Kristie Higgs v Farmor's School

The Respondent in these proceedings (to which I will refer as “the School”) is a secondary school in Fairford in Gloucestershire. At the time relevant to these proceedings the Claimant had been employed by the School for six years, latterly as a pastoral administrator and work experience manager. In the first of those roles she was responsible for overseeing students who had been removed from class for disruptive behaviour. She has two children, the elder of whom was a pupil at the School. She is a Christian.

On 26 October 2018 a parent at the School emailed the Head Teacher, Matthew Evans, complaining that the Claimant had expressed “homophobic and prejudiced views” on her Facebook page: I give details below. Following an initial interview with the Claimant, on 30 October Mr Evans asked a member of staff called Sue Dorey to conduct an investigation. The Claimant was suspended. On the basis of Ms Dorey’s report, disciplinary charges were brought against the Claimant. The charges were considered at a hearing on 19 December 2018 before a panel chaired by one of the governors, Stephen Conlan. By letter dated 7 January 2019 she was summarily dismissed for gross misconduct. An internal appeal was unsuccessful.

On 15 April 2019 the Claimant began proceedings in the Employment Tribunal (“the ET”). Although she initially raised other complaints, the complaints which eventually proceeded were of (direct) discrimination and harassment, within the meaning of sections 13 and 26 respectively of the Equality Act 2010, in both cases on the ground of religion or belief.

The claim was heard by the ET (Employment Judge Reed, Mrs England and Ms Maidment) in Bristol on 21-24 September 2020. The Claimant was represented by Mr Pavel Stroilov, of Andrew Storch Solicitors, and the School was represented by Ms Debbie Grennan of counsel. By a judgment sent to the parties on 6 October 2020 the Claimant’s claims of discrimination and harassment were dismissed.

The Claimant appealed to the Employment Appeal Tribunal (“the EAT”). The appeal was heard by the President, Eady J, on 16 March 2023. The Claimant was represented by Mr Richard O’Dair of counsel and the School again by Ms Grennan. The Archbishops’ Council of the Church of England was given permission to intervene and was represented by Mrs Sarah Fraser Butlin. By a judgment handed down on 16 June the Claimant’s appeal was allowed and the claim remitted to the ET.

Although the Claimant had to that extent succeeded in her appeal, she believes that the EAT should have gone further and have held for itself that her claim succeeded; and by an Appellant’s Notice dated 7 July 2023 she appealed to this Court on that basis. Permission to appeal was given by Elisabeth Laing LJ.

By a Respondent’s Notice dated 5 February 2024 the School sought permission to cross-appeal on the basis that, while it accepted that the claim had to be remitted, the EAT’s formulation of the question requiring determination was erroneous. By order dated 27 March Elisabeth Laing LJ refused that permission, and she subsequently refused permission to re-open that decision.

The Claimant has again been represented before us by Mr O’Dair. The School has been represented by Mr Sean Jones KC and Mr Christopher Milsom. The Archbishops’ Council has been permitted to intervene and has again been represented by Mrs Fraser Butlin (now KC). Permission to intervene has also been given to the Free Speech Union Ltd (“the FSU”), the Association of Christian Teachers (“the ACT”), the charity Sex Matters, and the Equality and Human Rights Commission (“the EHRC”), though in the case of all but the EHRC by written submissions only. They have been represented by, respectively, Mr Ben Cooper KC, leading Mr Spencer Keen, Mr Roger Kiska of Camerons Solicitors, Ms Akua Reindorf KC, and Ms Joanne Clement KC. (I should say that the EHRC’s written submissions were settled by Mr Tom Cross: Ms Clement was instructed to make the oral submissions when he fell ill, and she produced helpful short supplementary written submissions.) I am grateful for the work and thought that went into all the submissions, written and oral. I am also grateful for the immaculate way in which the bundles – particularly the bundles of authorities, which were voluminous – were prepared.

The email referred to at para. 2 above reads as follows:

“Dear Mr Evans,

I’ve noticed that a member of your staff who works directly with children has been posting homophobic and prejudiced views against the lgbt community on Facebook. I’m concerned that this individual may exert influence over the vulnerable pupils that may end up in isolation for whatever reason.  I find these views offensive and I am sure that when you look into it, you will understand my concern.  I’d rather remain anonymous as the person in question is … . I’ve attached a couple of screen shots so you can see what I’m referring to.”

The omitted words were redacted by the ET in order to preserve the anonymity of the complainant.

The attached screenshots were of the Claimant’s Facebook page showing the following post, apparently made on 24 October 2018:

“**PLEASE READ THIS! THEY ARE BRAINWASHING OUR CHILDREN!** On November 7^th the Government Consultation into making Relationships Education mandatory in primary schools, and Relationships and Sex Education mandatory in secondary schools closes.  Which means, for example, that children will be taught that all relationships are equally valid and ‘normal’, so that same sex marriage is exactly the same as traditional marriage, and that gender is a matter of choice, not biology, so that it’s up to them what sex they are.

At the same time it means that expressing and teaching fundamental Christian beliefs, relating to the creation of men and women and marriage will in practice become forbidden – because they conflict with the new morality and are seen as indoctrination into unacceptable religious bigotry.

Which means that freedom of belief will be destroyed, with freedom of speech permitted only for those who toe the party line!

We say again, this is a vicious form of totalitarianism aimed at suppressing Christianity and removing it from the public arena.

***Please sign this petition, they have already started to brainwash our innocent wonderfully created children and its happening in our local primary school now***”

There followed a link to a petition organised by CitizenGo.Org, an entity which describes itself as concerned to “uphold the right of parents to have children educated in line with their religious beliefs”. The main text was not drafted by the Claimant but cut-and-pasted from another source. She had added the words between the asterisks at the beginning and the end. The account was in her maiden name (rather than the married name the Claimant used for her work at the School) and contained nothing which suggested any connection with the School; but it is clear that her identity was apparent at least to the complainant, who must have known that she worked there and of her role with children “in isolation”. I refer to this as “the first post”.

In response to an enquiry from Mr Evans about whether they had any further information, the complainant on 29 October 2018 sent another email, attaching further screenshots and saying:

“I’m aware that not everyone has liberal views like myself but I do feel that people working directly with children should refrain from posting this type of view on social media.  I know of several children at Farmors who might fit into the category of person your staff member seems to find so obnoxious, friends of my children even.”

The additional screenshots were of what appear to be more than one (more recent) post by the Claimant, re-posting messages from campaigners in the U.S.A. objecting to materials used in schools there (“the re-posts”). I need not quote these in full. The following extracts sufficiently represent their gist and tone:

“While normal Americans are busy at work trying to provide for their families, liberal school systems are busy indoctrinating their children.  Kindergarten and first grade children are being primed for a gender fluid society.  Of course, the schools are introducing the propaganda in the name of anti-bullying campaigns, but we know better.

They are busy recruiting children for the transgender roster.  Their agenda is not about bullying.  They are using our children to promote their gender free society of madness.

… They are stealing the innocence of our children with a devious scheme to supplant traditional gender roles by differentiating a child’s gender assignment at birth with his perceived gender.

…

Not succumbing to the brainwashing of deranged educators is now a characteristic of bullying. The far-left zealots have hijacked the learning environment, and they insist on cramming their perverted vision of gender fluidity down the throats of unsuspecting school children who are a government mandated captive audience.

…

… Lying to children and convincing them that they can be anything they want to be when in reality they can’t is a form of child abuse, especially when it entails the changing of one’s genitalia or ingesting hormones.

The LBGT [sic] crowd with the assistance of the progressive school systems are destroying the minds of normal children by promoting mental illness. Delusional thinking is a form of psychotic thinking, and we have professionals promoting it to our young kids.”

I shall have to consider the content and language of those posts later. At this stage I will only note that they express two beliefs – (1) that gender is binary and not “fluid” (a view often labelled “gender-critical”) and (2) (though this only appears in the first post) that same-sex marriage cannot be equated with traditional marriage between a man and a woman – and accordingly that it is wrong to teach anything different to children, and particularly primary school children. Although all the posts are hyperbolically expressed, that is markedly more so in the case of the re-posts.

We do not have details of the School’s disciplinary process, but it appears to follow the conventional two-stage pattern comprising an “investigation” to establish whether there is a case to answer, followed if necessary by a disciplinary hearing based on formal allegations (in effect, charges). As noted above, the investigation stage started on 30 October 2018, and in a report dated 30 November Ms Dorey found that there was a case to answer on four allegations, which I summarise below. I need not give the full details of the report but it is important to record the substance of the Claimant’s response on the points raised. These are summarised at paras. 9-15 of the judgment of Eady J, the key parts of which I gratefully reproduce below.

In an initial interview with Ms Dorey on 30 October the Claimant confirmed that she had made the first post and the re-posts (save where it is necessary to distinguish I will refer simply to “the posts”). She accepted that it was possible that they might have been seen by parents of pupils at the School. She was asked whether other people might consider the posts offensive or prejudiced. She was recorded in the note of the meeting as responding:

“Yes.  I am not against gay, lesbian or transgender people.  It’s about making sure people are aware of what’s going on in the primary school.  It’s not about the schools, they are just following government policy, it’s about the government.”

The “primary school” referred to was a school attended by her younger child. She also said:

“I don’t regret making the posts, it’s about the children in the primary school. I don’t have any issues with gay, lesbian or transgender people, I love all people.”

At a meeting on 8 November the Claimant again told Ms Dorey that the posts were concerned with what was happening at her son’s primary school, as a result of Government policy, and that she had wanted them to be seen by other parents there.  She acknowledged that her term “brainwashing” was “not the best language to use”, and that as regards the re-posts she should have used her own words or included a link. She was again asked whether she thought other people might consider the posts offensive or prejudiced. She replied:

“I know that there are transgenders and gays who do have the same beliefs as me.  …  I am not against gay people, it doesn’t say that.”

She was asked whether she considered her posts might compromise her position of trust in working with children, some of whom might be LGBT. She answered:

“No I don’t.  Students know me and I know gay students, I wouldn’t treat any of them any different. … I wouldn’t bring this into school.”

As for the risk of reputational damage to the School, she said:

“People should know my belief … as people on [Facebook] … are my friends.  They would know me as a person and know I wouldn’t discriminate.  If anything I am being discriminated against as I have shared what the government is doing, this is what I stand for …”

It is important to record that in her report Ms Dorey referred to no evidence that the Claimant had ever expressed her views about gender fluidity or same-sex marriage to pupils or staff in the School or treated gay, lesbian or transgender pupils or staff differently.

The hearing before the disciplinary panel lasted several hours. The essential points about what transpired at it appear in the dismissal letter, which is full and carefully structured. For the most part it is sufficient to say that the Claimant adopted substantially the same position as in the investigation. But I should quote one passage, which is headed “The Language”. This reads:

“We discussed whether the two posts were yours and your position was that you were simply reposting existing articles and had only added a few words; they were not your posts. However, you confirmed that you had read the articles you re-posted and agreed with the content of them. Indeed, you wanted these articles to be circulated more widely as you felt it important for people to be aware of the content. We discussed at length whether the language used within these posts could be deemed as offensive or discriminatory and highlighted the specific words ‘brainwashing, transgender roster, madness, devious scheme, child abuse and mental illness’ amongst others (I will not repeat them all here). We were keen to understand whether, upon reflection, you understood that the use of such language could be deemed as offensive and that it has the ability to cause damage to the reputation of the school. You stressed that whilst you may not have chosen to use the same language as used in the articles, you agreed with the content and upon reflection you would not have acted differently. When asked specifically what language you would not use you identified the words brainwashing, delusional thinking and psychotic thinking. You did not say you would not endorse or use any of the other words.”

The letter finds allegations 1, 2 and 4 to be proved. Detailed reasoning is given as regards allegations 1 and 2 (though not allegation 4, which added nothing of substance). Those reasons can be sufficiently summarised for our purposes as follows.

Allegation 1 had two limbs – “illegal discrimination” and “serious inappropriate use of social media, e.g. Facebook or other online comments, that could bring the school into disrepute”.

As to the first limb, the Claimant’s posts were found to constitute “discrimination against [the] complainant in the form of harassment on the grounds of sexual orientation and/or gender reassignment”. It was common ground before us that the finding under the first limb of allegation 1 that the Claimant was guilty of unlawful harassment of the complainant was misconceived. Even if, which is unlikely but possible, the complainant was employed at the School, a post on the Claimant’s personal Facebook account could not be said to create, as required by section 26 of the 2010 Act, “an intimidating, hostile, degrading, humiliating or offensive environment” for them.

As to the second limb, the panel referred to the “inflammatory and quite extreme” language of the posts. It noted that, although the Claimant had said that she “did not agree with the wording” of the original messages, she had nevertheless copied/re-posted them because she agreed with their content, and that she had said that she did not regret doing so and had not taken the posts down. The letter says:

“You were unable to confirm that you would not do it again, only qualifying that you would not have used the word ‘brainwashing’ and would have added a link to the articles.”

I should quote in full what it says about the potential for damage to the School’s reputation:

“Regarding bringing the school into disrepute/damaging the reputation of the school, we agree that there is no direct evidence that as a matter of fact the reputation of the school has to date been damaged. The complainant did not appear to be criticising the school and Sue Dorey confirmed that to her knowledge, no other complaints had been received. You said that you were only sharing these posts with your Facebook friends, that you make no reference to working at Farmor’s on your Facebook page and use your maiden name. However, you also said that your friends know where you work, that you are a local girl, went to Farmor’s school as a child and have lived locally for years and people know you. Some of your friends are also members of staff at the school and others parents of pupils at Farmor’s. You affirmed that you were aware that your posts could be reposted and seen by others outside your immediate ‘Facebook Friends’ group. We considered that these posts have the potential to be reposted and therefore potentially reach a far wider audience, many of whom will know you and your maiden name and where you work due to your local connections and long history in the area. Once posted, you lose a degree of control. You had also expressed your view that it was your intention to get the message across to others and you hoped it would be reposted. You confirmed that you had not removed the posts.

We therefore concluded that whilst there was no actual evidence that the school’s reputation had to date been harmed, there was the potential that it could be. Whilst we also agree that it is unlikely that anyone reading these posts would consider that they represented the school’s views, we were concerned that there was sufficient association with the school by way of the fact that you and your place of work is well known locally and were therefore satisfied that your posts via facebook could have and may still bring the school into disrepute.”

Allegation 2 alleged breaches of the School’s Code of Conduct. Three breaches were identified:

The panel decided that the matters which it found proved constituted gross misconduct within the meaning of the School’s Conduct Policy. It then turned to the question of sanction. I should set out its decision and reasoning in full:

“Having made the finding of gross misconduct, we next considered the appropriate sanction. We had regard to the school’s Conduct Policy section 6.1 where it is stated that in the absence of exceptional mitigating circumstances, offences of gross misconduct will result in summary dismissal.

