Ethnic identity as a challenge to antidiscrimination law: protection, positionality and liminality

ABSTRACT

Constitutions and antidiscrimination provisions tend to compartmentalise and operationalise ethnicity through religion, ethnic origin or country of origin, nationality, language and race. The growing focus on religion and race contributes to the entrenchment of ethnicity in religion and race, and this is not always fruitful. This article argues that antidiscrimination law should explicitly and robustly integrate ethnic identity as a protected (personal) characteristic in domestic law, independent of any other descriptors that may be recognised, and independent of existing group-based protections. Furthermore, this article argues that anti-discrimination infrastructure needs a provision of positionality and liminality. The concept of positionality facilitates the recognition of power dynamics present in the ascription of otherness, as well as the potential incongruences between identity ascription and actual self-identification. The concept of liminality enables the consideration of fluid and more flexible forms of self-identification across multiple public and private spaces.

KEYWORDS:

• Ethnicity
• religion
• race
• antidiscrimination law
• positionality
• liminality

Introduction

Constitutions and antidiscrimination provisions tend to compartmentalise and operationalise ethnicity through religion, ethnic origin or country of origin, nationality, language and race – categories rooted in historical forms of othering vis-à-vis the nation or the Leitkultur. It would not be fair to say ethnicity is altogether absent, rather, ethnicity is approached through many different lenses, and increasingly so through the lenses of religion and race. This growing focus on religion and race contributes to the entrenchment of ethnicity in religion and race, and this is not always fruitful. For example, Assyrian Christians may be racialised as Arab or religionised as Muslim, even as neither of these labels would be descriptive of them (Sparre 2021). Moreover, secular “Muslims” may be racialised and religionised solely on the basis of their name or phenotype (Khosravi 2012), regardless of their personal commitment to Islam. The two most persecuted groups in Europe – the Jews and Roma – fit only awkwardly with the labels of race and religion. The idea of “the Jew” is often caught between racial and religious prejudices, and non-observing Jews might clarify that they are “secular Jews”. Whereas Jewish and Roma identities have historically been racialised – to the extent of persecution and elimination – their identity and self-perception is not exclusively descriptively or normatively expressed through the concept of race, while both remain vulnerable to exclusion and discrimination (Sepúlveda, van Banning, and van Genugten 2010). Moreover, ethnicity cannot be only or primarily understood through the lens of migration, ethnic origin or ancestry. Europe is the home to many ethnic identities, including Jewish and traveller identities, and sub-national communities, such as the Catalan in Spain, Hungarians in Romania, Russians in the Baltic states, Indian and Caribbean communities in the United Kingdom, or Turkish communities in Germany and the Netherlands.

This article refers to ethnicity as a web of cultural, social and linguistic practices, common descent or country of origin, religious cleavages, or shared histories within groups (Calhoun 1993; Van der Tol and Becker 2024). Second, ethnicity is a deeply personal matter, which might express in language(s), dress, cuisine, diet, observance of particular holidays, burial practices, partner choice, religious symbols or regional attachment. On a personal level, ethnic identity might be squeezed between different and potentially competing group identities. While theories of intersectionality have offered a pathway for understanding more complex minoritized identities, intersectionality in the law still operates on the basis of binaries that must be invoked and proven separately, and which may not always be of primary relevance (Cho, Crenshaw, and McCall 2013; Nash 2008). This article argues that domestic antidiscrimination infrastructure should explicitly and robustly integrate ethnic identity as a protected (personal) characteristic, independent of any other descriptors that may be recognised in domestic law, and independent of group-based protections. This is important, because ethnic identity is under increasing pressure from political, social and cultural othering emanating from the Far Right, who benefit from the ambivalent and inconsistent referencing of ethnicity in the law (Lazaridis and Campani 2016; Van der Tol 2024; Vieten and Poynting 2016). Not only does it jeopardise the protection of ethnicity in the law, it also compounds structures of societal inequality. Moreover, certain policies might unjustly target ethnic minorities through the use of artificial intelligence or ethnic profiling by police and security forces (cf Phillips 2011; Terlouw and Van der Pas 2024).

