2.1. Direct discrimination and genocide: the relationship

It is generally accepted that genocide and direct discrimination on the grounds of raceFootnote ⁷ and/or religion are related to each other in several respects.Footnote ⁸ Direct racial and/or religious discrimination is considered to mean that an individual suffers disadvantage (is treated differently) on the grounds of his/her race or religion.Footnote ⁹ One could argue that such discrimination is included in the concept of genocide, which is defined as committing certain acts, such as killing members of a group, causing serious bodily or mental harm to members of a group, and so on, with the intent to destroy, in whole or in part, this national, ethnic, racial, or religious group.Footnote ¹⁰ The discriminatory aspect of genocide is generally argued to relate to the special intent to destroy a group ‘as such’.Footnote ¹¹ The intent to destroy a group absorbs the intent to discriminate against this specific group. However, it can be asserted that the prohibited conduct (actus reus) is also discriminatory, as it concerns not merely killing someone, but killing members of a protected group. So in terms of discrimination, the conduct constitutes discrimination against the individual members of the group, while the special intent refers to the intent to eliminate the group against the members of which discrimination was carried out.

Persecution on political, racial, national, ethnic, or religious grounds, a crime against humanity,Footnote ¹² is often compared with genocide given the discriminatory component in both concepts; i.e., the intent to discriminate against a specific group in persecution cases is compared with the intent to destroy a protected group in case of genocide. Although both crimes incorporate the intent to discriminate against a group, it is generally accepted that the special intent of the crime of genocide constitutes a distinctive feature between these two crimes.Footnote ¹³ Notwithstanding this difference in the mens rea, in light of the above explanation of the relationship between discrimination and genocide, it can be argued that in both cases there is discrimination against the individual members of the protected group when a specific conduct is committed. This discrimination is an aspect stemming from but at the same time separate from the intent to destroy a specific group in genocide cases and from the intent to discriminate against a protected group in persecution cases. Whilst the mens rea is different in these two crimes, the actus reus in both cases includes discrimination. Accordingly, given this similarity, the approach toward the identification of members of the protected group as well as the protected group in persecution cases can be considered in the examination of genocide cases, and, for the sake of consistency, these approaches should be harmonized.

2.2. Building blocks of a model of review for genocide cases

Although the concept of discrimination has been clothed with different understandings and interpretations,Footnote ¹⁴ the following conceptual ‘building blocks’ of direct racial/religious discrimination can be identified: harm suffered (less-favourable treatment, disadvantage), protected characteristic (race and/or religion) and a causal link between the harm and the characteristic (less-favourable treatment on the grounds of race and/or religion).Footnote ¹⁵

For instance, if a Roma person is not employed on the explicitly stated grounds of his Roma origin, this is a clear case of overt direct racial discrimination: he suffered harm (rejection) on the grounds of (causal link) his racial origin (protected characteristic). However, very often the reason indicated for the refusal does not directly and explicitly refer to the Roma origin of the person; rather the potential employer gives the explanation that, for instance, the position has already been filled, or the applicant did not have enough experience; i.e., it is unclear whether there was a causal link between the harm and the racial origin. The suffered harm and the protected characteristic are present, but it is not clear what the actual reason of the employer was. In such cases, it has to be strictly scrutinized whether or not there is a causal link between the applicant's Roma origin and his unsuccessfulness in getting this job. This and similar cases can be categorized as covert direct discrimination.Footnote ¹⁶

Covert direct discrimination cases should, however, not be confused with indirect discrimination cases. Indirect discrimination focuses on the discriminatory effect of a seemingly neutral measure as opposed to the causal relationship in (covert and overt) direct discrimination cases.Footnote ¹⁷ A typical case of indirect discrimination is when, for instance, a rule is introduced that the lunch break of cleaners will be unpaid. If a high percentage of cleaners belongs to a certain racial or religious group, it can be proved that this measure has a disadvantageous effect on this group, and thus constitutes indirect discrimination against them. In this case the reason of the perpetrator is irrelevant; discrimination is established on the basis of the effect/impact of the measure at stake on a specific group. Due to its focus on the effect of a seemingly neutral measure, the concept of indirect discrimination is, arguably, not applicable to genocide cases, and will not be discussed further.

