Glorification of Terrorist Violence at the European Court of Human Rights

Abstract

This article examines the European Court of Human Rights’ approach towards restrictions on expression glorifying terrorist violence. This is done by situating the Court’s case law against two objections to respective criminal offences: their inherent overbreadth and their incompatibility with the restraining demands of the ‘harm principle’. In doing so, the article discusses how the ‘harm principle’ relates to the proportionality test and how the Court’s categorisation of expression glorifying violence responds to the objection of overbreadth. In arguing that the tool of categorisation has not been determinative in driving the outcomes in relevant decisions, the article suggests that engaging the existence of a competing public interest and reviewing the admissibility of reasons for such restrictions would appropriately elevate the Convention standard. Finally, the article argues that inconsistencies across decisions are best explained by the Court’s deference-giving practices, particularly in cases involving claims about the recency of terrorist violence.

1. Introduction

The past two decades have seen a proliferation of counterterrorism legislation and expansion of the scope of terrorism-related crimes. Criminalisation of radical speech under broadly defined offences of ‘indirect incitement’, ‘justification’, ‘glorification’ or ‘apologie’ of terrorist violence is one example. The terminology, forms and elements of such offences differ, but the purpose common to most such offences is the prevention of speech that is supportive, approving, celebratory or otherwise sympathetic to terrorist violence. The criminalisation of glorification has long been criticised by rights advocates, but their arguments have not found support from the European Court of Human Rights (ECtHR, the Court). The Court drew criticism for its finding of no violation of Article 10 in Leroy v France, a 2008 decision that implicitly approved of a glorification statute. This article discusses the developments in the Court’s case law and situates its current approach towards the issue in relation to the two objections against criminalisation of glorification of terrorist violence: the inherent overbreadth of glorification statutes and their incompatibility with the restraining principles of criminal law. In doing so, it suggests that mechanisms other than categorisation of glorification as a specific form of (un)protected speech should be determinative in decisions involving interferences against such expression. The Court can effectively elevate the Convention standard by establishing the existence of a competing public interest and reviewing the admissibility of reasons for such restrictions. The article argues that it is the deference-awarding practices employed by the Court in ‘hard cases’ that explain the object-level inconsistencies across relevant decisions.

The second part of the article briefly examines the emergence of glorification offences as a category of criminal law in international, regional and domestic legislative instruments and discusses the different forms in which such expression is criminalised.

Part three presents two primary objections levied against criminalisation of glorification of terrorism: the inherent overbreadth of glorification statutes—what has become and is termed a ‘standard’ objection, as well as a more novel argument that appeals to restraining principles of criminal law in general and the ‘harm principle’ in particular.

The fourth part presents the Court’s post-Leroy case law against the background of the two objections. It first discusses how restraining principles of criminal law relate to the proportionality test, suggesting that although the latter is agnostic to respective normative concerns, it can nonetheless produce results that similarly generate limits binding the employment of criminal law. The discussion then shifts to the role of categorisation of expression as (un)protected in the ‘standard’ objection and analyses the Court’s approach towards the categorisation of glorification. It is suggested that the role categorisation in driving the outcomes of the decisions might be overstated. Instead, it is the background mechanisms of establishing the existence of a competing public interest and reviewing the admissibility of reasons for such restrictions that can get the Court’s practice closer to a more demanding standard in addressing glorification. The final section suggests that the Court’s deference-giving practices explain the inconsistencies across relevant decisions. It is the Court’s treatment of the governments’ claims about the effects of recency of terrorist violence that appears to drive such inconsistencies. A brief conclusion follows.

2. The Emergence and Entrenchment of `Glorification of Terrorism' as a Matter of Criminal Law

One of the drivers for the expansion of the global counterterrorism regime following the 9/11 attacks was a ‘lacuna’ in the global counterterrorism law that the attacks were argued to have exposed.¹ In part, the legislators’ search for the now potentially dangerous and previously unregulated elevated a then-obscure notion of ‘glorification of terrorism’ into an object of international regulation and concern. This was the case with the work of the Multidisciplinary Group on International action against Terrorism that, in reporting to the Committee of Ministers of the Council of Europe in 2002, suggested the body to ‘further study the concept of “apologie du terrorisme” and of “incitement to terrorism”, and to identify, inter alia, the proper balance between freedom of expression and the need to prevent terrorism’.² A study identifying such offences in the legislation of the member and observer states followed, with the apologie offences understood to include ‘the public expression of praise, support or justification of terrorists and/or terrorist acts’.³ Out of 45 states, only three reported to have specific legislation criminalising conduct falling within that definition of apologie du terrorisme—Denmark, France and Spain, with such prohibitions dating back to 1894 in Spain and 1881 in France.⁴

The work that followed⁵ culminated in the introduction of the offence of ‘public provocation to commit a terrorist offence’ in Article 5 of the 2005 Council of Europe Convention on the Prevention of Terrorism (the 2005 Convention). It obliges state parties to criminalise ‘public provocation to commit a terrorist offence’, ‘when committed unlawfully and intentionally’, defined as

the distribution, or otherwise making available, of a message to the public, with the intent to incite the commission of a terrorist offence, where such conduct, whether or not directly advocating terrorist offences, causes a danger that one or more such offences may be committed.⁶

This definition does not strictly correspond to either of the two initial offences—‘apologie and/or incitement to terrorism’—but combines them into a third category, a ‘public provocation’, where the elements common to the crime of incitement (intent to incite and probability of harm)⁷ are combined with the language associated with offences of glorification/apologie (indirect advocacy of terrorist offences).⁸ The Explanatory Report to the Convention states that ‘a certain amount of discretion with respect to the definition of the offence and its implementation’ is allowed to the state parties. ‘[f]or instance, presenting a terrorist offence as necessary and justified may constitute the offence of indirect incitement’, but only insofar as the elements of intent and causality of danger are present.⁹

Several months after the adoption of the 2005 Convention, and in the wake of the terrorist bombings of 7 July 2005 in London,¹⁰ the United Nations Security Council adopted resolution 1624. The operative part of the resolution called upon states to prohibit by law and prevent ‘incitement to commit a terrorist act or acts’,¹¹ while the preamble condemned both ‘the incitement of terrorist acts’ and ‘attempts at the justification or glorification (apologie) of terrorist acts that may incite further terrorist acts’.¹² The United Kingdom sponsored the resolution,¹³ and some suggest that the distinction between what the instrument required and what it only condemned was made to accommodate the approach to freedom of expression in the United States.¹⁴

In some instances, the subtleties and ambiguities of the 2005 Convention and resolution 1624 have been exploited to justify criminalisation beyond their requirements. In reviewing the implementation of the resolution a decade after its adoption, the Security Council’s Counter-Terrorism Committee observed that in their pursuit to implement the resolution, some states had introduced prohibitions on glorification, ‘as well as statements that dehumanize the victims of terrorism … [arguing] that such expression creates a danger of the subsequent commission of terrorist acts’.¹⁵

The debate preceding the adoption of the UK Terrorism Act 2006—brought before parliament two months after the adoption of resolution 1624 and introduced in part to implement the 2005 Convention¹⁶—appears to have shaped the argument of the ‘radicalising’ effects that speech glorifying terrorist violence may have.¹⁷ The law introduced the offence of ‘encouragement of terrorism’ prohibiting both direct and indirect encouragement. The latter category encompassed statements that glorify ‘the commission or preparation (whether in the past, in the future or generally) of such acts or offences’, including when the speaker is ‘reckless as to whether members of the public will be directly or indirectly encouraged or otherwise induced by the statement to commit, prepare or instigate such acts or offences’.¹⁸ It went much further than Article 5 of the 2005 Convention it sought to implement by covering reckless statements and expanding the outcomes from the ‘commission of an offence’ to their preparation or instigation.¹⁹ This reflected how the law treated glorification: whereas acts of incitement are worthy of suppression because they cause violence, acts of glorification are to be suppressed because they can contribute to one’s ‘radicalisation’. Thus, neither objective causality nor subjective intent of the speaker needs to be required, nor such ‘contributions’ need to be explicit.²⁰ These concerns are now often expressed in terms of the effects that the presence of such speech has—one creating a ‘climate that is conducive to terrorism’²¹—rather than the direct effects of specific statements. One notable example is the 2014 French counterterrorism law that, among others, removed the offence of apologie from the 1881 press freedom statute to introduce it into the criminal code (with increased punishment specified for such acts committed online).²² Its explanatory memorandum states that ‘France cannot tolerate that on its own soil, messages can be disseminated with impunity calling for terrorism or glorifying it. These messages are part of ideological conditioning and are likely to lead to the commission of acts of terrorism’ (emphasis added)²³.