Furthermore, this article argues that anti-discrimination infrastructure needs a provision of positionality and liminality. The concept of positionality facilitates the recognition of power dynamics present in the ascription of otherness, as well as the potential incongruences between identity ascription and self-identification. The concept of liminality enables judicial discretion in the consideration of fluid and more flexible forms of self-identification across multiple public and private spaces. Discriminatory action embodies a power dynamic, whereby the enactor may operate on the basis of perception of identity, such as name, skin colour or appearance. The biases involved in this perception may or may not overlap with the self-identification of the other. Suppose someone is discriminated on the basis of religion without actually being religious (comp. Khaitan and Calderwood Norton 2019). What use is the freedom of religion and belief in this scenario? Or when someone is asked where they are from, without being a migrant, what use is ethnic origin in this scenario? How does constitutional theory understand ambivalences between the ground of discrimination and one’s self-conception? In these two examples, there is an indication that race might be the defining factor. However, it is possible that the concept of race could be functionally overburdened through the over-inclusion of culture, religion, language, dress, even as all of these may contribute to racialisation in the lived experience of ethnic minorities (Siebers 2017, 377). In such cases, it might be better to consider ethnicity and race alongside each other for analytical purposes, rather than to collapse ethnicity into race.

Ethnicity in anti-discrimination repertoire: somewhat of a lacuna

Constitutional law and theory imagine personal characteristics in fairly neat categories, including race, colour, religion, nationality, gender, disability, and in some countries, sexual orientation (Eyer 2021). They have often grown from emancipatory processes, prioritising equality in the realm of gender, sexuality, race and religion (Lee 2017). Constitutional theory has long underwritten the protection of personal characteristics on the basis of the conviction that these characteristics are stable, natural and even inherently unchangeable. The invocation of nature has been significant to the development of LGBT+ rights, for example, particularly through the recognition that someone gay had been “born this way” (Fineman, Jackson, and Romero 2016; Rahman 2020, 23). Similarly, the argument of nature has been crucial in the legal understanding of race. Yet here, it was the evolving understanding of race and nature which paved the way for the legal personality, and subsequently, the equality of Black and mixed-race citizens before the law. The shift from essentialism to constructivism has allowed constitutional theory to grapple with social processes of identity-formation, and the complexities that stem from social dynamics (cf Roth, Van Stee, and Regla-Vargas 2023). Even so, these characteristics often appear as isolated aspects of identity, and they may or may not receive additional protection through the articulation of specific rights and freedoms in the context of constitutional law as well as antidiscrimination law. Such additional protections might stem from domestic law as well as from international treaties and charters. When characteristics like these feature in court cases, their isolated nature is apparent from a range of legal provisions that may be cited, and upon which references to intersectionality still depend. Although connections between these characteristics may in many cases be facilitated through theories of intersectionality, such connections cannot hide the fairly rigid understanding of identity in constitutional theory (Nash 2008).

Ethnic identity as such is difficult to categorise because of the kaleidoscopic amalgamation of identity, which may include aspects of race, gender, religion, language and sometimes ethnic origin, as well as customs, culture, or history (Calhoun 1993; Hall 2017). That amalgamation makes ethnic identities more complex to protect as a whole, both collectively and personally. Examples of this legal in-betweenness are less obsolete than they might seem: those who partake in religious holidays for cultural reasons; the choice for religious dress for non-religious reasons; discrimination on the basis of one’s name when applying for employment, housing, or education; ethnic profiling by law enforcement; disadvantages in accessing healthcare and financial services. Discrimination on the basis of ethnic characteristics is oriented at the person in these cases; however, personal protection against such discrimination is weak in isolation from group-based protections (Siebers 2017). In some countries, discrimination on the basis of ethnicity might not even be explicitly illegal. For example, the Dutch border police argued that ethnic profiling would not be against the law, but promised to cease its practice following social outrage (Brouwer 2022; Schalkwijk and Abdoelhafiezkhan 2022; Terlouw and van der Pas 2024). Similarly, the Dutch tax authority as well as several local authorities relied on ethnic profiling in anti-fraud policies, including in the Childcare Benefits Scandal. But the effective reduction of ethnicity to race or religion in law may create new inequalities within those sharing in the same ethnicity: it could inadvertently lead to a priority of the religiously observant over those who are not, or it may protect those who could successfully appeal to race more than those for whom racial othering is less obvious, or perhaps irrelevant.