We had regard to the fact that we felt that throughout the hearing you were at times evasive, contradictory and despite having had time to reflect on your actions, appeared to have no insight into the impact that your posts had on the complainant, were dismissive of those that could take offence, calling them liberals and were unable to give us confidence that such actions would not be repeated. You had not removed the posts. We were concerned that upon reflection you demonstrated no understanding of the implications of your actions and how this may reflect upon your professional reputation as well as that of the school within the community which we serve. However, we also took into account that there were no complaints concerning your standards of work, we were unaware of any other disciplinary matters, that the posts had been made in the context of your personal facebook account and that you had six years’ service with the school.

Overall we were not satisfied that any lower level of disciplinary sanction would be appropriate in view of the nature of your misconduct and your lack of understanding of the potential impact upon the school. We concluded that there were no exceptional mitigating circumstances and therefore concluded that the correct sanction was summary dismissal.”

The letter concluded by telling the Claimant that she was dismissed with immediate effect.

In her original Claim Form the Claimant complained not only of discrimination and harassment but of unfair and wrongful dismissal. The latter claims were, however, dismissed by consent at a case management hearing on 13 December 2019 on the basis that they were out of time. The discrimination claim was of direct discrimination within the meaning of section 13 of the 2010 Act. An application to amend the claim to include an allegation of indirect discrimination was refused.

The Claimant was not required at the case management hearing to be precise about the “religion or belief” on which she relied. However, she identified her case on this aspect at the substantive hearing before the ET. At paras. 29-30 of its reasons the Tribunal records that she did not claim to have been discriminated against or harassed “for her Christianity per se” but for the following beliefs:

“(a)

Lack of belief in ‘gender fluidity’.

(b)

Lack of belief that someone could change their biological sex/gender.

(c)

Belief in marriage as a divinely instituted life-long union between one man and one woman.

(d)

Lack of belief in ‘same sex marriage’. Whilst she recognises the legalisation of same sex marriage she believes that this is contrary to Biblical teaching.

(e)

Opposition to sex and/or relationship education for primary school children.

(f)

A belief that she should ‘witness’ to the world, that is when unbiblical ideas/ideologies are promoted, she would publicly witness to Biblical truth.

(g)

A belief in the literal truth of the Bible, and in particular Genesis 1 v. 27: ‘God created man in His own image, in the image of God He created him; male and female He created them’”.

That is essentially an elaboration of the beliefs expressed in the posts. The belief that gender is binary is encapsulated in heads (a) and (b), but also (g). The belief that marriage can only be between a man and a woman is covered by heads (c) and (d). Head (e) is perhaps literally distinct, but the Claimant’s main concern about sex education for primary school children would appear to be that it was being used as a vehicle for the inculcation of views in favour of gender fluidity and same-sex marriage. Since it is (now) common ground that these beliefs fall within the definition of the protected characteristic of “religion or belief” in the 2010 Act, I will refer to them as “the protected beliefs”. Head (f) is of a rather different character: it does not state a substantive belief, but rather a commitment to publicly stating certain such beliefs.

The claim as originally pleaded identified numerous detriments/acts of harassment, but in his oral submissions Mr O’Dair made it clear that he was now relying on four – (a) the Claimant’s suspension; (b) the investigation/disciplinary proceedings; (c) the dismissal; and (d) the rejection of her appeal. I will refer to (a)-(b) as “the disciplinary process claim” and heads (c)-(d) (though in truth it is hard to see that head (d) has any independent existence) as “the dismissal claim”. The dismissal claim is evidently the primary claim in these proceedings, as Mr O’Dair acknowledged.

The Claimant’s claims are brought under Part 5 of the 2010 Act, which is concerned with discrimination at work. However, for reasons which will appear, it is necessary to consider also the effect of articles 9 and 10 of the European Convention on Human Rights (“the Convention”) and related articles; and it is more convenient to start with these and the relevant provisions of the Human Rights Act 1998.

Article 9 is titled “Freedom of thought, conscience and religion”. It reads:

“1.

Everyone has the right to freedom of thought, conscience and religion; this right includes freedom to change his religion or belief and freedom, either alone or in community with others and in public or private, to manifest his religion or belief, in worship, teaching, practice and observance.

2.

Freedom to manifest one’s religion or beliefs shall be subject only to such limitations as are prescribed by law and are necessary in a democratic society in the interests of public safety, for the protection of public order, health or morals, or for the protection of the rights and freedoms of others.”

There is good deal of case-law, both domestically and in Strasbourg, about the effect of article 9. For present purposes I need refer only to two passages from the judgment of the Grand Chamber of the European Court of Human Rights (“the ECtHR”) in Eweida v United Kingdom 48420/10, [2013] IRLR 231, both concerned with the issue of a person’s right to manifest a belief – that is, to express it publicly or otherwise demonstrate it, in their actions or their clothing or appearance or otherwise, to all of which I will refer compendiously as “conduct”.

The first passage identifies the limitations on the protection accorded to the manifestation of a belief. Para. 80 of the judgment reads:

“Religious freedom is primarily a matter of individual thought and conscience. This aspect of the right set out in the first paragraph of Article 9, to hold any religious belief and to change religion or belief, is absolute and unqualified. However, as further set out in Article 9 §1, freedom of religion also encompasses the freedom to manifest one’s belief, alone and in private but also to practice in community with others and in public. The manifestation of religious belief may take the form of worship, teaching, practice and observance. Bearing witness in words and deeds is bound up with the existence of religious convictions (see Kokkinakis, cited above, §31 and also Leyla Şahin v. Turkey [GC], no. 44774/98, §105, ECHR 2005-XI, 44 EHRR 5). Since the manifestation by one person of his or her religious belief may have an impact on others, the drafters of the Convention qualified this aspect of freedom of religion in the manner set out in Article 9 §2. This second paragraph provides that any limitation placed on a person’s freedom to manifest religion or belief must be prescribed by law and necessary in a democratic society in pursuit of one or more of the legitimate aims set out therein.”

In that passage the Court identifies a fundamental difference between the right to hold a religious belief and the right to manifest it. While the former is absolute, the latter is qualified by paragraph 2, which allows the right to manifest a belief to be limited if the limitation is (a) “prescribed by law” and (b) justified by reference to the other legitimate interests there specified. It is important to appreciate the reason given for that difference. Whereas the holding of a belief is in the nature simply of a characteristic (albeit one that is to a greater or lesser extent a matter of choice), its manifestation constitutes conduct which is outward-facing and for that reason, as the Court says, “may have an impact on others” and accordingly may require to be limited so as to take account of other interests.

As for the nature of the qualification imposed by article 9.2, both the elements which I have labelled (a) and (b) are familiar and they need no exposition here. I will only record, at the risk of stating the obvious, that the phrase “necessary in a democratic society in the interests of …” requires an assessment of the proportionality of the limitation in question. The classic exposition of the correct approach to such an assessment appears in the decision of the Supreme Court in Bank Mellat v Her Majesty’s Treasury (no. 2) [2013] UKSC 39, [2014] 1 AC 700. At para. 74 of his judgment Lord Reed identified four questions, as follows:

“(1)

whether the objective of the measure is sufficiently important to justify the limitation of a protected right,

(2)

whether the measure is rationally connected to the objective,

(3)

whether a less intrusive measure could have been used without unacceptably compromising the achievement of the objective, and

(4)

whether, balancing the severity of the measure’s effects on the rights of the persons to whom it applies against the importance of the objective, to the extent that the measure will contribute to its achievement, the former outweighs the latter.”

He noted that “in essence, the question at step four is whether the impact of the rights infringement is disproportionate to the likely benefit of the impugned measure”. I will in this judgment use the shorthand “objective justification” for the four steps taken as a whole and will also sometimes refer to the first step as requiring “a legitimate aim”. That terminology is well recognised from the EU and domestic discrimination legislation and connotes essentially the same exercise.

The second passage from the judgment in Eweida addresses the question of what may constitute the manifestation of a belief. At para. 82 the Court makes the point that, although, as it had said in para. 80, the manifestation of a religious belief may take the form of “bearing witness in words and deeds”, it does not extend to every act which is “in some way inspired, motivated or influenced by” that belief. I need not set out the whole of the paragraph. The essential point is made in the following passage:

“In order to count as a ‘manifestation’ within the meaning of Article 9, the act in question must be intimately linked to the religion or belief. An example would be an act of worship or devotion which forms part of the practice of a religion or belief in a generally recognised form. However, the manifestation of religion or belief is not limited to such acts; the existence of a sufficiently close and direct nexus between the act and the underlying belief must be determined on the facts of each case.”

I should mention one point about terminology. The Court in Eweida, following the language of article 9.2 refers to “limitations” on the right to manifest a belief. But of course that term covers not only rules or restrictions which prevent the exercise of the right but also detriments imposed on individuals by way of sanction for having done so; and I will sometimes use the term “interference” in order to reflect that.

Article 10 is entitled “Freedom of expression”. It reads:

“1.

Everyone has the right to freedom of expression. This right shall include freedom to hold opinions and to receive and impart information and ideas without interference by public authority and regardless of frontiers. This Article shall not prevent States from requiring the licensing of broadcasting, television or cinema enterprises.

2.

The exercise of these freedoms, since it carries with it duties and responsibilities, may be subject to such formalities, conditions, restrictions or penalties as are prescribed by law and are necessary in a democratic society, in the interests of national security, territorial integrity or public safety, for the prevention of disorder or crime, for the protection of health or morals, for the protection of the reputation or rights of others, for preventing the disclosure of information received in confidence, or for maintaining the authority and impartiality of the judiciary.”

Although paragraph 2 is not identically worded to article 9.2 it was not suggested that it was for our purposes materially different in its effect. I cite later in this judgment various decisions of the ECtHR relevant to its effect.

I should mention also the provisions of articles 14 and 17 of the Convention because although neither is directly in issue in this case they feature in the reasoning of authorities to which I will have to refer.

Article 14 provides that “the enjoyment of the rights and freedoms set forth in this Convention shall be secured without discrimination on any ground such as … religion, political or other opinion”. The case-law establishes that the reference to “discrimination” embraces both direct and indirect discrimination and (in this respect differing from the EU and domestic legislation) that direct as well as indirect discrimination may in principle be justified. Unjustified discrimination against a person because they had manifested a religion or belief, or had exercised their right to free expression, would be a breach of their rights under article 9 or 10 irrespective of article 14, and it is in such cases relied on secondarily if at all: in Eweida, for example, the Court declined to decide the claim under article 14 on the basis that it was unnecessary to do so in the light of its finding that there had been a breach of article 9 (see para. 95 of its judgment).

Article 17 provides that:

“Nothing in [the] Convention may be interpreted as implying for any State, group or person any right to engage in any activity or perform any act aimed at the destruction of any of the rights and freedoms set forth herein or at their limitation to a greater extent than is provided for in the Convention.”

This provision has been deployed in the Strasbourg case-law to limit article 10 rights in the most serious kinds of “hate speech”: see para. 126 below.

Section 3 (1) of the Human Rights Act 1998 provides:

“So far as it is possible to do so, primary legislation and subordinate legislation must be read and given effect in a way which is compatible with the Convention rights.”

“The Convention rights” are defined in section 1 as (most of) the rights set out in the Convention, including those in articles 9 and 10.

Section 6 renders it unlawful for a public authority to act in a way which is incompatible with a Convention right (unless obliged to do so by primary legislation). A claim for breach of section 6 may be brought in the ordinary courts: see section 7.

I will deal first with the relevant provisions of the 2010 Act and then identify two particular points arising out of the case-law concerning (1) the manifestation of a belief and (2) the so-called “separability principle”.

Proscription of discrimination against employees. Section 39 (2) of the Act proscribes discrimination by an employer against an employee by (among other things) dismissing them or subjecting them to any other detriment.

Direct discrimination. Direct discrimination is defined in section 13 of the Act. The only relevant subsection for our purposes is (1), which reads:

“A person (A) discriminates against another (B) if, because of a protected characteristic, A treats B less favourably than A treats or would treat others.”

The phrase “because of” in section 13 (1) connotes a causative link between the protected characteristic and the treatment complained of. There has been a fair amount of exposition of the nature of that link in the case-law. The line of cases begins with the speech of Lord Nicholls in Nagarajan v London Regional Transport [2000] 1 AC 501 and includes the reasoning of the majority in the Supreme Court in R (E) v Governing Body of the JFS  [2009] UKSC 15, [2010] 2 AC 728 (“the JFS case”). What it refers to is “the reason why” the putative discriminator or victimiser acted in the way complained of. In some cases that reason is inherent in the act complained of: these are often referred to as “criterion cases”. But in others it consists in the “mental processes”, conscious or unconscious, that caused the discriminator to act, often referred to as their “motivation” (though not their “motive”). Where convenient I will in this judgment sometimes use the phrase “on the grounds of” as an alternative to “because of”: this was the language of the predecessor legislation to the 2010 Act, as also of the EU Framework Directive referred to below, and it is recognised as having the same effect.

Protected characteristics: religion or belief. Section 4 of the 2010 Act sets out a list of “protected characteristics”. They include “religion or belief”. Section 10 contains further provisions about that characteristic, but I need only note that subsection (2) provides that “[b]elief means any religious or philosophical belief”.

Indirect discrimination. Although, as I have said, the Claimant in this case alleges only direct discrimination, the role of indirect discrimination in this field is material to the submissions before us. It is defined in section 19 of the 2010 as follows:

“(1)

A person (A) discriminates against another (B) if A applies to B a provision, criterion or practice which is discriminatory in relation to a relevant protected characteristic of B’s.

(2)

For the purposes of subsection (1), a provision, criterion or practice is discriminatory in relation to a relevant protected characteristic of B’s if—

(a)

A applies, or would apply, it to persons with whom B does not share the characteristic,

(b)

it puts, or would put, persons with whom B shares the characteristic at a particular disadvantage when compared with persons with whom B does not share it,

(c)

it puts, or would put, B at that disadvantage, and

(d)

A cannot show it to be a proportionate means of achieving a legitimate aim.

(3)

The relevant protected characteristics are —

…

religion or belief;

... .”

I will adopt the usual shorthand of “PCP” for the “provision, criterion or practice” referred to in subsection (1); and I will borrow from the Framework Directive referred to below the paraphrase “apparently neutral” for the more elaborate language of subsection (2) (a).

Harassment. “Harassment” is elaborately defined in section 26 of the 2010 Act. For present purposes it is not necessary to quote the definition in full but only to note that the conduct in question must be “related to a relevant protected characteristic”: the relevant protected characteristics include religion or belief. Section 40 (1) of the Act proscribes the harassment of an employee by their employer in relation to the employment.

The provisions of Part 5 of the 2010 Act, as regards discrimination on the grounds of religion or belief, have their origins in, and represent the United Kingdom’s implementation of, EU Council Directive 2000/78/EC “establishing a general framework for equal treatment in employment and occupation” (“the Framework Directive”). Recital (1) to the Directive records, among other things, that the European Union “respects fundamental rights, as guaranteed by the European Convention for the Protection of Human Rights and Fundamental Freedoms [i.e. the Convention]”.

