Peace by procedure: civil servants, metagovernance and the Northern Ireland peace process Open Access

The Belfast/Good Friday Agreement of 1998 and its supplementary accords [¹] marked the formal settlement of Northern Ireland’s protracted conflict. The institutional architecture of compromise and power-sharing that emerged generated cautious optimism for durable peace (Fenton, 2018; Byrne et al., 2023). Indeed, the Belfast/Good Friday Agreement of 1998 marked a new beginning after nearly 30 years of conflict between the Republican nationalist communities who wanted a united Ireland and the Loyalist unionist communities who aimed to keep Northern Ireland as part of the UK. The Republicans were primarily represented by the Provisional Irish Republican Army (IRA) and their political wing Sinn Fein, and the Loyalists were represented mainly by the Ulster Defence Association. Beyond the paramilitaries, the UK government had a range of actors working in Northern Ireland during the conflict, from the British Army, the Northern Ireland police force the Royal Ulster Constabulary (RUC), the UK Home Office and most significant for this paper: the Northern Ireland Office. Furthermore, from the 1980s onwards, the Irish Government attempted to have a greater say in Northern Irish affairs, primarily through its diplomatic wing the Irish Department of Foreign Affairs (DFA). In essence, the conflict was between the Provisional IRA and the UK state, however, it also encompassed a deadly contest between the Provisional IRA and the Loyalist paramilitaries who sought to counter the Provisional IRA and safeguard Northern Ireland’s union with Britain.

More than 25 years later, the settlement has held: Northern Ireland has not relapsed into conflict (O’Leary 2018). Yet optimism has been tempered by persistent fragilities. The power-sharing executive has repeatedly collapsed, most recently amid the destabilising effects of Brexit (Murphy, 2019; Murphy and Evershed, 2022; Whitten 2024). Paramilitary networks, contested legacies of violence and unresolved disputes over language and identity continue to obstruct reconciliation (Brown and Mac Ginty, 2003; Aiken, 2009; McVeigh, 2015; Costello, 2022). Belfast’s so-called “peace walls” remain in place, and integrated education has advanced only incrementally (Brooah and Knox, 2013). These limits underscore that the institutional design initiated in 1998 was neither crisis-resilient nor irreversible.

These challenges illustrate why the Northern Ireland peace process warrants renewed investigation. While the history of the process is often narrated through episodes of strategic political agency, it also reveals sustained engagement by the state apparatuses of both the Republic of Ireland and the UK. Throughout the conflict and its aftermath, departments within the British, Irish and Northern Irish civil services were deeply involved in bureaucratic practices aimed at ending violence. These included drafting texts designed to be deliberately ambiguous yet workable in practice and cultivating durable cross-border professional networks that became embedded within administrative cultures (Painter and Peters, 2010; MacCarthaigh and Saarniit, 2019). Yet, despite a rich body of scholarship on political leadership, diplomacy and security policy (Ruane and Todd, [1996] 2009; Gillespie, 2014; Murphy, 2014; Coakley, 2017; Leahy, 2020), the bureaucratic infrastructures that sustained cooperation and enabled political breakthroughs remain comparatively underexplored.

This article addresses that gap. It asks: how did Irish and British civil servants contribute to the construction of the political and procedural conditions necessary for peace? It argues that civil servants shaped the governance environment for peace not only through technical competence, but also by sustaining routines, networks and institutional memory that allowed political agreements to take root. In doing so, the article demonstrates that peacebuilding was not solely a political process but also a bureaucratic one.

To advance this claim, the article draws on the concept of metagovernance. Peacebuilding scholarship has recognised the “technocratic turn” (Mac Ginty, 2012), whereby bureaucratic logics of neutrality and efficiency increasingly shape interventions. Yet this literature has largely neglected the implementers themselves (Williams and Mengistu, 2015). Prevailing frameworks of governance, multilevel governance (MLG) and network theory shed light on structures and interactions but struggle to capture how actors orchestrate the conditions of interaction. Metagovernance, by contrast, allows this article to conceptualise civil servants as simultaneously embedded within governance as intermediaries and above it as designers of institutional arrangements. This framing illuminates how bureaucrats functioned as metagovernors, producing what this study terms as “structures of continuity”: routinised practices and professional networks that preserved institutional knowledge and provided stability amid recurrent political breakdowns. The authors are mindful of the risk that the concept of metagovernance might be interpreted as suggesting that civil servants acted in the absence of political direction. This is not the argument. On the contrary, the notion of metagovernance presupposes the existence of government and political authority, within which civil servants operate. Throughout the article, the analysis of civil servants as active policy actors should not be taken to imply autonomy from political leadership, but rather an embedded role within broader governmental strategies and frameworks. This is specified in the theoretical section.

This article thus contributes to debates in both Public Administration and Peace and Conflict Studies. It extends understanding of bureaucratic peacebuilding and provides a conceptual vocabulary for analysing how administrative routines sustain political settlements. While rooted in Northern Ireland, the analysis speaks more broadly to other divided societies where civil servants play a critical but often overlooked role in shaping pathways out of conflict.

The article proceeds as follows: it first develops the conceptual framework linking peacebuilding and metagovernance; it then sets out data and methods. The subsequent empirical analysis focuses on three domains of bureaucratic metagovernance: (1) the crafting of informal infrastructures; (2) the architecture of embedded cooperation, exemplified by the Maryfield Secretariat; and (3) informal bureaucracy and embedded intelligence, notably through the Irish DFA’s “traveller system”. The conclusion reflects on what it takes to metagovern the transformation from conflict to peace.

