The Prosecutor v. Vojislav Šešelj: A Symptom of the Fragmented International Criminalisation of Hate and Fear Propaganda

In 2016, the icty Trial Chamber found one of the greatest hate and fear propagandists of the Yugoslav wars, Dr Vojislav Šešelj, not guilty on all counts of the indictment. A full comprehension of the role the propaganda played was lost and the partial reversal of the judgment at the Appeals Chamber provided little improvement in this regard. Yet the blame does not solely rest with the Chambers but also with the Prosecution and an utterly fragmented law applicable to hate and fear propaganda. This article looks in depth at the Šešelj case in order to highlight the many hurdles to effective prosecution, some specific to the case and others symptomatic generally of propaganda trials. It then takes a multi-disciplinary approach in presenting the nature of hate and fear propaganda to suggest a broader way of looking at causality as well as to argue for reform of the current applicable law.


Introduction
Propaganda, in its various forms and manifestations, is as old as recorded history, and its philosophical origins can be traced back to ancient Greece.2 It is a deliberate and systematic attempt at shaping perceptions and manipulating the cognitions of the listener in order to direct their behaviour in a way that furthers the intent of the propagandist.3 The intended influence and manipulation distinguish it from dissemination of mere factual information.4 As one commentator put nearly fifty years ago, propaganda is: 'one of the most dangerous sources of international friction and war … [and that] the presence of unrestrained propaganda can sometimes make the difference between peace and war' .5 Because of its omnipresence, propaganda has largely come to be accepted as a fact of political life and its most pernicious forms remain inadequately addressed in international criminal law (icl).
In terms of propaganda that specifically spreads fear and hatred towards a particular out-group in order to foster crimes against its members6 there is a wide spectrum of approaches to it offered by a number of international 408 international criminal law review 20 (2020)  In Section 2, the article first examines how the concept of 'hate speech' has been defined by various international institutions, in the areas of international human rights law as well as international criminal law (icl). This Section further demonstrates the fragmentation of the law and the lacuna that is still present in terms of international criminalisation of 'hate and fear propaganda' as a particular phenomenon that includes a set of components, which cannot simply be equated with the basic notion of 'hate speech' .
Section 3 analyses the case of Vojislav Šešelj as a case study of the current problems in dealing with hate and fear propaganda at international criminal tribunals. It will delineate the propagandistic activities of Šešelj and demonstrates how the Trial Chamber (TC) failed in understanding the workings of his propaganda while simultaneously being derailed by the fragmentation of the applicable law. The Section will provide a critique of the prosecutorial method in selecting evidence as well as the overly restrictive interpretation of the law by both the TC and AC. The analysis of the Šešelj case is conducted with a view to avoiding mistakes in the future and understanding the underlying legal problems that require a reform of the existing law.
Section 4 takes a multi-disciplinary approach at identifying why the law and its current application at international criminal tribunals fails to approach hate and fear propaganda in an effective manner. It furthermore highlights the necessary shifts in legal perceptions of this phenomenon that would allow more meaningful conclusions in the jurisprudence in line with the findings of experts in the social sciences.
Section 5 assesses the work of the International Law Commission (ilc) on the new draft convention on crimes against humanity (cah) and the preventative function of international criminal law with the recommendation of adding a new inchoate crime -incitement to cah. The article concludes with proposing how hate and fear propaganda should be criminalised and adjudicated at the international level.

International Law Applicable to Hate and Fear Propaganda
This section will examine briefly the various domestic and international approaches to speech acts broadly termed 'hate speech' from the perspective of human rights and criminal law. It will further demonstrate the fragmentation of the law and the lacuna that is still present in terms of international criminalisation of 'hate and fear propaganda' as a particular phenomenon that includes a set of components, which cannot simply be equated with the basic notion of 'hate speech' .

412
international criminal law review 20 (2020) 405-491 however the final wording also now includes incitement to discrimination or hostility without a need for violence.30 The definitions of the terms remain somewhat of an enigma, as the hrc General Comment 34 is exceedingly summary, more detailed provisions in early drafts having been dropped.31 Nevertheless, as Schabas notes, discrimination is a 'familiar concept in international human rights law' thoroughly addressed in Articles 2 and 26 of the iccpr.32 It can be understood as any distinction, exclusion, restriction or preference of different categories of people, the list of which keeps expanding over time,33 with the purpose or effect of nullifying or impairing the recognition, enjoyment or exercise, on an equal footing, of human rights and fundamental freedoms in the political, economic, social, cultural or any other field of public life.34 In terms of the other two 'harms' referred to in Article 20(2), the Special Rapporteur described 'violence' as 'the use of physical force or power against another person, or against a group or community, which either results in, or has a high likelihood of resulting in, injury, death psychological harm, maldevelopment or deprivation' .35 On the other hand the term 'hostility' was described by the Rapporteur as 'a manifestation of hatred beyond a mere state of mind'36 in the form of actual harmful acts.37 The term is nevertheless vague and can blur the distinction between qualified and unqualified hate speech. For  For more on the work of the international ngo Article xix see www.article19.org. 40 ngo Article xix, supra note 34, Principle 12(1)(i) 'the terms 'hatred' and 'hostility' refer to intense and irrational emotions of opprobrium, enmity and detestation towards the target group' . Even narrower definitions of hate speech than that in Article 20(2) of the iccpr demand not solely incitement but instigation of the listener and thus a causal link with the discrimination, hostility, violence, or other harm subsequently committed, which is what is mainly the trend in icl, described in detail below. Approaches in the middle, require the likelihood of subsequent harm occurring, in which case context is essential in defining such likelihood. The European Court of Human Rights (ECtHR) has at times used a multifaceted test that takes into account a variety of factors including the likelihood and seriousness of the consequences of a particular expression and the intention of the speaker, while at other times it has disregarded this for an openended and context based so-called democratic necessity approach. 41 From the travaux préparatoires of iccpr Article 20(2) one can observe that the term 'hate propaganda' was often used throughout the drafting debates and when the term 'advocacy' was introduced, some delegates maintained that it should be understood as 'systematic and persistent propaganda'42 and others that it must mean 'repeated and insistent expression' .43 Interestingly, paragraph 20(1) still uses the term propaganda for war while there were no efforts to define it during the drafting.44 Delegates in the Third Committee indicated that the term had already been employed in national and international legal norms, in various General Assembly resolutions and in the judgment of the International Military Tribunal.45 The intent of the drafters of Art. 20(1) was to prohibit propagandistic incitement roughly comparable to that practised in the Third Reich. 46 Schabas thus understands propaganda in this context to mean 'intentional, well-aimed influencing of individuals by employing various channels of communication to disseminate, above all, incorrect or exaggerate allegations of fact… negative or simplistic value judgments whose intensity is at least comparable to that of provocation, instigation or incitement ' .47 Similarly, in terms of advocacy of hatred, the reference to Nazi views and Nazi-like propaganda was obvious. A text proposed by the representative of the Soviet Union, explicitly referred to fascist-Nazi views and the propaganda 41 S. Sottiaux, 'Leroy  46 Schabas, supra note 30, p. 581. 47 Ibid.
international criminal law review 20 (2020) 405-491 of racial and national superiority.48 Mr. Jevremović representing Yugoslavia stated that Nazi ideas had been the cause of the death of two million of his countrymen and that the people of the region 'knew what was meant by incitement to hatred' .49 He further noted that there were already laws in Yugoslavia prohibiting such incitement and that it was necessary to introduce the idea into the covenant.50 Despite this, the highly dangerous phenomenon of what can best be characterised as repetitive and systematic hate and fear propaganda cannot simply be equated with any subset of hate speech but rather includes a comprehensive set of specific components. As with all propaganda, it is designed to last longer than a single utterance or speech act.51 It is systematic and aimed at defining the very 'level of reality on which people think, discuss, and act' .52 The most precarious methods of hate propaganda 'are those that intersect with hate media, which stigmatizes and demonizes an out-group' .53 This is equally true of hate propaganda used by governments and other organised groups as part of a systematic process of persecution to prepare the public to commit atrocities against other members of the society.54 Crucially, it furthermore not only characterises the out-group in highly negative stereotypical terms (hate propaganda) but also as a threat to the survival or well-being of the in-group (fear propaganda), which in turn, is proposed to be managed with a 'solution' ranging from discrimination to physical separation or even annihilation. 'hate speech' and 'fear speech' are components of distinct but corresponding conduct geared towards achieving the same ends and goals. 56 Oberschall explains how 'the public makes sense of public affairs through a cognitive frame that establishes the truth value of perceptions, beliefs, opinions, attitudes, and action norms' .57 Threat messages in the mass media are the most effective in switching the public's cognitive frame of peaceful inter-group relations to a crisis frame justifying coercion and violence.58 This phenomenon, also called 'paranoia propaganda' , consists of fostering delusions of danger from external enemies and traitors at home and of complete dependency upon leadership, party and ideology.59 Its effects have been recognised in international criminal jurisprudence, for example in Nahimana et al. where most references to fear 'constitute inferences and links between the evidence of fear propaganda and the commission and perpetration of a wide range of physical crimes' .60 Referring to the newspaper Kangura, the TC stated that '[t]hrough fear-mongering and hate propaganda, [it] paved the way for genocide in Rwanda, whipping the Hutu population into a killing frenzy' .61 Furthermore, in Brđanin, the TC addressed the combination of spreading fear and hatred by stating the following: By his public statement the Accused created fear and hatred between Bosnian Serbs on the one hand and Bosnian Muslims and Bosnian Croats on the other hand, inciting the ethnic groups against each other. The Accused repeatedly used derogatory language to refer to non-Serbs, calling them 'Balijas' (Muslims), 'Ustaša' (Croats), 'Šiptar' (Albanians), 'vermin' , 'scum' , 'infidel' and second rate people.62 As noted by Predrag Dojčinović, 'through the distinct combination of fear, hate, derogatory and dehumanising, and indirect culture-specific references 56 Dojčinović, 'Introduction' , supra note 7, p. Hate and fear propaganda furthermore rest on a number of persuasion techniques to speed up this process, such as relentless repetition of its never changing narrative; speaking in the voice of trusted authorities including God,64 the majority of the in-group, ancestors, history, national heroes, and experts;65 and employing falsehoods, from a selective omission of facts, deliberate mischaracterisation of events and adversaries to out and out fabrication and lies. According to Oberschall, it is the latter that is the essential element of hate propaganda.66 An often-used example are accusations in a mirror, 'a rhetorical practice in which one falsely accuses one's enemies of conducting, plotting, or desiring to commit precisely the same transgressions that one plans to commit against them' .67 Despite the far broader and more complex nature of such propaganda, current international criminalisation is limited mainly to incitement of violence, leaving out much of the hate and fear propaganda surrounding it. Judgments that have dared to include it into the criminalisation of hate speech outside strict incitement and instigation have been reversed in later jurisprudence.

