A. Introduction
I. Little Research on Retaliation Against Bystanders Resisting Discrimination
Discrimination is a persistent problem throughout Europe. While great strides have been made to combat discrimination, with the adoption of anti-discrimination directives and the construction of equality bodies, victims still experience many difficulties in finding redress. The legislation’s individual complaints model insufficiently regards power dynamics and— notwithstanding the equality bodies’ invaluable work—leaves victims on their own all too often.Footnote 1 In turn, victims often do not even report discrimination, even though it falls within the scope of the legislation.Footnote 2 Reasons for this are a lack of proof, fear of negative consequences, and a lack of trust that reporting would change anything.Footnote 3
Policymakers increasingly recognize the potential of other actors to enhance the effectiveness of the legislation. One such actor is the bystander to instances of discrimination. This person—distinct from victim and perpetrator—becomes an “active bystander” when they intervene, by confronting perpetrators, supporting victims, or by reporting incidents.Footnote 4 Beyond informally changing social norms, they are particularly relevant for the enforcement of the legislation, playing a critical role in practice when they provide victims with otherwise unobtainable proof or when they file a complaint in the victim’s stead. In doing so, they can complement some of the pitfalls of the current legislative framework. However, “active bystanders” might also face legal or practical retaliation as a reaction to their support for the initial victim. They can be fired, lose the opportunity to get a promotion or to rent a property, face harassment, bullying, or legal liability. This underscores the need for intervening bystanders to be protected from these consequences.Footnote 5
The literature on this topic, especially in a European context, is scarce. Although “active bystanders” are already awarded some protection through EU provisions on protection against retaliation and their interpretation by the Court of Justice of the European Union, little scholarship exists on the theoretical justification for this protection. It remains unclear why this retaliation should be considered a moral wrong and how this relates to the initial discrimination. Additionally, considerable uncertainty remains about the legal protection an “active bystander” facing retaliation is afforded by current European legislation, and how this corresponds with theoretical insights.
II. Research Objectives and Methodology
The purpose of this Article is to explore the relationship between normative theories on the wrongness of retaliation against “active bystanders” and the legal protection currently afforded to them under European anti-discrimination law. Methodologically, the Article combines legal doctrinal analysis with normative legal theory. Section B starts by establishing the bystander’s potential to enhance the enforcement of the discrimination legislation, after which it applies recognition-based, desert-accommodating, and pluralist perspectives on the wrongness of discrimination to the specific phenomenon of retaliation against intervening bystanders. Section C conducts a doctrinal legal analysis of relevant provisions in the EU anti-discrimination directives, particularly those on victimization and retaliation, as well as the prohibition of discrimination itself, through the lens of associative discrimination. The European provisions are broadly framed and leave significant discretion to Member States. To illustrate how this discretion plays out at the national level, the Article draws on Belgian anti-retaliation protections as a concrete and instructive example. The choice for the Belgian framework is made in light of the Belgian origins of the landmark bystander case Hakelbracht by the Court of Justice of the European Union (CJEU). This has prompted amendments to the Belgian legislation that go quite far to protect bystanders from retaliation, both in terms of the protected conduct and the protection mechanisms. Particularly the provisions regarding burden of proof and the right to written evidence of the action opposing discrimination highlight tensions relevant beyond the Belgian legal context. This perspective has also been chosen given my background as a Belgian legal scholar. Section D then reflects critically on how well the current legal framework aligns with the theoretical justifications and where it may fall short. Section E ends with some concluding remarks.
Although the relevance of bystanders is not limited to employment, this Article focuses primarily on that context. Most legal provisions and developments concerning discrimination and retaliation relate to employment, reflecting the scope of the EU directives, which only protect certain grounds—such as religion or belief, disability, age, and sexual orientation—in the workplace. Moreover, the dynamics of bystander intervention and retaliation are particularly pronounced in employment due to power imbalances, information asymmetries, and the centrality of work to people’s lives and livelihoods.Footnote 6 Even where discrimination occurs outside work, the bystander is often situated in an employment context—for example, an employee in a housing agency witnessing discriminatory practices. Finally, the workplace includes specific anti-discrimination mechanisms, some mandated by instruments like International Labor Organization [ILO] Convention No. 111,Footnote 7 such as monitoring systems, bystander and leadership training, codes of appropriate professional conduct, hiring and dismissal protocols, and accountability structures like reporting mechanisms. Accordingly, many examples and legal provisions in this Article relate to employment, although they are also relevant beyond this context.Footnote 8
B. Framing the Active Bystander’s Role in Discrimination: Legislative Limits and Normative Justifications
Despite decades of anti-discrimination legislation, inequality remains a problem deeply entrenched in many social and institutional settings. While legal frameworks have evolved significantly, enforcement mechanisms struggle to provide all victims with effective recourse. In this context, the active bystander emerges as a potentially transformative actor. Section B starts by exploring the bystander’s potential in addressing the limitations of existing anti-discrimination law and the problems they themselves may encounter (Section B.I). It then continues by exploring the moral wrong of retaliation against bystanders from various theoretical accounts on the wrongness of discrimination (Section B.II). In doing so, it reframes bystander protection as an important component of an anti-discrimination regime.
I. Active Bystanders’ Potential to Address Limitations of Anti-Discrimination Legislation
In many laypersons’ understanding, which used to be reflected in earlier discrimination laws, people discriminate when they disadvantage a certain group of people because of an objectionable attitude towards, or belief about this group.Footnote 9 Discrimination is thus viewed as a behavioral expression of a person’s negative and prejudicial attitudes toward “out-groups,” that manifests when a prejudice reaches a level of intensity where the prejudiced individual chooses to deny those people or groups equality of treatment they themselves may wish.Footnote 10
This understanding does not account for implicit bias, or the effect discriminatory actions have on their victims regardless of intent. It also fails to account for the policies and practices that inadvertently impose disadvantages on members of particular social groups and reinforce stereotypes about them. For these reasons, this conception of discrimination is now generally regarded as inadequate. This has led to the abolishment of intent as a constitutive element of discrimination, and to the development of the concept of indirect or disparate impact discrimination, distinct from direct discrimination – which can be either intentional or unconscious as well.Footnote 11
The complaint model of contemporary discrimination legislation is focused on the individual: a victim must provide prima facie evidence of less favorable treatment by an individual perpetrator based on a prohibited ground. This individual complaint model has been criticized for not sufficiently taking into account the power imbalance between perpetrators and victims. For instance, relevant information is usually in the hands of an employer or landlord. This leads to difficulties for victims to even know about discrimination, never mind proving it. Moreover, the model leaves out the collective, societal dimension of discrimination, as victims still carry the bulk of the burden in a procedure, and the model does not account for the group dynamics at play.Footnote 12
Yet victims are not entirely alone. Throughout Europe, discrimination law is enforced largely through equality bodies or human rights institutes. Additionally, labor inspectorates, trade unions, and well-being services also contribute to preventing, detecting, and redressing discrimination. They do so in individual cases, for example, by supporting victims, or—in the case of the labor inspectorate—by exercising their competences to enforce the legislation, even in the courtroom. They also contribute structurally by influencing and advising policy and regulatory developments. The invaluable progress these actors have made notwithstanding, discrimination persists on a grand scale. Many cases are within the scope of the present legislation, but fail to be addressed or remedied.Footnote 13 The effective capability of human rights and equality institutions to engage in enforcement initiatives is also limited. This is at least partly due to the fact that many of these entities are under-resourced,Footnote 14 have a high volume of cases,Footnote 15 and face threats to their independence.Footnote 16
Victims still experience a plethora of barriers when seeking redress. Evidentiary difficulties, fear of reprisals, little prospect of a lasting solution, and the feeling that they cannot report every discrimination they experience —because it happens all the time—are just a few of them.Footnote 17 Therefore, only a fraction of all victims file a report. This leads to a general problem of underreporting, also called a “dark number.”Footnote 18
The difficulties and barriers victims experience when seeking redress have led to a growing recognition of the role other actors can or ought to play in these situations. Scholars like Sandra Fredman have argued for spokespersons other than victims themselves to be able to advance their rights.Footnote 19 The CJEU has demonstrated an openness for collective actors and interest groups to protect the principle of equality.Footnote 20 Arguably, the need for people to champion the rights of minority groups to which they do not belong themselves, applies to wider societal contexts than the courtroom alone.