Another notable place held by glorification offences is in the European Union (EU) counterterrorism instruments. Following the language of the 2005 Convention, the 2008 EU Framework Decision on combating terrorism included an offence of ‘public provocation to commit a terrorist offence’, criminalising acts ‘whether or not directly’ advocating terrorist offences, but requiring both the intent to incite and a causal link of the danger of such offences being committed.²⁴ The 2017 Directive that replaced it did not change the elements of intent and causality but specified that indirect advocacy includes conduct ‘such as … glorification of terrorist acts’.²⁵ Recital 10 states that the offence ‘comprises, inter alia, the glorification and justification of terrorism’.²⁶

The significance of including glorification as an example of indirect advocacy of terrorism in the 2017 Directive has been subject to differing interpretations,²⁷ with some actors renouncing criminalisation of indirect provocation in general²⁸ (that is, including glorification) and others criticising the usage of the terms like ‘glorification’ or ‘promotion’ of terrorism,²⁹ but, ostensibly, not acts of speech ‘whether or not expressly advocating the commission of terrorist offences’.³⁰ The Directive was set to be transposed by the EU member states into their national legislation by September 2018.³¹ In 2021, the EU’s Agency for Fundamental Rights reported that the requirement of danger was not present in 15 jurisdictions.³² Moreover, ‘even where the law specifically prescribes this assessment, it is not necessarily carried out in practice’.³³

3. Objections to the Criminalisation of Glorification of Terrorist Violence

A. The `Standard' Objection of Inherent Overbreadth and Leroy v France

The criminalisation of acts of glorification in a form that does not require a specific intent and fails to establish a causality requirement has long been an object of criticism from academic commentators, non-governmental organizations and international bodies. A 2008 UN Secretary-General report summarises what now represents the standard argument against it:

Incitement must be separated from glorification. If the first may be legally prohibited, the second may not. In the context of the fight against terrorism a troubling trend has been the proscription of the glorification of terrorism, involving statements which may not go so far as to incite or promote the commission of terrorist acts, but nevertheless applaud past acts. While such statements might offend the sensitivities of individual persons and society, particularly the victims of terrorist acts, it is important that vague terms of uncertain scope such as “glorifying” or “promoting” terrorism not be used when restricting expression. Incitement can be understood as a direct call to engage in terrorism, with the intention that this will promote terrorism, and in a context in which the call is directly causally responsible for increasing the actual likelihood of a terrorist act occurring. In their fight against terrorism and in order to respect the freedom of expression, States should be careful to differentiate the two notions.³⁴

This account deems the offences to be inherently vague and overbroad³⁵—a problem that can only be rectified by narrowing it to cases of intentional and possibly consequential calls to engage in terrorist violence, i.e. to the offence of incitement.³⁶ In other words, it does not suggest that reckless acts of glorification are worthy of protection as such or that they are necessarily inconsequential compared to incitement. Rather, the claim is that it is impossible to criminalise such forms of expression without it leading to ‘unnecessary or disproportionate interference with freedom of expression’.³⁷

The assertion underlying this objection finds support in the facts of the 2008 decision of the ECtHR in Leroy v France.³⁸ The applicant in the case was convicted for complicity in ‘apology of terrorism’ under an 1881 statute that had neither mens rea nor a requirement of causality of further violence and with the national court explicitly rejecting the defendant’s intention in the criminal proceedings.³⁹ The act of speech in question was a cartoon depicting the attack on the World Trade Center accompanied by a caption parodying an advertising slogan.⁴⁰ It was published on 13 September 2001 and later explained by the author to be an expression of his anti-Americanism.⁴¹ The ECtHR did not find a violation of Article 10. It dismissed the author’s stated motive in favour of the accompanying caption, finding it to be an expression supporting the attack’s perpetrators, not criticism of US imperialism.⁴² Further reasons adduced by the Court included the ‘temporal dimension’ of the publication, its impact ‘in a politically sensitive region’ (Basque country) and the reactions that followed, ‘which could fuel the violence and demonstrate its plausible impact on public order in the region’, as well as the severity of the punishment (a fine).⁴³ While upholding the decision of the authorities to punish the act of speech in question, the decision vindicates the objection against criminalisation of glorification by illustrating how its widely cast net inevitably suppresses speech, however distasteful, that either was not meant to be targeted or is not worthy of suppression otherwise.

The criticism of the decision largely aligned with the ‘standard’ objection in suggesting that the Court failed to uphold its previous position of finding restrictions failing to constitute incitement to violence to be unjustifiable.⁴⁴ Stefan Sottiaux argued that the Court’s preceding jurisprudence has moved towards a relatively fixed standard—formulated and advanced in a series of 1999 decisions that primarily focused on whether an expression ‘incites to violence against an individual, a public official or a sector of the population’⁴⁵—a question which the Court answered by engaging with various factors that included the content of the acts of speech, the intention of the speaker, possible consequences and their seriousness.⁴⁶ Against this background, the Court’s omission of the previously articulated standard and insufficient engagement with the facts was criticised as methodologically weak in its ‘ad hoc nature of the decision-making process’, where ‘the Court relapses into its earlier open-ended democratic necessity approach’.⁴⁷ Andrew Dyer qualifies this criticism by suggesting that the Court did engage with relevant elements of the ‘incitement’ standard but failed to do it accurately;⁴⁸ then, what Leroy demonstrates is that the ‘test is a vague one that allows judges considerable scope to prohibit speech merely because they consider it to be highly offensive’.⁴⁹

B. Glorification and Restraining Principles of Criminal Law

The second form of criticism is a more recent development. It emerged in response to the change brought to criminal law by criminalisation of preparatory and (pre-)inchoate terrorist offences that cover acts occurring much earlier than the actual infliction of harm.⁵⁰ The UK Terrorism Act 2000, for example, criminalised possession of ‘an article in circumstances which give rise to a reasonable suspicion that his possession is for a purpose connected with the commission, preparation or instigation of an act of terrorism’ and collection of ‘information of a kind likely to be useful to a person committing or preparing an act of terrorism’.⁵¹ Recently introduced offences targeting ‘foreign terrorist fighters’ function similarly and include, among many others, crimes of ‘attempting to travel for the purposes of planning or receiving terrorist training’⁵² or prohibiting ‘any act of facilitation that assists any person in travelling abroad for the purpose of terrorism’.⁵³ Such expansion of the body of criminal law is usually justified by the destructive nature of terrorist violence and the catastrophic harm a single act can bring, thus demanding its prevention and consequently sanctioning ‘early stage’ interventions.⁵⁴ The offences of glorification are part of this trend;⁵⁵ as discussed above, the primary rationale behind the suppression of such acts of speech is the prevention of only possible acts of violence.

Reflecting on this development in the context of control orders (a preventive counterterrorism tool that typically grants the executive authorities powers to impose various restrictions on individuals outside of criminal justice processes), Lucia Zedner suggests that while being a powerful mechanism, ‘there are limits to human rights discourse as a means of understanding and resisting’ preventive justice and that a ‘broader and denser set of interpretative frames is needed to analyse’ and ‘determine the principles by which it might be contested and constrained’.⁵⁶ Although not explicitly named, the proportionality test must be part of the ‘human rights discourse’ being critiqued.⁵⁷

One way of circumventing the proportionality test in criticising glorification offences that gained traction draws on normative constraints of criminal law.⁵⁸ In particular, the idea of ‘harm’ as a relevant restraining condition is often invoked to provide the limit on the exercise of coercive power by a state and suppression of speech. One form of the argument appeals specifically to the ‘harm principle’ as a matter of criminal law theory.⁵⁹ For example, Helen Duffy and Kate Pitcher suggest that criminal law can interfere to respond to (a real risk of) infliction of harm to a protected value, where the respective punishment must reflect the contribution made by the individual to the inflicted harm or risk.⁶⁰ ‘Conversely, “remoteness” is a constraining principle of criminal law, such that the individual cannot be prosecuted for speech with no meaningful proximate link to an ultimate wrong’.⁶¹ Another form of the argument rests on the idea of the ‘element’ or ‘requirement’ of harm being a necessary condition for restriction on expression.⁶² Anneke Petzsche and Manuel Cancio Meliá, for example, suggest that speech offences must contain ‘a strong nexus to the actual terrorist offence/act’ that ‘needs to be clearly defined’ to meet the requirement.⁶³ They find Brandenburg v Ohio⁶⁴ to be informative in clarifying what the content of such nexus can be, with its test of ‘clear and present danger’ incorporating ‘the requirements of imminence and likelihood which, if taken seriously and transferred to terrorist speech offences, provides for the establishment of a strong causal nexus’.⁶⁵ In a recent publication, Alice Gardoll finds that ‘Under international standards, speech cannot be restricted unless a state can demonstrate a specific type of harm that may arise from it’.⁶⁶ Consequently, ‘crimes such as the “glorification”, “justification” and “apology” of terrorism have been deemed to violate international law’.⁶⁷

