Mark Thakkar

New Wyclif polemic discovered! Only 250 words long, but not obviously detached from a larger work. It isn’t ascribed to him in the manuscript (no titulus or colophon) but the style is recognizably his. I’m posting it here because the content is eerily topical. #medievalLatin #medievalSky #genAI

Pre-lash: “The author is a software engineer who works extensively with AI agents, not a researcher, and has no training in astronomy.”

The FitMA team is (still) looking for a PhD-candidate with good Latin, a philosophical background, and a healthy curiosity for medieval manuscripts and digital editions. Don't procrastinate to apply! We might be looking for you! www.kuleuven.be/personeel/jo... #philsky #philosophysky #medievalsky

OK, since this thread has blown up a bit, I think it's worth clarifying in a new thread to what extent the 'GenAI is asbestos for the mind' analogy might hold up, and where it breaks down (as even useful analogies always do): 1/n

The same MS of Aristotle's Topics, with plenty of annotations, apparently ended up in a bookbinder's workshop. Plenty of bifolia were reused as pastedowns. (@ubleipzig.bsky.social, MS 376) resolver.staatsbibliothek-berlin.de/HSP000518E90...

The fallacy that University leaders pushing AI are falling into is thinking that the point of undergraduate research is to find things out. It isn't. What students find out is typically already generally known. The point of undergraduate research is to learn how to find things out.

I am finalizing the yearly update of my Supplementa of Aristoteles Latinus Codices. If you know (or own) manuscripts or fragments of Latin Aristotle texts that were not included before, now is the time to drop me a line!

My book ‘The Making of the OED’ was published on 25 Aug 2016, so I have missed the exact 10th anniversary; but I don’t care. I am going to indulge myself by quoting from the wonderful series of reviews I was lucky enough to get ten years ago. (Yes, I kept a file of them, of course I did.) 1/

Poe introducing his Marginalia (1844): “the chance and desultory comments of literary chit-chat … are not unfrequently “talk for talk’s sake,” … while the marginalia are deliberately penciled, because the mind of the reader wishes to unburthen itself of a thought”. 💯 www.google.co.uk/books/editio...

Flashback to when I witnessed an 'eminent' Oxbridge Prof give a totally wild lecture in which they presented slides arguing that 'global history' was more popular than 'taylor swift' despite having done google searches without having put these terms in quote marks... IFYKYK.

Lexicographer wanted at the TLL! Deadline end of August. Somewhat surprisingly, the advert doesn’t say how good your German needs to be – the TLL is based in Munich, and it’s in the nature of the job that you need to be in the office – but of course the dictionary itself is written in Latin anyway.

I have all the “treat me like an adult” options turned on by default in my browser, but sometimes I do a search directly on the Google website by mistake, and the results are so bad that I don’t understand how people can still use it in “treat me like a fuckwit” mode. Still get offered slop though:

A reminder of Joshi's Workslop Postulate

"Professor Jason Arday has resigned from the University of Cambridge and Jesus College with immediate effect, Black Current News can reveal, saying years of “relentless accusations, speculation and public commentary” have taken a profound toll on him and his loved ones."

Where should I be looking for solid coverage of AI changing conditions in translation work?

If you want to earn a PhD with a study of the medieval readers of the earliest MSS of Aristotle's De anima in Latin, only two days left to apply!

Continuing with the "someone made a horrific database of poor AI translations of early medieval letters" thing, I had a good laugh at these entirely hallucinated ones. They aren't even based on existing letters! www.lateromanletters.org/letters/epis... www.lateromanletters.org/letters/epis...

Some notes by little me on recent, hugely damaging, shamefully shameless events at Cambridge University. rorywaterman.substack.com/p/notes-on-a-cambridge-debacle If you read only one article about what has happened, try the new piece by Lanre Bakare in the Guardian, linked in my blog post.