The focal point of this Article, and an important collective dimension of discrimination, is the bystander, an actor already being responsibilized through policy initiatives. When a bystander refrains from intervening, they are called a “passive bystander.” They become an “active bystander” when choosing to intervene.Footnote 21 Potential bystanders are often trained in appropriate interventions in so-called “bystander intervention trainings.” These trainings serve to teach how to address a perpetrator, support or provide victims with the necessary information to prove discrimination, seek help, and report behavior.Footnote 22 These interventions can informally improve a situation for a victim, as well as challenge and change the social norms that facilitate discrimination.Footnote 23 They can also influence legal proceedings that otherwise wouldn’t have been initiated or feasible because of the lack of evidence.
Thus, this “active bystander” seems a promising ally for victims of discrimination, uniquely positioned to disrupt, prevent, or address discriminatory acts. Yet bystanders may also experience barriers to intervention and, occasionally, they themselves become a target of retaliation for supporting the victim.Footnote 24 This retaliation might materialize as a social or interpersonal cost, being perceived as a complainer or troublemaker.Footnote 25 Furthermore, retaliation can also have severe legal consequences, like in the Hakelbracht case, where an intervening bystander was dismissed for notifying a victim about a discrimination.Footnote 26 Therefore, to unlock the active bystander’s potential to improve the enforcement of the anti-discrimination legislation, they themselves might need protection as well. In that respect, people in power positions, even more so if they have a mandate to combat discrimination, are essential actors to prevent and resolve discrimination as active bystanders. They need to implement adequate structures to ensure that the culture in the workplace, school, housing agency, or nightclub deters rather than facilitates discriminatory conduct. In this Article, however, the focus lies on the hierarchic equals of victims, who willingly intervene in discriminations, and how they are protected.
II. The Bystander Vis-à-vis Contemporary Theories of Discrimination Law
Before examining how the European anti-discrimination directives protect bystanders, this section situates retaliation against them within contemporary theoretical perspectives on the wrongness of discrimination. It advances the normative claim that such retaliation constitutes a moral wrong, justifying legal protection, but the exact nature of this moral wrong is different depending on which theory is applied.
This section focuses on three strands of thought: recognition-based theories in Subsection 1; desert-accommodating prioritarian theories in Subsection 2;, and pluralist accounts in Subsection 3. These perspectives were selected with the Article’s objectives in mind, because they represent different moral foundations. These range from the (mis)recognition of equal moral status to fairness and equal moral desert in distribution. Because of these different foundations, each has distinct implications for the protection of active bystanders. Depending on the theory applied, retaliation may constitute a moral wrong auxiliary to the initial discrimination, or a moral wrong independently amounting to discrimination. It should be noted that the aim here is not to provide a comprehensive theoretical overview. Instead, the section seeks to highlight the complex role of bystanders, contrasting their different position across diverse accounts of discrimination’s wrongness.
1. Recognition-Based Theories
Recognition-based—or meaning-based—theories, such as Deborah Hellman’s, start from the observation that many actions and policies distinguish between people on the basis of some trait, and many policies or actions certainly have a disparate affect on a group of people defined by a particular trait. However, not all of those are morally wrong. In order to be morally wrong, the objective meaning of an action or policy must express denigration, or a failure to acknowledge an individual or group’s equal moral status.Footnote 27
This demeaning is objective in the sense that it does not depend on the actor’s intentions, but rather on the social meaning of the act. While this does not eliminate all disagreement about certain cases, it does channel the debate to particular questions, making some arguments stronger than others.Footnote 28
Furthermore, demeaning, or expressing a person’s unequal worth, has two aspects: an expressive dimension and a power dimension. A law, policy, practice, or action expresses that a person is not of equal moral worth, and the actor expressing this has the social power to enforce or communicate this degradation. This social power is both relational—for example,employer versus employee—and structural—such as dominant group versus marginalized group.Footnote 29 This demeaning, this expression of unequal moral worth, is morally problematic because people are of equal moral worth. People matter equally. Thus, morally troubling expressions are exactly those denying this fact.Footnote 30
Recognition-based theories have struggled to find an encompassing moral reasoning for the prohibition of indirect—disparate impact—discrimination. They contend that indirect discrimination is also morally wrong, insofar as they exhibit the expressive and the power dimension of demeaning.Footnote 31 If a law, act or policy that disparately affects certain groups is not demeaning, it could still be morally wrong because it “compounds past injustices,” exacerbating the harm caused by prior injustices, by entrenching it or carrying it into another domain.Footnote 32 But even if it does not compound previous injustices, it could still be justified to prohibit indirect discrimination on another basis, such as good social policy.Footnote 33
Applying this framework to the active bystander, it can be argued that they must be protected from retaliation for two reasons. First, they must be protected from retaliation because of the potential they have to counter the lack of recognizing the equal moral worth, on both the expressive and the power dimension of demeaning.Footnote 34 By confronting discriminatory acts or statements, bystanders affirm the equal moral status of the discriminated group or person, challenging both the expression of unequal moral status and shifting social norms and power dynamics.Footnote 35
Second, without legal protection against retaliation, the bystander’s own equal moral status may become conditional. Research suggests that people claiming discrimination on behalf of somebody else are met with derogations, depending on the endorsement of status justifying beliefs.Footnote 36 This means that in some cases, it could be argued that the retaliation against an intervening bystander creates a new, subordinate group, not based on a protected characteristic, but on the refusal of lowering other people’s social status. Only those who refrain from affirming marginalized groups’ equal moral standing have equal moral standing themselves—if in a given situation the social power dynamics allow for expressive acts to lower bystander’s status. This suggests that some subordinated groups oppression is so heavily policed, that a privileged person risks status loss when refusing to comply. Thus, some oppression is so intense that it functionally demands complicity.
In such a recognition-based view, it could therefore be argued that bystanders should not only be protected from retaliation as a mechanism secondary to the initial discrimination, but also that retaliation might be conceptualized as an autonomous form of discrimination. In the former case, retaliation impedes the bystander’s potential to recognize a victim’s equal moral status. In the latter, it directly violates the bystander’s own moral status, replicating the harm of discrimination even in the absence of a protected characteristics.
2. Prioritarian Theories
A second theoretical approach, found in desert-accommodating prioritarian theories, frames discrimination as a failure of distributive justice, because it fails to allocate social and economic goods in ways that realize moral value.Footnote 37 These theories begin from the premise that the right course of action is the one that maximizes moral value, which is greater when benefits are larger and when they go to those with a low prior expectation of advantage.Footnote 38 Thus, distributing benefits to those who are worst off—if they are deserving—yields the highest moral return.Footnote 39
These theories are both desert-accommodating and prioritarian. The former refers to the idea that benefits should be distributed on the basis of morally relevant criteria, such as merit or need, rather than arbitrary traits or luck. The latter holds that benefits to those worst off have greater moral weight, meaning that moral value is maximized when their interests are prioritized.
Central to these theories is the concept of socially salient groups: groups whose perceived membership shapes social interactions across a variety of social contexts.Footnote 40 Discrimination is distinguished from other forms of disadvantage because the harms inflicted upon one group member ripple out. These individual harms accumulate over time for the individual and spread to other group members. What constitutes a socially salient group can differ across different contexts.Footnote 41
One theory at the intersection of prioritarian and freedom-based accounts is Tarunabh Khaitan’s. According to Khaitan, the point of discrimination law is to enhance individual freedom by securing access to negative freedom, a sufficient range of valuable opportunities, and self-respect, needed for a good life.Footnote 42 At the systemic level, the theory is prioritarian, because it justifies efforts to remedy disadvantages that constrain freedom. At the individual level, it targets harms imposed on the basis of irrelevant group membership.Footnote 43 Anti-discrimination duties, in this view, are justified because they prevent freedom-denying impacts at both levels.