Article 1 defines the purpose of the Framework Directive as follows:

“The purpose of this Directive is to lay down a general framework for combating discrimination on the grounds of religion or belief [my emphasis], disability, age or sexual orientation as regards employment and occupation, with a view to putting into effect in the Member States the principle of equal treatment.”

Discrimination is defined in article 2. Paragraph 1 (a) defines direct discrimination as occurring “where one person is treated less favourably than another is, has been or would be treated in a comparable situation, on any of the grounds referred to in Article 1”: section 13 (1) (read with section 23) of the 2010 Act gives effect to that definition. Indirect discrimination is defined in paragraph 1 (b) in terms which have substantially the same effect as section 19 of the 2010 Act. Article 2.5 reads:

“This Directive shall be without prejudice to measures laid down by national law which, in a democratic society, are necessary for public security, for the maintenance of public order and the prevention of criminal offences, for the protection of health and for the protection of the rights and freedoms of others.”

As will be seen, that language is substantially to the same effect as, though not identical to, that of article 9.2 of the Convention.

I should mention, only because it is referred to in connection with the case-law cited below, that article 4 provides for a derogation from the right of equal treatment where a difference of treatment is a “genuine and determining occupational requirement”.

Manifestation of belief

It will be noted that, unlike article 9 of the Convention, the 2010 Act does not refer explicitly to discrimination on the grounds of the manifestation of a belief. However, it is clear, and was common ground before us, that the phrase “because of [the complainant’s] religion or belief” must be read as extending to such discrimination. That is authoritatively established by the decision of the Court of Justice of the European Communities (“the CJEU”) in Bougnaoui v Micropole SA C-188/15, [2018] ICR 139, which concerned an employer’s ban on the wearing of a headscarf by a Muslim employee (and more particularly whether it fell within the scope of the genuine occupational requirement exception in article 4). Having noted at para. 28 of its judgment that recital (1) to the Directive referred to the Convention, the Court said, at para. 30:

“In so far as the ECHR and, subsequently, the Charter (Footnote: ¹) use the term ‘religion’ in a broad sense, in that they include in it the freedom of persons to manifest their religion, the EU legislature must be considered to have intended to take the same approach when adopting Directive 2000/78, and therefore the concept of ‘religion’ in Article 1 of that directive should be interpreted as covering both the forum internum, that is the fact of having a belief, and the forum externum, that is the manifestation of religious faith in public.”

Since the 2010 Act constitutes the UK’s compliance with the Framework Directive, and the acts complained of occurred prior to “IP completion day” (31 December 2020) the Court’s decision is authoritative as to the scope of the protection afforded by section 13 read with section 4.

It is worth clarifying one point that came up in the submissions before us. There will be cases where the treatment complained of by the employee was ostensibly on the ground of conduct which manifested a religious or other belief but where it is found that the real reason was an animus against the belief in question. Such a finding may be straightforwardly because the employer’s account of its reasons is disbelieved; but it may also be because, as I put it in McFarlane v Relate Avon Ltd [2009] UKEAT 0106/09/3011, [2010] ICR 507, it is in the circumstances of the particular case “impossible to see any basis for the objection other than an objection to the belief which it manifests” so that “[the employer’s claim] to be acting on the grounds of the former but not the latter may be regarded as a distinction without a difference” (see para. 18). Neither kind of case is in truth a manifestation case at all, because the employer is motivated simply by the fact that the employee holds the belief (Footnote: ²). In a manifestation case proper the employer genuinely has no objection to the employee holding the belief and is motivated only by the conduct which constitutes its manifestation. Most claims of discrimination on the ground of religion or belief are likely to be genuine manifestation cases of this kind.

At the risk of stating the obvious, the fact that the 2010 Act gives employees a right not to be discriminated against on the ground of manifesting a belief does not mean that that right is unqualified; but the basis on which it should be treated as qualified is contentious in this appeal, and I return to it below.

“Separability”

In a case where the 2010 Act (or its predecessors), and other analogous legislation, affords protection to particular kinds of conduct by an employee – for example, in victimisation or whistleblowing cases, for making complaints of discrimination or making protected disclosures – the case-law recognises that it may be necessary to decide whether the real cause of the treatment is the conduct itself or is some properly separable feature of it. This is sometimes referred to as “the separability principle”. This line of authority is potentially applicable in a (true) manifestation case, since in such a case the court is concerned (untypically for a direct discrimination claim) with a motivation based not on the possession of the protected characteristic but on particular conduct on the part of the employee.

The most recent discussion of the separability principle can be found in the judgment of Simler LJ in Kong v Gulf International Bank (UK) Ltd [2022] EWCA Civ 941, [2022] ICR 1513. The case concerned an alleged dismissal for making a protected disclosure. At paras. 47-55 of her judgment Simler LJ considered a number of authorities concerned with protected conduct of various kinds and quotes various passages from them which I need not identify. At paras. 56-57 she says:

“56.

I would endorse and gratefully adopt the passages I have cited as correct statements of law. They recognise that there may in principle be a distinction between the protected disclosure of information and conduct associated with or consequent on the making of the disclosure. For example, a decision-maker might legitimately distinguish between the protected disclosure itself, and the offensive or abusive manner in which it was made, or the fact that it involved irresponsible conduct such as hacking into the employer’s computer system to demonstrate its validity. In a case which depends on identifying, as a matter of fact, the real reason that operated in the mind of a relevant decision-maker in deciding to dismiss (or in relation to other detrimental treatment), common sense and fairness dictate that tribunals should be able to recognise such a distinction and separate out a feature (or features) of the conduct relied on by the decision-maker that is genuinely separate from the making of the protected disclosure itself. In such cases, as Underhill LJ observed in Page (Footnote: ³), the protected disclosure is the context for the impugned treatment, but it is not the reason itself.

57.

Thus the ‘separability principle’ is not a rule of law or a basis for deeming an employer’s reason to be anything other than the facts disclose it to be. It is simply a label that identifies what may in a particular case be a necessary step in the process of determining what as a matter of fact was the real reason for impugned treatment. Once the reasons for particular treatment have been identified by the fact-finding tribunal, it must evaluate whether the reasons so identified are separate from the protected disclosure, or whether they are so closely connected with it that a distinction cannot fairly and sensibly be drawn. Were this exercise not permissible, the effect would be that whistle-blowers would have immunity for behaviour or conduct related to the making of a protected disclosure no matter how bad, and employers would be obliged to ensure that they are not adversely treated, again no matter how bad the associated behaviour or conduct.”

Simler LJ went on at para. 58 to refer to the decision of the EAT in Martin v Devonshires Solicitors UKEAT/86/10, [2011] ICR 352, in which I had said:

“Employees who bring complaints often do so in ways that are, viewed objectively, unreasonable. It would certainly be contrary to the policy of the anti-victimisation provisions if employers were able to take steps against employees simply because in making a complaint they had, say, used intemperate language or made inaccurate statements. An employer who purports to object to ‘ordinary’ unreasonable behaviour of that kind should be treated as objecting to the complaint itself, and we would expect tribunals to be slow to recognise a distinction between the complaint and the way it is made save in clear cases.”

The statement in that passage that the employer who purports to object to “ordinary” unreasonable behaviour “should be treated as objecting to the complaint itself” might suggest some kind of rule of law, but Simler LJ emphasised that that was not the case. The significance of the disproportionality of an employer’s response was evidential only: that is, as evidence that their motivation was in fact the protected disclosure itself and not to the manner in which it was made (see para. 60).

Elisabeth Laing LJ in her concurring judgment in Kong expressed some hesitation about references to “the separability principle”, given that, as Simler LJ had herself said, it did not connote a rule of law. I share that hesitation, but some label is required: I will in this judgment refer simply to “separability” or “the separability approach”.

The protection of the right of free speech, including speech expressing a person’s religious or other beliefs, has always been regarded as a cardinal principle of the common law, and it is of course now also protected by the incorporation by the 1998 Act of articles 9 and 10 of the Convention. There are many decisions of the highest authority expounding the relevant principles, but I do not need to recapitulate them here. I only note three points to which Mr O’Dair, and the FSU in its written submissions, attached particular importance.

First, freedom of speech necessarily entails the freedom to express opinions that may shock and offend. The most authoritative statement to this effect is probably that of the ECtHR at para. 46 of its judgment in Vajnai v. Hungary [2008] ECHR 1910, where it said:

“The Court further reiterates that freedom of expression, as secured in Article 10 §1 of the Convention, constitutes one of the essential foundations of a democratic society and one of the basic conditions for its progress and for each individual’s self-fulfilment. Subject to Article 10 §2, it is applicable not only to ‘information’ or ‘ideas’ that are favourably received or regarded as inoffensive or as a matter of indifference, but also to those which offend, shock or disturb; such are the demands of pluralism, tolerance and broad-mindedness, without which there is no ‘democratic society’ … . Although freedom of expression may be subject to exceptions, they ‘must be narrowly interpreted’ and ‘the necessity for any restrictions must be convincingly established’ (see, for instance, Observer and Guardian v the United Kingdom, 26 November 1991, §59, Series A no. 216).”

A very frequently-cited domestic authority to the same effect is Redmond-Bate v Director of Public Prosecutions [1999] EWHC Admin 733, where Sedley LJ, sitting with Collins J in the Divisional Court said, at para. 20:

“Free speech includes not only the inoffensive but the irritating, the contentious, the eccentric, the heretical, the unwelcome and the provocative … Freedom only to speak inoffensively is not worth having  ...”

Second, the protection of freedom of speech is particularly important in the case of “political speech” – that is, expression of opinion on matters of public and political interest. At para. 47 of its judgment in Vajnai the ECtHR stressed “that there is little scope under Article 10 §2 of the Convention for restrictions on political speech or on the debate of questions of public interest”.

Third, in any given case it is important to be alive not just to the effect of restrictions on freedom of speech in that case but to their chilling effect more widely. In R (Miller) v College of Policing [2021] EWCA Civ 1926, [2022] 1 WLR 4987, Sharp P said, at para. 68:

“The concept of a chilling effect in the context of freedom of expression is an extremely important one. It often arises in discussions about what if any restrictions on journalistic activity are lawful; but in my judgment it is equally important when considering the rights of private citizens to express their views within the limits of the law, including and one might say in particular, on controversial matters of public interest.” 

These are principles which any court or tribunal must have at the forefront of its mind in considering a case involving freedom of speech, including the expression of religious or other beliefs. It should be noted, however, that in each of those cases the Court was concerned with limitations on free speech imposed by a public authority. The present case is concerned with an interference with free speech on the part of an employer against an employee, and it is necessary to assess whether the interference was justified in the context of the employment relationship and the law applicable to it. The relevant principles in such a case were considered by this Court in the case of Page v NHS Trust Development Authority [2021] EWCA Civ 255, [2021] ICR 941, which was central to the decision of the EAT and to which I now turn.

The claimant in Page was a non-executive director of an NHS Trust whose role was terminated by the respondent authority (Footnote: ⁴) because in media interviews he expressed controversial views, derived from his Christian beliefs, about the morality of homosexual acts and about same-sex marriage and same-sex adoption. He brought claims in the ET of direct and indirect discrimination on the ground of religion or belief: the beliefs in question were characterised in the alternative either as Christianity tout court or as “belief in the traditional family”. It was accepted that this was a genuine manifestation case: the authority was not motivated by the fact that the claimant held those beliefs. The ET dismissed his complaint, and that decision was upheld both by the EAT and by this Court, in which I gave the leading judgment with which Peter Jackson and Simler LJJ agreed. We are only concerned with the reasoning as regards the direct discrimination claim.

The main thrust of the claimant’s argument was that his treatment had constituted a breach of his rights under articles 9 and 10 of the Convention. We upheld the decision of the EAT, applying the decision of this Court in Mba v London Borough of Merton [2013] EWCA Civ 1562, [2014] ICR 357, that the ET had no jurisdiction to entertain a claim for breach of Convention rights as such, and accordingly that the claim could only be advanced under the 2010 Act. The claimant’s counsel nevertheless argued that it was necessary in every case of belief discrimination under the 2010 Act to start by considering whether there had been a breach of the claimant’s Convention rights. As to that, I said, at para. 37:

“I do not think that there needs to be any such rule. It is, ultimately, the Act from which the claimant’s rights must derive, and there can be nothing wrong in a tribunal taking that as the primary basis of its analysis.” 

Despite that, we found it more convenient, because of the focus of the argument, to address first the claimant’s case that his Convention rights had been breached. The relevant part of my judgment is at paras. 38-67. We held that although there had been an interference with his right to manifest his beliefs under article 9 (and also his article 10 rights) the ET had been entitled to find that his termination was justified: see paras. 52-63. I need not summarise our reasons, which are specific to the facts of the case. At the end of this section of the judgment, having concluded that the ET had been entitled to find that the claimant’s Convention rights had not been infringed, I continued (at para. 67):

“It might be thought to follow that [the authority] cannot have discriminated against him on the grounds of his religion or belief, since the relevant protections under the Convention and the 2010 Act must be intended to be co-extensive. In my view that is indeed the case, but that does not absolve me from considering the issues through the lens of the 2010 Act, which must be the formal basis of the Appellant’s claim.”

Thus, although I repeated that the claim had to be based on the 2010 Act, I expressed the view that his rights under the Act were “intended to be co-extensive” with his Convention rights.

The claim of direct discrimination contrary to the 2010 Act is considered at paras. 68-80 of the judgment. The dispositive issue is treated as being the applicability of the separability approach. At para. 68 I said:

“In a direct discrimination claim the essential question is whether the act complained of was done because of the protected characteristic, or, to put the same thing another way, whether the protected characteristic was the reason for it … It is thus necessary in every case properly to characterise the putative discriminator’s reason for acting. In the context of the protected characteristic of religion or belief the EAT case-law has recognised a distinction between (1) the case where the reason is the fact that the claimant holds and/or manifests the protected belief, and (2) the case where the reason is that the claimant had manifested that belief in some particular way to which objection could justifiably be taken. In the latter case it is the objectionable manifestation of the belief, and not the belief itself, which is treated as the reason for the act complained of. Of course, if the circumstances (Footnote: ⁵) are not such as to justify the act complained of, they cannot sensibly be treated as separate from an objection to the belief itself.”

I went on at para. 69 to identify, and thus approve, the EAT cases referred to in that passage, being Chondol v Liverpool City Council [2009] UKEAT 0298/08 (a decision of my own), Grace v Places for Children [2013] UKEAT 0217/13 (Mitting J), and Wasteney v East London NHS Foundation Trust [2016] UKEAT 0157/15, [2016] ICR 643 (HH Judge Eady QC (as she then was)). These were all cases in which dismissal for inappropriate Christian proselytisation at work was held to be on a ground separable from “religion or belief”. Wasteney is the most fully reasoned. At paras. 54-55 of her judgment Judge Eady said:

“54.

