Dear Jonathan and Dominik,
I just had the very same thoughts. I'm not an expert of law either, but
technically speaking, the BORI Mahābhārata is not simply an edition, but a new text created by its editors between 1919 and 1966. The editors are, in this case, actually authors, who obviously transferred
their copyright to the still existing BORI. So unless an ancient and complete manuscript appears which contains the very same text as the BORI Mahābhārata (very unlikely, I would say), the BORI holds the copyright of its text. According to German law (mentioned
by Dominik), however, it does not 25 years have long gone past since the publication of the original edition. The co-owned copyright of Prof. Tokunaga (1994), too, would expire this year in Germany.
Of course, authors also have the copyright to transcriptions of their text just imagine someone would transcribe a talk you give and then publish it as their own text. I would argue that creating an electronic
transcription of a (copyrighted) Devanāgarī text isn't much different.
Best regards,
Dominik A. Haas