Introduction

Between 31 May and 2 June 2019, Pope Francis paid a visit to Romania and made a historical apology to all Roma in Europe. He apologised for the harm inflicted on Roma by the majority populations and institutions affiliated with the Roman Catholic Church. Prominent international media in Europe and around the globe reported on the Pope's apology, among them the New York Times: ‘Pope Francis, on the last day of his trip to Romania, on Sunday asked for forgiveness on behalf of his church for the suffering endured by the Roma people, saying his heart was “weighed down by the many experiences of discrimination, segregation and mistreatment” they have experienced’ (Gillet and Horowitz, 2019). This was not the first time Pope Francis expressed remorse for the mistreatment of Romani minorities. Earlier in May the same year, he spoke against the violence inflicted on the Omerović family, who fled from BIH to Italy during the post-Yugoslav wars. When the Italian state authorities relocated the refugee Romani family from the segregated camp on the fringes of Rome to a more centrally located apartment in the city, neofascist and extreme-right groups protested against the decision. There were numerous media reports about the protests: ‘The mayor of Rome, Virginia Raggi, also visited the family to show her support after some of the protesters said they wanted to see the family “hanged and burned”’ (Lakić, 2019). The Pope welcomed around 500 Roma and Sinti to a service at the Vatican, including the family in question, in his own protest against the xenophobic attacks, much to the disdain of the Italian Deputy Prime Minister at the time, Matteo Salvini. During the service, as reported by Vatican News, the Pope said the following words: ‘Yes, second-class citizens exist here, but those who are really second-rate are those who reject people because they are unable to accept them’ (Vatican News, 2019). With this gesture, along with his previous gestures of support for refugees, Pope Francis clashed with the global extreme far-right movement and became ‘a hate figure for the far right’ (Baldwin et al., 2019).

The news of the Pope's visit to Romania and his apology to Roma also made the national headlines in Europe. The media reported it in, among others, the newest EU Member State, Croatia, whose population predominantly identifies as Catholic (Perica, 2006). However, on exactly the same day, a local event in connection to Roma was in the media spotlight in Croatia. Whilst the Pope was visiting Romania on 1 June, a protest under the banner ‘I want a normal life’ (Telegram, 2019; my translation) ¹ was taking place in the Croatian county of Međimurje, where the highest number of Roma live (Šlezak, 2009: 67). The protest, which attracted around a thousand people, was organised by two men: an eighteen-year-old supporter of the radical right-wing parties’ coalition, and a Croatian war veteran who stated, ‘I have put my life in danger, sacrificed in war, but I do not want, speaking in the name of all peaceful and hard-working citizens of Međimurje, to suffer the Roma terror, which has lasted over twenty years’ (Međimurjepress, 2019; my emphasis and translation). ² The organisers claimed that their protest was to highlight the problem of petty crimes alleged to be mostly committed by the local Roma in Međimurje. They claimed that Romani petty criminals were violating the human rights of the local majority population, especially the right to a normal life (Vlasić, 2019). Whilst some local Romani representatives supported and even joined the protest, international NGOs, such as the ERRC, and the member of the Croatian parliament representing the Romani minority, Veljko Kajtazi, called it ‘pure racism’ (Duhaček, 2019). The Pope's apology in Romania and the local protest in the Međimurje county were reported side by side on the same day, and yet the connection between the two was never made, either by the media or by any of the local Croatian politicians.

The Croatian Interior Minister at the time, Davor Božinović, came to Međimurje county and promised extra police support to patrol Romani settlements, but emphasised that much more needed to be done to improve the situation in the county, especially in the education sector. Government websites too echoed Božinović's comments: ‘The key to solving the Roma problematics in Međimurje is education’ (Government of Croatia, 2019; my translation and emphasis). ³ Božinović declared that in this situation where only four out of 318 Romani high school pupils finished their schooling, ‘there must be something terribly wrong. These young people have disqualified themselves from the race of life so they are not equal any more, they do not fulfil the basic conditions for jobseekers, and this problem needs to be solved’ (Government of Croatia, 2019; my translation). ⁴ Some newspaper commentators agreed with the minister: ‘Božinović is right, Roma have disqualified themselves!’ (Vlasić, 2019; my translation). ⁵ Yet what was absent from these debates was the fact that Croatia had been charged with discrimination against Romani children in education in 2010 by the ECtHR, less than a decade previously. Similarly, as in other European countries, the minister and the public in Croatia engaged in the ‘politics of forgetting’ (Gándara, 2005) and returned to the view that Roma themselves were responsible for their own marginalisation. Whilst the Pope publicly apologised for the segregation of Roma, none of the political representatives of the states where Romani children were segregated in schools made a similar gesture, and they instead sought to legitimise it as a benevolent approach (O’Nions, 2010).

This chapter looks at the ‘citizens in the making’ in education and analyses the invisible edges of citizenship on which state authorities rely to argue that they offer equal opportunities for all of their citizens. By examining the segregation of Romani children, it highlights how through such invisible edges of citizenship the educational system not only maintains but also actively constructs the civic marginalisation of minorities such as Roma. The chapter shows how governments have legitimised the segregation of Romani children in public educational systems. It argues that policies based simply on either socio-economic disadvantage or ethnic discrimination are often myopic and cannot fully explain the dynamics and reasons behind such marginalisation. The chapter explores the justifications that state lawyers have used in ECtHR court cases on educational segregation of Romani children. Previous scholarly accounts have examined the successful arguments that advocates of Romani applicants have used to prove discrimination (Goldston, 2010). This chapter focuses specifically on the following ECtHR cases: 2007 D.H. and Others v. Czech Republic (henceforth the D.H. case); 2008 Sampanis and Others v. Greece (henceforth the 2008 Sampanis case); 2010 Oršuš and Others v. Croatia (henceforth the Oršuš case); and 2012 Sampani and Others v. Greece (henceforth the 2012 Sampani case). ⁶ In all these cases the states justified segregation as a legitimate non-discriminatory measure: they denied that spatial separation equals discriminatory segregation and that it had been organised specifically for Roma so that they could ‘catch up’ and secure the same chances of inclusion through education as the majority population. I claim that the practice of segregation has been structured through the invisible edges of citizenship in such a way that it actively creates the fringes of citizenship. Placing Romani children in separate educational facilities contributes to the positioning of Roma on the fringes of citizenship. Besides analysing governmental discourses that legitimised segregation, I also look at the dissenting opinions of ECtHR judges who disagreed that Romani children had been discriminated against in school systems, focusing in particular on the dissenting opinions in the D.H. and Oršuš cases. In the two cases from Greece, the ECtHR Chamber judges reached unanimous decisions.