In domestic law, the expression of right and limitation – as allowed by Article 9 of the Convention – is most easily discernible when addressing cases of indirect discrimination under section 19 EqA (which may be the more obvious route of challenge in most cases involving the manifestation of a religious belief).  Whilst there is no statutory means of ‘justifying’ direct discrimination or harassment, however, the Claimant accepts that the limitations permitted by Article 9.2 are relevant to the approach to be adopted to claims brought under sections 13 (direct discrimination) and 26 (harassment).  Although the Claimant relies on the protection of the right to manifest religious belief in the workplace, as recognised by the ECHR in Eweida, she (correctly) does not seek to suggest that right cannot be subject to limitation.

55.

The concession is in some senses easier to state than apply, but the task will always be made easier by having a clear understanding of the nature of the claim and how it is being put.  If the case is one of direct discrimination then the focus on the reason why the less favourable treatment occurred should permit an ET to identify those cases where the treatment is not because of the manifestation of the religion or belief but because of the inappropriate manner of the manifestation (where what is ‘inappropriate’ may be tested by reference to Article 9.2 and the case-law in that respect); see [Chondol] and [Grace].  Similarly, whilst the definition of harassment permits the looser test of ‘related to’, a clear sense of what the conduct did in fact relate to should permit the ET to reach a conclusion as to whether it is the manifestation of religion or belief that is in issue or whether it is in fact the complainant’s own inappropriate conduct (and that must be right, otherwise an employer’s attempt to discipline an employee for the harassment of a co-worker related to (e.g.) the co-worker’s religion or belief could itself be characterised as harassment related to that protected characteristic).”

That reasoning, which I evidently approved, explicitly treats the limitations in article 9.2 of the Convention as “relevant to” a claim of discrimination (or harassment) under the 2010 Act. I noted that Judge Eady had referred to the distinction as being between the manifestation of the religion or belief and “the inappropriate manner” of its manifestation; I described that as an acceptable shorthand, “as long as it is understood that the word ‘manner’ is not limited to things like intemperate or offensive language”.

I note in passing that Chondol and Wasteney had in fact already been approved by this Court in Kuteh v Dartford and Gravesham NHS Trust [2019] EWCA Civ 818, though this decision does not appear to have been cited to us in Page. At para. 64 of his judgment Singh LJ referred to them as setting out “an important principle, namely the distinction between the manifestation of a religious belief and the inappropriate promotion of that belief, which in turn reflects the jurisprudence of the European Court of Human Rights”.

At paras. 70-72 I analysed the ET’s reasoning. After dealing with one difficulty about how it had expressed itself, I said, at para. 72:

“Once that point has been clarified, the Tribunal’s reasoning is clear. Para. 71 applies the distinction which I have discussed at paras. 68-69 above. The Authority took disciplinary action against the Appellant not because he was a Christian or because he held the traditional family belief but because he expressed the latter belief (and his other views about homosexuality) in the national media in circumstances which, on the Tribunal’s findings, justified the action taken.”

At para. 74 I upheld the validity in principle of that distinction. I said:

“So far as I am aware the distinction applied by the Tribunal has not been endorsed in this Court, but it is in my view plainly correct. It conforms to the orthodox analysis deriving from Nagarajan: in such a case the ‘mental processes’ which cause the respondent to act do not involve the belief but only its objectionable manifestation. An analogous distinction can be found in other areas of employment law – see paras. 19-21 of my judgment in Morris v Metrolink RATP DEV Ltd [2018] EWCA Civ 1358, [2019] ICR 90 (Footnote: ⁶). Also, and importantly, although it gets there by a different route (because the provisions in question are drafted in very different ways), the recognition of that distinction in the application of section 13 achieves substantially the same result as the distinction in article 9 of the Convention between the absolute right to hold a religious or other belief and the qualified right to manifest it. It is obviously highly desirable that the domestic and Convention jurisprudence should correspond.”

The last two sentences are to essentially the same effect as para. 67 of the judgment.

At paras. 75-79 I addressed various particular arguments advanced on behalf of the claimant which I need not consider here. At para. 80 I concluded that there was no error of law in the ET’s decision on direct discrimination.

In summary, Page was decided on the basis that adverse treatment in response to an employee’s manifestation of their belief was not to be treated as having occurred “because of” that manifestation if it constituted an objectively justifiable response to something “objectionable” in the way in which the belief was manifested: it thus introduced a requirement of objective justification into the causation element in section 13 (1). Further, we held that the test of objective justification was not substantially different from that required under article 9.2 (and also article 10.2) of the Convention. I should clarify two points about language:

First, my formulation does not directly apply the four-step process identified in Bank Mellat, but it is a compressed version of the same exercise, involving (a) the identification of a feature of the employee’s conduct to which the employer could legitimately object (broadly corresponding to step (1)), and (b) an assessment of whether the employer’s response to that feature was proportionate (broadly corresponding to steps (2)-(4)). It is no doubt best practice to consider each of the Bank Mellat steps separately, but it is well recognised that there is a considerable degree of overlap between them.

Second, the equation of the applicable test with that under article 9.2 of the Convention appears to bring in not only the test of objective justification but also the requirement that the act in question be “prescribed by law”. The School sought to object to this element in its Respondent’s Notice (see para. 117 below), but since permission was refused the issue was not live before us. However, even if, absent Page, it would be unnecessary to import this element, I cannot see that it causes any conceptual problem in this context: the employer’s rights under the employment contract provide the necessary framework of “law”, in the sense in which that term is used paragraph 2 of articles 9 and 10.

Third, the burden of proof of objective justification is on the employer. I make this point in particular because in her reasons for giving permission to appeal Elisabeth Laing LJ expressed a concern that a consequence of the reasoning in Page might be that employees had to demonstrate that the treatment of which they complained was not in accordance with the law or proportionate to a legitimate aim. I do not believe that to be the case: as a matter of general principle a justification for interfering with a qualified Convention right must be proved by the party relying on it.

Fourth, although Page imports a test of objective justification into the separability approach, it does so only because of the protection conferred on the right to manifest a religious belief conferred by the Convention. It has no impact on the application of the separability approach in other cases.

Fifth, as regards a claim of harassment, section 26 of the 2010 Act requires the treatment to be “related to” the protected characteristic, rather than “because of” it as in section 13 (1). It was not suggested in argument before us that that difference renders the ratio of Page inapplicable in harassment cases, and I do not believe that it does. (Footnote: ⁷)

I accept that my judgment in Page does not clearly explain the jurisprudential basis of the decision that an objective element could be introduced into section 13 (1) of the 2010 Act. That may to some extent reflect the nature of the arguments before us in that case, and I hope that it may to that extent be venial. But in this appeal we have received extensive and helpful submissions, both from the parties and from the interveners, exploring the basis on which Page was, or must be taken to have been, decided. I believe that we should take advantage of the assistance which we have received in order to try to explain the reasoning more fully. Anything I say will be obiter, for one or both of two reasons. First, whatever the underlying reasoning, we are bound by the ratio in Page as identified above (a point which not all of the interveners appeared to appreciate). Second, even if it were otherwise open to us to reformulate that ratio, the EAT, as appears below, clearly decided the appeal by applying it and neither party has permission to challenge that aspect of its decision. But I think the exercise is worthwhile, not least because the interveners attach importance to having the most thorough understanding possible of how the law works in this area.

A conceptual underpinning for the ratio in Page might be thought to be most readily provided by section 3 of the 1998 Act – that is, that the incorporation of the test of objective justification in article 9.2 is necessary in order to render section 13 (read with section 4) compatible with Convention rights. That was the view of Judge Eady in Wasteney: see para. 48 of her judgment. Mrs Fraser Butlin advocated this analysis, relying by way of analogy on para. 57 of the judgment of Mummery LJ in X v Y, [2004] EWCA Civ 662, [2004] ICR 1634, where he referred to section 3 as enabling the “blending” of the requirements of the unfair dismissal legislation with the protection of employees’ rights under article 8 of the Convention.

However section 3 was not expressly relied on in Page itself, and the point is not as straightforward as it may appear at first sight. The section only operates so far as necessary to render the statutory provision in question “compatible with the Convention rights” – that is, with the rights of the persons who enjoy those rights (in our case persons wishing to manifest a religious belief). On the face of it, the protection conferred by section 13 is perfectly compatible with employees’ Convention rights: it may, because it is unqualified, go further than the Convention requires, and so place additional obligations on the employer, but that is not the same thing.

I initially thought that it followed that section 3 has no application in this case. On further consideration, however, I believe that a more sophisticated argument may be available, as follows:

On balance I think that that argument is correct and accordingly that section 3 does indeed provide a satisfactory basis for the ratio in Page. However, I acknowledge that it is not straightforward, and I should accordingly say that I believe that it can also be justified on ordinary principles of domestic construction, as summarised at paras. 85-88 below.

The starting-point is that the Court in Page believed that the legislature cannot have intended that an employer should be obliged to tolerate any conduct at all by an employee which constituted a manifestation of a belief, whatever form it took and whatever the circumstances; and that is reinforced by the fact that the drafters of the Convention thought it necessary to qualify the right to manifest a belief. It is not necessary to multiply examples of cases where an inability to prohibit particular kinds of manifestation would produce an obviously unacceptable result. One familiar case is the wearing of clothes or jewellery which would create a serious risk to health or safety at work. Another is that of religious proselytisation at work, as illustrated by the cases noted at para. 70 above: in that connection, I note that A-G Sharpston observed at para. 73 of her Opinion in Bougnaoui that proselytisation “has … simply no place in the work context”. It follows that it is in accordance with the legislative purpose to construe the Act, so far as possible, so as to incorporate some such limitation.

The next question is what the parameters of such a limitation should be. The Court’s view in Page was that the qualification which best achieved the legislative purpose was to permit a defence of objective justification substantially corresponding to the terms of article 9.2 of the Convention. There are a number of reasons why that should be so, even without resort to section 3 of the 1998 Act. We have seen that the CJEU in Bougnaoui referred to the Convention as an aid to the construction of the Directive. More specifically, article 2.5 of the Directive legitimises in principle a qualification reflecting the terms of article 9.2: that is significant, even though no such qualification was expressly included in the 2010 Act. It is in fact hard to see how the presumed legislative purpose of limiting the right to manifest a religious belief to the extent necessary to protect the rights and freedoms of others could be achieved more appropriately than by the balancing exercise required under the Convention. There is nothing to be gained by searching for some different formulation, and there are obvious practical advantages in employing a test with which practitioners and tribunals are familiar. It would also be anomalous if, in proceedings brought by an employee of a public authority in the ordinary courts under section 7 of the 1998 Act for breach of their right under article 9 to manifest a belief, the applicable test were different from that which would apply if they brought proceedings in the ET under the 2010 Act.

It remains to identify how such a limitation could operate within the structure of the provisions of the 2010 Act. The Court in Page believed, drawing on the separability case-law, that that was best done by treating it as going to the requirement of causation in section 13 (1). That is, where the act complained of was objectively justified it should not be treated as being done “because of” the manifestation in question.

A highly purposive construction of this kind is not objectionable in principle where the Court is satisfied that it is truly necessary. But the objections to implying a qualification which the legislature has failed to express may be less cogent in this case since the protection of the manifestation of belief is itself not express but is, as appears from Bougnaoui, the product of a “purposive” choice to prefer a wider interpretation of the words chosen by the legislature.

Very broadly, though not in detail, those explanations of the ratio in Page correspond to the submissions of Mr O’Dair for the Claimant and Mrs Fraser Butlin for the Archbishops’ Council. But Mr Jones and Mr Milsom for the School and counsel for the other interveners identified respects in which that ratio is said to be problematic or suggested ways in which it could be supported by better reasoning. Since, for the reasons already given, it is not open to us to decide this appeal on any different basis I do not propose to consider these submissions in detail, but I will briefly review them.

The fundamental objection advanced to the approach in Page is that it is said to undermine an essential feature of the law of direct discrimination. It is well established, at least in EU and UK law, that direct discrimination cannot generally be justified. That feature cannot, it is said, be circumvented by incorporating an element of objective justification into the requirement that the discrimination be “on the grounds of” the protected characteristic: it is well established that if the subjective mental processes of the putative discriminator had nothing to do with the protected characteristic they cannot be liable, however unreasonable or unfair the treatment in question may have been (see, classically, the judgment of Elias J in Law Society v Bahl [2003] UKEAT 1056/01/3107, [2003] IRLR 640, at paras. 93-101). Likewise, in separability cases the fact that an employer’s response to some objectionable feature of the protected conduct is disproportionate is relevant only if and to the extent that it supports a finding that that was not the real reason for the impugned act: see the observations of Simler LJ in Kong referred to at para. 59 above.

In support of this objection we were referred to paras. 58-67 of the Opinion of A-G Sharpston in Bougnaoui. In those paragraphs, which are headed “The differences between a restrictions-based approach and one based on discrimination”, she rejects as “simplistic” the suggestion that the requirements of the Strasbourg and EU jurisprudence as regards direct discrimination should be “blended” so as to allow the adoption of what she acknowledges to be the more flexible approach under article 9 of the Convention. She concludes, in para. 67:

“The distinction between [direct and indirect discrimination] is a fundamental element of this area of EU legislation. There is in my view no reason to depart from it, with the inevitable loss of legal certainty that would result.”

It is fair to say that that passage occurs in a part of her Opinion not directly concerned with the issue in the particular case and is not expressly, or so far as I can see implicitly, adopted in the judgment of the Court; but it remains of obvious persuasive authority.

I see the force of that objection, but if the incorporation of an objective test is required by section 3 of the 1998 Act it cannot prevail. And even if section 3 is not engaged, I do not in fact think that it is unanswerable. In the first place, there is nothing axiomatically objectionable in the proposition that direct discrimination may be capable of justification: that is the case under article 14 of the Convention, and the justification of direct discrimination is also permitted by the Framework Directive in the case of age discrimination and under article 2.5. Direct discrimination in manifestation cases is (uniquely) different from discrimination on the ground of other protected characteristics (and indeed from simple belief discrimination) because it is based, as the Court in Eweida identifies, not on the possession of the characteristic as such but on overt conduct, which thus has the potential to impact on the interests of society and the rights and freedoms of others. That distinction may be said to put it in a special category which requires a more flexible approach. As I have said, I find it hard to accept that the legislature intended employees to enjoy an absolute right not to suffer any adverse treatment on the basis of conduct manifesting their religious or other beliefs, whatever the nature of that conduct and whatever the circumstances. (Footnote: ⁸)

The submissions on behalf of the School, and of those of the interveners who addressed the point, acknowledged that it was important that employers should be entitled to prohibit or punish objectionable or inappropriate manifestations of religious or other belief; but it was contended that this could be achieved in alternative ways which did not do violence to the important principles of discrimination law identified above. These alternatives were essentially threefold. I take them in turn.