International Criminalisation of Hate Speech and Hate and Fear Propaganda
Neither hate speech nor systematic hate and fear propaganda are internationally criminalised as such, although they can in part constitute international crimes. The concept of propaganda is furthermore often used in the jurisprudence to describe 'behavioural patterns and forms of intent in a variety of ideological, political and military, individual or collectively coordinated efforts' .68 Since then, the Convention on the Prevention and Punishment of the Crime of Genocide (1948) established the international crime of direct and public incitement to commit genocide.69 This was incorporated verbatim into the statutes of the icty (1993) and the ictr (1994), while the icc Statute (2002) established in its 'General Part' , in respect of the crime of genocide, individual criminal responsibility for anyone who directly and publicly incites others to commit said crime.70 On the other hand, publicly uttered hate propaganda falling short of direct incitement to commit genocide was not criminalised in any of the statutes. In fact, during the drafting of the Rome Statute, Roger Clark and a few others suggested including incitement to all four major crimes, however the proposal gained little traction.71 Strongly opposed to the idea, some delegations even felt that incitement as a specific form of complicity in genocide should not be included in the 'General Part' of the Statute but only in the specific provision on the crime of genocide (Article 6) in order to underline that incitement was not recognised for other crimes.72 Yet such incitement is covered by several forms of complicity. to what has been described as a 'residual form of accessory liability' since it establishes 'the lowest objective threshold for participation' and thus constitutes a lower level of blameworthiness than other liability theories.77 While this mode of liability is based on the 'common purpose theories' from which the icty established the joint criminal enterprise (jce) doctrine78 it nevertheless differs from the latter doctrine and is unprecedented in icl.79 It applies to a person who intentionally contributes to a group crime, either with the aim of furthering the criminal activity or criminal purpose of the group or in the knowledge of the intention of the group to commit the crime. Yet membership of the person in the group or the existence of an agreement between the person and the group are irrelevant as it is rather the intentional contribution that is the basis of responsibility. The mens rea requirement reshapes part (d) into a version of the crime of joining a conspiracy, which is a crime based on individual responsibility, arguably formulated this way to overcome the collective responsibility implicit in the variations of jce, particularly jce iii.80 Dojčinović notes that 'any effective propagandistic campaign at the leadership level in modern times must be an enterprise and not merely a personal 76 icty Statute, Art. 7; ictr Statute, Art. 6(1 attempt at instigating groups and individuals to commit a crime' .81 He thus finds the concept of conspiracy to commit a crime or the corresponding doctrine of jce to best reflect this phenomenon that encompasses both a collective intent and a common objective and provides for the liability both of the inciter/ instigator and the liabilities of the incited/instigated.82 Importantly, for the establishment of liability in these instances, a causal link is necessary between the speech and subsequent crimes committed,83 while such a link is not necessary in terms of incitement to genocide.84 jce liability requires that the act or omission had a 'significant' contribution or effect on subsequent crimes, while aiding and abetting, instigating, soliciting and inducing, ordering and planning, all require a higher threshold, i.e. a 'substantial' effect or contribution.85 There are indications however, that the 81 Dojčinović, 'Introduction' , supra note 7, pp. 1, 7. 82 Ibid., p. 10. 'Conspiracy' is both an inchoate offence and a complicity doctrine (a basis for holding a person accountable for the consummated offences of another). A common law conspiracy is an agreement, express or implied, between two or more persons to commit a criminal act or series of criminal acts, or to accomplish a legal act by unlawful means. It has been frequently prosecuted particularly in the United States but more recently courts and scholars called for its reform or abolition, see  (Wilson,ibid.,p. 253). This is however an overstatement of the significance placed on an actual genocide by the two ictr judgments, which rather consider the existence of actual genocide as a contextual element that may help determine the mens rea of the speaker as well as how the intended audiences understood the speech and not a sine qua non in establishing the incitement. Other 422 international criminal law review 20 (2020) 405-491 after paragraphs (b) to (d) dealing with accessorial liability.97 The present authors, however, are of the view that the correct interpretation remains that incitement to genocide here also breaks with the dependence of the act of complicity on the actual crime, abandoning the accessory principle that governs paragraphs (b) to (d).98 Prominently, the same utterance that will constitute incitement to genocide under paragraph (e) will amount to soliciting or inducing under paragraph (b) when a causal link can be established between the speech and an actual genocide being attempted or committed.99 This further demonstrates the raison d'être of paragraph (e), which is to provide a tool for possible intervention before the speech has had its desired effect and a genocide actually takes place.100 In the opposite cases, however, it will be up to the Prosecutor to decide which form of liability should be pursued, under paragraph (e) or (b) in light of the available evidence. Most likely, all available forms would be argued in the indictment as is the standard practice.
Outside the context of incitement to genocide, modes of accessorial liability provide the basic tool for prosecuting hate propagandists and in practice they prove greatly challenging for a successful prosecution in several ways resulting in a high failure rate before the two ad hoc tribunals and the icc.101 Firstly, the law applicable to hate propaganda is too fragmented.102 The Prosecutorial approach in Šešelj reflected this problem as it employed the so-called 'catch-all practice' by firstly drawing a distinction between crimes physically committed by the Accused and other crimes committed by way of jce and then 'obscuring' this framework by alleging Šešelj's membership in a jce for all of the crimes.103 The Indictment furthermore claimed instigation, jce as well as aiding and abetting based on what the TC characterised as the same factual contextual elements may provide sufficient indicators of the mens rea and the 'directness' of the incitement as an element of the actus reus. basis.104 While the Prosecution was criticised by the Chamber for this 'circular approach in which practically each crime has multiple qualifications' , it is but a natural consequence of the unfortunate fragmentation of the law and quite common in practice.
It is also worth noting that, while the common law criminal theory does not distinguish between 'principals' , 'secondary participants' , 'accessories and accomplices' or 'individuals in a group acting with a common purpose' in terms of guilt or applicable penalty, many civil law systems perceive modes of participation through a hierarchical lens that implies a lesser penalty for participation falling short of principal perpetration.105 Thus, if we are to consider hate propaganda merely under accessorial liability, victims might feel that the perpetrators of such propaganda, who are in fact more morally culpable, are found less culpable in law than those who, under its influence, physically carried out the crimes. In German law, however, the instigator is punished in the same way as the principal, thus a hierarchy does not apply.106 Third, the causal link between the speech and subsequent crimes committed, which all modes of accessorial liability require, albeit to differing extents as mentioned above, poses the main challenge for the Prosecution. Many have thus questioned why incitement to international crimes such as war crimes and cah has not been criminalised in the same way as incitement to genocide, which would not only make prosecutions more straightforward but would also serve the preventative function of the law by providing a tool for intervention before subsequent crimes are committed.107 Fourth, as mentioned above, mostly left out of criminalisation is speech that does not strictly speaking amount to incitement to violence even if it constitutes a necessary part of the propaganda leading to the success of such incitement.108 Furthermore even direct incitement may not suffice and the added prerequisite of a causal link with subsequent crimes committed is necessary for the speech to be considered criminal.
The most encouraging development in terms of hate propaganda in recent jurisprudence has been in Ruggiu and Nahimana, where the ictr Appeal Chamber held that as part of a widespread and systematic attack against a civilian population, hate speech can also in and of itself constitute the physical 104  commission of persecution as a crime against humanity.109 The Appeal Judgment in Šešelj has now brought the jurisprudence of the icty in line with that of the ictr on the matter by confirming this possibility.110 It has nevertheless unjustifiably left almost all of Šešelj's hate propaganda outside of this characterisation, save for one speech in Hrtkovci, and even here the finding of persecution mainly rested on a finding of instigation showing a reluctance on the part of the Court to apply criminal responsibility sufficiently broadly with regards to hate propaganda, i.e. beyond successful incitement to violence, i.e. instigation.
Importantly, within the context of the crime of aggression, the icc Statute in Article 8 bis criminalises also its planning and preparation. While not yet tested in the jurisprudence, Dojčinović notes that there are no major evidentiary and jurisprudential obstacles inhibiting the investigation and prosecution of historical and political acts of propaganda and incitement, or instigation, already during the formative period of this act.111

The Almost Acquittal of Vojislav Šešelj as a Reflection of a Fragmented Law and a Poor Judicial Understanding of Hate and Fear propaganda
This section will analyse the case of Vojislav Šešelj as a case study of the current problems in dealing with hate and fear propaganda at international criminal tribunals. It will delineate the propagandistic activities of Šešelj, one of the greatest hate mongers of the Balkan wars of the 1990s, and demonstrate how the TC failed to understand the workings of his propaganda while simultaneously being derailed by the fragmentation of the applicable law that manifested itself in the Office of the Prosecutor (otp) bringing multiple potential qualifications for each crime, none of which could be applied by a straightforward approach. The section will provide a critique of the prosecutorial method in selecting evidences as well as the overly restrictive interpretation of the law by both the TC and the AC in the present case. The analysis of the case will be conducted with a view to avoiding mistakes in the future and understanding the underlying legal problems that require a reformulation of the law.

Vojislav Šešelj's Hate and Fear Propaganda
Vojislav Šešelj was the founder and president of the Serbian Radical Party from 23 February 1991, and a member of the Assembly of the Republic of Serbia. In 1989, Šešelj was declared Četnik duke by Momčilo Đujić, a Četnik leader from World War ii, with a mandate to make a unitary Serbian state where all Serbs would live, occupying all the Serb lands, the so-called Greater Serbia. Šešelj established a military wing of his party, created a War Staff, promoted the Četnik movement's militaristic traditions, appeared in military attire at frontlines and most importantly, relentlessly spread his fear and hate propaganda aimed mostly at Croats and Bosniaks. Šešelj studied the mass psychology of fascism and in his book, entitled Ideology of Serbian Nationalism, published in 2002, he expressed the belief that propaganda is based on the fact that the majority of people are ready to believe indiscriminately in everything they read, hear or see on television.112 Anthony Oberschall conducted a content analysis of Šešelj's 1990-1994 mass media propaganda that revealed that he massively used claims of past and on-going Serb victimhood coupled with claims of a threat being posed to Serbs by an endless list of victimisers, ranging from the West, the Vatican, Germany, the UN, to the communists, Slovenes, the Muslims, and, at the top of the list, the Croats in addition to numerous others.113 Sometimes the threat was presented as coming from within and Šešelj even accused Serb media of 'anti-Serbian propaganda' . 114 Šešelj's claims of Serb victimhood and threat messaging, coupled with a denial of any Serb responsibility formed the moral justification of collective violence and the core of the crisis frame in ethnic relations.115 He rejected the possibility of compromise and non-violent conflict management116 and propagated that the natural sentiment between ethnic groups was that of hatred, antipathy and rejection making co-habitation impossible.117 On top of this basis he outright advocated coercion and violence issuing numerous threats and warnings, particularly the phrase that 'rivers of blood will flow' .118 His claim 112  international criminal law review 20 (2020) 405-491 was that self-defence, revenge and retribution justified the violence.119 A big part of Šešelj's xenophobic and xenomisic narrative was a constant glorification of Serbs and the dehumanising of other nations. While in his theory Serbs were a historic nation, everyone else, i.e. Croats, Macedonians and Muslims, were 'artificial' and as such did not have any right to their own states and were not even worthy of consideration.120 Thus Šešelj characterised the Croats as the 'last trash of Europe' , a 'genocidal nation' , who has 'never been honest with anyone' , a 'poisonous snake' who bites Serbs on the heart and whose head Serbs need to smash.121 In numerous instances he called for the 'amputation' of Croatia. The snake analogy mirrors both Nazi propaganda against the Jews122 as well as Hutu propaganda against the Tutsis.123 Misuse of history and fabrication of facts were tools Šešelj constantly employed, using partial, biased, misleading and sometimes outright fabricated information or history (for example claiming that Dubrovnik had always been a Serb town)124 as well as the crimes taking place during the war (presenting ethnic cleansing as a consensual, spontaneous and civilised exchange of populations).125 In other words, Šešelj carefully used the main persuasion techniques employed in fear and hate propaganda, whereby the out-group was negatively stereotyped and, in its entirety, blamed for historical crimes while at the same time presented as a threat to the Serb nation and its historically legitimate aspirations.
Oberschall's analysis, which scientifically analysed and clearly demonstrated the workings of Šešelj's propaganda was however not given the weight it deserved at trial. Oberschall's expertise was challenged by Šešelj and the Court upheld his objections on two grounds. Firstly, on the ground that Mr Oberschall did not speak Serbian and was not an expert on Serbian mass media or nationalist propaganda and secondly on the absurd ground that he referenced eighty-seven books by authors other than himself showing, according to the Court, insufficient expertise in the subject.126 Oddly the Court nevertheless allowed the testimony, yet failed to grasp its usefulness for a conviction. As Oberschall himself explained the situation: No matter how I explained my methods to them, I was never going to convince the judges that my social-science approach could help them. Judges are interested in a specific causal sequence that leads to a specific crime… The question is, can you connect the speech of one person to the crimes? That person alone and not the others? No. It's the ensemble that does it. For propaganda to be effective, you don't need to influence 100 percent of the population…[yet] in the judges' way of thinking, if it's not 100 percent, then it's not a cause.127 At risk of oversimplification, it was indeed this narrow approach to causality that was at the crux of the TC's failure in finding individual criminal responsibility based on Šešelj's fear and hate propaganda, as analysed below. Yet, the TC shares the responsibility for the failure both with the Prosecution, which decided not to bring to the Court the most essential evidence relevant to the question of causation, as well as with the applicable law, which required the proof of a high level of causality, many times difficult or nearly impossible to demonstrate.

