Hidden barriers and divisive architecture: The role of “everyday space” in conflict and peacebuilding in Belfast

ABSTRACT Whilst “spatiality” and “architecture” have become recognized as important dimensions of urban conflict, contemporary forms of power push our gaze toward symbolic landmarks such as Belfast’s “peace walls.” This paper uses Belfast as a case study to instead highlight the fundamental role occupied by “everyday” urban space and architecture. It reveals evidence of an undisclosed body of divisive architecture put in place through a confidential process of security planning between 1977 and 1985 to physically segregate and spatially fragment Catholic and Protestant communities in contested areas of Belfast. Termed here as hidden barriers, they are formed from “everyday” roads, housing, shops, offices, factories and landscaping, and the ways in which they continue to promote division represents a crucially undervalued aspect of conflict-transformation planning. The paper examines the complex urban challenges that they pose, arguing for a reevaluation of the role of everyday architecture and space in conflict and peacebuilding processes.


Introduction
The conflict in and about Northern Ireland (often referred to as "the Troubles") has profoundly impacted the social, political, and economic structures of the region. Much less recognized is the wider architectural legacy that the conflict has left behind, particularly in urban areas. The signing of the 1998 Belfast Agreement brought a "normalisation of security arrangements and practices . . . consistent with the level of threat" (Northern Ireland Office, 1998, p. 25). This has resulted in the vast majority of militarized architecture commonly associated with the Troubles, such as the wide range of ad hoc army bases and observation posts, being removed from the landscape (for example, Addley, 2007;BBC News, 2000Wylie, 2007). In many ways, the "peace-walls" constructed between Catholic and Protestant communities 1 in many of Northern Ireland's most contentious residential areas, have come to embody the architectural representation of lingering sectarian conflict. This can be seen both in their place as a core facet of Belfast's vibrant tourist industry (Murtagh et al., 2017;Skinner, 2016;Wiedenhoft Murphy, 2010) and in their recognition within Together: Building a United Community, Northern Ireland's flagship conflict-transformation strategy. While the achievement of a policy ambition to remove all peace walls and associated "interface structures" by 2023 (Office of the First Minister and Deputy First Minister, 2013, p. 6) seems unlikely (at the time of writing, 46 of 59 officially registered interfaces remain), it is representative of the "top-down" policy approach often adopted when dealing with the architecture and spaces of conflict. These rationalizations tend to focus on the symbolic landmarks through which both "war-making and peace-making" activities take place (Koopman, 2011;Megoran, 2011). This article seeks to move beyond these conceptualizations by highlighting the value of micro-level interactions between people and their "everyday" built developments. The restrictions on pedestrian and vehicular movement imposed by these designs were germane to allegations of undisclosed military and political influence on what were purported to be objective social renewal programs. This was reinforced by discussions in both the popular press (Alcorn, 1982;Beresford, 1982) and academic literature (Cowan, 1982;G. M. Dawson, 1984), which emphasized the defensive nature of the redeveloped housing settlements, prompting an enduring understanding within many of Belfast's communities that a security-focused rationale informed the proposals T. Cunningham, 2014).
This article has four sections which set out to "tear away the virtual self-evidence" of the seemingly benign architecture constituting the hidden barriers at the center of this research and thus expose the "uncertainty of their structure, coherence and systematicity" (Foucault, 1972, p. 26). It begins by providing an historical overview of the Troubles-era confluence of urban renewal and securityplanning that brought these hidden barriers into being. The article then goes on to outline the research methodology and details three typologies of hidden barriers identified in the research: Inter-Community Barriers, Intra-Community Barriers and Invisible Boundaries. The third section of the article discusses the peacebuilding challenge posed by hidden barriers within a policy context largely driven by disparate and siloed urban policy and conflict-transformation initiatives. The article concludes by arguing that the legacy of divisive architecture in Belfast is more socially, politically, economically and environmentally sophisticated and entrenched than is suggested by both contemporary policy and popular narratives. It reflects on how latter-day power-relations continue to obscure the ongoing impact of these hidden barriers and stresses the need to reevaluate the vital role that the everyday built environment plays in conflict and peacebuilding processes.

