Mike Burke ✍Surveying the popular and scholarly literature reveals at least four key groups that currently make up the middle ground in the north of Ireland: (1) Alliance Party supporters, (2) people who say they are “neither” unionist nor nationalist, (3) those who identify as “Northern Irish”, and (4) those who have no religion (Breen, 2024b; Tonge, 2020). 

 They are “middle” in contradistinction to the two communities of Unionist/British/Protestant and Nationalist/Irish/Catholic. The four middle-ground groups are also collectively referred to as “others”. I’m primarily interested in the life of these groups in their guise as components of the middle ground. I examine the conceptualization, size, social composition, and constitutional preferences of these four middle-ground groups.

Recent events highlighted the centre of the north’s political spectrum. Alliance’s strong performance in the three 2019 elections—local, European and Westminster—focused attention on the party and its energized support base in the other middle-ground groups. Alliance reinforced its success in the 2022 Assembly election to become the third largest party in the north, surpassing the seat and vote totals of the UUP and SDLP (Tonge, 2020, 2022a & 2022b; Hayward, 2020; Whitten, 2020a & 2020b).

The consolidation of the middle ground relates to larger social and political processes in the north. The current surge in Alliance’s electoral success may herald a transformation of the northern party system in which three sizable blocs compete for votes from an electorate with a non-unionist majority (Ó Dochartaigh, 2021 & 2022; Raymond, 2019; Murphy, 2023). The rise of “neithers” relative to the traditional categories of unionists and nationalists is the result of a broader dealignment in communal-political identification. The steady popularity of a Northern Irish identity may signal a weakening of the historically strong affiliation of religion and politics that defined Protestant-and-British in opposition to Catholic-and-Irish. And the social weight of the no religion group emerges from wider developments in the deconfessionalization and secularization of northern society (Coakley, 2021; Hayes & McAllister, 1995).

The middle ground is typically and centrally defined by its presumptive location on the constitutional question. Just as the Unionist/British/Protestant grouping supports maintaining the union and its Nationalist/Irish/Catholic counterpart prefers reunifying Ireland, the middle ground has a distinctive view on the constitution. What academic Mary C. Murphy says in her study of the Alliance Party is true generally of all middle-ground groups:

... the middle ground represents a space which is neither unionist nor nationalist. It applies to parties and voters who are designated as ‘other’ as set out in the 1998 Belfast/Good Friday Agreement. In this way, the middle ground is not an indication of where a party lies on the left–right political spectrum; instead the label ‘middle ground’ or ‘centre party’ implies—first and foremost—neutrality on the constitutional question (Murphy, 2023, p. 95).

In conventional usage, then, the term “middle” applies to those groups that are intermediate or neutral or agnostic on the union-versus-unity divide. The middle ground stands halfway between the constitutional endpoints of remaining in the United Kingdom and reunifying Ireland.

I show that the conventional definition of the middle ground is highly inaccurate. Middle-ground groups do not in fact occupy a middle or intermediate position. Nor are they especially neutral or agnostic on the constitutional question, as Murphy and others contend. The conventional definition of the middle ground can also be exceptionally misleading. I argue that some conceptualizations of the middle ground are not constitutionally innocent but consciously partisan, enlisting the middle firmly in the service of maintaining the union. They in effect drag the middle to the pro-union end of the constitutional continuum, all the while claiming that this altered position somehow still represents the middle.

If constitutional politics is inherently bound up with the definition of the middle ground, it also accounts for the increasing public interest in middle-ground groups. As discussions of Irish unity and a border poll proliferate in post-Brexit Ireland, attention naturally turns to the constitutional decision that may confront northern (and southern) voters. If most unionists will vote for union and most but perhaps not as many nationalists for unity, what will the middle do? Many commentators assert that the middle-ground groups identified above—Alliance supporters, neithers, the N Irish and those with no religion—will play crucial roles in determining the outcome of any reunification referendum (Tonge and Evans, 2020; Murphy & Evershed, 2022; Breen, 2023a; Nolan, 2022; Whitten, 2023; Hanna, 2024). I demonstrate that this assertion is a reasonable use of the term “middle”, and examine the changing constitutional preferences of the middle-ground groups. I find that, in recent years, the constitutional opinions of the four groups have decidedly shifted, with support for union plummeting and support for unity burgeoning. Union still has the edge over unity, but its advantage is rapidly diminishing.

Other, less explicitly constitutional reasons also explain the rekindled interest in the middle ground. The congealing of the middle, especially Alliance’s recent successes in the context of Stormont dysfunction, helps to propel increasingly urgent voices in favour of reforming the north’s devolved institutions away from a “two-communities” perspective. In this paper, I’m concerned solely with the constitutional aspects of the middle ground and do not address the debate about institutional reform.

A Lopsided Middle

This section has two related aims: to show, first, that a middle ground actually exists but none of the middle-ground groups resides there and, second, that the place where middle-ground groups actually reside can in no way be considered a middle ground. These unhappy circumstances signify that Middle Earth is equal parts paradox and legend. It nevertheless remains worthwhile to probe the reasoning of those storytellers who continue to pretend that a densely populated Middle Earth does exist. And it’s useful to explore the social makeup and constitutional views of the peoples reputed to dwell there. For the inhabitants themselves are real, they are just not part of the actual middle ground.

The first aim is to identify the middle. Logically, the middle ground of the union/unity divide is some kind of joint authority or shared sovereignty in which Britain and Ireland act as co-sovereigns, each with binding powers and responsibilities over the north (O’Leary, 2022). This constitutional arrangement is a compromise position that stands between the poles of “United Kingdom” and “United Ireland.” Historically, various models of co-sovereignty enjoyed some fleeting governmental, party-political and academic interest but never reached the level of a popularly accepted option to the constitutional divide.[1] Currently, some political actors propose joint authority, sometimes called Plan B, as an interim alternative to direct rule from London during periods when devolved government is collapsed. In the governance breakdown just resolved, they used joint authority as an inducement (read threat) to convince the DUP to reenter Stormont. And the kind of joint authority they offered—giving the Irish government an enhanced consultative role within a system of unbridled British rule—fell well short of robust schemes for co-equal sovereignty. The most recent evidence suggests that an exasperated public might consider some undefined notion of joint authority in place of Stormont’s seemingly structural impairment (Breen, 2023b; LucidTalk, 2023).[2] This stance is not an active constitutional preference but seems very much a fallback position in the face of continual failures of devolution. In short, no one—least of all any of our four groups—actively advocates joint authority as a fair and reasonable compromise that could bring about a long-term resolution of constitutional divisions in the north.

Even though middle-ground groups are not located at the middle of the constitutional spectrum, I’ll continue to use the term “middle ground” because it’s a popularly accepted way to refer to the four groups I’m analyzing. For the moment, the reader should assume that my every use of middle ground is enclosed in ironic quotations, even though the quotations are not there. Let’s pretend they are. In the last substantive section on constitutional preference, I drop my ironic use of middle ground for reasons I explain there.

The second aim of this section is to identify where middle-ground groups actually reside, if not in the middle. I use “Northern Irish” identity as a case study to show the problematic conceptualization of the middle ground. I also challenge the Alliance Party’s contention—and the widespread acceptance of its claim—that it is part of the middle ground because it’s constitutionally neutral. We need to ask some serious questions about how the notion of the middle ground is used in analyzing the politics of the north.

The Northern Irish identity has been tracked in opinion surveys since the 1980s, but it caught the attention of the public only with the release, in December 2012, of the answers to the new 2011 census question on national identity. The new question asked “How would you describe your national identity?” and directed individuals to tick all the options that apply: “British, Irish, Northern Irish, English, Scottish, Welsh, Other” (NISRA, 2011, p. 5). Forty percent of northern residents said they were British only, 25 percent Irish only and 21 percent Northern Irish only (NISRA, 2012).

Media and scholarly responses to the appearance of the N Irish were, at first glance, starkly different. The media presentation was sensationalist; the academic response seemingly more measured. Reporting on the census results, the front-page headline of the Belfast Telegraph dramatically announced “The rise of the Northern Irish” (Belfast Telegraph, 2012a). The paper immediately drew a straight line between the N Irish identity and the maintenance of the constitutional status quo: many Catholics now regard themselves as “Northern Irish”; this new identity means that Catholics feel comfortable in the northern state; and this level of comfort means in turn that a united Ireland is inconceivable (Belfast Telegraph, 2012b & 2012c; Clarke, 2012).[3]

The usual unionist suspects, through the medium of the Belfast Telegraph, jumped in to pronounce on the implications of the new N Irish identity. Ruth Dudley Edwards triumphally remarked that “it doesn't matter how much republicans shout about a united Ireland and call for a border poll, the constitutional debate is over for generations” (Dudley Edwards, 2012). DUP leader Peter Robinson proudly proclaimed that nationalism is in crisis and its goal of a united Ireland “is further away than ever” (McAdam, 2012).

Temperate academic voices wondered if the N Irish label could prove to be a superordinate or inclusive “national” identity capable of effacing the old divisions separating the two communities. Could it be a middle-ground and shared identity that both traditions equally claim? “Northern Irish”, they reasoned, is an ambiguous identity capable of appealing, in different ways, to groups on either side of the constitutional divide. Catholics might see “Northern Irish” primarily in all-Ireland or territorial terms, as recognition that they inhabit the north eastern part of the island of Ireland. Protestants, on the other hand, might be attracted to the political meaning of “Northern Irish” as conferring constitutional legitimacy on the northern state. The speculation that N Irish could be an inclusive and shared identity crashed into some hard empirical realities. Survey data revealed that the number of Catholics identifying as “Northern Irish” fell away, and the label came increasingly to be claimed by disillusioned Protestants.[4] As I’ll show below, those identifying as N Irish are now the most heavily Protestant and unionist of the four middle-ground groups.

Survey results are not the only challenge to scholars who imagine N Irish as a possible shared identity located at the centre of the north’s political divide. Some conceptualizations of the Northern Irish identity take a different view; they seem to negate its potential as a meaningful cross-community allegiance. Academics Jonathan Tonge and Raul Gomez, for instance, offer a version of Northern Irishness that is far from shared or inclusive.[5] They conceive of N Irish as an identity that could help secure nationalist acceptance of the north’s place in the UK. In this sense, Tonge and Gomez emulate the Belfast Telegraph’s dramatic conclusion about Northern Irishness. The authors’ scholarly account is subtler and more nuanced—that is to say, more obscure and evasive—than is the media portrayal, but its constitutional implications are the same. By moving nationalists to recognize the north as a “country”—just as unionists and loyalists do—the development of Northern Irishness could erase constitutional division, stabilize politics and consolidate the north as a political entity (Tonge & Gomez, 2015). This conceptualization is heavily biased in favour of unionism. In effect, it sees Northern Irishness less as a shared or common identity than as a means of enticing nationalists to embrace the core element of a unionist identity: acceptance of the union.

Figures 1 and 2 demonstrate the form and degree of partiality in Tonge and Gomez’s theorization of the N Irish identity. Figure 1 shows a conventional understanding of unionism and nationalism as different combinations of cultural and political elements. The cultural dimension on the vertical axis ranges from British to Irish; the political dimension on the horizontal axis goes from Union to Unity. In this two-dimensional space, unionists are located at the “British” pole of the cultural dimension and the “Union” pole of the political dimension. Nationalists, their polar opposite, are located at the “Irish” and “Unity” ends of those dimensions.

Figure 2 adds N Irish to the mix, which in Tonge and Gomez’s formulation is located somewhere between British and Irish. Let’s assume for the sake of argument that as a shared or cross-community cultural identity, Northern Irishness draws equally from both cultural traditions and is therefore located exactly halfway between Britishness and Irishness.[6] But N Irish is not a shared or cross-community allegiance on the political dimension; according to the authors, it’s anchored firmly at the Union pole. The length of the journey from one identity to the next indicates the extent of bias in Tonge and Gomez’s view of Northern Irishness. To become N Irish, unionists must make one move: they traverse half the distance along the vertical cultural axis, from the British end to the middle. The nationalist journey to Northern Irishness is, in contrast, much longer and requires two moves. Nationalists, like unionists, travel half the distance along the cultural axis, but from the Irish end to the middle. In addition, nationalists move the entire distance along the horizontal political axis, from the Unity to the Union end. Unionists make no equivalent journey.


In summary, for unionists, adopting Northern Irishness is a change in the cultural dimension of their identity but an affirmation of their identity’s political dimension: the union remains. For nationalists, adopting Northern Irishness is likewise a change in the cultural dimension of their identity but a denial of its political dimension: Irish unity vanishes. In continuing to refer to the potential of Northern Irishness as a shared identity, Tonge and Gomez miss this striking asymmetry in their account. They
obfuscate two vastly different methods of reducing intergroup conflict: “one based on a shared compromise between the aspirations of the two traditions versus a reduction in division due to an acceptance of the dominant view” (Hayes, McAllister & Dowds, 2007, p. 476). Their conceptualization of Northern Irishness is not a shared accommodation occupying a middle ground between opposing traditions, but a nationalist affirmation of unionist political orthodoxy. As the admittedly schematic portrayal in Figure 2 makes clear, the Tonge-Gomez rendition of N Irish represents a lopsided middle, tilted heavily in favour of unionism.[7]

The Alliance Party’s placement on the middle ground is also problematic. Alliance began life as an avowedly liberal party attempting to attract voters from across the community divide, although it supported the north’s constitutional position as an integral part of the UK. It expressly positioned itself at the pluralist political centre, in contrast to parties defined by unionism and nationalism (Alliance Party, 1970; Farry & Neeson, 1999). Many commentators accepted Alliance’s claim of being a crucial part of the moderate middle ground in northern politics, even as it was an expressly unionist party. Alliance has since shifted from this explicit unionist position to pronounce itself neutral on the constitutional question. And commentators readily endorse the new basis for Alliance’s claim to the middle ground. Tonge believes that Alliance’s move to a position of neutrality or agnosticism consolidates its position as a centrist party (Tonge, 2020). Murphy, as mentioned above, views Alliance as part of the middle ground precisely because of its constitutional neutrality, which she says is rooted in the Good Friday Agreement’s principle of majority consent to constitutional change (Murphy, 2023).

Three clarifications are needed to appreciate Alliance’s complex and contradictory relation to constitutional change. First, neutrality based on the consent principle is not a policy position on the constitution. All the other main political parties in the north accept consent, at least notionally, but go further to articulate explicit positions on the constitutional question, which—it’s useful to recall—is whether the north should remain part of the UK or form part of a united Ireland. The DUP and UUP want to maintain the union, Sinn Féin and the SDLP to bring about Irish unity. Alliance’s consent-induced agnosticism is fundamentally the avoidance of making a constitutional decision (Bradley, 2024). Party leader Naomi Long actually acknowledges that neutrality is not a position on the binary constitutional question, but promises that Alliance will develop a position as the “facts and evidence” about all the implications of Irish unity become clear (News Letter, 2022). As this condition is unlikely ever to be met to everyone’s satisfaction, Alliance’s claim of constitutional agnosticism is likely to be with us for the foreseeable future and beyond.

