical and archived documents:
https://cain.ulster.ac.uk/ -CAIN is the main academic archive relating to the Northern Ireland conflict and peace process; it is maintained by the University of Ulster.
Other material: Blogs, newspaper articles, letters to newspapers, public events/speeches, government statements (34 substantive contributions).

Data analysis
Interviews were recorded, transcribed and coded in relation to NI Peace Process framing process.
Interviews were recorded and transcribed. Data was reanalysed and recoded in relation to NI Peace Process framing process.
Interviews were recorded and transcribed. Data was reanalysed and recoded in relation to NI Peace Process framing process.
Existing published studies and documents (above) were used to develop a detailed time line and then thematically analysed to identify evidence of framing over time.
the past is always shaping the emerging future. What happens, why and how it happens, and what result it brings about is dependent on when it happens' (Pettigrew, 2012(Pettigrew, , p. 1315. But the only way to reveal the relationship between multiple levels of context is to have a time series sufficiently long to expose the natural history of development of a system-wide change process over time. The NICPP affords this opportunity. Fourth, also crucial to our selection of the NICPP and to our general and analytical argument is the need to explore the multiple emergent outcomes of change processes (Pettigrew, 1979(Pettigrew, , 2000(Pettigrew, , 2003. We are interested in variations in the outcomes of system-wide changes. Ferraro, Etzon and Gehna (2015) remind us that examples of enacted system-wide change with successful outcomes are relatively rare and routes to success are complex (Ferraro, Etzon and Gehna, 2015). The NICPP is a relatively rare example of a system-wide change that for years was seemingly intractable, then became apparently resolvable and yet a change in the context, brought about by the Brexit environment, 1 threatens to overturn transient closure and has the potential to propel the situation into another cycle of tension and conflict.

Sources of data
Our study design was guided by our concern for developing a processual explanation of how an unfolding framing process yielded particular outcomes (Walsh and Bartunek, 2011). As a socially constructed phenomenon (Grint, 2005) that unfolds over time, framing represents a phenomenon best examined through processual analysis (Langley, 1999). The research design is unusual in that it represents an integration and synthesis of three separate but interconnected research projects which each have the NICPP as a central concern (see Table 1). Each study pursued an induction-driven research design (Walsh and Bartunek, 2011) and all draw on rich, qualitative data.

Secondary source material
We gathered several hundred pages of secondary source material (see Table 1). This material provided background about activities and events that occurred during the NICPP over an extended timeframe. Some of these information sources were available publicly, and our informants provided other material. Secondary resources included extensive material for British and Irish Government documentary archives, internal organizational publications and personal communications. These secondary records served as a guiding basis for constructing historical timelines about issues and events that transpired during the NICPP. In understanding ongoing, novel or extreme events, a combination of traditional and non-traditional data can be crucial, including sources regularly dismissed by researchers with different interests (Buchanan and Denyer, 2013). These include rapidly available blogs, newspaper articles, letters to newspapers, public events/speeches and governmental statements. Existing academic and analytical accounts were used to corroborate the findings. Key sources were identified which acknowledge the contested nature of interpretation and with this in mind, we were mindful to include academic sources from multiple perspectives (see Table 1).

Analysis procedures
We constructed a case history of the NICPP using data gathered from both the archival sources and interview transcripts from the three studies. We define the NICPP as system-wide change taking place over a 30-year period and beginning with what is generally regarded as the start of the 'Troubles' in 1969. We bracketed four key episodes (Langley, 1999) and singled out critical incidents and significant decisions that appeared to be turning points in the case. These events are both outcomes of the particular conjuncture of mechanisms and contextual conditions during the preceding episode but are also generative -in that they also spark changes in the configuration of mechanisms in the subsequent period.
As the archival sources were typically produced in 'real time', triangulation of interview data and academic accounts with the archival records not only enabled richer descriptions of the case (Denzin, 1989) but also reduced the risk of our analysis being impacted by interviewees' retrospective construction of 'new memories' (Loftus and Hoffman, 1989). We analysed the data using procedures recommended by Miles and Huberman (1994). This method entails repeated comparison of one's data to a nascent model emerging from analysis (Walsh and Bartunek, 2011). Tsoukas (2009) refers to this process as analytical refinement. Plowman et al. (2007) offer an illustration of this approach. Their case study of a single organization generated fresh insights into slow-moving radical change processes, based on an integration of traditional and unconventional information sources (interviews, documents, newspaper articles, observation), using content and timeline analyses to identify combinations (conjunctures) of organizational characteristics influencing the outcomes of interest.
Like much qualitative research, our analysis proceeded through cycles of inductive and deductive reasoning (Walsh and Bartunek, 2011). Our initial familiarity with the NICPP from the three studies and the existing literature led us to expect that framing would play an important role in experiences of the change process. We were thus attuned to issues related to framing processes and outcomes in our initial reading and coding of our cases. However, these expectations did not constrain our initial examination of the data. We grouped first-order codes into theoretical categories and organized these theoretical categories, or second-order codes, into aggregate theoretical dimensions that related to framing mechanisms, field and inter-field conditions, and outcomes. We identified and examined the unfolding and interplay of four framing mechanisms: frame contesting, reframing, frame reproducing and frame defending. Whilst our analysis reveals that each episode was characterized by a predominant framing mechanism, the four mechanisms intersect and overlap throughout the timeline.