Moreover, anti-discrimination categories are not identical in all jurisdictions: constitutional texts and anti-discrimination provisions make different choices in language, in meaning, and possibly in both (Magazzini 2024). The Dutch Equality Act does not refer to ethnicity at all, for example: it states that direct and indirect discrimination is prohibited on the grounds of “religion, fundamental convictions of life, political preferences, race, gender, nationality, hetero- or homosexual orientation or marital status” (Art 1 Dutch Equality Act). As a result, Jan-Peter Loof rightly observes that skin colour, language, ethnic ancestry, social background, wealth, birth are not yet included in domestic law. Although ethnicity is included in the category of race within the Dutch legal context, he notes that the usage of race for the purposes of law is not easy to grasp for ordinary citizens, and that explicit inclusion of ethnicity could be useful for the purposes of accessibility of the law (Loof 2020, 271–272). This lack of clarity also featured in the controversy over ethnic profiling. In Central Europe, ethnicity may be understood through the lens of nationality. The Preamble of the Hungarian Fundamental Law, for example, refers to “nationalities and ethnic minorities living with us” (Halmai 2021; Pap 2018). In the context of Hungary, the phrasing “with us” must be read in conjunction with the priority that the constitution implicitly ascribes to the ethnic Hungarian nation (Chronowski 2018). The UK Equality Act 2010 includes race, colour, language, religion, national or social origin, birth, as well as the “association with a national minority”. The latter recognises that discrimination is not constituted by identity per se, but also by the perception of that identity. The German constitution has a broader set of categories, including race, language, homeland, origin and ancestry. These arguably speak to the widest set of issues associated with ethnicity and might cover the identities of descendants of historical migrants. With the exception of the United Kingdom, these examples do not conceive of ethnic identity independent of migration, origin, or some form of “otherness”.

While ethnicity is inconsistently conceptualised in domestic law across Europe, ethnic minorities are afforded certain protections through European and international law. For example, the Council Directive 2000/43/EC of 29 June 2000 implementing the principle of equal treatment between persons irrespective of racial or ethnic origin speaks of both race and ethnicity, even as “origin” remains implied in its references to both race and ethnicity. Another example is the International Convention on the Elimination of All Forms of Racial Discrimination and Article 27 of the International Convention on Civil and Political Rights, which states:

in those States in which ethnic, religious, or linguistic minorities exist, persons belonging to such minorities shall not be denied the right, in community with the other members of their group, to enjoy their own culture, to profess and practice their own religion, or to use their own language (Bielefeldt, Ghanea, and Wiener 2016; Sepúlveda, van Banning, and van Genugten 2010).

Such provisions tend to speak to the needs of minority communities in contrast with the national political community, and tend to assume one’s membership or tangible association with a group (O'Nions 2016; Thornberry 1992). However, what tends to be overlooked here is that ethnic identity manifests in a range of contexts and relationships, and may have a different personal significance in each (Galal et al. 2016). Moreover, processes of de-institutionalisation and individualisation also relate to minority communities, and for this reason ethnic identity increasingly requires recognition at a personal level. A person may attach different significance to aspects of their identity depending on context, relationship, or space – or conversely, not attach significance to certain aspects of their identity at all.

The ethical value that is potentially expressed through the classification of ethnicity as a separate and independent personal protected characteristic is the assumption that ethnic identity could exist in isolation from a group or a community, at least for the purposes of the law. After all, many facets of ethnic identity may only be fostered communally, for example through the use of language, the sharing of customs or arguably adherence to religion. To some ethnic minorities, the idea that one can abstract one’s personal identity from this communal identity might be subversive, and even offensive. This criticism must be taken seriously: it is not for the state to determine what aspects of ethnicity are individual or indeed communal. The caveat is of course that the classification of ethnicity as an individual protected characteristic would complement group-based protections. Perhaps it is imperative to distinguish between the individual and the personal, as this article has tacitly done, to recognise that the personal can encapsulate both individual and communal dimensions, regardless of whether this concerns (ethnic) communities someone may be part of, or may have been part of in the past. This disctinction helps us to put into words that ethnic identification can be personal, that discrimination on the basis of ethnicity can cause personal harm, and that such experiences ought to carry legal significance.

One of the persistent questions in law is this relationship between the individual and the group, for example with regards to the relative vulnerability of women and children within the context of minority groups. Is it possible to argue for ethnic discrimination if someone faces a backlash over non-conformity, for example? For not wearing the right dress, for not marrying within the community, or for failing to teach the children the minority language? Such forms of intra-ethnic discrimination might benefit from additional individual protections. If we take the example of the hijab, it would be fair to say that some women wear the hijab for religious reasons; some women might wear it for cultural reasons, or to accommodate the wishes of the immediate family; and some choose not to wear the hijab, or would only do so in certain spaces (Bilewicz and Klebaniuk 2013; Pezzoli-Olgiati 2015). To consider the hijab as an essentially religious phenomenon is limiting, and so are arguments about religious coercion, which are at best speculative and irrelevant in several of the above examples. At the same time, men too may navigate a range of tensions, for example over religious adherence, partner choice, or decisions made with regards to family life. The examples used here reflect discursive practices that implicitly assume that the tension between the individual and the community pertains mostly to Islam. However, such tensions might very well arise within other groups, for example in the context of LGBT+ acceptance in religious groups, tensions over levels of adherence within ethnoreligious minorities, and yet, such tensions are not always considered on a par with Islam. Individual protection on the basis of ethnic identity could empower those who might be on the receiving end of intra-ethnic discrimination, and seek protection from it, insofar as such problems do indeed arise.