From a prioritarian perspective, the actions of active bystanders maximize moral value. Their intervention can lead to improved wellbeing for members of disadvantaged socially salient groups, not only in specific instances but through systemic spillover effects. Just as individual acts of discrimination contribute to broader group disadvantage, individual acts of solidarity can contribute to broader social recognition and remedy. Retaliation against active bystanders is therefore morally wrongful because it undermines beneficial dynamics and causes undeserved harm.
However, the moral wrong that is retaliation against active bystanders does not itself constitute discrimination. This is because individuals who intervene against discrimination do not form a socially salient group in the sense required by these theories. Although—as argued in the subsection on recognition—opposing discrimination comes at a cost,Footnote 44 being perceived as someone who opposes discrimination is unlikely to structure a wide range of social interactions across contexts in a way that generates systemic disadvantage.
Nevertheless, it remains a distinct moral wrong. It undermines efforts to correct existing distributive injustices, discourages solidarity with disadvantaged groups, and causes undeserved harm to individuals acting on morally relevant grounds. In this framework, retaliation might be understood either as a harm derivative of the initial discrimination or as a separate, non-discriminatory wrong that nonetheless warrants legal protection.
3. Pluralist Accounts of Discrimination
Pluralist theories challenge the assumption that discrimination is wrong for a single, foundational reason. Scholars like Sophia Moreau argue that discrimination is morally objectionable for multiple, equally significant reasons, such as its denial of equal moral status, its curtailment of freedom, and its role in perpetuating disadvantage.Footnote 45 Efforts to reduce all such harms to a single principle, pluralists argue, lead to unnecessarily reductive and monistic accounts. Rather than obscuring the complexity of discrimination by appealing to a single unifying principle, pluralists argue for embracing its inherent “messiness.”Footnote 46 In this way, the legal concept of discrimination is best understood not through a unitary moral theory, but as expressing a societal commitment to disavow and prevent entrenched patterns of behavior.Footnote 47
From a pluralist perspective, protecting active bystanders is part of maintaining the integrity of that institutional commitment. Just as the wrongness of discrimination is multifaceted, so too is the wrongness of retaliation against those who resist it. Retaliation undermines the moral recognition bystanders attempt to affirm, and may lower their own moral status. It compromises fairness in the distribution of opportunities by punishing actions on morally justified grounds. It decreases the freedom not only of structurally disadvantaged groups, but also of their allies who support their liberation. Lastly, it helps sustain structural hierarchies that depend on silence and complicity. A pluralist account can grasp all of these moral dimensions simultaneously, without establishing a hierarchy among them. Retaliation against active bystanders is thus a moral wrong both in itself and because it obstructs the broader fight against discrimination. Legal protection is therefore justified.
To conclude, the need to protect active bystanders from retaliation finds support across the three theoretical perspectives discussed here. Whether the wrongness of discrimination lies in the denial of equal moral status, unfair distribution or denial of freedom, people resisting discrimination must be protected. However, recognition-based or pluralistic theories more readily conceptualize retaliation as an autonomous act of discrimination, because it creates new subordinate classes, as it makes an individual’s privilege conditional upon their complicity in oppressing others. Contrastingly, prioritarian accounts are less likely to recognize retaliation as a discrimination in itself, because bystanders lack the social saliency to be discriminated as a group.
B. Bystander protection in European Directives
This section briefly considers the structure of the European discrimination legislation in Subsection I. Afterwards, it discusses possible bystander protection in current discrimination legislation by the secondary protection from retaliation in Subsection II. Subsection III then discusses whether retaliation against the active bystander could qualify as an autonomous form of discrimination.
I. Content Anti-Discrimination Directives
The legal basis is found in the Treaty on European Union (TEU),Footnote 48 the Treaty on the Functioning of the European Union (TFEU),Footnote 49 and the Charter of Fundamental Rights,Footnote 50 which enshrine the principles of equality and non-discrimination. Based on these, several key directives have been adopted: the Racial Equality Directive (2000/43/EC),Footnote 51 the Employment Equality Directive (2000/78/EC),Footnote 52 the Gender Equality Directives (2004/113/EC and 2006/54/EC),Footnote 53 and Directive 79/7/EEC on social security.Footnote 54 These anti-discrimination directives contain some concepts on combating discrimination, such as a definition of direct and indirect discrimination and a shift in the burden of proof.
“Direct discrimination” means that a person is treated less favorably than another person in a comparable situation on the basis of a protected characteristic and this distinction cannot be justified.Footnote 55 Such less favorable treatment can be either intentional or unconscious.Footnote 56 “Indirect discrimination” means that an apparently neutral provision, criterion or practice can put persons with a protected characteristic at a particular disadvantage, without justification.Footnote 57 To justify a less favorable treatment or particular disadvantage, one can show that it is objectively justified by a legitimate aim and the means of achieving that aim are appropriate and necessary. Additionally, there are specific justifications, for example regarding occupational requirements in employment, or reasonable accommodations for persons with a disability.Footnote 58
The shift of the burden of proof in civil proceedings was adopted because of the difficulties faced by victims of discrimination. When an alleged victim establishes “facts from which it may be presumed that there is a discrimination,” the alleged perpetrator must prove that there was no discrimination.Footnote 59 This can be done either by showing that there was no less favorable treatment or special disadvantage or by justifying it. Despite this mechanism, evidence remains a high hurdle for victims.Footnote 60 As previously mentioned, bystanders can be crucial by providing victims with evidence that would otherwise be unobtainable for them.
The directives are largely worded analogously, in both the recitalsFootnote 61 and the provisions.Footnote 62 The CJEU also contributes to their coherence by applying core concepts from one domain to other domains and using common interpretation techniques.Footnote 63 Still, it is untenable to completely apply and interpret them consistently. Sometimes the legal approach differs for each protected criterion.Footnote 64 Sometimes the different scope of the directives creates different levels of protection, prompting some scholars to speak of a hierarchy of grounds.Footnote 65
Summarizing, the European directives provide a solid legal basis for the fight against discrimination. They introduce and expand some core concepts, such as direct and indirect discrimination and the shifting of the burden of proof in civil cases. Yet they do not constitute a fully coherent discrimination legislation as there is some fragmentation, in both the scope of the directives, and the protected criteria.
II. Protection Against Retaliation or Victimization
While bystander intervention appears to be an effective tool against discrimination, it has not yet been systematically examined what the role of the bystander is or can be in discrimination cases, nor how the law should relate to it. Nevertheless, this actor is already somewhat present in current discrimination law—albeit not in name. Even though the bystander is not mentioned as such, the anti-discrimination directives contain provisions that protect active bystanders from retaliation as a reaction to their intervention, which is—as argued in Subection II of Section B—justified from various theoretical perspectives. More specifically, such protection applies to persons defending or giving evidence in favor of a victim.Footnote 66 In each case, the provision in question requires Member States to protect persons “from any adverse treatment or adverse consequence in response to a complaint or to proceedings aimed at enforcing compliance with the principle of equal treatment.”Footnote 67 In the directives that apply only to employment and occupation, the wording is stricter and only employees are protected against adverse treatment by their employer.Footnote 68 In doing so, Member States are given quite a lot of leeway to devise a protection regime, as the directives only require to take “the necessary measures.”Footnote 69 Therefore, a national implementation will be used as an example in Subsection 2, to showcase a possible implementation. This contribution has selected the Belgian legal framework for several reasons. First, it is very much up to date with the CJEU’s latest case law, as it has been amended post the Hakelbracht judgment, and has adopted a unique regime concerning burden of proof and obtaining written evidence of bystander interventions. Second, as shall be demonstrated in Section E, it is suited to highlight some tensions between the different legal regimes and theoretical frameworks applicable to bystanders. Finally, I have the most expertise on this framework given my background as a Belgian law scholar.