Contemplating the societal implications of educational discrimination

The link between educational opportunities and the social inclusion of Roma has been well documented and widely discussed in academic literature (Rostas and Kostka, 2014; O’Nions, 2015; Van Den Bogaert, 2018). More broadly, different liberal theories from Rawls (1971) to Nozick (1974) have underlined the importance of equal opportunities, as have international organisations (Gatti et al., 2016) and NGO activists (Albert, 2019), However, not everyone sees the greater value in embracing equal opportunities and striving for social inclusion, as Albert (2019: 44) notes: ‘It is regrettable that whilst many changes to education policy have been undertaken in response to the need to make educational opportunity more equal, the [Czech] Government has done a poor job of explaining to the public why these changes will actually improve education for all children, not just for those who have been excluded and educated in segregated, separate settings.’

International policymakers have taken equality of opportunity as a theoretical assumption, such as the World Bank, which published the report Being Fair, Faring Better: Promoting Equality of Opportunity for Marginalized Roma (Gatti et al., 2016). This report followed the logic that education in liberal democracies should serve as a corrective mechanism to mitigate the inequalities faced by children from disadvantaged backgrounds, who are still ‘citizens in the making’. In addition, it argued that equality of opportunity was not simply a moral responsibility for the state but also a rationale for economic growth, especially in Central and Eastern European (CEE) states with significant Romani minorities:

The report used economistic logic to justify the inclusion of Roma and portrayed it as a possible benefit for the entire wellbeing of wider society, including majority populations. This logic was also prevalent in EU policies on Romani integration, including the EU NRIS Framework (Goodwin and Buijs, 2013; Iusmen, 2018). The argument that desegregation has potential benefits for state economies was not an original one: in the US, different actors had made similar arguments around half a century earlier in the context of racially segregated education. In 1947 President Truman's committee issued a report in which it stated ‘three reasons, why civil rights abuses should be redressed: a moral reason – discrimination was morally wrong; economic reason – discrimination harmed the economy; and an international reason – discrimination damaged US foreign relations’ (Dudziak, 2011: 79).

According to the economistic perspective contained within the World Bank report, the main problem Roma face with regard to educational discrimination is the fact that it leads to their exclusion from the labour market later in life. The logic here is inherently neoliberal: the exclusion of Roma from the labour market is damaging for the local economy as it leads to more dependency on social benefits and welfare. At the same time, the description of the educational system as reproducing the labour force echoes a classic neo-Marxist text by Louis Althusser (1971) entitled The Ideological State Apparatus, where school serves as a mechanism for ideology production which is necessary for the continued functioning of the capitalist system. However, some anthropologists, such as Judith Okely (1994: 71), have argued that whilst Romani minorities have been outcasts from the capitalist system, this can be perceived as liberating rather than as a disadvantage: ‘Their outsider status is imposed, but also chosen.’ Okely refers specifically to Althusser and to schooling as a state's ideological apparatus, but claims that Romani minorities have actively avoided being included in educational and labour market structures; rather, they have formed alternative and sometimes competing niches in the official systems, and that is why state officials have particularly scrutinised their position (Okely, 1994: 66–7).

However, when considering these claims, I argue that the segregation of Roma children in schools needs to be analysed in terms of civic marginalisation in a European as well as a more global perspective, which goes beyond the question of socio-economic calculus and ethnic discrimination. It is also problematic to analyse it simply as an individual or even group choice without understanding the context in which these choices have been made. I claim that the case of school segregation shows that Romani minorities are not only discriminated against because of their socio-economic status and ethnic belonging but also discriminated against as citizens.

Global parallels

Legal activists and advocates focusing on the educational segregation of Romani children have drawn their inspirations from the earlier US Supreme Court cases on the school segregation of African American children, especially the 1954 case Brown v. Board of Education of Topeka (henceforth the Brown case) (Bader Ginsburg, 2005; Greenberg, 2010; Minow, 2012; Goldston, 2017). Indeed, the D.H. case has been referred to several times as the ‘European Brown case’ (ERRC, 2019; Chang and Rucker-Chang, 2020: 122). This chapter compares the reasoning of the states in question, which claimed that segregation does not mean discrimination and, in both cases, argued along the lines of ‘separate, but equal’. However, whilst there have been a number of works published that show the similarities between the two cases, I also highlight important differences that have not been raised until now. The most obvious difference lay in the question of addressing direct and indirect discrimination (see Hepple, 2006; Farkas, 2014; O’Nions, 2015; Arabadjieva, 2016, Chang and Rucker-Chang, 2020: 129): in the US, segregation was entrenched in law with the 1896 Plessy v. Ferguson decision of the Supreme Court, whilst the segregation of Roma did not have a legal basis but was present in practice. Nevertheless, some legal scholars have argued that the Roma segregation cases did amount to direct discrimination according to EU law (Farkas, 2014). Finally, from the perspective of connected sociologies (Bhambra, 2014), this chapter looks at the segregation cases in the frame of global politics at the time. Whilst some scholars point out that the US desegregation process has to be understood within the broader Cold War context (Dudziak, 2004; Bader Ginsburg, 2005; Goldston, 2017), Romani school segregation cases played an important role in shaping the global politics of postsocialist transitions and EU integration (Chang and Rucker-Chang, 2020: 38–51). Nonetheless, global politics has not had such a decisive role for these European cases as it has in the American context, even to the present day.

This chapter looks at how the invisible edges of citizenship are produced and reproduced through the education system and how they put marginalised minorities on the fringes of citizenship. It argues that discrimination created by the segregation of Romani children in state education systems can be considered one of the most prominent examples of the invisible edges of citizenship. In a similar vein to the pronouncements made by the Croatian minister Božinović discussed earlier, in the broader public discourse education is often described as a corrective mechanism for the inequalities that children inherit from their parents. However, in this chapter I show how the educational system can also create new inequalities and contribute to civic marginalisation despite claiming to use benevolent approaches towards marginalised children from minority backgrounds. Civic marginalisation does not simply happen to migrants or citizens with a migrant background as a result of their inherent cultural differences and their belonging to a different society incompatible with liberalism, as Taylor (1994: 62) claimed. The cases of both African Americans and Romani children show that liberal democratic states produce this difference by imposing invisible edges of citizenship against the minorities who have been part of these wider societies for centuries.