Ethnicity, religion and the nation state: otherness as “failed” assimilation

Whereas today the idea of the nation tends to be confidently understood as either self-evident or as an expression of privilege vis-à-vis certain minorities, historically, the idea of nationhood is no fait accompli. The possibility of imagining the political community as a nation presented an opportunity to move past the intra-Christian divisions of early modernity (Anderson 1983; Poggi 1978; Van der Tol 2020). These divisions have often been assumed to be primarily of a religious nature, and secondarily about political, social, or economic competition between various Christian groups. Inevitably, these differences were not necessarily religious, but at the very least interrelated with cultural, liturgical, political or linguistic practices. Nineteenth-century processes of nation building in West-Europe were therefore not limited to the mediation of religion in politics, but also entailed processes of administrative centralisation, linguistic and economic standardisation, and the development of a sense of a shared culture and history (Kaiser 2003; Knippenberg and De Pater 1988; Van der Tol 2020). In other words, the levelling of ethnic differences between local and regional communities that were subsumed in nascent stories of nationhood (Balibar 1990). These processes of integration, and in some cases, aggressive assimilation, are largely forgotten in contemporary discourses over religious, ethnic and ethnoreligious othering, which focus on “new” ethnoreligious others (Nussbaum 2012). It is clear, however, that on the socio-cultural axis, this is not a story of mere secularisation of belonging, especially as Jewish communities were not automatically included in new patterns of citizenship, irrespective of levels of religiosity or assimilation.

At the time, it was not necessarily controversial to expect that Jewish communities would assimilate, instead, it was part of a liberal narrative of cultural convergence to which the whole population was subject to some degree or another (Carter 2008; Van der Tol 2021). The axis of the secular-sacred only portrays part of the picture, and essentialising demands to assimilate in matters of religion, ethnicity or race, ignores the fundamental entwinement of a range of practices and customs that are embodied in various historical Jewish traditions. Moreover, one cannot underestimate the legacy of spatialised otherness through centuries of segregation, elimination and practices of toleration in shaping perceptions and lived experiences of difference. This is as it were the subconsciousness of the constitutional state: ethnic differences have been entrenched in spatialised communities and yet their distinctiveness was expected to decline over time. This progressivist narrative is in some ways akin to the secularisation thesis: the idea that religion would, over time, become less relevant, and eventually, possibly obsolete (Berger 1969; Neuhaus 1984; Van der Tol and Gorski 2022; Wilson 1985; cf “a people of the past” in Davis-Delano, Galliher, and Gone 2024). Ethnic differences are therefore tacitly understood as a temporary otherness within the context of the nation state, and as relying on personal efforts to integrate or even assimilate. This is of course not descriptive of reality: ethnic differences may or may not persist, and they do not always or necessarily constitute otherness. The extent to which differences constitute otherness is not a one-way street, and very much depends on attitudes within culturally and normatively dominant groups. The more otherness is emphasised politically or socially, the greater the concessions are that ethnic minorities must perform. Scholarship which objects to the secularisation thesis has proliferated, including in the field of constitutional theory. Much less attention has been paid to the development of ethnic identity: how ethnicity walked along a similar path, and how this impacts on our understanding of ethnic and ethnoreligious differences today.

European societies today are multi-ethnic, multi-religious and multi-cultural. They already were diverse since before the rise of nation states and their diversity has proliferated under the influence of decolonisation, post-war reliance on guestworkers, and other forms of migration. Therefore, negotiating differences is not something that is genuinely new, but an ongoing phenomenon that has been transforming over the course of intertwined socio-historical processes (Castles 2006; Chin 2007). Scholars like Will Kymlicka and Tariq Modood developed normative ideas about multiculturality since the 1990s, but multiculturality gained more traction in the realm of social and cultural policies than in the sphere of constitutional law and theory (Kymlicka 2001; Modood 2007; Modood, Triandafyllidou, and Zapata-Barrero 2006). Legally, the expectation is that migrants and their offspring integrate, almost as a one-dimensional process (cf Miera 2012). It assumes a division between those who are and those who are not considered to be sufficiently integrated: this division relies on ethnic and possibly racial distinctions, in which there remains an asymmetry between “the nation” and its others (Lauwers 2024). Such expectations may fluctuate, for example under the influence of major global developments. An example of this is the association of Muslim identity with terrorism: after 9/11 everyone who “looked” Muslim could be associated with terrorism and ordinary European Muslims were called to voice their personal rejection of terrorism (Hunter and McCallum Guiney 2023).