The second and third clarifications address Alliance’s practical and policy alignment as a unionist party, which belies its pretense to neutrality and agnosticism. Alliance’s current refusal to take a position on constitutional change renders change highly unlikely or impossible (Burke, 2020b). The GFA requires the Secretary of State to trigger a border poll if it appears likely to him that a majority of voters would support Irish unity.[8] The more support that Alliance gets in elections or opinion polls, the less likely is the Secretary to believe that there exists a majority for unity (Mac Ginty, 2023). The Secretary will interpret every expression of support for Alliance as condoning the party’s stance of constitutional neutrality, and count every such vote as a vote against a united Ireland. In these circumstances, barring the complete collapse of Alliance’s support or the electoral decimation of unionism, the Secretary is unlikely ever to trigger a border poll. And no border poll means the maintenance of the union. Alliance’s very lack of a formal position on constitutional change is in effect a crucial prop to the union. It provides a convenient rationale for the Secretary of State’s structural inclination to keep Irish unity off the political agenda. A mutual vested interest in the maintenance of the constitutional status quo is also at work here. Alliance—like the Secretary of State—does not wish to address the question of Irish unity. Being forced into making an explicit decision on the border is something the party obviously wishes to avoid as it would cause unrest and perhaps rifts in the party’s diverse constituency of supporters.

The final clarification concerns a conspicuous discrepancy in Alliance’s constitutional posture that, remarkably, no one seems to have examined. The party’s repeated calls for the development of a fully federal UK as part of a durable devolution settlement sit uneasily beside its claim of constitutional neutrality (Alliance Party, 2015, 2016, 2017a, 2017b & 2019). Alliance’s promotion of a federal United Kingdom reinforces British sovereignty in the north just as any other party’s advocacy of a federal united Ireland would underpin Irish sovereignty. Alliance has long maintained it wants the north to work, but it persists in wanting the north to work inside the union. All the party’s policy proposals and administrative plans assume only one constitutional future, an assumption that is inconsistent with the party’s and leader’s professed constitutional neutrality. Nor is that assumption consistent with the scholarly placement of Alliance on the “middle ground”.

Misinterpreting one end of the constitutional spectrum as the middle, partiality as neutrality, lopsidedness as balance, and asymmetry as symmetry is not unusual in discussions of the middle ground in northern politics. Brian Graham and Peter Shirlow, writing in 1998 during the talks that eventually led to the GFA, see the Catholic middle class in the north as one key element in the development of a middle ground. Why? Because that class, especially in Belfast, shows a heightened preference for retaining the union with Britain. Another main constituency in the growth of the authors’ middle ground is working-class Protestantism as represented by the Progressive Unionist Party. The PUP, the authors note, is increasingly pluralist and accepting of cultural Irishness, even though it remains highly committed to British sovereignty over the north. In this analysis, both elements of the developing middle ground—middle-class Catholics and working-class Protestants—deconstruct in their own way the two-communities model based on the stereotypes of “Catholic/Nationalist/Republican” and “Protestant/Unionist/British” (Graham & Shirlow, 1998, p. 245).

The authors fail to grasp the gaping imbalance and inequality in the accommodations that make their middle ground. For them, the loyalist tolerance of expressions of Irish culture equates to the nationalist abandonment of the political project of Irish reunification. This equivalence is false: the loyalist concession on culture pales in comparison to the nationalist capitulation on partition. Relatedly and more importantly, the authors employ unionist criteria to define the parameters of the middle ground: what unites loyalist workers in the PUP with relatively affluent Catholics in Belfast is their mutual acceptance of the north’s place within the UK. In this regard, Graham and Shirlow’s middle-ground groups fatally deconstruct the nationalist half of the two-communities dynamic as they reinforce its unionist half. The middle ground remains firmly entrenched on the sovereign territory of unionism.[9] It cannot be anything but lopsided.

The Northern Irish, the Alliance Party, well-off Catholics and loyalist workers are part of a middle so askew that perhaps it’s time to dispense with the use of the term “middle ground” altogether, or to specify explicitly and precisely why “middle” is an appropriate label for whichever group is being studied. In the next section, I identify circumstances in which middle-ground groups can legitimately be called “middle”. I’ll not examine in any detail Graham and Shirlow’s middle-ground groups. Middle-class Catholics are not now as much discussed as part of the middle ground as they were in the years surrounding the GFA; and the PUP, never big, is electorally insignificant, with no elected Assembly member since 2007 and only one local councillor. Instead, I’ll focus on the four middle-ground groups that feature in many contemporary discussions of northern politics: Alliance voters, neithers, the N Irish, and people with no religion.

Despite the obvious unionist bias in many specifications of the middle ground, examining the size, social profile and constitutional preference of middle-ground groups illuminates changes in the politics of the north, affords an opportunity for conceptual clarity and provides insights into the north’s constitutional future.

Data Analysis

I draw all the data for this analysis from the Northern Ireland Life and Times (NILT) surveys. These surveys, conducted annually since 1998 (except in 2011) under the auspices of Queen’s University Belfast, aim to yield representative samples of the north’s adult population (NILT, 2022a). The sample sizes range from 1800 in the survey’s first few years to 1200 between 2005 and 2020. In 2021 and 2022, the NILT surveyed some 1400 respondents. Margins of error are consistently within ±2.8 percent at the 95 percent confidence level (or 19 times out of 20). They fall to ±2.6 percent in 2021 and 2022.


Figure 3 shows the four questions in the 2022 survey that I use to identify the middle ground. This survey is the latest available but the NILT uses the same or equivalent questions in all other survey years.[10] The figure highlights how the questions can be interpreted as measuring each middle-ground group relative to its two-community counterparts:

1. Political Party Support: “Alliance Party supporters” versus those voting for the main unionist parties (DUP and UUP) or the main nationalist parties (Sinn Féin and the SDLP);

2. Political Identity: those respondents thinking of themselves as “neither” as opposed to unionist or nationalist;

3. National Identity: those describing themselves as “Northern Irish” rather than as British or Irish;[11] and

4. Religion: those saying they have “no religion” in contrast to those saying Protestant or Catholic.

The size of the middle ground

Figure 4 gives the 2022 distribution of responses to the four questions on party support, political identity, national identity and religion. The legend at the top of the figure emphasizes that the specifications of the middle-ground and comparison groups vary by question, as just discussed. The grey bars indicate the percentage size of middle-ground groups defined as Alliance/Neither/N Irish/No religion on the four variables. The orange bars show the size of the comparison groups correspondingly defined as Unionist/Unionist/British/Protestant. And green bars give the size of the comparison groups defined as Nationalist/Nationalist/Irish/Catholic.


Figure 4 underlines that the middle ground is a major player in the politics of the north. On the political identity variable, for instance, the middle-ground group is the largest group of all, with 38.3 percent of respondents saying they think of themselves as “neither unionist nor nationalist”. Some 30.7 percent identify as unionist and 26.3 percent as nationalist. This result is not unusual. The number of neithers has outstripped the number of unionists and nationalists in every NILT survey since 2006. Alliance voters, the middle-ground group on the party support variable, are also the largest group (26.7%) by a small margin over nationalist party supporters (25.6%) and unionist party supporters (23.8%). The other two variables confirm the significant size of middle-ground groups. On national identity, 30.6 percent of respondents say they think of themselves as N Irish as opposed to 31.9 percent saying Irish and 26 percent British. The no religion group is also of considerable size, with 29.2 percent of the sample having no religion compared to 37.8 percent who are Protestant and 33 percent Catholic.

The distributions in Figure 4 need to be interpreted carefully. The 2022 survey heavily overestimates the number of Alliance voters. It also continues an inaccurate pattern in many NILT surveys by underestimating the number of DUP and Sinn Féin supporters. These characteristics of NILT samples will tend to inflate the size of middle-ground groups. The 2021 census has two measures, religion and national identity, that we can compare directly to the 2022 NILT survey results. The census shows the no religion group at just 17.4 percent of the population compared to the NILT estimate of 29.2 percent (NISRA, 2022a). Some of this discrepancy, but not all of it, will be explained by measurement differences and survey sampling error. The census estimate of “Northern Irish only” (19.8%) is also considerably lower than the NILT figure for “Northern Irish” (30.6%). The census estimate rises to 31.5 percent if we add the number of people saying “Northern Irish only” to the number saying they had a Northern Irish identity along with other national identities—"Northern Irish and Irish” or “Northern Irish and British,” for instance (NISRA, 2022b). This larger census number is equivalent to the NILT’s 2022 estimate.[12]

The very size of the middle-ground groups, even using the lower census estimates, challenges the two-community model of northern politics around which much of the GFA’s institutional architecture has been constructed. It also partly explains the increasing references to three political blocs in the north, all of them minorities, which revise the conventional binary dynamic of majority-versus-minority (Doyle, 2022; Nolan, 2013 & 2021). Some of these discussions underestimate the resilience, political vigour and continuing institutional relevance of the two traditional communities, but middle-ground groups do add complexity to the northern polity (Ruane & Todd, 1992). The NILT data, with its measurement particularities and sampling imperfections, underline equality among the blocs. As the bottom set of bars in Figure 4 indicates, the average size of the three blocs, computed across the four measures listed, is in the ratio of 30:30:30.

Figure 5 demonstrates how the size of the middle ground varies over time. The figure introduces a new colour scheme that I use when comparing the four middle-ground groups to one another. I use a yellow line for Alliance voters, grey for neithers, red for N Irish and blue for the no religion group. The figure shows that the middle ground has long been a part of northern politics, but its social weight has increased, slowly and a little unevenly, since 1998. The average size of the four middle-ground groups rises from 18 percent of respondents in 1998 to over 31 percent in 2022. All four groups have grown in the period, with the steepest rates of growth in the no religion group and among Alliance supporters. For these two groups, a prolonged period of stagnancy and slow growth preceded a burst of relatively rapid growth since 2018. As examined above, Alliance’s spectacular rise in the 2019 elections, coinciding with a lengthy period of government breakdown, is an especially consequential change.

The census figures broadly confirm the trends for no religion and N Irish in Figure 5. They show that the no religion group has increased over time, with marked growth between the 2011 and 2021 censuses (Coakley, 2021; NISRA, 2022a). And they indicate that the numbers of N Irish stagnated in the last decade, as the NILT data also generally show (NISRA 2002b).



Social Composition

Knowing who the middle-ground groups are will help us understand their political attitudes and expected constitutional behaviour. What are their social background characteristics? How does their social makeup differ from the northern electorate as a whole?

In the section on the lopsided middle, I separately examined the N Irish and the Alliance Party. In this section, I’ll look at a different component of the middle ground, the neithers, people who say that they
think of themselves as neither unionist nor nationalist. As I’ll discuss, the social background characteristics of the neithers are generally similar to those of other middle-ground groups, and neithers can serve as a proxy for the rest.

Figure 6 profiles the neithers in respect of the three other middle-ground groups: Alliance supporters, N Irish and those with no religion. It also displays the profile of neithers on the social background variables of sex, age and occupation. Finally, the figure allows us to compare the social profile of neithers (the grey bars) to that of the sample as a whole (the black bars) to see how closely the two align.

Comparing the lengths of the grey and black bars is an easy way of determining group patterns of overrepresentation, underrepresentation and proportional representation. The first grey bar in Figure 6 shows that 44.3 percent of neithers are Alliance voters. The corresponding (and much shorter) black bar shows that only 26.7 percent of all respondents support Alliance. There are many more Alliance supporters in the ranks of the neithers than there are in the northern electorate as a whole. In other words, Alliance supporters are highly overrepresented among neithers. The same pattern of overrepresentation holds for the other middle-ground groups: the grey bar is much longer than is the black bar for both the N Irish and those with no religion. Those other middle-ground groups are also significantly overrepresented among neithers.


The pattern on the social background variables in Figure 6—sex, age and occupation—differs from the pattern we found for the middle-ground groups. Women are also overrepresented among the neithers but not to the same extent as are the middle-ground groups. The young and the old, and workers and management, are proportionally represented among neithers, with the grey and black bars of almost equal length.[13]

The bar chart for neithers displayed in Figure 6 closely resembles the bar charts for Alliance supporters, the N Irish and those with no religion, which I won’t reproduce here for reasons of space.[14] Two general similarities are noteworthy. First, no matter which particular middle-ground group we examine, the other middle-ground groups are strongly overrepresented in its ranks: the grey bars are much longer than the corresponding black bars, just as in Figure 6. That there is a robust relationship among the various middle-ground groups is not surprising. The Alliance Party, for instance, has long positioned itself as a voice for neithers and promoted a Northern Irish identity, and these groups along with people who have no religion are key elements in the party’s electoral successes (Farry & Neeson, 1999; Raymond, 2019; Tonge, 2020; Murphy & Evershed, 2022). The second similarity is that Alliance supporters, the N Irish and the no religion group show—again much like the neithers in Figure 6—a more or less proportional representation of women, youths and seniors, and the working and managerial classes. For these middle-ground groups, the corresponding grey and black bars are of equivalent lengths on the variables of sex, age and occupation.

There are some exceptions to this general profile. The N Irish stand out as distinct: of all the middle-ground groups, it is by far the most unionist and Protestant and the least nationalist and Catholic. One-third of the Northern Irish see themselves as unionist, for instance, compared to less than one-fifth of Alliance supporters and the no religion group. Some 44 percent of those with an N Irish identity are Protestants; the number of Protestants among Alliance voters and neithers is much lower, at about 29 percent. In the next section on attitudes towards the constitutional future of the north, we’ll see the effect of this distinctive profile of the N Irish. Regarding social background characteristics, Alliance is the most class-based group, with workers significantly underrepresented and managers strongly overrepresented among the party’s supporters. Those with no religion have the lowest representation of women, with men making up 58 percent of the group.

Constitutional Preference

Do middle-ground groups secure the north’s place in the UK, as some in the media and academy hope? This section examines the middle ground’s beliefs about the constitutional status of the north.


Figure 7 charts for 2022 the extent to which the middle ground and its comparison groups support union with Britain. The legend at the top of the figure reiterates that group definitions vary by question type. Figure 7 uses the NILT’s measure of long-term constitutional preference, which asks respondents: “Do you think the long-term policy for Northern Ireland should be for it to remain part of the United Kingdom, with direct rule; remain part of the United Kingdom, with devolved government; reunify with the rest of Ireland; be an independent state” (NILT, 2022b, p. 56). In the figure, I sum the responses of “remain with direct rule” and “remain with devolved government” to give an overall total of “support for union”. In calculating percentages, I use the full range of responses—including “don’t know”—although I show only the pro-union response in the figure.

On the political party variable, the orange bar indicates that fully 93.4 percent of unionist party supporters are in favour of the north remaining in the UK, the grey bar that 43.9 percent of Alliance voters support the union, and the green bar that just 14.1 percent of nationalist party voters prefer the union. The same pattern holds for the other variables of political identity, national identity and religion: high support for union among the Unionist/British/Protestant group, intermediate support among the various middle-ground groups, and low support in the Nationalist/Irish/Catholic camp. That is, Figure 7 underlines why in this instance the middle ground can legitimately be called “middle ground”. All four middle-ground groups are not as heavily in support of the union as are their orange counterparts nor as meagerly supportive as are their green comparison groups. They sit in the middle. Figure 8 gives the corresponding data for support for Irish unity, with equivalent results. Middle ground support for reunification stands in between the high support on the nationalist side and the low support on the unionist side. In sum, the middle ground is not distinguished from its orange and green comparison groups by having a different set of constitutional beliefs based on a wilful agnosticism or determined neutrality; rather, it’s distinguished by having a different distribution of the same set of beliefs. It’s levels of support of union and unity are middling.



