From 14 to 16 October 2025 the UK Supreme Court is due to hear an important case on the extent to which the Windsor Framework – the new name for the revised Ireland / Northern Ireland Protocol to the EU–UK Withdrawal Agreement of 2020 – protects ‘civil rights’ in Northern Ireland. In Dillon’s and Others’ Application, an appeal by the UK government against a decision by the Court of Appeal in Northern Ireland (and a cross-appeal by some of the applicants), the Justices’ decision will turn in part on how they interpret the phrase ‘civil rights’ in the Rights, Safeguards and Equality of Opportunity section (RSE section) of the Belfast (Good Friday) Agreement 1998). The Court of Appeal decided that certain provisions in the Northern Ireland Troubles (Legacy and Reconciliation) Act 2023 were in breach of article 2(1) of the Windsor Framework because they ‘diminished’ the ‘civil rights’ guaranteed by the 1998 Agreement. 

The judgment of the Court of Appeal addressed these issues in paragraphs [113]–[125]. It did so in the course of applying the six-stage test it had previously laid down (in SPUC Pro-Life Ltd’s Application) for deciding whether a breach of article 2 of the Windsor Framework had occurred. That case concerned a challenge to the Abortion (NI) Regulations 2021 and on 22 November 2023 the Supreme Court rejected SPUC’s application to appeal against the Court of Appeal’s judgment. In Dillon the Court of Appeal emphasised, at [96], that the SPUC test was ‘an aid and not a binding or rigid code’ but it nevertheless methodically applied it. The first stage of the six-stage SPUC approach is to ask whether the supposedly diminished rights are indeed rights (or equality of opportunity protections) included in the relevant part of the Belfast (Good Friday) Agreement. 

The rights relied upon by the applicants in Dillon were those set out in three different documents – the EU’s Victims of Crime Directive 2012, the ECHR and the Charter of Fundamental Rights of the European Union. As the ECHR rights are still available to everyone in Northern Ireland (and Great Britain) after Brexit, the focus here will be on the relevant rights in the Directive and Charter, which were as follows:

• the right conferred by article 11 of the Directive to have a review of a decision not to prosecute;
• the right conferred by article 16 of the Directive to have a decision made, in the course of criminal proceedings, on compensation from an offender;
• the right in article 1 of the Charter to have one’s human dignity respected and protected;
• the right in article 47 of the Charter to have an effective remedy before a tribunal if one’s rights and freedoms guaranteed by EU law are violated.

Reliance was also placed on the Charter’s Articles 2, 4 and other parts of 47, but the wording of those provisions is not materially different from those in Articles 2, 3 and 6 of the ECHR and so will not be addressed here.

Let us now consider whether these four ‘non-Convention’ rights are ‘civil rights’ for the purposes of article 2(1) of the Windsor Framework. On this point the Court of Appeal agreed with Colton J’s very broad interpretation of the phrase and, somewhat remarkably, said that they were prepared to do so whether or not the interpretative approach suggested by the Vienna Convention on the Law of Treaties was applied (or, for that matter, the rules of international customary law on which the Vienna Convention’s provisions are based). With respect, however, there may be some flaws in the Court of Appeal’s reasoning. 

First, the Court of Appeal placed considerable weight on paragraph 1 of the RSE section in the B(GF)A, where we read that ‘The parties affirm their commitment to the mutual respect, the civil rights and the religious liberties of everyone in the community’. To me this affirmation of a commitment to civil rights is a far cry from a declaration that, for example, the victim of a crime must be able to request a review of a decision not to prosecute someone for that crime. The fact that none of the eight specific rights subsequently set out in paragraph 1 make no reference whatsoever to criminal justice processes strongly suggests that the drafters did not have them in mind when composing the introductory wording in that paragraph. This is especially so since that wording explicitly situates the paragraph ‘[a]gainst the background of the recent history of communal conflict’ yet makes no mention of the right of access to justice or the right to hold people to account for their actions during the conflict. Therefore, the wording of the ‘human rights’ part of the RSE section in the Agreement looks as if it was intended to be aspirational and motivational in nature, not directional or in any way legislative. Christopher McCrudden’s informative analyses of the negotiation of article 2 of the Windsor Framework and of the ‘civil rights and religious liberties’ paragraph in the Good Friday Agreement do not, to me, suggest otherwise. And, of course, this part of the Agreement, unlike many other parts, was not transposed into the law of Northern Ireland by the Northern Ireland Act 1998. 