Given the relationship between the concepts of direct racial and/or religious discrimination and genocide, these building blocks can also be explored in the definition of genocide (and persecution) relating to the conduct element, where the individual discrimination takes place. The harm would be a conduct, a list of which is provided in Article II Genocide Convention (killing, inflicting serious bodily or mental harm, and so on); the protected characteristics are the following: nationality, racial or ethnic origin, or religion, and it has to be proven that killing of or inflicting serious bodily and mental harm and so on on certain individuals were carried out on the grounds of the victims’ national, racial, or ethnic origin or religious affiliation; i.e., that there is a causal link between the harm and the protected characteristic.

2.3. Model of review for genocide cases

As explained above, the conduct (e.g., killing of members of a group, actus reus) can be regarded as a specific form of direct discrimination on the grounds of race and/or religion: the perpetrator commits the specific act (e.g., killing) on discriminatory grounds (causal link) in addition to the intent to commit this act (e.g., intent to kill, dolus actus). So the formula of the conduct could be described as follows:Footnote ¹⁸

This formula is based on the identification of the different elements of the conduct in genocide cases; i.e., the causal link can be regarded as an inherent element in the definition of the conduct. Accordingly, the explicit mention of the discriminatory component does not create any additional burden of proof. The purpose of the explicit acknowledgement of the discrimination component is to realize that such a killing, inflicting serious injury, and so on conceptually constitute discrimination, which, in turn, has consequences for the approach judicial bodies should take toward the identification of the members of the protected group and the protected group itself. So realizing that the conduct constitutes discrimination has effect on who can be regarded as victims of genocide, and, hence, which deeds are parts of the genocidal charges.

If the conduct can be seen as a specific form of discrimination, the formula of genocide will be the following:

When examining genocide cases judicial bodies follow a two-step approach: first they look at the act in combination with the dolus actus required for this conduct (e.g., intent to kill) and then examine the dolus specialis (intent to destroy the group specified in the definition). Some considerations have been given to discrimination only with regard to the dolus specialis.Footnote ²⁰ In light of the relationship between genocide and discrimination and the formula described above, the suggested approach, i.e., the model of review, would, however, be the following: judges should first look at whether discrimination on the specified grounds has taken place (killing members of the group), and then examine whether this discriminatory conduct has been committed with the intent to destroy this particular group. The suggested review model is, thus, the following:

Conceptually, these two stages of investigating genocide cases can be well distinguished; however, in practice they are most of the time intertwined in that the factual evidence and the legal argumentation of the two stages often overlap.Footnote ²¹ In certain cases, there might be some direct evidence about the intent to destroy a group – for instance, some official documents might exist. This intent can then also be used to prove the causal link at the discrimination/conduct stage.Footnote ²² In other cases, there might not be any evidence about the intent to eliminate a group; however, inferences might be drawn from the first stage. For instance, the scale and extent of discriminatory killing might indicate the intent to destroy a specific group.Footnote ²³

One could argue that advancing such a discrimination-based approach would introduce a motive element in the crime of genocide, and for this reason it cannot be accepted. Indeed, it is generally accepted that ‘[t]he motive for which a crime is committed, as opposed to the intention with which it is committed, is ordinarily irrelevant to guilt in criminal law’.Footnote ²⁴ However, this argument against acknowledging the discriminatory component of the conduct of genocide can be rebutted for, at least, two reasons. First, as mentioned above, the explicit acknowledgement of the discriminatory nature of the conduct merely aims to provide further assistance in the identification of the protected groups and their members, and does not aim to impose any additional burden of proof in the investigation of and deliberation on genocide charges. Second, and irrespective of the explicit identification of the conduct as discrimination, the claim is commonly made with regard to the special intent that

In a similar vein, it has been argued that ‘it should be necessary for the prosecution to establish that genocide, taken in its collective dimension, was committed “on the grounds of nationality, race, ethnicity, or religion”. The crime must, in other words, be motivated by hatred of the group.’Footnote ²⁶ These claims take the position that a motive is inherently included in the special-intent requirement of genocide. If this is indeed the case, the explicit acknowledgement that the conduct element of genocide (as opposed to the special intent) constitutes discrimination would not add a further motive element, as discrimination against the members of the protected group stems from this collective hatred against a specific group, and not from the individual motive of the perpetrator.