Figure 9 clarifies the constitutional preferences of middle-ground groups and suggests why so many commentators claim the middle ground for the union. In each middle-ground group—Alliance voters, neithers, N Irish, and those with no religion—more respondents support union (orange bar) than support unity (green bar). Across all four groups, the average preference for union is 47.5 percent, for unity just 26.8 percent. For the N Irish, the gap in support for the two constitutional options is especially favourable to the union, with almost 60 percent supporting the constitutional status quo and merely 15 percent preferring a united Ireland. This kind of skewed distribution is to be expected given the distinctly unionist and Protestant social profile of the N Irish. Before I examine in more detail this pro-union snapshot of middle ground constitutional opinion, I want to look at a second notable finding evident in the figure.

Figure 9 also confirms that the middle ground is not particularly agnostic or neutral on the constitutional question. The number of “don’t know” responses among middle-ground groups, indicated by the grey bars, is very similar to the number of ‘don’t knows” among all respondents, shown at the bottom of the figure.[15] In other words, the middle ground makes up its mind about the constitution in roughly the same proportion as does the northern electorate as a whole. Middle-ground groups doubtlessly place less emphasis on constitutional issues than do their orange and green counterparts, and some individuals escape to the middle ground precisely to avoid the constitutional question altogether. But when asked to make a choice, middle-ground groups are prepared to commit to a constitutional position.


Alliance Party supporters have the fewest agnostics of all the middle-ground groups, although most of the differences are within the survey’s margin of error and cannot be regarded as statistically significant. But Alliance voters are consistent over time. In every NILT survey since 2007, they have the fewest “don’t know” responses compared to other middle-ground groups, and the differences are sometimes substantial. Alliance supporters seem to be ahead of the party and its current leadership in that they have a position on the constitution and are willing to express it well in advance of the presentation of facts and evidence on the possible ramifications of a united Ireland.

Let’s return now to the discussion of the middle ground’s relative preference for union over unity. This pro-union disposition probably gives some comfort to unionists and their allies in the media and academy. But recent trends in constitutional preference may cause them a measure of discomfort. The snapshot in Figure 9—a representation of constitutional preferences at a single point in time (2022)—says nothing of how opinions may be changing. Figure 10 depicts the middle ground’s support for union for the years 1998 to 2022, which allows us to see if levels of support vary over time. It resurrects the colour scheme used in Figure 5 to differentiate the four middle-ground groups and it adds a black line indicating support for union among all respondents.



Figure 11 shows the corresponding data on support for Irish unity; but note that it uses a different scale on the vertical (y) axis to give a sharper view of the yearly data points. I’ll pay particular attention to comparing opinions before and after the Brexit interruption (circa 2016), which appears as a watershed moment when constitutional trends begin observably to change (Burke, 2023). Generally speaking, the coloured lines of the middle-ground groups track the ups and downs of the black line of the electorate as a whole. Among both the middle ground and all respondents, the two figures chart steep declines in support for union and dramatic increases in support for unity beginning in the years 2015 to 2017.[16]

Focusing on the middle, Alliance voters’ current support for union is at its lowest ebb since 2003 and their support for unity is at its highest level ever in the NILT yearly series. In recent years, Alliance has cooperated with nationalist parties in opposition to unionism, most significantly on Brexit and relations with the EU, but also on such other issues as the flying of flags, same sex marriage and the status of the Irish language. This institutional alignment has not extended to the constitutional issue (Murphy, 2023; Ó Dochartaigh, 2022 ). Perhaps the new attitudes of Alliance voters evident in the figure are a portent of change. The trajectory of the no religion group is similar to that of Alliance supporters: the last few years witness historic lows in support of union and historic highs in support of unity. The neithers’ recent pro-union levels are the lowest since 2001 and their pro-unity levels the highest since 2006. The N Irish are a little anomalous. Compared to other middle-ground groups, the N Irish show the highest levels of support for union and the lowest levels of support for unity, especially in the past few years when their social profile has become increasingly unionist and Protestant. Even so, the extent of their support for union is lower in recent years than it was prior to 2017, and their support for unity higher. Overall, then, pro-union activists should be concerned about the middle ground’s recent direction of constitutional travel; and advocates of Irish unity might be encouraged but should not underestimate how much of the trip remains.


Figure 12, the final one, gives an alternative view of some of the data presented in the previous two figures. It is alternative in two senses: compared to Figures 10 and 11, it uses a different measure of constitutional preference to demonstrate the findings are robust, in that the patterns in the data remain stable across various measures. And it presents the yearly data in a different, condensed format to highlight how much change has occurred in a few short years.

Figure 12 charts the middle ground’s net preference for union using the “vote tomorrow” question asking respondents how they would vote on the north’s constitutional future if there were a referendum tomorrow. It compares constitutional attitudes in only two years, 2017 and 2022.[17] “Net preference for union” equals the percentage of a given group supporting union minus the percentage of that group supporting unity. Any positive difference indicates a group’s net preference for union, with more members of the group supporting union than unity. A negative difference indicates a net preference for unity. All the results in the figure are positive, showing a net preference for union in each middle-ground group. Let’s take Alliance supporters to illustrate the calculations. In 2017, 61.8 percent of Alliance voters support union on the “vote tomorrow” question and 15.4 percent support unity, yielding a difference (or net preference for union) of 46.4 percentage points, which is shown on the dark orange bar. The light orange bars shows the corresponding 2022 net preference for union falls to just 11.8%, computed as 45.8% union – 34% unity. Alliance voters’ net preference for union shrinks greatly between 2017 and 2022 both because their support for union falls and their support for unity rises.


The most notable feature of Figure 12 is the vast discrepancy in the lengths of the differently shaded orange bars: the light orange bars are much shorter than the dark ones. For all middle-ground groups but the N Irish, the net preference for union decreases markedly from 2017 to 2022. It drops especially rapidly among Alliance voters and the no religion group. These results are consistent with what we found previously using the NILT’s question on long-term constitutional preference, which showed the middle ground’s decreasing support for union and increasing support for unity. Union’s once large referendum lead over unity is becoming increasingly tenuous. It’s still a lead but it’s getting smaller and smaller.

Conclusion

The middle ground is malleable. Group opinions on the constitutional question are in a period of rapid and patterned change. There is everything to play for should a referendum be called. Which way middle-ground groups swing will help to determine the outcome of any border poll. If the trends I’ve documented continue, the outlook is promising for advocates of a united Ireland. The recent finding that—for the first time ever—more Alliance Party members favour Irish unity than support continued union is consistent with these trends (Breen, 2024a). But no one can know if such trends will persist. As the constitutional debate progresses, we’ll see at least two important developments that may reinforce or reverse current opinions. Facts and evidence about the implications of constitutional change will accumulate, with no doubt some reasoned, forceful and legitimate debate about the nature of the facts and evidence and their purported consequences. But that’s not all that will accumulate: so will partisan distortions, scaremongering misrepresentations, outlandish claims, and outright lies. The Working Group on Unification Referendums on the Island of Ireland frames the question as “how to ensure that voters have access to high-quality information and are not unduly exposed to misinformation” (Working Group, 2021, para. 14.41). I think we can expect that voters will have ready access to reliable information but will also be inundated with misinformation. The important question then becomes which will prevail, cogent argument or calculated deception.

The constitutional option of Irish unity has benefitted from a “Brexit bounce,” an upsurge in its popularity because of the complications and controversies of the UK leaving the European Union (Burke, 2023). Middle-ground groups, with the partial exception of the N Irish, are much more likely to say that Brexit has made them feel more rather than less in favour of a united Ireland.[18] And Figures 11 and 12 show concrete evidence that middle ground support for unity has increased since Brexit. This pro-unity bounce might moderate if the disruptive effects of Brexit abate and the attendant disputes subside. We’ll see. Regular opportunities for Brexit-related tumult will arise from pervasive trade issues, the Stormont Brake, the Assembly consent mechanism, continuing British government unpreparedness for leaving, and the Labour Party’s plan to renegotiate some of the terms of withdrawal from the EU. But Brexit’s intrusion on public debate in the north will ebb and flow. In the face of such fluctuations, advocates of Irish unity would be well-advised to refine their appeals to ensure that the middle ground continues on a trajectory favourable to reunification.

The middle is not the only key constituency in any coming border poll. Also important for advocates of unity will be firming up support among soft republicans and softer nationalists, appealing to that section of unionists who might waver in their current support for the constitutional status quo, engaging newly-eligible voters, and ensuring that all the disparate elements of their constituency turn out to vote on referendum day. A major problem is that the legal power to call a border poll rests solely in the hands of the Secretary of State, and his authority in this matter is as arbitrary as it is unaccountable. The “coming” border poll might never arrive.

Notes

[1] Brendan O’Leary and others proposed a system of “shared authority”, and explored a number of different models of co-sovereignty, including those of the New Ireland Forum, the Kilbrandon Committee, the SDLP and academics Basil Chubb, Bernard Crick and Frank Wright (O’Leary et al., 1993). O’Leary soon moved away from shared authority towards a “double protection” model, a scheme broadly consistent with the Good Friday Agreement (McGarry and O’Leary, 1995).

[2] The NILT surveys asked questions about joint authority only once, in 2018, more than one year into a three-year breakdown in devolved governance occasioned by the resignation of Martin McGuinness as Deputy First Minister in January 2017. The 2018 survey shows that 17 percent of respondents would find joint authority “almost impossible to accept”, 29 percent “would not like it, but could live with it if you had to”, 35 percent “would happily accept” it, and 20 percent say “don’t know”.

[3] Belfast Telegraph columnists continue to push the narrative that the growth of the Northern Irish identity means the end of Irish unity. In June 2023, John Laverty noted that the middle-ground supporters of Alliance eschew the constitutional extremes and “feel more at home in the middle of a road that appears to be leading somewhere; i.e., a Northern Ireland full of Northern Irish people content for things to stay just as they are” (Laverty, 2023).

[4] The following works typify the range of research on the “Northern Irish” identity: Moxon-Browne (1991); Hayes and McAllister (2009); McGlynn, Tonge and McAuley (2014); McKeown (2014); and Garry and McNicholl (2015).

[5] Tonge and Gomez actually illustrate both the empirical and theoretical limits of N Irish identity. Like others, they find that survey data do not support the characterization of Northern Irishness as a compromise, shared identity. But they continue to view N Irish as potentially a shared identity that, if developed, could be a key stabilizer of northern politics. My argument is that their very conceptualization of Northern Irishness actually rules out the potential of meaningful sharing. To the extent that they tie N Irish to maintenance of the union, they render the identity much closer to unionism than to nationalism.

Even the temperate academic renditions of the N Irish identity tend toward the unionist bias in Tonge and Gomez. These moderate voices explicitly point out how the Protestant version of the identity reinforces the constitutional status quo, but they remain silent on whether the Catholic version challenges current constitutional arrangements. That is, the Catholic territorial outlook on Northern Irishness is probably compatible with maintenance of the union, but the Protestant political interpretation is definitely incompatible with Irish unity. This moderate interpretation of the N Irish identity seems tilted in favour of unionism.

[6] I’ll leave aside the difficult question of what kind of cultural symbols of Northern Irishness might be considered part of a shared or cross-community allegiance. Recent research suggests that the level of hostility to or acceptance of cultural symbols like the Red Hand of Ulster depends on how they are presented to the public (Garry, O’Leary & Pow, 2024).

[7] Garry, O’Leary and Pow (2023) measure and conceptualize N Irish differently from Tonge and Gomez; they see the identity as possibly “a halfway house or bridging identity for Protestant unionists”. My analysis of NILT data shows little support for their view.

[8] I’ve argued extensively that the Secretary of State’s so-called mandatory duty to call a border poll is meaningless (Burke, 2020a). Here I assume, for the sole purpose of explicating the logical consequences of Alliance’s claimed neutrality, that the mandatory duty is a real obligation that would be implemented in the circumstances.

[9] The analyses of Graham and Shirlow and Tonge and Gomez are similar to earlier unionist arguments advocating a “non-political Irishness.” Tonge and Gomez even examine a form of non-political Irishness without using that term. These earlier arguments were not tied to the notion of the middle ground but like that notion suggested that nationalists must give up their aim of unity in order for the north to stabilize, normalize and prosper. The very act of abandoning reunification will ease unionism’s hostility to this newly-innocuous form of cultural nationalism (Haslett, 1995; Kennedy, 1995; Roche, 1995, Aughey, 1996). The asymmetry in non-political Irishness is obvious, not only because that kind of identity demands much of nationalists and next to nothing of unionists, but also because its advocates never even consider the equivalent notion of a “non-political Britishness.”

Cathal McCall’s discussion of how the GFA transformed identities in the north is similarly tendentious and asymmetrical. He contends that the Agreement’s North/South Ministerial Council (especially) and British-Irish Council (perhaps) helped to establish a model of transterritorial governance for the north. In this model, transterritorialism undermines conventional concepts of self-determination and indivisible territorial sovereignty, and points to a settled future of co-determination, co-sovereignty and pooling of sovereignty. The GFA’s transterritorial institutions in turn may lead to new forms of communal and political identities in the north. McCall sees republicanism’s acceptance of the legitimacy of the north and unionism’s acceptance of power-sharing as incipient forms of identity realignment that could potentially undermine old identity forms based on exclusion and nation-state territorialism (McCall 2001).

There is much wrong with McCall’s analysis. He greatly overestimates the extent to which globalization has weakened the capacity of national states. He exaggerates the importance and impact of the NSMC and BIC. The GFA’s Strand One institutions (Executive and Assembly) are the key institutions of the Agreement, not the weak and relatively insignificant institutions of Strand Two (NSMC) and Strand Three (BIC). McCall himself (2018) later acknowledged the limited success of the NSMC. Transterritorialism is largely a myth, as increasingly frequent unilateral applications of British sovereignty in the north repeatedly demonstrate. Britain’s resiling from the Stormont House Agreement, its Internal Market Bill, Protocol Bill, Legacy Bill, imposition of budget cuts, and Safeguarding the Union command paper—to name but a few recent examples of unilateralism—discredit the claim of movement towards a developing co-sovereignty in the north. Most importantly, for the purpose of this paper, is McCall’s identification of the nature of identity realignment. Despite all his talk of transterritorialism, McCall’s emphasis on Sinn Féin’s acceptance of partition leaves intact British sovereignty in the north. This massive Sinn Féin concession is by no means balanced by unionism’s halting, contradictory and limited acceptance of power-sharing in the north.

Coulter and others’ examine groups who disrupt the two-communities model of northern politics that pits Protestant/unionist/British against Catholic/nationalist/Irish (Coulter et al., 2021, chap. 5). Their analysis likewise reinforces the north’s place in the UK. Their disrupting groups—including the N Irish, those with no religion and those who identify as neither unionist nor nationalist—are distinguished by their preference for union over unity. In the authors’ view, disrupting the two-communities binary apparently involves adopting the constitutional preference of one community.