Second, the Court of Appeal said that later in the RSE section of the Agreement ‘victims’ rights were specifically recognised’. It added that this was ‘unsurprising’ given that the Agreement was designed to address, to a large degree, the legacy of the Troubles. But in fact the three paragraphs headed ‘Reconciliation and victims of violence’ are very general in nature and are anticipatory rather than confirmatory of existing rights. In paragraph 11 the participants ‘look forward to the results of the Northern Ireland Victims Commission’, while paragraphs 12 and 13 foresee ‘a peaceful and just society’, the ‘development of community-based initiatives’, the provision of supportive services, the importance of work being done ‘to develop reconciliation and mutual understanding and respect between and within communities’ and the essential need for ‘the promotion of a culture of tolerance at every level of society including initiatives to facilitate and encourage integrated education and mixed housing’. The Court of Appeal did not accept the Secretary of State’s submission that the victims’ rights mentioned in paragraph 12 are limited to a vague ‘right to remember’ and ‘to contribute to a changed society’, but the reality is that no other victims’ rights are specifically referred to in the Agreement. Yes, paragraph 11 says that ‘it is essential to acknowledge and address the suffering of the victims of violence as a necessary element of reconciliation’ but this does not imply that the conferment of additional rights is mandated. Even if the Court of Appeal is correct to say that an element of the reconciliation process ‘is plainly addressing these issues by means of legal remedies and avenues which have long been recognised as securing a measure of justice for victims’, it does not follow that the extent of those remedies and avenues must be measured by the term ‘civil rights’ in paragraph 1 of the RSE section. The blunt fact is that the Belfast (Good Friday) Agreement is notorious for not having addressed victims’ rights.  

Third, the Court of Appeal agreed with Colton J that the rights set out in articles 11 and 16 of the Victims of Crime Directive 2012 were encompassed within the notion of victims’ rights addressed set out in paragraphs 11 and 12 of the RSE section. It thought that the nature of the victims’ rights relied on in Dillon ‘are closely linked to acknowledging and addressing the suffering of victims’ ([117]). But is this really so when we look in more detail at articles 11 and 16 of the Directive? 

Article 11(1) provides that ‘Member States shall ensure that victims… have the right to a review of a decision not to prosecute’. The Independent Commission for Reconciliation and Information Recovery, even under the 2023 Act’s provisions before they were declared incompatible with the ECHR, was never empowered to ‘make a decision not to prosecute a person’. Instead an immunity requests panel could grant immunity from prosecution if certain conditions were met and the immunity from prosecution could be revoked if it transpired that the person granted immunity had made false statements in their application to the panel. The Commission (like the Police Service of Northern Ireland before the Commission was established) has to decide if someone should be referred to the Public Prosecution Service for prosecution; it has no power to decide whether to prosecute or not, any more than the police have. In any event, a decision not to pass papers to the PPS is presumably judicially reviewable and if papers are passed to the PPS but the latter decides not to prosecute that decision too is judicially reviewable. 

The Court of Appeal accepted Colton J’s statement, at [608], that the Victims of Crime Directive ‘pre-supposes the possibility of a prosecution. Any removal of this possibility is incompatible with the Directive’, but this surely goes too far. In the Victims of Crime Directive, the ECHR and the Charter of Fundamental Rights no right is conferred on anyone to have someone prosecuted. The European Court of Human Rights has made it clear that an ECHR-compliant investigation into a killing gives the victim’s family a right to a thorough investigation capable of leading to someone’s prosecution, but it does not give the family a right to have such a prosecution initiated (e.g. Armani Da Silva v UK, 2016: ‘the Court has not faulted a prosecutorial decision which flowed from an investigation which was in all other respects Article 2 compliant’, para 259, GC). In Northern Ireland, as in England and Wales, the Prosecution Service applies two tests when deciding whether to prosecute someone: is there enough evidence to make it likely that if prosecuted the person will be convicted and is the prosecution in the public interest. Surely article 11 of the Victims of Crime Directive was never intended to apply to situations where the conduct in question has been decriminalised or where some other legislation (such as a statute of limitation) makes a prosecution impossible?