1. EU Protection from Retaliation and Victimization
The content of retaliation protection has evolved over the years. The preparatory documents of Directives 2000/43/EC and 2000/78/EC show that the European Commission’s main concern in 1999 was to avoid scaring off victims when they report a case of discrimination.Footnote 70 During the drafting of Directive 2000/78/EC the Parliamentary Committee on Employment and Social Affairs already suggested extending protection to witnesses as well, but the Council did not include that amendment in the final text.Footnote 71 Only in Directive 2002/73/EC —now no longer in force —was it clarified in recital 17 that retaliation protection applied to all workers defending or testifying in favor of a person protected under the Directive.Footnote 72 The retaliation protection in the later Directives 2004/113/EC and 2006/54/EC also applied to third-party actions in support of victims. The preparation of Directive 2004/113/EC showed that this included defending or supporting a victim.Footnote 73 Directive 2006/54/EC even expressly included in recital (32): “An employee defending, or testifying for, a person protected under this Directive should be entitled to the same protection.”Footnote 74
In addition, there are two CJEU judgments that have contributed to the interpretation of retaliation protection. First, there is the 1998 Coote judgment, which concerned the question of whether the “adverse treatment” set out in Article 24 of Directive 2006/54/EC—then Article 6 of Directive 79/207/EEC—covers adverse action after the end of the employment relationship. In the case at hand, an employer refused to provide references to a former employee because she had filed a claim for sex discrimination during the employment.Footnote 75 Advocate General Mischo and the CJEU disagreed on the nature of that refusal. In his opinion, the Advocate General considered the refusal a discrimination in itself.Footnote 76 In contrast, the Court held that it was not discrimination but a retaliation for which the directive only provides secondary protection. One thing the judgment and the conclusion did agree on was the time frame: adverse treatment after the termination of the employment contract also falls within the scope of the directive.Footnote 77
The other judgment, Hakelbracht, concerned the question of whether retaliation protection also applied to informal actions to support or defend a victim of discrimination. Indeed, the Flemish and federal discrimination legislation had transposed the directives in such a way that only persons acting as witnesses during legal proceedings or persons who had filed a formal complaint could benefit from the retaliation protection.Footnote 78 In this case, the head of human resources at a clothing shop refused to employ a pregnant woman, Ms. Hakelbracht, because of her pregnancy. Ms. Vandenbon, an employee at the shop, told the head of human resources that this refusal violated the prohibition on discrimination, but even so, the company did not hire Hakelbracht. Vandenbon then informed Hakelbracht of the reason she had not been hired. Subsequently, the company dismissed Vandenbon.Footnote 79
Because she had not filed a formal complaint or acted as a witness in a legal procedure, Vandenbon could not rely on the retaliation protection under Belgian law, but the CJEU ruled that the law was flawed, as the protection provided for in the directive also applies to those who provide informal support to victims.Footnote 80 This ruling led the European Commission to start an infringement procedure INFR(2015)2012 against the Belgian state, for the incorrect transposition of the protection against retaliation. Subsequently, the provisions in both federal and Flemish legislation were amended.
It should be noted that in the Hakelbracht case, the CJEU only interpreted Directive 2006/54/EC, making it thus —strictly speaking —only applicable to the protected grounds sex and gender, in an employment context. However, the Court’s reasoning can mutatis mutandis be applied to other grounds and contexts as well. This is evident from the Court’s emphasis on the fact that benefiting the protection against retaliation should not be made on the basis of formal criteria, but rather on the basis of the role that a person—in casu: an employee—may have played for the benefit of the protected person.Footnote 81 This is—dixit the Court—an expression of the principle of effective judicial protection.Footnote 82 The Commission’s infringement procedure 2015(2012) also emphasized the protection of employees in cases involving other grounds than only sex and gender, but did not mention the protection of bystanders in other contexts.
2. Protection Against Retaliation in the Belgian Legislation
The European stipulations protecting bystanders and victims from retaliation leave a lot of discretion to Member States to implement a protection regime. Therefore, it is warranted to examine such a protection regime at the national level as well. Hereinafter, the Belgian legislation will be set out as an example. In Belgian federal legislation, protection against retaliation can be found in Articles 16 and 17 of the Anti-Discrimination Law, Articles 14 and 15 of the Anti-Racism Law, and Articles 21 and 22 of the Gender Law, all of which were amended in 2023 following CJEU case-law.Footnote 83 All these provisions are mutatis mutandis worded the same and are discussed together in this contribution.
The provisions protect individuals from adverse consequences for reasons related to a report, complaint or legal claim, or the content thereof. Since the amendments, all formal requirements regarding such reporting have been removed, to make the law compliant with the Hakelbracht judgment.Footnote 84 Notably, the Belgian legislature extended the Hakelbracht reasoning beyond the area of employment. Although the applicability of this reasoning in all protected areas might seem evident in light of the principle of effective judicial protection, the Hakelbracht judgment and subsequent infringement procedure INFR(2015)2012 did not explicitly require so. This is due to the fragmentation—both in terms of protected grounds and areas—of the EU Directives protection, as discussed in Subection I of this Section C.
European law clearly states who is protected against what kind of adverse treatment. But Member States are given a lot of freedom to determine the form this legal protection should assume. Indeed, the directives only require that “the necessary measures”Footnote 85 be taken. The Belgian legislature has diligently implemented this by constructing a comprehensive protection regime in the various legislative instruments. The most salient provisions —a shift in the burden of proof and a right to written evidence —are discussed in more detail hereinafter.
First, a presumption is inserted in Article 16, §3 and 17, §4 Anti-Discrimination Law, Article 14, §3 and 15, §4 Anti-Racism Law and Article 21, §3 and 22, §4 Gender Law that any adverse treatment, within twelve months after a complaint or report and within three months after a court decision has become final, is related to that report, complaint or legal proceeding. It is then up to the employer or person doing the adverse treatment to prove that it is unrelated. In doing so, the legislature tries to alleviate the heavy burden of proof that those involved would otherwise face.
This protection also applies to the active bystander, because Articles 16, §5 and 17, §7 Anti-Discrimination Law, Articles 14, §5 and 15, §7 Anti-Racism Law and Articles 21, §5 and 22, §7 Gender Law extend the protection against retaliation mutatis mutandis to persons who act as witnesses, file a report or complaint for the benefit of an alleged victim, or provide counsel, aid or assistance to a victim, or raise the violation of the law against the perpetrator. The draft states that indirect witnesses who became aware of the facts through another person who observed them are also protected. With this broad wording, the legislature seeks to protect all forms of third parties attempting to raise discrimination.Footnote 86 Seeing that retaliation protection applies to these persons, the shift in the burden of proof on retaliation will also apply to anyone who becomes the subject of an adverse treatment within twelve months of providing counsel, aid, or assistance to an alleged victim.Footnote 87
A novelty is that anyone filing a report, complaint, or legal action has the right to written evidence from the organization, agency, or institution with which the action is taken.Footnote 88 The relevant provision applies to both victims and active bystanders, by referencing the respective provisions applicable to the assisting bystander mentioned in the previous paragraph. This means that the active bystander is entitled to written and dated proof of formal and informal intervention. In the case of formal intervention, this is fairly obvious: when the bystander acts as a witness or files a report or complaint, the body where that action was taken gives them a written piece of evidence. For example, if a bystander witnesses discrimination taking place in a supermarket and decides to file a complaint with the supermarket’s complaint service, they have a right to receive written evidence of that complaint.