Historical context
From the late 19th century through to the 1970s the urban structure of Belfast was characterized by a dense network of Victorian terraced houses which had been built to house the workforce associated with the city's rapid industrial expansion. In common with many industrialized cities across the UK, the condition of Belfast's 19th century housing stock deteriorated rapidly during the opening decades of the 20th century. A series of Housing Acts (1919,1924,1930,1934) led to large scale public housing programs in Britain that tackled the worst of the urban slums. The Northern Ireland government (established in 1921) followed suit initially, but "drastically scaled down its ambitions" within a few years (Fraser, 1996, p. 272). By 1976, government surveys deemed Belfast to have the "worst housing stock in the UK and possibly in Western Europe." 2 At that time only 48% of Belfast's then 123,120 houses were deemed to be "sound." 3 A total of 30,940 houses needed to be "replaced," some 17,400 houses were to be "repaired" and a further 15,360 "improved." 4 This steady deterioration of the built environment played out amidst a protracted period of sectarian violence, population movement and widespread civil unrest which increased ethnic consolidation across the city (Boal, 1969;Hepburn, 1996;Jones, 1960). These processes were then considerably exacerbated by the onset of the Troubles in 1969 (Evelegh, 1978). Confidential government accounts from 1976 estimated that between 1969 and 1976, more than 60,000 people fled their homes in response to sectarian violence and "direct intimidation" or "fear of intimidation" from paramilitary groups. 5 Prolific sectarian confrontation in public housing areas would give rise to the damage of approximately 32,000 dwellings in Belfast and render more than 10,000 habitable dwellings effectively unusable due to persistent threats of paramilitary violence, 6 placing 12,136 people on the "Emergency Housing List." 7 This would, crucially, cause officials within the Department for the Environment (DoE), the ministry with overall responsibility for housing policy, to begin discussing these population migrations not just in terms of the problems that they posed for routine public housing management, but as something that presented new and pressing security ramifications. The bulk of these concerns were directly related to the inequalities in what was ostensibly a disproportionate and sectarian distribution of housing supply. This was most acutely evident where the "large numbers of Catholic refugee families" pouring into "already over-crowded Catholic areas" were perceived to create the threat of these areas "expanding and creating confrontation with the larger adjacent Protestant communities." 8 Despite the prevalence of military and paramilitary activity in residential areas, the NIO made no public acknowledgment of the impact that this was having on redevelopment programs. Records from November 1976 describe how this view was informed by a belief that "for security and policing reasons the balance of advantage lies against provoking widespread discussion on sectarianism in housing" and that any such discussion would be "unlikely to be actively supported by the army and the police who seem to find it easier to control violence in areas where community boundaries are clearly defined." 9 The ultimate decision that it would be better to "get on with the rebuilding of dilapidated areas . . . while trying to weaken sectarian boundary lines by stealth" would establish a duality that distinctively characterizes the abstruse nature of the initiatives that followed. In the first instance, the government would move in February 1977 to publicly announce a Ministerial Steering Group to oversee the comprehensive redevelopment of Belfast's inner-city: Belfast faces acute problems. In common with places like Liverpool and Glasgow the inner city has fallen into decay. In many inner areas people live in deplorable housing conditions. This plus the depressing nature of the environment undermines the quality of life for ordinary people. In addition, the Troubles have affected living conditions throughout the City. The job of the Steering Group will be to mount an attack to improve housing and environmental conditions in Belfast. 10 Whilst these redevelopment processes would directly engage with a wide range of inner-city areas where inter-communal boundaries were contested and where violence was extensive, the passing public reference made to the effects of the Troubles by government officials reflected a longstanding policy orthodoxy which kept housing policy and security policy quite separate in the public eye. To this end, a confidential Standing Committee on the Security Implications of Housing Problems in Belfast was established in June 1977 (Coyles, 2017) to enable a purposefully security-focused management of contentious redevelopment programs quite outside of public view. Operating in parallel to the public redevelopment process, this undisclosed committee brought together officials from the highest ranks of the British Army and Royal Ulster Constabulary (RUC) police force (collectively referred to as the Security Forces and commonly represented at Brigadier and Assistant Chief Constable level respectively), alongside department heads from the NIO and DoE, to assess proposals in areas where there were alleged security issues arising due to the proximity of Catholic and Protestant residents. As the minutes of the first committee meeting held in June 1977 demonstrate, the strategy adopted by the committee from the outset involved the exploitation of routine elements of the wider urban redevelopment process and utilized the "everyday" architecture of the public realm: It is the demolition of exiting buffers or the building of new houses in a previously neutral zone that develops flashpoints . . . Problems could be alleviated by the creation of a neutral zone. In positive terms industrial development or the construction of amenity or welfare buildings would have the added advantage of providing an area for neutral contact. Failing that the creation of a wide open space, a dual carriageway or the erecting of fencing would at lease serve to separate hostile communities. 11 These public and private processes therefore worked hand-in-hand to use hidden barriers to establish a dominant pattern of isolated single-identity inner-city residential communities that continues to define Belfast today. These hidden barriers remain largely unrecognized and the ways in which they promote division represents an important, but crucially undervalued, aspect of conflicttransformation planning.

Methodology
The methodology for this article is driven by a desire to place people and communities at the center of the analysis. It draws on research carried out as part of a wider interdisciplinary research project focused on hidden barriers in Belfast which first used extensive archival research and analysis of previously closed government records to identify a total of six case-study areas affected by their installation and where focused investigations could then take place ( Figure 1). Secondly, these areas underwent examination through a novel process of architectural-photographic analysis to document and illustrate the contemporary situation of the related hidden barriers. The research materials generated from this process were utilized in a series of in-depth interviews with architects and planners involved in the redevelopment of case-study areas and again in follow-up community engagement exercises. Third, an in-depth historical, cultural and socioeconomic profile was completed for each case-study area, collating longitudinal statistical data on demographics, housing, infrastructure, deprivation, and community, public and commercial service provision. Fourth, a review of all relevant contemporary policies and strategies was undertaken. Fifth, a total of five focus groups and 71 interviews were carried out with residents and community workers to gather reflections on hidden barriers in each case-study area. Sixth, a series of community engagement workshops were held across the case-study areas utilizing architectural mapping and photographic documentation. Drawing broadly on the "Q methodology" process (McKeown & Thomas, 2013), these workshops provided an extra layer of understanding and interpretation by presenting some of the research data to residents and community workers. Finally, the research data arising from the six steps outlined above was consolidated and coded using NVivo. It was then analyzed and evaluated by the researchers.

Findings
The evidence suggests three typologies of hidden barriers which act at different scales and in different ways to promote social and physical division. At the first level, there are Inter-Community Barriers. These are instances of the built environment being used on a larger scale to separate two communities that were once connected through the urban fabric before the onset of redevelopment in 1977, effectively leaving them fixed and isolated as single-identity areas. At the second level, Intra-Community Barriers are instances where the redevelopment proposals within these single-identity communities have transformed the existing network of interconnected terraced streets into spatially fragmented and disconnected clusters, creating urban areas that are extremely fractured and difficult to navigate. The third level of hidden barriers are what we call Invisible Boundaries. These barriers are not a direct consequence of the redevelopment programs or input from the Standing Committee on the Security Implications of Housing Problems in Belfast. Rather, they are elements of public space on the periphery of other hidden barrier areas that are identified locally as a recognized boundary between the two communities which has evolved at a local level and become entrenched over time.