Although this connection does not appear to be acknowledged, the two forms of the argument must be complementary, with the ‘harm principle’ functioning as a ‘first principle’ and the ‘harm requirement’ as a form of its operationalisation in criminal law.⁶⁸ The authors discussing the ‘harm principle’ in relation to the glorification offences do not appear to subscribe to a particular form of the principle or a specific idea of ‘harm’ but rather to a broader idea of it being a relevant normative constraint binding the law-making. The scarce detail with which the principle is normally presented need not as such be fatal for the validity of the objection. Indeed, the demand that criminal restrictions on expression are imposed only to protect an important interest (e.g. harm to others) stands as long as freedom of expression itself is recognised as an interest of such importance as to require a weighty reason to limit its enjoyment. This much is uncontroversial. What seems to be less well founded is that the ‘harm requirement’ necessarily follows from the ‘harm principle’. A requirement that a statute only targets expression likely to cause harm is certainly sufficient to implement the ‘harm principle’, but it does not seem necessary to do so. As noted in the preceding discussion, glorification offences treat praise of violence as harmful because the state in which it remains unsuppressed is understood to be more conductive to terrorist violence.⁶⁹ The collective interest in not creating a state that is more conductive to terrorist violence seems worthy of protection. Then, the expression that causes a loss to such interest seems to be properly ‘harmful’ as to warrant its suppression.⁷⁰ That the contribution made by any one such statement can be imperceptibly small need not mean that it is not real.⁷¹

The claim that the requirement of harm is a standard of international human rights law needs to be addressed separately, if only briefly, since it asserts its foundation in a principle different from the ‘harm principle’.⁷² The practice of the Human Rights Committee cited in support of the argument is best described as requiring that restrictions on rights are only imposed to protect a specified interest that would otherwise be threatened. The wording of the relevant paragraph⁷³ of the General Comment 34 reads as follows:

When a State party invokes a legitimate ground for restriction of freedom of expression, it must demonstrate in specific and individualised fashion the precise nature of the threat, and the necessity and proportionality of the specific action taken, in particular by establishing a direct and immediate connection between the expression and the threat.⁷⁴

The decision it cites in support concerned a conviction of an artist prosecuted under a national security statute for producing a painting that constituted an ‘enemy-benefiting expression’.⁷⁵ The government has failed to justify the restriction with reference to either of the purposes enumerated in Article 19(3) of the Covenant (e.g. right of others, national security or public order), with the Committee noting that the government ‘must demonstrate in specific fashion the precise nature of the threat to any of the enumerated purposes caused by the author’s conduct’ in finding a violation.⁷⁶ The other decisions cited in support of the claim equally concern a requirement that restrictions properly relate to one of the enumerated purposes;⁷⁷ the same is true of the later decisions explicitly citing the relevant part of the General Comment 34.⁷⁸ There appears to be no reason barring the use of term ‘harm’ to describe what the Committee describes as a ‘threat’ to an enumerated purpose, but its conversion into a ‘requirement of harm’⁷⁹ for criminal statutes proscribing expression is not well founded. It might be that a rights’ restriction imposed by a criminal offence that incorporates such a requirement is more likely to be justifiable, but it need not follow that this is a necessary condition. A restriction imposed by an incitement offence with no harm requirement enforced in circumstances where the expression poses a risk of harm would be equally sufficient to meet the standard articulated by the Human Rights Committee.

It is notable that despite having different starting points, both the ‘standard’ and the ‘criminal law’ objection, in both forms, reach an identical prescriptive conclusion that only acts of speech that are likely to be causally responsible for actual violence can be justifiably criminalised. On the one hand, this may be taken as evidence for the correctness of the conclusion since different accounts converging on the same normative standard speaks to the strength of said standard.

An alternative interpretation is that the convergence exists because the conclusion might be preferred for reasons other than the explicitly stated ones. Consider, for instance, that neither approach seems to engage with (the strongest) reasons for criminalisation of glorification—the assertion about the negative aggregate effect of non-sanctioning of expression glorifying terrorist violence. The dangers of any individual instances in which violence is glorified is not an issue motivating the instruments criminalising it; a suggestion that any one instance needs to be satisfy the requirements of intent and causality, or simply causality as the ‘harm principle’ objection would require, fails to capture the issue in question. Instead, doing so would require one to consider, for instance, if the effects of the presence of expression glorifying terrorist violence—absent a criminal sanction—create a more positive (however understood) state than that in which it is criminally sanctioned. For some, this might depend primarily on empirical questions (is the presence of such expression ‘conducive to terrorism’, and if so, how strong is the effect?); others might be more concerned with the normative implications (should we engage criminal law to punish acts of speech that are harmless individually but possibly dangerous in the aggregate?) or with some combination of both (how should the empirical uncertainty about the possible negative effects of non-sanctioning such expression be weighed against the normative uncertainly about the proper limits of freedom of expression?).

Crucially, this claim need not imply disingenuity on part of those presenting such objections. Rather, one’s preference for a categorical objection against such offences in forms that require neither intent nor causality might be determined, for example, by a higher-level engagement with respective costs and benefits. In particular, glorification statutes are known to be prone to abuse by way of enabling the targeting of political opponents as well as individuals whose expression—although falling within the offences’ scope—is not worthy of prosecution or harsh punishment.⁸⁰ Against this background, one can reasonably prefer to allow for some potentially dangerous expression (glorification of violence that fails to reach the requirements for incitement) to remain unpunishable if this norm against criminalisation of such expression would also serve to disallow its persecutory application.

Ultimately, this contribution’s primary interest lies in situating the ECtHR’s adjudicatory approach towards glorification of violence against the two objections discussed above. It is the outcomes they generate, not their validity, that is of relevance, as well as the question of how such outcomes compare to those produced by the Court’s reasoning methodology. The following discussion first briefly addresses the relationship between restraining principles of criminal law and the Court’s approach to proportionality and then proceeds to analyse it against the demands of the ‘standard’ objection.

4. Glorification of Violence at the European Court of Human Rights

The growing number of relevant decisions makes it difficult to locate the limits that restrain punishment for such expression in the Court’s case law. Indeed, as the following discussion demonstrates, the Court’s reasoning in relevant decisions seems to differ from case to case. Before moving on to present the Court’s relevant case law, it is worth highlighting that ‘case law’ as such might not be the unit of analysis capable of appropriately explaining the Court’s position on the matter. Instead, the proper explanation might be exogenous to the narrow issue of glorification; that is, the Court’s position might be better explained by broader factors that control how it engages with cases before it, such as the changing standards of review,⁸¹ its application of the ‘margin of appreciation’ or broader tactics guiding the employment of deference. In other words, these decisions might appear inconsistent when one looks at them through the prism of the broader proportionality test, more narrow standards establishing what satisfies the test of ‘necessity in a democratic society’, further more narrow Article 10 standards, principles governing violence-related speech, yet the same seemingly inconsistent decisions might be consistent from the standpoint of Court’s deference-giving practices, which, in turn, would explain the object-level inconsistencies. This is indeed the conclusion drawn here. What appears to better explain the inconsistencies across relevant decisions is the intersection between the Court’s chosen level of deference to national authorities and the prima facie justifiability of restrictions in particular cases. This argument is elaborated in the third section of this part; in what follows, the discussion first addresses the question of the Court’s approach to proportionality and its relation to criminal law principles as well as the ‘standard’ objection.

A. Restraining Principles of Criminal Law and Proportionality

Examining how the ‘harm’ objection against glorification offences interacts with the Court’s adjudicative method might seem rather futile. Indeed, the objection is at least in part an objection to proportionality-guided adjudication as it fails to account for considerations carried by such restraining principles—the justifiability of criminalisation of conduct as such. As far as the ECtHR is concerned, this is due to the two main obstacles.