Framing mechanisms in the NICPP
Given the scale, duration and complexity of the NICPP, we do not attempt to present a comprehensive account, which would be impossible given journal article word-length limitations. Instead, we offer an abridged case description and summary of the processes of framing mechanisms and the role of field and inter-field conditions across four bracketed episodes.

Outbreak of violence 1969-1973
The Northern Ireland 'Troubles' have been one of the longest violent conflicts in Europe's recent history (Fay and INCORE, 2001;Smyth and Moore, 1996). Its background is complex. The Irish War of Independence and the partition of Ireland in 1921 saw the retention of six northern counties within the UK. The result was an almost permanent political majority for the Protestant, Unionist and British-identifying community in the North, and a minority status for the Catholic and Nationalist community who generally identified as Irish. This mutually irreconcilable framing contest came under intense pressure in the 1960s when the Northern Ireland Civil Rights movement sought redress for perceived discrimination against the minority Catholic community. Street violence resulted in a breakdown of order. 2 Paramilitary activity surged and British troops were deployed in 1969. Deaths peaked in 1972, with 476 Troubles-related killings. The collapse of local devolved political institutions resulted in a return to direct rule from London in 1974. The Troubles are generally regarded as having lasted from 1969 to 1998. In that time there were 3,700 related deaths and an estimated 40,000 injuries. The theatre of violence was not confined to Northern Ireland itself, with violence occurring in Britain, mainland Europe and across the border into the Irish Republic (Hennessey, 2007).
The conflict was regarded as 'zero sum' (McDowell, Braniff and Murphy, 2017) and irreconcilable, with Northern Ireland often placed in the category of other high-profile ethnopolitical-divided societies, such as South Africa and Israel/Palestine. It was characterized, like other divided societies, by a pervasive dialectical framing -in this case of Unionism and Nationalism, containing two constellations of opposing logics -incompatible within the field as it was defined. The border was central to the frame contestation: defended by one constellation of actors (Unionists, state forces, loyalist paramilitaries and the British Government) and rejected by the other (Nationalists, republican paramilitaries and the Irish Government). 3 It was also a key area of contention and violence -fortified, closed at many local crossing points and marked by large British Army installations.
From the onset of the Troubles we see nascent and obscured attempts to transform the conflict away from violence and to craft new competing logics outside the two dominant opposing frames. These included the creation of new political parties, 4 new social movements, 5 covert attempts at engagement between the Irish Republican Army (IRA) and the British Government dating as far back as 1972 (Powell, 2010) and the creation of the Irish Government unit to facilitate cross-border engagement. There were also short-lived 6 and unsuccessful institutional efforts brokered by the British Government to bring moderates from the competing political constellations together (such as the Sunningdale Agreement of 1973).
For a summary of this period, see Table 2.

Search for and emergence of a solution 1973-1985
Among these disparate approaches evident in the first episode was a distinct reframing of the Unionist/Nationalist 'two-tribes' dialectic to a new frame, within a greater European setting, which centred on a dilemma of relationships. It disentangled the conflict into three interactions or strands: (1) between Britain and Ireland; (2) between Northern and Southern Ireland; and (3) between the divided people of Northern Ireland. These relationships rested on what became known as 'the principle of consent' (McLoughlin, 2010), which asserted that there could be no constitutional change in the status of Northern Ireland without the consent of the people of Northern Ireland. These three relationships were brought together under the slogan of An Agreed Ireland. This message was principally and first articulated by Nationalist politician John Hume, and quickly adopted and extended by other political actors within moderate nationalism, and the Irish Government.
The process of dispersing the reframing of An Agreed Ireland was adopted, promoted and sponsored by actors within the field, and in particular 4 Social Democratic and Labour Party and the much smaller Alliance Party (1970). 5 For example, the Peace People (1976). 6 The power-sharing executive of Sunningdale collapsed in 1974.
the Irish Government. For example, the development of the New Ireland Forum saw the creation of a temporary institutional form in Dublin, in the shape of a conference which was to run from 1983 to 1984, convened to discuss ways in which 'the Troubles' might be alleviated and televised daily on Irish TV. While northern Unionists by and large refused to participate, the forum was instrumental in establishing Irish institutional thinking on the conflict and providing a platform for new thinking which began to be articulated and would become a new narrative or 'master frame'. This thinking was fundamentally disruptive to the original contested logics of conflict and was diffused through London, America and crucially Europe by Hume himself and the Irish Government. The European field was vital as the geographic and political scaffolding of this new Frame: it had also emerged from war and its future avoidance (White, 2014). Similar activity was at work within Washington forging a bipartisan position on Northern Ireland, epitomized by Carter's statement of 1977, heavily influenced by newly engaged Irish America (McLoughlin and Meagher, 2018). The 'reframe' gained traction through the building of relationships in support of a future agreement, particularly at senior diplomatic levels in Dublin and London (Lillis and Goodall, 2010).
For a summary of this period, see Table 3.