Inter-community hidden barriers
Inter-community barriers represent the most direct evidence of interventions made by the Standing Committee on the Security Implications of Housing Problems in Belfast to specifically mitigate against a range of perceived security threats identified by the Security Forces. The minutes of meetings and memos from these cases and others describe how these perceptions commonly related to proposals which had the potential to alter the distribution of housing supply in ways which could expand Catholic territory at the expense of existing Protestant territory. This led to extensive permanent physical barriers being incorporated in ways which delineated fixed boundaries between adjacent Catholic and Protestant areas where the existing open borders were disputed. Whilst these barriers are widespread across the city and vary in type and application, it is possible to provide some examples which illustrate how this goal was achieved through co-opting a range of government departments and processes in order to manipulate the scope of the public redevelopment proposals. None of these examples appear on the list of interfaces currently recognized by the Northern Ireland government (Department of Justice, 2018).
In one such instance at the Lower Oldpark/Cliftonville area in the north of the city, by late 1976 large numbers of existing dwellings had been vacated by Protestant residents. These residents had fled the poor housing conditions in the area to avail of the surplus housing provision available in other recognized Protestant areas outside Belfast. This scenario starkly contrasted with the severe shortage of housing available for the Catholic population, yet the vacated houses in Lower Oldpark/Cliftonville could not be allocated to displaced Catholic residents because they were situated in what was still perceived by the authorities as Protestant territory: As mixed communities sort themselves out and become solidly of one colour or the other, the inter-communal inter-face becomes a front line. Murders occur, people are intimidated, families move out, and houses are vandalised, squatted in or at best bricked up. Paramilitary and other sectarian organisations focus upon such areas. Organised Protestantism or organised Catholicism may demand that a particular area should not "go over to the other side" even though individual Protestants or individual Catholics may no longer in practice be willing to live there with their families. In such a situation, an area-often containing good sound houses-may have to be held sterile for a time at least, until it is politic to let a new line of sectarian demarcation be established. It is the phenomenon of the "creeping inter-face" which gives rise to some of the gravest problems. 12 The committee response to the situation at the Lower Oldpark/Cliftonville interface was to recommend the removal of over 17 acres of disputed and vacated housing and work with the Department of Commerce to rezone the area for industrial use. 13 The consequent redevelopment of the area then brought with it the creation of a large industrial estate and business park which definitively separated and segregated these two areas whilst also decisively prohibiting any further future expansion of Catholic territory (Figure 2).
A similarly deliberate use of hard infrastructure can also be seen at the Twinbrook/Areema area to the southwest of Belfast, one of a range of high-profile suburban developments created by the Northern Ireland Housing Trust in the 1960s (Brett, 1986). The development is particularly distinguished in that it was originally envisaged as a model mixed-religion estate of 3,000 families with an even distribution of Catholic and Protestant residents when it was first constructed during the late 1960s. However, as a consequence of the substantial population migrations that were taking place across Belfast throughout the early 1970s, the population split had shifted to "three Catholics for every one Protestant" by 1973 (Darby & Morris, 1974). By 1977, the estate had become the site of rampant squatting and levels of paramilitary intimidation that forced the Northern Ireland Housing Executive (NIHE), the authority established in 1972 to manage Northern Ireland's public housing stock, to concede that it had "lost management control of the estate." 14 This presented yet a further perceived and unwanted expansion of Catholic territory: . . . the problem was that Twinbrook . . . was becoming increasingly regarded as a probable extension of Belfast's West Belfast Catholic areas. It became increasingly difficult to attract Protestant families . . . Many refugees, all Catholics [began] to squat in Twinbrook occupying homes that had been just built. The Twinbrook Tenants Association moved quickly and set up an anti-rumour service but the problem was clearly [the] desire of Protestants within the estate to leave it because of the influx of Catholic families . . . 15 The specific spatial problem presented by Twinbrook was the prospect that the outward flight of Protestant residents, and the practices of squatting that had been established, would spread to the adjacent and predominantly Protestant Areema estate. This smaller development was separated from Twinbrook by a large greenspace with communal playing fields and a shared link-road which cut across the intermediate zone to provide a vehicular connection between the two areas. In an attempt to address these concerns, the committee asked the Roads Division of the DoE "about the desirability on security grounds" of using a new road to "separate the Twinbrook and Areema estates." 16 The consequent dual-carriageway completed in 1983 now permanently divides these two communities ( Figure 3).