US school segregation

In 2018 two Harvard academics, Margareta Matache, a Romani scholar and activist, and Cornel West, an African American civil rights scholar and activist, wrote an article which argued that ‘Roma and African Americans share a common struggle’ (Matache and West, 2018). The piece mostly focused on the parallels with regard to the slavery that Roma endured on the territory of present-day Romania until the mid-nineteenth century:

Besides slavery, scholars have also highlighted other parallels between the position of Roma and other marginalised minorities, such as the spatial segregation of Aboriginal people in Australia (Armillei and Lobo, 2017; Taylor et al., 2018), the high incarceration rates of Māori in New Zealand (Takacs, 2017) and the coercive sterilisation of women belonging to First Nations of Canada (Stote, 2015). Yet the parallels that have received the greatest attention among scholars and activists is that with educational segregation. These parallels have been noted by Jack Greenberg (2010), a lawyer in the original Brown case, Ruth Bader Ginsburg (2005), an associate justice of the US Supreme Court until 2020, and the lawyer in both the D.H. and Oršuš cases, James Goldston (2017).

Whilst there have been several cases dealing with educational segregation in the US, the one receiving the most attention as relevant for the Romani movement and legal activism is the Brown case of 1954 (Goldston, 2017; Chang and Rucker-Chang, 2020). The Brown case overturned the decision of 1896 Plessy v. Fergusson (henceforth the Plessy case), which upheld the policy of ‘separate but equal’ (the understanding that the separate facilities for African American do not mean unequal treatment) as the legal doctrine on the federal level. Whilst the ‘separate but equal’ policy of the Plessy case did not deal with education, it did influence segregated schooling, and it soon not only became a practice but was institutionalised legally. The Brown case overturned Plessy with the decision that the ‘separate but equal’ position was in breach of the Fourteenth Amendment (the provision on equal treatment). The court in the Brown case reached the following conclusion:

Whilst it was a landmark legal case, Brown did not ignite a huge desegregation process. There was major opposition to the desegregation process, as witnessed, for example, in the case of the so-called Little Rock Nine, a group of nine African American students who were blocked from entering the previously all-white Little Rock High School in Arkansas in 1957 by the state governor, Orval Faubus (Epperson, 2014). The US President at the time, Dwight Eisenhower, had to federalise Arkansas's national guard so that it could protect those nine students from protestors at the entrance in order to execute the Supreme Court decision. As Epperson (2014) argues, Governor Faubus initially stated that he would not give up his state power to federal authority. Subsequent education laws in Arkansas were amended to prohibit segregation in theory, yet in practice the desegregation process never took place. One of the reasons for this was that some African American parents feared that their children would not be safe in mixed schools. In the subsequent 1958 case Cooper v. Aaron , Governor Faubus changed his tactics and attempted to suspend desegregation on the basis of the very protests themselves, claiming that the safety of the African American children in question could not be guaranteed. The court did not endorse these arguments:

The en masse desegregation process started only gradually, but it did yield positive results despite the initial disbelief, as Pettigrew (2004: 525) notes: ‘[f]rom 1970s to 1990s, Black high school completion rates rose sharply. Although less than half of Black students finished high school in 1950, the percentage now approaches that of White students. During these same years, the mean difference between Black and White achievement test scores steadily narrowed.’ However, a number of authors, including Danielle Holley-Walker (2004) and Wendy Parker (2000), have argued that several cases from the 1970s onwards almost prompted a re-segregation, starting with the 1974 Milliken v. Bradley case (hereafter the Milliken case) – a Supreme Court case on the planned desegregation of state school students across school districts by means of busing in Detroit, the city with the greatest African American majority in the US – that departed from Brown by arguing that de facto district segregation did not represent de jure segregation (Pettigrew, 2004: 523). As Orfield and his colleagues (1994) note, the separate de facto segregated facilities were a reality not only for African American students, but also for Hispanic pupils.

From American to European segregation cases

A number of prominent US lawyers have claimed that European court cases concerning the segregation of Romani children in education have been influenced by desegregation developments in the US. For example, Bader Ginsburg (2005: 501) declared: ‘Brown's example or inspiration on the tight tie between education and democracy, and on the role courts can play in advancing change in long-standing societal structures, is evident in a current controversy concerning the schooling of Romany children in Central and Eastern Europe.’ Parallels were also pointed out by Jack Greenberg, the original lawyer in Brown, who himself took an interest in the segregation of Romani children in his later life: ‘After D.H. was decided, many European lawyers and Roma rights advocates referred immediately to Brown’ (Greenberg, 2010: 940). However, Greenberg also underlined an important difference between the two cases: ‘whilst the Brown case was arguing against the existing legal doctrine of “separate but equal”, the European cases against Roma segregation were claiming that the states in question are not abiding by their anti-discrimination legislation’ (Greenberg, 2010: 940–1). Another American legal scholar, Bob Hepple, claimed that whilst Brown was about direct discrimination within the law itself, Roma segregation cases were about ‘disparate impact’ or, in the European context, ‘indirect discrimination’ (Hepple, 2006: 612; see also Chang and Rucker-Chang, 2020: 129). In this regard, Hepple argued that D.H. bore more resemblance to the Supreme Court case Washington v. Davis of 1976, which ruled that laws that were not established with racially discriminatory motives are valid (Hepple, 2006: 612). However, the EU's Race Equality Directive carries a slightly different view on indirect discrimination vis-à-vis US case law. According to Article 2 of the Directive, indirect discrimination occurs ‘where an apparently neutral provision, criterion or practice would put persons of a racial or ethnic origin at a particular disadvantage compared with other persons, unless that provision, criterion or practice is objectively justified by a legitimate aim and the means of achieving that aim are appropriate’ (European Commission, 2000, a. 2). Whilst many have celebrated the D.H. case as an achievement, other legal scholars have been particularly critical of how its decision was formulated and concluded that this case was in fact a missed ‘European Court of Human Rights at Brown v. Board of Education moment’ (Goodwin, 2009: 93). The ECtHR cases on the segregation of Romani children in schools did not raise the same question as Brown in the US context did: rather than examining whether segregation is an acceptable practice, all the ECtHR cases asked when segregation represents an unequal treatment and discrimination of the marginalised minority (Goodwin, 2009; Arabadjieva, 2016). In the next section, I will examine how the European governments did not deny the existence of segregation but, instead, argued that it was a justifiable practice because it was not based on ethnic or racial grounds. In some cases, like 2008 Sampanis and Oršuš, government representatives went further and argued that segregation is actually beneficial for children who belong to marginalised minorities.