This suggested review model might place the problems relating to two fundamental aspects of genocide cases in a somewhat different light and help to identify the most appropriate solutions that could address them. These problems relate to the difficulties with and inconsistencies in identifying the members of the protected group and identifying the protected group itself.

3.1. Objective versus subjective approach

In the legal literature as well as in the case law a discussion is going on over which approach should be applied by international criminal courts with regard to the identification of the members of a protected group: the so-called objective or subjective approach.Footnote ²⁸ The objective approach means that the judicial body examines the objective existence of the racial or religious identity of the victim; that is, whether or not the victim actually belonged to a certain racial or religious group or actually possessed the so-called ‘objective’ features that identify the members of these groups. In the Akayesu case, for instance,Footnote ²⁹ the point of departure of the ICTR seems to have been based on an objective approach. The Chamber stated that

More concretely, the question was raised whether serious bodily and mental harm inflicted on a Hutu woman can be regarded as forming part of the genocidal conduct against the Tutsi group. The factual description of the case explains that this should not be the case:

The application of the objective approach can also be observed in persecution cases.Footnote ³² Since the Krnojelac caseFootnote ³³ it seems now established that

‘Discrimination in fact’ has been argued to mean that ‘[t]here must be a discriminatory effect upon the victims’,Footnote ³⁵ i.e., persecution can be established when the act of the perpetrator results in discrimination against those actually belonging to the group.Footnote ³⁶

Opposed to this view, the subjective approach focuses on the identification of the victims by the perpetrator.Footnote ³⁷ In the Gacumbtsi case, for instance, the Trial Chamber of the ICTR explained that

So the Trial Chamber was willing to take into account the perception of the perpetrators for the identification of the members of the group. Nevertheless, the Chamber did not take a clear-cut position as to which approach should be used; in their view both approaches could, and probably at best their combination should,Footnote ³⁹ be used.

Similarly, starting from the Kvočka case,Footnote ⁴⁰ which was later confirmed in the Naletilić and Martinović case, the ICTY accepted in persecution cases that ‘persons suspected of being members of these groups are also covered as possible victims of discrimination’.Footnote ⁴¹ In other words, the ICTY did not require that ‘discrimination in fact’ occurred; i.e., they applied the subjective approach to persecution,Footnote ⁴² interestingly without any reference to related genocide cases.Footnote ⁴³

Examples can, thus, be found for the objective as well as for the subjective approach in both genocide and persecution cases. The effect of the objective approach is that only those are considered victims of genocide who actually belong to a specific group, so only such conduct will amount to genocide which was committed against the ‘actual’ members of a specific group. As opposed to this, the acceptance of the subjective approach results in that in addition to the above-mentioned conduct, the conduct committed against those who were mistakenly perceived as belonging to a group also forms part of genocide (provided the existence of the special intent is proven).

It has been argued in the legal literature that while previously the objective approach dominated in genocide cases, at present there is a tendency towards applying the subjective approach.Footnote ⁴⁴ The subjective tendency was considered inappropriate by some scholarsFootnote ⁴⁵ or was argued to be a mere misperception by others.Footnote ⁴⁶ It seems that at present, at least in the practice of the Tribunals, concerning charges of genocide ‘[t]he prevailing view is that determination of the relevant protected group should be made on a case-by-case [basis], relying upon both objective and subjective criteria’.Footnote ⁴⁷ Interestingly, with respect to persecution the objective approach seems to have gained more acceptance in both the legal literature and case law.Footnote ⁴⁸

3.2. Discrimination law: perception

The question can legitimately be posed: which approach ought to be applied in order to be conceptually correct and consistent, given the review model suggested above? Is it the subjective or the objective approach, or the combination of the two?Footnote ⁴⁹ The relationship between genocide and discrimination entails that the conceptual considerations of discrimination law relating to the protected group and the members of this protected group can be and are relevant to genocide cases.