3.2
Šešelj's Indictment and Trial at the icty Šešelj was indicted in January 2003 and surrendered voluntarily to the icty a month later.128 Under the Third Amended Indictment, issued in December 2007, the Prosecution charged Šešelj with persecution, deportation, and other inhumane acts (forcible transfer) as cah as well as five counts of war crimes. The Prosecution alleged that, acting individually, or as part of a jce with highranking Serb politicians and military leaders, such as Slobodan Milošević, Šešelj 'planned, ordered, instigated, committed or otherwise aided and abetted in planning, preparation or execution' of cah and war crimes that included persecution, murder, sexual assaults, torture, deportation and forcible transfer 428 international criminal law review 20 (2020) 405-491 of non-Serbs in Croatia and Bosnia.129 It alleged that Šešelj through his speeches in Vukovar (Croatia), Mali Zvornik and Hrtkovci (Serbia) in 1991 and 1992, 'physically committed' persecution because of 'direct and public ethnic denigration and forcible transfer' .130 The Prosecution did not rely on the totality of Šešelj's propaganda but focused on a particular set of speeches as described. In the speeches in Vukovar, Šešelj stated publicly that 'this entire area will soon be cleared of Ustaša;' and that 'not one Ustaša must leave Vukovar alive' .131 At Zvornik, Šešelj spoke at a rally, stating: 'Dear Chetnik brothers, especially you across the Drina river, you are the bravest ones. We are going to clean Bosnia of pagans and show them a road which will take them to the east, where they belong' .132 At Hrtkovci, Šešelj declared there was 'no room for Croats in Hrtkovci' . He proclaimed that 'we will drive them to the border of Serbian territory and they can walk on from there, if they do not leave before of their own accord' .133 He further stated that Serbs from Hrtkovci and the surrounding villages would 'promptly get rid of the remaining Croats in your village and the surrounding villages' .134 The TC also considered two other speeches, delivered in the Serbian Parliament on 1 and 7 April 1992, alleged by the Prosecution to have constituted clear appeals for the expulsion and forcible transfer of Croats.135 Thus, the Prosecution claimed, for what were essentially the same speech acts, several modes of liability in a bid to establish criminal responsibility under any of them, albeit in vain. Even claims under jce, which in the past proved a valuable tool for prosecuting hate and fear propaganda (see discussion infra section 3.3), were this time thwarted from the beginning as described below.
The actual trial did not begin until late 2007 and finally ended with an acquittal on 31 March 2016 making it the longest running trial in war crimes history. One of the judges sitting at the Šešelj trial, Judge Frederik Harhoff, was disqualified a few months before the Trial Chamber's judgment was expected While considered to be somewhat of a 'lone wolf' in his propagandistic endeavours by both the TC and AC, Šešelj's speeches in fact fit perfectly with the mainstream Serbian propaganda at the time, which almost entirely eliminated the marketplace of ideas.141 While previous judgments at the icty acknowledged the effects of such a combination, the Šešelj judgments were a significant step backward. the 'Strategic Plan') and 'Greater Croatia' .143 The implementation of these two political objectives, according to the icty trial records in the relevant cases, was carried out by criminal means, both military and political.144 Within the framework of the armed conflict in the former Yugoslavia, the grand narratives of 'Greater Serbia' and 'Greater Croatia' were the geopolitical, ideological and military equivalents of the Nazi narrative of 'Greater Germany' .145 Just as the latter employed the sub-narrative of Lebensraum, 'Greater Croatia' and 'Greater Serbia' meant that areas outside of Croatia and Serbia populated with Croats or Serbs should be used to create additional 'living space' .146 Dojčinović thus concludes: 'In as much as Hitler's grand narrative of Lebensraum was directly linked to the Endlösung (Final Solution), the extermination of the Jewish population, the Serbian and Croatian projects were directly linked to the 'ethnic cleansing' of non-Serbs or non-Croats in Croatia and Bosnia and Herzegovina' .147

Trial Chamber's Dismissal of jce − Setting Aside Previous Jurisprudence on Hate Speech in the Context of jce
While at the imt, the status of the relevant grand narrative was merely contextual evidence it shifted at the icty to constitute the forensic evidence and/ or the reasons for the commission of crimes.148 Most of the so-called 'Serb leadership cases' at the icty had the Greater Serbia narrative incorporated into their indictments, pre-trial and final briefs, opening and closing arguments, and judgments.149 Both in Šešelj and Milošević, the otp's theory of the case was largely framed within and by this concept.150 Thousands of exhibits, from speeches to maps, were introduced as part of the cases to evidence and support the significance of this grand narrative for the commission of crimes.151 Numerous witnesses, including combatants and principal perpetrators, confirmed and confessed that they had been inspired and prompted to join the struggle due to the narratives, implying that they had committed specific crimes precisely because of them.152 In their testimonies, they often made references to the exact pro-Greater Serbian phrases they may have read or heard from the accused or the accused's associates' .153 Dojčinović describes phrases such as the one invented by Šešelj, i.e. 'Karlobag-Ogulin-Karlovac-Vitrovitica line' as 'unique mental fingerprints' , i.e. when other politicians or principal perpetrators use such a phrase it can be concluded that a specific mind has been cognitively 'fingerprinted' by the originator.154 Such 'mental fingerprints' should fit into the evidentiary feedback loop as one of the links between the instigator and the instigated.155 In Šešelj, the Prosecution grounded its claim of participation in the jce amongst other things based on the fact that Šešelj 'espoused and encouraged the creation of a homogeneous 'Greater Serbia' , encompassing the territories specified in this indictment, by violence, and thereby participated in war propaganda and incitement of hatred towards non-Serb people' .156 While presenting great hope for an even stronger recognition of the role played by this grand narrative, the Šešelj Trial Judgment instead shifted into the regrettable direction of considering Šešelj's goal of creating a Greater Serbia to have been merely a political project and that the necessary common criminal purpose had not been proven.157 However, as the Prosecution noted, this finding was due to the fact that the Majority did not engage with the evidence about the substance of the Šešelj's Četnik ideology and goals that were predicated on ethnic cleansing and the forced expulsion of non-Serb ethnicities making this criminal element an essential part of the political project.158 While the TC acknowledged that crimes had been committed by Serbian forces in the process, it claimed these were not inherently linked to the fulfilment of the purpose of Greater Serbia. Judge Antonetti's individual opinion reveals the profound confusion of the may have been political, the latter was undoubtedly criminal, i.e. the forcible displacement of Croats and Bosniaks.160 By dismissing Šešelj's liability based on a contribution to a jce the Trial Judgment departed drastically from previous case law involving hate propaganda. Already in its first judgment, in Tadić, the TC clearly identified the importance of hate propaganda in fomenting ethnic discord and noted Šešelj, alongside Milošević and Brđanin, as fuelling conflict and crimes by use of such propaganda.161 Subsequently in Brđanin, the TC considered the accused's hate propaganda as his most substantial contribution to the so-called Strategic Plan.162 The defendant was found responsible for instigating forced transfers and deportations as well as aiding and abetting their execution through inflammatory and discriminatory public statements, such as advocating the dismissal of non-Serbs from employment, and stating that only a few non-Serbs would be permitted to stay on the territory of the Autonomous Region of Krajina.163 The judgment however fell short of framing the criminal responsibility arising from Brđanin's propaganda within the context of a jce. This was mainly due to the TC's understanding of the theory of jce as encompassing only small-scale cases and not enterprises as large as the one claimed in Brđanin164 as well as its view that, in order for the accused to be found responsible for committing a crime under the first category of jce, there should exist an agreement between the accused and the principal perpetrator of that crime (i.e. they should both be members of the jce).165 These understandings of the theory were however found to be erroneous by the AC, which confirmed that principal perpetrators need not be members of the jce166 and that jces can exist on a large scale.167 It was only for reasons of trial fairness that Brđanin's contributed to the commission of the crimes encompassed within through promoting an ideology of ethnic separation, using a rhetoric that amplified historical ethnic grievances and promoting propaganda to that effect, and creating a climate of impunity for criminal acts committed against non-Serbs177 in order to promote historical territorial claims and garner support for the creation of a largely ethnically homogeneous Bosnian Serb state in Bosnia and Herzegovina.178 Interestingly, Karadžić submitted that the TC erred in concluding that he was a member of a jce since the record allowed another inference: that he was part of a 'joint political enterprise' the aim of which was 'political autonomy, not physical separation through forced displacements' .179 His arguments were not accepted on appeal.
The twin goals of the propaganda in all these cases were to create an atmosphere of terror such that the victims would feel their only option was to leave their homes,180 as well as to create a climate where people were prepared to tolerate the commission of crimes and to commit crimes through the incitement of hatred, distrust, strife and fear of subjugation.181 In Stakić, the TC recognised that this atmosphere was of such a coercive nature as to satisfy the 177  condition of 'forced' in the definition of deportation as a crime against humanity.182 While all these defendants also committed other crimes and contributed to the jce in other ways, their propaganda activities have been relied upon as evidence of their knowledge of a common plan and their substantial contributions to the jces both in the first and third categories.183 Dojčinović has bemoaned how the concept of propaganda has been addressed in the judgments of the two ad hoc tribunals. He asserted the reader is 'often left with an intuition that propaganda plays an important role, although it is rarely explained why. The interpretation of the concept of propaganda seems to be taken for granted' .184 However, at a minimum, several judgments translated their sustained condemnation of propaganda and the identification of the crucial role it plays in creating the atmosphere necessary for the commission of the most serious international crimes into individual criminal responsibility through the theory of jce.185 The Trial Judgment in Šešelj however, took a decisive step backward by not acknowledging the defendant's contribution or indeed the existence of a jce in itself. The AC furthermore failed to rectify the matter. While it was of the view that there was a discernible pattern of crimes committed by cooperating Serbian forces, including 'Šešelj's men' in furtherance of a common criminal purpose to permanently forcibly remove a majority of the Croatian, Bosnian Muslim, and other non-Serbian populations, it did not consider that the coordination among them was necessarily in pursuance of this purpose but could have merely resulted from the necessities of the war effort.186 In this way, the effect of Šešelj's hate propaganda as a contribution to the jce was not even considered.

