November 23, 2023
by Sarah Ganty and Dimitry V. Kochenov[i]
In Valiullina and others v. Latvia and Džibuti and others v. Latvia, the Fifth Section of the European Court of Human Rights unanimously approved of further restrictions on Russian-language education in Latvia, thereby depriving a huge proportion of the population of the Republic of Latvia of education in their mother tongue. Constitutional identity was found to trump Convention rights and a very superficial assessment of proportionality was conducted. Supporting a bilingual school system in a bilingual society which pretends hard not to be one was proclaimed to amount to segregation, which needs to be dismantled at all costs. The Court agreed with the Latvian government that the existence of minority rights was the problem and not the violation of said minority rights. Valiullina and Džibuti thus follow the shame of Savickis in endorsing the reasoning of collective punishment by association. The Grand Chamber could still correct this double traducement of justice.
For twenty years now, Latvia has been undertaking structural reforms of its public and private education by increasing teaching of and in Latvian in minority schools, while reducing teaching in minority languages, mainly Russian. The cases of Valiullina and Džibuti concern the 2018 iteration of these reforms – respectively for public schools and for private schools – which stipulated that the new language proportions for teaching in both private and public primary and secondary schools should be: at least 50% of teaching hours in Latvian from first to sixthgrade; 80% from seventh to ninth grade; and Russian should disappear entirely between tenth and twelfth grade, with the exception of subjects related to the minority language, culture and identity.
As a result, Latvian children belonging to minority groups can no longer obtain education where substantial parts of the curriculum are taught in their mother tongue. Far-reaching exceptions are granted for international schools. Education in English, French and German, among others, is fine. It is only the Russian language – one of the two main languages spoken in the Republic of Latvia despite not being recognised officially – which is presented as a problem. The Latvian Constitutional Court unanimously blessed the reform in relation to public schools, critically discussed here and here. The same court examined the situation of the private schools in a separate case, ultimately reaching the same conclusion. Unlike the case concerning public schools, this decision was not unanimous (see the dissenting opinions here and here). So far this might all be considered fun and games, and might even not be unreasonable, but it is not the whole story. The new 2022 reform (here and here) de facto outlaws education in minority languages outright. September 2025 is the deadline set for all schools to comply. The reasoning of the ECtHR in greenlighting the 2018 reform is thus of crucial importance to the future scrutiny of the outlawing of minority language education in Latvia, with direct implications for the education of a huge proportion of the children in the bilingual republic.
The applicants in both cases are Latvian nationals and ‘non-citizens of Latvia’ (nepilsoņi, a special legal status reserved in Latvian law for the former USSR citizens and their offspring who were not granted Latvian nationality and who are settled in the territory), all belonging to the ‘Russian-speaking minority’, which includes ethnic non-Latvians of Russian, Jewish, Belarusian, Ukrainian and other extractions. They complained to the ECtHR that their right to private and family life (Article 8) and education (Article 2, Protocol 1) had been violated and that they had been discriminated against based on their language and ethnicity (Article 14 + Article 2 Protocol 1) because of the 2018 reform.
In both cases, the Article 8 complaint was considered inadmissible due to the framing of the case earlier before the Latvian Constitutional Court. Confirming the interpretation of the right to education developed in the famous Belgian Linguistic Case, the Court limited the right to access to education to the official languages. Since Russian is not an official language of Latvia, no violation was found. No discrimination on grounds of language was found either. The Court outright refused to examine ethnic discrimination.
In Valiullina, examining the discrimination claim, the Court followed the usual steps of the proportionality test. First, Russian-speaking pupils and Latvian-speaking pupils are in a similar situation when pursuing their education in public schools in light of the disputed reform. Second, the two aims of the disputed reforms are legitimate, namely: protection of the democratic order of the State – implying protecting and strengthening the Latvian language especially in light of the historical context and unlawful occupation and annexation of Latvia (the Court does not discuss the Russian minority in Latvia predating the occupation); and protection of the rights of the others – implying the unity of the education system to facilitate equal access for pupils to the State education system and the need to reverse the consequences of the segregation in education that had existed under the Soviet regime. Third, the Court found that the measure was not disproportionate to achieve the two aims sought: the reform had been implemented gradually and was foreseeable; and the Russian-speaking pupils could still learn their language and preserve Russian culture and identity. The Court acknowledged the absence of reliable statistics regarding the Latvian language proficiency of the Russian-speaking population of Latvia. However, in any case, according to the Court, if the applicants could speak good Latvian, then the reform does not affect them, but if it does affect them because they do not speak good Latvian, it is necessary.
