The Case of Sheikh Ali Salman [Bahrain]
Closed Mixed Outcome
Global Freedom of Expression is an academic initiative and therefore, we encourage you to share and republish excerpts of our content so long as they are not used for commercial purposes and you respect the following policy:
Attribution, copyright, and license information for media used by Global Freedom of Expression is available on our Credits page.
The European Court of Human Rights (ECtHR) found a violation of ECHR Article 3 (Prohibition of torture) taken in conjunction with Article 14 (Prohibition of discrimination) in a case of continued attacks on the applicant Ms Armine Oganezova, a prominent LGBTI activist from Armenia. Attacks included, inter alia, an arson attack on her club, online hate speech comments and death threats. Due to these circumstances the applicant had to flee to Sweden to seek asylum. The ECtHR opined that the state did not fulfill its positive obligations regarding the protection of the applicant from homophobic attacks and especially in the terms of criminal investigation and criminal punishment of the perpetrators. The Court noted in particular the “absence of [an] effective domestic criminal-law mechanism for investigating discrimination complaints.”
The post below by Margarita S. Ilieva originally appeared in Strasbourg Observers blog and is adapted for our database with permission and thanks.
In a high-profile case of sustained homophobic mobbing goaded by high-ranking officials, a leader (the applicant) of the LGBTI community (the community) was persecuted for publicly representing and empowering the community:
Thugs firebombed the community club she ran, destroying her livelihood. Others harassed her for weeks, vandalizing her property. They constantly threatened her life to stop her from restoring the club, and to exile her; they succeeded. Ruling-party parliamentarians, including the Deputy Speaker, publicly praised the arson and its hate motive, then bailed its perpetrators. The judiciary extended the arsonists lenient conditional sentences and eventually, amnesty. This was despite their admitted homophobic motive and ‘Fascist’ ideology and affiliation documented by the indictment. The law did not recognize homophobic crime. Police allowed the post-arson attackers, including online verbal aggressors, to proceed unfettered and enjoy impunity: the victim briefly received belated protection, withdrawn without reason or explanation. She was forced into self-isolation and ultimately, exile. Robbed of her homeland, she was afflicted by long-term depression.
Bearing in mind these facts, the applicant lodged two applications to the European Court of Human Rights (ECtHR). She complained of the violation of her ECHR rights and invoked Articles 3, 8 taken alone and in conjunction with Article 14 of the ECHR.
The ECtHR found a violation of Article 3 of the Convention taken in conjunction with Article 14.
Ten years after the cause of action, the Court unanimously adopted the applicant’s argument that she suffered a homophobic campaign. The ‘campaign’ finding acknowledged the harassment as systematic and deliberate, despite the perpetrators being numerous, not (clearly) organized, their attacks not apparently planned or coordinated – as opposed to seeing them as random or isolated, ignoring their continuum and unifying intention.
The Court termed the “hate-driven campaign” [para. 78], “sustained and aggressive” [para. 91], “continued” [paras. 94, 109],‘ongoing abuse” [para. 95], “continuous harassment” [para. 97], “premeditated and motivated by homophobic bias […] aimed at deterring the applicant from reopening the club” [para. 93], “evidently aim[ed at] frighten[ing] the applicant so that she would desist from her public expression of support for the LGBT community, including her community-oriented activism by running the club as a communal project” [para. 95].The arsonists and other perpetrators were neo-Nazi [paras. 11, 35].
This was examined simultaneously under Articles 3 and 14. The campaign met the severity threshold despite the applicant never being physically harmed. The community’s systemic victimization as per international equality data was relevant to the applicant’s vulnerability level. The police had acknowledged the risk to her physical integrity [paras. 94-5].
The arson’s investigation was ineffective despite the hate motive being overt and documented in the indictment [para. 102]. Effective criminal law, required to address this homophobic motive, was absent [para. 103]. The justice system treated the crime as an ordinary one, the hate having no legal significance. Armenia had not followed international recommendations to correct the law [para. 104].
The Court disbelieved that police had told the victim the arsonists were justified [para. 107]; the government (simply) contested it. The Court held it was “not possible […] to determine whether the police officers […] did indeed make the comments”, despite high-ranking officials having publicly made worse remarks [para. 107]. The burden of proof did not shift, no prima facie case was discussed: “there is no basis for the Court to find that it was a discriminatory state of mind that was at the core of the failure” to effectively investigate the arson.