The first was to contend that in most if not all instances where an employee suffers a detriment as the result of manifesting a religious belief the case can properly be characterised as one of indirect discrimination, in which case the employer can defeat the claim by showing that the impugned PCP was objectively justified. There are of course many manifestation cases where what the employee is complaining of is the effect of an apparently neutral PCP; indeed such cases may reflect the majority of claims brought, as Judge Eady suggests in Wasteney (see para. 54 of her judgment). But that point goes nowhere unless all cases of manifestation discrimination can be properly so analysed. I understood Mr Jones and Ms Clement to go so far as to contend that that was the case. I find that hard to accept. Take a case where an employee is disciplined for Christian proselytisation in the workplace. Even if it could be shown that the employer applied a PCP prohibiting such proselytisation (which might not be easy in the absence of an express rule) I do not see how the impact of such a PCP could be described as apparently neutral: it would be direct discrimination against anyone wishing to manifest their Christian faith. Chondol, Grace and Wasteney were all advanced as cases of direct discrimination, as indeed was the present case; and in my view that reflects a correct understanding. In this connection it is important to appreciate that whether a claim can be pleaded as one of direct or indirect discrimination is not a matter for the claimant’s choice: as Lady Hale made clear in the JFS case (see para. 57 of her judgment), direct and indirect discrimination are mutually exclusive.

The second suggested way of achieving consistency with recognised principle was to rely on the conventional separability approach. It was acknowledged that this approach would not afford protection in cases where the treatment complained of was genuinely because of the way the employee had manifested their belief but was disproportionate. But it was suggested that such cases would be rare, and that the difference in practice between the protection afforded by the two tests is accordingly slight. It is impossible to know whether that is the case.

The third alternative was to import the requirement of objective justification into the definition of the protected characteristic: that is, an objectionable manifestation of a belief should not be treated as falling within the phrase “religion or belief” in section 4 (cf. para. 83 (5) above). This approach has the attraction that it does not compromise the traditional subjective approach to the question of causation. But, like the second alternative, it does not on the face of it provide protection against disproportionate responses.

I have briefly reviewed these alternative approaches out of deference to the thought that went into the submissions before us. As I have said, the short answer to all of them is that they are inconsistent with the ratio of Page. But I should say that in my view none of them is obviously preferable to that ratio. I also believe that it is an advantage of the Page test that it uses legal concepts to which tribunals are well used, not least because they are familiar with applying them in cases of indirect discrimination.

The ET’s Reasons for dismissing the claim are presented carefully and systematically. I can summarise them sufficiently for our purposes as follows.

Paras. 1-7 deal with various interlocutory matters. Paras. 8-25 identify the witnesses from whom the Tribunal heard and contain its findings of fact. I need not add to what I have already summarised above, though I should note (because it is relevant to one of the grounds of appeal to the EAT) that Mr Conlan gave evidence but the other members of the panel did not. Paras. 26-32 summarise the relevant provisions of the 2010 Act and identify the nature of the Claimant’s claim (including her identification of the beliefs on which she relied as set out at para. 27 above).

At paras. 34-45 the Tribunal considers a submission by the School that the Claimant’s views about gender fluidity did not qualify for protection under section 10 of the 2010 Act because they did not satisfy the fifth of the requirements identified in the well-known judgment of Burton P in Nicholson v Grainger Plc [2009] UKEAT 0219/09/0311, [2010] ICR 360 – that is, that in order to qualify for protection a belief “must be worthy of respect in a democratic society, be not incompatible with human dignity and not conflict with the fundamental rights of others”. The Tribunal rejected that submission and held that the beliefs expressed in the Claimant’s posts constituted a protected characteristic. Its decision in that regard was subsequently vindicated by the decisions of the EAT in Forstater v GCD Europe UKEAT/0105/20, [2022] ICR 1, and Mackereth v Department for Work and Pensions [2022] EAT 99, [2022] ICR 1609, and it is not challenged by the School.

At para. 46 the Tribunal explains that it will address the claim of discrimination before considering the harassment claim. At paras. 48-52 it considers and rejects a preliminary submission by the Claimant that the School was not entitled to take any action against her because her Facebook posts were private.

Paras. 53-67 contain the ET’s reasoning on the direct discrimination claim. The gist appears at paras. 57-64. Paras. 58-60 focus on the language of the posts, which the Tribunal describes as “florid and provocative”, and not on the substance of the views expressed. At para. 60 it finds that the School believed that a reader of the posts

“might conclude that someone who associated herself with such a post (as [the Claimant] had done) not only felt strongly that gender fluidity should not be taught in schools but was also [my emphasis] hostile towards the LGBT community and trans people in particular”.

It thus found, at para. 61, that the act of which the Claimant was accused and found guilty was “posting items on Facebook that might reasonably lead people who read her posts to conclude that she was homophobic and transphobic”, which the School felt “had the potential for a negative impact in relation to... pupils, parents, staff and the wider community”. At para. 62 it quotes, and accepts, Mr Conlan’s oral evidence to the effect that “… had [the Claimant’s] beliefs been simply stated on her Facebook page in the form which they appear in paragraph 30 above [para. 27 of this judgment] no further action could or would have been taken against her”. Paras. 63-64 read:

“63.

We concluded that not only the dismissal but the entire proceedings taken against Mrs Higgs were motivated by a concern on the part of the School that, by reason of her posts, she would be perceived as holding unacceptable views in relation to gay and trans people – views which in fact she vehemently denied that she did hold.

64.

In short, that action was not on the ground of the beliefs but rather for a completely different reason, namely that as a result of her actions she might reasonably be perceived as holding beliefs that would not qualify for protection within the Equality Act (and, as we say, beliefs that she denied having).”

It is worth quoting also what the Tribunal went on to say at paras. 65-66:

“65.

It is important to bear in mind that this was not a claim of unfair dismissal. We were not concerned to decide whether the School’s actions were reasonable or not. It might be contended that there was a different course of action the school could have taken, in the light of the position made clear by Mrs Higgs in the disciplinary process. Since she denied being homophobic or transphobic, a reasonable employer might have taken the view that justice would be served by her (or the School) making it clear that if anyone thought she held those views they had got ‘the wrong end of the stick’ – that pupils and parents should not be concerned that she would demonstrate any sort of hostility to gay or trans pupils (or indeed gay or trans parents).

66.

That was not a subject canvassed before us, for the simple reason that it was irrelevant to our considerations. Our only task was to decide if there was a causal connection between the beliefs in paragraph 30 and the treatment meted out to Mrs Higgs.”

It is reasonably clear from that passage that the Tribunal thought it strongly arguable, to put it no higher, that the School’s treatment of the Claimant had been disproportionate.

At paras. 69-75 the Tribunal rejected the harassment claim for essentially the same reasons. Paras. 70-72 read:

“70.

It was possible to see some sort of connection between her beliefs and [the treatment complained of]. The posts in question clearly expressed those beliefs both in relation to same sex marriage and gender fluidity. However, as we have said, her treatment was not a consequence of her expressing those beliefs in a temperate and rational way. Rather, it was because the School felt that the language used in those posts might reasonably lead someone who read them to conclude that she held views (homophobic and transphobic) that she expressly rejected [emphasis supplied].

71.

The essence of the protection from harassment is that a claimant should be entitled to hold and express protected views without being mistreated as a consequence. It was not the protected views of Mrs Higgs that resulted in the disciplinary action but rather the School’s conclusion that her action in posting the items in question might reasonably (and in fact did) lead others to conclude that she held wholly unacceptable views.

72.

It follows that we also conclude that the causal nexus between the protected characteristic and the actions of the school was not made out. The School’s behaviour was not related to the relevant beliefs and it followed that the claim of harassment was not made out.”

In short, the Tribunal’s reasoning was as follows:

The Claimant’s eventual grounds of appeal from the ET to the EAT were as follows:

“(1)

The ET erred in law in failing to consider proportionality of the Respondent’s interference with the Appellant’s manifestation of her religious/philosophical beliefs.

(2)

The ET erred in law in failing to consider whether the interference with the Claimant’s Convention rights was ‘prescribed by law’.

(3)

The ET erred in law in holding that the employer could lawfully restrict the Appellants right to freedom of speech to the language of an ET pleading: see (ET 30 and 62).

(4)

The ET reached an impermissible conclusion and/or failed to properly explain its reasons for attributing Mr Conlan’s reasons to all other decision-makers; alternatively, misdirected itself in identification of the relevant decision-makers.

(5)

The ET erred in law in finding that the Respondent did not discriminate against the Claimant when it investigated and/or dismissed her by reason of the complainant’s objection to the Claimant’s beliefs.

(6)

The ET erred in law in finding that it was reasonable for third parties reading the Claimant’s posts to conclude that she was homophobic or transphobic. Alternatively, that finding is perverse.

(7)

The ET’s finding that the reason for dismissal was (or was solely) because of the views of third parties about the posts rather than the School’s own views about those posts (ET 60, 61) results from the following errors of law:

(a)

an employer cannot escape liability by pointing to pressure from a third party whose own motivation was discriminatory (Din v Carrington Viyella);

(b)

stereotyping a protected characteristic is a discriminatory reason (Aylott [2010] ICR 1278);

(c)

alternatively, this finding is perverse.”

The reasons why the EAT allowed the Claimant’s appeal appear from paras. 81-84 of Eady J’s judgment. These read (the emphases in paras. 82 and 83 are mine):

“81.

Returning then to the ET’s finding as to the reason for the respondent’s actions in this case, it stated that it considered that this was not because of, or related to, the claimant’s actual beliefs but because of the concern that her posts might be seen as evidence that she held other beliefs, which might be described as ‘homophobic’ or ‘transphobic’.  Putting to one side the dangers that can arise from the use of labels that might mean different things to different people (see the discussion at paragraph 250 R (oao Miller) v College of Policing and anor [2020] EWHC 225 (Admin), and the observations of Underhill LJ at paragraph 18 Page v NHS Trust Development Authority [2021] EWCA Civ 255), the difficulty with the ET’s analysis is that it did not engage with the question whether this was, nonetheless, because of, or related to, the claimant’s manifestation of her beliefs.  In answering that question, the views or concerns of the respondent were not relevant (Page, paragraph 49); applying the test laid down at paragraph 82 Eweida v UK (2013) 57 EHRR 8, the ET needed to consider whether there was a sufficiently close or direct nexus between the claimant’s protected beliefs and her posts (relied on by her as amounting to a manifestation of those beliefs).

82.

To the extent that the ET addressed the question identified in Eweida, it is apparent that it did so through the prism of the respondent’s view of the claimant’s posts.  The respondent’s views were relevant when determining whether there had in fact been any interference with the claimant’s right to manifest her beliefs and to freedom of expression - whether its treatment of her was because of, or related to, her exercise of those rights - but could not be determinative of the prior question, whether there was a sufficiently close or direct link between the claimant’s posts and her beliefs such as to mean that those posts were to be viewed as a manifestation of her beliefs.  If they were, then the ET needed to determine the ‘reason why’ question by asking itself whether this was because of, or related to, that manifestation of belief (prohibited under the EqA), or whether it was in fact because the claimant had manifested her belief in a way to which objection could justifiably be taken.  As was made clear in Page (see paragraph 68), in the latter case, it is the objectionable manifestation of the belief that is treated as the reason for the act complained of.  In order to determine whether or not the manifestation can properly be said to be ‘objectionable’, however, it is necessary to carry out a proportionality assessment: keeping in mind the need to interpret the EqA consistently with the ECHR, there can be nothing objectionable about a manifestation of a belief, or free expression of that belief, that would not justify its limitation or restriction under articles 9(2) or 10(2) ECHR (and see Page at paragraph 74; Wasteney v East London NHS Foundation Trust [2016] ICR 643 at paragraph 55).

83.

As the respondent acknowledged in its oral submissions, the ET’s reasoning demonstrates that, had it properly engaged with the Eweida question, it would have concluded that there was a close or direct nexus between the claimant’s Facebook posts and the beliefs that she had relied on in her claims: as it stated, ‘The posts in question clearly expressed those beliefs’ (ET, paragraph 70).  That did not mean that it was bound to find that the respondent’s actions necessarily amounted to direct discrimination or harassment, but, in determining the reason for the treatment complained of, the ET needed to assess whether those actions were prescribed by law, and were necessary for the protection of the rights and freedoms of others.  And, in carrying out that assessment, the ET needed to first recognise the essential nature of the claimant’s right to freedom of belief and to the freedom to express that belief (a recognition that must carry with it an understanding of the foundational nature of those rights for any democracy; see Sahin v Turkey (2007) 44 EHRR 5 and Handyside v UK 1 EHRR 737), before undertaking the proportionality assessment laid down in Bank Mellat v HM Treasury (No 2) [2014] AC 700 (see paragraph 54 above).

84.

The problem with the ET’s approach is that it by-passed any engagement with the nature of the claimant’s rights, and failed to carry out the requisite balancing exercise, when seeking to determine whether the mental processes which caused the respondent to act did not involve the claimant’s beliefs but only their objectionable manifestation.  As the claimant objects (ground 1 of the appeal), the ET’s approach meant that it impermissibly narrowed the task it had to undertake.  It was not enough to find that the respondent had been motivated by a concern that the claimant could be perceived to hold ‘wholly unacceptable views’ (ET, paragraph 70); the ET needed to consider whether that motivation or concern had arisen out of the claimant’s manifestation of her beliefs (accepted to be protected under the EqA) or by a justified objection to that manifestation.”

Eady J thus proceeded on the basis that Page had established that, in a case where, as here, the treatment complained of was in response to the manifestation of a protected belief, the question whether that manifestation was the reason for the treatment involved the application of a test of objective justification corresponding to that in article 9 (and article 10) of the Convention: see in particular the passages which I have italicised. The decision of the ET was overturned because it had not applied any such test. Because permission to cross-appeal was refused (see paras. 118-119 below) that self-direction is not challenged before us, but it will appear from my analysis of Page above that I believe that it was right.

Those reasons essentially correspond, as Eady J observes, to ground 1 of the Claimant’s grounds of appeal to the EAT. At para. 85 she rejects ground 3, but she says that she sees force in the point made in grounds 5-7.

Paras. 86-88 of Eady J’s judgment are essentially ancillary, and I need not set them out. However, I should refer to paras. 89-90. In those paragraphs she acknowledges that para. 65 of the ET’s Reasons suggests that it believed that the Claimant’s dismissal was disproportionate, but she says that there is no clear finding to that effect. She continued, at para. 91:

“While, therefore, the appeal should be allowed, this is not a case where it can properly be said that only one outcome is possible, and the appropriate disposal must be for this matter to be remitted for determination (Jafri v Lincoln College [2014] EWCA Civ 449).  That remission should be on the basis that it has already been found that the Facebook posts in issue had a sufficiently close or direct nexus with the beliefs relied on by the claimant in these proceedings such as to amount to a manifestation of those beliefs (per Eweida).  It will, however, be for the ET on the remitted hearing to determine, recognising the essential nature of the claimant’s rights to freedom of belief and freedom of expression: (1) whether the measures adopted by the respondent were prescribed by law; and, if so, (2) whether those measures were necessary in pursuit of the protection of the rights, freedoms or reputation of others.  Undertaking that analysis will enable the ET to determine whether the respondent’s actions were because of, or related to, the manifestation of the claimant’s protected beliefs, or were in fact due to a justified objection to the manner of that manifestation, in respect of which there was a clear legal basis for the claimant’s rights to freedom of belief and expression to be limited to the extent necessary for the legitimate protection of the rights of others.”