The Court fully embraced the argument that protecting and strengthening the official language goes ‘to the heart of the constitutional identity of the State’ and on this basis left a very wide margin of appreciation to Latvia, considering that it is not its ‘role to question the assessment made by the Constitutional Court in that regard unless it was arbitrary, which the Court does not find in the present case’ (§208). The Court dismissed the fact that the aims could have been reached by less restrictive means, citing the lack of European consensus on the principle of instruction in an individual’s mother tongue, emphasising again the historical context and the Russification policies under the Soviet regime and the need to reverse the consequences of unlawful occupation and annexation by the former Soviet Union.
The Džibuti judgment is in essence very similar to Valiullina. Following the government thesis that private schools form part of the State educational system, the Court found its conclusions concerning the absence of discrimination regarding public schools in Valiullina to be relevant to private schools. The number of private schools in Latvia is insignificant and they mainly operate in Latvian anyway (out of 58 private schools active in 2018/2019, only 11 were teaching in Russian and 18 were teaching only in Latvian). The Court expressly recognised that its stance departs from the Venice Commission’s opinion, according to which ‘persons belonging to a national minority have the right to set up and to manage their own private educational establishments’ (§96). Latvia intends to offer 25% of its population no such right.
In Džibuti the Court examined a specific argument, which was not at issue in Valiullina: the alleged discrimination against Russian-speaking pupils compared with pupils whose mother tongue was one of the official languages of the EU. This argument was dismissed on the grounds that the facilitation of learning EU languages is legitimate because it ensures fully-fledged participation in various processes within a united Europe and is not intended to develop the culture and identity of the pupils concerned. In this context, the choice to encourage the development of proficiency in official EU languages is not arbitrary or manifestly unreasonable because Latvia is a Member State of the EU: ‘this allows the linguistic diversity of all groups of pupils to further develop, and enables them to benefit from the advantages that come from Latvia’s membership of the EU’ (§156). Finally, the Court found that the applicants failed to substantiate their claims that they were discriminated against in comparison with pupils whose mother tongue is an official language of a country with which Latvia has concluded an international agreement in the field of education, as there is no evidence that they have been treated differently.
In what follows, we exclusively focus on the deficient, in our respectful opinion, reasoning of the Court under Article 14 ECHR together with Article 2 Protocol 1, commenting mainly on Valiullina, which is of primary importance in understanding the ECtHR’s ongoing retreat from principled minority protection, and touching upon Džibuti only for what concerns the issues not tackled in Valiullina. The reasoning of the Court in these two cases denies minority rights entirely, thus leaving no trace of what was once thought to be protected by the Convention. To make matters worse, the successful attack on minority rights comes with no analysis to provide sound reasons to underpin this move, which threatens to further undermine the legitimacy of the ECtHR, especially among continental ethnic and linguistic minorities, who are being repeatedly betrayed by the Court.
The ECtHR deferred to the stance of the Latvian Constitutional Court, citing Latvia’s ‘constitutional identity’. Can the ‘constitutional identity’ of a State consist in part in denying the rights to persons belonging to a historic minority? The Court hinted at providing a positive answer in Savickis, where Convention rights were denied on the basis of a lack of a local citizenship, and ‘constitutional identity’ consisted in not granting it. Dissenters underlined in that case that: ‘Europe knows only too well by now how some States may misuse or instrumentalise arguments relating to their constitutional identity for a variety of purposes’ (see dissenting opinion of Judges O’Leary, Grozev and Lemmens §24). The way the Court used this concept in Valiullina is just as problematic, further unleashing ‘constitutional identity’ as a trump card against the Convention rights, de facto paving the way to setting aside the Convention scrutiny of national violations: quite the ill-considered move, with respect, for a Court whose only task is to ensure that States respect their Convention obligations.
It is absolutely unclear both whether constitutional identity is an autonomous concept endowed with any specific meaning under ECHR law and what triggers it under the Convention as a means for setting aside the rights the Convention is designed to safeguard. In Savickis constitutional identity meant removing from certain Convention protections those Latvian residents who had under USSR law settled in Latvia during the period of its unlawful occupation by the Soviet Union. In Valiullina the Court makes clear that ‘protecting’ and ‘strengthening’ the State language is at the heart of constitutional identity – a term that the Court does not use in Džibuti, although explicitly embracing its reasoning from Valiullina on the substance of the protection and strengthening the Latvian language. Constitutional identity is thus deployed to increase greatly the State’s margin of appreciation, suddenly justifying infringements of Convention rights. The consequence is that constitutional identity has graduated into a magic spell first tried in Savickis and able to turn upside-down the ‘very weighty reasons’ test, while disengaging proportionality in Valiullina. By giving constitutional identity such a magic twist, the Court turns the respondent state into a judge in its own case: invoking a spell to transubstantiate violations of the ECHR. Quite astonishingly, the Court appears oblivious to the fact that the ECHR is a law designed precisely to hold States accountable for such dark arts when carried out under their usual noble pretexts.