Armenia failed to protect the victim from the post-arson attacks and to investigate those. It failed explaining why the protection was lifted [para. 113]. The authorities ignored the applicant’s complaints regarding the “highly abusive” online speech against her, including “numerous direct/ undisguised calls for violence” [paras. 117, 120]. Such “hateful comments” required criminal-law protection, which was absent [para. 120]. Instead, parliamentarians themselves made such comments. (Those were not discussed per se.) While currently hate speech is criminalised, sexual orientation and gender identity (SOGI) are not covered [para. 121].
As a consequence, Article 3 + Article 14 positive duties were violated. EUR 12 000 were awarded for non-pecuniary damages.
Despite recognizing the legislative fault – no protection against SOGI-based hate crime/speech– the Court rejected an Article 46 request for structural redress, i.e. legislative amendments. It provided no reasons beyond referencing state discretion regarding the means to execute its “essentially declaratory” judgments [para. 131]. Based on “the particular circumstances of the case, the Court [found] it appropriate to leave it to the Government to choose the means to be used […]” [para. 132].
Decision Direction indicates whether the decision expands or contracts expression based on an analysis of the case.
On The International Day Against Homophobia, Transphobia and Biphobia (IDAHO) 2022, the Court delivered its first LGBTI rights judgment against Armenia. This is Armenia’s first decided hate crime case, first hate speech case, and first case establishing failure to protect from discriminatory violence, as well as to investigate such violence. It is Armenia’s second case of established discrimination.
Oganezova‘s impact includes developing the Court’s systematization of ‘Sexual Orientation Issues’ – the Court introduced an ‘aggressive homophobic campaigns’ category in this factsheet, exemplified by this case. Oganezova is also a precedent as an Articles 3 + 14 breach was found due to failure to protect an individual from private-party homophobic attacks and hate speech with no physical violence against people. (Cf. WISG: ‘a mob […] randomly resorting to physical assaults, demonstrating the reality of the [death] threats […] had attacked the buses carrying […] applicants and managed to hit one on the head’ (§60). In Identoba, mobbing included physical assaults on some applicants (§69).) Oganezova is included in the Court’s 2022 Information Note on cases ‘of particular interest’.
Yet, did the Court acknowledge all layers of the homophobes’ crusade against the applicant, including its institutional enablers and forced exile agenda? Did the burden of proof shift as required in equality cases? Did the Court address the systemic issue – the environmental deprivation of the law concerning homophobic crime/ speech, or was its refusal to provide structural redress inconsistent, as well as non-transparent? It is argued that this strong judgment is deficient in several respects.
However, the judgment’s strengths outweigh several deficits. Assets include: contextual analysis built on equality data; recognition of persecution; an inferable requirement for ground-specific criminal protection against homophobia. Weaknesses consist of withholding: an injunction; discussion of a negative duty breach; a shift of the burden of proof.
a) Equality data steadily used to contextualize attacks
The Court consistently referenced the community’s overexposure to homophobia to gauge the acts’ severity for purposes of Article 3 applicability [paras. 92, 94, 97]. It integrated qualitative findings from international bodies, affirming that a context of systemic group victimization manifested the discrimination, and the level of physical threat, against the individual member [paras. 92, 94]. The risk thus derived sufficed to offset the absence of physical injury [para. 94]. This is advanced compared to cases not relying on equality data for the ‘severity threshold’ analysis [WISG, paras. 60-1]; Aghdgomelashvili and Japaridze, paras. 47-9].
Moreover, the Court referenced equality data to contextualise the need to investigate the arson as an attack on the community, elevating this duty to “essential” [paras. 104-5]. This resembles prior holdings of “a pressing need” to investigate given “well-documented hostility against the LGBT community” (Aghdgomelashvili, para. 40; WISG, para. 66). In Oganezova, investigation was framed as no less of an imperative than in cases, in which police were perpetrators (Aghdgomelashvili), or physical aggression took place (in WISG, mobbers stormed vehicles carrying the applicants and a stone injured one of them in the head, paras. 24, 26-7, 60). In Oganezova, private perpetrators destroyed property. Regardless, the equality data supported the same conclusion regarding severity. Thusly, in Oganezova, the Court built upon its respectable practice of heeding contextual information about group vulnerability.