Rare Book School mourns the loss of medieval manuscripts expert and RBS emeritus faculty member Albert Derolez. "His kindness as much as his brilliance will be remembered by those who knew him," said Director Michael F. Suarez, S.J. rarebookschool.org/resources/2026/07/rbs-mourns-albert-derolez

Martyn Hammersley's comments at the end of this article that Cambridge could have dealt with this from the outset are spot on - it will now be a field day for the right

We finished work on the R volume of the Thesaurus months ago. The last two parts are now in print and reached us today.

“The actual epistemological crisis [is] better exemplified by the Vanderbilt report’s AI-generated analyses: automated data collection, sloppy analysis, and a growing sprawl of LLM-generated pseudo-evidence. Meanwhile the AI ensloppification of [academic publishing] continues apace” archive.is/6qf0b

Academics who discreetly email journals to inform them of academic malpractice are doing those journals such a big favour. To see those journals shrug their shoulders in response is mystifying and unacceptable. I fear the only solution might be to call those journals out publicly

I don't suppose there's anyone here doing some work in the Archives nationales (Paris, not Pierrefitte) right now who wouldn't mind grabbing a couple of photos for me...?

Lexicographer wanted at the TLL! Deadline end of August. Somewhat surprisingly, the advert doesn’t say how good your German needs to be – the TLL is based in Munich, and it’s in the nature of the job that you need to be in the office – but of course the dictionary itself is written in Latin anyway.

The intricacies of the earliest stage of the mass production of printed books (Flood 2003):

A dark warning from a search results page with its listings upgraded from stodgy verbatim extracts to predigested robovomit. (Source: “When George Molland died in September 2002, it was a great loss to the history of medieval science and mathematics, especially of the fourteenth century.”) #progress

A dark warning from a search results page with its listings upgraded from stodgy verbatim extracts to predigested robovomit. (Source: “When George Molland died in September 2002, it was a great loss to the history of medieval science and mathematics, especially of the fourteenth century.”) #progress

Sad news I've only just heard: Albert Derolez has left us. He was 91, and one of the foremost palaeographers and codicologists in the world.

Last summer I gritted my teeth and paid 20 € postage to buy a 21.58 € paperback from Edizioni di Storia e Letteratura. Today I see they’re charging 50 € to deliver to the UK! Happily I can get the book I want elsewhere, but this isn’t the only ludicrously prohibitive postage fee I’ve seen recently 😠

Going to try to make this accidental 'livelihoo' on the OUP website into a thing: capricious quality of life and lifelikeness, essayistic esprit, careless brilliance.

Franciscan Studies 83 (2025): among other articles, a "total study" of the recently identified Franciscan collection in MS Paris, BNF, NAL 3245; Bonaventure's reception of Avicenna's psychology; and Jean de Roquetaillade's apocalyptic alchemy. buff.ly/YHi1npO

Today is as good a day as any other to think about giving back to something in your life that comes free to you. Do you have special skills or interests - for instance, you know Hebrew and are proficient at typing it? The digitized Lewis and Short could use your assistance with a few words.

Bulletin de Philosophie Médiévale 67 (2025): Ibn Ḥazm, Ibn Ṭumlūs, Durand on evils, Guy de l'Aumône's Sentences commentary, Jandun's commentaries, new MSS from the Paris arts faculty, Konrad Wimpina's citations, future contingents in Aurifaber. buff.ly/KhDj8Sb

So in summary from today’s unexpectedly viral thread: the methods and the tech here are not AI in the way that’s commonly understood right now. The methods are sound. The findings are important. HOWEVER 1/

Not even rare, old books are safe from AI companies

I understand the excitement about technology “rediscovering lost texts.” I do, & I get the media hype. But I wish people knew that there are hundreds of thousands of “lost” texts just sitting in manuscripts in libraries, which have never been published outside of those single manuscripts. 1/n

It’s not too late to book your place for the Palaeography Summer School in Birmingham in July! 📖✏️👨‍🎓👇 #MedievalSky #Skystorians #Latin palaeography.uk/study/short-...

This looks very interesting, with a strong review from Drew Jones. We don't often get new literary studies of Medieval Latin poetry these days. Plus: Baudri! I love Baudri. www.brepols.net/products/IS-...