In discrimination law the focus is not so much on the objective or subjective approach; rather the related discussion centres around the existence of the causal link between the harm and the protected grounds and the different ways to prove it.Footnote ⁵⁰ Proving that a person with a certain racial origin or religious affiliation (protected characteristic) has suffered a certain disadvantage (harm) is not sufficient for concluding that discrimination has taken place. It also has to be proven that there was a causal link between the protected characteristic and the harm. As explained above, in covert direct discrimination cases the perpetrators often try to negate the causal link by referring to grounds other than the racial or religious affiliation of the victims. In these cases a comparison between the treatment of this group and the treatment of another group, the members of which do not have the racial or religious traits possessed by the group harmed, can show this causal link. Namely, such a comparison can reveal the perception of the perpetrator as to the importance of the racial or religious affiliation of these individuals, irrespective of whether or not this perception was actually correct – from the perspective of the victim or otherwise. Accordingly, proving the causal link in direct discrimination cases means that it is proven that the perpetrator perceived the victims as belonging to a certain group and that this perception made the perpetrator commit the conduct; i.e., it is the perception that matters as far as the identification of the members of the protected group is concerned.

Given the discriminatory element of genocide cases, this perception-based approach can also be asserted to be applicable to the first step of scrutinizing genocide complaints. The perpetrator commits the conduct because he considers the victims to belong to a certain group. Accordingly, a discrimination-led conceptualization would call for the application of the subjective approach.

As the above elaboration also illustrates, the focus on the subjective approach does not, however, mean that the objective approach or ‘objective’ factors, which are themselves argued to be ‘the product of social processes, informed by subjective social beliefs which have been transferred to a broader socio-legal context’,Footnote ⁵¹ cannot and should not have any role to play. Based on discrimination law, it can be argued that they can assist in proving the perception of the perpetrator; i.e., they have evidential role. Most of the time relying on objective factors would imply comparison. With the help of comparison based on ‘objective’ factors the perception of the respondent can be proven. This implies that the so-called legal and evidential steps should be distinguished. While the legal step requires the finding of a causal link based on the perception of the perpetrator, in evidential terms this causal link can be proven with the help of comparison based on ‘objective’ factors. It is, thus, important to keep in mind that using comparison based on the ‘objective’ features of the protected group does not mean that the focus of the legal concept should shift from the subjective to the objective approach; the ‘objective’ factors have ‘mere’ evidential function.

3.3. Arguments against the subjective approach

Admittedly, the perception-based (or subjective) approach has already raised eyebrows and several arguments were brought up against it. If one accepts the suggested review model, these arguments can, arguably, be convincingly rebutted.

First, it has been argued that a strict textual interpretation of the concept and the prohibited acts in the definition of genocide seems to have reference to the ‘objective existence’ of the specific groups, rather than to the perception of the perpetrator.Footnote ⁵² For instance, ‘killing members of the group’, a prohibited conduct in the definition of genocide, seemingly advances the idea that individuals are killed because they actually belong to a certain group. The lack of clarity in this respect was also emphasized by the international criminal courts. In the Semanza case, for instance, the ICTR's Trial Chamber explained that ‘[t]he Statute of the Tribunal does not provide any insight into whether the group that is the target of an accused's genocidal intent is to be determined by objective or subjective criteria or by some hybrid formulation’.Footnote ⁵³ It can be argued that, as this formulation is very similar to the definition of direct discrimination, in that it also lacks an explicit reference to the perception of the perpetrator, conceptually there is no obstacle to accepting the subjective approach. Accordingly, the formulation does not rebut the application of the subjective approach.

Second, the view that the discrimination aspect is absorbed in the dolus specialis has resulted in the plea that only the chapeau of the definition of genocide (with its reference to the dolus specialis) pertains to the assessment of the subjective view of the perpetrator, while the conduct is seen as requiring the assessment of the facts.Footnote ⁵⁴ This argument can be translated as indicating that the perception of the perpetrator is relevant only with respect to the genocidal intent, and thus the conduct needs to be assessed without any reference to the subjective view of the perpetrator. In other words, the conduct has to be committed against the actual members of the protected group.Footnote ⁵⁵ This position is well illustrated by the Muhimana case,Footnote ⁵⁶ in which the Chamber adopted the subjective approach to the extent that it could prove the special intent to destroy the Tutsis. In this case, mistakenly, a Hutu woman, perceived as Tutsi, was raped, for which the respondent later apologized. While this rape was considered to provide evidence for the special intent to destroy the Tutsi group, the conduct against her was not taken into account in terms of the actus reus.Footnote ⁵⁷ The proposed approach, however, adopts the view that the actus reus also has a discriminatory component, which is different from the special intent to destroy a particular group. Accordingly, given the discriminatory component of the conduct, the examination of the subjective view of the perpetrator can and should play a role also when the conduct is examined. For the Muhimana case it means that this Hutu woman should have been seen as a victim of genocide, given the perception of the perpetrator.