The common element in these disparate discussions of non-political Irishness, transterritorial identity realignment, and binary disruption is that union not unity is the constitutional future of the north, a settlement which heavily privileges unionism over nationalism.

[10] The same questions are used in every survey except for the political party question. The NILT uses three slightly different but equivalent sets of questions to measure political party support. The different questions give rise to different variable names for political party support: NIPARTY from 1998 to 2006, POLPARTY in 2007 and 2008, and POLPART2 since 2009.

[11] The “Ulster” identity never polled more than 5 or 6 percent in NILT surveys. By 2022, the number of respondents identifying as Ulster is only 1.6 percent, which is typical of NILT results since 2014. As John Coakley notes, the Ulster identity shrinks “to the point of disappearance”, mostly because of movement into the Northern Irish category in the mid- and late-1980s (Coakley, 2021, p. 35).

[12] If we compare the same years—the 2021 NILT survey to the 2021 census—the same conclusions hold. The NILT estimates for neithers and the no religion group are virtually identical in 2021 and 2022.

[13] The NILT uses the National Statistics Socio Economic Classification (NSSEC) as a measure of occupation. Figure 6 shows the distributions of the lowest occupation, coded as “Semi-routine and routine occupations”, and the highest, coded as “Managerial, administrative and professional occupations.” NILT analysts derive the variable NSSEC1 from various questions about the work situation of respondents (NILT, 2022b, p. 70).

[14] The other figures are available on request.

[15] Especially in the last three years, the middle ground’s “don’t know” numbers have converged on those for all respondents. Neithers are an exception, just as in Figure 9.

[16] Middle-class Catholics are on a similar trajectory. In the text of the paper I argue that Graham and Shirlow’s middle-class Catholics cannot conceptually be considered part of the middle ground: the authors place them not in the middle but at the union end of the constitutional continuum. The empirical rationale for this conceptual confusion has collapsed. The proportion of middle-class Catholics favouring union as a long-term preference declines from 47.8 percent in 2017 to 16.7 percent in 2022. The 2022 number is markedly lower than any of the pro-union percentages reported for the class in the authors’ 1996 study. NILT data also reveal that the proportion of middle-class Catholics favouring unity rises from 39.1 percent to 62.3 percent in the years from 2017 to 2022.

[17] The NILT uses two slightly different questions. In 2017, the first time the question is asked since Brexit, the wording is: “If there was a referendum tomorrow about whether Northern Ireland should leave the UK and unite with the Republic of Ireland, how do you think you would vote? I would vote for Northern Ireland to remain part of the UK; I would vote for Northern Ireland to unite with the Republic of Ireland ...” BORDPOLL is the name of this variable in the NILT data set (NILT, 2017, p. M 27). In 2022, the wording of the question is: “Suppose there was a referendum tomorrow on the future of Northern Ireland and you were being asked to vote on whether Northern Ireland should unify with the Republic of Ireland. Would you vote ‘yes’ to unify with the Republic or ‘no’? Yes, should unify with the Republic; No, should not unify with the Republic ...” (NILT 2022b, p. 56). REFUNIFY is the variable name.

[18] In 2022, for instance, 52.3 percent of Alliance voters say that Brexit has made them feel more in favour of a united Ireland and only 4.9 percent say less in favour. The corresponding percentages for the other middle-ground groups are: neithers, 43 percent and 6.8 percent; N Irish, 33 percent and 11.2 percent; and the no religion group, 44 percent and 7.9 percent.

References

Alliance Party. (1970). Declaration of Intent – A New Political Charter. 21 April. Retrieved from.

Alliance Party. (2015). Step Forward Not Back: Manifesto for 2015 Westminster Election. Retrieved from.

Alliance Party. (2016). Forward Faster: [Assembly] Manifesto 2016. Retrieved from. 

Alliance Party. (2017a). How to Change Northern Ireland for Good: Manifesto 2017. Retrieved from.  

Alliance Party. (2017b). Change Direction: Westminster Manifesto 2017. Retrieved from. 

Alliance Party. (2019). Demand Better: Westminster Manifesto 2019. Retrieved from.

Aughey, A. (1996). “Obstacles to Reconciliation in the South.” In Building Trust in Ireland. Studies Commissioned by the Forum for Peace and Reconciliation, 2-51. Belfast: Blackstaff Press in association with the Forum for Peace and Reconciliation.

Belfast Telegraph. (2012a). “The rise of the Northern Irish.” 12 December. Retrieved from the LexisNexis electronic database of news articles. [Cited below as LexisNexis].

Belfast Telegraph. (2012b). “Foundations laid for shared future.” 12 December. Lexis/Nexis.

Belfast Telegraph. (2012c). “Now is a time for leaders.” Editorial. 12 December. Lexis/Nexis.

Bradley, D. (2024). “It’s shameful that Alliance will not come off constitutional fence.” Irish News. 5 January. Retrieved from.

Breen, S. (2023a). “The little-known group who could hold the key to the future.” Belfast Telegraph. 4 March. Retrieved from the Factiva (Dow Jones) electronic database of news articles. [Cited below as Factiva].

Breen, S. (2023b). “Sinn Fein is still the most popular party but DUP narrows gap to 3%: Poll reflects shift in unionist feeling to Stormont boycott.” Belfast Telegraph. 5 November. Retrieved from. 

Breen, S. (2024a). “More Alliance Party members would vote for Irish unity than to remain in UK: survey.” Belfast Telegraph. 3 March. Retrieved from. 

Breen, S. (2024b). “Middle ground will decide border poll... and it's now growing.” Belfast Telegraph. 4 March. Factiva.

Burke, M. (2020a). “Stealing Irish Unity: The Repertoire of Thieves,” Part 1. The Pensive Quill.

5 September. Retrieved from. 

Burke, M. (2020b). “Stealing Irish Unity: The Repertoire of Thieves,” Part 4. The Pensive Quill.

26 September. Retrieved from

Burke, M. (2023). “The Crisis of Democratic Legitimacy in the North.” The Pensive Quill. 18 July. Retrieved from. 

Clarke, L. (2012). “Link between our religion and our politics is fast weakening.” Belfast Telegraph. 12 December. Lexis/Nexis.

Coakley, J. (2021). “Is a middle force emerging in Northern Ireland?” Irish Political Studies 36:1 (March): 29-51.

Coulter, C., N. Gilmartin, K. Hayward and P. Shirlow. (2021). Northern Ireland a generation after Good Friday. Manchester: Manchester University Press.

Doyle, O. (2022). “The Irish Unification Debate After the Northern Ireland Assembly Elections.” IACL-AIDC Blog. 24 May. Retrieved from. 

Dudley Edwards, R. (2012). “Unionist Forum welcome in these desperate times.” Belfast Telegraph. 20 December. Lexis/Nexis.

Farry, S., and S. Neeson. (1999). “Beyond the band-aid approach: an Alliance party perspective upon the Belfast Agreement.” Fordham International Law Journal 22:4 (April): 1221-1249.

Garry, J. and K. McNicholl. (2015). “Understanding the ‘Northern Irish’ Identity.” Briefing Paper. Knowledge Exchange Seminar Series. Retrieved from. 

Garry, J., B. O’Leary and J. Pow. (2023). “Northern Ireland identity may serve as less adversarial for
politically moderate Catholics and Protestants.” Irish Times. 11 December. Retrieved from. 

Garry, J., B. O’Leary and J. Pow. (2024). “Can the Red Hand of Ulster be transformed into a unifying
image for the island of Ireland?” Irish Times. 22 January. Retrieved from. 

Graham, B., and P. Shirlow (1998). “An elusive agenda: the development of a middle ground in Northern Ireland.” Area 30:3 (September): 245-254.

Hanna, G. (2024). “The battle for the middle ground in a future border poll is under way... and this is what they want.” Belfast Telegraph. 13 March. Factiva.

Haslett, Edgar. (1995). “Historical Injustices.” In The Idea of the Union: Statements and Critiques in Support of the Union of Great Britain and Northern Ireland. ed. J.W. Foster, 121-127. Vancouver: Belcouver Press.

Hayes, B.C., and I. McAllister. (1995). “Religious independents in Northern Ireland: Origins, Attitudes and Significance.” Review of Religious Research 37:1 (September): 65-83.

Hayes, B.C., and I. McAllister. (2009). “Religion, identity and community relations among adults and young adults in Northern Ireland.” Journal of Youth Studies 12:4 (August): 385-403.

Hayes, B.C., I. McAllister and L. Dowds. (2007). “Integrated Education, Intergroup Relations, and Political Identities in Northern Ireland.” Social Problems 54:4 (November): 454–482.

Hayward , K. (2020). “The 2019 General Election in Northern Ireland: the Rise of the Centre Ground?” Political Quarterly 91:1 (January-March): 49-55.

Kennedy, D. (1995). “The Realism of the Union.” in The Idea of the Union, 27-36.

Laverty, J. (2023). “SF's so-called 'little helpers' have put Irish unity on back burner.” Belfast Telegraph. 1 June. Factiva.

LucidTalk. (2023). Main Report: LucidTalk NI Autumn ‘Tracker’ Poll – October/November. 3 November. Retrieved from.

McAdam, N. (2012). “DUP can entice more Catholics: Robinson.” Belfast Telegraph. 3 December. Lexis/Nexis.

Mac Ginty, R. (2023). “The Alliance Party of Northern Ireland: Are they really ‘other’?” Royal Irish Academy. 2 June: Retrieved from. 

McCall, C. (2001). “The Production of Space and the Realignment of Identity in Northern Ireland.” Regional & Federal Studies 11:2 (Summer): 1-24.

McCall, C. (2018). “The Travails of the North/South Ministerial Council.” Political Studies Association,

PSA Blog. 10 April. Retrieved from. 

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

McGlynn, C., J. Tonge and J. McAuley. (2014). “The Party Politics of Post-devolution Identity in Northern Ireland.” British Journal of Politics and International Relations 16:2 (May): 273-290.

McKeown, S. (2014). “Perceptions of a Superordinate Identity in Northern Ireland,” Peace and Conflict: Journal of Peace Psychology 20:4 (November): 505-515.

Moxon-Browne, E. (1991). “National Identity in Northern Ireland.” In Social Attitudes in Northern Ireland: The First Report. ed. P. Stringer and G. Robinson. Belfast: Blackstaff Press. Retrieved from. 

Murphy, M.C. (2023). “The Rise of the Middle Ground in Northern Ireland: What does it Mean?” Political Quarterly 94:1 (January/March): 95-103.

Murphy, M.C., and J. Evershed. (2022). A Troubled Constitutional Future: Northern Ireland after Brexit. Newcastle upon Tyne: Agenda Publishing.

News Letter. (2022). “Naomi Long: Alliance Party will take a stance on constitutional question in the future.” 6 March. Factiva

NISRA. (2011). Northern Ireland Statistics & Research Agency. 2011 Census: Individual Questionnaire Northern Ireland. Retrieved from. 

NISRA. (2012). Northern Ireland Statistics & Research Agency. Statistics Bulletin. Census 2011: Key Statistics for Northern Ireland. 11 December. Retrieved from. 

NISRA. (2022a). Northern Ireland Statistics & Research Agency. Census 2021: Main Statistics for Northern Ireland: Statistical Bulletin – Religion. 22 September. Retrieved from.

NISRA. (2022b). Northern Ireland Statistics & Research Agency. Census 2021: Main Statistics for Northern Ireland: Statistical Bulletin – National Identity. 22 September. Retrieved from. 

NILT. (2017). ARK. Northern Ireland Life and Times Survey 2017: Main Questionnaire. Retrieved from.

NILT. (2022a). ARK. Northern Ireland Life and Times Survey, 2022 [computer file]. Retrieved from. 

NILT. (2022b). ARK. Northern Ireland Life and Times Survey 2022: Questionnaire: CAWI (online). Retrieved from. 

Nolan, P. (2013). The Northern Ireland Peace Monitoring Report: Number Two. March. Belfast: Community Relations Council. Retrieved from. 

Nolan, P. (2021). “Running out of Road.” Dublin Review of Books 134 (June). Retrieved from.

Nolan, P. (2022). “Usuns, Themuns, Others.” Dublin Review of Books 148 (October). Retrieved from. 

Ó Dochartaigh, N. (2021). “Beyond the dominant party system: the transformation of party politics in Northern Ireland.” Irish Political Studies 36:1 (March): 7-28.

Ó Dochartaigh, N. (2022). “The rise and rise of a new majority in Northern Ireland.” UK in a Changing Europe. 18 May. Retrieved from. 

O’Leary, B. (2022). Making Sense of a United Ireland: Should It Happen? How Might It Happen? Sandycove an imprint of Penguin Books.

O’Leary, B., T. Lyne, J. Marshall, and B. Rowthorn. (1993). Northern Ireland: Sharing Authority. London: Institute for Public Policy Research. October.

Raymond, C. (2019). “Alliance Breakthrough Part of a Broader Trend.” Queen’s Policy Engagement. 28 May. Retrieved from. 

Roche, P.J. (1995). “Northern Ireland and Irish Nationalism,” in The Idea of the Union, 128-134.

Ruane J, and J. Todd. (1992). “Diversity, Division and the Middle Ground in Northern Ireland.” Irish Political Studies 7:1, 73-98.

Tonge, J. (2020). “Beyond Unionism versus Nationalism: the Rise of the Alliance Party of Northern Ireland.” Political Quarterly 91:2 (April–June): 461-466.

Tonge, J. (2022a). “The 2022 Northern Ireland Assembly Elections: Polling, Power-Sharing, Protocol.” Political Insight 13:2 (June): 10-13.

Tonge, J. (2022b). “Voting into a Void? The 2022 Northern Ireland Assembly Election.” Political Quarterly 93:3 (July/September): 524-529.

Tonge, J., and J. Evans. (2020). “Northern Ireland: From the Centre to the Margins?” Parliamentary Affairs 73: Supplement 1 (September): 172-88.

Tonge, J., and R. Gomez. (2015). “Shared Identity and the End of Conflict? How Far Has a Common Sense of ‘Northern Irishness’ Replaced British or Irish Allegiances since the 1998 Good Friday Agreement?” Irish Political Studies 30:2 (April): 276–298.

Whitten, L.C. (2020a). “#LE19 – a turning of the tide? Report of local elections in Northern Ireland, 2019.” Irish Political Studies 35:1 (January): 61–79.

Whitten, L.C. (2020b). “Breaking Walls & Norms: A Report on the UK general election in Northern Ireland, 2019.” Irish Political Studies 35:2 (April): 313-330.

Whitten, L.C. (2023). Constitutional change in Northern Ireland. Institute for Government and Bennett Institute for Public Policy, Cambridge. August. Retrieved from. 

Working Group. (2021). Working Group on Unification Referendums on the Island of Ireland. Final Report. The Constitution Unit, School of Public Policy, University College London. May. Retrieved from.

⏮ Mike Burke has lectured in Politics and Public Administration in Canada for over 30 years.