The differences that matter are of course the differences that present themselves or are perceived to be political. This depends as much on political context as it does on personal processes of othering, self-othering, integration and assimilation. This pertains clearly to “political Islam” (Linnemann and Bausback 2019), and is represented in particular by women and their children who resided in the bounds of Daesh, whose citizenship and consular services are at stake, as well as those who committed terrorism in Europe in the name of Islam. But it is also represented by the religious symbols that have come under scrutiny since the early 2000s, and which have led to a string of restrictions on perceived expressions of Islam in public space across European states. The legal discipline has largely focussed on the permissibility of such restrictions within the frameworks of contemporary law, and by necessity on definitions of religion and religious symbols. In isolation from the study of the formation and politicisation of ethnic and ethnoreligious differences, it is not difficult to see how these conflicts have become entrenched in debates over religious freedom.

Innovation 1: establish ethnicity as an independent category in domestic antidiscrimination infrastructure

Antidiscrimination infrastructure has largely grown from historical processes of emancipation: religious minorities, race, gender and sexuality would not have been recognised without these processes. However, antidiscrimination infrastructure often omits to classify ethnicity explicitly and independently, and this leaves ethnic minorities, such as Roma and Jews, relatively vulnerable to discrimination. It raises the question about the role of categorisations, who they include, and what happens with cases that do not quite fit (Türkmen 2024). The recognition of ethnic minorities in international law depends on communal status, which both assumes and reinforces the political institutionalisation of group-based identities. Ethnic groups may have specific needs, such as the preservation of a minority language, minority culture, or a minority religion, all of which depend on community-based reproduction. Whereas it is important that ethnic communities, where they operate as communities, receive such protections, domestic antidiscrimination infrastructures do not consistently incorporate the need to recognise ethnic identity as a protected (personal) characteristic. Sometimes this is because of the use of adjacent concepts, whereas in the Dutch case ethnicity is not (yet) included.

It is important to understand both the political and normative implications of the ambivalent use or omission of ethnicity. It expresses the symbolic ethical value that ethnic identity must exist in the context of a group or a community. It is a normative choice not to protect ethnicity on the level of the person: with reference to both communal and individual dimensions. It suggests the normative priority of race and religion over ethnicity, and attaches to this priority political consequence, namely the lack of protection of ethnicity. The operationalisation of ethnicity across Europe shows that ethnic inequality does not have the same moral status as racial or religious inequality. The phrase “ethnic origin” reflects a similar priority, in this case of migration status over ethnic status, even as many third- and fourth generation “migrants” – citizens – have no migration status themselves, while some minorities cannot be associated with migration at all. The priorities of race, religion and migration are not in themselves wrong, but their dominant representation in public and political discourse is problematic: in that they may compromise the avenues for redress when discrimination on the basis of ethnicity occurs; and problematic in that they invite liminal discrimination, which may not be fully anchored in religion or in race, and thus lack effective protection for ethnicity. Therefore, establishing ethnicity as an independent category in domestic law would have institutional value.

The recognition of ethnicity as an independent category signals that everyone has a duty not to discriminate on the basis of ethnicity, also when connections to religion and race cannot obviously be made. As such, the explicit inclusion of ethnicity can provide a basic clarity: ethnic discrimination is not permitted, not by the state, not by institutions, and not by citizens. Perhaps it is worth clarifying that the suggestion to recognise ethnicity explicitly and independently is to complement rather than replace other legal categorisations. Several governments have contemplated trading race for ethnicity, supposedly motivated by the idea that that there are no different races: everyone is human. It rings similar to the idea that ethnicity is irrelevant, because everyone is a citizen. Ethnicity might then function as a softer trap which makes prejudice sound more “respectable”. The political work that de-classification does, whether regarding race or ethnicity, is that it removes the legal basis for monitoring discrimination on the basis of race and ethnicity, which are both real in society. Moreover, it underlines how important it is for scholarship and legal practice to develop a robust understanding of ethnicity, both as an independent category as well as a contextual category that can be connected to discourses about race.