In the case of informal intervention, I contend that this is less obvious. Indeed, if the alleged perpetrator and the body that is legally required to provide the evidence are the same, one is entirely dependent on the good faith of the person whose toes have just been stepped on. For example, an employer may harass an employee on the basis of a protected characteristic. The bystander who supports the victim or addresses the employer must then ask that same employer for written proof of this support or outcall. Likewise, if a bystander witnesses a discriminatory refusal to sell certain goods and services in the supermarket, and calls out the perpetrator, that same perpetrator would then be legally required to provide written evidence of the bystander addressing them. If the perpetrator then refuses to sell the active bystander certain goods, that same written evidence would then serve to prove that the adverse treatment is related to the bystander calling out the perpetrator. As these examples illustrate, obtaining written evidence of an informal bystander intervention might be practically unfeasible if not impossible.
In the absence of written and dated proof, one can still prove the intervention by all means of law. Proving an informal intervention does remain more difficult than proving a formal intervention. Thus, there is a legal incentive to make the intervention as formal and large as possible, while scientific literature suggests on the contrary that bystanders, out of shame and fear of losing face, want to keep their intervention as small as possible.Footnote 89
In the Flemish Equal Opportunities and Treatment Policy Decree, protection against retaliation is stipulated in Articles 37 and 38. A number of amendments to the Equal Opportunities and Treatment Policy Decree were recently published in the Belgian Official Gazette to align it with European legislation and infringement procedure INFR(2015)2012. As in federal legislation, the formal requirements in paragraph 2 of Article 37 have been removed.Footnote 90 The decree also introduces the presumption that an adverse treatment is related to a complaint or notification if it is taken within twelve months. In the case of a legal action, this is within three months of the court decision becoming final. In such a case, the burden of proof weighs on the person who took the adverse action.Footnote 91 Novel is the addition of a third paragraph to Article 37 §3, which, analogously to federal law, provides for the possibility of requesting a written and dated proof.Footnote 92
Regarding the protection of the active bystander, the text of the Equal Opportunities and Treatment Policy Decree is worded differently. Article 38 stipulates that whoever provides support to a victim by defending, representing, or testifying is entitled to the same reprisal protection mentioned in Article 37.Footnote 93 Though unclear in the decree itself, the parliamentary preparations of the amending decree are unambiguous that this refers to all persons who have formally or informally provided support to a victim of discrimination.Footnote 94
Although the wording of the retaliation protection in the Flemish and federal discrimination legislation differs somewhat, both the parliamentary preparations and their European-law origins show that these provisions offer an equal level of protection, both to victims and to active bystanders. These are thus protected against all adverse consequences as a reaction to both formal and informal intervention in support of the victim, in all areas protected by the Belgian legislation. Furthermore, there is a rebuttable legal presumption that adverse treatment within the three- or twelve-month period discussed above are related to that intervention, which leaves the bystander primarily with the burden of proving the intervention. To ease that burden of proof, a written, dated proof can be requested, but as explained above, this is less evident in case of a—more frequently occurring— informal bystander intervention.Footnote 95 Therefore, it will be interesting to see the extent to which these provisions have practical relevance. At the very least, the law and decree do signal clearly that this collective dimension of discrimination deserves attention and protection.
III. Discriminating the Active Bystander: Associative and Attributive Discrimination
Two related concepts that give recognition to the complex, collective reality of discrimination are “associative discrimination” and “attributive discrimination.” These concepts could provide an avenue for an autonomous discrimination claim for the retaliation suffered by the active bystander. The first refers to the situation where a person who does not possess a protected characteristic becomes a victim of discrimination because of the association or connection with a person who does possess it.Footnote 96 In the second concept, “attributive discrimination,” “misperception discrimination” or “discrimination based on a perceived characteristic,” there is no link with a person with a protected characteristic, but a person is disadvantaged because of a characteristic that is wrongly attributed to them.Footnote 97 Hereinafter, Subsection 1.1 starts by explaining the legal concept of associative discrimination, as developed by the CJEU and implemented in Belgian law. Afterwards, the legal scope of attributive discrimination will also be explained in Subsection 1.2. Because the actions covered by the protection against retaliation can also be considered associative or attributive discrimination,Footnote 98 the question to what extent the bystander could directly invoke the prohibition of discrimination by means of associative or attributive discrimination when becoming the object of less favorable treatment is addressed further in Subsection 2. To this end, the types of association according to Sottiaux and Vrielink are explained first, followed by some reflections on the link.
1. Development and Codification of the Protection Against Associative and Attributive Discrimination
1.1. Associative Discrimination
Protection against associative discrimination was developed by the CJEU in the Coleman and CHEZ cases, although the Court did not use the term “associative discrimination.” This development was possible due to a creative reading of the guidelines, that stipulate that a discrimination is a less favorable treatment on the basis of “a” protected characteristic, and not on the basis of “her/his/their” protected characteristic.Footnote 99 In the Coleman case, a woman had been treated less favorably than other employees because of her son’s disability. The treatment consisted of the employer refusing flexible working hours and working conditions, subjecting Coleman to misguided comments, and telling her that she would be thrown out on the streets if she arrived late, while her colleagues who were sometimes late were subjected to no such threats.Footnote 100
In its judgment, the CJEU followed the reasoning of Advocate General Maduro, that the employer’s reasoning being based on the protected ground was decisive for the qualification as discrimination, regardless of who possesses this protected characteristic.Footnote 101 The CJEU further clarified that both direct discrimination and harassment can occur by association, as in casu the woman also suffered harassment due to her son’s disability.Footnote 102
In the CHEZ Razpredelenie Bulgaria AD case an electricity supplier placed electricity meters in a neighborhood inhabited mainly by persons of Roma origin at a height of six to seven meters, while in other neighborhoods they were placed at a normal height of 1.70 meters, presumably to prevent fraud. A resident of the neighborhood who herself was not of Roma origin claimed that she was a victim of racial and ethnic discrimination.Footnote 103
The CJEU followed the opinion of Advocate General Kokott who stated that associative discrimination does not necessarily require a personal connection. Indeed, the association in this case occurred because the measure, by its general and collective nature, affects persons other than those with the protected criterion.Footnote 104 In doing so, the Court clarified that indirect discrimination by association can also occur. Anyone who suffers from an indirectly discriminatory measure can—directly—invoke the prohibition of discrimination. Although it is therefore not required that the person with a protected criterion initiates a claim, the person without a protected criterion who initiates a claim does have to prove that a protected person is put at a particular disadvantage.Footnote 105
By the way, the protection against associative discrimination was always implicitly present in Flemish and federal legislation, because analogously to the directives, there was talk of “a” protected criterion and not “her/his/their.” The legislature wanted to codify the jurisprudential developments by stipulating that there is both direct and indirect discrimination if someone is treated differently because of a protected characteristic of their own or an associated one.Footnote 106 The parliamentary preparation of the Law of 28 June 2023,Footnote 107 which added the concept to federal legislation, also explicitly refers to the Coleman and CHEZ judgments.Footnote 108 Thus, this clarification has no additional legal effect, given that the Court of Justice’s interpretation was already binding.
1.2. Attributive Discrimination
Attributive discrimination occurs when a person is treated adversely on the basis of a perceived—attributed—characteristic. An example is the situation where a southern Italian person would not be hired because an employer mistakenly assumed that this is a person of Arab origin—about whom they have negative prejudices. Unlike associative discrimination, as of yet no CJEU cases have been ruled in which the Court explicitly applies attributive discrimination. Nevertheless, the wording “a protected characteristic” in the directives allows for an interpretation that includes attributive discrimination, similar to the previously discussed associative discrimination. Moreover, both the Belgian and the European legislature consider this included in the definition of discrimination. The Commission already stated in the preparatory works of the Racial Equality Directive that the principle of equal treatment should apply regardless of whether the discrimination based on racial or ethnic origin is real or presumed.Footnote 109 When adopting the federal discrimination legislation, the Belgian legislature followed the Commission and clarified further that a prohibited motive, falsely attributed to a victim, is also captured by the discrimination prohibition.Footnote 110
In 2023, the Belgian legislature opted to enshrine this reading in federal and Flemish law. Indeed, the definitions now contain the wording “actual or alleged” protected characteristics.Footnote 111 Thus, Belgian federal and Flemish discrimination legislation explicitly, and European legislation implicitly, protect against attributive discrimination.