Progress towards agreement 1985-1998
The successful diffusion of the new Frame first became evident in the Anglo-Irish Agreement in 1985, which adopted aspects of the Frame by enshrining legally, for the first time, the significance of the relationship between Northern Ireland and the Republic of Ireland and between Ireland and Britain -two of what became known as the three strands (Guelke, 2012). The new settled Frame was further reinforced by the 'Downing Street Declaration' of 1993, in which the UK and Irish Governments reaffirmed the principles of selfdetermination for the Irish people, the 'principle of consent' and state bipartisanship. The IRA ceasefire of 1994 and the subsequent loyalist ceasefires moved the conflict into a new, less violent phase. At this point the new master frame of the three sets of relationships had been adopted by the main institutional actors (British, Irish, US Government and EU structures). It was still contested within Northern Ireland itself -that process would take They don't get it' (R3I67). 'One of the features of NI during the troubles was the almost peculiar veneer of normality which existed in many parts of NI at that time, and how it co-existed with the dysfunction and the violence which was also present. But the troubles could also make interventions in people's lives that could completely disrupt their normal lives in ways that were completely unpredictable' (R2WS1).

Division
Defended by one constellation of actors (Unionists, state forces, loyalist paramilitaries and the British Government) and rejected by the other (Nationalists, republican paramilitaries and the Irish Government). a For example, 'Why was the Department of Foreign Affairs involved in this, at all? We claimed Northern Ireland in the constitution. Under the constitution, it was part of our territory . . . ' (R1I6). Characteristics of a deeply divided society (Lustick, 1993).

Conflict
One of the longest violent conflicts in Europe's recent history (Fay and INCORE, 2001;Smyth and Moore, 1996).
Intractable problem Many short-lived and unsuccessful attempts to broker peace (e.g. Sunningdale Agreement).
a Through Articles 2 and 3 of the Irish Constitution which upheld the Irish State's right to the six northern counties. substantial further negotiations. These talks were constructed around the three strands, with both governments acting as joint conveners, and chaired by former US Senator George Mitchell. The historic Belfast Agreement was signed on Good Friday, 1998. While this combination of framing and the active utilization of inter-field brokers set the direction towards peace, it was not until the political system was aligned (including a strong Westminster government under the premiership of Tony Blair that was not reliant on Unionist parliamentary votes) and stakeholders committed (including both governments, and a generation of Northern Ireland political parties and paramilitaries) that the sufficient conditions for system-wide change were present. Very significant EU financial support for peace and reconciliation was also vital. Post agreement, the structures of the three strands were established -including the Northern Ireland Assembly and Executive, the North South Ministerial Council, the North/South inter-parliamentary Association, the British Irish Intergovernmental Conference and the British Irish Council (O'Leary, 2016). The IRA and loyalist ceasefires which had preceded the 1998 Agreement also precipitated widespread demilitarization throughout Northern Ireland -and especially on and around border communities. In practice, this meant a reduction in what had become known as the 'architecture of the Troubles' -watch towers, fortified police station and border posts, and an opening of previously blocked border roads and pathways which created a distinct change in how people were able to travel, shop and go about normal activities. This reduction in military infrastructure visibly reinforced the benefits of peace. The EU Customs Union underpinned what was now an open border. Operation Relational frame/three strands 'We had a problem and we were desperate for a solution, and he (Hume) had a solution' (R1I34). 'There was one school of thought that said that we have to fight it out, literally fight it out . . . and then Hume came along with this idea that the solution was to be found in human relationships and he designed a process addressing the three sets of relationships that impact the conflict. I do not say this lightly and I am a unionist' (Jeffrey Donaldson, DUP MP, quoted in a letter to the Irish Times, 3 June 2019).
Frame settlement 'The two governments will work together with the parties to achieve a comprehensive accommodation, the implementation of which would include interlocking and mutually supportive institutions across the three strands, including: (a) structures within Northern Ireland . . . (b) North/South intuitions . . . (c) East West structures' (Article 13, The Framework documents, 22 February 1995, A New Framework for Agreement). 'I mean whether it was described as three strands from the beginning, I don't remember. But it was the basis on which we engaged to Senator Kennedy and others in Washington' (R1I23).
Deep frame (cutting across multiple institutional orders) 'I was asked to go and talk to Blair, which I did, and what he wanted to do was . . . to send a message to Dublin because he hadn't seen Bertie Ahern at the time, that he was really interested. He was up for it, so to speak, to get this agreement over the line' (D1I8).
Leverage (across fields) 'Irish America and with Clinton very much kind of setting the tone, you know, had rowed in pretty much in a united kind of approach' (R1I16). 'The importance of the European Union could not be exaggerated in relation to the development of closer British/Irish relations and beyond that to the improved atmosphere which made it possible for the two Governments to drive the peace process in Northern Ireland' (R1I3).