Third and closely related to the previous point, those arguing in favour of the objective approach often refer to the position taken in persecution cases.Footnote ⁵⁸ As indicated above, in persecution cases the Krnojelac approach seems to be the generally accepted position,Footnote ⁵⁹ where the position has been taken that the actual members of the group should suffer harm because without ‘actual’ discrimination ‘an accused could be convicted of persecution without anyone actually having been persecuted.’Footnote ⁶⁰ However, it is commonly acknowledged that a typical feature of discrimination law and hate crimes is that, although mistakenly harmed, the selection of certain individuals on the grounds of their racial origin or religious affiliation still causes harm to the protected group and its members by sending out stigmatization signals.Footnote ⁶¹ Accordingly, this position gives rise to the convincing argument that even in the absence of ‘actual’ discrimination, the perpetrator should be found responsible for persecution if it can be proven that he perceived the victims as belonging to a certain group and this belief made him commit the harmful acts. The Chamber in the Naletilić and Martinović case also confirmed this position, though using a slightly different reasoning. In their view, mistakenly harmed victims are also victims of persecution because they ‘have no influence on the definition of their status’, and they ‘are discriminated in fact for who or what they are on the basis of the perception of the perpetrator’.Footnote ⁶²

Fourth, it has also been argued that responsibility for a conduct committed mistakenly against a specific person can be excluded on the grounds of mistake of fact, which in turn would support the prevalence of the objective approach.Footnote ⁶³ Indeed, it can be argued that the mistake-of-fact argument and the subjective approach are mutually exclusive.Footnote ⁶⁴ Schabas, for instance, explained with regard to the ‘factual elements of the actus reus’ (conduct) that ‘[a]n accused might also argue mistake of fact, and even mistake of law, with respect to such knowledge-related aspects of the crime as membership in the targeted group’.Footnote ⁶⁵ If courts accept evidence proving that the perpetrator was mistaken as to the actual belonging of the alleged victim to the group targeted, it supports the view that the objective approach is preferred for the identification of the members of the protected group. However, the applicability of this ground of exclusion with regard to the identification of the members of the protected group can be rejected on, at least, three grounds. First, if one accepts the argument that hate crimes are specific in that they will be realized once stigmatization signs are sent out even if committed on the ‘wrong object’, the mistake-of-fact argument seems difficult to accept. Second, the mistake-of-fact argument raises the question what the ‘objective’ membership means: who defines the objective belonging, and what is the relevance of the victim's own identification? Racial (and religious) identities are social constructs, which are difficult to define in objective terms, and thus it is difficult to construct how one could be ‘factually’ mistaken. Still, one could say that if the perpetrators identify some features considered ‘objective’ by them based on which they can select the members of the protected group, and the person who suffered harm does not actually possess these specific features, the mistake-of-fact argument should be accepted. However, in such cases, it seems difficult to see why a criminal court should accept the lack of those ‘objective’ features that were identified by the perpetrators as grounds of excluding responsibility. Similarly, probably the victim was attacked because some of the ‘objective’ features of the victim made the perpetrator ‘believe’ that this person belonged to the group targeted. Again, the question can be posed why the lack of one of the many identified features should exclude this conduct from the crime of genocide. Third, conceptually the mistake-of-fact argument ‘serves to show that the defendant did not have the mens rea.’Footnote ⁶⁶ The example par excellence in this respect is bombing a civilian object, e.g., hospital, believing that it was a military facility. The mistake-of-fact argument can, in such cases, be accepted if the accused proves that he believed that this object was a military facility, even though in actual fact it was not; i.e., it aims to prove the perception of the perpetrator as opposed to the objective facts. In genocide cases this would mean that the perpetrator will not be held responsible if it can be proved that the perpetrator believed the victim was a member of the protected group, but, given the lack of certain ‘objective’ features, he, in actual fact, was not. However, in addition to the issues raised above, the following question can also be posed: if the victim did not possess those features, what made the perpetrator believe that they did belong to this group in the first instance? All in all, one needs to realize that due to this discrimination component the logic of the mistake-of-fact argument is difficult to apply and is different from the factual circumstances in other crimes. Given all these reasons, this specific ground of exclusion, arguably, does not support the prevalence of the objective approach over the subjective approach.