Trial Chamber's Push towards Instigation as the Main Form of Liability − Proving a Substantial Contribution to Subsequent Crimes
Committed Essentially, the Prosecution was compelled to direct its main focus on instigation. This meant that it needed to prove the necessary causal link between the 436 international criminal law review 20 (2020) 405-491 speeches and subsequent crimes committed according to a higher standard than that for the jce, i.e. it had to prove a 'substantial contribution' .
Instead of the totality of his hate and fear propaganda, the judgment analysed and qualified certain statements made by Vojislav Šešelj and their potential effect on the perpetrators of the crimes charged in the Indictment, in the light of their context. It specifically examined the substance and context of a) a speech on the Vukovar road on 7 November 1991 where he allegedly stated that 'Very soon, there will not be a single Ustaša left in this area'; b) a speech in Vukovar between 12 and 13 November 1991 when he is said to have stated that 'Not a single Ustaša must leave Vukovar alive' and 'Ustašas, surrender! There is no need to lay down your lives anymore'; c) a speech in Mali Zvornik in March 1992 in which he called on the Serbs to take revenge and 'clear up' Bosnia from the 'pogani' ( filth) and the 'balijas' (a derogatory term for Muslims), and to push them back towards the east, far beyond the Drina river; d) a speech in Hrtkovci (a village in Vojvodina, Serbia) on 6 May 1992 in which Šešelj called for the expulsion of the Croatian population and where he stated among other things, that there was no place for Croats in Hrtkovci.187 Only the speech in Hrtkovci and the two speeches in the Serbian Parliament were considered by the majority (Judge Antonetti dissenting) as a clear call for the forcible transfer of Croats from the village. For all the other speeches the TC bizarrely concluded that they were not necessarily a call not to spare anyone, but were rather meant to boost the morale of the troops of his camp.188 A particularly unreasonable interpretation, considering the events that followed. Furthermore, even for the speech in Hrtkovci, the TC nevertheless found that instigation could not be established because the Prosecution failed to show that it caused the departure of the Croats or the campaign of persecution alleged by the Prosecution.189 Similarly, two other speeches, delivered in the Serbian Parliament on 1 and 7 April 1992, were found to have constituted clear appeals for the expulsion and forcible transfer of Croats190 however they were considered to have formed part of an opposition to the official Serbian policy, and as such an expression of an alternative political programme that would never have been put into practice.191 Here too, the majority considered that there was no causal link or measurable impact established of the speeches on 187  This judgment reflects the fact that proving the causal link between the utterances of a leader and the subsequent crimes committed by his followers is extremely challenging for reasons of both law and logistics. The Šešelj TC furthermore raised the threshold by considering that it was necessary to prove, in addition, that the Accused had resorted to various forms of persuasion, such as threats, enticement or even promises, before concluding that he was responsible for instigating crimes.193 By doing so, the TC departed from previous case law,194 which established that instigation means 'urging, encouraging, or prompting' another person to commit an offence195 and that the instigator must, in one way or another, have influenced the physical perpetrator by soliciting, encouraging or otherwise inducing him or her to commit the crime.196 The Šešelj Trial Judgment eventually showed that the Chamber considered the contribution requirement not to have been met because the speeches were not 'at the root of the departure of the Croats or the persecution campaign ' .197 This was an additional raising of the threshold for causality to be established and looks like a 'but for' test contrary to previous jurisprudence and sound reasoning.198 For example, in Akayesu the ictr TC ruled that instigation is punishable when the speech is followed by subsequent criminal conduct.199 In the same token, the Blaškić TC found that there must be a causal connection between the instigation and the commission of the crime at issue.200 This does not inevitably signify a proof of 'but for' causation, as according to the Kordić Trial Judgment it is not necessary to prove that the crime would not have been perpetrated without the accused's involvement' what needs to be proved however is 'that the contribution of the accused in fact had an effect on the commission of the crime.201 This contribution has been described as a 'clear contributing factor' in Kvočka202 and in Kordić as a 'substantial contribution' to the conduct of the principal perpetrator. 203 In Ndindabahizi and Mpambara the ictr found that the contribution also needs to be direct. 204 Nevertheless, as determined in the Orić Trial Judgment, the instigator need not be the original author of the idea or plan. Even if the principal perpetrator was already pondering the commission of the crime, but did not yet reach the final determination to do so, such a decision can still be brought about by persuasion or strong encouragement of the instigator.205 Only 'if the principal perpetrator is an 'omnimodo facturus' meaning that he has definitely decided to commit the crime, further encouragement or moral support' cannot constitute instigation, however, it may qualify as aiding and abetting.206 Furthermore, the instigator only needs to possess a certain capability to impress others and does not need to be in a position of superiority in relation to the physical perpetrator. 207 The icc definition of inducing as provided for in Article 25(3)(b) of the icc Statute and as set out in the Ntaganda case confirmed the jurisprudence of the ad hoc tribunals: the following objective and subjective elements must be fulfilled: (a) the person exerts influence over another person to either commit a crime which in fact occurs or is attempted or to perform an act or omission as a result of which a crime is carried out; (b) the inducement has a direct effect on the commission or attempted commission of the crime; and (c) the person is at least aware that the crimes will be committed in the More recently at the icc, in her dissenting opinion to the 2019 Gbagbo and Blé Goudé decision, Judge Carbuccia referred to instigation as a concept including both soliciting and inducing and relied on the Orić Trial Judgment as providing the best definition and the most accurate standard for the 'substantial contribution' test.209 The challenging task of proving a causal link is echoed in other international prosecutions, such as the collapse of three cases against Callixte Mbarushimana, William Samoei Ruto and Joshua Arap Sang at the icc. Mbarushimana was charged under 25(3)(d), but the Court found that there were no reasonable grounds to believe that he had encouraged the troops on the ground and thus there could be no significant contribution on his part to those crimes.210 William Ruto was charged with co-perpetration of cah thus requiring the proof of an essential contribution, and the icc Trial Chamber articulated the requirement for a causal nexus between his speech act and the criminal acts thus: Even if it were accepted that Mr Ruto's speeches contained a sufficiently clear message that he wanted others to engage in conduct that would, in the ordinary course of events, constitute any of the crimes charged, it still has to be established that this message was actually heeded by the physical perpetrators or that his speeches had a direct effect on their behaviour.211 On the other hand, Mr Sang was charged with solicitation or inducement and aiding and abetting. Judge Fremr noted that with regard to solicitation or inducement there had to be evidence that any of the physical perpetrators were influenced by the alleged words even if it would have been established that he called upon listeners to engage in conduct that would, in the ordinary course of events, result in the commission of one or more of the crimes charged.212 It is important to note that the Chamber could not even assess whether the speech amounted to inflammatory or instigating speech as a record of the exact words that were used was not provided by the Prosecution but merely witness testimonies.213 This was particularly problematic since the speech was claimed to be coded and there was great risk that the witnesses provided 'their own, incorrect, interpretation of the obscure wording' .214 The fact that the Prosecution could not obtain such evidence was further seen as a sign that Mr Sang was not as popular and influential as was claimed.215 The lack of sufficient reliable evidence prevented the possibility of establishing any contribution to the commission of the charged crimes and thus also liability under aiding and abetting.216 Sang was further charged under Article 25(3)(d), which requires merely a significant contribution, yet this option was excluded for lack of evidence that would prove the existence of the necessary group sharing a common plan.217 Even in the landmark case for international criminalisation of speech crimes, i.e. Nahimana et al., the AC acquitted Ferdinand Nahimana of instigating genocide in terms of pre-6 April broadcasts due to what they considered to be at the very least a tenuous link between the speeches and specific acts of genocide considering the time lapse between the speeches and the killings.218 The blame for this rests also with the Prosecution as they did not bring a single perpetrator to testify that they had been influenced by the defendants. Most recently in the case of Gbagbo and Blé Goudé, the majority of the Trial Chamber held that there was 'no case to answer' including on charges in relation to the contribution of the accused's speech acts to post-election violence.
One of the biggest challenges in proving a causal link to subsequent crimes is bringing witnesses as evidence. Calling as witnesses the very people who had been influenced by the incitement and subsequently committed crimes or who knew others that did, would open the doors for questioning their 212  credibility.219 In the Muvunyi case before the ictr,220 the credibility of a witness was doubted, based on their being an Interahamwe militiaman who may have been motivated to exaggerate the defendant's role in acts of genocide in order to minimise his own role. More importantly, the phenomenon of recanting by prosecution witnesses, although generally common in international criminal proceedings, is especially endemic in speech crimes cases, in part because the accused are often high-level politicians with continued government backing and a zealous following and witnesses are often subjected to bribery and intimidation.221 Another problem is that international criminal tribunals do not have the leverage of encouraging testimony from subordinates against superiors in exchange for reduced sentences, such as is the practice in the United States.
It is important to note that during the second icty investigation into Šešelj's propaganda, which was led by the German prosecutor Hildegard Uertz Retzlaff, a large number of potential fact witnesses were interviewed, including many different types of insiders.222 Some of the witnesses later retracted their statements to the otp under threats by Šešelj's men. However, apart from them, there was a dozen or so paramilitary troops, mainly members of Šešelj's forces who gave statements corroborating direct influence on their crimes as charged in the indictment.223 As Dojčinović notes, '[t]he proof of direct instigation in these statements was truly overwhelming' .224 Unfortunately, the prosecution teams and their strategies drastically changed during trial and with a reliance on jce the view was eventually adopted that instigation or a causal link with crimes committed did not need to be proven.225 The statements highly regrettably thus ended up not being used. They now sit in the otp archives and due to their confidentiality cannot be disclosed to the public.226 In terms of instigation the AC corrected the TC judgment in several ways. Firstly, it made clear that it was not necessary for the instigator to have used different forms of persuasion such as threats, enticement or promises to the  physical perpetrators of the crimes.227 Secondly, it also clarified that it was not necessary to establish that the crimes would not have been committed without the instigator's involvement but merely that his utterances were a substantial factor contributing to the conduct of the perpetrators.228 For Šešelj's speech in Hrtkovci, the AC found him responsible for instigating deportation, persecution (forcible displacement) and other inhumane acts (forcible transfer) as cah. This in turn was based on the AC finding that the speech substantially contributed to the conduct of the perpetrators in light of a) Šešelj's influence over the crowd; and b) the striking parallels between his inflammatory words and the acts subsequently perpetrated by, inter alia, members of the audience.229 In a positive twist, the AC thus provided two additional criteria for trial chambers to take into consideration when determining the link between speech and subsequent crimes that are more than mere chronology. As far as the mens rea is concerned, the AC noted that Šešelj intended to prompt the commission of the crimes or, at the very least was aware of the substantial likelihood that they would be committed in execution of his instigation. 230 Regrettably, the AC dismissed the Prosecution's appeal regarding all other speeches, that is in the towns of Mali Zvornik and Vukovar, his statements before the Serbian Parliament, and other statements encouraging the creation of a Greater Serbia, claiming that the evidence presented by the Prosecution was insufficient to discern any specific impact on the commission of crimes.231 Considering his speech at Mali Zvornik, the AC found that a substantial contribution to the commission of the relevant crimes there did not exist. The AC acknowledged that Šešelj's statements that 'rivers of blood' would follow a Bosnian declaration of independence, were clearly inflammatory when viewed in the context of the events that were unfolding at Zvornik at the time, including the detention, torture and murder of Muslim civilians by several factions of the Serbian forces.232 The AC further acknowledged that Šešelj exerted influence over members of his party and some combatants.233 It found that the speech in Mali Zvornik where Šešelj implored the Četniks to clear up Bosnia from the 'pagans' and show them the road to the east where they belonged, were a call for ethnic cleansing, which could have prompted other persons to commit 227  crimes against non-Serbian civilians.234 Despite this, however, the AC did not find him responsible for instigating the crimes charged in the Indictment as the evidence was insufficient to discern whether the impact on the commission of the crimes amounted to a substantial contribution.235

3.5
Aiding and Abetting With regard to the alleged responsibility of Vojislav Šešelj on the count of aiding and abetting, the Chamber, by majority, Judge Lattanzi dissenting, rejected it, noting that the underlying allegations of the Prosecution have the same factual basis as its allegations on his responsibility under the jce and instigation. The TC held that the defendant did not show sufficient criminal intent, and it was not proven that his speeches calling for expulsions and forcible transfers had had a substantial effect on the perpetration of war crimes.236 As mentioned above, on paper, aiding and abetting requires the same threshold for causality as instigation or ordering and planning, i.e. a 'substantial' effect or contribution, yet the jurisprudence has shown more flexibility in this regard for aiding and abetting.237 For example, in Simić, the AC stated: 'It is not required that a cause-effect relationship between the conduct of the aider and abettor and the commission of the crime be shown, or that such conduct served as a condition precedent to the commission of the crime' .238 Whether the Court will consider acts as aiding and abetting without a causal link to the offense in question, will depend on the nature of the contribution and the pattern of facts in the case.239 Furthermore there is no specific threshold for the level of assistance required to establish liability for aiding and abetting other than that it 'must have furthered, advanced or facilitated the commission'240 of the offense.241 In Charles Taylor it was also confirmed by the AC of the Special Court for Sierra Leone (scsl) that each individual act does not have to exert a ' substantial 234  international criminal law review 20 (2020) 405-491 effect' on the outcome to qualify as aiding and abetting, but rather '[t]he facts of a case may involve multiple acts or conduct which, considered cumulatively, can be found to substantially contribute to the crime charged' .242 Yet despite the seemingly lower threshold for causation found in the international jurisprudence, Šešelj's acquittal for aiding and abetting cah was upheld by the AC, on the grounds that Šešelj did not show the requisite awareness that crimes were being committed and his speech acts did not rise to the requisite level of making a substantial contribution to the commission of crimes.243 The decisions of both chambers thus displayed a departure from previous jurisprudence against a just ruling and again in favour of acquittal.244