1.3. Conceptual Distinction
It is important to conceptually distinguish associative and attributive discrimination from direct and indirect discrimination, harassment and instruction to discriminate, as they relate to other aspects of discrimination. Associative and attributive discrimination relate to the “who” aspect: who is being subjected to the less favorable treatment and what is the link between the victim and the protected characteristic? Associative discrimination, where one is discriminated against on the basis of someone else’s characteristic, is opposed to discrimination on the basis of one’s own protected characteristic. Attributive discrimination, where the characteristic is alleged, is in turn opposed to discrimination based on an actual characteristic. In contrast, whether it concerns a direct or indirect discrimination, harassment, or instruction to discriminate, relates to the “how” aspect: how does the unequal treatment manifest: as a direct discrimination, indirect discrimination with a disproportionate impact, unwanted conduct that affects a person’s dignity, or an instruction to discriminate?
The who-question and how-question should always be addressed separately to avoid mixing a number of related concepts. Thus, a person may be discriminated against on the basis of a characteristic of their own, granted by attribution or by association. Additionally, that same discrimination can be either direct or indirect, or a harassment or instruction to discriminate. Indeed, in Coleman, there was direct discrimination and harassment, because the person in question was directly treated unfavorably and harassed. This happened by association, because it was on the basis of her child’s disability.Footnote 112 In CHEZ, there was also discrimination by association because the person in question was discriminated against on the basis of someone else’s characteristic. Here, it was an indirect discrimination because the less favorable treatment was not directly directed against persons from the Roma community, but against the whole neighborhood, putting those persons at a particular disadvantage.Footnote 113 The messiness of these concepts when occurring simultaneously highlights the importance of separately addressing the who- and how-question.
2. An Autonomous Discrimination Claim by the Active Bystander?
As argued in Section B.II, the harm inflicted upon a bystander resisting discrimination can be considered a discrimination in itself from a recognition-based or pluralist perspective, as failure to comply with other people’s oppression can rob a person of their own privilege, creating a new subordinate class. However, a prioritarian would disagree because this retaliation does not create a new socially salient group that structures social interactions in a wide array of contexts.
Moving beyond this theoretical field of tension, the next sections dive deeper into the question whether the legal texts and the CJEU’s interpretation thereof, actually provide a possibility for such an interpretation.
2.1. Relevant Discrimination Types
For an active bystander to be able to seek an autonomous discrimination claim against the adverse treatment as a reaction to an intervention, the treatment must be taken “on the basis of” or “on grounds of” a protected characteristic. This link between the protected characteristic and the adverse treatment of the active bystander will manifest by attribution in some cases and association in others. Below, these concepts are explored further and applied to the situation of the active bystander.
All in all, the link may arise through attribution if there is an invisible protected characteristic. For example, if an offender believes that a bystander who stands up for persons with a queer identity is also queer and subsequently takes adverse action against that bystander, those actions may qualify as attributional discrimination. That bystander can then invoke the discrimination prohibition directly. The legislature was already aware of such dynamics when amending discrimination laws. Indeed, the parliamentary preparation mentions the situation where a person committed to an LGBTQIA+ organization is presumed to have a particular sexual orientation.Footnote 114 From there it is only a small step to apply the protection against attributive discrimination to those who stand up for LGBTQIA+ persons in the context of bystander intervention, and are treated less favorably for doing so because of their alleged belonging to that group.
From the perspective of the bystander seeking redress, this path would entail a disadvantage, though. Namely, in such a case, they would have to prove that the perpetrator, when taking the adverse action, was actually under the assumption that the bystander possessed the protected characteristic. It will be extremely difficult for the bystander to know that assumption, let alone prove it. Furthermore, certainly not all retaliation against active bystanders will be taken on the basis of such an assumption. Therefore, the conditions of attributive discrimination will only be met in specific cases.
Nonetheless, the active bystander could also argue that the link between their less favorable treatment and a protected characteristic did not arise by attribution, but by association. To find out under what conditions this is possible, it is necessary to delve deeper into the different types of association as discerned by legal doctrine. Scholars Jogchum Vrielink and Stefan Sottiaux distinguish three types: separate associative discrimination, passively transferred associative discrimination and actively transferred associative discrimination.Footnote 115
The first type, “separate associative discrimination,” also known as “associative discrimination sensu stricto,” occurs when the person with a protected characteristic is not directly affected by the less favorable treatment targeted at a person without a protected characteristic.Footnote 116 The discrimination in the Coleman case is an example of this: a mother is treated unfavorably because her child has a disability. As Advocate General Maduro points out, the son here also suffers from the discrimination: he sees his dignity and autonomy affected by his mother’s disadvantage.Footnote 117 Nevertheless, this is a “separate associative discrimination,” because the person with the protected criterion would not be able to claim discrimination, while the person without the protected criterion could. In this view, the notion of “separate associative discrimination” may, on the one hand, compensate the absence of possibilities of legal redress for the person with the protected criterion, by allowing persons without the protected criterion to claim their disadvantage.Footnote 118 On the other hand, this form of associative discrimination recognizes the own disadvantage suffered by the person without a protected criterion and thus gives that person an independent right of action. This then seems to presume a close connection between the person without and with protected characteristic.Footnote 119
When a person becomes victimized together with a person who does possess a protected criterion, this is called a “transferred associative discrimination” or “co-victimization.” In the case of “passively transferred discrimination,” one is co-victimized by circumstances beyond one’s control. This is the case in CHEZ, where the discriminatory measure targets the whole neighborhood.Footnote 120 In such a case, the association consists of persons without a protected characteristic becoming “collateral damage.” Therefore, no further close connection is required.Footnote 121
The third form of association, most relevant to active bystanders, is “actively transferred discrimination.” Here, the own action of a person without a protected characteristic triggers the association. Sottiaux and Vrielink give the example of a bouncer who is fired for refusing to cooperate with his employer’s instruction to keep out persons of foreign origin.Footnote 122 In such a case, there is no complete “co-victimization” because the discriminatory instruction and the dismissal are separate acts. This situation is reminiscent of the Hakelbracht judgment, discussed above at the protection against retaliation, which Petra Foubert also named associative discrimination.Footnote 123 But the person in that case did not autonomously rely on the prohibition of discrimination. Instead, they invoked the secondary protection against retaliation.Footnote 124 Moreover, there are no other CJEU judgments in which the person without a protected criterion directly invoked the prohibition of discrimination pursuant to such an “actively transferred discrimination.”
2.2. Maximalist or Minimalist Causality?
The question therefore arises whether a court could also qualify this “actively transferred discrimination” as associative discrimination. Such qualification would allow an active bystander to directly claim the protection of the discrimination prohibition against the adverse treatment as a reaction to a bystander intervention. To determine this, what matters most is whether the protected criterion is the ground for the less favorable treatment, because that consideration was decisive in the Coleman and CHEZ judgments.Footnote 125 But the exact interpretation of that causality is unclear. How strong should the link between the protected characteristic and less favorable treatment be? Neither the legislation nor the case law provide an unequivocal answer to the question of what conditions the link must satisfy.Footnote 126
On the one hand, according to a “maximalist” interpretation, a less favorable treatment “on grounds of” a protected characteristic exists from the moment that it would not have occurred, if the protected characteristic had not been present. In such a case, the protected characteristic is a conditio sine qua non for the disadvantage. The CJEU has already used this test in other discrimination cases.Footnote 127 For the active bystander, this would mean that the secondary adverse treatment as a reaction to resisting the original less favorable treatment almost always arises “on grounds of” the protected characteristic. This is because if the protected characteristic were not present, neither the original treatment, nor the resistance against it, nor the secondary disadvantage would have occurred.