The Inhabitants Of Middle Earth: Exploring The ‘Centre Ground’ In Northern Politics

Mike Burke ✍Surveying the popular and scholarly literature reveals at least four key groups that currently make up the middle ground in the north of Ireland: (1) Alliance Party supporters, (2) people who say they are “neither” unionist nor nationalist, (3) those who identify as “Northern Irish”, and (4) those who have no religion (Breen, 2024b; Tonge, 2020). 

 They are “middle” in contradistinction to the two communities of Unionist/British/Protestant and Nationalist/Irish/Catholic. The four middle-ground groups are also collectively referred to as “others”. I’m primarily interested in the life of these groups in their guise as components of the middle ground. I examine the conceptualization, size, social composition, and constitutional preferences of these four middle-ground groups.

Recent events highlighted the centre of the north’s political spectrum. Alliance’s strong performance in the three 2019 elections—local, European and Westminster—focused attention on the party and its energized support base in the other middle-ground groups. Alliance reinforced its success in the 2022 Assembly election to become the third largest party in the north, surpassing the seat and vote totals of the UUP and SDLP (Tonge, 2020, 2022a & 2022b; Hayward, 2020; Whitten, 2020a & 2020b).

The consolidation of the middle ground relates to larger social and political processes in the north. The current surge in Alliance’s electoral success may herald a transformation of the northern party system in which three sizable blocs compete for votes from an electorate with a non-unionist majority (Ó Dochartaigh, 2021 & 2022; Raymond, 2019; Murphy, 2023). The rise of “neithers” relative to the traditional categories of unionists and nationalists is the result of a broader dealignment in communal-political identification. The steady popularity of a Northern Irish identity may signal a weakening of the historically strong affiliation of religion and politics that defined Protestant-and-British in opposition to Catholic-and-Irish. And the social weight of the no religion group emerges from wider developments in the deconfessionalization and secularization of northern society (Coakley, 2021; Hayes & McAllister, 1995).

The middle ground is typically and centrally defined by its presumptive location on the constitutional question. Just as the Unionist/British/Protestant grouping supports maintaining the union and its Nationalist/Irish/Catholic counterpart prefers reunifying Ireland, the middle ground has a distinctive view on the constitution. What academic Mary C. Murphy says in her study of the Alliance Party is true generally of all middle-ground groups:

... the middle ground represents a space which is neither unionist nor nationalist. It applies to parties and voters who are designated as ‘other’ as set out in the 1998 Belfast/Good Friday Agreement. In this way, the middle ground is not an indication of where a party lies on the left–right political spectrum; instead the label ‘middle ground’ or ‘centre party’ implies—first and foremost—neutrality on the constitutional question (Murphy, 2023, p. 95).

In conventional usage, then, the term “middle” applies to those groups that are intermediate or neutral or agnostic on the union-versus-unity divide. The middle ground stands halfway between the constitutional endpoints of remaining in the United Kingdom and reunifying Ireland.

I show that the conventional definition of the middle ground is highly inaccurate. Middle-ground groups do not in fact occupy a middle or intermediate position. Nor are they especially neutral or agnostic on the constitutional question, as Murphy and others contend. The conventional definition of the middle ground can also be exceptionally misleading. I argue that some conceptualizations of the middle ground are not constitutionally innocent but consciously partisan, enlisting the middle firmly in the service of maintaining the union. They in effect drag the middle to the pro-union end of the constitutional continuum, all the while claiming that this altered position somehow still represents the middle.

If constitutional politics is inherently bound up with the definition of the middle ground, it also accounts for the increasing public interest in middle-ground groups. As discussions of Irish unity and a border poll proliferate in post-Brexit Ireland, attention naturally turns to the constitutional decision that may confront northern (and southern) voters. If most unionists will vote for union and most but perhaps not as many nationalists for unity, what will the middle do? Many commentators assert that the middle-ground groups identified above—Alliance supporters, neithers, the N Irish and those with no religion—will play crucial roles in determining the outcome of any reunification referendum (Tonge and Evans, 2020; Murphy & Evershed, 2022; Breen, 2023a; Nolan, 2022; Whitten, 2023; Hanna, 2024). I demonstrate that this assertion is a reasonable use of the term “middle”, and examine the changing constitutional preferences of the middle-ground groups. I find that, in recent years, the constitutional opinions of the four groups have decidedly shifted, with support for union plummeting and support for unity burgeoning. Union still has the edge over unity, but its advantage is rapidly diminishing.

Other, less explicitly constitutional reasons also explain the rekindled interest in the middle ground. The congealing of the middle, especially Alliance’s recent successes in the context of Stormont dysfunction, helps to propel increasingly urgent voices in favour of reforming the north’s devolved institutions away from a “two-communities” perspective. In this paper, I’m concerned solely with the constitutional aspects of the middle ground and do not address the debate about institutional reform.

A Lopsided Middle

This section has two related aims: to show, first, that a middle ground actually exists but none of the middle-ground groups resides there and, second, that the place where middle-ground groups actually reside can in no way be considered a middle ground. These unhappy circumstances signify that Middle Earth is equal parts paradox and legend. It nevertheless remains worthwhile to probe the reasoning of those storytellers who continue to pretend that a densely populated Middle Earth does exist. And it’s useful to explore the social makeup and constitutional views of the peoples reputed to dwell there. For the inhabitants themselves are real, they are just not part of the actual middle ground.

The first aim is to identify the middle. Logically, the middle ground of the union/unity divide is some kind of joint authority or shared sovereignty in which Britain and Ireland act as co-sovereigns, each with binding powers and responsibilities over the north (O’Leary, 2022). This constitutional arrangement is a compromise position that stands between the poles of “United Kingdom” and “United Ireland.” Historically, various models of co-sovereignty enjoyed some fleeting governmental, party-political and academic interest but never reached the level of a popularly accepted option to the constitutional divide.[1] Currently, some political actors propose joint authority, sometimes called Plan B, as an interim alternative to direct rule from London during periods when devolved government is collapsed. In the governance breakdown just resolved, they used joint authority as an inducement (read threat) to convince the DUP to reenter Stormont. And the kind of joint authority they offered—giving the Irish government an enhanced consultative role within a system of unbridled British rule—fell well short of robust schemes for co-equal sovereignty. The most recent evidence suggests that an exasperated public might consider some undefined notion of joint authority in place of Stormont’s seemingly structural impairment (Breen, 2023b; LucidTalk, 2023).[2] This stance is not an active constitutional preference but seems very much a fallback position in the face of continual failures of devolution. In short, no one—least of all any of our four groups—actively advocates joint authority as a fair and reasonable compromise that could bring about a long-term resolution of constitutional divisions in the north.

Even though middle-ground groups are not located at the middle of the constitutional spectrum, I’ll continue to use the term “middle ground” because it’s a popularly accepted way to refer to the four groups I’m analyzing. For the moment, the reader should assume that my every use of middle ground is enclosed in ironic quotations, even though the quotations are not there. Let’s pretend they are. In the last substantive section on constitutional preference, I drop my ironic use of middle ground for reasons I explain there.

The second aim of this section is to identify where middle-ground groups actually reside, if not in the middle. I use “Northern Irish” identity as a case study to show the problematic conceptualization of the middle ground. I also challenge the Alliance Party’s contention—and the widespread acceptance of its claim—that it is part of the middle ground because it’s constitutionally neutral. We need to ask some serious questions about how the notion of the middle ground is used in analyzing the politics of the north.

The Northern Irish identity has been tracked in opinion surveys since the 1980s, but it caught the attention of the public only with the release, in December 2012, of the answers to the new 2011 census question on national identity. The new question asked “How would you describe your national identity?” and directed individuals to tick all the options that apply: “British, Irish, Northern Irish, English, Scottish, Welsh, Other” (NISRA, 2011, p. 5). Forty percent of northern residents said they were British only, 25 percent Irish only and 21 percent Northern Irish only (NISRA, 2012).

Media and scholarly responses to the appearance of the N Irish were, at first glance, starkly different. The media presentation was sensationalist; the academic response seemingly more measured. Reporting on the census results, the front-page headline of the Belfast Telegraph dramatically announced “The rise of the Northern Irish” (Belfast Telegraph, 2012a). The paper immediately drew a straight line between the N Irish identity and the maintenance of the constitutional status quo: many Catholics now regard themselves as “Northern Irish”; this new identity means that Catholics feel comfortable in the northern state; and this level of comfort means in turn that a united Ireland is inconceivable (Belfast Telegraph, 2012b & 2012c; Clarke, 2012).[3]

The usual unionist suspects, through the medium of the Belfast Telegraph, jumped in to pronounce on the implications of the new N Irish identity. Ruth Dudley Edwards triumphally remarked that “it doesn't matter how much republicans shout about a united Ireland and call for a border poll, the constitutional debate is over for generations” (Dudley Edwards, 2012). DUP leader Peter Robinson proudly proclaimed that nationalism is in crisis and its goal of a united Ireland “is further away than ever” (McAdam, 2012).

Temperate academic voices wondered if the N Irish label could prove to be a superordinate or inclusive “national” identity capable of effacing the old divisions separating the two communities. Could it be a middle-ground and shared identity that both traditions equally claim? “Northern Irish”, they reasoned, is an ambiguous identity capable of appealing, in different ways, to groups on either side of the constitutional divide. Catholics might see “Northern Irish” primarily in all-Ireland or territorial terms, as recognition that they inhabit the north eastern part of the island of Ireland. Protestants, on the other hand, might be attracted to the political meaning of “Northern Irish” as conferring constitutional legitimacy on the northern state. The speculation that N Irish could be an inclusive and shared identity crashed into some hard empirical realities. Survey data revealed that the number of Catholics identifying as “Northern Irish” fell away, and the label came increasingly to be claimed by disillusioned Protestants.[4] As I’ll show below, those identifying as N Irish are now the most heavily Protestant and unionist of the four middle-ground groups.

Survey results are not the only challenge to scholars who imagine N Irish as a possible shared identity located at the centre of the north’s political divide. Some conceptualizations of the Northern Irish identity take a different view; they seem to negate its potential as a meaningful cross-community allegiance. Academics Jonathan Tonge and Raul Gomez, for instance, offer a version of Northern Irishness that is far from shared or inclusive.[5] They conceive of N Irish as an identity that could help secure nationalist acceptance of the north’s place in the UK. In this sense, Tonge and Gomez emulate the Belfast Telegraph’s dramatic conclusion about Northern Irishness. The authors’ scholarly account is subtler and more nuanced—that is to say, more obscure and evasive—than is the media portrayal, but its constitutional implications are the same. By moving nationalists to recognize the north as a “country”—just as unionists and loyalists do—the development of Northern Irishness could erase constitutional division, stabilize politics and consolidate the north as a political entity (Tonge & Gomez, 2015). This conceptualization is heavily biased in favour of unionism. In effect, it sees Northern Irishness less as a shared or common identity than as a means of enticing nationalists to embrace the core element of a unionist identity: acceptance of the union.

Figures 1 and 2 demonstrate the form and degree of partiality in Tonge and Gomez’s theorization of the N Irish identity. Figure 1 shows a conventional understanding of unionism and nationalism as different combinations of cultural and political elements. The cultural dimension on the vertical axis ranges from British to Irish; the political dimension on the horizontal axis goes from Union to Unity. In this two-dimensional space, unionists are located at the “British” pole of the cultural dimension and the “Union” pole of the political dimension. Nationalists, their polar opposite, are located at the “Irish” and “Unity” ends of those dimensions.

Figure 2 adds N Irish to the mix, which in Tonge and Gomez’s formulation is located somewhere between British and Irish. Let’s assume for the sake of argument that as a shared or cross-community cultural identity, Northern Irishness draws equally from both cultural traditions and is therefore located exactly halfway between Britishness and Irishness.[6] But N Irish is not a shared or cross-community allegiance on the political dimension; according to the authors, it’s anchored firmly at the Union pole. The length of the journey from one identity to the next indicates the extent of bias in Tonge and Gomez’s view of Northern Irishness. To become N Irish, unionists must make one move: they traverse half the distance along the vertical cultural axis, from the British end to the middle. The nationalist journey to Northern Irishness is, in contrast, much longer and requires two moves. Nationalists, like unionists, travel half the distance along the cultural axis, but from the Irish end to the middle. In addition, nationalists move the entire distance along the horizontal political axis, from the Unity to the Union end. Unionists make no equivalent journey.


In summary, for unionists, adopting Northern Irishness is a change in the cultural dimension of their identity but an affirmation of their identity’s political dimension: the union remains. For nationalists, adopting Northern Irishness is likewise a change in the cultural dimension of their identity but a denial of its political dimension: Irish unity vanishes. In continuing to refer to the potential of Northern Irishness as a shared identity, Tonge and Gomez miss this striking asymmetry in their account. They
obfuscate two vastly different methods of reducing intergroup conflict: “one based on a shared compromise between the aspirations of the two traditions versus a reduction in division due to an acceptance of the dominant view” (Hayes, McAllister & Dowds, 2007, p. 476). Their conceptualization of Northern Irishness is not a shared accommodation occupying a middle ground between opposing traditions, but a nationalist affirmation of unionist political orthodoxy. As the admittedly schematic portrayal in Figure 2 makes clear, the Tonge-Gomez rendition of N Irish represents a lopsided middle, tilted heavily in favour of unionism.[7]

The Alliance Party’s placement on the middle ground is also problematic. Alliance began life as an avowedly liberal party attempting to attract voters from across the community divide, although it supported the north’s constitutional position as an integral part of the UK. It expressly positioned itself at the pluralist political centre, in contrast to parties defined by unionism and nationalism (Alliance Party, 1970; Farry & Neeson, 1999). Many commentators accepted Alliance’s claim of being a crucial part of the moderate middle ground in northern politics, even as it was an expressly unionist party. Alliance has since shifted from this explicit unionist position to pronounce itself neutral on the constitutional question. And commentators readily endorse the new basis for Alliance’s claim to the middle ground. Tonge believes that Alliance’s move to a position of neutrality or agnosticism consolidates its position as a centrist party (Tonge, 2020). Murphy, as mentioned above, views Alliance as part of the middle ground precisely because of its constitutional neutrality, which she says is rooted in the Good Friday Agreement’s principle of majority consent to constitutional change (Murphy, 2023).

Three clarifications are needed to appreciate Alliance’s complex and contradictory relation to constitutional change. First, neutrality based on the consent principle is not a policy position on the constitution. All the other main political parties in the north accept consent, at least notionally, but go further to articulate explicit positions on the constitutional question, which—it’s useful to recall—is whether the north should remain part of the UK or form part of a united Ireland. The DUP and UUP want to maintain the union, Sinn Féin and the SDLP to bring about Irish unity. Alliance’s consent-induced agnosticism is fundamentally the avoidance of making a constitutional decision (Bradley, 2024). Party leader Naomi Long actually acknowledges that neutrality is not a position on the binary constitutional question, but promises that Alliance will develop a position as the “facts and evidence” about all the implications of Irish unity become clear (News Letter, 2022). As this condition is unlikely ever to be met to everyone’s satisfaction, Alliance’s claim of constitutional agnosticism is likely to be with us for the foreseeable future and beyond.