In fact, the ECtHR repeated the mistake of the European Court of Justice (ECJ), when it used the ‘national identity’ of Article 4(2) TEU as an easy fix to justify depriving minorities of rights, turning EU law into an additional tool for undermining minority protection, as we have seen in Runevič-Vardyn, for instance (as analysed here). More recently, in the Cilevičs case the concept began as mere ‘boilerplate’ ‘to reassure the Member States that their specificities have been duly taken into account’ in the words of Giacomo Di Federico and Giuseppe Martinico. Only the issue is trickier, of course: the ECJ and the ECHR are now weighing in not on the specificities, but on the struggle against specificities. Neither the ECJ nor the ECtHR provide any explanations as to why protecting and strengthening the official language against minority languages is part of ‘national’ or ‘constitutional’ identities – two concepts which also need to be distinguished from each other – which appears counterintuitive in light of Benedict Anderson’s insight that nations can be successfully imagined without linguistic commonality, even beyond the impossible reading that a constitutional identity is about suppressing minority rights.
In any case, the outcome of the constitutional identity trend is messy and incomprehensible: a broad catch-all pretext to suspend rights guaranteed by law. The only safeguard emerging from the case law is that this suspension should not be arbitrary. But since constitutional identity appears to encompass most of the elements of the proportionality test – i.e. the aim sought and the means to reach it, as expressed in the descriptive observation on the consequences of the unlawful annexation and the normative claim that Latvian should be increased and minority languages decreased (see Yordan Nugraha at 158) – there are not many other elements left to evaluate the arbitrariness of a measure, implying a ‘blanket acceptance of all measures that are claimed to pursue this particular aim’ (Id. at 157). Bad arguments are often heard, but the ECtHR could do better.
The majority’s stance in Valiullina ignores the Latvian constitutional history interbellum and is based on the assumption that protecting and strengthening the Latvian language amounts to eliminating the consequences of Latvia’s unlawful occupation and annexation by the former Soviet Union, which should in turn lead to decreasing the use of Russian at school. Education in English, French or German is not targeted by the reform. In a fascinating twist, the Court was clear that it is irrelevant whether the Russian-speaking minority speaks good Latvian – no reliable statistical data about the actual Latvian proficiency of the Russian-speaking pupils was provided. The Court does not care either that this minority’s alleged poor command of Latvian could be due to the ‘lack of high-quality teachers, inadequate methodology of teaching the Latvian language, low quality and insufficiency of educative support for language teaching as well as ineffective implementation of the compulsory proportions for the use of Latvian’ as emphasised by the Venice Commission (§71). Whether decreasing the use of Russian in schools will help increase the general population’s proficiency in Latvian does not matter either, even if it risks, in fact, altering the average quality of the teaching in the Russian-speaking minority schools. Data do not matter because Russian is to be decreased at all costs: the flippancy with which the Court dismisses the rights of the Russian-speaking minority is quite astonishing (§208). Simply put, the Russian-speaking minority is punished for speaking the language of the former oppressor in a bilingual country that pretends not to be one. This painfully echoes the outdated ‘enemy aliens’ approach of the First World War, which has been revived against Russian citizens since Putin launched a full scale invasion of Ukraine in February 2022. The Schengen visa and entry ban adopted inter alia by Latvia, is a paradigmatic example of this. It is the perceived ‘Russianness’ that emerges as a problem to be solved and all the other considerations are secondary. The only problem here is that this approach is repugnant to the letter and spirit of the ECHR, whether the Court votes for it or not. The Court thus departs from the very raison d’être of the Convention, which is a barrier against ‘the tyranny of the majority’, by adopting a skewed reasoning based on cherry-picking who should be entitled to Convention rights in the Latvian population based on a particular reading of Latvian history, which has nothing to do with Convention rights.
When dismissing the alternatives available to the authors of the reform, the Court considered that there is no European consensus concerning teaching in a minority’s mother tongue – a strange argument to be applied to a country where a huge proportion of the population speaks Russian and where there is a considerable tradition of education in this language since its proclamation of independence in 1919 (as well as before). Even in 1991, the restored Latvia, whose ethnic composition had changed tremendously since 1919, opted to provide education for minorities mainly in their mother tongue, including the Russian-speaking minority. This had been the rule and the practice since the creation of the country and until 2018. The 2018 reform changed this paradigm, justifying the setting aside of minority rights on the basis of considerations of constitutional identity related to the ‘protection’ of the Latvian language, presumably in full knowledge of the fact that Latvian has never been spoken in Latvia as much as today (see also ECRI report, §63).