3.6
Committing Persecution through 'Incitement to Hatred' It is not however solely under instigation, aiding and abetting or the doctrine of jce that liability can be attributed to hate propagandists. In Šešelj, the Prosecution alleged physical commission of persecution as a cah through the accused's speeches in Hrtkovci, Vukovar and Mali Zvornik based on the violation of the right to dignity and security of the targeted groups. The Prosecution strategy was heavily influenced by the Nahimana et al. case where the possibility of such liability was most notably confirmed when taking place in a context of a widespread and systematic attack against a civilian population.
While this position has been criticised as 'essentially ignoring the decision of icc States parties not to criminalise incitement to war crimes and cah as such'245 it addresses an apparent gap in icl in terms of war propaganda. Furthermore, it would not be entirely accurate to equate incitement to war crimes or cah with incitement to hatred as persecution per se. The latter does not require a causal link to any subsequent crimes. However, it is nevertheless defined also in terms of impact and is thus not an inchoate crime as detailed below. 246 As an inchoate crime, it is a wrong in itself (a 'malum in se') and the inciter is being held criminally responsible for his own actions, not those of the incitee. 247 Notably, however in his partly dissenting opinion in the Nahimana case, Judge Meron clearly disagreed with the majority, insisting that mere hate speech could not be criminal. 248 Similarly in the Trial Judgment in Šešelj the majority dismissed the notion that incitement to hatred constituted a crime in international law.249 Also against previous icty jurisprudence,250 the TC in Šešelj furthermore denied the existence of a widespread and systematic attack against the non-Serbian population that meant that the consideration of Šešelj's hate speech as a cah per se was a priori impossible. Erroneously, in para. 192 of the judgment the TC referred to the element of 'widespread or systematic' cah as conjunctive rather than disjunctive, raising the bar further and contradicting the existing law.
In general, the crime of persecution consists of an act or omission that discriminates in fact against protected groups in the context of a widespread or systematic attack on a civilian population. What is required is not merely the intent to discriminate but 'the act or omission must have discriminatory consequences' . 251 The persecutory act or omission must deny -or infringe upona fundamental human right laid down in international (customary or treaty) law.252 There must be a 'gross or blatant denial' of a fundamental right. 253 Article 7(2) (g) of the Rome Statute states that the act of persecution must result in an 'intentional and severe deprivation of fundamental rights contrary to 446 international criminal law review 20 (2020) 405-491 international law' . The aim of such a deprivation of an individual's rights is 'the removal of those persons from the society in which they live alongside the perpetrators, or eventually even from humanity itself' .254 An act is discriminatory when a victim is targeted because of his or her membership in a group defined by the perpetrator on a political, racial or religious basis.255 It is thus the perpetrator's subjective perception of the group or community and who belongs to it, which is determinative and not some objectifiable characteristics that connect the group.256 A number of accused have been found guilty of persecution by denigration through speech. In Prosecutor v. Ruggiu the ictr found that hate speech, even without proof of causally related violence could be the basis for charging persecution as cah. The Trial Judgment in that case characterised the acts of persecution as consisting of: radio broadcasts all aimed at singling out and attacking the Tutsi ethnic group and Belgians on discriminatory grounds, by depriving them of the fundamental rights to life, liberty and basic humanity enjoyed by members of wider society. The deprivation of these rights can be said to have as its aim the death and removal of those persons from the society in which they live alongside the perpetrators, or eventually even from humanity itself.257 As Gordon puts it, the Chamber thus considered that the words themselves assaulted the victims with the consequence of denying them their fundamental rights and were not merely a means through which to inspire others to commit violent acts.258 The TC in Nahimana expressly recognised that such speech 'is not a provocation to cause harm. It is itself the harm. Accordingly, there need not be a call to action in communications that constitute persecution' .259 This approach is reflected also in the Supreme Court of Canada in Mugesera.260 Until the Šešelj Appeal, the jurisprudence at the icty was, however, the opposite of the above. The first time the icty dealt with incitement to hatred as a crime within the subject matter jurisdiction of the Tribunal was in the Kordić & Čerkez case in which the indictment included 'encouraging and promoting hatred on political and other grounds' as a cah.261 However the TC did not find this act to constitute a cah due to the fact that such incitement was not enumerated as a crime in the Statute and according to the Chamber, its prohibition had not reached customary international law either.262 Furthermore it considered the gravity of such an act not to rise to the level of other acts enumerated in Article 5 of the icty Statute.263 All three justifications were dismissed in the Nahimana appeal judgment, which noted the consistent AC jurisprudence by which it does not matter whether a specific act of persecution was criminalised or whether its gravity amounted to other cah as the underlying acts of persecution could be considered together.264 The ictr in the Nahimana appeals judgment set a two-step test to determine whether incitement to hatred amounts to persecution. Firstly, the hate speech has to involve the denial of a fundamental right and discriminate in fact and secondly, the violation of these rights needs to be of equal gravity to the other cah. Where the hate speech targets a population based on ethnicity, or any other discriminatory grounds, it violates the right to respect for the human dignity of the members of the group, whereas where the speech constituted a call for violence against a population on such grounds, it violated the group members' right to security.265 Both violations constitute discrimination in fact.266 According to the Chamber, speech itself cannot however constitute a violation of the right to life, freedom or physical integrity, without a third person's interference. 267  Where the hate speech did not incite to violence, the ictr determined that it is not necessary to decide whether it was in itself of a gravity equivalent to the other cah. This is because, as mentioned, it is only necessary for the cumulative effect of all the underlying persecutory acts to be of equal gravity to the other cah and not for each underlying act.269 The Chamber therefore considered that hate speeches and calls for violence uttered after the beginning of a systematic and widespread attack against a population themselves constitute underlying acts of persecution, and when they substantially contribute to the commission of other acts of persecution they further constitute instigation to such crimes.270 After 6 April 1994, in Rwanda, the context was that of calls for genocide and a massive persecutory campaign against the Tutsis, therefore the gravity threshold was clearly met.271 The Šešelj Trial Judgment however dismissed hate speech as persecution per se entirely. In terms of the speech in Hrtkovci it stated that the mere use of insulting or defamatory language was insufficient to amount to persecution.272 On the other hand, in her dissenting opinion, Judge Lattanzi considered that at least the speech in Hrtkovci constituted a physical commission of the denigration of the non-Serb populations, particularly Croats, as an underlying act of persecution.273 The AC confirmed this opinion and considered that said speech denigrated the Croatians of that town on the basis of their ethnicity, in violation of their right to respect for dignity as human beings.274 Furthermore it violated the right to security of the Croatians since Šešelj's instigation of their forcible expulsion incited violence that denigrated and violated this right.275 Based on this the speech in Hrtkovci constituted commission of persecution as a crime against humanity. This finding of commission of persecution as cah was thus partly based on initially finding Šešelj responsible for instigating deportation, persecution (forcible displacement) and other inhumane acts (forcible transfer) as cah. Other speeches that violated the right to dignity and the right to security of targeted groups went unrecognised as acts of persecution.
One of the most contentious issues when it comes to hate speech as persecution per se is the question whether it can amount to said crime without an  actual incitement to violence. The Nahimana Trial Judgment concluded that calls to violence were not necessary and in Ruggiu276 the Tribunal relied on the Streicher judgment in its reference to 'the poison that Streicher injected into the mind of thousands of Germans' .277 Hate speech as persecution per se was thus considered as notably wider than instigation or direct and public incitement to include advocacy of ethnic hatred in other forms.278 Thus in the Nahimana Trial Judgment examples such as describing Tutsis as cunning and tricky or the portrayal of Tutsi women as femmes fatales or seductive agents of the enemy were all considered to constitute persecution.279 In support of this approach Gordon states: speech in service of widespread and systematic attacks directed against civilian populations… not calling directly for action should nevertheless be criminalized. It cannot wrap itself in the mantle of democracy-pro motion, self-governance, diversity-enhancement, community-building, or individual-empowerment in light of its link to a pervasive or well-organized attack on ordinary citizens in a community.280 On the contrary and mostly influenced by the American theory on the matter, those opposing such a wide definition have relied on the fact that Streicher was only convicted because of incitement to murder and they furthermore consider the acquittal of Fritzsche on grounds that his hate speeches did not seek 'to incite the Germans to commit atrocities against the conquered people' .281 They furthermore understand the Ruggiu Trial Judgment as showing that it is only speech whose ultimate aim is to destroy life that constitutes persecution282 and criticise the TC for having failed to follow the Čerkez and Kordić Trial Judgment, which had found that mere hate speech could not constitute persecution.283 The Nahimana Appeal Judgment refrained from settling the matter in its entirety, while the partially dissenting opinions were split between Judge Meron and Judges Shahabudeen and Pocar. Relying on Kordić and US jurisprudence, Meron simply dismissed the possibility of hate speech ever forming a basis for criminal conviction save for incitement to violence.284 On the contrary, Judges Shahabudeen and Pocar were of the opinion that the Appeal judgment should have made it clear that speech not inciting violence can constitute an underlying act of persecution.285 Pocar felt that in the case at hand the circumstances were a perfect example of where hate speech fulfils the conditions necessary for it to be considered as an underlying act of persecution and that the relevant speeches taken together and in their context amounted to a violation of sufficient gravity.286 Furthermore Shahabuddeen claimed that in the case of Fritzsche his acquittal was based on the fact that he did not form part in originating or formulating the propaganda campaign and not on the fact that he did not call to violence.287 Citing furthermore the decisions in Flick and Einsatzgruppen he interpreted Nuremberg jurisprudence as not necessitating incitement to violence for persecution as a cah but rather 'harm to life and liberty' or 'acts committed in the course of wholesale and systematic violation of life and liberty' that may include economic and political discrimination and not necessarily a physical attack. In his argument: '[T]he court may well regard the 'cumulative effect' of harassment, humiliation, and psychological abuse as impairing the quality of 'life' , if not 'liberty' , within the meaning of the tests laid down in Einsatzgruppen' .288 Shahabuddeen also criticised the Kordić judgment, which he claimed contradicted the Ministries case, other cases of the icty and customary international law by excluding the hate act from its context: it is not possible fully to present a campaign as persecutory if integral allegations of hate acts are excluded… The subject of indictment is the persecutory campaign, not the particular hate act. This was why noncrimes were included with crimes in the Ministries case.289 Thus, despite the strong dissents, the AC in the Media case left the issue unresolved.290 The Šešelj Appeal did not settle the matter either, as it does not give a definitive answer but rather implies that a call to violence is necessary. There is scope to understand the mention of a violation of dignity in the Šešelj Appeal Judgment as allowing for a finding of persecution without a call for violence, yet there are two issues that arise in this context. Firstly, as mentioned, the Appeal Judgment does not mention the violation of dignity in the actual disposition, and secondly, it is not clear what is meant by 'dignity' .291 As David Weisstub theorises, human dignity can be violated through disrespect whereby a person is 'reduce [d] […] to behave in a less than dignified manner' .292 As Timmermann writes, it is also violated when a person is made to lose 'selfrespect or the feeling of self-assurance that results from social recognition that individuals experience when they are respected as equal members of the community' .293 Margalit explains that humiliation 'shows the victims that they lack even the most miniscule degree of control over their fate -that they are helpless and subject to the good will […] of their tormentors' .294 These definitions would seem to imply a policy of persecution beyond mere words to include instead a significant degree of coercion. In the context of speech, they thus suggest words that carry at least an implied threat of physical force or violence. This would seem to loosely correspond with the Appeal Judgment since it did not find a violation of dignity in Šešelj's other speeches despite clearly using severely derogatory terms for the out-groups. 289  Limiting the scope of persecution to hate speech inciting to violence leaves essential parts of any successful hate propaganda campaign out of criminalisation. In this sense it mirrors the situation with incitement to genocide, where only words that directly call for genocide fall within the definition of the crime, whereas all other propaganda surrounding such calls can merely provide context for the purpose of determining the intent of the speaker or the directness of his calls to genocide.295 In her dissenting opinion in Gbagbo and Blé Goudé, Judge Carbuccia referred to Féret v. Belgium to highlight that a politician's comments constituted public incitement to racial hatred against outsiders, and without requiring a call to this or that act of violence or another delinquent act, violated the dignity and security of the effected groups of people, posing a danger for social peace and the political stability of democratic states.296 This said, the Judge did not clarify what implications she wanted to infer from this for the purpose of interpreting the icc Statute on such incitement.
If there is no need for a direct call to violence or causality between the uttered speech and any subsequent atrocities committed, proving guilt of hate speech as persecution per se is inherently easier. However, it nevertheless does not solve the issue of the potential preventative function of hate speech as a crime in itself. This is because hate speech as persecution per se still demands the contemporaneity of the speech with a widespread or systematic attack directed against a civilian population, since such contemporaneity has always been a part of the definition of cah. The AC in Nahimana thus insisted that the hate speech in question be contemporaneous with the atrocities committed against the Tutsis after 6 April 1994. This leaves out all the propaganda leading up to the genocide, which was essential for the success of the actual incitement to genocide. Judge Pocar drew the distinction between hate speeches prior to 6 April 1994, that is, prior to the beginning of the systematic and widespread attack against the Tutsi ethnic group, and hate speeches that occurred after that date. In his opinion the later speeches were per se underlying acts of persecution,297 whereas the former ones could be considered instigation.298