The classification of ethnicity is important to the future of antidiscrimination infrastructure. Ethnic self-identification increasingly transcends what may have appeared to be relatively stable ethnic or ethnoreligious communities. For Millenials and Generation Z, the classification of ethnicity as a group-identity is insufficiently descriptive as the texture of self-identification has become more complex and context-dependent (Jiménez 2010). Experiences of ethnic discrimination do persist in applications for jobs, internships and housing, for example, as well as in more casual micro-aggressions, and may still target those who are otherwise regarded as well-integrated. Under the influence of de-institutionalisation, the omission of ethnicity becomes more and more problematic, and especially as third- and fourth-generation “migrants” – citizens – navigate increasingly complex spaces compared to their parents or grandparents (Galal et al. 2016). The more individualistic society becomes, the more important it is to anchor protection from ethnic discrimination in provisions that pertain to the person. Furthermore, ethnic and ethnoreligious communities have been subject to secularisation, like many other religious communities in Europe. Khaitan and Calderwood Norton argue that protections stemming from the freedom of religion should be available to “committed non-adherents”: those who participate in religious practices for other than religious reasons, signalling the value of secular belonging to a religious minority (Khaitan and Calderwood Norton 2019). While this could be a viable strategy, reliance on the category of religion will become increasingly awkward for younger generations, especially with respect to the de-institutionalisation and fluidity of (ethno)religious belonging. Of course, part of the younger generation will remain religious in a more traditional sense, and they can continue to invoke the freedom of religion.

The absence of an explicit and independent category of ethnicity from a number of European constitutions and antidiscrimination provisions is striking given historical records of ethnic discrimination in Europe. The legal fragmentation of ethnicity indicates an inequality in the protection of minorities, and sometimes creates room for states to engage in discriminatory practices, for example through policies that may negatively impact on minorities (Magazzini 2024; Messing and Pap 2024). Such practices might be legal at face-value, as in they are based on the law and not in conflict with the law, but they may still not be legitimate. One must also take into consideration the cumulative effects of constant vulnerability to discrimination – both structurally and on a personal level. The doctrine of structural racism and ethnic inequality has put the spotlight on institutions and structures, yet its generality tends to lack the specificity that the law requires for effective redress. In order to build a portfolio of instances of individual ethnic discrimination, ethnicity as such must first be fully acknowledged in the law. This is what elevates anecdote and personal experience to usable legal arguments. Otherwise, its absence contributes to the difficulty with which structural ethnic inequality can be proven or may be deliberately instrumentalised by political movements that deny even the existence of these inequalities.

Innovation 2: the introduction of “positionality” and “liminality” as a judicial discretion

The distance between law and reality can be pronounced, as Terlouw and van der Pas (2024) argue in this issue. One of these distances is the rigidity with which particularly identities may be asserted in the context of the law. This is partially explored in Minow’s dilemma of difference, which captures the discrepancy between the legal relevance and the personal relevance of aspects of one’s identity (Anthias 2020; Minow 1990). Another distance is that of real and ascribed identity, and this is where the notion of “positionality” comes in. The notion of positionality can provide for the need to acknowledge, as part of the legal dynamics, that discrimination can occur on grounds that may or may not have a grounding in objective or subjective aspects of one’s identity. Positionality takes into account the uneven relationship between the ascription of identity and modes of self-identification, with specific reference to the power dynamic of those ascribing particular identities, whether personally or institutionally. For example, when discrimination occurs on the basis of associations with a minority ethnicity, without the victim actually belonging to this minority ethnicity – say, when someone “looks Jewish”, but is not in fact Jewish, or when Assyrians are misidentified as Arab Muslim. Further complications arise where discrimination pertains to hybrid and mixed-ethnic or ethnoreligious identities, where phenotype and self-identification may or may not overlap with commonplace binaries. In these instances, an appeal to antidiscrimination clauses would not be based on characteristics that are indeed personal to the victim, but on identities that are largely or solely ascribed.

If the concept of positionality would be taken into account, the assessment of the facts and circumstances of a case could explicitly acknowledge both the act of ascription and the relevant power-dynamic. This is particularly important to cases in which institutions are implicated, including state institutions. But even in horizontal cases between small businesses and individuals, or between individuals, power dynamics can still play a role. Arguably, the mere ascription of identity is in itself an assertion of power, which carries consequence through the act of discrimination. The concept of positionality could help grasp that assertion of power, assess the ascription of identity as inherent to that assertion of power, and acknowledge that the victim’s actual self-identification may not be relevant to the very act of discrimination.