In sum, reliance on the review model in genocide cases implies that judicial bodies need to find ample evidence concerning the perception of the alleged perpetrator in order to construct the causal link between the conduct committed and the protected characteristic. The so-called objective approach can be used to prove the perpetrator's perception; i.e., this approach has a mere, though very important, evidential role. So a hybrid approach, which is often argued to be present in the case law of the tribunals, needs to be understood in this light. The focus on the perpetrator's perception has the effect that if an individual is proven to be perceived as being a member of a specific national, racial, ethnic, or religious group, he/she needs to be taken as a victim of the discriminatory conduct, irrespective of his/her actual belonging to this group.

4.1. Coverage of the protected group

The first issue, the coverage of the protected group, relates to situations where individuals are not members of the specific national, ethnic, racial, or religious group and are not perceived by the alleged perpetrator as being so. Still, these individuals suffer the same or similar harm as those belonging to the group because they give some form of support to the group actually attacked. It has been explained above that non-members can be regarded as ‘covered’ by the protected group if perceived by the perpetrator as belonging to such a group. Now the question is whether the coverage of the protected group can be stretched so far as to include mere ‘supporters’ of the actually protected group.Footnote ⁶⁷

In the Nahimana et al. caseFootnote ⁶⁸ the Appeals Chamber was confronted with this dilemma.Footnote ⁶⁹ The Prosecutor argued that

The Chamber, however, did not share this position. It explained that ‘the perception of the perpetrators of the crimes may in some circumstances be taken into account for purposes of determining membership of a protected group’; however, ‘in the present case Hutu political opponents were acknowledged as such and were not ‘perceived’ as Tutsi’.Footnote ⁷¹ They went further and stated that

While the Prosecutor tried to argue that the perpetrators perceived the Hutu political opponents as belonging to the Tutsi group, the Chamber concluded that in this case those individuals were not conceived as members of the Tutsi group, instead they were seen as a political group, which is not a protected ground under the Statute and the Convention. These different positions illustrate that there are some uncertainties as to the coverage of the protected group. At first sight, both arguments seem reasonable: the Hutu political group supported the Tutsis, they had a close link to them, and they also contributed to the protection and survival of Tutsis; still, this Hutu political group was clearly not seen as ethnically belonging to the Tutsi group. So how should the judicial bodies proceed in such cases?

Discrimination law might provide an answer to this dilemma, where the concept of ‘discrimination by association’ has emerged recently, and has gained increasing acceptance both in practice and in academic circles.Footnote ⁷³ The first remarkable point in this process was the Maduro AG Opinion in the Coleman case before the European Court of Justice relating to discrimination against the mother of a disabled child. The question was whether this discrimination could be seen as discrimination on the grounds of disability. In the view of Maduro AG ‘[o]ne way of undermining the dignity and autonomy of people who belong to a certain group is to target not them, but third persons who are closely associated with them and do not themselves belong to the group’.Footnote ⁷⁴ Conceptually, ‘discrimination by association’ would also result in regarding the members of the protected group as less ‘valuable’, the prevention of which is one of the aims of discrimination law. How could this concept be applied to the examination of the discrimination stage of genocide cases?

The Nahimana et al. case shows that the arguments of the Prosecutor as well as the Appeals Chamber were based on the perception of the perpetrators, which, in light of the conceptual elaboration above, can be praised. In the Chamber's view there was no evidence that could support the position that the perpetrators considered the Hutu political opponents as belonging to the Tutsi ethnic group. Indeed, based on the available facts and the inferences that could be drawn from them, one could argue that in this case the Hutu political opponents were not regarded as belonging to the Tutsi ethnic group. However, this does not mean that they cannot be considered victims of the conduct constituting genocide.Footnote ⁷⁵ The concept of ‘discrimination by association’ could cover this group: they suffered harm because they were closely associated with the Tutsi. Turning the argument around, one could say that had they not been associated with the Tutsi ethnic group, the Hutu political opponents would not have become victims of killings, injury, and so on. The most relevant question to be answered is: what was the main reason to attack them? It seems that they were not attacked because they were an opponent political group, but because they were closely associated with an ethnic group that was to be destroyed.