Before outlining the theoretical approach guiding this analysis, a brief clarification is in order. In this article, the term civil servant refers to career officials employed within the public bureaucracies of the UK, the Republic of Ireland and Northern Ireland. This includes staff in the Northern Ireland Office, the Northern Ireland Civil Service (NICS), the Irish DFA and the British Home and Cabinet Offices. It does not include elected politicians, political advisers or temporary envoys. The focus is on career administrators whose permanence and institutional embeddedness shaped their ability to sustain continuity and influence the governance environment. These are also the actors placed under the spotlight within the emerging literature on routine peace (Mac Ginty, 2012; Auteserre, 2014).

Debates on the “technocratic turn” in conflict management often assume that public bureaucracies provide neutrality and efficiency in managing issues within deeply-divided societies. As Mac Ginty (2012) notes, the logics of efficiency, neutrality and bureaucratic routine increasingly shape not only the design of peacebuilding interventions but also the forms of peace they produce (Richmond, 2010; Pankhurst, 2007; Bjorkdahl and Höglund, 2013; Mac Ginty and Firchow, 2016; Brett, 2022).

Paradoxically, however, the very actors most closely associated with these logics – public bureaucrats – have attracted relatively little systematic attention. While scholars acknowledge the near-universal reliance on bureaucracies to deliver peace policy (Mengistu and Vogel, 2006; Williams and Mengistu, 2015; O’Connor, 2014), research has largely highlighted their shortcomings in implementation rather than their constitutive role in shaping political outcomes. This neglect produces a paradox: peace is never only negotiated or imposed; it is also designed, administered and sustained. Yet little is known about how civil servants behave in practice or how their routines, values and professional networks shape the environments in which compromise becomes possible. Calls for closer engagement between Political Science and Public Administration (Esmann, 1999; O’Connor, 2014, 2017) have noted that the design and performance of public bureaucracies matter as much as the policies they implement. Despite such interventions, the study of bureaucratic agency in peacebuilding remains marginal.

One reason for this lacuna lies in the limitations of the frameworks most often mobilised to analyse peacebuilding through technocratic lenses. Governance approaches (Bollens, 2000; Banche and Flinders, 2004; Brinkerhoff, 2005, 2007; Stephenson, 2013; O’Connor, 2014; Loizides, 2016; Beevers, 2019; Peters et al., 2022) illuminate the rules and practices through which interaction is structured, but they typically portray actors as embedded within arrangements rather than as designers of them. MLG (Marks, 1996; Hooghe and Marks, 2002) highlights scalar complexity, but tells us little about how actors actively orchestrate across levels. Network theory (Börzel, 1998; Blanco et al., 2011; Bsisu and Murdie, 2021; Lagana, 2025) captures relational ties, but struggles to explain how the rules of interaction themselves are constructed or reconfigured.

These shortcomings become especially acute in cases such as Northern Ireland, where governance is both institutionalised and contested. Here, civil servants did not merely implement policy within fixed structures: they moved between formal and informal domains, blurred scalar hierarchies and managed layered territorial logics (Piattoni, 2009). Furthermore, much of the literature treats governance as a neutral “container”, but in practice it can also represent a strategic tool: one that enables or constrains political agency depending on how it is configured (Jessop, 1990; Hay, 2014; Lagana and Pearce, 2025).

To theorise this strategic dimension, the article turns to the concept of metagovernance: the orchestration of governance modes – hierarchical and networked – without reliance on direct command (Jessop, 2001; Sørensen and Torfing, 2009). Metagovernance highlights the agency of actors who shape the rules, boundaries and norms within which governance unfolds. Initially developed as a managerial concept, it has since been extended to encompass both political and administrative dimensions (Sørensen and Torfing, 2020; Lagana, 2021; Lagana and Pearce, 2025). Politically, it captures how elected leaders foster collaborative governance; administratively, it illustrates how bureaucrats use legal frameworks, incentives, discourse and institutional design to influence behaviour and organisational performance (Sørensen, 2020).

Applied to conflict management, metagovernance does not suggest that civil servants directly negotiated political settlements. Rather, it shows how they shaped the environment in which such negotiations became possible. Bureaucratic routines – deploying administrative rules, policy inducements, narrative framing, and organisational architecture – combined with practices of purposeful ambiguity, institutional memory and continuity across political cycles. These strategies did not substitute for political leadership but underpinned it, providing a procedural and organisational architecture within which compromise could be pursued.

In Northern Ireland, metagovernance was less an abstract theory than a practical necessity (Lagana, 2021; Knox et al., 2023; Lagana and Pearce, 2025). British and Irish civil servants were compelled to craft cross-jurisdictional policy environments capable of accommodating nationalist, unionist and cross-community actors without undermining state authority or breaching constitutional constraints. Their toolkit included backchannel diplomacy (Ó Dochartaigh, 2021) institutional choreography, discursive innovation and spatial design, all deployed to sustain the possibility of compromise (Lederach, 1997; Lagana, 2023).

Where governance and network theories gesture towards “structural selectivity”, metagovernance demonstrates how such selectivity is operationalised. It shifts the analytical focus from governing to orchestrating governance, foregrounding the bureaucratic labour that renders negotiated settlements viable. In deeply divided societies, formal authority alone is rarely sufficient. Civil servants acted as metagovernors by designing relational, procedural and spatial frameworks that enabled interaction, constrained conflict and sustained cooperation. Crucially, these practices were not improvised but reproduced through structures of continuity: an ensemble of bureaucratic memory, routinised practices and long-term expertise that preserved institutional knowledge, carried forward cross-border networks and maintained a shared orientation towards stability and devolution even amid recurrent crises.