Equality data on the legislative deficit – hate crime law not covering SOGI – supported an iteration of this deficit as a well-documented concern, implying that Armenia persistently ignored it [paras. 104, 121]. This was more comprehensive than Stoyanova v. Bulgaria (discussed below), which did not reference ECRI highlighting a similar legislative gap.
In conclusion, Oganezova’s integration of equality data builds on precedents, consolidating a context-heedful approach.
b) Recognition of premeditated homophobic persecution
Oganezova thoroughly acknowledged the anti-community purpose of the numerous, various, seemingly unrelated off – and online perpetrators as defining of a continuum of aggression. It consistently articulated their overt intent to disempower the community by obliterating the applicant’s voice and the safe space she provided [paras. 93, 95, 109]. Oganezova iterated the diverse violent acts and expressions as a continuous extent, despite their perpetrators not being apparently associated or organised [paras. 91, 94, 97, 109]. These findings related to the severity threshold being met despite the absence of bodily damage [para. 97].
This is the first holding that violent homophobia amounted to a campaign. It is significant, as this campaign was crowd-sourced, and not structured. Perpetrators lacking a defined organisation were recognised as capable of sharing premeditated intent and acting consistently, in a sustained manner over a period of time long enough to render their harassment ‘continuous’.
The Court did not acknowledge the perpetrators’ forced exile agenda. The applicant argued her harassers’ overt intention was to exile her, as well as to stop her from rebuilding the community club. This was sufficiently evidenced to warrant discussion. The exiling was only factored in as an outcome, regarding severity [para. 91].
c) Ground-specific criminal law against homophobia required?
The Court held that criminal law must respond to proven (overt) homophobic crime by documenting and addressing its homophobic nature [paras. 102-3). Previously, this requirement was posited in ground-unspecific terms (cf. “discriminatory motives”, “hate-motivated crimes” in WISG, para. 63). The Oganezova language is possibly a basis to require SOGI-specific criminal-law protection. This would reflect a natural-right demand for homophobia to be named rather than implied through catch-all provisions.
a) Denial of structural redress for deficient law
Article 46 relief was withheld without adequate reasoning. It is unclear which “particular circumstances of the case” were areas on to trust Armenia to address its own homophobia. Several circumstances warrant otherwise. The community’s heightened vulnerability was a reason to grant them better execution chances and thus hate crime/speech prevention through criminal-law deterrence. The lack of alternative execution means –amendments being the only way to prevent the law from further ignoring homophobia – was a reason not to leave Armenia the “choice”. Oganezova indicated criminal law protection was required; accordingly, there are no other options.
An Article 46 injunction was appropriate to steer Armenia– over a decade, it had ignored consistent international advice to correct its law. It has meanwhile extended hate speech bans, excluding homophobia [para. 121]. (According to PINK, the aggravating circumstances law and public incitement to violence law are now non-exhaustive–not specifically including and not necessarily excluding SOGI, depending on judicial construction.) Such deliberate non-inclusion (accidental lawmaking being unlikely), will be (perceived as) an enabler for further homophobia. The Court, in withholding an injunction, deferred to badly exercised state discretion that should be restrained, rather than rewarded. That (high-ranking) legislators acted as unchecked homophobes, perpetrating hate speech, advocating hate crime [paras. 33, 37, 39) with impunity, equally signaled that the establishment is unlikely, unassisted, to make law to protect its own victims from abuse.
Armenia should have been directed regarding curbing homophobia as it enabled a “sustained”, “ongoing”, ”continued”, “continuous” homophobic campaign, overtly “aim[ed at] frighten[ing] the applicant so that she would desist from her public expression of support for the LGBT community’ [paras. 78, 91, 93, 95, 97, 109]. It allowed a homophobic neo-Nazi group to operate, which fostered the arsonists and other perpetrators [paras. 11, 35]. The government did not claim having tried to curb it. This pertains to Armenia’s record, not the case’s severity. This government hardly deserved discretion regarding the plagued community. The community deserved structural relief.
Generally, under using international human rights injunctions serves no legitimate aim. Based on practical experience, they (likely) help states do better, avoiding liability, while giving victims better access to prevention. The Court’s injunctions are a public interest tool, not to be construed strictly. Their use should not be narrowly defined, especially where the remedy entails no redistribution. Furthermore, there execution depends on specific reform, an injunction wouldn’t preclude consideration of alternative means (there being none). It wouldn’t curtail states’ discretion as they wouldn’t have discretion if there are no options.