The first one is the explicit permissibility of restriction of rights in favour of various matters of public interest. The limitation clauses vary slightly for each limitable right in the Convention, but all are broader than prevention of harm, at least insofar as they allow for rights to be restricted for ‘the protection of morals’.⁸²

The second obstacle is the object of review: the Court’s review is primarily case-specific, that is, it typically does not address the compatibility of the underlying legislation with the Convention, but only the individual application of such legislation.⁸³ Decisions in which the Court has indicated that a specific offence is incompatible with the Convention as such are rare,⁸⁴ and the reasoning varies greatly.⁸⁵ In other words, it is conceivable that the properties of a criminal offence that improperly criminalises some conduct would consistently render its application disproportionate, but for the offence itself not to be found incompatible with the Convention. Consider, for instance, the Court’s decision in Erkizia Almandoz v Spain, in which the applicant was convicted under the glorification offence for the speech he delivered as a keynote speaker at the event in tribute to a former member of the Euskadi Ta Askatasuna (ETA).⁸⁶ Further aspects of this decision are discussed below; of most relevance here is the Court’s finding that the applicant’s speech did not incite violence either directly or indirectly and that ‘there can be no question of concluding that he had intended to incite the use of violence while justifying and praising terrorist violence’ with respect to the more ambiguous parts of the speech.⁸⁷ Note that the applicant was convicted under a statute that establishes no such requirements.⁸⁸ Further, the Court appears to conclude that in addition to failing to constitute incitement to violence, the applicant’s speech should not be understood to constitute glorification of violence.⁸⁹ Despite the Court’s analysis explicitly addressing the applicant’s conduct through the requirements that the underlying legislation ignores, nothing in the decision suggests that the legislation itself is incompatible with the Convention. Rather, it is the authorities’ intervention that is found disproportionate for not being ‘necessary in a democratic society’.⁹⁰ In a concurring opinion, Judge Lemmens suggested that it would have been preferable for the Court to say explicitly that the problem lies in the law itself, but even absent such a finding, the authorities will have to draw consequences from the judgment.⁹¹

Still, the inability of proportionality-based reasoning to fully capture the normative principles of criminal law is not necessarily a flaw. The primary function of the proportionality test, however construed, is to answer whether a limitation of a right in question is justifiable.⁹² In turn, although not expressly termed in rights language, a negative ‘harm principle’ is effectively constitutive of an immunity-right that is a corollary of a disability on part of the state to employ regulation against (non-)harmful and inconsequential acts of speech that such ‘harm principle’ would demand.

However, to concede that proportionality-based reasoning fails to recognise normative constraints of criminal law principles is not to say that it does not recognise limits to the employment of criminal law. While being agnostic about, e.g. whether the law in question requires an appropriate mens rea or an expectation of causality, proportionality can be sensitive to the presence of these elements in the specific instance in which the law is being applied. As will be suggested in what follows, some of the basic principles generated by the logic of proportionality in relation to glorification offences are the inadmissibility of restrictions in instances where one’s expression fails to engage a competing interest or where one’s expression has been interfered with for inadmissible reasons. These limits might not be as robust as those provided by the constraining principle of criminal law, especially where the adjudicator engages them only so mildly, yet they are limits nonetheless and should inform the elements of respective criminal offences.

B. Glorification as a Category and the `Standard' Objection

First is the question of the Court’s approach to glorification as a category. One possible approach is that of the 2005 Convention and the EU Directives, which treats glorification as a form of indirect incitement, thus subjecting it to the requirements of intent and causality that bind the ‘standard’ incitement offences (‘glorification-as-incitement’). The other is to treat it as a separate category that seeks to punish such acts of speech for reasons other than their inciting nature, but rather because the presence of such speech is believed to be harmful as such (‘glorification-as-glorification’).⁹³ On this account, the respective offences need not require special intent or causal links to violence, as exemplified by the statute in, e.g. France.

It is worth briefly noting the Court’s position on the issue before Leroy. While it was the first decision in which the Court has addressed a glorification statute, it was not the first one to discuss glorification of violence as a category, with the preceding decisions explicitly treating glorification as subject to the ‘incitement standard’.⁹⁴ For example, in Sürek (no. 1), the Court states the ‘incitement standard’, finds that the speech in question met the standard and concludes by underlining that it was not the shocking or disturbing qualities of it that render it unprotected under Article 10, but that it constituted ‘hate speech and the glorification of violence’.⁹⁵ In Şener, a decision of the same period, the categorisation of glorification of violence as subject to the ‘incitement standard’ is less pronounced due to the use of the conjunction ‘nor’ in finding the speech to be protected because ‘the article taken as a whole does not glorify violence, [nor] does it incite people to hatred, revenge, recrimination or armed resistance’,⁹⁶ but the following paragraph underlines that ‘the views expressed in the incriminated article cannot be read as an incitement to violence, nor could they be construed as liable to incite violence’,⁹⁷ with no discussion indicating that a different standard would apply to non-inciting acts of glorification.

It was then the supposed non-application of that standard in Leroy that drew criticism.⁹⁸ Arguably, then, the Court’s approach to the categorisation should inform the relevant standards to be applied and therefore be an issue of some importance. In other words, should ‘glorification-as-incitement’ be preferred, it would align the Court’s approach with the demands of the ‘standard’ objection to glorification: the Court would treat the interferences brought about by glorification statutes under the strict ‘incitement’ standards (intent and causality) and would then only find violations in cases where acts of glorification do not constitute incitement,⁹⁹ effectively establishing that non-inciting glorification (‘glorification-as-glorification’) is not to be interfered with. This would not perfectly align with the ‘standard’ objection as the latter rejects that ‘glorification’ as a legitimate category, but the results of the Court’s findings would closely reflect the remaining outcomes demanded under the ‘standard’ account.

i. Glorification as a category in the Court's case law

The post-Leroy case law and the Court’s broader speech-related jurisprudence appear to treat glorification as a separate category of speech unprotected by Article 10 (‘glorification-as-glorification’) rather than as a subcategory of incitement to violence (‘glorification-as-incitement’). Despite a non-negligible number of relevant decisions, it should nonetheless be noted that providing an accurate account of the Court’s approach towards the issue is difficult due to the overlap in the scope of both approaches: the cases where acts of glorification do constitute incitement and are addressed by the Court as such offer no explanation of the Court’s stance on ‘glorification-as-glorification’. The Court’s position is then to be accounted for primarily through the varying use of conjunctions, examples of unprotected categories of speech that it provides and a smaller number of cases where acts of speech were explicitly deemed to constitute non-inciting glorification (‘glorification-as-glorification’). One example falling into the first category is the ‘key case’ of Perinçek v Switzerland, where the Grand Chamber notes that it is ‘expression that promotes or justifies violence, hatred, xenophobia or another form of intolerance’ that normally does not obtain the protection of Article 10,¹⁰⁰ which would suggest that glorification need not constitute incitement. Similarly, glorification and incitement are independent categories in Altintaş, where ‘hate speech, apology for violence or incitement to violence’ are noted to be forms of expression that are not normally protected.¹⁰¹

Taşdemir v Turkey is an admissibility decision that appears to be the first one to deal with ‘apology of terrorism’ after Leroy.¹⁰² The applicant was prosecuted for shouting slogans in support of the armed wing of the PKK (‘HPG (the armed wing of the PKK) to the front line in retaliation!’) at a rally. The Court found that the slogan amounts ‘to an apology of terrorism’ and the application was deemed inadmissible. Erkizia Almandoz is one out of a small number of decisions involving interferences driven by glorification statutes, with the applicant convicted under a glorification offence on account of the speech he delivered at an event in tribute to a former member of the ETA.¹⁰³ Although the Court notes that it needs to establish whether the applicant’s expression constituted ‘direct or indirect call for violence, or justification of violence’ (emphasis added),¹⁰⁴ it concludes that the applicant did not ‘intend to incite the use of violence while justifying and praising terrorist violence’ (emphasis added).¹⁰⁵ If the former assumes that justification need not incite to be punishable (‘glorification-as-glorification’), the concluding sentence appeals to the higher standard (‘glorification-as-incitement’). This is followed by a finding seemingly independent from the preceding one, with the Court disagreeing with the reasoning of the domestic authorities on the characterisation of the speech as glorification.¹⁰⁶ Judge Lemmens’ concurring opinion in the case is informative in this regard, suggesting that the Court’s use of the phrase ‘justification of violence, hatred or intolerance’ is unfortunate as it can be understood to incorrectly equate calls for violence—that are unprotected—and statements that ‘praise’ or ‘justify’ terrorism that are not unprotected as such: ‘the “simple” act of praising terrorism or justifying acts of terrorism, without such remarks being considered as calls for violence or hate speech, is not sufficient to exempt these opinions from the protection of Article 10’.¹⁰⁷ It concludes with a suggestion for the Court to avoid wording that implies that glorification is a category that does not obtain the protection on part with calls for violence and hate speech (i.e. to treat it as ‘glorification-as-incitement’).¹⁰⁸ Importantly, the concurring opinion does not seem to appeal for a change in the Court’s approach to glorification as a category. Indeed, Judge Lemmens finds the preceding case law to treat glorification as a subcategory of incitement and only suggests that this approach is not to be abandoned in the future and that the textual ambiguity of Erkizia Almandoz is not to be understood to set a new standard. As the preceding discussion shows, the ambiguity of the Court’s approach appears to predate Erkizia Almandoz and what Judge Lemmens seeks to prevent seems to have already happened.

ii. Categorisation, competing interests and exclusion of wrong reasons

It would then be difficult to interpret the ambiguity of the Court’s approach towards the categorisation of glorification as being in line with the requirements of the ‘standard’ objection against interference with expression by glorification statutes; indeed, the ambiguity conflicts with the clarity that the ‘standard’ account seeks to uphold.

