No Through Road / Ní Trébhóthar É Seo: Belfast’s Dual-Language Street Sign Impasse…

When Declan Kearney, Sinn Féin MLA and former party chairman, accused Gordon Lyons, the DUP Communities Minister responsible for the Irish language strategy, of presiding over a ministry hostile to the language, Lyons responded with an inversion. It was Sinn Féin, he told the Assembly on 17 February, that had done more to damage Irish language development — and his evidence was not some executive delay or budgetary shortfall but a local government signage policy: the 15% resident-signature threshold that Belfast City Council applies before erecting dual-language street signs. A policy designed, by Sinn Féin’s account, to normalise the Irish language in public space had become, by the Communities Minister’s account, proof that the party was willing to override local majorities and corrode community relations in the language’s name. Whether or not Lyons’s attack was opportunistic, the threshold he targeted deserves the scrutiny he declined to provide. However, a Slugger O’Toole piece by DeBeer, published last August, supplies us with just that.

DeBeer opened with a double-take: Sarah Bunting, the DUP group leader at Belfast City Council, had criticised the Council’s decision to erect dual-language signage on Shandon Park, an affluent residential street in outer east Belfast. The signs had been approved after 16.8% of residents voted in favour, with 49.6% against. What made the case unusual was that Alliance, which typically votes with nationalists on signage questions, opposed the application on this occasion because of the size of the disparity. The application had initially been blocked at committee on unionist and Alliance votes; a full Council meeting then overturned that decision.

DeBeer read the Shandon Park outcome as evidence that the Council’s 15% threshold is anti-democratic, and argued for a 50% requirement. The 15% figure is defensible on its own terms. It is the mechanism that delivers it — a street-by-street consent procedure that transforms a constitutional question into a series of neighbourhood contests — that is the real structural problem. The case for reform is not that the threshold should be higher, but that the survey mechanism should be replaced by a statutory framework — something Northern Ireland’s parties had already agreed in principle in New Decade, New Approach (2020).

The 15% Threshold

Belfast City Council did not set 15% arbitrarily. Its Dual Language Street Signs Policy, published on the Council website, cites two sources. The European Charter for Regional or Minority Languages, to which the UK is a signatory, does not prescribe numerical thresholds; it establishes flexible obligations, expressed through ‘where numbers justify’ formulations, that require state signatories to avoid creating barriers to minority language use in public life. Separately, the UN Special Rapporteur on minority issues has offered, as comparative guidance rather than a binding requirement, the suggestion that trigger thresholds for minority language signage fall within the range of 5 to 20% of the local population. The Policy document quotes both sources. At 15%, Belfast sits at the upper end of that non-binding range — more demanding of Irish speakers than the Rapporteur’s guidance suggests is necessary, and considerably more demanding than the comparable arrangements in Wales or Scotland, which require no resident survey at all. A 50% threshold of the kind DeBeer proposes would place Belfast well outside even this non-binding comparative range and would sit uneasily with the Charter’s barrier-avoidance obligations.

The point is routinely lost in the public argument. The 15% threshold is not an invention of nationalist politics. It is the upper bound of a range drawn from comparative experience across multiple jurisdictions about what kinds of trigger mechanisms succeed in protecting minority languages without generating excessive administrative burden. The Council’s policy is, in this respect, more conservative than international guidance suggests it should be.

There is, however, a dimension to the mechanism that the Charter’s principles and the Rapporteur’s comparative guidance do not resolve. The Good Friday Agreement commits sovereign government to exercise power ‘with rigorous impartiality on behalf of all the people in the diversity of their identities and traditions,’ grounded in ‘parity of esteem and of just and equal treatment for the identity, ethos, and aspirations of both communities.’ That principle creates a constitutional baseline within which Belfast’s street sign policy must be evaluated, and it exposes a structural asymmetry that the 15% figure, however defensible on other grounds, does not cure. English-language street signs require no affirmative petition from residents; they are the administrative default. Irish-language signs require 15% support. Whatever the number, the mechanism itself treats the two languages unequally: one is presumed, the other must be earned. The parity of esteem commitment does not automatically entail that both languages receive identical administrative treatment in every context, but it raises a legitimate question about whether a consent threshold that applies to Irish and not to English can be squared with the Agreement’s foundational commitments — a question that defenders of the current policy have rarely addressed.

The Democratic Objection

The claim that 15% is anti-democratic depends on a particular — and mistaken — conception of what democracy requires. DeBeer’s argument is essentially majoritarian: a decision affecting a street should reflect the wishes of the majority on that street. Applied consistently, this logic would eliminate most minority rights protections. Disability access modifications to buildings are not subject to the majority approval of a building’s users. The entire point of rights-based protections is that they are not conditional on the goodwill of the majority.