Narratives on justifiable segregation in Europe

The D.H. case received an ECtHR Grand Chamber ruling in 2007, after the Czech Republic had joined the EU. The case was initially examined by the Constitutional Court in the Czech Republic in 1999, but this court did not find discrimination on ethnic grounds. The representatives of the applicants from the town of Ostrava claimed they had been discriminated against on ethnic grounds as they had been placed in special schools for children with learning disabilities solely because they were Romani. The applicants’ lawyers argued that the Czech Republic violated Article 14 (Prohibition of Discrimination) and Article 2 of Protocol 1 (Right to Education) of the European Convention on Human Rights (ECHR). Whilst officially there was no law stating that Romani children should be placed in such schools, the applicants presented the court with the following statistics:

The Czech government argued that the placing of Romani children in separate schools was not a violation of the European Convention. Initially, it argued that school placement was based solely on the ‘intellectual capacity’ of the child applicant rather than their belonging to a particular ethnic or social group. Furthermore, it maintained that it had the consent of the parents of the children who were placed in special schools for this action. The government argued that according to the Czech legislation, oral parental consent was adequate in these cases (D.H. and Others v. the Czech Republic, 2007, para. 46). The court heard opinions from a number of organisations working in human rights protection, including the CoE's Commissioner for Human Rights, who underlined that such practices of discrimination diminish equality of opportunity for Romani children because they hinder their success in the labour market later in life (D.H. and Others v. the Czech Republic, 2007, para. 50). It also cited a UK asylum law case heard in the House of Lords that effectively confirmed that discrimination against Roma had occurred when the UK's immigration officer did not allow Roma to travel to the UK more often than other Czech citizens, because of an assumption that they would seek asylum in the UK (para. 105). ⁸ The ECtHR referred to a US Supreme Court decision in the case of Griggs v. Duke Power Co., which stated that there was a disparate impact on African American applicants with lower education (D.H. and Others v. the Czech Republic, 2007, para. 43).

The Czech government argued that the applicants did not use all the domestic legal instruments in this case. For example, it claimed that none of the parents appealed against the placement of their children in special schools. However, its main argument was that although there were many Romani children in special schools, these schools had been established for all children with specific learning disabilities. The government argued that this did not discriminate against Romani children but that, rather, it did the opposite: it gave more equal access to education for all children in a similar position (D.H. and Others v. the Czech Republic, 2007, paras. 110–23). However, the applicants replied that although the intention to discriminate could not be directly proven, Romani children were disproportionately affected and the government had to prove that it did consider protection against discrimination (para. 130) rather than the applicants having to prove they had been subjected to unequal treatment. The court agreed with this and shifted the burden of proof onto the Czech government.

The government's lawyer argued that the state representatives were aware that it was not necessarily the case that having anti-discrimination legislation was always effective and was always implemented in practice. However, he also repeated that in the case of special schools, ethnic background did not play any role in the placement of children. He reiterated that the main criteria for placement were intellectual capacity and parental consent and, on that basis, he also claimed that Romani parents were not particularly interested in the education of their children (D.H. and Others v. the Czech Republic, 2007, para. 197). He rejected the statistics presented by the applicant's lawyers, arguing these did not constitute adequate or conclusive evidence of discrimination since they were taken from interviews with head teachers as opposed to official state data (para. 148).

According to the numerous reports cited in the court, all the applicants belonged to a vulnerable minority, and so the ECtHR decided against examining the individual appeals of applicants but concluded that all the children's parents lacked capacity to give informed consent about their children's schooling. The Grand Chamber decided that the Czech government had violated Article 14 of the ECHR in conjunction with Article 2 of Protocol 1:

Thirteen ECtHR judges agreed with this decision, with four expressing their dissenting opinions. Boštjan Zupančič, the judge appointed by the Slovenian government at the time, expressed a dissenting opinion in defence of the Czech policies. He suggested that it was inappropriate to blame the Czech Republic for discrimination against Romani children when it was the only country willing to work to improve their position. He claimed that the ECtHR was being abused for political games rather than serving justice (Zupančič in D.H. and Others v. the Czech Republic, 2007). Karel Jungwiert, a judge appointed by the Czech Republic, presented a more elaborate dissenting opinion defending his own country. He claimed that the Czech Republic had guaranteed the right to education for Romani children where older EU Member States before 1989 had failed to do so. Citing CoE statistics, he argued that a large proportion of children belonging to Romani minorities in Western Europe were never included in any education system at all, but that his country had managed to overcome this exclusion: ‘In a way, the Czech Republic has thereby established an educational system that is inegalitarian. However, this inegalitarianism has a positive aim: to get children to attend school to have a chance to succeed through positive discrimination in favour of a disadvantaged population’ (Jungwiert in D.H. and Others v. the Czech Republic, 2007). He then asked the following question: ‘which country in Europe has done more, or indeed as much, in this sphere?’ (Jungwiert in D.H. and Others v. the Czech Republic, 2007).

In his dissenting opinion, Judge Borrego Borrego highlighted the discrepancies between the case and the court's decision. He criticised the court's decision not to examine individual cases and its conclusion that none of the parents in this case had the capacity to decide about their children's schooling. He claimed that the latter echoed earlier practices of rulers in Europe who had taken Romani children away from their parents, who they proclaimed to be unfit to decide about their children's education. To fight discrimination, Borrego Borrego underlined, the court itself was essentialising the position of Romani parents as all being unfit to make decisions about their children's education. Furthermore, he questioned whether the lawyers themselves really represented the interests of the applicants in question, asking whether they were too far removed from them and had ulterior motives that had very little to do with the applicants themselves: ‘An example of the sad human tradition of fighting racism through racism … How cynical: the parents of the applicant minors are not qualified to bring up their children, even though they are qualified to sign an authority in favour of British and North American representatives whom they do not even know!’ (Borrego Borrego in D.H. and Others v. the Czech Republic, 2007).

