Private (Transnational) Power without Authority

On 9 September 2019 Facebook banned from its platforms all pages and profiles related to the Italian far-right organization “CasaPound”, for the violation of its Community Standard no. 12 (hate speech and incitement to violence). CasaPound – whose political program is a classic combination of totalitarianism, militarism and nationalism – is a relatively small but extremely active group, targeting with violent actions minorities in suburban slums, political opponents and sometimes left-wing journalists. On 11 December 2019, the Tribunal of Rome (ToR) adopted the precautionary measure ordering Facebook Ireland Ltd. to restore the pages and their content and to pay the losses. The decision raises significant issues in several respects and might serve as a model to courts beyond Italy.

Although many members have been convicted in past years, CasaPound as such is yet to be dissolved via existing anti-fascist legislation. It is therefore a “legitimate” actor in the Italian political landscape, regularly taking part in local and national elections.

Despite its poor election results, the political danger posed by CasaPound and similar organizations should not be underestimated. It results from their contribution to the normalization of fascist-like language and ideas, an ongoing transnational process in which social media play a crucial role. Given the features of contemporary political discourse, CasaPound’s very “existence” as a political actor of any significance depends on its +40,000 Twitter followers and – before the ban – 280.000 Facebook “likes”, without which it would probably be a little less than a politically-flavored criminal organization. However, and quite importantly, CasaPound hardly posted online explicitly fascist content, rather saving this kind of speech for other channels of communication.

The Tribunal’s line of reasoning

The first issue relates to the quite unconventional line of reasoning, which did not focus on the parties’ contractual obligations, but rather on those arising directly from art. 49 Constitution (“Any citizen has the right to freely establish parties to contribute to determining national policies through democratic processes”). Importantly, the ToR’s decision can only be understood in the context of the right to (collective) political participation, rather than the – admittedly overlapping but distinct – right to freedom of expression. Keeping in mind these starting points, the reasoning can be summarized in three steps.

1. Facebook has de facto reached a systemic role to the purposes of political participation under art. 49 Const.
   
2. Given such systemic role – and departing from a relatively well-established case law of the Supreme Court (e.g. judg. no. 31022/2015) qualifying Facebook as a mere “private messaging service establishing a network of relationship among people in the same system” – the ToR held that Facebook has a “special position” towards private individuals.
   
3. As a further consequence, besides the obligations arising from its contractual will, Facebook is bound by those arising from constitutional principles (principi costituzionali e ordinamentali). Therefore, although it did not explicitly refer to such doctrine, the measure constitutes an example of direct horizontal effect (unmittelbare Drittwirkung), whereby private actors are directly bound by constitutional duties, i.e. without resorting to constitution-oriented interpretation of private law or contractual clauses; or to the declaration of unconstitutionality of applicable legislation.

Having re-shaped the range of Facebook’s obligations, the ToR found that the complaint concerning CasaPound’s right to political participation had some prima facie ground (fumus boni iuris) and was potentially subject to irreparable damage pending ordinary proceedings (periculum in mora), thus granting the precautionary measure. Although adopted in a proceeding with limited evidence-collection and prima facie merits assessments – and the ToR itself stressed that the ordinary proceedings could well turn out differently – the decision still provided several interesting elements of reflection, especially as concerns the power relationship between Facebook and CasaPound.

Facebook in the eyes of the ToR: bound by the Constitution but no “keeper of the Constitution”

Already at the linguistic level, the opinion significantly switches from “Facebook Ireland Ltd.” (the procedural defendant) to “Facebook” tout court, somehow downsizing the relevance of corporate personality to the purposes of the decision. Put differently: The ToR’s language gives the impression that it is not Facebook as an individuated corporate person to infringe upon the constitutional obligation, but rather Facebook as a (business) enterprise, a transnational actor in control of a good amount of the communication flow relevant to political discourse and participation. Further, the ToR explicitly qualified the ban as a “sanction”, rather than a countermeasure to a breach of contract.

