After the referrals in Mio [IPKat here and here] and USM Haller [IPKat here], another referral asking about the meaning of originality in EU copyright law has been made to the Court of Justice of the European Union (CJEU): it is the referral from Romania in Institutul G. Călinescu, C-649/23.

Unlike the Swedish and German referrals, the Romanian one has not been made in the context of a dispute concerning works of applied art (which is refreshing), but rather in relation to the protectability of a critical edition of a work.

The notion of ‘critical edition’ refers to the edition of an ancient, classical or other text produced as the authoritative version by comparing the variants, and accompanied by comments and the requisite critical apparatus. The author of a critical edition is required to ‘reconstruct’ the meaning of the text and convey the original message, given that such text is likely to be missing ‘pieces’.

Translated into copyright language: a critical edition is an example of derivative work.

Despite (or rather because of?) the academic subject-matter of the background proceedings, this new referral is likely to give rise to yet another seminal CJEU ruling. If I were to make a prediction, I would say that this Romanian case is going to make one of the most relevant CJEU copyright rulings ever issued.

In my view, this is so even for reasons beyond the question of originality. Indeed, in Institutul G. Călinescu, the CJEU will have the opportunity to clarify, once and for all, at least the following:

1. The treatment of derivative works under EU copyright, not from the perspective of an infringement claim (because those have been tackled already: read on), but rather from the perspective of the very protectability of works that stem from another’s work;
2. Whether a critical edition is the type of work that qualifies as original in an EU sense. In this sense, a somewhat similar – though dated (by that meaning: pre-Infopaq [IPKat here and here]) – case litigated both in France and the UK cannot but come to mind: Hyperion Records v Sawkins [IPKat here], concerning copyright protection of a ‘reconstructed’ baroque music score.

Let’s break the referral down before discussing its potential relevance and implications, including for Artificial Intelligence (AI).

Background

Professor Dan Slușanschi was a Romanian classicist and specialist in Indo-European linguistics.

In 2001, the professor published a critical edition of Demetrii principis Cantemirii. Incrementorvm et decrementorvm avlae othman(n)icae sive aliothman(n)icae historiae a prima gentis origine ad nostra vsqve tempora dedvctae libri tres (‘History of the rise and fall of the Ottoman or Aliottoman court from the origins of the lineage to the present day, in three books’), based on the original Latin manuscript discovered at Harvard in 1984. A second (revised) edition followed in 2008.

The professor also prepared a Romanian translation of the Latin text.

In 2017, the Regional Court of Bucharest held that the defendants had infringed the professor’s moral right of attribution. The decision was upheld on appeal.A further appeal to the High Court of Cassation and Justice followed. This now centres on the assessment of originality of the critical edition. According to the defendants, the Court of Appeal misapplied CJEU case law. Specifically, the defendants question whether freedom and creativity may be truly exercise in 'reconstructing' a text so that is intended to be close as possible to the original (in the sense of initial) version.

The Court of Cassation thus decided to stay the proceedings and refer the following question to the CJEU:

Comment

As stated, this is going to be a key referral for the (further) construction – and de facto unification (!) – of copyright in the EU.

Derivative works under EU law

So far, the CJEU has tackled derivative works from the perspective of infringement, not copyright subsistence. Examples of cases concerning derivative works abound: some concern actual transformations, e.g., Painer (photo-fit) and Deckmyn (parody); others relate to incorporation, e.g., Pelham (music sampling) and Renckhoff (downloading and use of photograph).

When the CJEU decides Institutul G. Călinescu, it will be required to tackle the question of what makes a derivative work protectable and what ‘freedom’ and ‘creativity, both cumulative requirements under the EU originality test, mean in this context.

While there is little doubt that ‘translations, adaptations, arrangements of music and other alterations of a literary or artistic work shall be protected as original works without prejudice to the copyright in the original work’ (see Article 2(3) Berne), the CJEU has also indicated that the EU originality test requires more than skill, labour or effort. And this leads me to the second point:

What is originality?

Since BSA, the CJEU has been adamant (see, more recently, Brompton) that choices dictated by technical considerations, rules and constraints do not confer originality.

But when it comes to the work of a translator or a scholar like Slușanschi or Lionel Sawkins, is what they do purely ‘technical’?

If that was true and as an example, then any translation of a work would do, while we know that that is not true. There are objectively translations that are better than others: but is that so exclusively because of technical choices?

Similarly, the choice of words has an impact on the meaning: for example, Thomas Mann’s Der Zauberberg has been long known in Italian as La Montagna Incantata but, recently, the use of 'incantata' ('enchanted') has been questioned as a suitable translation, with 'magica' ('magical') being indicated as preferable to 'incantata'. Similarly, a title like The Cather in the Rye, translated in a literal way, would be incomprehensible to an Italian reader, with the result that JD Salinger’s famous novel is (creatively) known to the Italian public as Il Giovane Holden ('Young Holden').

As noted, the Sawkins case was litigated both in France and the UK: despite that, historically, the French understanding of originality as 'la marque d’un apport intellectual' has entailed a creative aspect that was not part of the old UK standard of labour, skill, and effort, courts on both sides of the Channel held that the work of reconstruction conducted by the musicologist would warrant protection.

All this, of course, does not imply that a faithful reproduction is protectable: for example, in line with Article 14 of the DSM Directive, a simple (verbatim) reproduction of a work is not going to be considered sufficiently original and, thus, eligible for copyright protection. See here, here, here, here.

Broader implications: AI-generated outputs and beyond (including right of adaptation)

The referral is also one to watch regarding the treatment, under EU copyright law, of AI-generated outputs.

If we leave the question of authorship aside and we focus on originality instead, then the guidance that the CJEU will provide in this case is going to be key also to determine the ‘originality’ of outputs created by or with the help of AI and their eligibility to be regarded as copyright works.

Indeed, if a large-language model is trained on existing content – by default – a lot if not all that it generates at a later stage is going to be derivative. This should not shock anyone, as human creativity is inherently derivative too: what is produced today borrows from and elaborates upon the creativity of those who came before us - nani gigantum humeris insidentes!

On a separate and final note, the case might also offer a chance to clarify once and for all whether the right of adaptation, which is relevant to derivative works, is de facto part of the broader right of reproduction under Article 2 of the InfoSoc Directive.

This right has been only formally harmonized in relation to databases (Article 5(b) of the Database Directive) and software (Article 4(1)(b) of the Software Directive). However, as I also discuss here, the CJEU has likely de facto recomprised such right within the broader right of reproduction. It is clear to me that the InfoSoc right applies to instances of both literal and non-literal copying (a clear example being the photo-fit at issue in Painer!) and, therefore, also to situations that, formally, would fall within the generally unharmonized right of adaptation. Yet, not everyone agrees.

In sum, this referral is one to closely watch. A classicist’ work might pave the way for a very meaningful series of legal developments: who would have though? 🤓 😅 Stay tuned!

[Originally published on The IPKat on 17 January 2024]