The same reasoning can be applied to challenge the Court of Appeal’s conclusion (differing from Colton J this time) that the protection afforded by article 16 of the Victims of Crime Directive was also diminished by the 2023 Act: if no criminal proceedings are possible, a victim cannot receive compensation in the course of them. (The Court of Appeal then accepted that there was no underpinning in EU law for the right to bring civil actions against the state (as opposed to an offender), so it did not declare section 43 of the 2023 Act to be invalid under article 2 of the Windsor Framework, only incompatible with Convention rights.)

Fourth, the Court of Appeal, like Colton J, was strongly influenced by the UK government’s Explainer Document issued in August 2020.  In that document, at section 13, the government conceded that victims’ rights are protected by article 2 of the Windsor Framework. Moreover, the Court of Appeal suggested, at [89], that, in line with article 31(3)(b) of the Vienna Convention on the Law of Treaties, the government’s subsequent practice in the application of the EU–UK Withdrawal Agreement helped to confirm the Court’s interpretation of the treaty. But, of course, neither of those realities necessarily obliges the Supreme Court to hold that article 2 embraces the Victims of Crime Directive if it chooses to interpret the term ‘civil rights’ in a less expansive way.

At this point it is worth stepping back from the detail of the case and asking ourselves whether the Court of Appeal’s generous and purposive approach to the interpretation of the Belfast (Good Friday) Agreement is in line with previous case law involving the interpretation of that document. Without going into detail, it seems that it is not. The Vienna Convention on the Law of Treaties has not been cited to any significant extent in previous cases concerning the interpretation of the Agreement. Article 2 of the Windsor Framework – itself part of an international treaty – has prompted new thinking in this context. Likewise the European Union (Withdrawal) Act 2018 and the European Union (Withdrawal Agreement) Act 2020 are being interpreted purposively as constitutional statutes in a manner which is hard to reconcile with statements by Law Lords and Supreme Court Justices in cases such as Watkins v Home Office (2006) and Re Agricultural Sector (Wales) Bill (2014), where they said that so-called constitutional statutes (such as the Northern Ireland Act 1998) must be interpreted in the same way as any other statute.

In conclusion, one is entitled to wonder what other rights, if the Court of Appeal’s interpretation of the term ‘civil rights’ is endorsed by the Supreme Court, may then gain greater protection in Northern Ireland thanks to article 2 of the Windsor Framework. The High Court of Northern Ireland has already ruled in Re Northern Ireland Human Rights Commission’s Application (2024) that immigration rights fall into that category and an appeal in that case is on hold pending the Supreme Court’s decision in Dillon’s and Others’ Application. It may be that the protection of a number of rights – such as the right to equal opportunity in all social and economic activity regardless of class, the right to freedom from sectarian harassment and the right of women to full and equal political participation (all mentioned in paragraph 1 of the RSE section) – will be immediately enhanced. Moreover, in the High Court at [543], Colton J relied on the Oxford English Dictionary’s definition of civil rights and the Court of Appeal did not demur from that. The OED defines the term as including ‘the political, social and economic rights which are recognised as the entitlement of every member of a community, and which can be upheld by appeal to the law’. Even bearing in mind that article 2 of the Windsor Framework applies only if all six conditions set out in the SPUC case have been met, applying such a definition of ‘civil rights’ could (contrary to the intentions of the drafters of both the B(GF)A and the Windsor Framework) open the door to a whole host of rights being guaranteed in a wholly unanticipated fashion Though that outcome might be of benefit to many in society, it would, in my view, be an example of inappropriate judicial activism in such an important constitutional context as the peace settlement in Northern Ireland.    

Brice Dickson, Emeritus Professor of International and Comparative Law, Queen’s University Belfast.

The author is a non-executive Commissioner on the Independent Commission for Reconciliation and Information Recovery but is writing here entirely in a personal capacity. He is grateful for the opportunity to first air these points at a conference in Belfast on 27 September 2024 organised by the Irish Centre for European Law. He greatly benefited from the discussions at that event.

(Suggested citation: B. Dickson, ‘The meaning of ‘civil rights’ in the Belfast (Good Friday) Agreement’, U.K. Const. L. Blog (13th October 2025) (available at https://ukconstitutionallaw.org/))