The concept of positionality can further distinguish between geographies and spaces in which this power dynamic occurs. The distance between law and reality is that antidiscrimination provisions make few spatial distinctions, even though space is a significant indicator for the social relevance of identity. Not unlike religious symbols, the social relevance of dress, phenotype and names – as highly visible aspects of identity – depend on their spatial embeddings (Klocker and Tindale 2021). For example, the social relevance of ethnic difference is higher if representation is relatively low, and vice versa, the social relevance of ethnic (and other) differences may be lower if representation is high. The obvious example may be a plural urbanised space and a relatively homogenous rural region. Distribution of religious adherence may play a role as well. For example, when a relatively religious community understands its others primarily on the basis of religion, ethnic identity might primarily be understood along the lines of religious otherness – and vice versa. Weighing geographic and demographic aspects as part of an assessment of discrimination, and not only as part of the facts or circumstances of a case, can be important to the recognition of discrimination, and the context in which discrimination may have taken place (cf Smith, Edwards, and Caballero 2011).

Spaces can be subject to myriad ascriptions of meaning beyond the demarcation of sacred or public (secular) space, which can render aspects of identity relatively more important in some spaces compared to others (Van der Tol and Gorski 2022). Attention to ascription of meaning to space may elucidate minimal differences within ethnic, religious or racial minorities. For example, the particular way in which a headscarf is worn, the length of a skirt or dress, hairstyles, or language may signal sub-ethnic or sub-religious identification. Those differences may have a certain significance to those living in a neighbourhood or town that outsiders may or may not recognise. Forms of intra-religious discrimination may pertain to ethnic or cultural differences. Every day or micro-level forms of discrimination and inequality are not usually on the horizon of the general legal discourse on discrimination in Europe, which tend to think along the lines of religion, race and ethnic origin (Phillips 2011). Spatial differentiation may also inform personal relevance of identity. Whereas first migration migrants tended to live fairly segregated lives, third- and fourth-generation “migrants” – citizens – tend to navigate a plethora or spaces in which aspects of their identity may have distinct and even contradictory relevance (Galal et al. 2016).

This is where the concept of “liminality” comes in: defined as the peripheral or liminal relevance of two or more protected characteristics in antidiscrimination law, without a clear of exclusive mooring in either. The concept of liminality allows for the recognition of discrimination where multiple protected characteristics may be considered relevant, yet not comprising the core of the issue, and where the legal ground of discrimination may not otherwise allow for the effective recognition of discrimination. Such situations have previously been described in the context of the United States as pertaining to the “but for” doctrine (cf Eyer 2021; Lee 2017). For example, ascriptions of identity, as assertions of power, may or may not be traced to the rigid categorisations of religion, race, and where recognised, ethnicity. Indeed, certain forms of discrimination may be based on ascriptions of identity that have relatively weak moorings in two or more protected characteristics. Whereas intersectionality relies on the ability to articulate the relevance of two or more personal characteristics to an act of discrimination, liminality may recognise (intentionally) vague grounds of discrimination. This could potentially be operationalised as a judicial discretion in the application of the categories of antidiscrimination law.

The concept of liminality might mirror a personal experience of liminality, as potentially existing on the edge of particular group identities, or referring to multiple anchorings of belonging (cf transliminality in Artaud-de la Ferrière 2020). Hyphenated identities, like being Turkish-German, Moroccan-Dutch, Hungarian-Jewish, American-Born-Chinese are examples of this (Becker 2024). Whereas they may be regarded as “not quite” German, Dutch, Hungarian or American, they may equally be “not quite” Turkish, Moroccan, Jewish or Chinese (Hall 1996). These hyphenated identities hide experiences of contested belonging, as well as assertions of belonging which may continue to inform their experience with discrimination and inequality (Beaman 2017; Becker 2021; Guadeloupe 2022). Compared to when such ethnicities were considered “new”, ethnic identification has become more ambiguous through forms of hybrid belonging, as would be relevant for mixed-ethnicity families (Arweck and Nesbitt 2010; Smith, Edwards, and Caballero 2011). Jiménez describes this as “affiliative ethnic identity”, in which there “elastic” connections between ancestry and culture (Jiménez 2010). The concept of (trans)liminality can be an effective tool for recognising the relative significance of aspects of identity – aspects which heavily depend on context, and which inform how one experiences discrimination.