The Court completely ignored the effects on the Russian-speaking minority of the discriminatory context they have been living in since the restoration of independence. The Soviet-era immigrants were initially excluded from Latvian society, falling ‘outside of both the discursive and real body of Latvian citizenry’. Still today, 9.3% of the population hold the status of ‘non-citizen of Latvia’ meaning that in practice, they are stateless and face restrictions of their rights, including to pensions and employment and all the crucial rights offered by EU law, since they are not EU citizens because the Latvian government decided to exclude this minority from the scope of EU law as not being connected to any EU Member State. NGOs and UN special rapporteurs (here and here) and European and International Committees (see here and here) have expressed their concerns for this group more generally, and in particular for the Russian-speaking minority as such in light of the 2018 and 2022 reforms. Pretending that the aim of the measure is to ‘facilitate equal access for pupils to the State education system’, is quite ironic if not cynical as it only decreases the rights of the Russian-speaking minority members, without actually having any impact on the ethnic composition of schools. It is as fake as refusing fathers the parental leave granted to mothers in the name of gender equality. It is even more cynical in light of the fact that linguistic diversity is glorified when it comes to EU citizens in Džibuti, while Russian-speaking schools imply a segregation that should be removed at all costs.
The Court felt the need to specify that what distinguishes Valiullina from ‘other cases examined by the Court as regards allegations of discrimination in access to education stemming from the existence of segregated schools or classes for members of historically and socially disadvantaged groups such as Roma people’ is the need to ensure unity in education and ‘securing equal chances to all pupils’ (§201).
It is peculiar that Valiullina is about eliminating the consequences of an alleged ‘segregated system of education’ under the Soviet Union – also implying, in the words of the Latvian Constitutional Court, that today’s Latvian educational system is segregated – considering that ‘the implementation of minority rights must not be aimed at segregation of society and endanger the unity of society’ (pt. 15.2 at 40). The Constitutional Court obviously misunderstands the meaning of the word ‘segregation’, particularly in light of the Belgian Linguistic Case. The ECtHR’s repetition of this argument brings even more confusion to a case which already fails to convince.
In the Court’s case law segregation in education designates a policy (often neutral) that keeps one group in the education system apart and treats it differently and unfavourably on the basis of a prohibited criterion such as race, sex or religion. The designated ‘segregated system of education’ under the Soviet Union does not quite fit this reality, as the russification policies, as traumatic as they were, aimed to spread the official language – Russian – among the peoples of the Soviet Union (Russians and other minorities attended schools in Russian, while Latvians went to schools in Latvian, where Russian was part of the curriculum). In this context, it is absurd to qualify as ‘segregated’ a system which protects minority rights in education, even more so as Latvia is bound by the UNESCO Convention against Discrimination in Education and the Framework Convention for the Protection of National Minorities. Aleksejs Dimitrovs is right that ‘this idea can become a universal reservation against the respect for any kind of minority rights: the unity of society is above all’.
Supporting tuition in different languages cannot in itself be an indicator of a segregated system in education. Otherwise, many countries in Europe could be accused of segregation. In fact, segregation is fed by other factors – such as ethnicity, race and economic status – with damaging consequences for the targeted group. In casu, if there were a risk of segregation, it was triggered by the 2018 amendments and reinforced by the 2022 reform, which imply disadvantaging several groups – Russians, but also Poles, Ukrainians and Belarusians, to say nothing about the Jewish – by significantly decreasing their chances to be educated in their mother tongues, with a ‘negative impact on their potential for successful socio-economic integration’ according to ECRI. In fact, had the Court approached the case through the prism of discrimination on the ground of ethnicity, as raised by the applicants, implying very weighty reasons and therefore a narrower margin of appreciation, the reasoning would have probably been quite different, as the Court would have had no choice but to take into account the strong ethnic aspect of the case, where ethnic minorities are unfavourably treated because they speak the language of a former oppressor in a multiethnic country that chooses to build its ‘constitutional’ identity by denying the fact of its own societal composition.
[i] The authors are grateful to Aleksejs Dimitrovs, Vadim Poleshchuk and Tobias Mortier for their help with this text.
Strasbourg Observers is an academic blog that discusses recent developments at the European Court of Human Rights. The editorial team of Strasbourg Observers is based at the Human Rights Centre of Ghent University and the Centre for Government and Law of Hasselt University.
Strasbourg Observers would not be possible without the valued work of countless guest bloggers. Visit the “Contact” page on our website to find out how to become one!