The next case dealing with the segregation of Romani children that reached the ECtHR was the 2008 Sampanis case. It was heard by the ECtHR Chamber, which decided that Greece, an older EU Member State, had discriminated against Romani children when it put them in a separate annex of the school they attended. The government's argument in this case resembled that in the US Supreme Court case Cooper v. Aaron of 1958. In 2004, with the support of the ERRC amongst other NGOs, Romani parents from the Psari camp wanted to enrol their children into primary schools in the nearby town of Aspropyrgos. According to the applicants, the schools refused to enrol the children: the schools’ justification was that the relevant ministry did not require them to enrol any Romani children, but if the order was issued, they would send the invitations to the parents in question. After a visit by the ombudsperson, who responded that there was no justification for the Romani children from the Psari camp not to be enrolled in those schools, their enrolment did begin in 2005 (Sampanis and Others v. Greece, 2008, para. 8). However, at the start of the academic year, when Romani children arrived to attend the Aspropyrgos school that they were assigned to, they were met with protests from non-Romani parents who prevented them from entering. On one occasion the non-Romani parents put up a sign on the school that stated, ‘The school will remain closed because of the Gypsy problem’ (para. 19, my emphasis). On other occasion, the Romani children could enter the school only when escorted by the police (para. 20). The school authorities decided to move the children to a separate building to protect them from the protests of the majority population (para. 23). In this separate building they attended ‘preparatory classes’ that would help them to attain the level required for mainstream schooling in the future (para. 29).

The argument for such a placement was that it was implausible from a ‘psycho-pedagogical’ perspective that these children would mix with other children: most of the Romani children were older than the other children in the same grades. As well as not having attended school previously, they were, according to the school authorities, not able to mix with other children. The official reason that the school gave to the ministry was that there was not enough space and that the parents had agreed that the children could be in a separate building in order to be closer to their camp. The government lawyer repeated these arguments and claimed that the classes in question were not put there to segregate Romani children, but to offer additional help so that they could catch up with other children before being placed into ordinary classes (Sampanis and Others v. Greece, 2008, para. 61). However, the court noted that the Greek government had not proven that any of the children would be transferred into the mixed classes after successfully attending the preparatory classes (para. 90). The Chamber unanimously decided that the Greek government was not providing an acceptable justification for the differential treatment of Romani children, as the preparatory classes were based mainly on the criteria of children being of Romani origin. In this case, there were no dissenting opinions from any judges (para. 98).

Both the D.H. case and the 2008 Sampanis case were referred to in the Oršuš case. At the time both of the hearings and of the decisions (in 2009 and 2010 respectively), Croatia was a not an EU Member State but a candidate country. Following the 1993 Copenhagen criteria, the protection of minorities such as Roma was one of the topics highlighted during Croatia's accession period (Bračič, 2016; Sardelić, 2019c). In the 1990s, during the post-Yugoslav wars, which significantly affected Croatia, Roma were not the main minority of concern for international organisations. Instead the organisations focused more on the Serbian minority, because of the potential territorial instability and because this group had been recognised as the ultimate Other in Croatia (Sardelić, 2015: 168; Vermeersch, 2006: 3). The position of Roma was also not a primary concern for minority protection in the post-conflict and postsocialist transition, despite the fact that a number of international organisations and NGOs (such as the ERRC, Open Society Foundation and CoE, for example) had urged Croatia to stop the educational segregation of Romani children. This, perhaps unsurprisingly, changed during Croatia's negotiations for EU membership (Sardelić, 2011; 2019).

The Oršuš case was initially addressed by the local courts in Croatia and finally by the Constitutional Court. The representative of fifteen applicants of Romani origin from Međimurje county claimed that their right not to be discriminated against (Article 14 of the ECHR) and their right to education (Article 2 of ECHR Protocol 1) had been violated (Oršuš and Others v. Croatia, 2010, para. 4). In contrast to the D.H. case and the 2008 Sampanis case, here Romani children were not placed in separate educational facilities (different schools or separate annexes), but in Roma-only classes in the same schools as majority children (para. 10). The main argument from the government was that they were not placed in these classes because of their ethnicity but because they did not speak the majority language (that is, Croatian) competently. According to the government's lawyer, these were so-called ‘catch up’ classes to give Romani children extra support with their Croatian language skills (para. 123).

Like the Croatian courts, the ECtHR Chamber concluded in 2008 that there had been a violation of ECHR Article 6 (Right to Fair Trial) owing to prolonged judicial procedures (Oršuš and Others v. Croatia, 2010, para. 99). However, the ECtHR Chamber did not find discrimination or the violation of the right to education. In 2009 the Grand Chamber revisited the case. The applicants represented at the ECtHR Grand Chamber attended two primary schools in Međimurje county: Podturen and Macinec. In the school year of 2001, 47 Romani children were enrolled in the Podturen. The school had the 463 students in total, of whom 47, or around 10 per cent, were Roma. Of these 47 Romani students, 17 were placed in a Roma-only class and 30 in mixed classes. The school authorities reported that there were no more Roma-only classes two years later. The school included a statement from a Romani assistant who was working in a Roma-only class to prove that members of the Romani community themselves were aware of the language issues among pupils. The assistant stated that in addition to their inadequate command of the Croatian language, she had observed many behavioural issues and poor school attendance which were due to the disinterest of parents (paras. 11–17). Macinec primary school also called upon Romani assistants, but the statistics for Roma-only classes here were different: there were 194 Romani pupils out of a total intake of 445 (43.5 per cent) in the school year of 2001. Of these, 142 Romani pupils were in six Roma-only classes whilst the remaining 52 were in mixed classes. One of the justifications the school offered was that in the lower four grades, Romani children represented 50 to 75 per cent of the total number of all students in those classes (paras. 11–17).