Note, however, that the difference in premises underlying the ‘standard’ objection and the Court’s approach need not result in different outcomes. This might seem counterintuitive; after all, a principled rejection of criminalisation of glorification under the ‘standard’ account ought to cover more ground than the assessment of justifiability of such interferences on the basis of ‘necessity in a democratic society’, and the Court’s decision in Leroy would indicate as much. The variety of ‘factors’ taken into account by the Court in establishing the existence of a ‘pressing social need’ for the respective restriction, e.g. the existence of a political dimension;¹⁰⁹ the medium of the expression¹¹⁰ and its form;¹¹¹ time of the publication;¹¹² possible effects of the speech in the future;¹¹³ the identity of a speaker (e.g. an association legally operating to advocate for the ideas in question¹¹⁴); spread of the publication;¹¹⁵ ‘the problems linked to the prevention of terrorism’;¹¹⁶ or popularity of slogans,¹¹⁷ would also speak against the determinacy that the strict ‘incitement’ standard seeks to provide.

It is not, however, the choice of such factors in a specific case or their assessment that minimises the relevance of the categorisation of glorification, but that the engagement with such factors serves as a proxy which allows the Court to identify, if any, the interests engaged by the expressive conduct in question. That is, any such analysis first establishes a relationship between two prima facie competing principles, where both must be worthy of protection. In the Court’s practice, this can be found in its holding, for example, that ‘where the views expressed do not comprise an incitement to violence … [states] cannot rely on protecting territorial integrity and national security, maintaining public order and safety, or preventing crime, to restrict the right of the general public to be informed of them’.¹¹⁸ In other words, where an act of expression does not meaningfully engage an interest that is worthy of protection, there exists no reason to interfere with it.¹¹⁹ This element of the Court’s engagement with the facts is not necessarily explicit and should be distinguished from the analysis of whether the inference pursued a ‘legitimate aim’. The latter, in the Court’s practice,¹²⁰ is a weak requirement that can be described as a check on the overall directionality of the interference towards the stated aims. The identification of the competing interests serves a different purpose, often carried out in the background, that of a check on whether one interest (freedom of expression) meaningfully engages another.

What includes that ‘another’ interest is the second relevant mechanism, one of excluding wrong reasons for the restriction of rights.¹²¹ Where freedom of expression is concerned, such ‘wrong’ reasons for the interference would include concerns that the speech in question is distasteful, wrong or otherwise objectionable.¹²² In the Court’s terms, for instance, ‘the fact that statements are made by somebody who is considered to be an outlaw cannot in itself justify an interference with the freedom of expression of those who publish such statements’.¹²³

When applied to glorification offences, these two functions then must rule out the restrictions on utterances that fail to engage a competing interest (‘national security’, ‘public order’) and restrictions imposed simply due to such utterances’ objectionable character. Since ‘glorification-as-glorification’ offences are purposefully broad in their scope, they are likely to capture instances where restrictions fail to meet either of these ‘standards’ by, for instance, allowing for the prosecution of inconsequential or low-impact utterances or expressions that only seek to provoke a reaction and not to encourage or incite. Then, these two ‘requirements’ would be expected to render such offences closer to ‘glorification-as-incitement’ by restricting their use to instances where both are met.

To illustrate, consider the Court’s findings in Yavuz and Yaylali v Turkey, Stomakhin v Russia and Müdür Duman v Turkey. The applicants in Yavuz and Yaylali¹²⁴ were prosecuted for propagandising in favour of a terrorist organization on account of their shouting of slogans during a demonstration in protest of the killing of members of an armed organization that occurred during a clash with government forces. The slogans included, among others, ‘the murderous state will be held accountable’, ‘the martyrs of the revolution are immortal’ and ‘we have paid the price, we will make it pay’.¹²⁵ The Court’s finding of a violation of Article 10 is primarily driven by the applicants’ speech constituting criticism of the authorities that neither sought nor was likely to encourage violence.¹²⁶

The applicant in Stomakhin v Russia¹²⁷ was convicted for his writings, whose contents varied significantly—from criticism of the Russian Government’s actions in the context of the armed conflict in the Chechen Republic, to ‘generalised negative statements’ characterising ethnic and religious groups, to statements in support of acts of violence perpetrated by the Chechen separatists—and the Court addressed them separately.¹²⁸ Some of the statements that the Court placed in the latter group¹²⁹—the one of most relevance here—included acts of speech that can be reasonably described as glorification of terrorist violence: in his monthly newsletter publication, the applicant described the 2002 hostage-taking at the Moscow theatre as ‘the action of … heroic Chechen rebels’¹³⁰ and the 2004 Moscow metro bombing as ‘justified, natural and lawful … Chechens have a moral right to blow up everything they want in Russia, after what Russia and Russians have done to them …’.¹³¹ Other texts that the Court included in this category appear to be qualitatively different.¹³² For example, they included a headline mentioning ‘national liberation struggle of the Chechen people against the colonial expansion of Russia’,¹³³ attribution of titles like ‘President’ and ‘Commander-In-Chief’ to separatist leaders’ names¹³⁴ and others.¹³⁵ The paragraph following the listing of extracts in question opens with the Court stating that they ‘promote, justify and glorify terrorism and violence’ in romanticising and idealising the cause of the separatists and their leaders while stigmatising the other party to the conflict.¹³⁶ Analysis continues, and the initial characterisation is repeated prior to the analysis of the reasons adduced by the domestic courts: ‘the impugned statements went far beyond the acceptable limits of criticism and amounted to glorification of terrorism and deadly violence’.¹³⁷ The Court then highlights that the domestic courts characterised the applicant’s statements in part as justification and glorification of terrorist violence¹³⁸ and finds their conclusion valid as it ‘is prepared to accept that the impugned statements … incited hatred against the members of the federal armed and security forces and exposed them to a possible risk of physical violence’.¹³⁹ This allowed the Court to find that the interference with the applicant’s freedom of expression met a ‘pressing social need’ with regard to these acts of speech.¹⁴⁰

The factual backgrounds in the two decisions are similar in that the acts of speech in both serve to communicate a degree of support towards a terrorist group—at the lower end of the spectrum with slogans like ‘the martyrs of the revolution are immortal’ in Yavuz and Yaylali and higher up the scale with the justification of specific acts of terrorist violence in Stomakhin—that is sufficient to code them both as glorification of terrorism.¹⁴¹ The relevant difference lies in the existence of a competing public interest, with the relevant statements in Stomakhin engaging it as they ‘incited hatred against the members of the federal armed and security forces and exposed them to a possible risk of physical violence’,¹⁴² and lacking in Yavuz and Yaylali as their slogans ‘amounted to a criticism of the acts committed by the official authorities but did not encourage the use of violence or armed resistance or uprising’ and unlikely ‘to promote violence by instilling a deep and irrational hatred towards identified persons, in particular members of the security forces and their families’.¹⁴³

Müdür Duman v Turkey further illustrates an interference that failed to engage a competing interest. The applicant was convicted for ‘praising or condoning’ acts publishable by law on account of ‘illegal publications and flags and symbols of the PKK’ as well as ‘pictures, articles and books pertaining to Mr Öcalan’ found by the police in the branch office of a political party that the applicant was the director of.¹⁴⁴ The Court disagreed that the presence of such materials in the office can be properly construed as an expression of support for unlawful acts since the authorities failed to show that the materials in question advocated violence,¹⁴⁵ thus finding a violation of Article 10. In other words, there was no competing public interest that the mere presence of materials engaged.

This serves to highlight that the Court’s approach to categorisation of glorification is not, and need not be, necessarily determinative in decisions concerning acts of glorification. By establishing the existence of a competing public interest and reviewing the admissibility of reasons for restrictions on such forms of expressions, the Court effectively raises the standards above those of ‘glorification-as-glorification’: it is necessary but not sufficient for an act of speech to simply constitute glorification of violence in order to be justifiably restricted as such, with other conditions needing to be met. Or so it should. As the following section will demonstrate, the effects of these functions are negated by deference afforded to domestic authorities in questions of the impacts and consequences of one’s utterances.

C. Deference in Review and the Impact of Expression

The Court’s jurisprudence goes beyond the decisions discussed so far and the outcomes of the other decisions would count against the restraining effects of the Court’s adjudicatory methods described above. Rather, at least insofar as object-level inconsistencies are concerned, what appears to bear some explanatory power for the apparent and controversial divergence in outcomes is the Court’s deference to the domestic authorities on the issue of the expression’s impact. In particular, it is the Court’s deferential acceptance of the domestic authorities’ assertions concerning the dangerous effects of instances of glorification that drives the inconsistencies, with a focus on one recurring element in such assertions, that of the recency of violence.