The Northern Ireland conflict, commonly known as “the Troubles”, spanned three decades of sectarian violence between the nationalist/republican/Catholic and unionist/loyalist/Protestant communities and resulted in more than 3,500 deaths (Williams, 2023, p. 47). The conflict formally ended in 1998 with the signing of the Belfast/Good Friday Agreement, but the long trajectory of institutional and administrative change that preceded it is essential to understanding how civil servants came to act as metagovernors of peacebuilding. Their agency cannot be reduced to technical implementation. It was conditioned by the evolving structures of governance within which they were embedded and through which they strategically exercised influence.

Northern Ireland’s distinctive institutional history provided the foundations for this role. Created in 1921, the region occupied a peripheral position within the British state (Fenton, 2018). Sovereignty was divided: Westminster retained authority over defence and foreign affairs, while devolved institutions in Belfast oversaw domestic services in what Bulpitt (2008 [1983]) described as a “dual polity” (see also Miller, 2007; Stolz, 2024). This arrangement produced two parallel bureaucracies. In Great Britain, executive authority was centralised in the Prime Minister’s Office, the Cabinet Office and the Whitehall civil service (Kenny, 2023). In Belfast, the NICS inherited much of its organisational DNA from the pre-partition Irish administration (Ó Beacháin, 2019). Despite shared meritocratic norms, these bureaucracies diverged in composition and culture. The Northern Ireland Civil Service was long characterised by sectarian imbalance: no Catholic was appointed as Permanent Secretary until the late 1960s (O’Connor, 2017; Rouse, 2018). By contrast, the Home Civil Service in London had comparatively greater Catholic representation (O’Connor, 2017). The outcome was a quasi-federal system with weak vertical integration between Belfast and Whitehall, reinforced by Northern Ireland’s geographic marginality and the legacy of partition.

This governance model was fundamentally restructured with the imposition of Direct Rule in 1972. Escalating violence and political deadlock led to the suspension of the Stormont Parliament and the transfer of authority to Westminster. Although initially presented as temporary, Direct Rule endured for nearly three decades (Ó Dochartaigh, 2005; Birrell and Gormley-Heenan, 2015; Williamson, 2017). Politically, it undermined local accountability: Northern Ireland was governed by ministers without mandates in the region, and no Secretary of State for Northern Ireland has held a Northern Irish constituency since 1972 (Bloomfield and Lankford, 1996). Administratively, however, the transition was marked more by continuity than disruption (Carmichael, 2002). The most significant innovation was the creation of the Northern Ireland Office, a new Whitehall department charged with administering the region and preparing for future devolution. As Sir Quentin Thomas, NIO Political Director between 1991 and 1998, later recalled:

The Northern Ireland Office did not emerge in isolation. It evolved from existing Home Office structures, inheriting personnel and administrative habits shaped by earlier internal arrangements. As [Joseph] Joe Pilling, appointed as Northern Ireland Office Permanent Secretary in 1997, recalled of this formative moment:

This origin story highlights the improvised and pragmatic ethos that characterised the Northern Ireland Office’s early operational culture. Initially a small unit working with minimal oversight on emergency legislation and politically sensitive measures such as internment, the Northern Ireland Office gradually expanded to encompass security, constitutional affairs and political negotiations (Mann et al., 2025). Its location within Whitehall ensured proximity to central government, while its flexible structure allowed adaptation to Northern Ireland’s unique demands. Over time, it evolved into a politically engaged bureaucracy, occupying an ambiguous space between administration and strategic policy leadership (Rhodes, 1997).

The Northern Ireland Office’s emergence created initial tensions with the NICS, which regarded it as an external imposition. Yet pragmatic cooperation soon developed. Secondments, informal exchanges and procedural coordination fostered alignment. While the Northern Ireland Office retained strategic direction, it adhered to the Whitehall principle that local policy delivery should remain with local officials. This hybrid arrangement created space for civil servants to innovate governance mechanisms tailored to Northern Ireland’s contested realities.

The first major test of these arrangements came with the Sunningdale Agreement of December 1973, an early attempt to institutionalise power-sharing and cross-border cooperation. Yet the agreement lacked support from Irish republicans and the majority of unionists and it collapsed in May 1974 amid a general strike led by the Ulster Workers’ Council. Its failure underscored the limits of top-down political engineering and highlighted the necessity of sustained bureaucratic collaboration. From the mid-1980s, cooperation between the Northern Ireland Office and NICS deepened, with informal relationships, shared procedures and a culture of pragmatic adaptation that gradually produced what we describe as structures of continuity: recurrent patterns of intergovernmental coordination sustained through institutional memory and routine practices.

This intra-UK collaboration was soon mirrored in cross-border developments. The 1985 Anglo-Irish Agreement, signed by Prime Minister Margaret Thatcher and Taoiseach Garret FitzGerald, institutionalised a formal Irish role in Northern Ireland’s affairs. It provided mechanisms for cooperation between UK and Irish officials on security and cross-border issues, most notably through the creation of the Maryfield Secretariat. Maryfield functioned as a shared administrative hub that enabled continuous cooperation between the Irish DFA and the Northern Ireland Office. It provided Irish civil servants with continuous access to Northern Ireland’s political and social dynamics.

By the early 1990s, a core group of officials on both sides had accumulated the institutional memory, relational capital and political judgement required to operate across formal hierarchies, informal networks and contested territorial spaces. Their role extended beyond implementing decisions taken elsewhere, shaping procedures and practices that structured interaction and made political compromise more feasible. In this sense, Irish and British civil servants acted not merely as administrators but as metagovernors: architects of the institutional environment that made peace possible.