The second and third clarifications address Alliance’s practical and policy alignment as a unionist party, which belies its pretense to neutrality and agnosticism. Alliance’s current refusal to take a position on constitutional change renders change highly unlikely or impossible (Burke, 2020b). The GFA requires the Secretary of State to trigger a border poll if it appears likely to him that a majority of voters would support Irish unity.[8] The more support that Alliance gets in elections or opinion polls, the less likely is the Secretary to believe that there exists a majority for unity (Mac Ginty, 2023). The Secretary will interpret every expression of support for Alliance as condoning the party’s stance of constitutional neutrality, and count every such vote as a vote against a united Ireland. In these circumstances, barring the complete collapse of Alliance’s support or the electoral decimation of unionism, the Secretary is unlikely ever to trigger a border poll. And no border poll means the maintenance of the union. Alliance’s very lack of a formal position on constitutional change is in effect a crucial prop to the union. It provides a convenient rationale for the Secretary of State’s structural inclination to keep Irish unity off the political agenda. A mutual vested interest in the maintenance of the constitutional status quo is also at work here. Alliance—like the Secretary of State—does not wish to address the question of Irish unity. Being forced into making an explicit decision on the border is something the party obviously wishes to avoid as it would cause unrest and perhaps rifts in the party’s diverse constituency of supporters.

The final clarification concerns a conspicuous discrepancy in Alliance’s constitutional posture that, remarkably, no one seems to have examined. The party’s repeated calls for the development of a fully federal UK as part of a durable devolution settlement sit uneasily beside its claim of constitutional neutrality (Alliance Party, 2015, 2016, 2017a, 2017b & 2019). Alliance’s promotion of a federal United Kingdom reinforces British sovereignty in the north just as any other party’s advocacy of a federal united Ireland would underpin Irish sovereignty. Alliance has long maintained it wants the north to work, but it persists in wanting the north to work inside the union. All the party’s policy proposals and administrative plans assume only one constitutional future, an assumption that is inconsistent with the party’s and leader’s professed constitutional neutrality. Nor is that assumption consistent with the scholarly placement of Alliance on the “middle ground”.

Misinterpreting one end of the constitutional spectrum as the middle, partiality as neutrality, lopsidedness as balance, and asymmetry as symmetry is not unusual in discussions of the middle ground in northern politics. Brian Graham and Peter Shirlow, writing in 1998 during the talks that eventually led to the GFA, see the Catholic middle class in the north as one key element in the development of a middle ground. Why? Because that class, especially in Belfast, shows a heightened preference for retaining the union with Britain. Another main constituency in the growth of the authors’ middle ground is working-class Protestantism as represented by the Progressive Unionist Party. The PUP, the authors note, is increasingly pluralist and accepting of cultural Irishness, even though it remains highly committed to British sovereignty over the north. In this analysis, both elements of the developing middle ground—middle-class Catholics and working-class Protestants—deconstruct in their own way the two-communities model based on the stereotypes of “Catholic/Nationalist/Republican” and “Protestant/Unionist/British” (Graham & Shirlow, 1998, p. 245).

The authors fail to grasp the gaping imbalance and inequality in the accommodations that make their middle ground. For them, the loyalist tolerance of expressions of Irish culture equates to the nationalist abandonment of the political project of Irish reunification. This equivalence is false: the loyalist concession on culture pales in comparison to the nationalist capitulation on partition. Relatedly and more importantly, the authors employ unionist criteria to define the parameters of the middle ground: what unites loyalist workers in the PUP with relatively affluent Catholics in Belfast is their mutual acceptance of the north’s place within the UK. In this regard, Graham and Shirlow’s middle-ground groups fatally deconstruct the nationalist half of the two-communities dynamic as they reinforce its unionist half. The middle ground remains firmly entrenched on the sovereign territory of unionism.[9] It cannot be anything but lopsided.

The Northern Irish, the Alliance Party, well-off Catholics and loyalist workers are part of a middle so askew that perhaps it’s time to dispense with the use of the term “middle ground” altogether, or to specify explicitly and precisely why “middle” is an appropriate label for whichever group is being studied. In the next section, I identify circumstances in which middle-ground groups can legitimately be called “middle”. I’ll not examine in any detail Graham and Shirlow’s middle-ground groups. Middle-class Catholics are not now as much discussed as part of the middle ground as they were in the years surrounding the GFA; and the PUP, never big, is electorally insignificant, with no elected Assembly member since 2007 and only one local councillor. Instead, I’ll focus on the four middle-ground groups that feature in many contemporary discussions of northern politics: Alliance voters, neithers, the N Irish, and people with no religion.

Despite the obvious unionist bias in many specifications of the middle ground, examining the size, social profile and constitutional preference of middle-ground groups illuminates changes in the politics of the north, affords an opportunity for conceptual clarity and provides insights into the north’s constitutional future.

Data Analysis

I draw all the data for this analysis from the Northern Ireland Life and Times (NILT) surveys. These surveys, conducted annually since 1998 (except in 2011) under the auspices of Queen’s University Belfast, aim to yield representative samples of the north’s adult population (NILT, 2022a). The sample sizes range from 1800 in the survey’s first few years to 1200 between 2005 and 2020. In 2021 and 2022, the NILT surveyed some 1400 respondents. Margins of error are consistently within ±2.8 percent at the 95 percent confidence level (or 19 times out of 20). They fall to ±2.6 percent in 2021 and 2022.


Figure 3 shows the four questions in the 2022 survey that I use to identify the middle ground. This survey is the latest available but the NILT uses the same or equivalent questions in all other survey years.[10] The figure highlights how the questions can be interpreted as measuring each middle-ground group relative to its two-community counterparts:

1. Political Party Support: “Alliance Party supporters” versus those voting for the main unionist parties (DUP and UUP) or the main nationalist parties (Sinn Féin and the SDLP);

2. Political Identity: those respondents thinking of themselves as “neither” as opposed to unionist or nationalist;

3. National Identity: those describing themselves as “Northern Irish” rather than as British or Irish;[11] and

4. Religion: those saying they have “no religion” in contrast to those saying Protestant or Catholic.

The size of the middle ground

Figure 4 gives the 2022 distribution of responses to the four questions on party support, political identity, national identity and religion. The legend at the top of the figure emphasizes that the specifications of the middle-ground and comparison groups vary by question, as just discussed. The grey bars indicate the percentage size of middle-ground groups defined as Alliance/Neither/N Irish/No religion on the four variables. The orange bars show the size of the comparison groups correspondingly defined as Unionist/Unionist/British/Protestant. And green bars give the size of the comparison groups defined as Nationalist/Nationalist/Irish/Catholic.


Figure 4 underlines that the middle ground is a major player in the politics of the north. On the political identity variable, for instance, the middle-ground group is the largest group of all, with 38.3 percent of respondents saying they think of themselves as “neither unionist nor nationalist”. Some 30.7 percent identify as unionist and 26.3 percent as nationalist. This result is not unusual. The number of neithers has outstripped the number of unionists and nationalists in every NILT survey since 2006. Alliance voters, the middle-ground group on the party support variable, are also the largest group (26.7%) by a small margin over nationalist party supporters (25.6%) and unionist party supporters (23.8%). The other two variables confirm the significant size of middle-ground groups. On national identity, 30.6 percent of respondents say they think of themselves as N Irish as opposed to 31.9 percent saying Irish and 26 percent British. The no religion group is also of considerable size, with 29.2 percent of the sample having no religion compared to 37.8 percent who are Protestant and 33 percent Catholic.

The distributions in Figure 4 need to be interpreted carefully. The 2022 survey heavily overestimates the number of Alliance voters. It also continues an inaccurate pattern in many NILT surveys by underestimating the number of DUP and Sinn Féin supporters. These characteristics of NILT samples will tend to inflate the size of middle-ground groups. The 2021 census has two measures, religion and national identity, that we can compare directly to the 2022 NILT survey results. The census shows the no religion group at just 17.4 percent of the population compared to the NILT estimate of 29.2 percent (NISRA, 2022a). Some of this discrepancy, but not all of it, will be explained by measurement differences and survey sampling error. The census estimate of “Northern Irish only” (19.8%) is also considerably lower than the NILT figure for “Northern Irish” (30.6%). The census estimate rises to 31.5 percent if we add the number of people saying “Northern Irish only” to the number saying they had a Northern Irish identity along with other national identities—"Northern Irish and Irish” or “Northern Irish and British,” for instance (NISRA, 2022b). This larger census number is equivalent to the NILT’s 2022 estimate.[12]

The very size of the middle-ground groups, even using the lower census estimates, challenges the two-community model of northern politics around which much of the GFA’s institutional architecture has been constructed. It also partly explains the increasing references to three political blocs in the north, all of them minorities, which revise the conventional binary dynamic of majority-versus-minority (Doyle, 2022; Nolan, 2013 & 2021). Some of these discussions underestimate the resilience, political vigour and continuing institutional relevance of the two traditional communities, but middle-ground groups do add complexity to the northern polity (Ruane & Todd, 1992). The NILT data, with its measurement particularities and sampling imperfections, underline equality among the blocs. As the bottom set of bars in Figure 4 indicates, the average size of the three blocs, computed across the four measures listed, is in the ratio of 30:30:30.

Figure 5 demonstrates how the size of the middle ground varies over time. The figure introduces a new colour scheme that I use when comparing the four middle-ground groups to one another. I use a yellow line for Alliance voters, grey for neithers, red for N Irish and blue for the no religion group. The figure shows that the middle ground has long been a part of northern politics, but its social weight has increased, slowly and a little unevenly, since 1998. The average size of the four middle-ground groups rises from 18 percent of respondents in 1998 to over 31 percent in 2022. All four groups have grown in the period, with the steepest rates of growth in the no religion group and among Alliance supporters. For these two groups, a prolonged period of stagnancy and slow growth preceded a burst of relatively rapid growth since 2018. As examined above, Alliance’s spectacular rise in the 2019 elections, coinciding with a lengthy period of government breakdown, is an especially consequential change.

The census figures broadly confirm the trends for no religion and N Irish in Figure 5. They show that the no religion group has increased over time, with marked growth between the 2011 and 2021 censuses (Coakley, 2021; NISRA, 2022a). And they indicate that the numbers of N Irish stagnated in the last decade, as the NILT data also generally show (NISRA 2002b).



Social Composition

Knowing who the middle-ground groups are will help us understand their political attitudes and expected constitutional behaviour. What are their social background characteristics? How does their social makeup differ from the northern electorate as a whole?

In the section on the lopsided middle, I separately examined the N Irish and the Alliance Party. In this section, I’ll look at a different component of the middle ground, the neithers, people who say that they
think of themselves as neither unionist nor nationalist. As I’ll discuss, the social background characteristics of the neithers are generally similar to those of other middle-ground groups, and neithers can serve as a proxy for the rest.

Figure 6 profiles the neithers in respect of the three other middle-ground groups: Alliance supporters, N Irish and those with no religion. It also displays the profile of neithers on the social background variables of sex, age and occupation. Finally, the figure allows us to compare the social profile of neithers (the grey bars) to that of the sample as a whole (the black bars) to see how closely the two align.

Comparing the lengths of the grey and black bars is an easy way of determining group patterns of overrepresentation, underrepresentation and proportional representation. The first grey bar in Figure 6 shows that 44.3 percent of neithers are Alliance voters. The corresponding (and much shorter) black bar shows that only 26.7 percent of all respondents support Alliance. There are many more Alliance supporters in the ranks of the neithers than there are in the northern electorate as a whole. In other words, Alliance supporters are highly overrepresented among neithers. The same pattern of overrepresentation holds for the other middle-ground groups: the grey bar is much longer than is the black bar for both the N Irish and those with no religion. Those other middle-ground groups are also significantly overrepresented among neithers.


The pattern on the social background variables in Figure 6—sex, age and occupation—differs from the pattern we found for the middle-ground groups. Women are also overrepresented among the neithers but not to the same extent as are the middle-ground groups. The young and the old, and workers and management, are proportionally represented among neithers, with the grey and black bars of almost equal length.[13]

The bar chart for neithers displayed in Figure 6 closely resembles the bar charts for Alliance supporters, the N Irish and those with no religion, which I won’t reproduce here for reasons of space.[14] Two general similarities are noteworthy. First, no matter which particular middle-ground group we examine, the other middle-ground groups are strongly overrepresented in its ranks: the grey bars are much longer than the corresponding black bars, just as in Figure 6. That there is a robust relationship among the various middle-ground groups is not surprising. The Alliance Party, for instance, has long positioned itself as a voice for neithers and promoted a Northern Irish identity, and these groups along with people who have no religion are key elements in the party’s electoral successes (Farry & Neeson, 1999; Raymond, 2019; Tonge, 2020; Murphy & Evershed, 2022). The second similarity is that Alliance supporters, the N Irish and the no religion group show—again much like the neithers in Figure 6—a more or less proportional representation of women, youths and seniors, and the working and managerial classes. For these middle-ground groups, the corresponding grey and black bars are of equivalent lengths on the variables of sex, age and occupation.

There are some exceptions to this general profile. The N Irish stand out as distinct: of all the middle-ground groups, it is by far the most unionist and Protestant and the least nationalist and Catholic. One-third of the Northern Irish see themselves as unionist, for instance, compared to less than one-fifth of Alliance supporters and the no religion group. Some 44 percent of those with an N Irish identity are Protestants; the number of Protestants among Alliance voters and neithers is much lower, at about 29 percent. In the next section on attitudes towards the constitutional future of the north, we’ll see the effect of this distinctive profile of the N Irish. Regarding social background characteristics, Alliance is the most class-based group, with workers significantly underrepresented and managers strongly overrepresented among the party’s supporters. Those with no religion have the lowest representation of women, with men making up 58 percent of the group.

Constitutional Preference

Do middle-ground groups secure the north’s place in the UK, as some in the media and academy hope? This section examines the middle ground’s beliefs about the constitutional status of the north.


Figure 7 charts for 2022 the extent to which the middle ground and its comparison groups support union with Britain. The legend at the top of the figure reiterates that group definitions vary by question type. Figure 7 uses the NILT’s measure of long-term constitutional preference, which asks respondents: “Do you think the long-term policy for Northern Ireland should be for it to remain part of the United Kingdom, with direct rule; remain part of the United Kingdom, with devolved government; reunify with the rest of Ireland; be an independent state” (NILT, 2022b, p. 56). In the figure, I sum the responses of “remain with direct rule” and “remain with devolved government” to give an overall total of “support for union”. In calculating percentages, I use the full range of responses—including “don’t know”—although I show only the pro-union response in the figure.

On the political party variable, the orange bar indicates that fully 93.4 percent of unionist party supporters are in favour of the north remaining in the UK, the grey bar that 43.9 percent of Alliance voters support the union, and the green bar that just 14.1 percent of nationalist party voters prefer the union. The same pattern holds for the other variables of political identity, national identity and religion: high support for union among the Unionist/British/Protestant group, intermediate support among the various middle-ground groups, and low support in the Nationalist/Irish/Catholic camp. That is, Figure 7 underlines why in this instance the middle ground can legitimately be called “middle ground”. All four middle-ground groups are not as heavily in support of the union as are their orange counterparts nor as meagerly supportive as are their green comparison groups. They sit in the middle. Figure 8 gives the corresponding data for support for Irish unity, with equivalent results. Middle ground support for reunification stands in between the high support on the nationalist side and the low support on the unionist side. In sum, the middle ground is not distinguished from its orange and green comparison groups by having a different set of constitutional beliefs based on a wilful agnosticism or determined neutrality; rather, it’s distinguished by having a different distribution of the same set of beliefs. It’s levels of support of union and unity are middling.