The term ‘deference’ here is not equivalent to the notion of ‘margin of appreciation’. Indeed, as far as the Court is concerned, deference is often understood to be a matter of ‘margin of appreciation’, a doctrine that formalises the Court’s practice of allowing national authorities a degree of discretion in applying the Convention.¹⁴⁶ The content of the doctrine, however, remains quite ambiguous,¹⁴⁷ with its various accounts attributing it differing functions. Such functions include, among others, addressing the discretion afforded to national authorities in conceptualising the relationship between individual freedoms and collective goals and regulating the intensity of the Court’s review;¹⁴⁸ serving as a tool assigning an appropriate standard of proof;¹⁴⁹ or being a tool of second-order reasoning and assigning weight to the reasons provided by the national authorities on the basis of considerations external to such reasons.¹⁵⁰ At the same time, a recent quantitative study finds no relationship between the invocation of the doctrine and actual deference to national authorities, but only between the doctrine and outcomes of decisions involving matters with no settled case law.¹⁵¹ Similar findings have been made in qualitative contributions.¹⁵² Then, equating the term ‘deference’ with ‘margin of appreciation’ would require an account of the latter. This not a task to be undertaken here. Rather, in what follows, deference is understood to relate to broader, unspecified and implicit tactics guiding the intensity of review and its variation within specific decisions, e.g. choices made with regard to the examination of some assumptions made by the authorities, but not others, or the choices to assert the authority over the justifiability of applicant’s punishment but not over the interference itself, or vice versa. Admittedly, this account of defence is likely to fall within the notion of margin of appreciation on most, if not all, of its accounts. Still, it is this lower-level deference—as opposed to higher-level doctrine of margin of appreciation—that appears to properly explain the diverging conclusions across relevant decisions of the Court, with some lower-level assumptions tested and rebutted and other ones accepted unexamined. To suggest that this is driven by the reliance on the doctrine of margin of appreciation is to offer a high-level explanation to what appears to be a lower-level phenomenon. Instead, a lower-lever account of deference, as suggested here, explains the divergence of the outcomes at a level at which it appears to be caused.

Not all relevant decisions are subject to complicated deference-granting practices. Indeed, some are ‘easy cases’, with their facts describing interferences that exist on the less controversial side of the spectrum. These decisions demand little deference from the Court and the outcomes are equally ‘easy’ to justify, whether they result in the Court finding a violation or not.

It is in the ‘hard cases’ where deference-giving is primarily employed. ‘Hard cases’ concern instances in which the conclusion about the justifiability of the interference is not relatively straightforward in the sense of clearly signalling that an interference is abusive or disproportionate, or, to the contrary, appropriate and uncontroversial; neither a finding of a violation or a finding of no violation would necessarily be unacceptable given the complexity of factors shaping a case. A specific issue on which the Court appears to grant it in glorification-related cases, and one that appears to drive their outcomes, is the expression’s impact. It has not been ignored in the post-Leroy criticism, but it is the overall ‘vagueness’ of the requirement of necessity in a democratic society that appears to have been its central element, with strict adherence to a ‘standard’ incitement approach being the solution.¹⁵³

The following subsection will first briefly address the decisions that fall into the category of ‘easy cases’. The subsequent one will address deference-giving in accepting the assertions about the effects of the recency of violence in the ‘hard cases’. What follows is meant to be a descriptive exercise; the question of the justifiability of the Court’s tactics of deference is outside of this contribution’s scope.

i. Non-deferential review in `easy cases'

The relevant decisions falling in the category of ‘easy cases’ include Altintaş v Turkey, Mehdi Tanrikulu v Turkey (No 2) and Erkizia Almandoz v Spain. What allows for them to be categorised as such is that their underlying facts demanded no deference from the Court, which explains their outcomes.

In Altintaş, the publication edited by the applicant had a piece discussing the 1972 Kizildere hostage-taking incident perpetrated by members of violent revolutionary organizations to stop the execution of death sentences against the groups’ leaders. The negotiations with the security forces failed and following a fire exchange, the militants executed the hostages. All but one hostage-taker were killed as the armed forces stormed the house.¹⁵⁴ The title of the article described one of the hostage-takers ‘and his friends’ as still living ‘like the idols of youth’, with the incident presented as an assault of ‘a group of young revolutionaries’ ‘who wanted to prevent the execution’ ‘of their friends’ by carrying out the kidnapping, but who were ‘massacred without achieving [their goal]’; they ‘still live as the idols of youth’.¹⁵⁵ The applicant was prosecuted and found guilty of the offence of glorification of crime and sentenced to a fine.¹⁵⁶ In analysing whether the interference was ‘necessary in a democratic society’, the Court first points to its authority to analyse the writings in the case on its own and not only rely on the reasoning provided by the domestic courts, ‘in particular when it is clear and manifest that the disputed remarks can only be qualified as hate speech, apology for violence or incitement to violence’ as such forms of expression do not attract the benefit of protection.¹⁵⁷ It then proceeds with its analysis of ‘the terms used in the disputed article, the context of its publication and its capacity to cause harm’,¹⁵⁸ finding that the piece was published in a ‘tense social context’ (on the occasion of the incident, distributed in the region where it took place and against the background of commemorative events organised by groups sympathetic of the hostage-takers),¹⁵⁹ that the analysis offered in the piece constitutes ‘an apology or, at the very least, a justification of violence’ (the approving terms describing the hostage-taking and qualification of the abductors’ deaths as a ‘massacre’),¹⁶⁰ and that the risk of violence inspired by the piece ‘should not be minimised (as it ‘could encourage or incite certain young people, in particular members or supporters of certain illegal organizations, to commit violent acts similar to those committed … so that they too become “the idols of youth”’).¹⁶¹ No violation of Article 10 was found.

The decision in Mehdi Tanrikulu concerned the prosecution of an editor for stories that described ‘A.Ö.’, ‘the leader of the illegal armed organization PKK’ as ‘leader of the Kurdish people’, the PKK as ‘Kurdish people’s liberation movement’,¹⁶² and its members presented in various heroic terms.¹⁶³ The applicant was found guilty of the offense of ‘propaganda in favour of an illegal organization’ and sentenced to a prison term of seven years and six months, the execution of which was then suspended for three years.¹⁶⁴ The Court examined the issues of the publication containing the writings in question separately, finding a violation of Article 10 with respect to the pieces that described Öcalan as ‘leader of the Kurdish people’ (no reason to depart from ‘numerous cases against Turkey’ establishing that this characterisation does not incite violence on its own),¹⁶⁵ and with respect to one that described the PKK as ‘Kurdish liberation movement’ since the Court found that the article in question contains no reference to PKK.¹⁶⁶ The conclusion concerning these two issues is primarily driven by the failure of the domestic authorities to provide ‘relevant and sufficient’ reasons justifying the interference.¹⁶⁷ However, the Court found no violation with respect to the second category of published articles, ones that included statements made by PKK leaders and those titled in the fashion of ‘The Kurds will defend themselves’, ‘Pioneering role of the PKK and popular heroism’ and ‘The week of heroism is an epic of freedom’.¹⁶⁸ The two relevant statements of the group’s leaders were: ‘as long as the attacks continue, the Kurds will defend themselves by all the means’ and ‘if they continue to attack the Kurds, [the latter] have the right to defend themselves’; further, the articles presented the group’s activity in various laudatory terms.¹⁶⁹ The Court found that the leaders’ statements can only be read as an attempt to incite violence;¹⁷⁰ ‘if the domestic courts did not take into account the intention of the applicant, the Court cannot overlook the potential impact of such messages containing glorification of violence in a politically sensitive region’.¹⁷¹ The Court then agreed with the domestic authorities that these writings ‘presented a concrete danger to public order’.¹⁷²

Some aspects of the decision in Erkizia Almandoz have been discussed in the preceding sections. Of most relevance here is that the decision of the domestic courts in upholding the applicant’s conviction for a glorification statute appear to have been driven primarily by the overall setting of the event where the speech was delivered—a venerating tribute of a ‘deceased terrorist’ ‘who was known only for his notable role within ETA’¹⁷³—and only secondarily by the applicant’s ‘deliberately ambiguous’ call to reflect on choosing the most appropriate path, ‘the path that will do the most harm to the state: the path that leads this people towards a democratic scenario’ or his shouting ‘Long live!’ to the name of the celebrated individual.¹⁷⁴ As noted above, the ECtHR found that the speech did not intend to incite violence;¹⁷⁵ rather, it found that the applicant’s speech sought to encourage reflection towards a democratic path, and neither the risk of ETA violence that existed at the time nor the dramatic arrangements of the event were to justify the conviction.¹⁷⁶ Thus, the majority found a violation of Article 10.