Metagovernance encompasses more than the formal design of governance systems; it also involves the cultivation of informal infrastructures – backchannels, trusted relationships and procedural workarounds – that quietly enable cooperation behind the scenes (Klijn and Koppenjan, 2016). In the context of Northern Ireland, these informal infrastructures proved essential in creating the conditions for formal agreements to emerge. This section explores one pivotal episode from the early 1980s, when Irish and British civil servants initiated structured engagement during a period of acute political hostility.

At the time, relations between British Prime Minister Margaret Thatcher and Taoiseach [³] Charley J. Haughey were deeply strained. The 1981 hunger strikes had cast a long shadow (Guelke, 1988; Cox et al., 2006; Lagana, 2025), and Taoiseach Charles Haughey’s controversial stance during the 1982 Falklands War had further damaged trust. His successor, Garret FitzGerald, a much more positive relationship with Thatcher.

While informal diplomacy had long been part of Anglo-Irish relations, what unfolded in the run-up to the 1985 AIA was qualitatively different: it was deliberately initiated and strategically managed by senior civil servants with support, authority and direction from their respective heads of states (Todd, 2011). These officials also benefitted from the quiet encouragement of international actors – particularly the USA and the then European Community – who applied subtle pressure on the UK to engage more constructively with Dublin (Mac Ginty 1997; Tannam, 1999; Cochrane, 2007; Kelly, 2020, 2021; Lagana, 2021; Litter, 2023).

In this context, a small group of Irish DFA officials began reopening channels of communication with British counterparts. Rather than wait for a political green light, they moved first, re-establishing informal contact to rebuild trust and test new ideas. These efforts quickly coalesced into a discreet series of face-to-face meetings between Irish and UK civil servants, most notably Michael Lillis (Irish DFA) and David Goodall (UK Cabinet Office). While neither held a formal negotiating mandate, both were trusted insiders with the authority to speak informally and interpret political signals. Their goal was not to negotiate an agreement, but to create the scaffolding that might allow one to emerge.

The first concrete step in this backchannel came during a walk along Dublin’s Grand Canal in September 1983, after a formal meeting of the Anglo-Irish coordinating committee (Goodall and Sheridan 2021, p. 161–163). Lillis invited Goodall for a stroll and used the opportunity to sketch out what he described as possible new arrangements for Northern Ireland. As he made clear, these were not official proposals from the Irish government, but reflections of the Taoiseach’s evolving thinking (interview with Michael Lillis, online, 25/01/2025):

This informal conversation was anything but casual. It was a tactical use of unstructured space, outside the constraints of protocol, to float radical ideas, test boundaries and prepare the ground for more formal exchanges. It was, in short, an act of metagovernance: shaping the political environment, not by issuing commands, but by subtly expanding the field of possibility.

Goodall’s report drew Thatcher’s attention and prompted further exploration. Despite her scepticism, she authorised further exploration and eventually asked her Cabinet Secretary, Robert Armstrong, to lead the British side of what would become the Anglo Irish Agreement negotiations (Thatcher, 1993, p. 395). As Armstrong later recalled:

For nearly two years, Armstrong and his Irish counterpart Dermot Nally ran this channel, while quietly orchestrating the architecture of the Anglo Irish Agreement. Crucially, the Northern Ireland Office was excluded from these early negotiations, due to fears that its close ties to unionism might obstruct progress (Maryfield Witness Seminar 2015, p. 25). This exclusion would not last, but it underscores how the civil servants leading the early process operated in the shadow of formal hierarchy (Scharpf, 1994; Medzini, 2021).

This episode reflects a core dynamic of metagovernance: creating low-risk, low-visibility arenas in which dialogue can be initiated and trust can be built, even amid deep political antagonism. By shifting early discussions into informal arenas, officials enabled exploration away from immediate political pressures. They did not dictate outcomes, but they framed the conditions under which meaningful negotiations could begin.

Their efforts also helped shift institutional mindsets within Whitehall. Civil servants such as Goodall and Armstrong later stressed that mutual trust, not political affinity, enabled cooperation during moments of tension:

In these reflections, trust functioned less as a personal virtue than as an operational resource, sustaining working relationships amid political volatility.

By decoupling early engagement from formal institutions, these officials carved out a space in which compromise could be imagined: the genesis of what we call a structure of continuity. When overt relations between political figures failed, the structures that tied civil servant’s work together allowed for a continuity in progress. For example, in 1982 when relations between the UK and Irish leaders were at an all-time low after a controversial speech by then Taoiseach Charles Haughey, UK-Irish relations were completed frozen. Indeed, O’Malley (2025, p. 139) mentions the only part of UK-Irish relations that continued during the diplomatic disaster was the “informal talks between Dermot Nally, the Irish government secretary and Robert Armstrong, the UK cabinet secretary”. This illustrates how senior civil servants were able to raise above the noise of politics and continue their work covertly through the structures of continuity. Their work was invisible to the public and often unofficial, but it was foundational. These practices established relational and procedural patterns that informed later negotiations and introduced a reframing of the conflict in terms of institutional accommodation, mutual recognition and shared responsibility over zero-sum sovereignty and hard security.

In this sense, the civil servants involved were not simply intermediaries. They acted strategically, using discretion and sensitivity to timing to expand what was politically thinkable.

Civil servants operated as metagovernors during the Northern Ireland peace process particularly when institutional arrangements allowed them to embed cooperation into the everyday routines of governance. These embedded settings sustained the relational, procedural and spatial configurations that enabled Irish and British officials to shape, rather than merely implement, the evolving peace architecture. Among these, the Anglo-Irish Secretariat at Maryfield stands out.