This interpretation maximally protects the active bystander, given that they can first make an associative discrimination claim and, subsidiary, a protection against retaliation claim. Given such an interpretation, any subsequent link in the chain of association could ad infinitum give rise to an autonomous discrimination claim against retaliation. To illustrate, one might imagine a situation where a person with a protected characteristic wants to access a nightclub but is refused entry by the bouncer. A bystander who was just entering the club sees this happening and addresses the bouncer, calling out the discrimination. The bouncer then says, “now you’re not getting in either,” which can be seen as a retaliation against an intervening bystander. That retaliation would also qualify as associative discrimination. If another person then speaks up for the bystander, and is in turn also refused entry, that would again constitute discrimination, because the entire situation would not have happened if the initial discrimination did not occur. Of course, this chain of associative discrimination can simply be broken if the bouncer does not discriminate or take retaliatory measures.
On the other hand, according to a “minimalist” interpretation, the association would have to be stronger to speak of an adverse treatment “on grounds of” the protected characteristic. Requiring that the perpetrator intentionally treats somebody adversely because of the association with the protected characteristic would be manifestly wrong because proving intent is not required in civil discrimination law and an act in good faith is not accepted as a defense to a discrimination claim.Footnote 128 Instead, a minimalist interpretation could mean that the protected characteristic is—unwittingly or not—a reason for the less favorable treatment.Footnote 129
Such an interpretation would mean for the active bystander that they can only autonomously invoke the discrimination prohibition if the protected characteristic is part of the reasoning on which the retaliation is based. For example, if the bouncer in the previous example does not tolerate dissent and retaliates against the intervening bystanders, this would not be associative discrimination. In that case, only secondary protection from retaliation would apply. If that same bouncer believes that allies of certain protected groups are “just as bad” as those groups themselves, there would be associative discrimination, given that the protected criterion is part of its reasoning. Thus, in a minimalist interpretation, fewer situations would be simultaneously protected by the prohibition of associative discrimination and the protection against retaliation.
The existence of the concept of indirect discrimination could also be an indication to give a broader, maximalist interpretation to the association. In the example of the employer not tolerating dissent, a minimalist interpretation means that the protected criterion is not a reason for the disadvantage. But it could be justifiable to argue that not tolerating dissent will have a disproportionate impact on vulnerable groups, as it is precisely they who benefit most from the prosocial behavior of bystanders who stand up for them.
In addition, a potential evidence problem arises in a minimalist interpretation. An intentional “actively transferred associative discrimination” can still be proven if the perpetrator regularly expresses the sentiment that combating discrimination is nonsense. Yet if the perpetrator never makes such statements or is unaware of any negative sentiments towards persons taking a stand against discrimination, proving that the protected characteristic plays a role in the reasoning will be almost impossible - even with the shift of the burden of proof. In that case the active bystander can still rely on secondary retaliatory protection.
The CJEU’s Coleman and CHEZ judgments leave room for both interpretations of association. Thus, in a maximalist reading of Coleman, one could argue that the mother is disadvantaged “on the grounds of disability” because she would not be disadvantaged if her child did not have a disability. In a minimalist reading, she is disadvantaged because prejudice against persons with disabilities seeps into how she is treated. The employer assumes that persons with disabilities are in need of care, thus making their parents bad employees because they have to look after their child more than other parents.
In the CHEZ judgment, according to a maximalist interpretation, the woman would not have been disadvantaged if her neighbors were not of Roma origin. In a minimalist reading, the disadvantage of the person without a protected characteristic is also based on prejudice against persons of Roma origin, but the association exists simply because the person without a protected characteristic is affected by the same disadvantage.
Some nuance to these interpretations is in order. When discrimination legislation was developed, the issue of the active bystander was by no means on the agenda. As a result, the position of this figure within the benchmarked outlines is unclear, to say the least. Moreover, the bystander illustrates that concepts such as attributive and associative discrimination are not mutually exclusive. Thus, in the example of the bystander who stands up for LGBTQIA+ persons, the connection between the adverse treatment and the protected characteristic could be attributive, but also an actively transferred association, because the perpetrator takes a detrimental action towards the bystander who stands up for the victim and thus associates with it.
It can be concluded that for the active bystander the actively transferred discrimination is the most relevant type of associative discrimination. Nevertheless, it remains unclear whether and to what extent the retaliation in response to resisting discrimination can also be directly qualified as discrimination. Both maximalist and minimalist interpretations offer insights into the matter, each with advantages and disadvantages. Ultimately, the question of the exact scope of “on the grounds of” remains unanswered for now.
C. Different Legal Regimes, Same Theoretical Reflections?
Due to some overlap between retaliation and associative discrimination, and the legal regimes protecting individuals against them, it is warranted to articulate some of the differences between them. The retaliation protection from the Belgian legislationFootnote 130 will be used to illustrate some of these, because of the brief wording of the stipulations in the respective EU directives.Footnote 131 The following part delves deeper into some of these differences, namely who is protected against what conduct, what needs to be proven to shift the burden of proof, what possible defenses can be raised and what remedies are available in Subection I. Attributive discrimination will not be part of this comparison, because it seems only relevant to the active bystander in very specific instances. Afterwards, Subsection II will compare these insights with the theoretical foundations of bystander protection that were developed in part B, highlighting some tensions between them.
I. Comparison Between the Legal Regimes
An important difference pertains to the legal basis and autonomy of the claim. The protection against retaliation is derivative, protecting persons because they filed a complaint, supported a victim or opposed discrimination, either formally or—as ruled in the Hakelbracht judgment—informally. Their legal claim is parasitic on the original act of discrimination and their resistance against it. Contrastingly, the protection against associative discrimination is more direct, protecting the bystander as an autonomous victim of discrimination that directly invokes the prohibition of discrimination, based on their association with someone who holds a protected characteristic.
This has implications for the causal link to the protected characteristic. In case of protection against retaliation, the adverse treatment must be a reaction to the bystander’s intervention, without the need for a protected characteristic to be involved in the legal debates. In case of associative discrimination, the protected characteristic must be at least a partial cause of the adverse treatment, be it minimally or maximally.
Consequently, this influences what exactly needs to be proven in order to make a successful claim. To claim protection against retaliation, a person would need to prove that they made a complaint, supported the victim, or asserted the right to equal treatment in any other way, and that the adverse treatment they are subjected to is in response to that complaint, support or assertion. However, because this would be very hard to prove, the Belgian and Flemish legislatures have implemented a shift in the burden of proof. Therefore, if there is an adverse treatment either within twelve months of the formal or informal complaint or intervention, or within three months after a legal procedure has been finished, the burden of proof shifts to the person enacting the adverse treatment or the person against whom the legal action has been brought. The defendant must then prove that the adverse treatment was not related to the complaint or intervention.Footnote 132 Although the legislation does not explicitly provide for this possibility, I contend that the defendant could also attempt to demonstrate that the alleged complaint, support or assertion of the right to equal treatment, was no complaint or intervention at all, thus not activating the shift in the burden of proof. This would occur mostly in case of an informal complaint or intervention.
To make an autonomous discrimination claim, a person would need to prove the existence of “facts from which it may be presumed that there has been a discrimination.” This means that the claimant would have to demonstrate that there is an adverse treatment and that there is a, minimal or maximal, causal link between this treatment and a protected criterion. The defendant can then either prove that there was no adverse treatment, that the adverse treatment was not related to the protected criterion, or that the adverse treatment was somehow justified.
Thus, in case of a formal complaint or an intervention with a paper trail, the evidence regime of retaliation protection makes it easier to bring a claim. In case of a complaint or intervention without written evidence —which is quite likely, as argued before in Subsection II of Section C — this is more complicated, but it would still be easier to demonstrate that there was a bystander intervention and an adverse treatment, than it would be to demonstrate a causal link between adverse treatment and protected criterion.