Perhaps some of the most latent forms of intervention made possible by the oversight of the committee are those that were put in place at the Ligoniel/Squire's Hill area on the outer fringes of north Belfast. This is an especially notable case-study in that the housing proposals for the area preceded the inauguration of the committee and were at an advanced stage of development when they fell under committee review. This meant that there were considerable limitations on any direct interventions that could be made in what were detailed plans for a new and bespoke development of Catholic houses. These proposals offer an important historical footnote in that across all the programs that would come under the scrutiny of the committee, those at Ligoniel are quite unique. The area identified for this housing was itself an exceptional commodity, described by the DoE in 1977 as "the last green-field site for Catholics within the building stop-line" which could be used to provide new housing for Catholic residents displaced from other parts of Belfast. 17 As a consequence, not only did the proposals not involve any form of apparent trade-off between Protestant and Catholic territory, they also precluded any intercession in preexisting local conflicts, two factors that were common to the other areas across the city evaluated by the committee. However, the Ligoniel site was located next to the Squire's Hill estate, an area of privately owned and predominately Protestant housing that had been established in the 1960s. An extensive petition of concerns raised by residents from the Squire's Hill area prompted the Security Forces to object to the new scheme on a number of security grounds: The Protestant Squire's Hill estate consisted of private housing and the residents were taking a firm line against the scheme, in particular very strong objections would be made against the second phase which would bring Roman Catholic houses to the edge of the Squire's Hill estate . . . If the Roman Catholic houses were built as far as the edge of Squire's Hill estate a point of conflict would be established. It was likely that the majority of the residents would move and would be replaced by hard-line Protestants (thus creating a confrontation zone) or the estate would become a 'no man's land' with perhaps some Roman Catholic squatters. 18 To provide some manner of redress to these perceived threats, the NIO Minister for Housing, Ray Carter, directly intervened to instruct the DoE to include a series of "planning conditions" in the legal planning permission documentation that had been granted for the Ligoniel development: (i). that there should be particularly heavy planting, to a minimum depth of twelve metres, in the proposed landscape strip between the two developments; (ii). that a physical barrier should be erected before any of the Executive houses immediately adjacent to the private development were occupied; (iii). that the relevant Executive houses should not face towards the private development; and (iv). that there should be no access whatever, vehicular or pedestrian, between the two developments. 19 The "landscape strip" referred to within these conditions is today well-established as a dense and impenetrable zone of mature trees and assorted vegetation which provides a permanent physical division between Ligoniel and Squire's Hill ( Figure 4). Moreover, whilst the advanced procurement of the specifics of the scheme had meant that any direct reworking of its numbers of houses and their arrangement was not possible, the especially troubling assumptions inherent to condition (iii) of the stipulations were achieved, at least in part. This was accomplished by leaving the approved layout of the dwellings untouched and revising the designation of rooms and spaces within the curtilage of the applicable houses. As a consequence, "living-rooms" and "front-gardens" would instead be fitted-out respectively as "kitchens" and "drive-ways," thus reducing the opportunity for these occupants to "face towards the private development" (Figure 5).

Intra-community hidden barriers
These barriers arise from the intricate mosaic of residential designs that characterized the redevelopment of Catholic and Protestant areas between 1977 and 1985. Quite distinct from the workings of the Standing Committee on the Security Implications of Housing Problems in Belfast, responsibility for these initiatives fell to the NIHE. In a typical scenario, a single-identity community such as Lower Oldpark or Cliftonville would be divided into a series of Redevelopment Areas which would be assigned to different teams of NIHE architects and planners. The new proposals put in place drew heavily from the residential typologies that were in vogue at that time, notably the Dutch woonerf (Ben-Joseph, 1995;Kraay, 1986), the cul-de-sac (Essex Planning Officers Association, 1973) and defensible space (Newman, 1972(Newman, , 1976. These typologies are now widely established and are diverse in their conceit and application. However, they also share a common ambition that was something of a rebuke to the mass-produced and anonymous nature of both the historic and more modern-era forms of public housing that had gone before. This shift centered on the new idea of now empowering local residents with a sense of ownership over the communal public spaces which served their dwellings. Over the course of the redevelopment period in Belfast, the introduction of these design tropes typically resulted in open-ended networks of Victorian-era streets being replaced by small clusters of courtyard housing and dead-end streets where vehicular connectivity was eliminated and pedestrian permeability significantly reduced.
These revisions to the existing urban form had particular significance in Belfast. On the one hand, the spatial strategies put in place to imply notions of ownership over space can be seen to align with their uncontroversial and widespread application in routine residential developments across the United States and Western Europe. However, on the other hand, they also imparted new spatial criteria that can be seen to directly address pertinent security issues explicitly evident during the Troubles. This is borne out in the findings from research interviews carried out with a range of planners and architects involved in these programs. A number of excerpts are quoted here at length in order to illustrate this ambiguous synonymity between security-orientated design and more conventionally aspirational design ideals: We tried to make a clear divide between pedestrians and cars, trying to make areas much more secure and controlled for the residents, to try and give people privacy and security. 20 We tried to break the areas down and make them much more private for those that lived there. We built in quite crude ways, building walls maybe here and here to stop through-traffic and through-movement and give the folks that lived here more ownership. That certainly was the theme . . . The design thinking around these redevelopment areas was how do you move back to something that gives local people control, and I suppose, keeps those that aren't from the area out. 21 There was an expressed desire to reduce the number of through-routes . . . so that there was safety for kids to go out on the streets . . . they wanted us to create residential areas that were quiet. It wasn't totally their views but it has had the unfortunate effect of leaving all the street-ends blocked off . . . The houses were deliberately difficult to find. You need to know it's there and I think that was the plan, that it is only the residents who would know the route. 22 The Northern Ireland Office did have a hand in the security aspects of the redevelopment. You would have been shown plans and they would have been looked at and the opinion would have come back stating "we'd prefer if you didn't do this." You would have found back entries to terraced houses with gates on them and it was simply because people were using them for means of terrorist escape. The government didn't only use the terrorist thing. There was also the fact that the streets were a short cut for people to get from one road to the other in busy traffic times. You didn't want that either. 23 The quotations reflect broader interview findings which appear to suggest three main ways in which the myriad range of design responses can be seen as a complex set of intra-community hidden barriers. In the first instance, there is the elimination of vehicular through-flow. To some designers, this has been enacted to address a straightforward public safety problem; that of "rat-runs," the local vernacular for how the lattices of older residential streets provided a network of vehicular short-cuts between nearby arterial routes. For others, removing the opportunity for vehicles to pass through the area is more directly allied to reducing the potential for facilitating terrorist escape. The vast majority of inner-city paramilitary attacks were carried out using cars. The travel to an area to carry out an attack, and the escape from an area or within an area after that attack, were greatly facilitated by the preexisting Victorian-era gridiron layout. Either way, both design responses entail an indistinguishable and consistent rudimentary closure of existing and formerly open and connected residential streets (Figures 6 and 7).