Stoyanova v. Bulgaria, a homophobic murder case decided after Oganezova, included an injunction of the type needed in Oganezova–criminal law amendment to recognize homophobic crime. Admittedly, in Stoyanova, a man died, while in Oganezova, no person suffered physically. However, the individual violation’s gravity is no criterion for Article 46 applicability. In Stoyanova, the criterion was that the “breach found […] appears to be of a systemic character […] result[ing] from the [law]” [para. 78]. Because of this, it was “appropriate for the Court to give some indications on how breaches of this kind are to be avoided in the future” [para. 78]. This applies to Oganezova, which equally concerns a breach stemming from the law. The Court did not term it “systemic”, yet, such a breach is systemic by nature. Unrectified, the law will reproduce the breach. ILGA-Europe ranks Armenia lower than Bulgaria in LGBTI equality – 47th (out of 49 countries) v. 40th. It was questionable, treating these two countries – these LGBTI communities – differently, disadvantaging the more vulnerable one. The Armenian systemic issue is more pressing, the societal context being more homophobic. In Stoyanova, there is no finding of a community “precarious situation”. Granted, legislative deterrence of SOGI-based murder was at stake, but Armenian law equally withholds such deterrence.
b) No mention of a negative duty breach
The Court only discussed the Deputy Speaker’s and MPs’ acquiescence – inciting homophobic violence and publicly bailing the arsonists – as contextual to the private-party aggressions, but not as acts per se with implications for the duty not to discriminate. It omitted the apparent causality between these officials’ acts and the ensuing private-party aggressions. On the day a parliamentarian publicly justified bailing the arsonists, justifying their crime, thus promising impunity, homophobes broke into the community club, demolished property and perpetrated hate graffiti [paras. 33-4]. Continuous mobbing followed, publicly goaded by other MPs on two occasions in the next days. The MPs stoking the hate campaign were critical actors: the state was responsible for active involvement.
A reckoning for ill-treatment by the state itself is indispensable. Ignoring such ill-treatment may amount to revictimisation. A state encouraging minority persecution should be held liable. In WISG, the Court declared the authorities’ procedural failure “contribute[d] to official acquiescence or connivance in hate crimes” [para. 76]. It documented violence as possibly “conditioned, at least in part, by the authorities’ failure” to prevent impunity for prior violence. The Court regarded no official passivity “as having contributed to the subsequent proliferation […] of hate crimes against the LGBT community”. In Oganezova, the MPs’ virulent homophobia warranted similar findings.
Additionally, the MPs’ acts constituted discrimination per se, to be examined as such. In WISG, police officers’ homophobic remarks “went against the State’s substantive negative obligations” [para. 77]. The MPs’ behavior in Oganezova, by the same token, constituted a negative duty breach.
Moreover, based on the overtness of the high-profile official hate speech, including the (justification of the) financial support for the perpetrators’ release, a discussion of a prima facie case of institutionalized homophobia was possible. Elected officials would not risk such behavior if they didn’t rely on the passive tolerance, if not backing, of (a critical mass in)the establishment. The Court neglected this, missing an opportunity to elevate the framing of LGBTI victimhood issues to the level of their severity in reality. Homophobic persecution will proliferate, as acknowledged in WISG, for as long as perpetrators, officials or others, enjoy institutional sponsorship.
Ultimately, the outcome of the judiciary’s handling of the arson fulfilled the MPs’ promise of impunity: the perpetrators received conditional, lenient sentences, then, complete amnesty. This would not be the expected outcome for highly destructive arson potentially endangering lives, without the judges, i.e. the state, being biased. But for the applicant’s identity, would the case outcome have been the same? The Court doesn’t comment, seemingly resisting the growth of its own discrimination case law. (The “but for” test in equality cases was settled House of Lords’ case law already in the last century, while the Court has never mentioned it.).
c) No shifting burden of proof
The Court disbelieved that police expressed the support for the arsonists that MPs did. The government had but to deny this, not proving anything that would constitute negation beyond a mere declaration – for example, demonstrating that the force as such and/or its officers take homophobic victimization seriously. The officers’ homophobic statements were consistent with their Court-noted neglectful (in)actions in this case of continued aggression. Despite this context, the Court did not mention a possible inference of discrimination by the police. It did not refer to the shifting burden of proof rule, which should be by default in discrimination cases (Handbook on European Non-discrimination Law, 6.1).