Para. 1 of the EAT’s order reflects the contents of para. 91 of the judgment, though the material parts are not identically worded. It reads:

“The appeal be allowed and this matter remitted (in accordance with the reasons provided in the Judgment handed down this day) for the determination of the question whether the respondent’s actions were because of, or related to, the manifestation of the claimant’s protected beliefs, or were due to a justified objection to the manner of that manifestation, in respect of which there was a clear legal basis for the claimant’s rights to freedom of belief and expression to be limited to the extent necessary for the legitimate protection of the rights of others.”

Finally, at paras. 92-94 of the judgment Eady J considered what guidance she could properly give to the ET as regards its decision on the remitted questions.  Paras. 93-94 read:

“93.

For my part, I consider that a danger can arise from any attempt to lay down general guidelines in cases such as this.  Experience suggests that issues arising from the exercise of rights to freedom of religion and belief, and to freedom of expression, are invariably fact-specific.  Although the public debate around these issues tends to be conducted through the prism of categories and labels, that is not an approach that can properly inform the decisions taken in individual cases.  The values that underpin the right to freedom of religion and belief and of freedom of expression – pluralism, tolerance and broadmindedness (per Sahin v Turkey (2007) 44 EHRR 5; Handyside v UK 1 EHRR 737) – require nuanced decision-making; there is no ‘one size fits all’ approach. 

94.

All that said, I can see that, within the employment context, it may be helpful for there to be at least be some mutual understanding of the basic principles that will underpin the approach adopted when assessing the proportionality of any interference with rights to freedom of religion and belief and of freedom of expression.

(1)

First, the foundational nature of the rights must be recognised: the freedom to manifest belief (religious or otherwise) and to express views relating to that belief are essential rights in any democracy, whether or not the belief in question is popular or mainstream and even if its expression may offend.

(2)

Second, those rights are, however, qualified.  The manifestation of belief, and free expression, will be protected but not where the law permits the limitation or restriction of such manifestation or expression to the extent necessary for the protection of the rights and freedoms of others. Where such limitation or restriction is objectively justified given the manner of the manifestation or expression, that is not, properly understood, action taken because of, or relating to, the exercise of the rights in question but is by reason of the objectionable manner of the manifestation or expression. 

(3)

Whether a limitation or restriction is objectively justified will always be context-specific.  The fact that the issue arises within a relationship of employment will be relevant, but different considerations will inevitably arise, depending on the nature of that employment. 

(4)

It will always be necessary to ask (per Bank Mellat): (i) whether the objective the employer seeks to achieve is sufficiently important to justify the limitation of the right in question; (ii) whether the limitation is rationally connected to that objective; (iii) whether a less intrusive limitation might be imposed without undermining the achievement of the objective in question; and (iv) whether, balancing the severity of the limitation on the rights of the worker concerned against the importance of the objective, the former outweighs the latter.  

(5)

In answering those questions, within the context of a relationship of employment, the considerations identified by [the Archbishops’ Council] are likely to be relevant, such that regard should be had to: (i) the content of the manifestation; (ii) the tone used; (iii) the extent of the manifestation; (iv) the worker’s understanding of the likely audience; (v) the extent and nature of the intrusion on the rights of others, and any consequential impact on the employer’s ability to run its business; (vi) whether the worker has made clear that the views expressed are personal, or whether they might be seen as representing the views of the employer, and whether that might present a reputational risk; (vii) whether there is a potential power imbalance given the nature of the worker’s position or role and that of those whose rights are intruded upon; (viii) the nature of the employer’s business, in particular where there is a potential impact on vulnerable service users or clients; (ix) whether the limitation imposed is the least intrusive measure open to the employer.”

Mrs Fraser Butlin invited this Court to approve the guidance given in para. 94 (parts of which at least are acknowledged by Eady J to have been based on the submissions of the Archbishops’ Council to the EAT). She drew our attention to a number of ET decisions in which it had been acknowledged to be helpful, and to an observation to the same effect in Harvey on Industrial Relations and Employment Law. Mr O’Dair and some of the other interveners, on the other hand, while not asserting that it was positively erroneous, proposed various amplifications or refinements. Eady J herself, at para. 93, sounds a strong note of caution about the value of guidelines in this field. I agree, and I would echo in particular her statement that the relevant principles require nuanced decision-making and that there can be no one-size-fits-all approach. But, provided para. 94 is read as being limited to, as she says, a summary of the underlying principles, I would respectfully endorse it. I would only say, consistently with the caution already expressed, that it will not be necessary – or always even useful – for a tribunal to structure its reasoning by reference to the nine “considerations” enumerated in head (v). All of them are potentially relevant; but in practice, as the EHRC observed, the focus of the issues in any given case will only be on some of them, and there may be some cases where other considerations – or considerations which do not neatly fit into her formulation – may be relevant.

The Claimant’s grounds of appeal begin with a summary which acknowledges that the appeal was allowed but says that it is her case that the claim should not have been remitted to the ET because “the EAT was bound in law to reach its own conclusion and allow the Claimant’s claim for direct discrimination”. As to that, there is no issue about the principles applying to the EAT’s decision whether to remit a case in respect of which it has found that the ET made an error of law. As I put it at para. 45 of my judgment in Jafri v Lincoln College [2014] EWCA Civ 449, [2015] QB 781:

“If, once the ET’s error of law is corrected, more than one outcome is possible, the authorities are clear that it must be left to the ET to decide what that outcome should be, however well-placed the EAT may be to take the decision itself.”

(As I said in the following paragraph, I think it is regrettable that the test is so inflexible, but the law is settled short of the Supreme Court.)

The Claimant then pleads four specific grounds of appeal, as follows:

“GROUND 1: On the factual findings of the ET, supplemented by undisputed and indisputable facts of this case, the EAT was bound to conclude that the Respondent’s interference with the Appellant’s rights cannot be justified under Article 9(2) or 10(2), because:

(a)

The interference was not ‘prescribed by law’;

(b)

The interference was not justified by protecting the Respondent’s reputation;

(c)

The interference was not justified by the protection of rights and freedoms of others. There is no right not to be offended, and the offence taken by the audience (‘heckler’s veto’) can never justify interference with Convention rights.

(d)

The interference was not proportionate; and/or

(e)

The interference could not be justified as necessary in a democratic society in the light of the essential principle of pluralism which underpins the Convention.

GROUND 2: The EAT has failed to direct itself, or to provide guidance to the ET on remission, on the principle that the Convention protects not only the substance of a manifestation/expression, but likewise the language and manner.

GROUND 3: The EAT has failed to address Grounds 5-7 of the Grounds of Appeal before it. Had it done so, it was bound to conclude, on the unchallenged factual findings of the ET supplemented by undisputed and indisputable facts of this case:

(a)

that the complainant was guilty of unlawful stereotyping and therefore of discrimination

(b)

Respondent adopted the discriminatory views of that third party and was thus had been [sic] guilty of direct discrimination.

GROUND 4: EAT has erred in failing to uphold Ground 4 of the Appellant’s appeal: The ET reached an impermissible conclusion and/or failed to properly explain its reasons for attributing Mr Conlan’s reasons to all other decision-makers; alternatively, misdirected itself in the identification of the relevant decision-makers.”

In her reasons for giving permission to appeal Elisabeth Laing LJ found each of those grounds to have a real prospect of success, but she added that, even if they did not, the appeal raised “at least three important questions about the dismissal of an employee for the expression of her beliefs”. I have addressed these in the course of my consideration of the issues.

As already noted, the School sought in its Respondent’s Notice dated 5 February 2024 to challenge the terms of the EAT’s order. Section 6 of the Notice reads:

“The Respondent accepts that the EAT was entitled to uphold the appeal and remit to the same employment tribunal to reconsider the effect (if any) of the Facebook posts as constituting a manifestation of a protected belief. It challenges, however, the importation of a Convention-based proportionality assessment - including a gateway criterion that the impugned conduct must be ‘prescribed by law’ - into s13 EqA 2010.

The Respondent’s Grounds are set out in the Answer to Claimants Appeal and Notice of Cross-Appeal.”

I need not set out the full grounds referred to in the final sentence. The School’s essential point was that there was no warrant for introducing the requirements of article 9.2 into the exercise required by the 2010 Act. Its position is sufficiently summarised by para. 9 of the Notice of Cross-Appeal, which reads:

“An ET should apply heightened scrutiny where the decision-maker asserts that the reason for detrimental treatment is not the protected characteristic but a feature separable from it. A similar approach is taken to separability in victimisation or whistleblowing complaints. R accepts that the ET’s enquiry at first instance is not sufficiently reasoned and consents to remission accordingly. That heightened enquiry, however, is confined to those matters known to and operative upon the mind of the decision-maker. It is subject to neither a ‘prescribed by law’ test nor a proportionality exercise. Both the EAT’s terms of remission and Ground One are thus founded on errors of law.”

Elisabeth Laing LJ refused the School permission to cross-appeal on that basis for two reasons – (a) that the School had conceded the point below “by accepting in both tribunals that A’s Convention rights were relevant and by arguing that R’s interferences with those rights were justified”, and (b) that its proposed argument was contrary to the ratio of Page, by which the Court was bound. Although the School sought permission to re-open that refusal, Elisabeth Laing LJ found that the stringent requirements for granting permission to re-open had not been met; and she in any event maintained her view that this Court had in Page “treated the discrimination claims and the Convention rights arguments as co-extensive, or as virtually co-extensive”.

The result is that it is not open to the School on this appeal to argue that Eady J was wrong to proceed on the basis that its treatment of the Claimant had to be justified by reference to the criteria in article 9.2.

Although ground 1 ostensibly comprises a number of distinct points, most of them are simply different formulations of the submission that the EAT was bound to hold that the treatment complained of by the Claimant was not an objectively justifiable response to her having made the posts. Ground 2 makes a further particular point which feeds into that submission, and it is convenient to take it together with ground 1, as indeed Mr O’Dair did in his oral submissions.

The School did not in its original Grounds of Resistance plead any case on objective justification and, because of the view that it took on the law, the ET did not consider any such case. Even in the School’s skeleton argument before us no case is clearly articulated about the features of the posts which are said to justify the action taken against the Claimant. I accordingly focus on the justification advanced by Mr Jones in his oral submissions. That justification broadly reflected the reasons given by the School in the dismissal letter, but it does not correspond to them entirely. In so far as it differs, that is not fatal to the School’s case: the test is objective, and an employer can in principle justify an act complained of on a basis that it did not articulate at the time.

Mr Jones began by referring to two decisions of the ECtHR as identifying the relevant principles – Giniewski v France, 64016/00, [2006] ECHR 82, and Lilliendahl v Iceland, 29297/18, [2020] ECHR 931. Neither is concerned with the limits of free speech in an employment context, but Mr Jones submitted that they remained useful as statements of principle.

Giniewski concerned a newspaper article criticising a papal encyclical. It included a statement that certain Catholic teachings had led to antisemitism and “prepared the ground in which the idea and implementation of Auschwitz took seed”. The publisher had been found liable in civil proceedings for “public defamation against a group of persons on account of their membership of a religion”. He claimed (and the Court found) that that decision infringed his article 10 rights. Mr Jones relied on the decision only for the following statement of principle (at para. 43 of the judgment), which he said applied equally to the expression of religious beliefs protected by article 9:

“As the Court has stated on many occasions, freedom of expression constitutes one of the essential foundations of a democratic society and one of the basic conditions for its progress and for each individual’s self-fulfilment. Subject to para. 2 of Art. 10, it is applicable not only to ‘information’ or ‘ideas’ that are favourably received or regarded as inoffensive or as a matter of indifference, but also to those that offend, shock or disturb.  As para. 2 of Art. 10 recognises, however, the exercise of that freedom carries with it duties and responsibilities. Amongst them – in the context of religious opinions and beliefs – may legitimately be included an obligation to avoid as far as possible expressions that are gratuitously offensive to others and thus an infringement of their rights, and which therefore do not contribute to any form of public debate capable of furthering progress in human affairs.”

Importantly, that passage recognises the distinction between, on the one hand, expressing views that may “offend, shock or disturb” and, on the other, the way in which the views are expressed. Mr Jones relied in particular on the Court’s reference to expressions of the belief that are “gratuitously offensive to others” and which therefore do not contribute to any form of public debate capable of furthering progress in human affairs. I do not think that the Court intended that those phrases should be treated as stating the definitive criterion of the limits of acceptable speech, but I accept that they give a useful general indication. To the extent that the Court’s language is relied on, the word “gratuitously” should not be overlooked: the Court was evidently referring to language which was offensive for the sake of offence.

In Lilliendahl the applicant had posted a comment underneath an online newspaper article about the promotion of education about LGBT issues, in which he described homosexual activity in crude and highly offensive language and referred to it as disgusting. He was convicted of a criminal offence and fined. In its judgment the Court reviewed its previous case-law on the subject of “hate speech”. This distinguishes between (a) speech which “seeks to stir up hatred or violence”, such that it can be treated as being “aimed at the destruction of the rights and freedoms [sc. of others] laid down in [the Convention]” and thus to fall within the scope of article 17 (see paras. 24-25 of the judgment), and (b) “less grave forms of ‘hate speech’ which the Court has not considered to fall entirely outside the protection of Article 10, but which it has considered permissible for the Contracting States to restrict” (see para. 35). Para. 36 reads:

“Into this second category, the Court has not only put speech which explicitly calls for violence or other criminal acts, but has held that attacks on persons committed by insulting, holding up to ridicule or slandering specific groups of the population can be sufficient for allowing the authorities to favour combating prejudicial speech within the context of permitted restrictions on freedom of expression (see Beizaras and Levickas v. Lithuania, … §125; Vejdeland and Others v. Sweden, … §55, and Féret v. Belgium, … §73). In cases concerning speech which does not call for violence or other criminal acts, but which the Court has nevertheless considered to constitute ‘hate speech’, that conclusion has been based on an assessment of the content of the expression and the manner of its delivery.”

The Court’s conclusion was that the applicant’s conduct fell into the second category and that his former conviction and fine were objectively justifiable.

Mr Jones asked us to note in particular that the Court in Lilliendahl found that the language used by the applicant, coupled with the clear expression of disgust, was such as “to promote intolerance and detestation of homosexual persons”. He relied on the case as authority for the proposition that article 10 rights, and thus also rights under article 9, did not extend unqualified protection to “insulting, holding up to ridicule or slandering specific groups of the population”: whether speech of that kind was protected would depend on “the content of the expression and the manner of its delivery”. I have no difficulty with that proposition, which is consistent with the approach endorsed in Page.