Established under the 1985 Anglo Irish Agreement, Maryfield exemplified a spatially embedded structure of continuity: a physical site that combined formal mechanisms of joint coordination with informal practices of trust-building across state boundaries [⁴] The Secretariat was located on the outskirts of Belfast [⁵] and was commonly referred to by its staff as “the bunker” [⁶]. Interestingly, it was a NICS secondee to the Northern Ireland Office, Chris Maccabe, who was tasked with finding the building that would eventually house the Secretariat. He later recalled being asked to identify a site large enough to accommodate extensive security requirements for a new organisation [⁷] but when he inquired who in Northern Ireland knew of these plans, the answer was: “nobody” [⁸]. The rule was that no one locally should be aware of the impending arrangements [⁹].

This secrecy was deliberate. British officials feared that knowledge of the new framework would mobilise unionist opposition before it had even begun [¹⁰]. The Secretariat was to be staffed and jointly led by British and Irish officials, meaning that Irish civil servants would, for the first time, be based in Northern Ireland – an innovation unlikely to be welcomed by large sections of the unionist population. Nevertheless, the arrangements moved forward. While each delegation maintained its own offices, staff worked in close proximity on a daily basis, shared tasks and meals and jointly hosted receptions. This enforced cohabitation blurred institutional boundaries, encouraged interpersonal familiarity and facilitated pragmatic cooperation.

As former Irish Head of Secretariat David Donoghue reflected:

Donoghue’s account illustrates how spatial co-location helped gradually pull the Northern Ireland Office into a more deliberative and cooperative mode of bureaucratic engagement. As Irish and British civil servants worked side by side, on both technical and political matters, the boundaries between administration and strategy began to soften. This shift, though subtle, had significant implications.

Former Irish diplomat and traveller Eamonn McKee also emphasised the strategic importance of Maryfield’s design. He noted that the Secretariat’s 24/7 staffing enabled it to respond to immediate crises while also monitoring longer-term dynamics. Its presence in Northern Ireland allowed Irish officials to gain intimate, real-time knowledge of developments on the ground, an arrangement that marked a sharp departure from traditional diplomatic practice. In McKee’s words, this was both symbolically and practically significant: it embedded Irish governance capacity into the daily life of Northern Ireland, and forced the UK government to engage with the conflict in ways that went beyond a narrow security lens (interview with Eamonn McKee, online, 27/03/2024).

This embeddedness – physical, procedural and epistemic – was central to Maryfield’s metagovernance function. Unlike the more centralised institutions of Whitehall, Maryfield enabled both proximity and responsiveness. It generated strategic proximity, allowing British and Irish officials to anticipate challenges, align responses and gradually build a shared understanding of the political terrain.

Dáithí Ó Ceallaigh (2015, p 31) similarly highlighted how Maryfield shifted perceptions within the British state:

Co-location and sustained contact enabled officials on both sides to learn from one another, recalibrate their assumptions and develop an ethos of joint problem-solving. Over time, these working relationships transcended the institutional frameworks that produced them.

Maryfield, then, was not a neutral coordination mechanism. It was a metagovernance site, a place where new bureaucratic habits were formed and new possibilities for cross-border engagement were tested. It illustrated an alternative mode of statecraft. It helped stabilise governance not by imposing new structures from above, but by embedding habits of cooperation into the everyday functioning of administration.

While Maryfield did not produce headline breakthroughs, it transformed the texture of Anglo-Irish engagement. It shifted the environment in which decisions were made, lowered the cost of collaboration and enabled civil servants to play an increasingly active role in shaping the peace process. In this way, it exemplifies the power of bureaucratic cohabitation in divided societies: not as a symbolic gesture, but as a practical tool for managing complexity and sustaining peace.

Alongside formal institutions like the Maryfield Secretariat, Irish and British civil servants developed additional tools and routines that deepened cross-border cooperation and sustained bureaucratic trust. These included social rituals, decentralised knowledge-gathering and informal communication networks that, over time, evolved into a system of embedded frontstage and backstage intelligence. These practices exemplified metagovernance in practice: they enabled civil servants to respond to complexity not through command, but through adaptive, relational and often asymmetric infrastructures of coordination.

One such innovation was the deliberate cultivation of informal routines, which some officials jokingly referred to as the “affective infrastructure” of diplomacy (Williams, 2024). From 1990 onwards, these included the Maryfield jointly hosted receptions [¹¹], where hospitality created opportunities to ease political tensions. Prior to 1990, the hospitality was confined to small lunches and dinners. Though seemingly trivial, these social gatherings provided important spaces for relationship-building among political, security and bureaucratic actors. As multiple participants recalled, the shared informality helped build familiarity and defuse mistrust (interviews with Prof. Chris Maccabe, online, 04/11/2024; Sir Nigel Hamilton – former Head of the NICS – Belfast, 23/ 01/2025). Senior Irish officials such as Michael Lillis and Daire Ó Criodáin noted the importance of these events in enabling frank, off-the-record exchanges with sceptical unionists and RUC officers, who were the police force in Northern Ireland at the time (Maryfield Secretariat Witness Seminar, 2016, p. 11).