Banner (the British Army's operation in Northern
Ireland) ended and Northern Ireland's police were tasked with a radical process of change and reform (Murphy, 2013). The institutional outworking of the Belfast Agreement further cemented inter-field engagement, by establishing the reinforcing institutional infrastructure further supporting the Frame. However, significant differences over the 'decommissioning' of paramilitary weapons stalled for a time the creation of new democratic institutions in Northern Ireland itself. Both governments, the USA and the EU engaged to provide a mechanism whereby the decommissioning of weapons could be verified. For an extended period of time, significant concerns that paramilitary organizations were still operating dominated the political landscape and key constellations of actors, including both governments and the international community more widely, were still investing time, energy and attention in the process. This included the appointment of high-profile British and Irish officials and politicians acting as brokers, and the appointment of a new US special envoy in 2003. However, ongoing disagreements and a perceived failure of the centre parties to negotiate effectively led to an electoral drift to the extremes. Once again, investment from key actors in frame reproduction and process bricolage (a new set of negotiations) resulted in the St Andrew's Agreement of 2006: a recombining of existing elements in a new way to suit an evolving political landscape and to preserve the successful framing. This was an indication of the continuing bipartisanship of the two governments and sustained engagement from them and the USA and EU. However, ongoing street conflicts over identity (such as the Flags protest of 2012), continued activity of dissident republicans, the murders of two police officers and a prison officer, and the 'legacy of the past' continued to require the Frame to be constantly repositioned. These elements of challenge reasserted the previous contested framing of Unionist predominance or a forcibly united Ireland.
For a summary of this period, see Table 4.

Power sharing, Brexit and the Irish Border 1998-ongoing
The NICPP has required significant political, institutional and cultural reinforcement to keep it on track. However, until recently, the master frame of the three strands or relationships, established over a long time period and reinforced by substantial Frame reinforcing 'I've always taken the view that it's going to take at least three generations and that the two governments need to be like that over those three generations and that's why I'm so worried about Brexit. It's not the North, it's the London relationship' (R1I5). 'I think the biggest initiative that came out of that period, other than the summit, was, well, obviously the wins were being protected. We protect the (North South) bodies. We continued routine work like the all-Ireland pension scheme . . . So that was a small institution-building thing' (R1I2).
Truce (see Rao and Kenney, 2008, p. 356) 'On the one hand, the whole thing is incredibly, was and is incredibly precarious . . . but on the other hand, there's no doubt that individual people played a role, a very important role, at different times, and there were some people without whom it's hard to imagine the whole thing going as well as it did' (R1I8).
Dissemination and translation (see Campbell, 2004) 'We're all entrepreneurial and out and about and we can be going to a conference here and there and here they are sort of locked into making the process work' (R1I2). 'Policing is about much more than crime fighting. Why did I brief the American president for 7 years in a row about policing in Northern Ireland? Because America was very important to the peace process in Northern Ireland' (R3I32).
Socio-political legitimacy 'It had been established and that frame needed to be, you know, respected because it was part of the tradition. So, I kind of started to think then of, you know, the Northern Ireland process as, like, a tradition that has to be slowly developed, you know, linguistically and a couple of new ideas in context, and slowly move forward step by step, peace comes dropping slow' (R1I6). 'There is something about quiet conversations; people kept saying "why haven't you written a book?" But the book that would be interesting is the one I'd never want to write. Because it's the stuff around how do you learn about a place. And it's through all sorts of third parties, for a better description. You get to talk to people who historically have never talked to cops and people who wouldn't want those conversations made public, even now, it would be unfair and wrong, but you got a sense of their perception of this and my understanding of the reality of this. And how you narrow that gap' (R3I32). ' . . . what does it mean in terms of good policemen -it means nothing. But in terms of the symbolism of Northern Ireland about change itself it was highly significant in that it showed that change could be achieved without walking all over both traditions, either or all of the traditions in the North of Ireland and secondly it showed that this Board could actually bring that about and drive it through and achieve a win-win situation for everyone' (R3I29). inter-field activity, was not under significant threat. However, the Brexit vote of 2016 represented the first significant hazard to the Frame and its outworking. At the core of the three strands was the elimination of the Irish border as a significant factor in political discourse on the island of Ireland, therefore removing it as a source of contention, violence and ideological radicalization. By making the border invisible, the concerns of Irish Nationalists were diminished and it no longer acted as a metaphorical call to arms for republicans. The peace process and removal of army installations normalized the border and the underlying customs union ensured that it was invisible. Brexit places the border at the very heart of British-Irish relations once again. Other factors have threatened the frame. A change in political leadership within all the UK political parties saw a new generation of leaders with different priorities in the new century and the emergence of an element of peace process fatigue. The financial crisis refocused the attention of both the British and Irish Governments away from peace-making and towards economic stability. A resurgence of identity politics and the manipulation of these contested issues of identity and allegiance by the extremes within Northern Ireland has also hardened the political environment. The Trump White House has none of the focus on Ireland apparent in previous administrations. While the original contested frames of Nationalism and Unionism were suppressed, their resilience as ideological cyphers is not eliminated (Hayward, 2017). Traditionally, frame defence and reproduction were managed by the two governments, engaged in further talks with the political parties, with the occasional addition of an independent chair. Post the Brexit referendum, the British Government's movement from the bipartisanship of the peace process to a reliance on Democratic Unionist Party (DUP) votes in Westminster disrupted that model and created new 'logics of action'. Frame defence has also shifted and now sits principally with the alliance of the Irish Government and the other 26 European states on the issue of the border. The eponymous 'Irish backstop' has become a key sticking point of the Brexit negotiations. To date, the united front of the EU 27 represents not just support for Ireland as a member state, but a defence of the significant amounts of EU funding invested in the Northern Ireland peace process -1.3 billion euros between 1995 and 2013. However, Ireland and the EU are not the only actors engaged in frame defence. Frame advocates such as former British Prime Ministers John Major (signatory to the Downing Street Declaration) and Tony Blair (signatory to the Good Friday Agreement) have acted as Frame protectors both in relation to the border and the importance of governmental bipartisanship. Major has spoken out about both an open border as a necessary requirement of the ongoing peace process but also, importantly, the British Government's 'confidence and supply' agreement with the DUP. 7 Blair has also spoken about the danger to the process. 8 It remains to be seen if the Frame will evolve to manage the disruptions.