This case ended up at the ECtHR after the expedited case at the Constitutional Court of Croatia in 2007. The initial case was lodged in the lower courts in 2002. The Constitutional Court acknowledged that the placement of Romani children in separate classes was an exception rather than a rule, since 93 per cent of Romani children in Međmurje county attended school with their non-Romani peers. It argued that the qualified experts had placed the children in the classes that they believed would benefit the children's development the most: ‘[s]eparate classes were not established for the purpose of racial segregation in enrolment in the first year of primary school but as a means of providing children with supplementary tuition in the Croatian language and eliminating the consequences of prior social deprivation’ (Oršuš and Others v. Croatia, 2010, para. 60). In the case at the ECtHR Grand Chamber, the court implicitly agreed with the assumption that spatial separation itself does not necessarily equate to a breach of the right to education and discrimination. The parties involved had to prove or disprove discrimination, which according to the ECtHR meant ‘treating differently, without an objective and reasonable justification, persons in relevantly similar situations’ (para. 149).

The Croatian government representatives at the ECtHR argued that there was no discrimination and sought to show that there was in fact an ‘objective and reasonable justification’ for putting the children in question in separate classes from their non-Romani peers. First, she claimed that the school was not enrolling Romani children and Croatian children as such, but was testing their knowledge of the majority language, which was the necessary prerequisite for attending the school in Croatia. She built on the argument of the Constitutional Court that these classes were an exception that applied only to the schools in question. The claim was that in the case of these two schools, the board of experts, after careful deliberation and after assessment of each individual child, had decided that the applicants’ knowledge of Croatian was insufficient to allow them to be placed in the same classes as the majority of children. The government representative tried to prove that the best remedy for addressing the language inequality of the children in question was to apply differential treatment in the form of separate classes (Oršuš and Others v. Croatia, 2010, para.123). She stated that the quality of education in Romani classes did not vary beyond that permissible by legislation, which permits the omission of up to 30 per cent of the curriculum (para. 163). Furthermore, the government representative argued that such classes were needed to develop the basic social skills of Romani children, as, in comparison with Croatian children, they were lacking in such skills (para. 123), and stated that the teachers in these particular schools had to deal with poor attendance, low interest among parents in their children's education and disruptive behaviour from the children who were put in the Roma-only classes (para. 13). The representative also claimed that the parents of the children in question did not object to the placement of their children in the Roma-only classes. She also asserted that Romani children in Roma-only classes had numerous opportunities to mix with other children, especially during extra-curricular activities (para. 122), and that the schools in question had showed respect for Romani culture by celebrating International Romani Day (para. 135).

However, the applicants’ representatives argued that putting children in Roma-only classes translates into other forms of spatial segregation in later life as it normalises parallel societies of Romani and the majority population. They presented statistics showing that only 16 per cent of Romani individuals finished primary schooling in comparison to 91 per cent of the total population in the school year 2006/07 (Oršuš and Others v. Croatia, 2010, para. 18) and also that in the present environment the vast majority of Romani children did not have friends who were non-Roma (para. 53). They argued that for the Romani children to efficiently learn a language they needed to have maximum opportunities to mix with all other children, which in this case they did not have. The applicants’ representatives also stated that their knowledge of the Croatian language was never properly tested and that their right to education was violated since the curriculum in Roma-only classes was significantly reduced. The representatives doubted that the separate classes helped the Romani children to progress in the Croatian language when they only sporadically included additional Croatian lessons. Further, the lawyers questioned whether these classes were transitional given that no Romani children transferred to mixed classes throughout their schooling period. The government representative responded to that point by claiming that classes were not broken up, because evidence suggested it was important for the children's development to ensure the homogeneity and stability of the class throughout their (lower-grade) schooling ( para. 119).

The ECtHR Grand Chamber agreed with the government representative and the previous conclusion of the Constitutional Court that spatial separation does not necessarily mean discrimination or a breach of the right to education. However, it questioned whether proper safeguards had been put in place to ensure that these measures would not disproportionally affect Romani children. It reiterated that the special position of Romani minorities needed to be considered (Oršuš and Others v. Croatia, 2010, paras. 147–8). It brought forward statements from various international expert bodies, including the CoE's Human Rights Commissioner at the time, who had visited Croatia in 2002. His statement was presented to the court:

The ECtHR Grand Chamber asked whether it was only Romani children who were placed in separate classes or whether there was any evidence that other children who lacked Croatian language skills were also placed in such classes. The Grand Chamber could find no evidence that there were any other children in separate classes at the time besides Roma (para. 158). It also concluded there were no tests for specifically assessing the children's knowledge of the Croatian language, but that the assessment was based on a broader psychological test on school entry (para. 159). Another contradiction was that some of the applicants were initially in the mixed class but were only later put into the Roma-only class, which seemed to indicate that the teachers did not notice initial difficulties with the Croatian language, but this seemed unlikely to the court (para. 161). Furthermore, when the government representative claimed that the curriculum was the same in both cases, the court asked why, if that was the case, the children needed to attend separate classes. The court indicated that there was no evidence of progress reports for children who lacked Croatian language skills, as well as no evidence that they could be transferred into the mixed class if their progress was sufficient:

However, eventually the Grand Chamber made the following decision:

In a marginal vote of 9 to 8, it decided there had been a breach of ECHR Article 14 in conjunction with Article 2 of Protocol 1. Eight judges – from the Czech Republic, Croatia, the Russian Federation, Armenia, Germany, the Netherlands, Monaco and Montenegro – wrote a joint partial dissenting opinion that stating that they did not agree that there was no justifiable reason for the separation of the Romani children in question and their placement into Roma-only classes, as the applicants did not state that their command of the Croatian language was adequate. They also reiterated that these classes were established to address the ‘special needs’ of children and were benevolent, and could be understood as a type of affirmative action. They argued that these actions applied only to the children who had difficulties with the language, not to all Romani children. The dissenting judges emphasised that the efforts of the schools in question could not be denied: the schools organised parent–teacher meetings to deal with low school attendance, employed Romani assistants and redirected the blame for the disinterest in schooling to the parents. They argued that the authorities had done their best to address a culturally specific minority and that the judgement did not mention the rights of majority ‘Croatian-speaking children’ and how their education would have been disrupted if they had been in a classroom with a large number of children with ‘special language needs’ (dissenting opinion in Oršuš and Others v. Croatia, 2010). The dissenting judges stated that, rather than the Romani minority being looked at as a whole, there should be more focus on specific cases. They claimed that the ECtHR Grand Chamber had overstepped its role in these matters and that it went against a well-reasoned judgement of the Constitutional Court of Croatia (dissenting opinion in Oršuš and Others v. Croatia, 2010).