Law Not Fit to Deal with War Propagandists
As it stands, the sort of hate and fear propaganda that was employed in the former Yugoslavia is not criminalised in its complex entirety. Unless it is a matter of direct incitement to genocide, large chunks of speech employing persuasion techniques that prime a society for the commission of atrocities remain outside the ambit of criminalisation. The consideration of the totality of one's hate propaganda and the context in which it is uttered is crucial for recognising the role it plays. This section will take a multi-disciplinary approach to identifying the reasons behind the failure of international criminal law to adequately recognise the effects of propaganda and effectively address it. It will furthermore highlight the necessary shifts in legal perception of this phenomenon that would allow more meaningful conclusions in the jurisprudence in line with the findings of social sciences experts. Predrag Dojčinović noted already before the decisions in Šešelj, that there was a conceptual vacuum in the interaction of propaganda and international law.299 In his words: 'Propaganda is used in a manner which leaves the impression of an intellectual embellishment, a prosthesis of deliberative thinking, rather than a concept which is first clearly understood and only then situated in law' .300 He further notes that after reading judgments from the icty and ictr 'we are often left with an intuition that propaganda plays an important role in the text, although it is rarely explained why' .301 Richard Wilson goes a step further and claims that while international tribunals use forceful declarations regarding the deleterious effects of propaganda, they are seldom supported by extensive and reliable evidence 'and it could be argued that the chasm between the claims regarding propaganda's consequences and the evidence presented to support such claims is wider than in any other area of  international criminal law' .302 In their view, a closer insight into its mechanism would contribute much required precision and logical coherence to the analysis in the judgments. 303 Wilson warns against relying on common sense reasoning when trying to construct causal models of speech at international tribunals.304 In his words, 'our ability to understand the effects of speech through intuition and a priori reasoning alone is actually quite limited. Many recent findings in persuasion research and communication theory are counterintuitive, or at least not immediately obvious, even to seasoned political and legal observers' .305 While regretting the lack of social research expertise during trials of hate propagandists, Wilson at the same time notes that not even such research can determine with absolute certainty the concrete effects of a particular speech act uttered in the context of widespread and systematic violations, however it can identify which types of speech acts are most likely to elevate the risk of violence in a particular context. 306 The above indicates that proving a causal link between a speech and an act from a member of the audience beyond reasonable doubt may be a near impossible task, particularly since, as Wilson notes, there is no guidance on how to understand such a causal link in icl jurisprudence or scholarly works.307 Aksenova observes the intrinsic problem in relying on causation in the strict sense of the word in the context of speech acts. In the physical world, a causal chain is considered broken when interrupted by the free choice of an individual, even one acting under an accomplice's influence.308 In this sense it cannot be said that an accomplice should be understood as a necessary prerequisite for harm caused by the primary perpetrator; rather he merely affects the freely chosen behaviour of the principal perpetrator. merely chance-raising.310 On the other hand, many have questioned how much 'free will' there can be in a propagandistically determined context.311 Regardless of the theory, a more appropriate approach than searching for causation in the strict sense would be to establish that a certain speech is or was likely to elevate the risk of violence in a particular context including the cultural, political and historical context. This way the inquiry could be done through an objective expert evaluation of outwardly manifestations and would not necessitate the testimony of principal perpetrators as to the effects the propaganda had on them. As Joseph Jaconelli notes generally on incitement, it may prove impossible to ascertain the number, location or identity of the incited and how incited they were.312 In fact, in the present context, it would require bringing to the court perpetrators who: a) have a perfect recollection of their own thought process behind the commission of the crimes; b) have the willingness to honestly speak about it; and c) have an extraordinary ability of self-analysis to the extent that they can determine with surgical precision how far their thought process at the time was 'their own' and to what degree it was influenced by the subconscious effects of propaganda or the general context created or permitted by other individuals under its influence.
Considering the potential risk, instead of strict causation, would mirror the approach taken by human rights courts in assessing the legitimacy of limiting freedom of expression in the sense that it would consider the dangerousness of the speech.313 It is important to note that the cases in question at international criminal tribunals generally do not consider the legitimacy of a restriction of the freedom of speech of a minority group versus the national security considerations of the government as is the situation at the human rights courts. Rather it is the opposite: the speech under consideration is that of the so-called 'majority population' , in support of the government and as a supposed defence of national security.314 Thus the standards that have evolved in international law largely to deal with the former situations required some adaptation in the sense of lowering protections, 'so that ethnically specific expression would be more rather than less carefully scrutinized to ensure that minorities without equal means of defence are not endangered' .315 A broader understanding of causation in propaganda cases would thus allow for more effective prosecution of hate and fear propagandists, even if a substantial reform of the law in terms of expanding liability remains unattainable.
Whether propaganda can lead to atrocities depends on the context more so than particular words uttered. Several studies have confirmed that ordinary people are capable of committing atrocities in the name of nationalist imperatives, religious principles and righteous ideology.316 When it comes to ethnic collective violence, Oberschall analyses three instances spanning three different continents and concludes that this occurs under the following conditions: when there are contentious issues between [the groups], when their political leaders advocate hostility and aggression in ethnic conflict management, and when threat and hate messages in the mass media amplify danger to the group, incite hostility to adversaries, and justify collective violence against them as a solution to outstanding issues.317 As Gordon suggests, of all the elements to consider in incitement, context is the most important.318 The Hartford Guidelines advance a risk assessment framework informed by the latest social science research that has identified many of the key ingredients of mass persuasion.319 They provide a checklist of indicative factors known to elevate the risk of speech acts prompting mass crimes, including: (a) the political context of the speech act, (b) the emotional state of the audience, (c) the historical and political context of the country, (d) the perceived charisma, credibility and authority of the speaker, (e) their use of graphic and dehumanising language, (f) the degree to which they summon up cultural symbols and cultivate historical grievances, (g) their calls for revenge, 315   While the TC in Šešelj acknowledged its duty to analyse and qualify Šešelj's statements and their potential impact in light of the cultural, historical and political context,321 it essentially considered it as a sort of mitigating circumstance. The Chamber construed the 'context' in such way that any effects and any responsibility that went beyond the concrete situation were dismissed. As Posselt describes, hate speech and hate crime necessarily have a historical and iterative character: For every performative requires for its functioning that it can be reiterated, that it evokes and affirms the historical sedimented norms, structures and conventions that ultimately enable it. This becomes particularly apparent, when we recall that the one who performs hate speech or commits hate crime never acts 'alone' , but as a representative of social structures, attitudes and dispositions that his or her (speech) acts invoke and actualize. Hence, hate speech and hate crime are never simply expressions of an individual, but always also a reflection and reiteration of 'inbuilt tendencies and predispositions of societal structures that make those acts possible. 322 Posselt continues however, that this should not be an excuse for diminished but rather a strengthened concept of responsibility. Quoting Butler, he concludes that responsibility is not linked with speech as origination, but precisely with its citational and repetitive character, i.e. its invocation and reiteration of norms, conventions, traumas, and exclusions that are sedimented in language with which the speaker renews the linguistic tokens of a 'community' , reissuing and reinvigorating such speech.323

A Broader Approach to Causality Considering the Context of the Speech
In her dissenting opinion in Šešelj, Judge Lattanzi claimed there was ample admitted evidence to establish instigation, which if taken into account in its entirety, analysed meticulously and in accordance with previous ICTY/ICTR jurisprudence, would lead to the only reasonable conclusion that instigation did take place as well as material perpetration of the crime of persecution through the underlying act of denigration.324 Her assessment is correct despite the fact that a vast amount of evidence remained unused as previously mentioned.325 What evidence was available was simply not factually analysed.326 Dojčinović points out for example how the word 'Ustaša' referred to by Šešelj, had already been recognised as one of the strongest derogatory hate speech references to Croats in the icty jurisprudence.327 Yet the TC did not even mention it, so the AC could not have taken it into account.328 Unfortunately, according to Lattanzi, in the context of instigation the Chamber accorded no weight at all to the evidence showing that the Accused exerted a great influence on his followers and the srs volunteers involved in the crimes… [and] failed to take into account important elements such as the means used by the Accused to influence the behaviour of the perpetrators of crimes, the repetition of the same incriminating discourse over time, the general background of the disintegration of the former Yugoslavia and the extreme inter-ethnic tensions against which these acts took place, and the fact that the existing situation worsened following the speeches in dispute.329 Judge Lattanzi, as well as the Prosecution's Appeal Brief, faulted the TC for not having taken into account more of Šešelj's statements in order to deliberate on his instigation liability; however, it is important to note that while there were 324   at least over five hundred exhibits presented to address the mens rea, the indictment only mentioned five speeches as the specific crime base.330 As Wilson notes, 'a full comprehension of the ensemble of conditions that were jointly sufficient [for crimes to take place] is often lost as prosecutors and defense counsel seize upon opposite poles of the same continuum of causation' .331 In the same way, the Prosecution in Šešelj sought to show how those five speeches were each 'directly causal and set in motion criminal actions' .332 Yet such an approach misses the most important findings of recent social science research, i.e. that certain types of speech, while not attaining a sine qua non threshold, nevertheless elevate the probability conditions that a criminal harm or injury will ensue333 for example by lowering empathy towards an out-group,334 heightening fear, disgust,335 anger336 and an indifference to their well-being337 as well as morally justifying particularly brutal violence against its members.338 While some find it 'problematic to suggest speeches pleaded and not pleaded in the indictment, 'taken together' , satisfied the requirements of instigation' , this is precisely what should have been done in the Šešelj case, for a more appropriate consideration of the effects of his propaganda.339 It is rather shocking that among all of his speeches and writings, only one, that is the 460 international criminal law review 20 (2020) 405-491 speech in Hrtkovci, was eventually found on Appeal to be criminally relevant. In other words, had Šešelj not given that particular speech, he would have still been considered innocent of any international crime. In view of the entirety of his propaganda portfolio, this would undoubtedly have been a grave injustice and would send a message that hate propaganda will go unpunished unless a particular speech can be linked with overwhelming evidence of causality to a specific crime committed as a result.
In the Gbagbo and Blé Goudé case at the icc, the Prosecutor brought a much larger sample of potentially relevant speech to the Court and pursued a holistic approach to proving causation taking into consideration the combined effects of the words and actions.340 In delivering his reasons for joining Judge Tarfusser in deciding to end the case against Mr Laurent Gbagbo and Mr Charles Blé Goudé, Judge Henderson acknowledged that the Prosecutor was 'correct in emphasising the importance of evaluating all the evidence together and to take into consideration the combined effect' of the accused's actions and words.341 However, he was unconvinced by the implications drawn therefrom and essentially expected evidence proving straightforward causation.
The Prosecution claimed, for example, that Mr Blé Goudé's words led to violence by creating a climate of fear with threats of genocide posed by France, the unoci and the supporters of the opposition, who he branded as the 'rebels' .342 According to the Prosecutor, this primed the youth to be ready for a 'mot d'ordre' to be issued by him in response to this threat.343 The latter referred to a speech in which Mr Blé Goudé instructed the audience to (a) prevent unoci from moving in the neighbourhoods, and (b) to coordinate with the 'présidents de quartier' in order to be aware of and verify the comings and goings in their neighbourhoods and to 'denounce' strangers entering their neighbourhoods.344 While the said 'mot d'ordre' did not include calls to violence, according to the Prosecutor explicit language did not need to be used for the message to be understood after the audience was primed with fear and in the context of preexisting tension between the relevant neighbourhoods that made it foreseeable that inflammatory words stigmatising 'foreigners' would lead to violence.345 Henderson rejected such a general presentation of a causal link and stated that while the speech contained potentially inflammatory passages, which, in a volatile context, may well have caused some individuals to act in a manner going beyond what the words actually called for, such an effect must be established through evidence and cannot simply be assumed.346 Again the best evidence in this regard was considered to be examples of those who attended the speech and subsequently committed crimes and who were demonstrably influenced by the words.347 What exactly is meant by 'demonstrably influenced' is, however, unclear.
Contrary to this approach, in her dissenting opinion Judge Carbuccia considered that a reasonable Trial Chamber could find, from the evidence, that Blé Goudé's speeches, 'mots d'ordré' and calls during the post-election violence, considered as a whole, and in the light of the particular social and political context of that post-election violence, contributed substantially to the acts of violence committed.348 While direct perpetrators testifying to the effects that certain speeches had on them may be the ideal scenario for judges, it should not a be a necessary requirement for the Prosecution for the reasons described above. When it comes to causality, it is furthermore important to acknowledge that the link not only runs directly from the words uttered to the perpetrators in the form of persuasion or other direct effects on their cognition. Rather, there exists an important indirect causal link through intermediaries as well as through a signalling of a new societal order and morality which individuals either feel compelled to join or are willing to join voluntarily for personal reasons. In other words, most perpetrators do not need to be persuaded into the ideology being spread in order for propaganda to have an effect on them and thereby contribute substantially to the commission of the crimes.