A second and closely related theme, but one that is less direct, can be seen in the desire to create residential areas that are "quiet," "private" and "secure." This is closely linked to the eradication of vehicular through-flow and the associated reduction in pedestrian passage which then establishes local residents as the dominant and near-exclusive users. This also encourages the passive supervision of these communal spaces by residents which in turn helps to reinforce the sought-after sense of ownership. This might be quite innocuous in a context not impacted by sectarian strife. In the context of Troubles-era Belfast, it is difficult to untangle notions of "privacy" and "security," and their day-today meaning in lived experience, from the gamut of paramilitary activity and cross-community  violence then taking place. This is reflected in how latter-day expressions of ownership often continue to be articulated in distinctly partisan territorial terms ( Figure 8).
A third and final theme can be discerned in the deliberate attempts to circumvent easy wayfinding and establish residential areas that are completely counterintuitive to normal everyday access. Design aspirations to "keep those that aren't from the area out" and to have "only the residents" knowing the route, might hinder entry from paramilitary agents. But these intentions are only made possible through the methodical isolation of these new houses from their once connected neighbors. This is an innately divisive spatial tactic that considerably limits opportunities for routine local community engagement and social integration with the wider general public.
The ultimate consequence of this series of thematic overlaps is that, irrespective of the motivation driving isolated individual redevelopment programs, the subsequently revised inner-city persists as a realm where community fragmentation and spatial disconnection continue to be promoted by these everyday residential spaces (Figures 9 and 10).  . A supervised pedestrian passage through a 1980s west Belfast redevelopment. Here the "woonerf approach" that sought to break down large continuous streets into a series of smaller spaces has been somewhat reimagined to create a disjointed series of spaces that obscure the route through. Photograph: Donovan Wylie.

Invisible boundaries
These barriers emerge out of a multi-layered process combining the impacts of contiguous physical barriers and the psychological consequences of historical and contemporary lived experience. As such, they exist in the movements and behaviors of residents and can be understood in the context of "activity segregation" (Boal, 1969(Boal, , 1971(Boal, , 1987. According to Boal's understanding of activity segregation in the very different context of pre-1969 Belfast, residents living in adjoining ethnically defined neighborhoods experienced minimal interaction with each other. For example, this meant using only the bus routes that traversed the neighborhood they most identified with. Activity segregation only intensified as the events of the Troubles resulted in greater fear and mistrust of "the other" in Belfast, and this has continued to a large extent in the post-Troubles period. Building on these ideas and further calls for related research (Doherty & Poole, 1997;Franzén, 2009;Grimes, 1993;Huck et al., 2019;McCafferty, 2001;Robinson, 1969;Schnell & Benjamini, 2005), we consider here the impacts of the inter-community and intra-community barriers discussed above in activity segregation in and around a number of the case-study areas.
The extent to which movement in contemporary Belfast is impacted by memory, trauma, and perceptions of "the other" has been considered by Shirlow and Murtagh (2006). According to this analysis, knowledge of past violent events has permeated through generations to create fear induced movement patterns that see people use "mental maps" to move through the city, influenced by avoidance of "spaces of fear" and gravitation toward "sanctuary spaces." G. Dawson (2016) draws on the work of Tumarkin (2005) and De Jong and Rowlands (2008) to consider the interplay of "traumascapes" and "memoryscapes" respectively with segregation, movement, and spatial legacies of violence in Belfast. The truamascape concept suggests, as Dawson notes, that "an analysis of temporal disturbance, associated with the psychic and physical experience of '[t]raumatised people [who] have to live the past that refuses to go away' is transferred to particular geographical spaces 'where events are experienced and re-experienced across time'" (Tumarkin as cited in G. Dawson, 2016, p. 136). Similarly, memoryscapes as articulated by Rowlands and De Jong refer to "spaces and sites . . . that are inscribed with 'old' meanings and memories, but may be 'reworked' or reinscribed, becoming 'palimpsest memoryscapes that are both real and imaginary'" (Rowlands and De Jong as cited in G. Dawson, 2016, pp. 136-137). We contend that the invisible boundary examples identified in some of the case-study areas can be partly understood in the terms outlined above. However, we also note that such long-established internalized practices in these areas have been extended, or indeed refined, as a result of ongoing interaction with inter-community and intra-community barriers. Invisible boundaries tend to exist in public spaces situated between two ethnically defined neighborhoods, such as parks, commercial centers and transport infrastructure. Critically, they condition how people move around their neighborhoods despite the absence of immediate physical barriers, and reinforce the idea that freedom of movement and access to goods and services in the affected areas are restricted. For example, in the largely Catholic Ligoniel case-study area in north Belfast, some residents expressed unwillingness to use a public park adjoining the area and neighboring majority Protestant Ballysillan. 24 One interviewee made the point that despite the park's proximity to her family home, none of her 10 grandchildren had ever used it. 25 An innocuous traffic island (Figure 11) by the entrance to the park demarcates the boundary between these Catholic and Protestant areas. Community workers and residents in the locality point to the significance of this traffic island as a boundary marker, despite the absence of any identifying communal features. The specific space has a long history of territorial demarcation. Ligoniel was once an industrial village, peppered with linen mills staffed by Catholic and Protestant workers. One such mill, now demolished and replaced with housing, exited onto the main Ligoniel Road in the vicinity of where the traffic island now sits. A community worker in the area recalled that the workers tended to exit the mill at quitting time and proceed north or south along the road depending on their community backgrounds. 26 Ligoniel residents point squarely to this traffic island and the space around it as an invisible boundary that can be understood in the terms discussed above: It's not a wall, it's a silent wall. 27 What would you call a wall that's there, but not there? . . . It used to be that you couldn't walk down to the turn [beyond the traffic island.] It was just a no-no. You can do it now. I'm going to the wee butchers now up at the turn of the road. I'm up here 30 years, and I only started going there just before Christmas. That's something I wouldn't have done before. 