This is not a peripheral point but a constitutional one. Article 27 of the International Covenant on Civil and Political Rights states that individuals belonging to linguistic minorities shall not be denied the right to use their own language in community with other members of their group. The European Charter operates on the same premise. Neither instrument permits a state to comply with its obligations only in areas where a local majority happens to approve. To argue, as DeBeer does, that 15% in favour at a street level should not be sufficient is to argue that Irish speakers may exercise their language rights only where they constitute a local majority — which is to say, only in areas that are already predominantly nationalist. This is not a democratic principle; it is a territorial one.

A further internal contradiction follows directly from this. Under Belfast’s policy, if 14.9% of residents petition for a sign and no one objects, the application still fails. DeBeer does not describe this outcome as anti-democratic. The consistency of the democratic objection turns out to be selective: it is invoked when the threshold produces an outcome that unionists dislike, and set aside when it produces outcomes they prefer. This is not a principled democratic argument; it is a veto dressed in democratic language.

The Structural Problem

Shandon Park is not the only case that reveals the mechanism’s deficiencies. A resident in Derry posted to the Slugger thread that had prompted DeBeer’s piece, quoting the result of a survey on their own street: 61% had voted in favour of dual-language signage, 39% against — but the application had failed because the relevant council’s threshold was set at 66%. A clear majority wanted Irish signage and did not get it, not because their right was denied, but because a locally set arithmetic rule happened to fall the wrong way. These two outcomes — signs going up in Shandon Park at 16.8% support in one city, signs blocked on a Derry street at 61% support in another — are not the product of a coherent rights-based framework. They are the product of a mechanism that distributes outcomes arbitrarily depending on which council has set which percentage for which reason.

The material consequences are also documented. BBC News reported in December 2024 that vandalism of Irish street signs had cost councils £60,000. On 11 October 2025, a bilingual street sign was reported to have been damaged with an angle grinder. These are foreseeable risks of a process that frames each application as a local contest with winners and losers, conducted at a level of granularity where the symbolic stakes are highest, and community relations are most exposed.

Shandon Park itself is instructive in one further respect. The name is an anglicisation of Seán Dún — old fort — and refers to the Norman motte that stands in the area, built on the site of an earlier Gaelic Irish fortification. The street whose dual-language sign generated weeks of political controversy, media coverage, and DUP condemnation carries, in its English-only name, a phonetic approximation of the Irish original. This was noted in the Slugger comments but not in DeBeer’s article.

The structural critique has sharper teeth than the Welsh and Scottish comparisons alone provide, because Northern Ireland’s own parties, when they actually negotiated a settlement on Irish-language rights, chose a different instrument entirely. NDNA committed the Executive to legislation establishing an Irish Language Commissioner, whose primary function was to develop best-practice standards for the use of Irish by public authorities — covering correspondence, websites, and public signage in official buildings — through a tiered, proportionate framework agreed with each body. That model was expressly institutional in scope, directed at public authorities rather than residential streets. The argument is not that NDNA was intended to resolve the street-sign question, but that when Northern Ireland’s own parties turned their minds to how Irish-language rights should be delivered, they chose a legislative, public-authority-facing model and did not propose street-by-street petitions for any category of public signage.

Belfast City Council’s survey mechanism is anomalous not merely by comparison with Wales and Scotland but against the grain of what Northern Ireland’s own political settlement identified as the appropriate instrument for this class of decision.

This points to a question the structural argument implies, but the debate rarely confronts directly: whether Belfast City Council is the right institution to be handling this at all. The question of whether Irish has a recognised place in the public realm of Northern Ireland is a constitutional one, settled in principle by the Good Friday Agreement and given statutory expression by the Identity and Language (Northern Ireland) Act 2022. A council committee adjudicating individual streets is not resolving that question; it is, in effect, declining to implement the logic of an answer already given at a higher level. The perceived legitimacy problem is also real: given the Council’s current political composition, any decision in favour of Irish signage will be read by a significant section of the unionist community as a partisan outcome rather than an authoritative institutional settlement — which is precisely what successive legal call-ins and petitions of concern have been designed to signal. The Council finds itself in this position not because it is the constitutionally appropriate venue but because Stormont’s prolonged dysfunction has left a legislative vacuum that councils have been forced to fill inadequately.

Wales, Scotland, and Northern Ireland

Wales offers the clearest demonstration that a workable alternative exists. Welsh bilingual signage today results from a combination of successive Welsh Language Acts, Welsh Language Standards developed from 2012, and decades of administrative practice — but none of it has ever required a residential petition at any threshold. Scotland has pursued a comparable course through the Gaelic Language (Scotland) Act 2005 and subsequent implementation by local authorities and transport bodies. In both cases, the critical feature is not the specific legislative vehicle but the level at which the decision was taken: national and institutional rather than residential and territorial. Statutory bilingualism depersonalises and deterritorialises the question, removing it from the domain of neighbourhood contests about whose language belongs on a given street — and it is this quality that explains why neither jurisdiction has produced the territorial flashpoints that Belfast’s consent mechanism reliably generates.