4.2
The Context of a Severely Skewed Marketplace of Ideas It is not merely the totality of an individual's propaganda that should be taken into account, but rather the totality of all state and non-state propaganda with essentially the same messaging and its constant repetition that surrounded the speech of the accused and that makes such speech more effective. As was the case with Streicher, Šešelj was seen as a lone wolf. Yet the mere fact that their propaganda campaigns took place in the context of a complete breakdown of the marketplace of ideas, where all opposing views were strictly marginalised and to a large extent eliminated, invalidates the idea that their acts could not be considered as part of the broader campaigns and were mutually reinforcing. 346 Ibid., para. 1994 The US model for freedom of speech, which arguably allows the fewest exceptions to it, includes the important precondition that the state remains neutral and protects public discourse as a sphere that remains equally open to all communities.349 Šešelj's words were not uttered in a vacuum or in the context of a rich marketplace of ideas. Rather they were spread in an environment in which the media and, in particular, national television, laid the necessary emotional foundations for conducting a wartime propaganda campaign by building up hateful stereotypes of Croats and Muslims.350 There was generally no marketplace of ideas during the Yugoslav wars in the 1990s as each side controlled its media to give its public a distorted version of reality and cut off broadcasts from the other side. 351 Renaud de la Brosse presented an expert report detailing the methods employed by those controlling the media during the Yugoslav wars and the use of media for ultra-nationalist goals.352 In 1986 Slobodan Milošević began his efforts to gain control over the state media until by the summer of 1991 he held a monopoly over Serbian state media, albeit that there was a growing number of niche independent media created by liberal, progressive and independent journalists. Milošević and his regime waged an intense media battle of hateful propaganda and biased and untrue information. According to Eric Gordy, Under the pressure of the political authorities, the media had to attempt to convince the citizens of Serbia that they were the victims of an international conspiracy to eliminate them and remove Serbia from the face of the earth. The fact that the media and, in particular, national television built up such hateful stereotypes incontestably laid the necessary emotional foundations for conducting a wartime propaganda campaign.353 Milošević 'personally appointed editors-in-chief of the newspapers and news programs, especially directors-general of the radio and television' and was 'in direct communication with all editors who 'fed' the public with the news, comments and generally with information' .354 Milošević-controlled media reached more than 3.5 million people every day, while he used several ways of controlling independent media to ensure that their reach was minimal. It is important to note that due to significant demonstrations against regime control of the state media, the largest of which were in Belgrade in March of 1991, the regime left some room for independent media to survive albeit with limited means, circulation, viewers and listeners and subject to constant harassment and attacks.355 A Special Report by the Belgrade Institute of Political Studies found that in 1994 'the Milošević regime controlled 90% mass media penetration, i.e. 90% of information on public affairs and news reaching the public was through regime media' .356 By 1995, the independent media were largely restricted to Belgrade, thus the majority of Serbs only received regime propaganda.357 As one of Milošević's closest collaborators explains: 'He was deeply convinced that citizens formed their view of the political situation on the basis of what they were presented and not on the basis of their real material and political position. What is not published has not happened at all -that was Milošević's motto' . 358 During a strike at the cultural and music service of Radio Belgrade that broke out in protest of political control, Šešelj himself held a press conference to disclose a list of radio and TV journalists he wanted eliminated because of their lack of patriotism and obedience to political masters.359 Most were subsequently put on 'compulsory leave' or marginalised360 while other purges followed, eventually getting rid of 200 journalists and thousands of staff. 361 In the Serb majority territories of Bosnia, Oberschall describes the control of mass media as physical and coercive.362 Eight months before the start of war in Bosnia-Herzegovina, Serb paramilitaries supported by the Yugoslav army seized the Kozara mountain transmitter, followed by other such seizures in 464 international criminal law review 20 (2020) 405-491 areas with Bosnian Serb populations in order to cut off Sarajevo TV signals and ensure exposure to Belgrade TV only. 363 Tadeusz Mazowiecki, Special Rapporteur appointed by the UN Human Rights Commission described how the media of the former Yugoslavia was mainly spreading nationalist discourse and generalised attacks and insults aimed at the other peoples.364 He added that 'it comes as no surprise that the phenomenon has directly led to the perpetration of horrible atrocities on the fields of battle and throughout the entire territory' . 365 In understanding the diminished marketplace of ideas, it is also essential to note that in Yugoslavia, the national question overshadowed all other issues after the end of communism and all Serb political parties, including the opposition, subscribed to the national crisis frame, though less strident and more nuanced than Milošević's sps or Šešelj's srs. 366 Genocide and mutatis mutandis other atrocities against a targeted group, arise from a pattern, or gestalt, rather than from any single source.367 A measure of the effectiveness of a defendant's use of propaganda is how their forms of speech shaped the language of the time. The question to ask is whether the accused's vocabulary and forms of expression became part of the general public's universe after the propaganda campaign began. 368 In the former Yugoslavia, words like 'Ustaša' and 'Balija' , derogatory terms for Croats and Muslims respectively, became common in the documents of Serbian armed forces and Serbian government bodies, as did the use of words like 'Četnik' to denote Serbs in Bosnian Muslim army and civilian records. This kind of evidence demonstrates how effective propaganda becomes part of the public's ideology and daily life.369

4.3
Ideological Context and Historical Anchors Related to this is the need to consider the ideological context of particular words uttered as it indicates how they are meant to be understood and how they are in fact understood by the followers of said ideology (see discussion on Greater Serbia above). Furthermore, the history of the peoples involved is 465 The Prosecutor v. Vojislav Šešelj international criminal law review 20 (2020) 405-491 essential context to consider. In Tadić, the TC described how the media was very effectively directed towards stirring up Serb nationalist feelings and converting an apparently friendly atmosphere as between Muslims, Croats and Serbs in Bosnia and Herzegovina into one of fear distrust and mutual hostility.370 De la Brosse described how history and the revival of ancient Serbian myths were used in this sense to keep the masses mobilised. 371 Historical facts were imbued with mystical qualities to evoke a feeling of desire for revenge directed at the 'others' , i.e. the Croats and Muslims.372 In this de la Brosse found parallels with Nazi propaganda: Nazi propaganda had shown that myths bind the masses together tightly. Indeed, it was through myths and, therefore, the appeal to the forces of the unconscious, to fear and terror, the instinct of power and the lost community that the propaganda orchestrated by Goebbels had succeeded in winning over the Germans and melding them into a compact mass. The Serbian regime would use a similar technique. To weld the population together official propaganda drew on the sources of the Serbian mystique, that of a people who were the mistreated victims and martyrs of history and that of Greater Serbia, indissolubly linked to the Orthodox religion.373 Political, historical and cultural propaganda each played their part.374 Oberschall notes in this respect the private aspect of the propaganda campaign, in the sense that Yugoslavs experienced ethnic relations through the frames of personal and family experience and through community memories of the Balkan wars.375 These memories were repeated in culture and in public life in collective myths and in history books and literature.376 The television and newspapers were filled with historical features glorifying medieval Greater Serbia and listing, many times in an exaggerated manner: 370  [p]eople are reluctant to change; thus, to convince them to do so, the persuader has to relate the change to something in which the persuadee already believes. This is called an anchor because it is already accepted by the persuadee and will be used to tie down new attitudes or behaviours.383 377

4.4
Shifting the Cognitive Frames Paranoia propaganda was extremely effective in Yugoslavia. As the head of the International Committee for the Red Cross mission in Croatia in 1991-92 stated: 'the conflict [in Bosnia] was the first time I have seen such strong and effective propaganda on both sides. When you are talking to either side, they are absolutely convinced they will be slaughtered by the other side' .384 Jovan Rašković, the Serb leader in Croatia, whose campaign for Serb autonomy highlighted Ustaša atrocities against Serbs in World War Two, reflected thus on the effect it had: 'I feel responsible for this [Croatian] war. If I hadn't created this emotional stress in the Serb people, nothing would have happened. My party was the fuse of Serb nationalism' .385 As Oberschall observed, Yugoslavs experienced ethnic and nationality relations through two competing frames, both culturally available: a normal frame in peaceful times expressing cooperative relations in workplaces, neighbourhoods, and in public affairs, and a rival crisis frame for times of tension and conflict.386 The crisis frame was grounded in the experiences and memories of the Balkan wars, the first and second world wars -and other wars and conflicts before those in which civilians were not distinguished from combatants.387 Politicians, and other elites pursuing the nationalist agenda did not invent the crisis; they activated and amplified it while discrediting the normal frame.388 This was first achieved over the plight of the Kosovo Serbs who were subject to intimidation and violence.389 Serb intellectuals activated the crisis frame with fears of extinction.390 The public was fed exaggerations of sexual assault and rape numbers in Kosovo and a false claim that they were a matter of Albanians against Serbs, instead of the reality, which was that they mostly occurred within, and not across nationalities, and where they were overall lower than in central Serbia or the rest of Yugoslavia.391 At the icc in the Gbagbo and Blé Goudé case, the Prosecutor sought to establish a causal link between propaganda of an impending genocide and the spread of violence in Côte d'Ivoire. Yet the majority opinion dismissed the gravity of such a narrative. Judge Henderson recognised that some of the 384  speeches could have constituted fear-mongering that instilled a certain level of fear and resentment among the listeners, however he considered it 'possible to understand these utterances first and foremost as an effort to delegitimise political opponents and their international backers' .392 He thus stated that: the fact that the pro-Gbagbo population was projected as being the potential victim of genocide does not imply an approval or encouragement to reciprocate. Even though the threat and actual instances of use of violence by the opponent's side was repeatedly emphasised, speakers systematically reassured their audiences that the situation was under control and that their side would prevail. This reinforces the impression that talk about atrocities was mainly designed to foster unity among Mr Gbagbo's supporters and loyalty to his regime.393 Furthermore, Judge Henderson reached the same conclusions regarding speeches that characterised the opposition supporters as rebels and bandits or those that insisted on crimes allegedly committed by the pro-opposition forces.394 From this opinion one can again observe how a dangerous propaganda technique is downplayed by the judges as normal political talk and not properly evaluated through expert opinion. There is also an insistence on express encouragement of violence as the only indicator of speech that can lead to atrocities.

4.5
Looking Beyond Isolated Words While historical and present facts used in hate propaganda can be true, the intent of the speaker and the context in which they are conveyed to the public can mean the difference between such propaganda and neutral reporting. As Gordon observes, out of all the elements considered in evaluating hate speech as incitement, context is the most important.395 Some of the Kangura and rtlm articles and broadcasts conveyed historical information, political analysis or advocacy of an ethnic consciousness regarding the inequitable distribution of privilege in Rwanda. 396  The Court thus clearly drew the line between legitimate educational messages involving historical facts and political analysis on the one hand and statements intended to inflame ethnic resentment, calling on history as an aide in this effort.407 As mentioned above, unfortunately the AC took a much narrower approach in considering what criminally relevant hate speech was.
Another neglected aspect of hate propaganda in the search for a causal link is the indirect effects it has on criminality beyond direct persuasion. There is a significant difference between micro-and macro-criminality in the way a person examines the morality of their own actions. When all one can hear in the media is denigrating speech and exhortations of violence from authority figures, they may conclude that the usual strictures have been lifted and that now assaulting members of other ethnical or religious groups will go unpunished.408 Le Bon describes this as such: The isolated individual may be submitted to the same exciting causes as the man in a crowd, but as his brain shows him the inadvisability of yielding to them, he refrains from yielding. This truth may be physiologically expressed by saying that the isolated individual possesses the capacity of dominating his reflex actions, while a crowd is devoid of this capacity.409 Further, Le Bon notes: The violence of the feelings of crowds is also increased, especially in heterogeneous crowds, by the absence of all sense of responsibility. The certainty of impunity, a certainty the stronger as the crowd is more numerous, and the notion of a considerable momentary force due to number, make possible in the case of crowds sentiments and acts impossible for the isolated individual. In crowds the foolish, ignorant, and envious persons are freed from the sense of their insignificance and powerlessness, and are possessed instead by the notion of brutal and temporary but immense strength.410 Propaganda helps produce public support for collective violence in inter-group relations that would normally be condemned, prosecuted and limited. [I]n hate crime, a person is attacked not randomly, but precisely for being perceived as X. In other words, hate crime identifies, categorizes, and labels persons according to real or supposed features such as sex, race, class, sexual orientation etc. This act of labelling a person as someone or something is in itself already a linguistic act of positing, an act of denomination and determination that attributes a social status to a person. This tendency to differentiate and to discriminate can go as far as to restrict the use of the term 'human' exclusively for the designation of one's own social group.416 Thus, before considering the effects of hate propaganda on the 'reasonable man' with no previous inclinations towards violent action, we need to recognise the fact that in any society there are also men (or women) eager to take on any cause that would allow them to act out their aggressive tendencies. Although it is hard to determine exactly the percentage of psychopaths, sociopaths and individuals with other relevant personality disorders in a society, what we do know is that there exists a significant number of them. One could say, they are the main reason criminal law and the outwardly affirmations of basic principles of morality exist, since these individuals lack an internal 'golden rule' . Propaganda, on the other hand, creates the perception that the law no longer applies or, better yet, that there is a new law and a new morality established in the society thereby enabling rampant criminal behaviour. 'Propaganda defines or creates our reality and describes what is socially acceptable and considered to be morally right. It determines how people's actions are viewed and evaluated' .417 Arguably psychopaths or sociopaths do not need such a 'new law or morality' in order to act out their evil tendencies, as they tend to do it regardless, however the impression of a suspension of all societal sanction empowers them to act out these tendencies on an exponentially larger scale. However, even people who become perpetrators as a result of their personality (self-selected or selected by their society for the role) evolve along a continuum of destruction, while others who were initially bystanders become involved with the destructive system and become perpetrators.418 In most societies there are those who are prepared to turn against other groups, however it is the population as a whole that provides or denies support for this. Incitement to hatred is an integral, necessary and inevitable part of any persecutory campaign. It is needed to convince those carrying out the physical elements of the campaign of its necessity, and of the lesser value and perniciousness of the group to be ostracized. It serves to create a common bond between those carrying out the persecutory campaign in that it creates a barrier between those who are part of the community or 'in-group' and those excluded from the community, the 'out-group' .424 A good example of this from the Yugoslav wars can be observed in the video depicting the events surrounding the murder of 6 Bosnian civilians by the Scorpion paramilitary unit. One of the Scorpions turns to a 17-year old victim on the ground and asks him in a mocking and vulgar manner whether he had ever had sexual intercourse, proceeding to answer his own question with a blunt answer 'well, you never will' .425 Interestingly, the same man who spoke so brutally to the petrified young man does not shoot when the time for it comes just minutes after.426 He keeps his barrel raised in the air, and does not fire, suffering ridicule and humiliation from his comrades. 427 An important study by Scott Straus into the effects of rtlm Radio on the genocide in Rwanda shows a far less straight forward causal link between the broadcasts and the crimes than generally assumed. As he points out, there were examples of rtlm broadcasts giving specific names and places, which were followed by attacks on those individuals and locations, however they were a tiny fraction of the total violence and the Nahimana Trial Judgment lists a mere 10 instances. 428  Provocation may have also existed in the context of Šešelj's speeches, however it was by no means more important or effective than hate propaganda and indeed the success of the provocations rested on the background of extensive propaganda. Had the population not been already primed through systematic and repetitive propaganda, any individual speech would not have had the explosive effect the judges were looking for. As the expert described in Šešelj, propaganda techniques are used 'to convince, persuade and influence public opinion' and they 'focus on manipulating emotions and preconceptions, at the expense of the faculties of reasoning and judgment' .439 In the words of Gustave Le Bon: When it is wanted to stir up a crowd for a short space of time, to induce it to commit an act of any nature … it is necessary that the crowd should have been previously prepared … When, however, it is proposed to imbue the mind of a crowd with ideas and beliefs … the leaders have recourse to different expedients … affirmation, repetition, and contagion. Their action is somewhat slow, but its effects, once produced, are very lasting.440 Judge Henderson delivering the majority opinion in the Gbagbo and Blé Goudé case, confirmed the focus on the temporal and special link when adjudicating propaganda charges. He stated that: [a]lthough there is no fixed level of spatio-temporal proximity between the act(s) of inducement and the induced act(s), there must be a clear connection between the two. Simply arguing that everything Mr Gbagbo [the accused] said during a particular time-period influenced all criminal conduct that occurred during the same period is unacceptably vague.441 Such an approach fails to grasp the nature of how propaganda works and its many subtle and subconscious effects that accumulate in time.