These informal rituals complemented a more systematic form of embedded governance: the Irish DFA traveller system. This decentralised network assigned Irish officials to specific domains – politics, security, human rights and economics – and tasked them with regular site visits across Northern Ireland. Travellers built relationships with local leaders, gathered real-time intelligence and translated social and political signals back to Dublin. As former diplomat Eamonn McKee noted:

The system’s value lay not only in information collection but also in interpretation: Irish officials translated lived realities into strategic insight. This produced an asymmetry in bureaucratic awareness. As former traveller and Irish diplomat Ray Bassett put it:

Because many Irish DFA officials involved in formal negotiations had previously served as travellers, they brought with them grounded knowledge of both nationalist and unionist communities. David Donoghue described this dual exposure as essential:

British officials, particularly within the Northern Ireland Office, also relied on Irish-generated intelligence and on the informal interactions that Irish officials, travelling across the border, had with Northern Irish civil servants such as Chris Maccabe. Maccabe – at the time the only NICS secondee permanently retained within the Northern Ireland Office – was frequently deployed to facilitate communication between the British civil service and the unionist community. His longstanding professional experience in Northern Ireland and his embeddedness in local institutional networks, lent him credibility in sustaining dialogue and maintaining open channels of communication. As Irish diplomat Daithí Ó Ceallaigh observed, Irish officials were often well positioned to brief their British counterparts on developments within their own policy environment (Ó Ceallaigh 2015, p 34). Daire Ó Criodáin characterised this interface as a filter and feedback loop: traveller networks gathered information, Dublin curated it and channelled it back into the Northern Ireland governance process (Ó Criodáin 2015, p. 29). These dynamic underscores the reciprocal nature of cross-border engagement: while Irish officials gained access to the workings of British policy, British officials in turn benefitted from curated intelligence and informal access to local political and community sentiment, particularly when mediated through figures like Maccabe who could navigate institutional and societal divides with relative ease.

In parallel to these social and intelligence routines, both British and Irish civil servants also sustained informal lines of direct contact with political actors, particularly during high-stakes periods. Ó Dochartaigh (2011) labels the space in which these interactions occurred as the “backchannel”. The defining feature of such interactions was their covert character; the exclusion of wider audiences was intended to prevent interference, political fallout or reputational damage. Interacting via the backchannel often required trusted intermediaries, and recently released archives and private papers have revealed, among other things, the role played by Brendan Duddy – the intermediary who acted as the primary contact between the Provisional IRA and the British government over two decades (Ó Dochartaigh, 2011).

However, the character of the backchannel evolved over time. By mid-to-late November 1993, the messages that British officials (notably Northern Ireland Office civil servants) had been exchanging throughout the year were leaked and soon published in the Observer, significantly diminishing the veil of secrecy that had previously defined this mode of contact [¹²]. The backchannel, in this case, was no longer secret – what had once been deniable became, in effect, undeniable.

From a Public Administration perspective, what is most striking is the image of a public bureaucratic body like the Northern Ireland Office engaging in informal dialogue with so-called “terrorists” (Powell, 2015), while publicly upholding commitments to transparency and accountability (Steven et al., 2018). To explore how these dynamics contributed to the metagovernance of pathways from conflict to peace, we examine the case of the evolving relationship between Northern Ireland Office civil servant Tony Beeton and Sinn Féin representative Siobhán O’Hanlon in the mid-1990s. Archival records from 1995 show that this channel – though no longer secret – functioned as a quiet conduit for de-escalation, troubleshooting and confidence-building.

On 15 August 1995, for instance, O’Hanlon contacted Beeton to raise concerns about Northern Ireland’s police forces’ behaviour. Beeton, in turn, flagged the political sensitivity of the issue and reflected on how Sinn Féin might seek to leverage the incident around the ceasefire anniversary [¹³]. When tensions flared again in September, Beeton suggested that Minister Michael Ancram personally call Republican leader, Martin McGuinness to steady the ground ahead of their meeting. This informal recommendation reflected the kind of situational judgement that formal channels often lacked.

By 27 September, the Beeton–O’Hanlon relationship showed signs of growing rapport. In one call, O’Hanlon laughed when Beeton joked that McGuinness wouldn’t dare snub Northern Ireland Office senior official Quentin Thomas [¹⁴], who at the time was the Northern Ireland Office political director [¹⁵]. That levity carried into the formal meeting the next day. When discussions became strained, particularly around decommissioning, it was Thomas who tactfully intervened, reframing the conversation and helping move the parties out of impasse [¹⁶].

This episode exemplifies how bureaucrats quietly managed the micro-politics of contact. They were not simply conduits for political decisions; they were actors exercising relational judgement, who understood when to step in, when to de-escalate and how to preserve the integrity of the process. Their actions kept negotiations afloat during moments when the political climate might otherwise have derailed progress. This example was selected for three main reasons. Firstly, it remains relatively underexplored in existing analyses, allowing this article to highlight new documentary evidence. Secondly, it provides a clear illustration of the work of civil servants and is broadly representative of their modes of interaction. Thirdly, it was frequently cited by our interviewees when asked what it took to shape the environment that made the peace process possible.

Importantly, these acts of metagovernance operated asymmetrically. Irish officials often wielded broader societal reach, thanks to the traveller system and the embeddedness of the DFA in Northern Ireland’s daily life. British officials, by contrast, held tighter control over institutional levers but had to rely more heavily on informal contacts and delegated judgement. Yet this asymmetry sometimes created complementarities: it allowed each side to compensate for the other’s blind spots, reinforcing a complementary ecosystem of trust-based governance.

Together, these practices – social diplomacy, embedded intelligence and informal dialogue – constituted a durable infrastructure of metagovernance. They enabled civil servants to stabilise interaction, test possibilities and adapt to fast-moving developments. While often invisible to the public eye, these mechanisms were central to the architecture of peace. The evolution of the backchannel – from deniable to undeniable – demonstrates how civil servants adapted informal instruments of influence into sustained tools of governance.