Multi Party Agreement
For a summary of this period, see Table 5. 7 He warns that: 'A fundamental part of that peace process is that the UK government needs to be impartial between all the competing interests in Northern Ireland' (The Spectator, 13th June 2017). 8 'A hard border between the countries would be a disaster and I am sure everyone will and must do all they can to avoid it' (Belfast Telegraph, 13th May 2017).

Discussion
This analysis of the NICPP reveals how the framing of system-wide change by institutional actors unfolds over time. Frame contesting inhibited the mobilization of cooperation and the possibility of collective action in system-wide changes that transcend strategic action fields over an extended timeframe, resulting in long periods of 'relative incoherence' (Schneiberg and Clemens, 2006, p. 210) and conflict. Reframing of the NICPP around three sets of relationships enabled actors to negotiate, settle and promote a particular frame (Furnari, 2018;Rao, Morrill and Zald, 2000). Settlement and reproduction on a master frame (Snow and Benford, 1988) resulted in a 'relatively durable truce' among field actors (Rao and Kenney, 2008, p. 356; see Helms and Oliver, 2015;Litrico and David, 2017) and an extended period of relative peace. Actors are engaged in a process of frame defending against the exogenous threat to the frame produced by Brexit, which has the potential to propel the system into another episode of contestation. We now explain the influence of field and inter-field conditions on the four framing mechanisms.

Frame contesting
A dialectical frame emerged from a 'dynamic, purposive and politically-charged process' (Kaplan, 2008, p. 730) embedded in the context of violent ethno-political conflict. Adversarial frame contesting was played out through ongoing interactions (Gray, Purdy and Ansari, 2015). The NICPP supports previous work which has suggested that frame contestation can be heightened in fields that are fragmented rather than centralized (Furnari, 2018). Within the NICPP, actors and actions were widely dispersed (Powell, 1990;Powell, Koput and Smith-Doerr, 1996) and uncoordinated (Meyer, 1982), with elites 'disorganized and possess little influence to change the system' (Rao, Morrill and Zald, 2000, p. 259). Multiple regulatory bodies and state agencies with overlapping jurisdictions (e.g. Suddaby, Cooper and Greenwood, 2007) made it difficult to govern due to the lack of an overarching authority (Schüssler, Rüling and Wittneben, 2014). This frame contesting endured because it was impossible for one side or the other to win and, as such, the field could not Cross-field threats 'That's the real thing that's in danger now is the London/Dublin relationship, because they can't get together in the North' (R1I5). ' . . . unilateral action by one actor (the British or Irish government, the Northern Ireland parties, or the EU itself) inevitably has consequences for the overall "ecological balance" and sustainability of peace on the island of Ireland' (LSE blog post, 17 May 2018).
Purposeful efforts at defence 'I said to X, "It's ridiculous. We have to get a grip on the North", and that's why I went to Belfast' (I2).
Mobilization of defenders 'But it was very difficult to get government departments down here on board with a North/South agenda. They were very suspicious and it was as if you were a raving republican sometimes, the way they would react when you'd say, "Well, what about the North/South dimension and have you considered the implications?" That said, it has grown incrementally and that's why a hard border would be so devastating' (I14). ' . . . if there would be any weakening of the Good Friday accord, then there would be no chance whatsoever -a non-starter -for a US-UK trade agreement' (speech of Nancy Pelosi -Speaker of US House of Representatives at LSE, 16 April 2019).
reform around a winning logic (Reay and Hinings, 2009). Framing contests are typically explained as competition between two logics rather than, as Goodrick and Reay (2011) put it, between 'constellations' of logics. However, similar to previous work, the NICPP shows that fragmentation led to institutional complexity, with a diverse network of actors (e.g. Unionists, Nationalists, state forces, republican paramilitaries, loyalist paramilitaries, Irish Government and British Government) each with their own logic, agenda and micro-level framings. These multiple logics led to contestation as stacks of contradictory and contested frames or 'laminations' (Ansari, Fiss and Zajac, 2010) built up over time. As such, actors confronted 'incompatible prescriptions from multiple institutional logics' (Greenwood et al., 2011, p. 317) that 'are interdependent and yet also contradictory' (p. 250). Institutional complexity impedes attempts at system-wide change (Currie and Spyridonidis, 2015), when competing logics exist over an extended period of time (Reay and Hinings, 2009). Contestation inhibited system-wide change and created a long period of 'relative incoherence' (Schneiberg and Clemens, 2006, p. 210) and conflict.