At a more substantial level, the decision can be framed as a review of the proportionality of Facebook’s conduct. According to the ToR, the contents shared by CasaPound did not reach a degree of gravity such as to justify the outright ban, i.e. a de facto exclusion from political debate: a more proportionate (and possibly legitimate) reaction would have been removing the single contents, with a more circumscribed limitation of CasaPound’s single exercises of freedom of speech.

Facebook’s argument, that the ban “sanctioned” the fact that CasaPound is a political organization intrinsically against the Constitution and human rights law, was also rejected. Neglecting such role of “keeper of the Constitution”, the ToR held that it is not up to Facebook to determine if CasaPound is a legitimate actor in the Italian political landscape, taking also into consideration that it has been active for more than ten years without being outlawed by competent Italian authorities. Admittedly, such statement goes in a direction opposite to that taken in Germany with the contested 2017 Network Enforcement Law (NetzDG) and in France with the bill no. 1785, designing systems whereby major social media platforms are tasked with – and potentially held responsible for – gate-keeper functions against hate speech and fake news on the Internet.

In a nutshell, the ToR “saw” in legal terms the private power and accordingly attached obligations to it, but did not recognize Facebook as a legitimate actor to autonomously take decisions with potentially far-reaching political consequences. The ToR thus denied Facebook’s authority – at least as long as it uses its power in a way deemed disproportionate. In this last regard, looking at Facebook as a global institution and its Standards as a transnational private regulation or even a semi-autonomous legal system, it could be argued that the decision features some elements of the Italian ‘counter-limits’ doctrine, the (in-)famous tool of management of inter-systemic relationship aimed at preventing external legal sources to be applied into the domestic system, when they risk to jeopardize the “fundamental principles of the constitutional order”.

The quest for a (transnational) horizontal effect and the need for supranational regulation

The ToR’s measure stands out for its bold application of (direct) horizontal effect. Indeed, while the horizontal effect is not a novelty in the practice of Italian courts, the Italian Constitution – contrary to arts. 1(3) and 9(3) of the German Basic Law or § 8(2) of the South African Constitution – has no express provision concerning the scope of application of constitutional rights. As a result, the horizontal effect has been applied in quite an unpredictable way, mainly on a case-by-case basis. The differences with the US functional equivalents – the “State action” and “public forum” doctrines – are especially striking, as even the most advanced application of these latter require some “color of State law” (see Knight First Amendment Institute v. Trump, 9 July 2019; Manhattan Community Access Corp. v. Halleck, 587 US __ (2019)). Certainly, the US solutions give better guidance to judges and more certainty to private individuals, but remain largely blind to social de facto powers that might be able to infringe upon constitutional rights just like State actors. Therefore, while the dangers to legal certainty should not be underestimated, such admittedly ambitious attempts must probably be welcomed, in an age where transnational private (especially economic) powers are more and more able to escape constitutional constraints.

However, the decision also raises the conflict-of-laws issues typical of any horizontal effect towards transnational actors. Firstly, the measure is to be enforced by Irish courts, and it cannot be excluded that an order to restore the access to a social media platform of a far-right organization may run contrary to the Irish ordre public under art. 45 Reg. 1215/2012/EU. Secondly, the order may give rise to overlapping and conflicting obligations under the laws of other States, e.g. the above-mentioned German NetzDG. Faced with conflicting impulses coming from different States, a transnational actor such as Facebook has to select and re-elaborate them in its internal law- and decision-making. This process inevitably results in a form of legal Darwinism, penalizing weaker States.

The ToR’s decision therefore highlights the need for a comprehensive hard law regulation (at least) at the European level, which would go beyond the piecemeal approach adopted by the EU so far with Dir. 2000/31/EC on E-Commerce; Framework Decision 2008/913/JHA requiring Member States to sanction by means of criminal law public incitement to violence or hatred; Dir. 2010/13/EU on Audiovisual Media Services; and lastly with the 2016 Code of conduct on countering illegal hate speech online, a soft law instrument of co-regulation, elaborated by the Commission together with four leading companies of the industry.

