An interesting post on Twitter by Gregory Crane of the Perseus project leads to some interesting finds:
The Scaife viewer build on work by @PonteIneptique. It would be easy to import the texts into gitlab.inria.fr/almanach/cllg/ from @ThesLingGraecae. We have chosen not to do so but there is no copyright at play. Here are the biggest 20 not yet in Perseus….
Thibault Clérice seems to be the main man here. Inria seems to be doing some very useful things. The Corpus Liberatum Linguae Graecae site at gitlab:
The CLLG Freed Corpus is a free XML-TEI corpus of Ancient Greek texts derived from the Thesaurus Linguae Graecae (TLG) E CD-ROM.
It has been transformed using Diogenes and enriched with Dapytains-compatible citation structures.The corpus is developed within the CLLG project, funded by the Programme Inria Quadrant at Inria.
The corpus is tested with HookTest.
With the interesting postscript:
Texts are distributed under French law, which holds that the texts of critical editions are not subject to copyright protection, as they do not constitute an expression of the editor’s personality (CA Paris, Pôle 5, 2ème Ch., 9 June 2017, no. 16/00005).
The Cour d’appel de Paris judgement is here, but I have not tried to read it. It seems to be a case where a publisher, Libraire Droz, sued another publisher, Libraire Classiques Garnier. Droz lost. AI Overview:
In CA Paris, Pôle 5, 2ème Ch., 9 June 2017, no. 16/00005, the Paris Court of Appeal confirmed that a publisher cannot claim copyright protection over public-domain medieval texts when the transcription involves no original creative input.
Lack of Originality: The court held that Librairie Droz did not perform a creative or transformative editorial adaptation on the underlying medieval texts. Transcribing, retranscribing, or applying standard paleographic methods (like fixing punctuation or abbreviations) to public domain material does not imprint the work with the author’s personality.
Copyright Infringement Rejected: Because the texts lacked the originality required to qualify as works of the mind (œuvres de l’esprit), the infringement (contrefaçon) and damages claims by Librairie Droz were dismissed.
Even more interesting is CoMMA:
…a large-scale corpus of medieval manuscripts produced through automatic text recognition. The corpus contains around 2.5b tokens drawn from more than 23,000 digitized manuscripts in Latin and Old French, harvested via IIIF. Unlike other resources, it is made of raw, non-normalized text enriched with layout analysis in various formats.
A PDF describing this is here. It does say:
For medieval manuscripts in Western languages, recognition accuracy is increasingly limited, not by models but by the availability of ground-truth data, with current error rates often below 10%.
10% is a big error rate, tho. Even 1% is painful in OCR. But it’s a massive step forward.
I see other projects which are bootlegging the Thesaurus Linguae Graecae. The TLG is run by the excellent Maria Pantelia, who relies on subscriptions to hire people to transcribe texts. So I have very mixed feelings here. The TLG is wonderful and necessary. It needs that revenue stream until the project is complete.
Much as I want to see everything available to everyone freely, I agree with Gregory Crane. I’m not going to do things that might injure the TLG. Nobody else should do so either. Let Dr Pantelia carry out her mission. The world will not end if we have to manage for a few more years until she’s done.
In general, if there are people out there producing stuff that promotes the classics, patristics, the ancient world, let them enjoy the income from their labour. Heaven knows there isn’t huge money to be made. A few PDFs for use by the poor people, of books that are sold at hundreds of dollars; that is one thing. That won’t affect the publisher revenue stream. But let’s not wreck projects like the TLG.
Update (26 Sep 2026): I learn from a twitter post by Gregory Crane that “the EU allows nations to copyright editions for up to but not more than 30 years.”
The regulation in question is article 5 of the EU Copyright Term Directive (Directive 2006/116/EC, originally 1993/98/EC) (Download PDF here). The regulation is optional, however. Relevant extract:
An AI search (caution!) tells me that only Germany – inevitably – and Estonia have implemented it. No doubt the German government – which seems to be completely subservient to the publishing industry – drew up this foolish clause.