The survey mechanism is not necessary to manage community relations; it is a political accommodation that, paradoxically, has generated more visible conflict than a statutory policy would have. A Council-wide policy of bilingual street signs would have been contested at the chamber level, but it would not have produced a street-by-street series of Shandon Parks, each one a local flashpoint and a fresh occasion for vandalism. The mechanism does not reduce conflict; it relocates it to the residential level, where it is more personal and harder to resolve.

Northern Ireland’s own trajectory since 2020 reinforces this. The Identity and Language (Northern Ireland) Act 2022 legislated for the commissioner framework that NDNA had committed to, creating the statutory Irish Language Commissioner and the Office of Identity and Cultural Expression, without the residential survey model appearing anywhere in its provisions. When the Council voted in October 2025 to adopt a wider Irish-language policy covering bilingual corporate identity, Council facility signage, and staff uniforms — carried by Sinn Féin, the SDLP, Alliance, People Before Profit and the Greens, with unionist parties opposed — it did so through a single institutional decision. The DUP’s response was a legal call-in at council level. The conflict migrated to the institutional forum where, on the argument advanced here, it belongs. An equality screening by a Northern Ireland public authority into Irish-language signage at Belfast Grand Central Station found that bilingual signage at a public transport hub did not diminish the Agreement entitlements of those with a British identity, and cited Wales and Scotland in support. That finding came from a Northern Ireland public authority applying its statutory equality obligations. It does not automatically resolve the residential street question, but it formally removes the premise that Irish-language signage is inherently incompatible with the Agreement’s guarantee of British identity.

The Objections

Two objections to a statutory solution deserve direct engagement. The first is that Stormont legislation on Irish-language street signage would face cross-community resistance, and that unionist parties would deploy a petition of concern or equivalent procedure to block it. This is a genuine practical obstacle. But its force is weaker than it appears. The DUP signed NDNA in January 2020 — under the direct pressure of imminent Westminster legislation, had the parties failed to reach agreement — and accepted the commissioner framework. That agreement was contingent and hard-won, and the political landscape has shifted since. But the claim that unionism will never accommodate statutory Irish-language provision is contradicted by the fact that it already has. The objection from Stormont’s consent mechanisms is an argument about current political conditions, not a permanent constitutional bar.

The objection also founders on its implied alternative. What is being defended in its place is a policy that has produced repeated council deadlocks, legal challenges, a documented £60,000 vandalism bill, and unresolved controversy — not a functioning alternative, but the absence of one. The difficulty of legislating at Stormont is a reason to pursue the legislative route carefully, with attention to sequencing and framing, not a reason to persist with a mechanism that demonstrably fails on its own terms.

The second objection is that a statutory framework would not neutralise the territorial anxieties that fuel the controversy; it would merely displace them from the street to the Assembly chamber. This has genuine analytical force. The objection to Irish-language signs in unionist areas is not primarily to the 15% mechanism; it is to Irish script in public spaces that loyalist communities regard as their territory. No change to the administrative machinery removes that objection. A statutory framework does not eliminate that anxiety; it contains it within a forum that has the democratic authority and institutional capacity to manage it, rather than distributing it across hundreds of residential streets where it produces heat without resolution. The territorial anxiety does not disappear under a statutory scheme; it is simply required to make its case in the forum designed to adjudicate it.

Conclusion

The 15% threshold for Irish-language street signs in Belfast is defensible. It falls within the comparative range the UN Special Rapporteur has suggested, it sits at the conservative end of what the European Charter’s barrier-avoidance obligations require, and the democratic objection to it rests on a majoritarian premise that, applied consistently, would dismantle minority rights protections across the board. The critique that a 15% minority should not be able to impose on an unwilling majority misunderstands the function of minority rights, which exist precisely because they are not subject to majority approval.

The harder point is that the threshold is the wrong thing to argue about. The mechanism that delivers it is constitutionally asymmetric — English requires no petition, Irish requires 15% — in a jurisdiction whose foundational agreement commits to parity of esteem between the two traditions. And it is being administered by an institution — Belfast City Council — that lacks the constitutional standing and cross-community legitimacy to settle a question of this character, and is doing so only because Stormont has failed to legislate. The NDNA framework and the 2022 Act demonstrate both the legislative pathway and the correct institutional logic: language rights belong in statute, administered at the appropriate level, not in neighbourhood surveys conducted street by street.

The debate will remain a cul-de-sac for as long as the wrong institution is asked to resolve the wrong question. Street-level mechanisms cannot settle constitutional questions about the status of a minority language; they can only reopen them, one petition at a time, in the places where the stakes feel most personal, and the least resolution is available. The right forum is Stormont — the one that Northern Ireland’s own parties, when the argument was conducted seriously, already chose.