However, any form of supranational regulation, especially if restrictive of the rights of freedom of speech and to political participation, 1) would require a difficult political consensus, among States interpreting in extremely different ways their identities of “militant democracies”; and 2) must take into account existing human rights law. Therefore, it is also worth briefly assessing the ToR’s decision against this background.

Varieties of “militant democracies” and compatibility with international law

The obstacles concerning the reach of a consensus among different States on banning parties from political arenas emerge entirely, if one sees the ToR’s decision as a manifestation of the Italian (version of) “militant democracy”. The latter has been famously categorized by Niesen as a “negative Republicanism”, i.e. the self-definition of a democracy in contrast to a particular authoritarian past. Such “negative Republicanism” differs markedly from the abstract and quasi-universal German approach, labelled as “anti-extremism”, not just because it is past-oriented, but also because it assumes a relatively more fragmented political-ideological landscape, and thus a reactive rather than proactive attitude towards anti-democratic tendencies.

Such “negative Republicanism” is based on the idea that – instead of being expelled outside the political arena – extremist movements can be re-absorbed into a constitutional (“Republican”) consensus through the hidden but constant “normalizing effect” of democratic procedures. Right or wrong, such attitude gives further explanations as for why the Italian anti-fascist legislation led to the dissolution of a far-right organization only once (Ordine Nuovo in 1973); and for the stark rejection by the ToR of Facebook’s (self-attributed) “gatekeeper functions”. To sum up, the ToR’s decision does not run against the tenets of “militant democracy” but rather represents a specific implementation of them.

At the same time, the ToR’s measure cannot be said to infringe upon international law. After all, there is no positive obligation for States to generally prevent fascist or any other extremist organizations to access social media platforms. On the contrary, under human rights law the right to free speech can only be restricted according to narrowly defined parameters. Hate speech regulation must carefully weigh freedom of expression (art. 19 ICCPR) against prohibition of incitement to hatred (art. 20 ICCPR). Indeed, there is only a fine line between the two, visible from the “safety net” of freedom of expression that human rights instruments often include. The above-mentioned Framework Decision 2008/913/JHA or the Additional Protocol to the Cybercrime Convention, for instance, underline that they shall not interfere with freedom of speech standards. Where such a provision is not explicitly comprised, States have often brought forward reservations on its basis.

Of course, freedom of expression cannot be understood as to per se cover hate speech. Art. 10 ECHR, for instance, has traditionally been interpreted by the ECtHR as to comprise an internal barrier to incitement of hatred (see e.g. Perinçek v. Switzerland, 15 October 2015). Only very recently has art. 17 on the prohibition of abuse of a right been applied in the context of anti-Semitic speech (M’Bala M’Bala v. France, 20 October 2015). Overall, limiting the freedom of expression nonetheless “must remain an exception”, as even the Rabat Plan of Action on the prohibition of hatred puts it. For the specific case of online publications, the Human Rights Committee has also clarified in its General Comment No. 34 that this entails only “content-specific” removals being permissible. Further, from the case law of the ECtHR it emerges that, far from being obliged to restrict the overall-access of an extremist political organization to social media, States are only allowed to ensure that the respective postings are deleted (Delfi v. Estonia, 16 June 2015) in cases where single content constitutes unlawful hate speech or incitement to violence. Without the pivotal element of hate speech or “clearly unlawful” content the right to free speech is to be ensured (Phil v. Sweden, 7 February 2017; Magyar Tartalomszolgátatók Egyesülete and Index.hu Zrt v. Hungary, 2 February 2016). A positive obligation has recently been found to exist (Beizaras and Levickas v. Lithuania, 14 January 2020) but, once again, it concerned only the issue of investigating whether a posting might constitute hate speech, and not the overall access to social media.

After this brief excursus, it is fair to conclude that there is no incompatibility between existing international law and the ToR’s decision. This is especially true, if one considers its stark distinction between the right to freedom of expression concerning the single content/posting and the right to political participation concerning the overall access to social media platforms, as well as the related proportionality assessment of Facebook’s “sanction”. As formal, problematic and far-fetched as this distinction may be, it adds further elements of complexity to a decision that probably will not remain isolated, and whose line of reasoning might serve as a model to other courts worldwide.