Figure 9 clarifies the constitutional preferences of middle-ground groups and suggests why so many commentators claim the middle ground for the union. In each middle-ground group—Alliance voters, neithers, N Irish, and those with no religion—more respondents support union (orange bar) than support unity (green bar). Across all four groups, the average preference for union is 47.5 percent, for unity just 26.8 percent. For the N Irish, the gap in support for the two constitutional options is especially favourable to the union, with almost 60 percent supporting the constitutional status quo and merely 15 percent preferring a united Ireland. This kind of skewed distribution is to be expected given the distinctly unionist and Protestant social profile of the N Irish. Before I examine in more detail this pro-union snapshot of middle ground constitutional opinion, I want to look at a second notable finding evident in the figure.

Figure 9 also confirms that the middle ground is not particularly agnostic or neutral on the constitutional question. The number of “don’t know” responses among middle-ground groups, indicated by the grey bars, is very similar to the number of ‘don’t knows” among all respondents, shown at the bottom of the figure.[15] In other words, the middle ground makes up its mind about the constitution in roughly the same proportion as does the northern electorate as a whole. Middle-ground groups doubtlessly place less emphasis on constitutional issues than do their orange and green counterparts, and some individuals escape to the middle ground precisely to avoid the constitutional question altogether. But when asked to make a choice, middle-ground groups are prepared to commit to a constitutional position.


Alliance Party supporters have the fewest agnostics of all the middle-ground groups, although most of the differences are within the survey’s margin of error and cannot be regarded as statistically significant. But Alliance voters are consistent over time. In every NILT survey since 2007, they have the fewest “don’t know” responses compared to other middle-ground groups, and the differences are sometimes substantial. Alliance supporters seem to be ahead of the party and its current leadership in that they have a position on the constitution and are willing to express it well in advance of the presentation of facts and evidence on the possible ramifications of a united Ireland.

Let’s return now to the discussion of the middle ground’s relative preference for union over unity. This pro-union disposition probably gives some comfort to unionists and their allies in the media and academy. But recent trends in constitutional preference may cause them a measure of discomfort. The snapshot in Figure 9—a representation of constitutional preferences at a single point in time (2022)—says nothing of how opinions may be changing. Figure 10 depicts the middle ground’s support for union for the years 1998 to 2022, which allows us to see if levels of support vary over time. It resurrects the colour scheme used in Figure 5 to differentiate the four middle-ground groups and it adds a black line indicating support for union among all respondents.



Figure 11 shows the corresponding data on support for Irish unity; but note that it uses a different scale on the vertical (y) axis to give a sharper view of the yearly data points. I’ll pay particular attention to comparing opinions before and after the Brexit interruption (circa 2016), which appears as a watershed moment when constitutional trends begin observably to change (Burke, 2023). Generally speaking, the coloured lines of the middle-ground groups track the ups and downs of the black line of the electorate as a whole. Among both the middle ground and all respondents, the two figures chart steep declines in support for union and dramatic increases in support for unity beginning in the years 2015 to 2017.[16]

Focusing on the middle, Alliance voters’ current support for union is at its lowest ebb since 2003 and their support for unity is at its highest level ever in the NILT yearly series. In recent years, Alliance has cooperated with nationalist parties in opposition to unionism, most significantly on Brexit and relations with the EU, but also on such other issues as the flying of flags, same sex marriage and the status of the Irish language. This institutional alignment has not extended to the constitutional issue (Murphy, 2023; Ó Dochartaigh, 2022 ). Perhaps the new attitudes of Alliance voters evident in the figure are a portent of change. The trajectory of the no religion group is similar to that of Alliance supporters: the last few years witness historic lows in support of union and historic highs in support of unity. The neithers’ recent pro-union levels are the lowest since 2001 and their pro-unity levels the highest since 2006. The N Irish are a little anomalous. Compared to other middle-ground groups, the N Irish show the highest levels of support for union and the lowest levels of support for unity, especially in the past few years when their social profile has become increasingly unionist and Protestant. Even so, the extent of their support for union is lower in recent years than it was prior to 2017, and their support for unity higher. Overall, then, pro-union activists should be concerned about the middle ground’s recent direction of constitutional travel; and advocates of Irish unity might be encouraged but should not underestimate how much of the trip remains.


Figure 12, the final one, gives an alternative view of some of the data presented in the previous two figures. It is alternative in two senses: compared to Figures 10 and 11, it uses a different measure of constitutional preference to demonstrate the findings are robust, in that the patterns in the data remain stable across various measures. And it presents the yearly data in a different, condensed format to highlight how much change has occurred in a few short years.

Figure 12 charts the middle ground’s net preference for union using the “vote tomorrow” question asking respondents how they would vote on the north’s constitutional future if there were a referendum tomorrow. It compares constitutional attitudes in only two years, 2017 and 2022.[17] “Net preference for union” equals the percentage of a given group supporting union minus the percentage of that group supporting unity. Any positive difference indicates a group’s net preference for union, with more members of the group supporting union than unity. A negative difference indicates a net preference for unity. All the results in the figure are positive, showing a net preference for union in each middle-ground group. Let’s take Alliance supporters to illustrate the calculations. In 2017, 61.8 percent of Alliance voters support union on the “vote tomorrow” question and 15.4 percent support unity, yielding a difference (or net preference for union) of 46.4 percentage points, which is shown on the dark orange bar. The light orange bars shows the corresponding 2022 net preference for union falls to just 11.8%, computed as 45.8% union – 34% unity. Alliance voters’ net preference for union shrinks greatly between 2017 and 2022 both because their support for union falls and their support for unity rises.


The most notable feature of Figure 12 is the vast discrepancy in the lengths of the differently shaded orange bars: the light orange bars are much shorter than the dark ones. For all middle-ground groups but the N Irish, the net preference for union decreases markedly from 2017 to 2022. It drops especially rapidly among Alliance voters and the no religion group. These results are consistent with what we found previously using the NILT’s question on long-term constitutional preference, which showed the middle ground’s decreasing support for union and increasing support for unity. Union’s once large referendum lead over unity is becoming increasingly tenuous. It’s still a lead but it’s getting smaller and smaller.

Conclusion

The middle ground is malleable. Group opinions on the constitutional question are in a period of rapid and patterned change. There is everything to play for should a referendum be called. Which way middle-ground groups swing will help to determine the outcome of any border poll. If the trends I’ve documented continue, the outlook is promising for advocates of a united Ireland. The recent finding that—for the first time ever—more Alliance Party members favour Irish unity than support continued union is consistent with these trends (Breen, 2024a). But no one can know if such trends will persist. As the constitutional debate progresses, we’ll see at least two important developments that may reinforce or reverse current opinions. Facts and evidence about the implications of constitutional change will accumulate, with no doubt some reasoned, forceful and legitimate debate about the nature of the facts and evidence and their purported consequences. But that’s not all that will accumulate: so will partisan distortions, scaremongering misrepresentations, outlandish claims, and outright lies. The Working Group on Unification Referendums on the Island of Ireland frames the question as “how to ensure that voters have access to high-quality information and are not unduly exposed to misinformation” (Working Group, 2021, para. 14.41). I think we can expect that voters will have ready access to reliable information but will also be inundated with misinformation. The important question then becomes which will prevail, cogent argument or calculated deception.

The constitutional option of Irish unity has benefitted from a “Brexit bounce,” an upsurge in its popularity because of the complications and controversies of the UK leaving the European Union (Burke, 2023). Middle-ground groups, with the partial exception of the N Irish, are much more likely to say that Brexit has made them feel more rather than less in favour of a united Ireland.[18] And Figures 11 and 12 show concrete evidence that middle ground support for unity has increased since Brexit. This pro-unity bounce might moderate if the disruptive effects of Brexit abate and the attendant disputes subside. We’ll see. Regular opportunities for Brexit-related tumult will arise from pervasive trade issues, the Stormont Brake, the Assembly consent mechanism, continuing British government unpreparedness for leaving, and the Labour Party’s plan to renegotiate some of the terms of withdrawal from the EU. But Brexit’s intrusion on public debate in the north will ebb and flow. In the face of such fluctuations, advocates of Irish unity would be well-advised to refine their appeals to ensure that the middle ground continues on a trajectory favourable to reunification.

The middle is not the only key constituency in any coming border poll. Also important for advocates of unity will be firming up support among soft republicans and softer nationalists, appealing to that section of unionists who might waver in their current support for the constitutional status quo, engaging newly-eligible voters, and ensuring that all the disparate elements of their constituency turn out to vote on referendum day. A major problem is that the legal power to call a border poll rests solely in the hands of the Secretary of State, and his authority in this matter is as arbitrary as it is unaccountable. The “coming” border poll might never arrive.

Notes

[1] Brendan O’Leary and others proposed a system of “shared authority”, and explored a number of different models of co-sovereignty, including those of the New Ireland Forum, the Kilbrandon Committee, the SDLP and academics Basil Chubb, Bernard Crick and Frank Wright (O’Leary et al., 1993). O’Leary soon moved away from shared authority towards a “double protection” model, a scheme broadly consistent with the Good Friday Agreement (McGarry and O’Leary, 1995).

[2] The NILT surveys asked questions about joint authority only once, in 2018, more than one year into a three-year breakdown in devolved governance occasioned by the resignation of Martin McGuinness as Deputy First Minister in January 2017. The 2018 survey shows that 17 percent of respondents would find joint authority “almost impossible to accept”, 29 percent “would not like it, but could live with it if you had to”, 35 percent “would happily accept” it, and 20 percent say “don’t know”.

[3] Belfast Telegraph columnists continue to push the narrative that the growth of the Northern Irish identity means the end of Irish unity. In June 2023, John Laverty noted that the middle-ground supporters of Alliance eschew the constitutional extremes and “feel more at home in the middle of a road that appears to be leading somewhere; i.e., a Northern Ireland full of Northern Irish people content for things to stay just as they are” (Laverty, 2023).

[4] The following works typify the range of research on the “Northern Irish” identity: Moxon-Browne (1991); Hayes and McAllister (2009); McGlynn, Tonge and McAuley (2014); McKeown (2014); and Garry and McNicholl (2015).

[5] Tonge and Gomez actually illustrate both the empirical and theoretical limits of N Irish identity. Like others, they find that survey data do not support the characterization of Northern Irishness as a compromise, shared identity. But they continue to view N Irish as potentially a shared identity that, if developed, could be a key stabilizer of northern politics. My argument is that their very conceptualization of Northern Irishness actually rules out the potential of meaningful sharing. To the extent that they tie N Irish to maintenance of the union, they render the identity much closer to unionism than to nationalism.

Even the temperate academic renditions of the N Irish identity tend toward the unionist bias in Tonge and Gomez. These moderate voices explicitly point out how the Protestant version of the identity reinforces the constitutional status quo, but they remain silent on whether the Catholic version challenges current constitutional arrangements. That is, the Catholic territorial outlook on Northern Irishness is probably compatible with maintenance of the union, but the Protestant political interpretation is definitely incompatible with Irish unity. This moderate interpretation of the N Irish identity seems tilted in favour of unionism.

[6] I’ll leave aside the difficult question of what kind of cultural symbols of Northern Irishness might be considered part of a shared or cross-community allegiance. Recent research suggests that the level of hostility to or acceptance of cultural symbols like the Red Hand of Ulster depends on how they are presented to the public (Garry, O’Leary & Pow, 2024).

[7] Garry, O’Leary and Pow (2023) measure and conceptualize N Irish differently from Tonge and Gomez; they see the identity as possibly “a halfway house or bridging identity for Protestant unionists”. My analysis of NILT data shows little support for their view.

[8] I’ve argued extensively that the Secretary of State’s so-called mandatory duty to call a border poll is meaningless (Burke, 2020a). Here I assume, for the sole purpose of explicating the logical consequences of Alliance’s claimed neutrality, that the mandatory duty is a real obligation that would be implemented in the circumstances.

[9] The analyses of Graham and Shirlow and Tonge and Gomez are similar to earlier unionist arguments advocating a “non-political Irishness.” Tonge and Gomez even examine a form of non-political Irishness without using that term. These earlier arguments were not tied to the notion of the middle ground but like that notion suggested that nationalists must give up their aim of unity in order for the north to stabilize, normalize and prosper. The very act of abandoning reunification will ease unionism’s hostility to this newly-innocuous form of cultural nationalism (Haslett, 1995; Kennedy, 1995; Roche, 1995, Aughey, 1996). The asymmetry in non-political Irishness is obvious, not only because that kind of identity demands much of nationalists and next to nothing of unionists, but also because its advocates never even consider the equivalent notion of a “non-political Britishness.”

Cathal McCall’s discussion of how the GFA transformed identities in the north is similarly tendentious and asymmetrical. He contends that the Agreement’s North/South Ministerial Council (especially) and British-Irish Council (perhaps) helped to establish a model of transterritorial governance for the north. In this model, transterritorialism undermines conventional concepts of self-determination and indivisible territorial sovereignty, and points to a settled future of co-determination, co-sovereignty and pooling of sovereignty. The GFA’s transterritorial institutions in turn may lead to new forms of communal and political identities in the north. McCall sees republicanism’s acceptance of the legitimacy of the north and unionism’s acceptance of power-sharing as incipient forms of identity realignment that could potentially undermine old identity forms based on exclusion and nation-state territorialism (McCall 2001).

There is much wrong with McCall’s analysis. He greatly overestimates the extent to which globalization has weakened the capacity of national states. He exaggerates the importance and impact of the NSMC and BIC. The GFA’s Strand One institutions (Executive and Assembly) are the key institutions of the Agreement, not the weak and relatively insignificant institutions of Strand Two (NSMC) and Strand Three (BIC). McCall himself (2018) later acknowledged the limited success of the NSMC. Transterritorialism is largely a myth, as increasingly frequent unilateral applications of British sovereignty in the north repeatedly demonstrate. Britain’s resiling from the Stormont House Agreement, its Internal Market Bill, Protocol Bill, Legacy Bill, imposition of budget cuts, and Safeguarding the Union command paper—to name but a few recent examples of unilateralism—discredit the claim of movement towards a developing co-sovereignty in the north. Most importantly, for the purpose of this paper, is McCall’s identification of the nature of identity realignment. Despite all his talk of transterritorialism, McCall’s emphasis on Sinn Féin’s acceptance of partition leaves intact British sovereignty in the north. This massive Sinn Féin concession is by no means balanced by unionism’s halting, contradictory and limited acceptance of power-sharing in the north.

Coulter and others’ examine groups who disrupt the two-communities model of northern politics that pits Protestant/unionist/British against Catholic/nationalist/Irish (Coulter et al., 2021, chap. 5). Their analysis likewise reinforces the north’s place in the UK. Their disrupting groups—including the N Irish, those with no religion and those who identify as neither unionist nor nationalist—are distinguished by their preference for union over unity. In the authors’ view, disrupting the two-communities binary apparently involves adopting the constitutional preference of one community.