Having said that, I do not think that the Strasbourg concept of “hate speech” (or the equivalent domestic jurisprudence) can be straightforwardly applied in the present context, and Mr Jones did not invite us to do so. Lilliendahl (like the other cases referred to in it) is concerned with restrictions imposed by the state, indeed with criminal sanctions; and the context of the employment relationship is different.

In his skeleton argument Mr O’Dair relied on observations in the decision of the ECtHR in De Haes v. Belgium 19983/92, [1997] ECHR 7, to the effect that “journalistic freedom … covers possible recourse to a degree of exaggeration, or even provocation” (para. 46) and that “article 10 … protects not only the substance of the ideas and information expressed but also the form in which they are conveyed” (para. 48). He cited by way of illustration a number of Strasbourg and domestic cases in which the exercise of free speech has been held to be protected notwithstanding the use of offensive language. I have no difficulty accepting that in particular cases even very offensively expressed statements may be protected by article 10. But that is not inconsistent with the proposition clearly recognised in Giniewski and Lilliendahl that in some cases the offensive language of a statement of fact or opinion will nevertheless justify an interference with the speaker’s article 10 rights. I also observe that the authorities in question are, again, concerned with interference by the state and not in the context of the employment relationship.

Against that background, Mr Jones submitted that an ET could properly conclude that the language of the posts was indeed gratuitously offensive and insulting to homosexual and trans people and did nothing to contribute to constructive public debate; that it was potentially damaging to the reputation of the School that one of its employees should post or re-post messages that used such language; and that dismissal was a proportionate response. The essential points that he made in support of each of those submissions were as follows.

As regards the language of the posts, Mr Jones referred in particular to the reference in the re-posts to “the LGBT crowd” perpetrating a form of “child abuse” by subjecting schoolchildren to what is described as the “mental illness” of believing in gender fluidity. That was grossly insulting to gay and trans people as a group. He acknowledged that it was the language of the original posts, rather than the Claimant’s own, but he pointed out that she had chosen to re-post them and that, despite her concession that it would have been better to use her own language, she had not taken them down.

As regards potential damage to the reputation of the School, Mr Jones emphasised that this was not a case where the Claimant was using the kind of language that she did about some issue which had no bearing on her employer’s business. On the contrary, the posts made serious allegations about what was said to be going on in schools, including secondary schools. In those circumstances, a reader of the posts who knew that she worked at the School might well be concerned that she was expressing those views, and using the same offensive language about LGBT people, in the school environment, including to pupils. At least one reader of the posts, the complainant, had identified her and knew that she worked at the School, and there was no reason to suppose that there might not be many others: indeed the passage from the dismissal letter quoted at para. 22 above records the Claimant’s acknowledgment that that was so.

As for whether dismissal was a proportionate response, Mr Jones relied on the fact that the Claimant had told the panel that, despite her concessions about the language of the posts, she would not have acted differently: there was thus no reason to believe that she would not post similar material in the future. He said that a tribunal would be entitled to accept the panel’s conclusion, recorded in the dismissal letter, that she had demonstrated no understanding of the implications of her actions or of how they might reflect on her professional reputation and that of the School within the community.

Accordingly, he submitted, following the Bank Mellat approach, it would be open to the ET to find that the School’s aim was to protect itself from the reputational harm which it was liable to suffer from a member of its staff using the gratuitously offensive and insulting language that she had; that her dismissal was a rational way of achieving that aim; that any measure short of dismissal would have unacceptably compromised the achievement of that objective; and that its achievement outweighed the interference with the Claimant’s article 9 (and article 10) rights.

I need to emphasise two things about the School’s case on justification as so advanced.

First, it is no part of its case that it was entitled to object to the Claimant publicly expressing her protected beliefs. That is so even if parents or others who knew where she worked might have found those beliefs offensive and thought the worse of the School for employing her. Mr Conlan accepted that in his evidence in the ET (see para. 102 above), and he was right to do so. This is not therefore a case about whether an employee can be dismissed simply for expressing those views on Facebook: it is about the terms in which she did so. This point is worth making because, on at least one reading of the complainant’s emails (Footnote: ¹⁰), what they found offensive was indeed the substance of the Claimant’s protected beliefs: if so, that is not the basis of the School’s justification.

Second, it is not part of the School’s case that there was in fact a risk that the Claimant would express those views, or exhibit any prejudice against gay or trans people, in the work environment. She had said in terms that she would not do so, and that was not controverted by the School in its dismissal letter and appears to have been accepted by the ET. The justification is focused squarely on reputational damage – that is, on the risk that readers of the posts might think the worse of the School for employing, in the Claimant’s position, someone who expressed themselves about “the LGBT crowd” in the way that she did.

I do not propose to summarise Mr O’Dair’s submissions in full, partly because some were not material to the School’s case on justification as it eventually emerged, and partly because some of his points will be incorporated into my own reasoning. But there are three points which it is useful to address at this stage.

“Reputational Damage”

Mr O’Dair submitted that the effect of the reasoning of the ET summarised at para. 105 (2) above was that the School had been justified in dismissing the Claimant simply on the basis that the complainant had been, and other readers of the posts might be, offended by her expression of her protected beliefs and think the worse of it for employing her. He submitted that reliance on reputational damage of that kind was unacceptable in principle. (Footnote: ¹¹) It is part of the price of a pluralist society that employers may sometimes have to take a hit for employing someone who expresses unpopular beliefs. He pointed out that it is well established that an employer cannot avoid liability for race or sex discrimination by showing that the act complained of was done in order to avoid adverse reactions from racist or sexist customers or clients (see, e.g., Din v Carrington Viyella Ltd [1982] ICR 256 (Footnote: ¹²)); he submitted that that principle applies equally in the present case. The FSU likewise in its written submissions expressed concern about the indiscriminate use of reputational harm as a justification for interference with employees’ freedom to manifest their beliefs.

The ET was not of course addressing the issue of justification as such, and I am not sure that the effect of its reasoning is indeed as Mr O’Dair characterised it; but I agree with him that, if it is, it is wrong in principle. An employer does not have carte blanche to interfere with an employee’s right to express their beliefs simply because third parties find those beliefs offensive and think the worse of it for employing them. Nor, however, does the employee have carte blanche about what they can say in public or how, or in what circumstances, they say it. The characteristics of the employment relationship may entitle the employer to impose limitations on the employee’s rights to manifest their beliefs and of free expression in accordance with articles 9.2 and 10.2 of the Convention. The particular characteristic with which we are concerned here is the employer’s legitimate interest in protecting its reputation – not only with customers or consumers or users of its services, but also with other third parties, including other employees, whose perception of it may affect its work or business. The extent to which it will be justified in interfering with the employee’s article 9 or article 10 rights will depend on the circumstances, and it is not possible to give any kind of general exposition. For present purposes I need only note three considerations which will be relevant to the proportionality of any such interference.

The first consideration is the subject-matter of the expression of opinion or belief. A statement of the employee’s views about matters which have nothing to do with the employer’s business is self-evidently less likely to damage its reputation than a statement about matters which are central to it. In the present case there is, as Mr Jones pointed out, a connection between the posts and the School because they related to sex education in schools.

The second consideration is the way in which the employee expresses their beliefs. This of course is the same kind of distinction as discussed above in connection with the separability cases, and also acknowledged in Giniewski and Lilliendahl. Even where the belief itself is protected, an employer may suffer reputational harm from being associated with an employee who expresses it publicly in an inappropriate way. The paradigm of such a case will be where the views are expressed in egregiously offensive or insulting language, as illustrated in cases like Lilliendahl (though there might in principle be cases where the reputational harm is done by some other feature). I would emphasise that the threshold of offensiveness should be high: protection should not be lost merely because the employee has expressed themselves intemperately. As we have seen, Mr Jones submits that the threshold is crossed in this case.

The third consideration is whether it is clear that the views expressed are personal to the employee. It is one thing to be entitled to express your own views on sensitive topics, but another to risk them being imputed to your employer, to whom it may be important to maintain institutional neutrality on the issue in question. This consideration may be particularly (though not only) important in the case of senior employees. As I said at para. 59 of my judgment in Page,

“[t]he extent to which it is legitimate to expect a person holding a senior role in a public body to refrain from expressing views which may upset a section of the public is a delicate question which can only be decided by reference to the facts of each particular case.”

A principal reason why this Court upheld the tribunal’s decision in that case was that it had found that the trust legitimately feared that the expression by one of its directors of a view that homosexual acts were wrong might discourage some gay patients from taking up its services: see para. 61 of my judgment, where I distinguished a concern of that kind from “generalised perceived reputational damage”. I mention this consideration because it is quite clear that it was not present in this case. As we have seen, the panel expressly accepted that it was unlikely that readers of the posts would believe that the Claimant was speaking for the School (see the final paragraph of the passage in the dismissal letter quoted at para. 22 above).

I should emphasise that those are no more than relevant considerations. They should not be treated as criteria all of which an employer must satisfy in order to justify an interference with an employee’s article 9 or article 10 rights on the basis of reputational harm. I should also emphasise that even where reputational harm, or a risk of it, is shown the interference in question must be proportionate.

It follows from that discussion that the School’s reliance on reputational damage in this case is open to it in principle. Mr Jones does not rely on any damage that might be caused simply by the Claimant’s expression of her gender-critical views, or her beliefs about same-sex marriage, but on the damage done by what he says was her gratuitously offensive and insulting language about gay and trans people in relation to an issue of relevance to the work of the School.

Before leaving this topic, I should refer to the decision of Briggs J in Smith v Trafford Housing Trust [2012] EWHC 3221 (Ch), on which Mr O’Dair placed some reliance. In that case a manager in a housing trust had expressed views on his personal Facebook page disapproving of churches conducting same-sex marriages. He was disciplined for, among other things, bringing the trust into disrepute. He brought proceedings in the County Court for breach of contract: the proceedings were subsequently transferred to the High Court. In paras. 55-64 of his judgment Briggs J held that the claimant’s conduct was incapable of bringing the trust into disrepute, for essentially two reasons. First, he found that, although his Facebook page did identify his work, there was no risk that any reader would understand his views to be those of the trust. Second, even if the views that he expressed had the propensity to upset fellow-employees or customers of the trust who read them that did not constitute the bringing of it into disrepute. As he put it at para. 62:

“On the assumption that Mr Smith was not (as I have found) reasonably to be taken as seeking to express the Trust’s own views, I cannot envisage how his moderate expression of his particular views about gay marriage in church, on his personal Facebook wall at a weekend out of working hours, could sensibly lead any reasonable reader to think the worst of the Trust for having employed him as a manager.”

I respectfully agree with Briggs J’s conclusions in those paragraphs, but I do not believe that they advance the argument in this case, since Mr Jones’s case is that the Claimant’s expression of her opinions was far from “moderate”, and the opinions expressed were relevant to the work of the School.

Assumptions and stereotyping

It will have been noted that a central element in the Claimant’s ground 3 is that the complainant had been guilty of discrimination against her in the form of “unlawful stereotyping”: see para. 116 above. As will appear, I do not believe that it is necessary for us to determine ground 3, but this aspect of it is potentially relevant in the context of the School’s justification in the present case.

The foundation for the Claimant’s case on this point is the complainant’s accusation that she had expressed “homophobic” views. Mr O’Dair proceeded on the basis that the effect of the accusation was that she had an “animus” against gay people. That, like “homophobic” itself, is a rather imprecise term, the meaning of which may be affected by the context, but fortunately all that matters for our purposes is that the complainant meant to describe an attitude on the part of the Claimant towards gay people that might lead her to treat them differently. The complainant did not in terms also describe the Claimant as being “transphobic”, but it seems from paras. 15-16 above that Ms Dorey and/or the Claimant herself understood that to be part of the case against her.

Mr O’Dair’s submission is that the accusation that the Claimant had homophobic and (apparently) transphobic attitudes was not based on any actual expression of such attitudes in the posts. Rather, it was an assumption that anyone who expressed the protected beliefs must be homophobic or transphobic. He submitted that there was no basis for such an assumption. Holding gender-critical beliefs, or believing that same-sex marriage is not equivalent to traditional marriage, cannot be equated with an animus against gay or trans people. The Claimant herself put the point pithily to Ms Dorey, when she said (see para. 16 above):

“I know that there are transgenders and gays who do have the same beliefs as me.  …  I am not against gay people, it doesn’t say that.”

As noted at para. 105 (1) above, the ET accepted her evidence on that point and appears to have accepted the distinction in question.

In this connection Mr O’Dair referred us to para. 250 of the judgment of Julian Knowles J at first instance in the Miller case ([2020] EWHC 225 (Admin)), where he accepts evidence that “some involved in the [transgender] debate are readily willing to label those with different viewpoints as ‘transphobic’ or as displaying ‘hatred’ when they are not”. There is also a parallel with the observation of this Court (Irwin and Haddon-Cave LJJ and Sir Jack Beatson) at para. 5 (10) of its judgment in R (Ngole) v University of Sheffield [2019] EWCA Civ 1127. The claimant in that case had been removed by the defendant university from his MA course in social work because he had expressed on Facebook the opinion that homosexual acts were sinful. The Court said:

“The mere expression of views on theological grounds (e.g. that ‘homosexuality is a sin’) does not necessarily connote that the person expressing such views will discriminate on such grounds. In the present case, there was positive evidence to suggest that the Appellant had never discriminated on such grounds in the past and was not likely to do so in the future (because, as he explained, the Bible prohibited him from discriminating against anybody).”

The next steps in the argument advanced in ground 3 are that the assumption that a person holding the protected beliefs must be homophobic and/or transphobic constitutes an unlawful stereotype applied to people holding those beliefs; and that the School “adopted” that discrimination by acting on the complaint.

Those steps are irrelevant under grounds 1 and 2 because we are concerned only with whether the School’s response to the posts was objectively justified. But it is just as necessary in that context to judge an employee’s statement by what they actually say (albeit including any necessary implications) rather than by what some readers might choose illegitimately to read into them. That is particularly important in the current social media climate, where messages are often read hastily and sometimes by people who are partisan or even ill-intentioned or (more likely) simply succumb to the common human tendency to find in a communication what they expect to find rather than what is actually there.

I have addressed this question because the point made in the previous paragraph may be of some general importance. It was indeed one of the three questions identified by Elisabeth Laing LJ when granting permission. She said:

“Where the objection is based on the words used by the employee, it is arguable that the defence should only be available if objectively, the employer can legitimately complain about the meaning of those words, and that it should not be available because of the reaction to those words of a person which derives, not from the objective meaning of the words, but from subjective inferences some people might draw, or which the complainant has drawn, from those words.”

As will be seen, I agree with that argument; and thus also with Falk LJ’s judgment below. However, in the light of the School’s case on justification as developed by Mr Jones, the point no longer directly arises: that justification does not depend on any stereotypical assumption that the Claimant was homophobic or transphobic, whatever the complainant may have thought, but on the actual language of the posts.