Reframing
A critical aspect of the NICPP was a reframing of the issues away from specific nameable actors (e.g. Unionist and Nationalist), instead regarding the source of contestation to be 'actorless entities' (Gamson, 1992, p. 32) in the form of three interactions or strands: Britain and Ireland, Northern and Southern Ireland, and the divided people of Northern Ireland. The NICPP also highlights the critical importance of inter-field influences on reframing processes, such as the involvement of US politicians forming a bipartisan position on Northern Ireland. The case contributes to a nascent body of work that suggests reframing can be stimulated when fields have permeable boundaries, making them susceptible to influence from other fields (Zietsma et al., 2017).
Permeable boundaries allow ideas and actors from outside to enter with relative ease, bringing with them practices rooted in logics from other fields (Maguire et al., 2004). Reframing that brings about radical change can emerge or be enabled in the space between fields (Phillips, Lawrence and Hardy, 2000;Rao, Morrill and Zald, 2000) and can spark concurrent, co-constitutive contestation and change in another field (Mora, 2014a). Similar to Evans and Kay (2008), the Northern Ireland case highlights the importance of 'leverage' mechanisms in system-wide change involving the 'judicious use of opportunities available at the intersection of multiple fields'. In previous work, alliance brokerage, rulemaking, resource brokerage and frame adaptation enabled activists to strategically leverage advantages across fields (Evans and Kay, 2008). Similar to previous work, actors who are embedded in multiple fields might be more likely to experience and reflect on contradictory and competing meaning systems (Friedland and Alford, 1991), liberating them and compelling them to act (Seo and Creed, 2002). As such, framing mechanisms that operate across fields can generate concurrent, co-constitutive change within and across fields in ways that are both intended and unintended.

Frame reproducing
The 'three strands' frame received socio-political legitimacy through the endorsement and support of key actors (Carroll and Hannan, 2000;Rao, Morrill and Zald, 2000). Socio-political legitimacy refers to the 'value placed on an activity by cultural norms and political authorities' (Aldrich and Fiol, 1994, p. 648). The master frame based on three relationships was wider in scope and influence than a typical social movement frame (Benford and Snow, 2000) and was 'sufficiently elastic, flexible, and inclusive enough so that any number of other social movements can successfully adopt and deploy it in their campaigns' (Benford, 2013, p. 1), thus affording the opportunity to contribute to system-wide change.
In the NICPP, the master frame was crucial to system-wide change because it cut across multiple institutional orders (Thornton, Ocasio and Lounsbury, 2012, p. 176). Actors recombined symbolic principles and practices that became acceptable and legitimate (March and Olsen, 1989). A process of frame reproduction, involving dissemination (Scott, 1994), diffusion and translation, ensured that the master frame became deeply embedded within and across fields (Campbell, 2004). During the period of settlement and reproduction, the master frame (Gray, Purdy and Ansari, 2015) was pervasive, touching every aspect of society. Rather than a 'surface frame', the case reveals a 'deep frame' that structured actors' 'moral system' and their 'worldview' (Lakoff, 2006, p. 12). Over time, the dominant 'three strands' frame became reconstituted and (re)institutionalized (Gray, Purdy and Ansari, 2015;Greenwood, Suddaby and Hinings, 2002). Similar to Greenwood, Suddaby and Hinings (2002), the NICPP reveals diffusion on the basis of increasing objectification and perceived legitimacy and re-institutionalization as the new arrangement becomes established. Over time the frame became a taken-for-granted (Greenwood, Suddaby and Hinings, 2002) cultural framework (Jepperson, 1991) that made alternatives 'literally unthinkable' (Zucker, 1983, p. 5).