Will the Court standardize the shifting burden of proof rule by applying it in each discrimination case? Without it, anti-discrimination protection could never be effective. The shifting burden of proof has been an integral part of EU law for decades (the ECJ applied it in 1989 to indirect sex discrimination in Danfoss), as well as at the national level (see King, 1991). In Oganezova, shifting the burden of proof could have enabled an inference of institutional discrimination by the police and possibly the courts. It would have exposed the government for not attempting to rebut such a possible inference. Instead, the government was allowed to simply deny the applicant’s experience of revictimisation by the police.
Discrediting victims of hate crime is not consistent with the due victim-centered approach (ODIHR, Hate Crime Victims in the Criminal Justice System, 1.4: “Victims need to have their experience acknowledged and validated by law enforcement and criminal justice agents. They need to be believed and have their experiences recorded in the case file.”) Victims’ rights to respectful treatment (cf. Directive 2012/29/EU) should be integrated in a case like Oganezova.
Given all the established facts in Oganezova, it was less probable that the victim lied than that the police did victim-blame her, which the government would deny. Arguably, the standard of proof (for the applicant) in discrimination cases should be “on the balance of probabilities” rather than “beyond a reasonable doubt”, in keeping with the effectiveness principle. The ECJ settled the shifting burden of proof rule based on this principle (for example, Enderby). Firma Feryn established that discriminatory remarks suffice for a prima facie case [para. 34].
What is the Court’s record of finding covert direct discrimination? If substandard, the burden of proof should be (re)considered.
Oganezova is good enough (cf. Winnicott) to reflect the compelling survivorship of Ms. Oganezova. Facing the above, she continued her advocacy, knowing visibility equaled targeting. The judgment pays her respect. It corresponds to Jessica Gavron’s definition (at 3:45): “A positive judgment provides validation for the applicants of their experience, which they won’t have received at the domestic level […]. It also creates an account of state behavior, which is unassailable.”
Oganezova is a robust litigation debut for the community, acknowledging their victimhood. It advances the case law by recognizing that informal grouping scan mount a discrimination campaign, i.e. systematic, purposive, sustained victimization of a minority (member). Despite not broaching the subject of institutional discrimination, Oganezova demonstrates capability to register continua of abuse in contexts of oppression.
In addition to Margarita S. Ilieva’s analysis above, the judgment has a two-fold impact regarding the freedom of expression.
First, its purpose is to protect several interlinked rights promoting sexual orientation, including the freedom of expression of the LGBTI community. Albeit the case is not about ECHR Article 10, the judgment is based on the requirement to provide a framework where everyone, including LGBT+ persons, may freely and without any fear express themselves. Ilieva observed based on a close reading of the facts that there was clear causality – following each interview Oganezova or her associates gave to the media, or public demonstration they took part in, there was a retaliatory attack on her club.
Second, the judgment is an example of the continued ECtHR’s approach towards hate speech, indicating that it enjoys no or little protection under the ECHR. However, criminal-law responses are only “for the most serious expressions of hatred, inciting others to violence”, “only as an ultima ratio measure” [para. 120]. The ECtHR, for instance, previously found “that criminal-law measures were required with respect to direct verbal assaults and physical threats motivated by discriminatory attitudes” (Beizaras and Levickas v. Lithuania, paras. 110-111]. The same type of attack motivated by sexual orientation must be regulated in the same manner – by criminal law protection (as noted above referencing para. 120 of the decision). Armenia failed to enact such laws, and even today “the sexual orientation and gender identity are still not included in the characteristics of victims of the offence of hate speech” [para. 121].
In sum, if a state fails to prohibit hate speech based on sexual orientation or gender identity by the virtue of criminal legislation, it will most likely be a losing party in eventual cases before the ECtHR providing the speech “constitute[s] serious offences […] directed against a person’s physical or mental integrity [such as] undisguised calls for violence against the applicant” [para. 120].
Global Perspective demonstrates how the court’s decision was influenced by standards from one or many regions.
Case significance refers to how influential the case is and how its significance changes over time.
Let us know if you notice errors or if the case analysis needs revision.