The architecture of metagovernance described throughout this article was not accidental. It was made possible by structures of continuity: the institutional memory, interpersonal familiarity and procedural durability that only long-serving bureaucracies could provide. In the case of Northern Ireland, the endurance and adaptability of civil servants, on both the Irish and British sides, created conditions in which trust could develop, innovations be carried forward and cooperative environments sustained. Far from being passive custodians of policy, these officials were long-term strategists whose permanence allowed them to maintain, recalibrate and embed practices of cooperation. Their ongoing presence meant that lessons learned from previous failures (such as the collapse of the 1974 Sunningdale Agreement) were not lost to political amnesia.

This bureaucratic continuity underpinned every mechanism detailed in this article. Whether through the institutional routines of the Maryfield Secretariat, the embedded intelligence of the traveller system or the informal channels maintained with Republican interlocutors, civil servants made metagovernance work not by starting from scratch, but by layering experience over time. Each initiative built upon prior efforts, slowly knitting together a web of relationships and expectations that could withstand volatility and political rupture. Structures of continuity were essential within the he Irish DFA specifically too. According to Eamonn McKee (interview, online, 27 March 2024) there was a very deliberate policy of circulating the same officials through Anglo-Irish so that when they came for postings they would be sent to Anglo-Irish related missions, such as the Embassy London and the Embassy and Consulates in the USA. Afterwards they would then return to Anglo-Irish division. According to McKee this approach meant that the Irish side had a real depth of knowledge in specialized areas, including human rights, security and policing, administration of justice and rule of law, not to mention the politics of Northern Ireland, London and the USA and Northern Ireland psephology. This proved valuable when the Talks Team was formed.

These practices also reveal that metagovernance is highly context-specific. It does not operate through abstract templates or fixed hierarchies, but through strategic improvisation, grounded in place, procedure and personality. In the Northern Ireland context, metagovernance had one overriding aim: to foster coordination among actors with historically opposed identities, mandates and narratives. This required more than formal agreement: it required the sustained orchestration of intergovernmental relations, affective trust and discursive re-framing.

By the late 1980s, the cumulative effects of these embedded practices began to show. A shared infrastructure of knowledge, interpretive cues and strategic language emerged across Irish and British bureaucracies. A defining moment in this evolution was the 1990 Whitbread speech delivered by Secretary of State Peter Brooke (Wincott, 1992). Widely regarded as a catalytic moment in the peace process, its core message – that the British government had “no selfish strategic or economic interest” (quoted e.g. in Wincott, 1992, p. 3; Spencer, 2008, p. 460) in Northern Ireland – resonated deeply with Republican audiences. But the significance of the speech lay not only in its content, but in its construction.

Behind the scenes, civil servants on both sides worked meticulously to shape its legal, constitutional and symbolic dimensions (interview with former Northern Ireland Office civil servant Prof. David Cooke, London, 14/12/2024). This was a clear example of discursive metagovernance: the use of bureaucratic expertise to reframe conflict narratives, lower rhetorical temperatures and open new political horizons. As revealed in Todd and Coakley’s (2020, p. 254–255) participant seminars, the wording of the speech was the product of intense negotiation and cross-jurisdictional sensitivity. Senior Irish civil servant Dermot Nally, for instance, famously observed that “there is no ‘political’ in the Whitbread speech, there is no comma in the Whitbread speech” (Todd and Coakley, 2020, p. 254–255). A reminder of how textual precision was used to navigate contested meanings.

This interpretive care was mirrored by British counterparts. Sir Robin Butler, Cabinet Secretary at the time, explained that even the placement of a comma carried weight:

Such grammatical choreography reveals the depth of bureaucratic fluency in crafting texts that could simultaneously reassure unionists, signal openness to nationalists and pass constitutional muster in both London and Dublin.

In this instance, civil servants acted not merely as speechwriters or legal advisors, but as key actors in discursive reframing that supported the peace process. The Whitbread speech was not simply a political gesture; it was the culmination of sustained bureaucratic collaboration that had developed through years of informal contact, shared spatial infrastructures and cross-border embeddedness. It illustrates how co-production and bureaucratic drafting became tools of peacebuilding, allowing carefully calibrated messages to reach diverse audiences and support convergence beyond the limits of formal diplomacy.

In sum, structures of continuity enabled more than institutional resilience: they enabled creativity, adaptability and discursive innovation. Civil servants were able to carry lessons forward, maintain relational capital and operate across political transitions. Their permanence was not a relic of bureaucratic inertia, but the foundation of a metagovernance that was flexible, relational and contextually attuned. In deeply divided societies, where political leadership is often unstable or polarised, such bureaucratic memory becomes not only useful but indispensable. It is a foundation upon which more durable peace processes could be built.

This article has shown that peace in Northern Ireland was not only negotiated at summit tables or forged through grand political bargains: it was cultivated in corridors, sustained in offices and embedded in routines. Irish and British civil servants, through their steady presence, procedural dexterity and adaptive strategies, underwrote the conditions of possibility for peace. By applying the lens of metagovernance, we have demonstrated how these actors operated not simply as bureaucratic implementers, but as designers of interaction, shaping relational, spatial and discursive infrastructures that enabled cooperation across conflict lines.

Central to this analysis is the concept of structures of continuity: the informal, adaptive and enduring bureaucratic practices that functioned as connective tissue in the peace process. Continuity, however, was not only procedural but also institutional. In contrast to present-day patterns of accelerated turnover, many civil servants – particularly within the Northern Ireland Office – tended to spend the entirety of their careers within the same office. Advancement to the highest ranks usually required at least a decade of service and experience across multiple policy domains. This long tenure generated a depth of expertise that was both cumulative and transferrable: senior officials passed knowledge to newly posted colleagues, while also developing locally grounded understandings of Northern Ireland’s political, social and security environments. Without this embedded expertise and the steady transmission of institutional memory, bureaucratic continuity would have been far weaker and the capacity to sustain peace-enabling practices far more fragile. Through these routines of cohabitation, embedded intelligence gathering, trust-building and discursive calibration, civil servants managed volatility, interpreted shifting political signals and co-produced the frameworks through which peace could be imagined, tested and institutionalised.