After Oršuš had been concluded, the ECtHR Chamber had to decide on another Greek case, which implied that the Greek government was not abiding by the initial 2008 Sampanis judgement. The segregation of Romani children continued, and in 2009 a new Sampani case was presented to the ECtHR. The 140 applicants were represented by the Greek Helsinki Monitor, and some of them were the same applicants as in the initial Sampanis case (lodged in 2005). After the initial 2008 Sampanis case had been concluded, the non-Romani parents and the Mayor of Aspropyrgos (the town where the original case took place) addressed a letter to the Greek Ministry of Education. The letter was presented to the ECtHR Grand Chamber in the Sampani case:

The ECtHR Chamber unanimously decided that Greece had violated ECHR Article 14 in conjunction with Article 2 of Protocol 1.

Shifting global contexts of racial segregation: from Cold War to postsocialist and EU integration narratives

Comparing the two examples of strategic litigation, James Goldston, who was one of the lawyers in both the D.H. and Oršuš cases, commented that in the European context segregation was met with less political disapproval than in the US: ‘While the European Court's condemnation of segregation was important, it is less revolutionary than Brown – and may have less immediate impact on the public consciousness – at least in part, because the law had not been entwined with segregation in Europe as it has been in the United States’(Goldston, 2017: 182). Yet the other reason, as Goldston mentions too, was the Cold War as the geopolitical framing in the period of the Brown case (Goldston, 2017: 182). Dudziak (2004) has similarly confirmed in their legal historical research that the Cold War played a significant role in cases such as Brown and the American civil rights movement more generally (Dudziak, 2011): segregation of African Americans did not put an impressive international light on the US as a global moral leader, especially on the African continent, where it was rivalling the then Soviet Union for influence. The Soviet Union used the case of US segregation for its political gains in the Cold War as well. Whilst Brown itself did not directly cause the desegregation process, it served as a signal to the world that the US had deemed segregation to be ‘unAmerican’ (Dudziak, 2004: 34–5). The American media described the Brown case as ‘the blow to Communism’ (Dudziak, 2004, 35). At the same time, there was initially massive public resistance to desegregation at the local level, some even describing desegregation as a communist plot (Dudziak, 2011).

The question is whether the two contexts of desegregation can be directly compared, as they occurred in different time periods. However, it is clear that countries on the other side of the Berlin Wall were similarly concerned with their public image with regard to discrimination. That is why, according to Donert (2017: 54), socialist Czechoslovakia gave special attention to how Roma were treated and to the public image of their position within Czechoslovak society. Other socialist countries, such as the non-aligned Yugoslavia, had divergent policies towards Romani children's schooling and integration: whilst the wealthier republics of Slovenia and Croatia clearly practised segregation of Romani children, North Macedonia and Kosovo embraced ethnic integration policies (Sardelić, 2016). Media at the time, however, portrayed the position of Roma as improving as part of the development of socialism (Sardelić, 2016). However, since school segregation was not enshrined in law, even during the socialist period, it remained invisible. After the fall of the Berlin Wall, the US State Department scrutinised the position of Roma in countries such as the Czech Republic, describing it as the remains of human-rights-violating communist regimes (Schlager, 2017). However, as Donert (2017: 247) showed, the 1990s human rights violations were also a product of the human rights vacuum created during the transition period.

As segregation was not enshrined in law and did not happen systematically to all Romani children, the European cases resembled the 1974 Milliken case in the US, which is now studied as a case that deems de facto segregation permissible. The European cases confirmed that the court deemed ‘spatial separation’ per se unproblematic it had a justifiable cause. Another important difference is the distinction between the power of the US and the federalism of the EU. First, in the US it was not the state itself that was on trial but either individuals or educational institutions, whilst in the European cases the countries themselves were on trial. The US could turn the story of desegregation into a victorious story of progress to enhance its international image as the protector of human rights (Dudziak, 2011), whereas the European countries charged with discrimination did not have the same opportunity and were shamed as human rights violators in the international arena. There was no consensus that the ECtHR represented the ultimate legislator in Europe as the Supreme Court did in the US. Furthermore, there was no consensus in the countries in question that discrimination was taking place; but even if it did take place, there was no agreement that it is morally wrong or that it is bad for the economy, foreign relations or the international standing of the countries in question.

The desegregation process involving Romani in a number of postsocialist Central and Eastern European countries began in the anticipation of EU accession at the turn of the millennium, and not during the postsocialist transition in the 1990s:

Yet the efforts for school segregation remained mainly in the hands of Romani activists (Rostas, 2012: 353) rather than the actual burden being extensively taken by the state authorities, which treated it more as a tick-box in the EU accession negotiation processes (Rostas, 2019).

In terms of strategic litigation, all but one of the cases at the ECtHR were decided on when the countries in question had already joined the EU; the Oršuš judgement was delivered a year and a half before all the negotiation chapters were closed for Croatia at the end of June 2011 (Croatia joined the EU two years later). Just after the decision had been made, the two principals commented on it in some of the Croatian newspapers: ‘I categorically deny there was ever anyone discriminated in this school, but if there was ever anyone discriminated, it was the children of non-Roma nationality’ (Jutarnji, 2010; my translation). ⁹ Another newspaper opinion piece from the region stated that the problem in Međimurje was ‘not segregation but Roma terror’ (Međimurje, 2010; my translation and emphasis), ¹⁰ showing that these phrases had been used before the protests in 2019.