28 Despite an emerging willingness by some residents to venture beyond the traffic island on foot, there exists an anecdotal history of threat and intimidation experienced by Ligoniel residents who decided to use the shops and services beyond it. This serves to reinforce the impact of this invisible boundary to the extent that some people refuse to use the more frequent bus services that run in the area beyond. Some residents also referred to the impact of civil disorder arising from the 1995 "Drumcree standoff," a particularly contentious period of competing protests across Northern Ireland surrounding ceremonial marches by the Protestant "Orange Order" through the majority Catholic Drumcree area of Portadown, some thirty miles south of Belfast (O 'Neill, 2000). During this time, unofficial roadblocks were constructed in the Ligoniel area and communications infrastructure was destroyed in arson Figure 11. A traffic junction at Ligoniel which represents the local boundary between what is acknowledged as Catholic territory (above) and Protestant territory (below). attacks. The cumulative effect was that Ligoniel was essentially cut off from both the city and the rural area beyond, leaving the community feeling isolated and threatened. 29 Further historical research postulates that the roots of this invisible boundary perhaps stretch back to as early as the 1920s, when temporary peace wall style structures were erected throughout Belfast as a means of community protection during a 2-year period of intense sectarian conflict in which almost 500 people were killed (N. Cunningham, 2013;Parkinson, 2004). In that earlier period, a large gate could be pulled across a narrow point in the road to protect the area from attack. 30 In the predominantly Protestant Lower Oldpark case-study area, again in north Belfast but much closer to the city center, some residents noted the impact of their experiences as children: " . . . constantly my brother was jumped on, nearly every day . . . They used to wait on us coming home from school. I was quite frightened at the time . . . " 31 While another resident spoke about how the fear of attack or intimidation, although unlikely to occur in the majority of occasions, still precluded them from venturing outside of their own area: You could go up there. Don't get me wrong, there's plenty of good people, I know . . . You could go up there nine times out of ten, nobody would say boo. You could go up there some day, and some wee cheeky teenager could come over: "Hey, aren't you a Proddy from . . . " That's you dead. You'll never get back home again. 32 At the predominantly Protestant Suffolk case-study area in southwest Belfast, an enclave enclosed by a near continuous belt of Catholic housing, an assortment of prominent physical barriers such as buildings, road, fences and landscaping, segregate Protestant houses from their surroundings. Many Suffolk residents refuse to use the local bus service as it is seen to be a Catholic route not safe for a Suffolk resident to travel on. 33 Such choices and behaviors are reminiscent of how residents in Ligoniel consider bus routes or bus stops, and indeed as Boal (1969) has shown, how people in the Falls and Shankill areas organized their movements even prior to the 1969 outbreak of hostilities.
It must be stated, however, that such attitudes and behaviors were not general across all those engaged with as part of the research. For instance, at a series of group interviews in Suffolk, one participant noted the extremity of how her movements are restricted in a psychological manner: I would be too afraid to go to that garage [shop] now. After it gets to teatime. I live the same way I lived during the Troubles. Once it gets to teatime, that's it. That's me in the house . . . it used to be . . . if you'd no cigarettes, or you'd no milk, or none of this or none of that . . . that's it, you done without until the next day. And really, that's the way I live now. 34 Yet, other residents made clear that they and others frequently use a local taxi service due to its convenience and despite its strong Catholic "republican" identity. Another participant summed up the strength of internal and external perception in terms of how Suffolk residents move around and use services in the wider area. In this particular case, he needed to use an unnamed public service and went to the closest available outlet, which happened to be in an area with a similarly strong republican identity. He stated that on seeing his home address, the service provider asked if he was comfortable with the location and made clear that they would have expected him to use one of the alternative outlets in Protestant areas, despite the longer distance and greater difficulty in getting there. This was not communicated in a hostile fashion, rather the experience of the service provider was that Suffolk residents tended not to use the most convenient location and would instead travel the longer distances in order to feel safe. 35 The ways in which people move through the city and within and beyond their neighborhoods are highly complex, especially the areas we have identified as displaying hidden barriers across the three typologies. Individuals move in different ways, with some continuing the movement patterns they observed during the conflict, others moving freely but with some self-consciousness in that activity, and a spectrum of differences between these. It is impossible to isolate an exact motivation for these movement patterns, but it is fair to say that the changes made to the built environments of the casestudy areas as outlined in this article, contribute to the construction of mental maps in a similar way to symbolic markers (flags, murals, memorials) or memories of violence and trauma.

Discussion
Our research has uncovered a range of hidden barriers in Belfast. As our archival research has shown, these barriers have a long and deep history and an insidious impact on daily life and movement across the city. However, what is most striking is that hidden barriers as we have defined them are a peripheral, or non-existent, concern at a policy and planning level.
Some researchers have noted that "communities are also kept apart in less obvious ways, where motorways, shopping centers, dense foliage and/or vacant and derelict landscapes have been used to define the perimeters of particular communities" (Byrne et al., 2012). But major policy approaches, and academic research, has not sought to address these divisions or to categorize them in any systematic way. For example, in terms of addressing inter-community divisions, Together: Building a United Community (OFMDFM, 2013), as noted earlier, is the leading relevant government strategy. Naturally, the bulk of the strategy deals with issues such as relationships and attitudinal change at a community level, but it acknowledges that physical segregation is related to such issues, and "shared space" is a major theme of the document. Minimal recognition of what we term hidden barriers is made. Although it notes "segregation in housing and our education system" and "physical divisions and invisible lines of separation that exist in both urban and rural settings," these "invisible lines" are not defined (OFMDFM, 2013, p. 34). In relation to rural areas, it is noted "rural towns and villages are divided not by walls or fences-but by an invisible line in the road or a local landmark" (OFMDFM, 2013, p. 55). This would broadly fit our description of an invisible boundary, but again little is said about how to address such issues and their relevance in urban areas. Where physical division is tackled, Together: Building a United Community, like much of the discussion in Belfast around division as we noted earlier, focuses on the well-known goal of removing interfaces and peace walls entirely by 2023.