Accepting the concept of ‘discrimination by association’ would, thus, mean that the coverage of the protected group can, in addition to those neither belonging to a protected group nor perceived by the perpetrator as such, be extended to those who provide support to this protected group.Footnote ⁷⁶ It should be clear that ‘discrimination by association’ does not mean that a new protected ground, namely political view or opinion, is added in a hidden manner to the enumerated grounds in the Genocide Convention and the statutes of the international courts. This concept does not focus on the political opinion of the group, though in certain situations it is definitely closely related to providing support for a group. In order for these attacks to constitute genocidal conduct, it has to be proven that these individuals were not targeted because of their political belief, but because their support of this group hindered the execution of the genocidal plan. Such reliance on the concept of ‘discrimination by association’ would also provide higher protection for those being actually targeted by the genocidal conduct. Interestingly, the broad coverage of the protected group was also accepted with respect to persecution in the Naletilić and Martinović case,Footnote ⁷⁷ where the Chamber concluded that ‘[t]he targeted group must be interpreted broadly, and may, in particular, include such persons who are defined by the perpetrator as belonging to the victim group due to their close affiliations or sympathies for the victim group’.Footnote ⁷⁸

4.2. Approach to the identification of the protected group

The second issue relates to the approach adopted to the identification of the protected groups.Footnote ⁷⁹ The question, among others,Footnote ⁸⁰ has been raised in the case law whether or not one could commit genocide against a group, which is identified by the lack of certain characteristics, e.g. non-Serbs. Put differently, the dilemma is whether the discriminatory conduct (and thus genocide) needs to be directed against a specifically identified national, ethnic, racial, or religious group or whether it can also be aimed at all individuals who are excluded from the group to which the perpetrators consider them to belong. In the case law and in the literature these two methods are referred to as ‘positive approach’ and ‘negative approach’, respectively.Footnote ⁸¹ There is a general consensus that the negative approach is not appropriate in genocide cases for several reasons, among which are the arguments that it extremely broadens the victim group and that it does not fit the definition.Footnote ⁸²

Conceptually, the so-called ‘negative approach’ would, indeed, be very broad, as it includes each and every individual who does not belong to the perpetrator's group. However, the fact that a group of individuals is described by the perpetrators as non-Serb does not necessarily imply that a negative approach is applied. As explained, conceptually, discrimination is based on the perception of the perpetrator, which perception is constructed upon certain traits of individuals. Accordingly, the judicial bodies need to look for those features that the perpetrators took into account when they targeted certain groups, and identify the group based on these traits.Footnote ⁸³ This approach might lead to the creation of a group the members of which do not necessarily regard themselves as a coherent group. This construction, as mentioned above, is based on the idea that, strictly speaking, racial and ethnic groups are psychological and social constructs,Footnote ⁸⁴ and do not have an ‘objective’ existence.Footnote ⁸⁵ These ethnic or racial groups ‘are subjectively established, depending on particular conceptions of in-groups and out-groups in society. Since these conceptions vary in time and space, different proxies are used to single these groups out.’Footnote ⁸⁶ In other words, the perpetrators can create a group; i.e., a group that does not necessarily have an ‘objective’ existence.Footnote ⁸⁷ It is, however, essential that the features the perpetrators perceive are based on national, ethnic, racial, or religious proxies; e.g., language, skin colour and so on. Unlike the Akayesu method, with a clear-cut definition for each protected group,Footnote ⁸⁸ the suggested approach supports the view that these four groups should be considered to ‘help to define each other, operating much as four corner posts that delimit an area within which a myriad of groups covered by the Convention find protection.’Footnote ⁸⁹ Thus, one does not need to fit a group into one of the groups, rather the focus should be on the traits that constitute the foundations of these groups enumerated in the Genocide Convention. Accordingly, it follows from the discrimination component of the review model that a protected group can also be constructed by the perpetrators themselves, which means that any reference to individuals not belonging to a protected group should be strictly scrutinized in order to identify the actually targeted group.