Frame defending
Whilst the NICPP appeared resolvable for a time through cooperation and collective action, a change in context, brought about by Brexit, overturned transient closure and propelled the system into another cycle of contestation and possible volatility. Similar to previous work, external threats to a frame can arise from pressure from forces or parties outside the field (Gray, Purdy and Ansari, 2015). Previous work has shown that the threats to frames include crises, shocks, jolts or discontinuities (Pettigrew, 1985;Tushman and Anderson, 1986), whereby old routines, interests and existing frames are disrupted, requiring a fundamental cultural readjustment (Thornton and Ocasio, 1999). In times of crisis, new 'cultural frames' or 'logics of action' can come into existence (Fligstein, 2011). Previous studies have also highlighted actors engaging in 'disruptive work', attempting to 'loosen' or 'weaken' (Greenwood and Suddaby, 2006) dominant frames by 'undermining assumptions and beliefs', 'disassociating moral foundations' and 'disconnecting sanctions' through changes in legal or professional regulations (Lawrence and Suddaby, 2006, pp. 235, 238). In the NICPP we found that actors became 'institutional defenders' (DiMaggio, 1988) and attempted to maintain the status quo frame. These actors defended the frame against 'actors embedded in multiple fields or their interstices' who can draw on multiple threatening logics that could 'produce transformation and new paths' (Sneiberg, 2007, p. 59). It is important to acknowledge that this is a highly political and even controversial case selection. Many of the details of the Northern Ireland situation are disputed by both actors involved in the NICPP and scholars. Our retrospective interviews relied on individuals' memories and hindsight bias and retrospective sensemaking may alter or obscure details that are important to researchers (Loftus and Hoffman, 1989). It is possible that some of our interviewees, for instance, overstated their participation in the change process or omitted relevant information. We triangulated secondary sources, archival records and interview data to reduce these risks. It is also impossible to provide a comprehensive, rich description of a case that transcends levels of analysis and an extended time horizon within the word length limitations of a journal paper. Our data presentation and narrative is necessarily selective and summarizes key mechanisms, contextual conditions and outcomes. It is also important to recognize that framing is just one mechanism in system-wide change. No doubt, alternative approaches could identify other important mechanisms. Framing is just one part of the picture, but our analysis revealed to be an important part and one which deserves to be better understood.