Numerous government strategies replicate this approach, where hidden barriers and invisible boundaries as we define them are alluded to but ultimately ignored in favor of a focus on a designated register of such interfaces. For example, Building Safer, Shared and Confident Communities (Department of Justice, 2012, p. 25) notes the presence of geographic barriers, such as rivers and major roads, that can reinforce a sense of separation, but specifically indicates that it is only responsible for 59 officially recognized interfaces formerly under the control of the Northern Ireland Office. Likewise, the Strategic Planning Policy Statement for Northern Ireland: Planning for Sustainable Development (Department of the Environment, 2015) recognizes the "insular layouts and poor connectivity" in certain communities and calls for more "balanced" housing estates, but ultimately concludes with a discussion of peace-lines, contested spaces and removing large-scale barriers. In a similar vein, when an issue such as connectivity across the city is discussed, a macro approach is favored over a micro analysis of inter and intra community connection as we have outlined. The Living Places: An Urban Stewardship and Design Guide for Northern Ireland notes, in line with some of our findings, the divisive role of cul-de-sacs and how this imposes "long journeys on people needing to travel short distances, leading to increased usage of the car" (Department for Infrastructure, 2014, p. 36). The Community Cohesion Strategy 2015-2020, recognizes the need to improve connections within individual residential areas and promote "interaction between segregated, single identity estates" (Northern Ireland Housing Executive, 2015, p. 34). Both documents however, provide little specific detail on how to tackle these issues. In contrast, the Urban Regeneration and Community Development Policy Framework focuses purely on the macro and seeks to "improve linkages between areas of need and areas of opportunity" noting that it is necessary to improve "the linkages, physical and otherwise, between these areas and urban assets and employment opportunities" (Department for Social Development, 2013, p. 16). Similarly, the Regional Development Strategy 2035: Building a better future, makes gestures at the larger scale, aiming to "improve connectivity to enhance the movement of people, goods, energy and information between places" (Department for Regional Development, 2010, p. 21).
In short, our analysis of the policy documents reveals that the over-arching approach (despite some mention of micro-level connectivity issues due to hidden barriers) tends to focus on connectivity and movement in and out of interface areas. New policies are mainly concerned with facilitating access from the local community into the city or commercial areas of high economic activity. Of course, this attention to issues of access to the city and commercial areas and the associated macro lack of connectivity is essential and necessary. But it is our contention that this is often at the expense of intercommunity and intra-community connectivity that has significant everyday impacts. These manifest in such routine experiences as having to drive around a neighboring housing estate to get to the shops, or lack of access to nearby parks, playing fields and other local amenities. Similarly, tackling large-scale divisions such as peace walls is vital to addressing some of the urban legacies of the conflict. But, given the widespread and deceptive nature of the hidden barriers we have identified, such an approach may have limited impact over time in terms of the actual lived experience of those surrounded by hidden barriers. Broadly speaking therefore, the ways in which division plays itself out at the local level, and is consistently reinforced by hidden and invisible barriers, largely falls below the policy radar.
The less visible barriers that promote division (say a shopping center, a road or an industrial park) clearly cut across a range of policy areas and responsibilities. As a consequence, they are not clearly defined as conflict-related divisions or marked as the designated responsibility of any one government department. This sits in stark contrast to the register of "interface structures" designated to the Department of Justice and now targeted through its Interface programme. 36 Catch-all policy statements such as promoting "shared space" or "connectivity" substitute for a specific set of actions geared toward the particular types of hidden barriers we have identified. The deeply ingrained and embedded physical infrastructure of inter-community and intra-community division at the everyday level evidenced in our research (in addition to the inter-community tensions around housing in the past and present) does not sit comfortably with the macro approach to policy in Northern Ireland which fails to fully acknowledge or tackle head-on the lived realities of hidden barriers. Strategies for addressing conflict are written into many policy documents, and peacebuilding work is supported particularly in terms of local community groups funded to connect and work with "the other" community. However, taken in the round, this does not challenge the reason for segregation in the first place. Nor does it seek to fundamentally transform architectural division and social space at a local level. In other words, the reluctance to openly challenge the sectarian segregation that dominates inner-city social-housing, as was plainly evidenced in the 1970s and 1980s, still remains. This is not to say there have been no changes since the 1970s. There is little doubt that in the interceding 2 decades since the signing of the Good Friday Agreement, the reduction of political violence has been dramatic and there is evidence of gradually improving inter-community perceptions and attitudes (Byrne et al., 2015(Byrne et al., , 2012Kelly et al., 2018;Nolan, 2012Nolan, , 2014Wilson, 2016). Yet the enduring division fundamentally reinforced by the hidden barriers illustrated within this article presents a legacy of conflict that remains broadly unchallenged by current policy. The ways in which the related vehicular infrastructure and housing layout enforce segregation and deliberately eliminate opportunities for integration echoes the sectarian spatial planning policies of the apartheid-era South African government (Berrisford, 2011;Hart, 1988;Nagle & Clancy, 2010) and longstanding Israeli government policy (Yiftachel, 1998(Yiftachel, , 2010(Yiftachel, , 2015Yiftachel & Yacobi, 2003). However, at the level of policy implementation there is an inclination to view the peacebuilding utility of architecture from a predominantly instrumental perspective. Tourist-orientated signature projects in Belfast such as the Crumlin Road Gaol adjacent to the Manor Street peace wall in north Belfast (Muir, 2014), the Titanic Quarter and visitor attraction center in east Belfast (Coyles, 2013;Neill et al., 2013), as well as international examples such as the reconstruction of the historic Stari Most in Mostar after its destruction during the Bosnian war (Forde, 2016), or shared schooling on the Cypriot "Green Line" (Shaw, 2014), use emblematic associations to call attention to peacebuilding progress in a manner that resonates at an international level. However, this article demonstrates that there is also a danger where the spectacle of these architectures encourages a reductionist understanding of what are, in fact, highly specialized conflicts with complex and localized spatial dimensions. The architectural interventions of Israeli military policy implemented to control the daily rituals and behaviors of the Palestinian population (Segal et al., 2003;Weizman, 2010Weizman, , 2012) are a particularly well substantiated example of how the everyday urban fabric can be utilized as a "technology of violence" against a civilian population (Weizman, 2010). These forms of control represent a means of exercising power which effects a "politics of life" (Anderson, 2012) over daily rituals and behaviors that can be conceived of as reproducing and duplicating the conditions of conflict. As in the Israeli-Palestinian conflict, overarching local understandings of Catholic and Protestant territory are central to ongoing ethnic tensions in Northern Ireland. The hidden barriers and invisible boundaries highlighted in this research have the subtle capacity to latently reinforce conflict-era attitudes and behaviors in ways which continue to shape and strengthen this informal knowledge of territory. This is compounded by how contemporary forms of government power serve to push our gaze toward peace walls, interface spaces and large-scale infrastructure projects, whilst at the same time drawing attention away from more nuanced considerations of other forms of divisive architecture. If we hold that "power and resistance are together the governance machine of society" (Hunt & Wickham, 1994, p. 83), the conflict-era attitudes that are reinforced at the micro-level by these hidden barriers (or outsourced to community workers and civil society to address) provide evidence of the limits of governmental power and the limited effectiveness of current policy to make transformative changes to social and residential space.