These three decisions illustrate the instances where the reasons for the interference are not subject to much controversy, whether they succeed (Altintaş and with regard to the second category of statements in Mehdi Tanrikulu) or fail (Erkizia Almandoz and the first category in Mehdi Tanrikulu) at that task. The Court need not defer to the domestic authorities in reaching its conclusions. Altintaş is exemplary of the sort of expression targeted by glorification offences, and the article’s publication against the background of commemorative events held by groups sympathetic to the hostage-takers is a valid reason in favour of the interference. The group of relevant statements made by PKK leaders in Mehdi Tanrikulu appear to be communicated in a form that is difficult to read as something other than encouragement of violence. Finally, the finding of a violation in Erkizia Almandoz is the least controversial one as the applicant’s speech, however ambiguous, does not seem to justify or appeal to violence in any form.

ii. Deference on the impact of expression in `hard cases'

The decisions in Z.B. v France, Rouillan v France and Jorge López v Spain are the ‘hard cases’, with the interferences imposed in a less clearly (non-)justifiable manner, and thus with more deference expected from the Court.

Z.B. is the first of two decisions concerning the French glorification offence delivered after Leroy.¹⁷⁷ The applicant gifted his three-year-old nephew, a boy born on 11 September 2009 named Jihad, a T-shirt that read ‘I am a bomb!’ on the front and ‘Jihad, born on September 11’ on the back and asked that he wears it to his preschool.¹⁷⁸ The applicant later explained that this was intended in jest.¹⁷⁹ The text on the T-shirt was seen solely by the preschool director and another adult responsible for dressing the boy, but invisible to others.¹⁸⁰ The applicant and his sister were found guilty of the charge of glorification of ‘crime of intentional attacks on life’.¹⁸¹ The Court’s analysis of the necessity of the interference is carried out primarily through the pronouncements of the domestic authorities. The pertinence of the threat of terrorism in France was one contentious issue of relevance here. In the proceedings before the Court, the government submitted that the applicant’s conviction was handed down ‘in the global context of terrorist threat’ but also took account of the particular circumstances of involving a child in a school: the T-shirt was worn on 25 September 2012, months after a series of attacks were perpetrated in March 2012 ‘in the name of the same ideology that inspired the attacks of 11 September 2001’.¹⁸² One of the attacks in March 2012 targeted a school, ‘and the applicant was convicted less than a year later’.¹⁸³ Therefore, it is this context of the threat of terrorist violence that is of relevance, not the long period between the 11 September attacks and the day the T-shirt was worn.¹⁸⁴ Similar arguments concerning the recency of the March 2012 attacks were made in the domestic proceedings,¹⁸⁵ and the Court found no reason to disagree: ‘in the view of the terrorist ideology underlying’ the 11 September and March 2012 attacks, the passage of time was not likely to attenuate the impact of the message, and nor does the fact that the applicant has no links to terrorist groups or does not subscribe to a terrorist ideology.¹⁸⁶ The Court then finds that the limited visibility of the message—two adult staff of the school who discovered it while dressing the child—was to be negated by the fact that the applicant did not deny having requested that the T-shirt is worn at school or wanting to share his message.¹⁸⁷ With respect to the statement’s impact, the Court deferred to the arguments made by the government in domestic proceedings, where ‘the emotion and the tensions caused by this message’ were underlined ‘to demonstrate its violence as well as the danger for social peace that messages associating a tragic event with a cause for celebration could constitute’.¹⁸⁸ A finding of no violation of Article 10 is reached unanimously.¹⁸⁹

Rouillan is the second of the two post-Leroy decisions addressing the French glorification statute.¹⁹⁰ The applicant in the case is a former member of the Action Directe—a far-left terrorist group active in France in the 1980s—who was convicted for various terrorism-related offences and sentenced to life imprisonment but conditionally released in 2012 subject to obligations of refraining from disseminating his works relating to his offences or publicly discussing them.¹⁹¹ In February 2016, he gave an interview broadcast on a local radio station and published online.¹⁹² The journalists interviewing the applicant presented his background, and a part of their talk was dedicated to the state of emergency introduced in France after the November 2015 terrorist attacks; in their discussion of ‘the phenomenon of Islamist radicalisation in prison and [the applicant’s] exchanges with radicalised prisoners during his own detention’, the applicant expressed disagreement with ‘developing anti-terrorist clichés’ that describe the perpetrators as cowards, continuing to say that he ‘found them very brave, they fought bravely, they fight in the streets of Paris, they know that there are two or three thousand cops around them’.

Often they don't even plan their exit because they think they will be killed before they finish the operation. … So, there you go, we can say we are absolutely against their reactionary idea, we can go and talk about a lot of things against them and say it was idiotic to do that. But not to say that it is [these] kids who are cowards.¹⁹³

The applicant was then prosecuted for apologie of terrorism; the first instance court noted that the offence does not require the speaker to support the act of the person in question directly, but ‘must be understood as constituted by any remarks aiming at valorising, justifying, excusing a terrorist fact and must consider the context in which the remarks or acts are carried out and the personality of the speaker’.¹⁹⁴ Given the recency of the November 2016 attacks and the applicant’s own background, he was found guilty of committing the offence.¹⁹⁵ The court of appeal ruled that the provision only allowed the applicant to be prosecuted as an accomplice (with the broadcasters being the primary offenders) and found him guilty of complicity in the commission of the offence.¹⁹⁶ In proceedings that followed, the Conseil constitutionnel found the provision constitutional.¹⁹⁷

Preceded by a finding that cases involving statements that ‘glorify violence and thereby indirectly incite its use’ are afforded a broader margin of appreciation, as are statements that do so directly,¹⁹⁸ the Court establishes that the applicant’s remarks were made in the context of a debate of general interest.¹⁹⁹ This is followed by a restatement of the findings made by various national courts in the litigation, with the focus on the courts’ emphasis on the recency of the November 2015 attacks and the applicant’s own past involvement in political violence.²⁰⁰ An aspect of the reasoning offered by the first instance court that was not referred to in the preceding parts of the judgment is presented: the court emphasised that the applicant himself admitted that the radio station broadcasting the interview was ‘listened to by many young people in working-class neighbourhoods in Marseille’ and ‘that even if his intention was to encourage membership in extreme left-wing groups, he admitted that these listeners constituted a fragile public easily seduced by the discourse of supporters of radical Islamism that could lead to terrorist actions’.²⁰¹ The Court’s own finding is that the applicant’s remarks were made at a time when ‘the turmoil caused by the deadly attacks of 2015 was still present in French society’ and when ‘the level of the terrorist threat remained high’, and were made by means that made it likely to reach a large audience;²⁰² in these circumstances, ‘statements at issue must be regarded, in view of their laudatory nature, as indirectly inciting the use of terrorist violence’ and therefore, the Court saw no reason to depart from the reasons adduced by the first instance court, that were then accepted in appellate and cassation proceedings.²⁰³ Where the Court disagreed with the domestic authorities was in the sufficiency of the reasons provided for the imposition of a severe penalty against the applicant, finding a violation of Article 10 on that basis.²⁰⁴

The final and the most recent decision of relevance is Jorge López.²⁰⁵ The applicant is a rapper who was convicted under the Spanish glorification offence on account of some of his lyrics positively referring to Grupos de Resistencia Antifascista Primero de Octubre (GRAPO)—a left-wing terrorist group—and ETA. Relevant passages in the songs included, e.g. ‘enjoy imagining a terrorist attack in the Congress’, ‘I’m back but I had not left, I was planning to blow up the Valle de los Caídos [Franco’s grave]’, ‘My heroes are not the capos [mafia bosses], my heroes are the GRAPOs’, ‘It’s legitimate as expropriating at Carrefour / entering the Popular Party headquarters with a gun and [sound of a shot]’, as well as the applicant saying ‘when I speak about hip-hop I’m talking about armed fight’ at a concert ‘accompanied by the image of a graffiti reading “Let the GRAPO come back!”’.²⁰⁶ Upholding the applicant’s conviction, the appeals court found that

the content of the songs is unequivocally aimed at praising terrorist organization GRAPO, its members and activities, indirectly inciting terrorist violence, like that perpetrated by ETA, identifying musical performance with the language of hatred … We understand that the lyrics ... go beyond the expression of support to political objectives or ideological statements, being undoubtedly a justification of violent means and an invitation to the use of terrorist methods, representing terrorism as deserving of praise ….²⁰⁷