What does this tell us about the nature of metagovernance in deeply divided societies? The Northern Ireland case reveals that to act as a metagovernor is to sustain continuity amid disruption; to provide procedural stability when politics is volatile; and to translate uncertainty into routines that allow adversaries to interact without collapse. It requires leveraging institutional memory, cultivating cross-cutting relationships and embedding practices of responsiveness into everyday governance. In short, the metagovernor is neither a neutral administrator nor a political negotiator, but an actor whose authority derives from the ability to orchestrate the conditions under which political bargains can endure.

The implications extend beyond Northern Ireland. While this article has not sought direct comparison, the theoretical lessons suggest ways to analyse the often-overlooked role of bureaucrats in other conflict-affected contexts such as Colombia, Ukraine and Kosovo. Recognising the strategic labour of civil servants allows scholars of Public Administration and Peace and Conflict Studies to better account for how the state apparatus can contribute to conflict management. The framework advanced here provides a vocabulary for systematically examining how bureaucratic memory, informal infrastructures and adaptive routines stabilise interaction where political leadership is unstable or polarised.

In rethinking bureaucracies as agents of transformation rather than neutral backdrops, we highlight conflict management not as a technocratic fix but as a cultivated process: one that depends on invisible scaffolding, long-term trust and embedded practices of cooperation. In contexts of deep division, metagovernance should therefore be understood not as a managerial ideal but as a peacebuilding necessity.

Bassett, Ray. He is a former Irish diplomat who served as Ambassador to Canada, Jamaica, and the Bahamas, and as Joint Secretary to the British–Irish Intergovernmental Conference in Belfast. A graduate of Trinity College Dublin, he holds a PhD in biochemistry. Following his retirement in 2016, he became Senior Fellow for EU Affairs at the think tank Policy Exchange. Bassett is the author of several works on Brexit and Irish foreign policy, including Ireland and the EU Post Brexit (2020), in which he offers a critical assessment of Ireland’s approach to Brexit negotiations. Interviewed in Dublin on 23 February 2024.

Cooke, David. He is a former Director of the British Board of Film Classification (BBFC), a role he held from 2004 until March 2016. Before that, he was a senior civil servant involved in areas including the Northern Ireland peace process, devolution, asylum, criminal justice, and broadcasting policy. He played a key role in drafting the Downing Street Declaration, working closely with Sir Quentin Thomas. Interviewed in London on 14 December 2024.

Donoghue, David. As part of his long career in the DFA, he was as involved for many years in the Northern Ireland peace process. He was one of the Irish Government’s negotiators for the ground-breaking Belfast/Good Friday Agreement (1998), which has provided a political framework for lasting peace and stability in Northern Ireland. Interviewed online, on the 21 November 2023.

Lillis, Michael. He was a senior Irish diplomat central to Anglo-Irish relations. In the 1970s, he helped draft the Carter Initiative with John Hume and key US figures, despite British and US opposition. As Diplomatic Advisor to Taoiseach Garret FitzGerald, he launched talks with Sir David Goodall in 1983 that paved the way for the 1985 Anglo-Irish Agreement, where he served as one of Ireland’s lead negotiators. Interviewed online, on the 25 January 2025.

Hamilton, Nigel. He was a senior civil servant form the NICS who played a significant role as Head of Central Secretariat, liaising between the NIO, NI Departments, and the Irish during some of the most formative periods of both Direct Rule and political discussions. When devolution was restored in 2005, with Ian Paisley and Martin Mc Guinness as First and Deputy First Ministers (to major local, national and international surprise) his role was to make the new institutions work as Head of the NICS and Secretary to the newly established Executive. He was awarded a knighthood in 2005 for his contribution to public service. Interviewed in Belfast on 23 January 2025.

Maccabe, Chris. He retired in 2008 after more than eight years as Political Director of the Northern Ireland Office (NIO) and nearly four decades in public service. He began his career in 1971, later serving as special assistant to the head of the Royal Ulster Constabulary and as Director of the Northern Ireland Prison Service. Returning to the NIO in 1992 as Head of Political Affairs, he played a key role in the negotiations leading to the 1998 Belfast/Good Friday Agreement and the establishment of power-sharing in 2007. He has since advised on peace processes in Sri Lanka, Kosovo, and Iraq, and was appointed Companion of the Order of the Bath (CB) in 2004. Interviewed online on 04 November 2024.

McKee, Eamonn. He joined the Irish DFA in 1986, beginning the first of three postings in the Anglo-Irish Division (1986-89). He worked as a Traveller in Northern Ireland. He served at the Irish Embassy in Washington (1990–96). He returned to the Anglo-Irish division (1996-2000). Eamonn served as press officer at the New York Consulate (2000–2001). Eamonn once again returned to the Anglo-Irish division in 2001 (2001-04). After leading the Justice and Security Section, he became head of Irish Aid’s Emergency and Recovery Section in 2005, and in 2006 was appointed UN Director and Director of the Conflict Resolution Unit. McKee was part of the talks team for the 1998 Belfast/Good Friday Agreement and later worked on its implementation. Interviewed online on 27 March 2024.

Pilling, Joseph. He served as the director-general of the Prison Service from 1991 to 1993, and then Director of Resources and Services in the Department of Health, before being appointed as Permanent Secretary of the Northern Ireland Office in 1997, where he served until 2005 when he retired. Interviewed online on 18 December 2024.