Regarding remedies, the Belgian legislation refers to the same provision in case of remedies against a discrimination or remedies against retaliation. In all contexts, these are either fixed compensation or the exact damages suffered, if proven.Footnote 133 In an employment context, these could additionally amount to six months of wages or in case of dismissal, re-integration in the organization.Footnote 134
Ultimately, bystanders would receive the same remedies, regardless of whether their claim is framed as retaliation or discrimination. However, a bystander will have more difficulties discharging the burden of proof when they pursue an autonomous discrimination claim, than they would have when pursuing the secondary retaliation protection. The question might therefore arise what exactly the added value is of pursuing an autonomous discrimination claim for active bystanders if there is no strategic incentive. To answer this question, it is useful to turn to the theoretical insights we have articulated in Section B.
II. Theoretical Reasons for a Distinct Recognition of the Active Bystander’s Harm
Previously this Article contended that from various theoretical perspectives, retaliation against active bystanders can be considered a moral wrong. A prioritarian account foregrounds its distributive impact, understanding it as morally wrong either because it undermines dynamics to remedy earlier wrongs, or because it punishes individuals acting on morally relevant grounds. Either way, prioritarians would have difficulty considering it discrimination in itself. Contrastingly, a recognition-based account focuses on the demeaning expression of discrimination, not recognizing individual’s equal moral status. In this way, the retaliation can be considered either a secondary wrong, as it impedes the bystander’s potential to recognize victim’s status, or an autonomous wrong that violates the bystander’s own moral status. In turn, a pluralist account reflects the multidimensional wrongness of retaliation against bystanders, just like it recognizes the multidimensional wrongness of discrimination itself. Thus, pluralists could consider bystander retaliation both an autonomous and a derivative wrong.
On the one hand, the current legal regime on protection against retaliation primarily conceptualizes harm to intervening bystanders as a derivative harm. That is, retaliation against bystanders is viewed as problematic mainly because it undermines their role in supporting the enforcement of equality law. The justification for legal protection in this case is largely instrumental: it is aimed at safeguarding the broader effectiveness of anti-discrimination measures. This is clearly reflected in the recitals of EU anti-discrimination directives, which articulate the need for protective measures to ensure the effective implementation of the principle of equality.Footnote 135
On the other hand, an evolutive interpretation of the concept of associative discrimination offers a path toward recognizing retaliation against bystanders as inflicting an autonomous harm, that directly violates the moral and legal status of the bystander themselves. Retaliation is not merely a secondary consequence of discrimination, but a distinct harm that reinforces systems of oppression. Seen this way, structural injustices such as racism, sexism, homophobia, ableism, and religious intolerance do not only harm those who are their primary targets, but also members of dominant groups who resist or refuse to participate in discriminatory practices. The active bystander, then, is not merely a witness or supporter, but a target in their own right for stepping out of line with dominant norms.
The figure of associative discrimination thus opens up the possibility of legally acknowledging bystanders as victims of discrimination in their own right, not only as agents of enforcement. However, this recognition creates a tension. As discussed in the previous section, the legal remedies for retaliation and discrimination—at least in Belgium—are functionally the same. Paradoxically, it would actually be easier for a bystander to prove that they supported a victim and were retaliated against, than it is to prove that they themselves were discriminated against through associative discrimination.
This raises a potential injustice. If bystanders are recognized as victims of discrimination—via associative discrimination—on the one hand, but face lower evidentiary burdens and still access the same remedies on the other hand, they might be treated more favorably than others who must meet the more demanding standards of a primary discrimination claim. The equal treatment of different types of victims in terms of remedies may appear fair on the surface, but it obscures the uneven burdens involved in establishing victim status.
E. Conclusion
The European directives provide an extensive legislative framework to combat discrimination. They have been criticized for the fragmentation with regards to their scope and protected grounds, and for the focus on the individual dimension of discrimination: the less favorable treatment of a victim at the hands of a perpetrator. Nevertheless, they introduce important legal mechanisms in Europe.
Furthermore, they are not completely blind to the collective dimensions of discrimination. For instance, the procedure through equality bodies and human rights institutions recognizes the social need to support victims. Another collective dimension, focused on in this Article, is the “active bystander,” who is involved in discriminatory situations alongside perpetrator and victim. Their current and potential role in discrimination cases has not yet been clarified, nor the role of legislation to incentivize their prosocial behavior.
Section B of this Article has argued that, besides practical reasons, there are theoretical justifications to protect this active bystander from retaliation. From a recognition-based perspective, retaliation against bystanders undermines their equal moral status and impedes their capacity to affirm the moral worth of victims. From a prioritarian standpoint, bystanders can play a key role in advancing the wellbeing of those most disadvantaged, making retaliation against them harmful. A pluralist theoretical lens brings these strands together and highlights the multidimensional wrongness of such retaliation: it is both a derivative harm—insofar as it impedes anti-discrimination enforcement—and an autonomous harm—insofar as it punishes morally commendable acts of solidarity.
There is already some legal attention for this actor, as shown Section C. “Bystander interventions” fall within the scope of the secondary protection against retaliation as provided for in the directives and Belgian legislation. The Hakelbracht judgment clarified that this protection does not only apply to formal interventions, such as a complaint, report, or legal proceeding. Indeed, the CJEU ruled that informal interventions, such as offering help, counselling, or assisting a victim, also activate the protection.
As bystander interventions could be a crucial component of the fight against discrimination, this Article reflected on whether the retaliation against bystanders could also be deepened through recognition of bystanders as autonomous victims of discrimination. One way is through attributive discrimination. But this seems only applicable to the active bystander in very specific situations. More promising is associative discrimination, developed by the CJEU in the Coleman and CHEZ cases. In this regard, a person without a protected characteristic becomes a victim of discrimination because of the connection with a person with a protected characteristic. This Article explored the question to what extent this legal mechanism would allow the bystander to invoke the prohibition of discrimination against the adverse treatment as a reaction to a “bystander intervention.”
For this, it was necessary to find out the exact meaning of discrimination “on grounds of” a protected characteristic. In a maximalist interpretation, this is the case once the disadvantage would not have occurred, had the protected characteristic not been present. In a minimalist interpretation, a disadvantage is only “on grounds of” a protected criterion if it played a substantial role in the discriminator’s reasoning, intentional or not. In the former interpretation, bystanders would be more likely to invoke the discrimination prohibition themselves, while in the latter interpretation there is less concurrence between the situations captured by the associative discrimination and the protection against retaliation.
Finally, Section D discussed how both retaliation and discrimination claims currently give access to the same remedies under Belgian law. This parity conceals a deeper normative tension. The evidentiary burden for retaliation claims is often lower than for an autonomous discrimination claim. If bystanders can be recognized as victims of discrimination in their own right, but require a heavier burden of proof to claim that status, a form of injustice arises. The remedies for the bystander claiming retaliatory protection would be the same as those for the initial victim and the bystander claiming autonomous victim status, while the procedural burdens are not.
The bystander already receives some protection in anti-discrimination law, numerous questions remain unanswered. As this Article has shown, the moral and legal wrongs suffered by active bystanders warrant their recognition not just as agents of enforcement, but as rights-holders and potential victims in their own right. One question that this Article did not ask is whether bystanders could also be duty-holders. Is there a theoretical and legal foundation to responsibilize passive bystanders, and attach legal consequences to that passivity? Some would argue for the imposition of positive non-discrimination duties for the actor or body in the best position, backed by a liability.Footnote 136
Beyond a duty to act or report, these positive non-discrimination bystander duties can include procedural obligations for the persons in power positions. The employer, school principal, or housing agency administrator could implement performant reporting mechanisms and bystander intervention guidelines, while the nightclub owner could implement anti-discrimination trainings for bouncers and bartenders. These actors can complement and facilitate the fight against discrimination of the hierarchically equal bystanders, which should also be reflected in the legislation.
This Article, however, limited itself to bystanders that intervene voluntarily, already creating tensions and pushing against the boundaries of current discrimination law. Establishing the bystander’s position in future discrimination cases and legislation will require action by judges and legislatures informed by further research.
Acknowledgements
The author declares none.
Funding Statement
No specific funding has been declared in relation to this Article.
Competing Interests
The author declares none.
Open access