The EU had warned Croatia previously that if it did not fully cooperate with the International Criminal Tribunal for Former Yugoslavia it might jeopardise its accession to the EU. There was a similar opportunity for the EU to force Croatia to develop an effective strategy to end the segregation of Romani children (Sardelić, 2019c). However, three months before the negotiations were concluded, the interim report on the progress of Croatia on judiciary and fundamental rights barely mentioned Roma (European Commission, 2011c). The final Croatia report commented: ‘There has been progress as regards the protection of minorities and cultural rights. The commitment to the rights of minorities, reaffirming their place in Croatian society, continues to be expressed at the highest level’ (European Commission, 2011a: 12). The report continued:

In November 2015 the FCNM Advisory Committee published an opinion that despite progress in the field of education, the segregation of Romani children in schools was increasing, even after Croatia had joined the EU: ‘the number of classes where only Roma are educated has increased, despite the 2010 Oršuš judgment of the European Court of Human Rights that bans segregation of Roma in schools’ (FCNM, 2016b: 25). Taking into account all the chapters in the EU negotiations, the European Commission did highlight the position of Romani children who were segregated, but presenting a viable plan for desegregation was not a condition for Croatia to join the EU.

Most of the scholarly literature and policy reports at least implicitly connect school segregation to the legacies of communism. However, this does not explain the persistence of segregation that has been reported in older EU Member States such as Greece and others. Whilst there was some reluctance among the ECtHR judges to decide on discrimination in the D.H. and Oršuš cases, in the Greek cases the decision was already unanimous at Chamber level. Yet in its NRIS Framework for Roma, the Greek government did not mention segregation even once, merely stating that drop-out rates and low attendance were among the most serious issues in the education of Romani minorities. In 2014 the European Commission's evaluation of Greece's NRIS highlighted that desegregation remained an issue in Greece. In 2014, ten years after the decision in the D.H. case, the European Commission started infringement procedures against the Czech Republic for not abiding by the Race Equality Directive. Similar procedures have been started against Hungary and Slovakia (Chang and Rucker-Chang, 2020: 119). Yet in 2018 the evaluation report on the NRIS noted that whilst there had been some improvement in the field of education, at the same time segregation of Romani children was increasing at the European level, not only in Central and Eastern Europe but also in the older EU Member States. Another European Commission report on the implementation of anti-discrimination law (Chopin et al., 2017) showed that patterns of segregation exist in seven post-2004 EU Member States as well as in the seven pre-2004 EU Member States.

Taken together, this all shows that the segregation of Romani children cannot be understood simply as a postcommunist legacy but is a broader reality of the wider EU today. The segregation of African American children (as well as other minorities in the US) was not simply in the past and did not end with the Cold War, but continues in the present day in a different form. Similarly, the segregation of Romani children has continued, even after the postsocialist countries allegedly complied with the Copenhagen criteria test for EU membership. Additionally, the older EU Member States were never subjected to the same test as the postsocialist countries seeking to join the EU.

Conclusion

This chapter has explored the cases of educational segregation of Romani children in Europe both in some newer EU Member States (Croatia and the Czech Republic) and in an older one (in Greece). It has particularly scrutinised the ECtHR cases on educational discrimination, focusing specifically on the states’ justification of why segregation does not equal discrimination. It has also contextualised the educational discrimination of Romani children within international and global narratives on the segregation of other marginalised children, that is, our future citizens.

The socio-legal analysis of the ECtHR cases and especially the states’ justifications have revealed both an embeddedness in and a detachment from global narratives when it comes to the segregation of Romani children. I started the chapter with the discrepancy between the local and the global: whilst the Pope was apologising on behalf of the Catholic Church and the majority population for the segregation of Roma in Europe, a local political elite in Croatia blamed Roma for their own lack of education. The fact that Croatia was charged with educational segregation and discrimination of children in Roma-only classes seemed absent from the public's memory after a decade. According to the Croatian public discourse in 2019, it is Roma who allegedly choose their own lack of education and therefore it is Roma who actively resist any involvement in the formal labour market.

In all four cases, the states followed a similar line of reasoning on why spatial segregation was not only permissible for the Romani children in question, but also beneficial as it addressed their ‘special needs’: either their intellectual capacity (as in the Czech Republic), their incomprehension of the majority language (as in Croatia) or the security risks and need for preparation before integration (as in Greece). In all the cases, the state authorities argued that this was not racial segregation since it did not apply explicitly to Romani children and, moreover, it was not only Romani children who were subjected to such treatment. According to the states’ arguments, the main difference from the earlier Supreme Court cases on segregation in the US was that the latter was entrenched in laws and applied to all African American children. In the case of Roma, the states’ representatives argued that it was qualified pedagogical experts who decided on the placement of children and that their decisions were not based on ethnicity but on objective educational criteria. However, the states’ representatives could not explain the invisible edges of citizenship, that is, why Roma were disproportionally represented in schools for children with special needs (as in the Czech Republic) and why only Romani children had to be in separate classes (Croatia) or in a separate annex (Greece). No other children were treated in such a manner. In the NRIS, the Czech Republic and Croatia did recognise Romani educational segregation as a challenge that needed to be addressed. The Czech Republic even linked African American desegregation with the plight of Romani children. Yet the Greek NRIS did not mention segregation as a challenge in the educational system, despite the three ECtHR cases proving that segregation did exist. This also shows that the global context in the case of Roma segregation cannot be simply understood through dichotomies of new versus old EU or socialist versus capitalist Europe. Segregation and discrimination had, on one hand, local specificities, while, on the other hand, they transcended previous Cold War divisions. Neither did the EU accession represent a clear-cut ‘carrot’ for the postsocialist countries to stop segregation.

The case of Romani segregation in educational systems provides a window through which to explore a broader question of how state authorities and international organisations create citizenship fringes in which they situate marginalised citizens: how Croatian politicians in 2019 ‘forgot’ that ten years previously the ECtHR had confirmed the existence of segregation and discrimination against Romani children, so that Roma could be again blamed for their own exclusion from the labour market. There is a broader neoliberal idea, too, promoted by the World Bank, that marginalised citizens should be included as they can enhance the economy, without questioning the basic premise: should their potential economic contribution be the measurement of whether rights are granted to marginalised minorities? Such questions point to the ways in which citizenship is constructed through its fringes. In this process, as Bhambra (2015: 110) explains, it is not enough simply to grant equal rights to marginalised minorities previously deprived of such rights:

This chapter shows that even introducing the practices of multicultural citizenship and the politics of recognition – such as celebrating International Romani Day – does not necessarily address the underlying discrimination that Roma face as citizens. It is at its fringes that citizenship establishes its boundaries, and it is there that it has to be reconstructed.