The common element in these disparate discussions of non-political Irishness, transterritorial identity realignment, and binary disruption is that union not unity is the constitutional future of the north, a settlement which heavily privileges unionism over nationalism.

[10] The same questions are used in every survey except for the political party question. The NILT uses three slightly different but equivalent sets of questions to measure political party support. The different questions give rise to different variable names for political party support: NIPARTY from 1998 to 2006, POLPARTY in 2007 and 2008, and POLPART2 since 2009.

[11] The “Ulster” identity never polled more than 5 or 6 percent in NILT surveys. By 2022, the number of respondents identifying as Ulster is only 1.6 percent, which is typical of NILT results since 2014. As John Coakley notes, the Ulster identity shrinks “to the point of disappearance”, mostly because of movement into the Northern Irish category in the mid- and late-1980s (Coakley, 2021, p. 35).

[12] If we compare the same years—the 2021 NILT survey to the 2021 census—the same conclusions hold. The NILT estimates for neithers and the no religion group are virtually identical in 2021 and 2022.

[13] The NILT uses the National Statistics Socio Economic Classification (NSSEC) as a measure of occupation. Figure 6 shows the distributions of the lowest occupation, coded as “Semi-routine and routine occupations”, and the highest, coded as “Managerial, administrative and professional occupations.” NILT analysts derive the variable NSSEC1 from various questions about the work situation of respondents (NILT, 2022b, p. 70).

[14] The other figures are available on request.

[15] Especially in the last three years, the middle ground’s “don’t know” numbers have converged on those for all respondents. Neithers are an exception, just as in Figure 9.

[16] Middle-class Catholics are on a similar trajectory. In the text of the paper I argue that Graham and Shirlow’s middle-class Catholics cannot conceptually be considered part of the middle ground: the authors place them not in the middle but at the union end of the constitutional continuum. The empirical rationale for this conceptual confusion has collapsed. The proportion of middle-class Catholics favouring union as a long-term preference declines from 47.8 percent in 2017 to 16.7 percent in 2022. The 2022 number is markedly lower than any of the pro-union percentages reported for the class in the authors’ 1996 study. NILT data also reveal that the proportion of middle-class Catholics favouring unity rises from 39.1 percent to 62.3 percent in the years from 2017 to 2022.

[17] The NILT uses two slightly different questions. In 2017, the first time the question is asked since Brexit, the wording is: “If there was a referendum tomorrow about whether Northern Ireland should leave the UK and unite with the Republic of Ireland, how do you think you would vote? I would vote for Northern Ireland to remain part of the UK; I would vote for Northern Ireland to unite with the Republic of Ireland ...” BORDPOLL is the name of this variable in the NILT data set (NILT, 2017, p. M 27). In 2022, the wording of the question is: “Suppose there was a referendum tomorrow on the future of Northern Ireland and you were being asked to vote on whether Northern Ireland should unify with the Republic of Ireland. Would you vote ‘yes’ to unify with the Republic or ‘no’? Yes, should unify with the Republic; No, should not unify with the Republic ...” (NILT 2022b, p. 56). REFUNIFY is the variable name.

[18] In 2022, for instance, 52.3 percent of Alliance voters say that Brexit has made them feel more in favour of a united Ireland and only 4.9 percent say less in favour. The corresponding percentages for the other middle-ground groups are: neithers, 43 percent and 6.8 percent; N Irish, 33 percent and 11.2 percent; and the no religion group, 44 percent and 7.9 percent.

References

Alliance Party. (1970). Declaration of Intent – A New Political Charter. 21 April. Retrieved from.

Alliance Party. (2015). Step Forward Not Back: Manifesto for 2015 Westminster Election. Retrieved from.

Alliance Party. (2016). Forward Faster: [Assembly] Manifesto 2016. Retrieved from. 

Alliance Party. (2017a). How to Change Northern Ireland for Good: Manifesto 2017. Retrieved from.  

Alliance Party. (2017b). Change Direction: Westminster Manifesto 2017. Retrieved from. 

Alliance Party. (2019). Demand Better: Westminster Manifesto 2019. Retrieved from.

Aughey, A. (1996). “Obstacles to Reconciliation in the South.” In Building Trust in Ireland. Studies Commissioned by the Forum for Peace and Reconciliation, 2-51. Belfast: Blackstaff Press in association with the Forum for Peace and Reconciliation.

Belfast Telegraph. (2012a). “The rise of the Northern Irish.” 12 December. Retrieved from the LexisNexis electronic database of news articles. [Cited below as LexisNexis].

Belfast Telegraph. (2012b). “Foundations laid for shared future.” 12 December. Lexis/Nexis.

Belfast Telegraph. (2012c). “Now is a time for leaders.” Editorial. 12 December. Lexis/Nexis.

Bradley, D. (2024). “It’s shameful that Alliance will not come off constitutional fence.” Irish News. 5 January. Retrieved from.

Breen, S. (2023a). “The little-known group who could hold the key to the future.” Belfast Telegraph. 4 March. Retrieved from the Factiva (Dow Jones) electronic database of news articles. [Cited below as Factiva].

Breen, S. (2023b). “Sinn Fein is still the most popular party but DUP narrows gap to 3%: Poll reflects shift in unionist feeling to Stormont boycott.” Belfast Telegraph. 5 November. Retrieved from. 

Breen, S. (2024a). “More Alliance Party members would vote for Irish unity than to remain in UK: survey.” Belfast Telegraph. 3 March. Retrieved from. 

Breen, S. (2024b). “Middle ground will decide border poll... and it's now growing.” Belfast Telegraph. 4 March. Factiva.

Burke, M. (2020a). “Stealing Irish Unity: The Repertoire of Thieves,” Part 1. The Pensive Quill.

5 September. Retrieved from. 

Burke, M. (2020b). “Stealing Irish Unity: The Repertoire of Thieves,” Part 4. The Pensive Quill.

26 September. Retrieved from

Burke, M. (2023). “The Crisis of Democratic Legitimacy in the North.” The Pensive Quill. 18 July. Retrieved from. 

Clarke, L. (2012). “Link between our religion and our politics is fast weakening.” Belfast Telegraph. 12 December. Lexis/Nexis.

Coakley, J. (2021). “Is a middle force emerging in Northern Ireland?” Irish Political Studies 36:1 (March): 29-51.

Coulter, C., N. Gilmartin, K. Hayward and P. Shirlow. (2021). Northern Ireland a generation after Good Friday. Manchester: Manchester University Press.

Doyle, O. (2022). “The Irish Unification Debate After the Northern Ireland Assembly Elections.” IACL-AIDC Blog. 24 May. Retrieved from. 

Dudley Edwards, R. (2012). “Unionist Forum welcome in these desperate times.” Belfast Telegraph. 20 December. Lexis/Nexis.

Farry, S., and S. Neeson. (1999). “Beyond the band-aid approach: an Alliance party perspective upon the Belfast Agreement.” Fordham International Law Journal 22:4 (April): 1221-1249.

Garry, J. and K. McNicholl. (2015). “Understanding the ‘Northern Irish’ Identity.” Briefing Paper. Knowledge Exchange Seminar Series. Retrieved from. 

Garry, J., B. O’Leary and J. Pow. (2023). “Northern Ireland identity may serve as less adversarial for
politically moderate Catholics and Protestants.” Irish Times. 11 December. Retrieved from. 

Garry, J., B. O’Leary and J. Pow. (2024). “Can the Red Hand of Ulster be transformed into a unifying
image for the island of Ireland?” Irish Times. 22 January. Retrieved from. 

Graham, B., and P. Shirlow (1998). “An elusive agenda: the development of a middle ground in Northern Ireland.” Area 30:3 (September): 245-254.

Hanna, G. (2024). “The battle for the middle ground in a future border poll is under way... and this is what they want.” Belfast Telegraph. 13 March. Factiva.

Haslett, Edgar. (1995). “Historical Injustices.” In The Idea of the Union: Statements and Critiques in Support of the Union of Great Britain and Northern Ireland. ed. J.W. Foster, 121-127. Vancouver: Belcouver Press.

Hayes, B.C., and I. McAllister. (1995). “Religious independents in Northern Ireland: Origins, Attitudes and Significance.” Review of Religious Research 37:1 (September): 65-83.

Hayes, B.C., and I. McAllister. (2009). “Religion, identity and community relations among adults and young adults in Northern Ireland.” Journal of Youth Studies 12:4 (August): 385-403.

Hayes, B.C., I. McAllister and L. Dowds. (2007). “Integrated Education, Intergroup Relations, and Political Identities in Northern Ireland.” Social Problems 54:4 (November): 454–482.

Hayward , K. (2020). “The 2019 General Election in Northern Ireland: the Rise of the Centre Ground?” Political Quarterly 91:1 (January-March): 49-55.

Kennedy, D. (1995). “The Realism of the Union.” in The Idea of the Union, 27-36.

Laverty, J. (2023). “SF's so-called 'little helpers' have put Irish unity on back burner.” Belfast Telegraph. 1 June. Factiva.

LucidTalk. (2023). Main Report: LucidTalk NI Autumn ‘Tracker’ Poll – October/November. 3 November. Retrieved from.

McAdam, N. (2012). “DUP can entice more Catholics: Robinson.” Belfast Telegraph. 3 December. Lexis/Nexis.

Mac Ginty, R. (2023). “The Alliance Party of Northern Ireland: Are they really ‘other’?” Royal Irish Academy. 2 June: Retrieved from. 

McCall, C. (2001). “The Production of Space and the Realignment of Identity in Northern Ireland.” Regional & Federal Studies 11:2 (Summer): 1-24.

McCall, C. (2018). “The Travails of the North/South Ministerial Council.” Political Studies Association,

PSA Blog. 10 April. Retrieved from. 

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

McGlynn, C., J. Tonge and J. McAuley. (2014). “The Party Politics of Post-devolution Identity in Northern Ireland.” British Journal of Politics and International Relations 16:2 (May): 273-290.

McKeown, S. (2014). “Perceptions of a Superordinate Identity in Northern Ireland,” Peace and Conflict: Journal of Peace Psychology 20:4 (November): 505-515.

Moxon-Browne, E. (1991). “National Identity in Northern Ireland.” In Social Attitudes in Northern Ireland: The First Report. ed. P. Stringer and G. Robinson. Belfast: Blackstaff Press. Retrieved from. 

Murphy, M.C. (2023). “The Rise of the Middle Ground in Northern Ireland: What does it Mean?” Political Quarterly 94:1 (January/March): 95-103.

Murphy, M.C., and J. Evershed. (2022). A Troubled Constitutional Future: Northern Ireland after Brexit. Newcastle upon Tyne: Agenda Publishing.

News Letter. (2022). “Naomi Long: Alliance Party will take a stance on constitutional question in the future.” 6 March. Factiva

NISRA. (2011). Northern Ireland Statistics & Research Agency. 2011 Census: Individual Questionnaire Northern Ireland. Retrieved from. 

NISRA. (2012). Northern Ireland Statistics & Research Agency. Statistics Bulletin. Census 2011: Key Statistics for Northern Ireland. 11 December. Retrieved from. 

NISRA. (2022a). Northern Ireland Statistics & Research Agency. Census 2021: Main Statistics for Northern Ireland: Statistical Bulletin – Religion. 22 September. Retrieved from.

NISRA. (2022b). Northern Ireland Statistics & Research Agency. Census 2021: Main Statistics for Northern Ireland: Statistical Bulletin – National Identity. 22 September. Retrieved from. 

NILT. (2017). ARK. Northern Ireland Life and Times Survey 2017: Main Questionnaire. Retrieved from.

NILT. (2022a). ARK. Northern Ireland Life and Times Survey, 2022 [computer file]. Retrieved from. 

NILT. (2022b). ARK. Northern Ireland Life and Times Survey 2022: Questionnaire: CAWI (online). Retrieved from. 

Nolan, P. (2013). The Northern Ireland Peace Monitoring Report: Number Two. March. Belfast: Community Relations Council. Retrieved from. 

Nolan, P. (2021). “Running out of Road.” Dublin Review of Books 134 (June). Retrieved from.

Nolan, P. (2022). “Usuns, Themuns, Others.” Dublin Review of Books 148 (October). Retrieved from. 

Ó Dochartaigh, N. (2021). “Beyond the dominant party system: the transformation of party politics in Northern Ireland.” Irish Political Studies 36:1 (March): 7-28.

Ó Dochartaigh, N. (2022). “The rise and rise of a new majority in Northern Ireland.” UK in a Changing Europe. 18 May. Retrieved from. 

O’Leary, B. (2022). Making Sense of a United Ireland: Should It Happen? How Might It Happen? Sandycove an imprint of Penguin Books.

O’Leary, B., T. Lyne, J. Marshall, and B. Rowthorn. (1993). Northern Ireland: Sharing Authority. London: Institute for Public Policy Research. October.

Raymond, C. (2019). “Alliance Breakthrough Part of a Broader Trend.” Queen’s Policy Engagement. 28 May. Retrieved from. 

Roche, P.J. (1995). “Northern Ireland and Irish Nationalism,” in The Idea of the Union, 128-134.

Ruane J, and J. Todd. (1992). “Diversity, Division and the Middle Ground in Northern Ireland.” Irish Political Studies 7:1, 73-98.

Tonge, J. (2020). “Beyond Unionism versus Nationalism: the Rise of the Alliance Party of Northern Ireland.” Political Quarterly 91:2 (April–June): 461-466.

Tonge, J. (2022a). “The 2022 Northern Ireland Assembly Elections: Polling, Power-Sharing, Protocol.” Political Insight 13:2 (June): 10-13.

Tonge, J. (2022b). “Voting into a Void? The 2022 Northern Ireland Assembly Election.” Political Quarterly 93:3 (July/September): 524-529.

Tonge, J., and J. Evans. (2020). “Northern Ireland: From the Centre to the Margins?” Parliamentary Affairs 73: Supplement 1 (September): 172-88.

Tonge, J., and R. Gomez. (2015). “Shared Identity and the End of Conflict? How Far Has a Common Sense of ‘Northern Irishness’ Replaced British or Irish Allegiances since the 1998 Good Friday Agreement?” Irish Political Studies 30:2 (April): 276–298.

Whitten, L.C. (2020a). “#LE19 – a turning of the tide? Report of local elections in Northern Ireland, 2019.” Irish Political Studies 35:1 (January): 61–79.

Whitten, L.C. (2020b). “Breaking Walls & Norms: A Report on the UK general election in Northern Ireland, 2019.” Irish Political Studies 35:2 (April): 313-330.

Whitten, L.C. (2023). Constitutional change in Northern Ireland. Institute for Government and Bennett Institute for Public Policy, Cambridge. August. Retrieved from. 

Working Group. (2021). Working Group on Unification Referendums on the Island of Ireland. Final Report. The Constitution Unit, School of Public Policy, University College London. May. Retrieved from.

⏮ Mike Burke has lectured in Politics and Public Administration in Canada for over 30 years.

1 comment:

  1. As ever, Mike, an excellent piece of work. Well thought out and presented.

    ReplyDelete