It also pointed to ‘the dangers in disseminating such message to young audiences, who due to “their lack of full knowledge of the [past] terrorist events, in situations of ... economic precariousness, or for similar reasons” could find it appealing and fall to the incitement’.²⁰⁸

Importantly, GRAPO and ETA are both defunct, with the decision stating that the last deadly attacks committed by GRAPO took place in 2006, followed by arrests of some of the group’s members in 2007 and with 18 individuals being imprisoned at the time of the proceedings against the applicant in 2018, but with no statement made by the group concerning its dissolution.²⁰⁹ In turn, ETA announced a ‘definitive cessation of its armed activity’ in 2011 and gave up its weaponry in 2017, with 234 of its members being imprisoned in 2018.²¹⁰ It appears that the court of appeals addressed this only briefly, suggesting that ‘it shouldn’t be forgotten that [ETA and GRAPO] had not disappeared and that some of the crimes committed were still pending to be judged’ and that the ‘suffering of many victims and their relatives should also be taken into account’.²¹¹ This was further addressed by the Supreme Court (that dismissed the applicant’s cassation appeal) by ‘recalling that the traces of ETA and GRAPO acts were still active in many victims and they were remembered as a traumatic experience in the Spanish society’.²¹²

The Court’s assessment of the interference in the case is explicitly procedural, reviewing the decisions of the national authorities and determining whether the reasons they adduced are ‘relevant and sufficient’.²¹³ In doing so, it addressed three following aspects of the applicant’s conviction. First, with regard to the defunct nature of the terrorist groups in question, the Court admits that their latest attacks date ‘some years prior to the events’ but highlights that ‘one could not ignore that both had carried out terrorist activities in Spain for decades, causing numerous deaths and injuries’, with some criminal proceedings against the members still pending and remaining in the focus of the society and the media, concluding that the past terrorist attacks were ‘fresh in the country’s collective mind, justifying an enhanced degree of regulation of statements relating to them’.²¹⁴ Second, the Court agreed with the domestic courts’ finding that the lyrics justified violence by ‘directly suggested injuring or killing politicians, judges, security forces, rich, royal family and those perceived as ideological opponents’ and therefore falling outside the acceptable limits of criticism.²¹⁵ Third, the Court also deemed acceptable the domestic courts’ findings on the songs’ capacity to cause harmful consequences, with the music ‘conveyed to the wide audience through a YouTube channel, a Facebook profile or in concerts’.²¹⁶ With the applicant’s punishment deemed proportionate, the complaint was found inadmissible.²¹⁷

What unites the three decisions is the Court’s deferential acceptance of controversial assertions for the interference that, in significant part, appear to be rationalisations, rather than good faith justifications. One key issue across the three decisions is the recency of terrorist violence and its overall threat at the time of expression. In Z.B., the controversial T-shirt was worn several months after a series of terrorist attacks were perpetrated in France, including one targeting a school—a fact that appears to be at the core of the Government’s submission.²¹⁸ What remains unclear is how this relates to the aim of ‘prevention of disorder or crime’ (as submitted by the Government and accepted by the Court)²¹⁹ in the scenario where the text in question could only be read by the adult staff of the kindergarten. The concern with the possible effect of the T-shirt on violence, however improbable or indirect, is lost in both the Government’s justification or the Court’s engagement with it, as both invoke the past violence only to assert that the significance of the shock experienced by the adults exposed to the T-shirt—as done by the Government,²²⁰ or to highlight the ‘significance of the message’—as is done by the Court first in the abstract, but then followed by deferral to the national authorities’ claim concerning the T-shirt’s effect on the ‘social peace’.²²¹

The recency of violence was similarly central in Rouillan.²²² Yet, if the fact of both the March 2012 attacks in France and the 11 September attacks in the United States (invoked by the T-shirt) were guided by the same ideology was deemed relevant in the Z.B.,²²³ it was the recency of violence as such that was accepted in Rouillan. The applicant’s far-left outlook was irrelevant since he spoke positively of recent terrorist violence, regardless of his support for the ideology claimed to be behind it.²²⁴ However, can the concern with the recency of ideological violence be of relevance in the assessment of possible effects of one’s non-ideological apologie of violence as such? Would the applicant’s expression be interfered with if the violence he discussed was not recent? A negative answer would require that such a determinative role of time is further justified; an affirmative one would indicate that the recency as such is of insufficient importance and that its invocation should have little or no bearing on the conclusion reached. The reasoning of the domestic authorities in the case takes the opposite form: the recency is constantly emphasised but its relevance is never explained, and it is the applicant’s personality that appears to be the real factor driving the interference. That is, he would have been liable to be prosecuted regardless and the invocation of the recency of violence can serve as a justificatory reason only superficially.²²⁵

The facts in Jorge López further illustrate the fallible role of recency as the case deals with what must be a far end of a possible temporal spectrum, with the violent groups being glorified having been defunct for more than a decade. This was deemed irrelevant on account of numerous deaths and injuries caused by the groups over prolonged periods in the past and because criminal ‘proceedings against the members of both terrorist organizations were pending before the Spanish courts and remained in the focus of attention of the society and the media’ and therefore ‘still fresh in the country’s collective mind, justifying an enhanced degree of regulation of statements relating to them’.²²⁶ This approach does find some support in the Court’s case law,²²⁷ but it is notable that it effectively renders the factor of the recency of violence meaningless as it is replaced with a much broader factor of it being in the focus of society and the media.

None of this need imply that recency of violence should or should not be a factor relevant to determining the justifiability of counterterrorism-related interferences with freedom of expression or that the outcomes of these decisions are incorrect. Rather, this is to demonstrate how one element common to the three decisions is utilised in deferential review. For a court to defer is to accept the superior democratic legitimacy or epistemic position of the primary decision-makers and engage with their arguments only superficially. It appears that simply some discussion of the temporal aspects of one’s expression by the domestic authorities is sufficient, however unpersuasive or internally inconsistent. The lower the intensity of the review, the lower the strength of its diagnostic power of correctly identifying the instances of unjustifiable interferences:²²⁸ it might be capable of doing so in cases where the interference is apparently justifiable (Altintaş, Mehdi Tanrikulu) or unjustifiable (Erkizia Almandoz), but is liable to unsatisfactory in those where the lines are not as clearly drawn (Z.B., Rouillan and perhaps to a lesser extent in Jorge López). Still, it is not to be ignored that the more ‘controversial’ cases appear to be such at least in part because of the controversial statutes that underly the interferences. Where an apologie statute operates in the ‘glorification-as-glorification’ mode, it is liable to result in the prosecution of expression that might not be worthy of it.

5. Conclusion

If such can be said to exist, the Court's approach to glorification of terrorist violence is then mostly at odds with the constraints suggested in the objections to unconstrainted criminalisation of glorification. Or so it would appear to be if one were to judge only by the outcomes of the relevant decisions. In situating the Court’s position in relation to the accounts critical of glorification, this article sought to demonstrate not just the differences but the similarities between the different approaches. Although different in their underpinnings, both the ‘standard’ and the ‘new’ objections to criminalisation of ‘glorification-as-glorification’ conclude that only incitement as such is a legitimate object of criminal law. Given the broad coalition of actors and legal instruments that, with some variation, find the ‘glorification-as-glorification’ statutes to be unjustifiable—and one that includes not only human rights groups and the Special Rapporteur on the promotion and protection of human rights and fundamental freedoms while countering terrorism but also the claims articulated by the UN Secretary-General and the texts of the relevant EU Directives—one could expect the Court to be closer aligned with that position. As was suggested, this divergence might at least in part be explained by the Court’s adjudicatory method and the objects of its review, neither of which allows for explicit findings of statutes’ incompatibility with the Convention as such. Yet, both are capable of yielding results similar to those that would have been generated if such statues were indeed deemed per se incompatible with the Convention. It was argued that this can be achieved by scrutinising the existence of a public interest competing with expressions that glorifying violence as well as the reasons advanced for the imposition of respective interferences. As far as the Court’s past decisions are concerned, the inconsistencies in the outcomes of the decisions appear to be explained by the Court’s deferential treatment of national authorities’ claims concerning their expectations of the impacts that such expression might have on future violence.

Acknowledgements

An early draft of this paper was presented at a workshop ‘Old and new threats to freedom of expression: Can the European Court of Human Rights meet the challenges?’, held online in June 2021 and organised by the Centre for Fundamental Rights at the Hertie School, the European Convention on Human Rights Law Review and the International Law and Human Rights Unit of the University of Liverpool School of Law and Social Justice. I thank Vassilis Tzevelekos for comments and feedback at the workshop. I thank Gloria Gaggioli for feedback that helped further improve the paper. This work was supported by the Swiss National Science Foundation grant number 176781.

Footnotes

Author notes