4.7
Undue Emphasis Placed on an Identity between Words Uttered and Acts Committed Furthermore, the fact that the judges were looking for the crimes committed to be nearly identical to what was said in the speeches shows a reluctance to consider any speech whose message may not be as direct as possible instigation. 439  This sends the message that hate propaganda is acceptable as long as it is clever enough not to directly call for crimes.
In the Babić case, the defendant pleaded guilty for his participation in persecution against Croats and other non-Serbs as a cah in accordance with jce by making 'ethnically based inflammatory speeches that added to the atmosphere of fear and hatred amongst Serbs living in Croatia and convinced them that they could only be safe in a state of their own… [u]ltimately this kind of propaganda led to the unleashing of violence against the Croat population and other non-Serbs' . 442 In Prlić et al. the TC concluded that 'in several official and public statements, Jadranko Prlić did indeed engender fear, mistrust and hatred of the Muslim population among Bosnian Croats… and exacerbated nationalist sentiments among the Bosnian Croats, thus contributing to the realization of the jce' . 443 It is clear that the causal link between the accused's messages and the effect on his audience was more broadly understood in these cases than the very narrow look at causality in Šešelj. This kind of reasoning is more similar to the case law of the Nuremberg Tribunal than the very narrowly construed causal link in Šešelj. This is in line with Wilson's study that suggests that revenge speech has the most powerful effects overall, and references to past atrocities (conceptually related to revenge speech) enhance moral justifications for violence. 444 Babić claimed to have been strongly influenced and mislead himself by Serbian propaganda, which repeatedly referred to an imminent threat of genocide by the Croatian regime against the Serbs in Croatia, thus creating an atmosphere of hatred and fear of the Croats. 445 The argument was unsuccessfully claimed as a mitigating circumstance the same as in the Banović case, where the defence argued that 'with his low education and modest intellectual capabilities, the Accused easily succumbed to the war propaganda that spread collective hatred and rumours about the enemy's brutality' .446 These arguments show how such ideas spread from one person to another and how the causal link with the subsequent crimes committed may not be straightforward and may be impossible to prove, however it nevertheless exists.

4.8
Position of the Speaker A further important contextual element to be considered in evaluating hate propaganda and its potential impact is the position of the speaker and the influence he exerts on an audience. Importantly, a number of social science studies also confirm a pervasive and deep obedience to authority figures.447 For example, Šešelj portrayed himself as a military leader and established a military wing of his party, created a War Staff, promoted the Četnik movement's militaristic traditions, appeared in military attire at frontlines, and took on the title Vojvoda (duke) for himself, as well as bestowing it on Šešeljevci, including those found to have committed serious crimes.448 As Gordon details, analysis of the context of the speech should: embrace aspects of the speaker herself: her background and professional profile, her previous publications/broadcast history, and her personal manner of transmission (including tone of voice). It would also include the authority of the speaker. Are we dealing with a high-level government official (or even lower-ranking but with sufficient stature to have a significant impact on public opinion) or a private person with other indicia of authority, such as media personality, tycoon, or political activist?449 In Gacumbitsi, the TC considered the impact of the accused driving around, using a megaphone, asking Hutu young men whom girls had refused to marry, to have sex with young Tutsi girls, and killing them in an atrocious manner, if they refused to do so. It concluded that: Placed in context, and considering the attendant audience, such an utterance from the Accused constituted an incitement, directed at this group of attackers on which the bourgmestre (mayor), had influence, to rape Tutsi women. That is why, immediately after the utterance, a group of attackers attacked Witness taq and seven other Tutsi women and girls with whom she was hiding, and raped them.450 480 international criminal law review 20 (2020) 405-491 While the paragraph does not really delve into the reasons for his influence, except for mentioning his position as bourgmestre, his exceedingly high status in the community went beyond this title. One witness described him in an interview as: a big man; a pillar of church and community. He was the man you went to if you had problems with education for your children, or disputes over farmland. Mr Gacumbitsi was the man who preserved order in the crowded hills. Who kept a signed blessing from the Pope on the walls of his home. His was the face of authority.451 The importance of the position of the speaker for the ability to instigate largescale atrocities is already recognised in the icty and ictr jurisprudence. While no hierarchical relationship is required for instigation, cases related to instigation that have set precedents have either been those of military commanders or political leaders who wielded substantial military and political authority, such as Gacumbitsi, Šešelj or Kordić.
In terms of aiding and abetting, in the case of Brđanin the TC considered that his inflammatory and discriminatory statements amounted to encouragement and moral support to the physical perpetrators of the crimes, in light of the positions of authority that he held. 452 As Judge Lattanzi noted in her dissenting opinion in Šešelj, in Féret v. Belgium the ECtHR stressed as a general rule that: 'it is vitally important that in their public speeches, politicians should avoid making comments likely to foster intolerance' .453

Failing the Preventative Function of the Law
Regrettably, the drafters of the Charter of the imt did not consider that the significant role played by hate speech during the Holocaust 'might necessitate formulation of a specific offence accounting for the unique characteristics of incendiary rhetoric in the atrocity context' . 454 Similarly, when the drafters of the 1948 Convention on the Prevention and Punishment of the Crime of Genocide considered hate speech, they did so exceptionally within the context of a plan to destroy in whole or in part particular groups. Such narrow focus animated the drafting and the inclusion of Article iii in the Genocide Convention and ultimately the criminalisation of direct and public incitement to commit genocide. To date, there is no international treaty like the Genocide Convention declaring direct and public incitement to cah or war crimes to be, in itself, a crime under international law.455 Focusing on verbal acts before they escalate to physical violence will directly enhance the preventive function of international criminal courts and tribunals. This will be discussed in subsequent sections. As Fletcher notes: [w]e should generalize from the Rwandan case as the Rome Statute has developed a general aversion to impunity as expressed in the Preamble. The argument should be that 'directly and publicly inciting' any of the major four crimes should be punishable.456

The Proposed International Convention for the Prevention and Punishment of cah
The Proposed Convention on cah that was concluded in 2010 by a group of international law experts meeting at the Washington University School of Law in St. Louis (Proposed Convention)457 contains extensive requirement to outlaw incitement of various forms and requesting States 'to endeavour to take measures' to 'prevent crimes against humanity' , including, but 'not limited to, ensuring that any advocacy of national, racial, or religious hatred that constitutes incitement to discrimination, hostility, or violence shall be prohibited by law' .458 In the same fashion as in Article iii(c) of the Genocide Convention 455  that makes the '[direct] and public incitement to commit genocide' punishable, Article 4(2)(e) of the Proposed Convention states that a 'person shall be criminally responsible and liable for punishment for a crime against humanity if that person … directly and publicly incites others to commit crimes against humanity' . If we compare the preventive measures included in both the Genocide Convention and the Proposed Convention we find that the latter contains more detailed language regarding the obligation to prevent.459 As noted by Leila Sadat, 'the Proposed Convention included 'incitement' in Article 4(2)(e), to enhance the treaty's preventive dimension' . 460 Sadat recalled what was noted by the former US Ambassador for War Crimes Stephen Rapp in his Keynote address to the Crimes against Humanity Initiative, 'incitement is often a key precursor to the commission of crimes against humanity' .461

The International Law Commission's Draft Articles on a Global Convention for cah -The Missing Inchoate Crimes
The Proposed Convention discussed above has prompted the International Law Commission (ilc) to include, at its 66th session in 2014, the topic of cah on its long-term work programme. 462 Sadat has noted that during the Commission's 2016 plenary of the Second Report submitted to the Commission by the Special Rapporteur Sean Murphy, some Members were of the position to include 'incitement' as a mode of liability in the ilc's draft articles on cah along the lines of Article iii(C) of the Genocide Convention.463 It is significant in this regard to draw a distinction between incitement or instigation as a mode of participation in a criminal conduct i.e. cah and 'direct and public incitement to commit cah' . The latter stands as an inchoate offence.
Surprisingly, the Commission's draft articles on cah which were adopted by the Drafting Committee in 2017 lacks any provision criminalising 'incitement international criminal law review 20 (2020) 405-491 Commission.469 When the ilc invited feedback from States, Iceland (on behalf of the Nordic countries) and Sierra Leone suggested that the forms of liability should not be interpreted narrowly and should include conspiracy and incitement for all core crimes.470 Yet the reaction from the Special Rapporteur was merely to note that the Rome Statute does not refer to incitement with respect to cah, 'and hence the Commission elected also not to use such terms' .471 If this implies a stance that the law of 1998 should remain as it is, just because it is formulated this way, it would defeat entirely the mission of the ilc to progressively develop the law in light of the new challenges that arise with time.
The Commission's position has been criticised by eminent scholars and by a member of the ilc who actively participated in the drafting process.472 From the Summary records, it is clear that the ilc's reasoning for the non-inclusion of the term 'incitement' in the draft articles was based in part on the fact that the term 'incitement' had not been included in certain international treaties, such as the Rome Statute, and in part on the fact that 'the concept did not exist in some national legal systems' . 473 Schabas comments that this 'is not a good reason to omit direct and public incitement and conspiracy to commit crimes against humanity from the draft articles' and reminds that 'there are unfortunate gaps in Article 25 of the icc Statute' which should be addressed.474 Van Sliedregt calls the absence of 'incitement' as an inchoate form of criminal responsibility noteworthy, especially in light of the Genocide Convention where it is punishable as such. 475 From the point of view of morality the criminalisation of incitement to cah is essential. Van Sliedregt argues that 'promoting an individual to commit a crime may be even more reprehensible than assisting someone who has already decided to commit a crime' . 476