Conclusion, contributions and future research agenda
Here, we outline five ways in which our study addresses current gaps in our understanding of framing mechanisms in system-wide change. First, the existing literature has yielded a wealth of insights about the 'top-down' organizational and institutional factors that structure and constrain transformation and reinforce persistence rather than enable change. A central concern has been the question of 'how field-level logics shape frames, schemas, and narratives (a top-down approach) while failing to recognize that frames, schemas, and narratives can also originate through bottomup processes that may aggregate and "amplify" to challenge and reshape extant logics' (Purdy, Ansari and Grey, 2019) and influence system-wide change. Framing mechanisms involve 'interactions and processes through which meanings and practices are initiated, reconstituted, or instantiated at multiple levels of the system' (Purdy, Ansari and Grey, 2019). Our study has helped to elucidate the 'bottom-up' framing processes that help explain how the instantiation of an idea can unlock a period of contestation, and how framing subsequently diffuses, reproduces, scales up and opens up possibilities of system-wide change where stasis had been a strong legacy feature.
Second, frames are not static entities, but develop through an unfolding process of ongoing change. We have sought to link theoretically the context, mechanisms, processes and outcomes of challenging change. More specifically, we examine how and why such complex and precarious processes emerge over time through four mechanisms: frame contesting, reframing, frame reproduction and frame defending. Each of these mechanisms is agentic, dynamic, purposive and politically charged. We expose our data through four critical phases illustrating how, why and when the four types of framing mechanisms created the possibility of system-wide change. We show how microlevel (actor) framing can amplify into widely shared, taken-for-granted and more enduring meaning systems. We build on a tradition of processual change research that has given emphasis to time, history, process and action and has examined the links between change processes and outcomes (e.g. Pettigrew, Woodman and Cameron, 2001). Our study highlights the role of ideas and framing mechanisms in delivering change outcomes and the importance of time, history, context, configuration, power and process in shaping the development and resolution of big societal issues of the day. System-wide changes are historical processes and products. Crucially for any processual analysis, the 'interchange between agents and contexts over time is cumulative. The legacy of the past is always shaping the emerging future. What happens, why and how it happens, and what result it brings about is dependent on when it happens' (Pettigrew, 2012(Pettigrew, , p. 1315. The only way to reveal the relationship between multiple levels of context is to have a time series sufficiently long to expose the natural history of development of a systemwide change process over time. Examining the NICPP by integrating studies from the findings from several studies afforded this opportunity. Third, the study reveals the importance of reframing in addressing fragmentation and institutional complexity, with a diverse network of actors (e.g. Unionists, Nationalists, state forces, republican paramilitaries, loyalist paramilitaries, Irish J. Murphy et al. Government and British Government) each with their own logic, agenda and micro-level framings. In this case, the reframing of the perception of the problem from 'two tribes' to 'three strands' created both more complexity but also more malleability and allowed other mechanisms to play out in the subsequent process. We also demonstrate how and why the mobilization of cooperation and the delivery of contextually appropriate collective action are critical to the achievement of outcomes in system-wide change processes. A critical aspect of the NICPP was a reframing of the issues away from specific nameable actors (e.g. Unionist and Nationalist), and instead regarding the source of contestation to be 'actorless entities' (Gamson, 1992, p. 32) in the form of three interactions or strands: Britain and Ireland, Northern and Southern Ireland, and the divided people of Northern Ireland. Reframing of the NICPP around three sets of relationships enabled actors to negotiate, settle and promote a particular frame (Furnari, 2018;Rao, Morrill and Zald, 2000). Settlement and reproduction on a master frame (Snow and Benford, 1988) resulted in a 'relatively durable truce' among field actors (Rao and Kenney, 2008, p. 356;see Helms and Oliver, 2015;Litrico and David, 2017) and an extended period of relative peace. In the NICPP we also found that actors became 'institutional defenders' (DiMaggio, 1988) and attempted to maintain the status quo frame. These actors defended the frame against 'actors embedded in multiple fields or their interstices' who can draw on multiple threatening logics that could 'produce transformation and new paths' (Schneiberg, 2007, p. 59).
Fourth, we highlight how framing mechanisms are shaped by field and inter-field conditions. The NICPP case well illustrates the role of multiple fields in both inhibiting and propelling systemwide change. The NICPP is an example of a strategic action field (as defined by Fligstein and McAdam, 2012, p. 9) that is embedded in a system of fields -the Northern Ireland conflict and peace field itself and border counties in the Republic of Ireland. As such, the NICPP field permeates and intrudes on every other field in Northern Ireland. We identify the field and inter-field conditions that shape these framing processes and draw attention to the transient nature of the outcomes of system-wide change. The NICPP also highlights the critical importance of inter-field influences on reframing processes, such as the involvement of US politicians forming a bipartisan position on Northern Ireland. The case contributes to a nascent body of work that suggests reframing can be stimulated when fields have permeable boundaries, making them susceptible to influence from other fields (Zietsma et al., 2017).
Fifth, our intention is to encourage management and organizational scholars to elevate their gaze to the system-wide changes so emblematic of contemporary society and to offer a workable future research agenda for the study of system-wide changes. This kind of research demands access to multiple stakeholders, sometimes political and other elites, and above all time-series process data across various levels of context. The present study was substantially enabled by its location in three related studies in the same setting and by its rich complement of primary and secondary data. As such, the paper may provide some methodological guidance and encouragement for other scholars wishing to study system-wide change processes. We suggest that any programme of research on systemwide change might build over two phases of learning by doing. In phase one we envisage single case studies examining the role of framing system-wide change in particular settings, for example health, environmental change or mass migration and initially delivering pattern recognition studies of why, how and when change does or does not occur in those settings. The goal of such research will be to reveal the natural history of system-wide change processes over retrospective and real time. The analytical base of such 'phase-one' studies should be informed by antecedent questions, context questions, evolution and development questions, sequencing questions, pace questions and eventually in the 'phase-two' work consequence and outcome questions (Pettigrew, 2012;Pettigrew, Woodman and Cameron, 2001).
Phase-two studies could then accrue from the comparative analysis of such case studies across settings. These kinds of studies could reveal empirical patterns in the sets of mechanisms present or absent in such change processes, the sequencing and timing of the appearance of the mechanisms and the relationship between variations in contextual conditions in the strategic action field and the related patterns in the change process (Fiss and Zajac, 2006). This kind of pattern in the process change study could then move on to the much more ambitious and revealing work which seeks to link changes in context and process to variations in outcomes (Pettigrew, 2012;Pettigrew, Woodman and Cameron, 2001). Here there will be the usual challenges about the conceptualization, operationalization and measurement of the outcome variables, the weighting of factors at both the context and mechanism level of analysis and the linking of analysis of the processes to the outcomes of the change processes.
We argue that the understanding and practical appreciation of system-wide change needs to become an essential part of the field of management and organization studies. These kinds of research themes and questions will demand data-sensitive and date-intensive studies likely to involve, amongst others, societal and policy elites. Whereas some of these studies can and will be attempted by aspirational individuals, others may necessitate the recruitment and perpetuation of teams of scholars sometimes located across disciplines and fields in particular societies and sometimes on a cross-national scale and scope. The study of big themes may necessitate the recruitment and development of big teams. These kinds of leadership and intellectual challenges will be culturally and structurally demanding in the present international scholarly environment (Pettigrew and Starkey, 2016). But if our research endeavours are to be phenomenon driven (Hambrick, 2005, p. 124), prescient (Corley and Gioia, 2011) and impactful in academia and more broadly in society, then some will have to challenge the constraining effects of contemporary academic cultures.