The concept of liminality will be most effective where multiple protected characteristics are at play, but it cannot be retrospectively determined which aspect of their identity was most fundamental to a particular instance of discrimination. The concept of liminality could play an important role in recognising structural or institutional discrimination, where a mix of ethnicity, race, religion, culture, marital status may have been important, but it may not be possible to determine the reason for the selection of each individual case. This may be the case with the use of Artificial Intelligence, where the “distance” between algorithms and individual cases may not be directly traceable (Brouwer 2022; Kleinberg et al. 2018). An example of this might be the use of algorithms in recruitment (Köchling and Wehner 2020). Whereas the precise ground for selection- and de-selection may be unclear to any individual, one can facilitate the analysis of discriminatory patterns through auditing algorithms or data collection on equality and diversity (Kim 2017). Another example is the Dutch Childcare Benefit Scandal in the 2010s, which unravelled discriminatory investigative practices concerning fraud, and which targeted Dutch citizens with dual nationalities as well as those with a minority ethnicity background. The concept of liminality might, in extra-ordinary circumstances, facilitate collective action to address structural discrimination on the basis of a range of relevant characteristics (cf Meuwese 2021; Prins 2021).

The concepts of positionality and liminality may not always be necessary, and in some cases could make the legal analysis unnecessarily complex. However, these concepts could be a welcome judicial discretion in the context of more complex cases on discrimination – with or without the explicit recognition of ethnicity as a protected (personal) characteristic. Positionality and liminality might function similarly to the concepts of reasonableness and fairness in the law. These concepts allow for a more tailored and personal approach to the adjudication of a discrimination claim, one that can potentially be closer to people’s personal experiences and needs. They might provide a way of holding different perspectives on allegations of discrimination: to acknowledge the incongruence between identity ascription and reality, to recognise the gap between legal expression of discrimination and the real-life experience thereof, or to offer a more precise account of power-dynamics that inhere in an alleged act of discrimination. Instead of reducing real-life experiences to legal abstractions, the concepts of positionality and liminality bring the law closer to real-life experiences. As such, they could play a significant role in closing the gap between people and the law, institutions or the state.

Conclusion

This article has argued that generally, ethnicity should be explicitly recognised in the law independent of religion, (ethnic) origin, migration or race. Constitutions and antidiscrimination infrastructures do not consistently include such references to ethnicity, while anti-discrimination frameworks are logistically fragmented between domestic, European and international mechanisms and standards. This article has argued that ethnicity ought to be included in the most basic domestic mechanisms, especially in constitutions and antidiscrimination infrastructure. This expresses the recognition that ethnicity is not a temporal or transient otherness, or indeed simply a foreign addition to the body politic that otherwise awaits the assimilation of its minorities. Moreover, it recognises ethnicity on a par with other categories that have emanated from processes of emancipation: such as gender, sexual orientation, religion and race. Historic conflations of ethnicity with religion have contributed to the political stigmatisation of Islam and migration, and the privileging of belief over cultural practices, seemingly legitimising the political regulation of cultural practices as desired by far-right actors. Newer conflations of ethnicity with race have put the concept of race under discursive pressure, which detracts from analytical precision with regard to ethnic inequality and contributes to the entrenchment of polarised discourses on ethnicity, religion, and race in Europe.

This article has also argued for the possibility of judicial discretion in taking into account positionality and liminality in assessing claims of discrimination. The concept of positionality would facilitate explicit legal recognition of discrepancies between self-identification and ascribed identities, with reference to the power-dynamics that shape such attributions. Discrepancies might arise from fluid relevances of identity in different spaces, taking into account spatialised and geographical dimensions to the ascription of certain identities. The concept of “liminality” would allow for the recognition of discrimination in cases of two or more relevant protected characteristics, where it is difficult to establish which one was more foundational to an act of discrimination. For example, where ethnic discrimination has racial as well as religious overtones, but discrimination cannot be exclusively grounded in either religion or race, or where acknowledged, in ethnicity as such. The two proposed innovations are not mutually exclusive: whereas categorical clarity warrants the inclusion of ethnicity in first instance, the use of the concepts of positionality and liminality would not be dependent on the prior recognition of ethnicity.

Acknowledgements

The author would like to thank the editors and reviewers for their insightful comments, as well as all the contributors to this special issue, who met during several workshops on Structural Ethnic Inequality and the Law at the Universities of Oxford and Heidelberg in 2022 and 2023. The author is especially grateful for the support of Philip Gorski and Jonathan Wolff, the Alfred Landecker Programme at the Blavatnik School of Government, and for the collegial collaboration with Elisabeth Becker.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This work was supported by Alfred Landecker Programme at the Blavatnik School of Government, University of Oxford.