Use of personal Facebook account

Mr O’Dair emphasised in his submissions, as the Clamant had in the investigation and at the disciplinary hearing, that the posts were on her personal account, to which only her Facebook friends had access and which did not identify that she worked for the School. Likewise, in her grant of permission to appeal Elisabeth Laing LJ identified an important issue as being “the extent to which an employer may lawfully dismiss an employee for expressing views which are based on her religious beliefs in a forum which is not in the workplace, is not controlled by the employer, and which has a limited number of members”.

As to that, I accept that the facts that the posts were published in a forum which had nothing to do with the School, and would be likely to be seen by only a small number of people not all of whom would know of the Claimant’s connection with it, are highly relevant to the extent and gravity of any reputational damage that the School might suffer. But I did not understand Mr O’Dair to be submitting that they afforded a complete answer to the School’s case, and I do not believe that they do. As both the panel and the ET pointed out, there is no guarantee that Facebook posts will only be seen by those to whom the owner of the account grants access; and in any event publication even to her Facebook friends might in principle damage the School’s reputation among the local community who had close connections to it as parents or otherwise.

That conclusion is not inconsistent with Briggs J’s conclusion in Smith quoted at para. 145 above. The fact that in that case the claimant’s views were posted only on his personal account was not treated as decisive of the question of reputational damage: Briggs J takes into account equally that they were “moderately expressed”. They also of course had no relevance to his work.

It follows from my observations on Mr O’Dair’s submissions that there is no threshold objection to the School’s case on justification. The question is whether it would be open to the ET, if the case were remitted, to accept that case on the facts.

It is simpler to state my conclusion first and then give my reasons. In my opinion the ET would be bound to find that the Claimant’s dismissal was not objectively justified and accordingly that it constituted unlawful discrimination. That is of course her principal complaint, and no separate finding is required about the rejection of her internal appeal. As regards her secondary complaint about the disciplinary process leading to her dismissal, I do not believe that we are in a position ourselves to decide whether that also was unjustified, but I hope and expect that it will not be necessary for the case to be remitted for the determination of that question. I take the two complaints in turn.

As regards the dismissal, I am prepared to assume, but without deciding, that the School was entitled to take objection to the posts for the reasons relied on by Mr Jones – that is, that their language was gratuitously offensive to gay and/or trans people because of the way that it described the conduct of “the LGBT crowd” and that it was used in the context of sex education in schools and was accordingly relevant to the Claimant’s work. Even on that assumption, however, I believe that dismissal was unquestionably a disproportionate response. My reasons are as follows.

First, even if the language of the re-posts passes the threshold of objectionability, it is not grossly offensive. In context, the description of the promoters of gender fluidity as “the LGBT crowd” does not appear to be primarily intended to incite hatred or disgust for homosexuals or trans people. Rather, it is one of a series of derogatory sneers, alongside “liberal school systems” and “far-left zealots”. The accusations of “child abuse” and the promotion of mental illness may be stupidly rhetorical exaggeration but they are not likely to be taken literally. I do not mean to downplay the offensiveness of the language of the re-posts, but we are a long way from the kinds of direct attack on homosexuality found, for example, in Lilliendahl.

Second, the language is not the Claimant’s own (except for her repetition of the word “brainwashing”). It appears only in messages from others which (as would be evident to the reader) she had re-posted. She made clear to the School that she did not agree with the language used: that does not of course absolve her of responsibility for re-posting it, but it is relevant to the question of the degree of any culpability.

Third, the panel accepted that there was no evidence that the reputation of the School had thus far been damaged: its concern was about potential damage in the future (see, again, the final paragraph of the passage in the dismissal letter quoted at para. 22 above). As it also accepted, there was no possibility that, even if readers of the posts associated the Claimant with the School, they would believe that they represented its own views. Any reputational damage would only take the form of the fear expressed by the complainant, namely that the Claimant might express at work the homophobic and transphobic attitudes arguably implicit in the language used. I accept that if that belief became widespread it could harm the School’s reputation in the community, as the panel clearly thought. But the risk of widespread circulation was speculative at best. The posts were made on her personal Facebook account, in her maiden name and with no reference to the School. By the time of the hearing, several weeks after the posts were made, only one person was known to have recognised who she was.

Fourth, even if readers of the posts might fear that the Claimant would let her views influence her work, neither the panel nor the ET believed that she would do so. There was no reason to doubt her assertion that her concern was specifically about the content of sex education in primary schools; that she “wouldn’t bring this into school”; and that she would never treat gay or trans pupils differently (see para. 16 above). There had indeed been no complaints about any aspect of her work for over six years. It would have been open to the School, if it really thought it necessary, to issue a statement making it clear that it was confident that there was no risk that the Claimant’s views would affect her attitude towards gay or trans pupils or parents: it will be recalled that the ET itself floated this possibility at para. 65 of its Reasons.

Taking those reasons together, I do not believe that dismissal was even arguably a proportionate sanction for the Claimant’s conduct. It was no doubt unwise of her to re-post material expressed in (to use the ET’s words) florid and provocative language with which she did not agree, and in circumstances where people were liable to realise her connection with the School. But I cannot accept that that can justify her dismissal, and still less so where she was a long-serving employee against whose actual work there was no complaint of any kind.

My decision on this point is reinforced by the strong indication in the ET’s Reasons that it would be likely to have found that the Claimant’s dismissal was unfair if that issue had been before it: see para. 103 above.

In reaching that conclusion, I have not lost sight of the point, emphasised by Mr Jones, that the panel believed that the Claimant had no “insight” into the consequences of her actions, as illustrated not only in what she said to it but in her failure to take down the posts: see the second paragraph of the passage from the dismissal letter quoted at para. 24 above. This view was obviously central to its decision to dismiss. I accept that in an appropriate case lack of insight may justify an employer in choosing dismissal rather than a less severe sanction; but there can be no universal rule. There are understandable reasons why in some cases an employee may not be willing to admit that the conduct in question was wrong, or seriously wrong, particularly if it was the manifestation of a deeply-held belief. If the case is not one that would otherwise justify dismissal, it is hard that it should be marked up in seriousness because of a failure to make an acknowledgement of fault which the employee would genuinely find difficult. The position may be different where the employer needs to be confident that the employee understands what they have done wrong in order to prevent a more serious or damaging occurrence of the same conduct in the future; but for the reasons already given this was not that kind of case.

I would add that the judgment of this Court in Ngole contains useful observations on the dangers of placing inappropriate weight on “lack of insight”. The university appears to have acknowledged that the claimant’s expression of his view that homosexual acts were sinful was not inherently inconsistent with working as a social worker. But it believed that it was necessary for him to change the way in which he expressed himself and make clear that his views would not affect his work; and that that would not be possible because he had adopted an intransigent position in defence of his posts which showed no insight into why they were problematic. At paras. 109-112, headed “Lack of ‘insight’ and entrenched positions”, the Court held that the claimant’s termination was not justified, both because the university was itself partly responsible for his intransigence but also, “crucially”, because it never made it clear

“that it was the manner and language [emphasis in original] in which [he] had expressed his views which was the problem or discuss or offer him guidance as to how he might more appropriately and moderately express his views on homosexuality in a public forum and in a way in which it would be clear that he would never discriminate on such grounds or allow his views to interfere with his work as a professional social worker.”

(see para. 111). I do not suggest that the facts in Ngole are directly comparable to those of the present case; but the Court’s approach to the issue is nevertheless instructive.

I can give my reasons on this point shortly. I have no doubt that the School was entitled to carry out an investigation of some kind in response to the complaint. The posts unquestionably used offensive language, even if it was evidently not the Claimant’s own, and they had been seen by, and caused concern to, at least one parent who knew that she worked at the School. It would frankly have been irresponsible not to try to ascertain whether there was a risk of serious reputational damage or of the Claimant “bringing into school” the issues that she raised in the posts or holding attitudes that might affect how she treated gay or trans children. It is debatable whether that investigation needed to be disciplinary in character or, if it did, whether Ms Dorey was justified in finding a case to answer at the end of it. It is still more debatable whether it was necessary to suspend the Claimant: as to this, see the observations of Elias LJ in Crawford v Suffolk Mental Health Partnership NHS Trust [2012] EWCA Civ 138, [2012] IRLR 402, at para. 71. However, these are not questions to which I believe that this Court can provide a confident answer in circumstances where they were not explored by the ET or in the submissions to us.

Formally, therefore, the Claimant’s case on these elements of her claim will have to be remitted to the ET. But I would strongly discourage that course. The real complaint in this case has been about her dismissal, and any points of principle which it raises have been decided in the context of that claim. Even if, as to which I express no view, the quantum of her compensation might be (slightly) greater if she succeeded on these elements, the additional sums would be wholly disproportionate to the costs. Mr O’Dair sensibly said in his oral submissions that he expected that the parties would take a “pragmatic” view about the need to remit these issues if the Claimant succeeded on the dismissal claim; and I am confident that they will do so.

For the reasons given above I believe that the EAT was wrong to order remittal on the dismissal claim and we should ourselves hold that the Claimant’s dismissal constituted unlawful discrimination on the ground of religion or belief. Remittal of that claim will only be necessary in order to determine the question of remedy. I hope and expect that remittal of the disciplinary process claim will not be necessary.

My conclusions on grounds 1 and 2 make it unnecessary to decide grounds 3 and 4, which are incapable of affecting the outcome of the appeal; and, like Eady J, I prefer not to do so. I can explain my reasons very summarily.

As to ground 3, the significance of the Claimant’s case based on unlawful stereotyping is that if it were established it would, at least as Mr O’Dair submitted, bypass any defence of objective justification because the complainant, and through them the School, would have been shown to be motivated, at least to a significant extent, not by the language of the posts, or any other separable feature, but directly by the beliefs expressed in them. This submission was not addressed in the oral submissions, but I am inclined to think that it is analytically correct. But I am not satisfied that it would be right for this Court to make findings on the complainant’s motivation, or on the issue of whether it was adopted by the School, particularly since this analysis was not considered by the ET and not fully addressed in the oral submissions before us. There is no advantage in our wrestling with an issue which is far from straightforward and cannot lead to a different finding than we have already made on grounds 1 and 2.

It is right to record that we had helpful written submissions from EHRC on the law concerning direct discrimination by stereotyping in the context of religion or belief. So that their work is available in a case where it may be directly useful, I quote the following passage:

“50.

… One particular species of [direct discrimination] in which an employee’s holding or manifesting their belief might have a significant influence on their treatment is through the putative discriminator’s adoption of a stereotype. In such a case, although the reason for the treatment given by the employer is not the protected characteristic, it is nonetheless positively relied on by the claimant. The claimant argues that the decision-maker has, consciously or unconsciously, adopted a stereotype and was significantly influenced by it in deciding on the treatment complained of i.e. on the basis that the claimant would share the perceived attributes of the group, rather than relying on evidence about the particular individual. If that is so then the treatment will be ‘because of’ the protected characteristic. It does not matter if the stereotype is very likely to be true: see e.g. R (European Roma Rights Centre) v Immigration Officer at Prague Airport … [2005] 2 AC 1 per Lady Hale at §82; Commerzbank AG v Rajput UKEAT/154/18 [2019] ICR 1613 … at §77 per Soole J.

51.

…

52.

In relation to religion or belief, a decision-maker will accordingly discriminate where the reason given for the treatment is significantly influenced, consciously or unconsciously, by a stereotype that persons who hold or manifest the relevant belief will share attributes of a group which they might not in fact possess. Examples of stereotypes in relation to religion or belief explicitly recognised in the case law are that: persons who hold/manifest certain gender-critical beliefs have animus towards trans persons (see e.g. R (Miller) v College of Policing [2020] EWHC 225 (Admin) [2020] 4 All ER 31 at §§250, 281); or that persons holding/manifesting the belief that same-sex sexual activity is sinful have animus towards gay persons (see e.g. Ngole at §115). It may be that some persons who hold/manifest such beliefs have such animus, but it is stereotyping to assume that all do. Given the above cases, the EHRC considers that a Tribunal is likely to be able to proceed on the basis that the stereotypes which they identify exist (Commerzbank, §§79-80), although it may need to give prior notice to the parties of a proposed use of the principle: ibid., §84.

53.

The stereotype must significantly influence the decision-maker’s decision. That is irrespective of whether, as in a case such as the present, the employer is acting following a third-party objection/complaint about the claimant.”

I am prepared to say, albeit entirely obiter, that my provisional view is that that is a correct summary of the law.

As to ground 4, this was the subject only of the briefest oral submissions from Mr O’Dair, and of none from Mr Jones or the interveners, and since it cannot affect the outcome I see no advantage in my prolonging this judgment by addressing it.

For the reasons given above I would allow the Claimant’s appeal against the EAT’s decision to remit to the ET the issue of whether her dismissal was unlawfully discriminatory and I would substitute a finding that she succeeds on that claim. I would dismiss her appeal against the decision to remit the remaining elements of her claim, but, as I have said, I hope and expect that a decision on those issues will not be necessary.

This has been a regrettably long, and long-delayed, judgment. It may assist non-lawyers or skim-readers if I summarise my essential conclusions in broad terms:

I agree with both judgments.

I am grateful to Underhill LJ for his detailed exposition of the issues, and agree with his judgment in its entirety. I add this concurring judgment only to emphasise one aspect.

As Underhill LJ explains at paragraphs 151 and 152 of his judgment, it is necessary in this context to judge a statement by what it actually says, and not by reference to a concern about what some readers might wrongly read into or infer from it. While this will, as he notes, include necessary implications, it must be emphasised that the test is an objective one. In other words, what meaning do the words used actually have? What message would they convey to a reasonable reader? In the event of a dispute, this will be a matter that the tribunal must decide for itself.

The ET did not undertake that exercise in this case. Rather, it accepted that the School felt that the posts could reasonably lead someone to conclude that the Claimant held certain views. This is understandable because the ET was focusing only on the School’s motivation. But it highlights a real concern that arises from such an approach. Namely, that if unchecked it would risk a judicially-endorsed silencing of the legitimate expression of views due to a concern about the conclusions that some might choose to draw from what is said, even if an impartial tribunal would not agree that the conclusions could reasonably be drawn and would also recognise the force of the point that freedom of speech entails the freedom to express opinions that may offend.

This point is closely linked to the ratio of Page, and in particular the use of the words “to which objection could justifiably be taken” in paragraph 68. It is only when a tribunal has determined what the words mean that it can then proceed to determine whether what was (actually) said was, despite being a manifestation of a belief, expressed inappropriately and whether, if it was, the response was objectively justifiable.

As Underhill LJ has explained (at paragraph 74 above) the words “to which objection could justifiably be taken”, and its shorthand version “objectionable”, should be taken to have the same effect as the word “inappropriate”. Speaking for myself, I find “inappropriate” the more helpful term, for two reasons. First, it reduces the risk of the words “to which objection could justifiably be taken” being incorrectly read as permitting regard to be paid to the risk that some readers might wrongly read what is said in a way that a reasonable reader would not. Secondly, it more obviously conveys that the forum and context for what is said is relevant, as well as the content and manner. For example, something that might be unproblematic on a private Facebook page could justify different treatment if communicated in a work setting.