Conclusion
Communities in Belfast are no longer forced apart by top-down planners, security officials and architects using roads, parks, shops and tree lines, as in in the 1970s and 1980s. Instead, we contend, the divisions of the past are now maintained and managed by communities, politicians and policymakers who shy away from addressing micro intra-community and inter-community division. This is achieved by ignoring (or failing to see) the full historical legacy and persistence of hidden barriers and invisible boundaries on the one hand, and by the implementation of wide-ranging policies often geared toward the most visible forms of separation such as peace walls. Of course, the removal of "visible" barriers such as peace walls would be welcomed, but our research suggests that the everyday reality of a multitude of other deeply entrenched local hidden barriers would mitigate against the real impact of removing peace walls, a fact largely ignored at a policy and political level.
In our view, this is in part due to the fact that these hidden barriers have not been cataloged in the past, but also because these hidden barriers and invisible boundaries sit "right at the point where relations of power are exercised" (Foucault, 1980, p. 142) in the contemporary context. Arguably, "tackling" division in Belfast through generic policies about "mixing," "shared space," "connectivity" and the removal of iconic walls of separation, whilst promoting some form of change, are also powerful instruments that maintain the status quo, keep communities divided, and leave local powerrelations unchecked. In this sense, specific communities have become both subjects and objects as they are tied into the spectacle of an emblematic peace (a Belfast without peace walls). All the while, this obscures the fact that most individuals continue to live in their historically ethically carved up space, wherein many of its hidden barriers are now normalized and almost rendered imperceptible.
This effects a problematic condition where these policy gaps become "devolved" and diverted downwards, largely to local community groups and civil society charged with improving local community relations but not essentially disturbing the spatial and architectural reality of their communities. Communities themselves are in the "position of simultaneously undergoing and exercising power; they are not only its inert or consenting targets; they are always also the 'elements of its articulation'" (Foucault, 1980, pp. 78-108). An investigation into hidden barriers therefore helps to illuminate a more complete mapping of how contemporary power dynamics operate in the post-Troubles city. This suggests that this realm of architecture and space can work against, and at the same time in tandem with, "top-down" policy as well as promote change and/or maintain the status quo.
Architecture produces "positive effects when the liberating intentions of the architect coincide with the real practice of people in exercising their freedom" (Rabinow & Foucault, 1984, p. 246). It can also become a fundamental arbiter in the exercising of this freedom when those intentions are predicated instead on using the urban fabric to promote security and maintain political power blocs. Considered alongside a peace process that is now more than 2-decades old, the architecture and spaces of the hidden barriers exposed in this article illustrate an afterlife in the unfolding peace that remains distinctly undervalued. This firmly establishes these hidden barriers as a discrete peacebuilding problem and suggests the need for a more sophisticated evaluation of what constitutes the architectural legacy of conflict. Notes 1. We have chosen to the use the terms Protestant and Catholic to describe the case-study areas. The historical documents analyzed for the research in this article use these broad descriptive terms and as such we have used these descriptors. Urban and housing division is also still broadly referred to in these terms (see Kelly, G., Gray, A., Hamilton, J., Melaugh, M., Lynn, B. & Robinson, G. (2018). Northern Ireland Peace Monitoring Report: Number Five. Belfast: Community Relations Council.) More recently terms such as Catholic/Nationalist/ Republican (CNR) and Protestant/Unionist/Loyalist (PUL) are being used in policy circles to better capture the political diversity within these communities. There are also increasingly individuals who might not ascribe to any of these identities. By using the descriptors Protestant and Catholic in this article we do not wish to convey a lack of diversity or to narrowly homogenize communities, but rather the terms reflect a historical delineation that is locally understood and typifies how certain urban areas are still perceived despite differing levels of internal diversity. 2. CJ (NIO) 4/1559: Press release on the setting up of the Ministerial Steering Group on Belfast Housing, 8 February 1977. 26. O'Reilly, C. (2018. Personal interview, Ligoniel case-study area. 27. LG02 (2018)

Disclosure of potential conflicts of interest
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by the Arts and Humanities Research Council [AH/M001342/1].