Author: Roger Pearse

A curious story: the 1972 sales of the library of the King’s Inns, Dublin.

Posted on by Roger Pearse

The 1597 edition of the Opera Omnia of Tertullian, edited by Franciscus Junius, is one of the rarer early editions.  A few days ago, I learned that there was a copy at the Tresoar library in Frisia, in the Netherlands, shelf-mark “F 1532”.  The volume has a clear stamp showing that at one point it belonged the library of the King’s Inns, in Dublin.

The four Inns of Court in London are professional associations, which every barrister must join in order to practice, and which provide various facilities for the legal profession.  But Dublin did and does possess its own inn of court, the King’s Inns, which fulfils a similar role in the Irish legal system.  This was established in 1541, and is an important part of the cultural heritage of Ireland.  In particular it built up a valuable general library of antiquarian books and maps, including incunables, only a portion of which were related to the law.  But much of it was of crucial importance for Irish heritage and cultural identity.  Often the book was the only copy in the country, in some cases almost the only copy known to exist.

In 1971, however, something strange happened.  The King’s Inns was in serious financial difficulty.  This had been made worse by spending significant sums on its catering arrangements.  The ruling council of King’s Inns, the “benchers”, had to do something.  So what they decided to do was to sell those books in the library which were not related to the law, books “under the classification of scientific, religious and geographical.”

As news of the proposed sale drifted out into the library community in Ireland, concerns began to appear, and increasing numbers of questions were asked.  Scenting trouble, the then Irish Prime Minister obtained a promise from the benchers not to sell books which contained “any reference to Ireland.” 1  The barristers themselves began to ask questions about what was happening.  But Sotheby’s were instructed, and duly proceeded to remove all the books.  Shockingly it seems that the benchers did not even pay very much attention to just which books were removed.  A short hand list existed of a few hundred volumes, but this was only a fraction of what was removed:

It is clear that, as the van from Sotheby’s drove down Henrietta Street laden with some 7,000 volumes comprising almost three thousand titles, nobody at King’s Inns knew exactly which books had gone from the society’s library. Not even Sotheby’s had a list.2

Inevitably this meant that a number of valuable legal books were also taken, as well as books clearly of cultural importance.  Around 90,000 volumes were left, about 50% of which are non-legal.

Despite frantic attempts by Irish libraries to rescue at least some of the books, this collection was disposed of in 1972 in three sales, raising a total of £138,975 for the King’s Inns, plus a commission of approximately £24,000 going to Sothebys. 3  In modern money this would be approximately £2.38m and £411k.  In these days, when the expenditure of a few million pounds is almost petty cash, the income from the sale seems very paltry compared to the damage that it did to Ireland. 4

Title page of the Sothebys catalogue of the King’s Inns sale, 20 November 1972.
  1. Kenny, p.54.[↩]
  2. Kenny, p.62.[↩]
  3. Kenny, p.137.[↩]
  4. For many more details, see Colum Kenny, King’s Inns and the Battle of the Books, 1972: Cultural Controversy at a Dublin Library (2002).[↩]

Mithras for sale! – A Cautes figure at auction in Monte Carlo (CIMRM 645 / 646)

Posted on by Roger Pearse

Exciting news for those with rather more money than myself – a two foot (67 cms) tall statue of Cautes, one of the torch-bearers (“dadophores”), the attendants of Mithras, is up for sale at the Interencheres sale “From Caesar to Caesar” in Monte Carlo on the 8th August 2026.  The lot is here on the auction house website.  Here’s one of the images:

The inscription is “Hymnus inbicto” (for “invicto”), “Hymnus (gives this) to the unconquered (god)”.

The estimate is 30,000 euros – about $34,000 – although of course it will undoubtedly go for more, and the auction house commission and the tax add 30%.

The item must have come originally from a Mithraeum somewhere.  It’s immediate origin is from a “French private collection”.  Earlier restorations indicate 18-19th century work.

The website tells us that it was published in F. Cumont, Textes et Monuments figurés relatifs aux Mystères de Mithra, vol. 2 (1898).  The inscription is no 162 (on p.122), who gave its origin as from Capena.  It’s in CIL XI, 3865.  In Cumont the monument is no. 105 (on p.260):

*105. CAPENA. Statue de marbre blanc [H. Om,65] trouvée au commencement de ce siècle à Turrita. Aujourd’hui au même lieu dans la maison de M. Trasi.
Décrite: Tomasetti, Archivio della società Romana di storia patria, VII(1894), p. 420, n° 1, cf. CIL. XI,3865.

The CIMRM reference is 645 and 646, in Etruria, with location instead as “Torrita”:

645
Marble statue (H. 0.65), found at Torrita near Nazzano in the beginning of the 19th century. Formerly in Trasi’s house at Torrita, now at Rome (Tomasetti) but I do not know where.

Tomasetti in Arch. Storia Patria VII, 1884, 420 n. 1; Campagna Romana, III, 328f; MMM II 260 No. 105; Taylor, Cults Etr., 58.

Standing torchbearer in Phrygian cap, tunica, cloak and anaxyrides. In the l.h. (broken off) a torch; the rh. is missing. On the base beside the r. foot an inscription:

646

ClL XI 3865; MMM II No. 162.
Hymnu/s inbic/to.

I hope whoever buys this item enjoys it!

More Coptic on St Colluthus

Posted on by Roger Pearse

A few days ago I wrote about locating the 18th century publication of an early Coptic fragments of the acts of St Colluthus.  This was important, not for the content, but for the preface explaining how Cardinal Stefano Borgia came to own some parchment pages from the White Monastery at Sohag in Egypt; the first in Europe to do so.

I’ve not had time to read this preface as yet.  But by coincidence I learned today that those parchment pages have just been digitised and indeed uploaded to the Vatican website.  The Vatican collection is being thoroughly turned into digital copies, and of course this includes the Vatican Borgia collection.  After Cardinal Borgia’s death his collection of Coptic manuscripts was divided into two halves.  The first, containing mainly biblical texts, ended up in the Vatican.  The rest went to Naples.

Here’s the BlueSky post by Aaron M., who maintains the Wiglaf website which indexes what has been uploaded.  It’s a far, far better way to locate Vatican manuscripts than the official pages.  Here’s the post on his blog, “Week 25 of 2026“:

Twenty-five manuscripts were digitized the past week, unusually for the Vatican, the majority of these were uploaded on Sunday. Most of the manuscripts this week, twenty-two, came from Borg.Copt. The last three manuscripts came from. Barb.lat, giving two, and one volume only from Ott.lat. It is not clear if this is the start of a new pattern in digitization work, or an outlier week.

To the right is f.11v from Borg.copt.109.cass.XXVII.fasc.141, a 9th Century Sahidic copy of The Acts and Martyrdom of St. Coluthus.  First a note about the shelfmarks and the identification: Borg.Copt.109 is a collection of approx. 168 disbound leaves and partial manuscripts. They are shelved in 29 boxes, marked as “Cass” in the shelfmark, but numbered sequentially. Thus cass.XVII.fasc.141 is number 141 of the entire collection, and found in box 27. These were originally catalogued by Georgio Zoega1 in the early 19th C, and then there is a later catalogue/concordance between his catalogue and the current organization found in

J.-M. Sauget, “Introduction historique et notes bibliographiques au catalogue de Zoega,” _Museon 85_ (1972) pp. 25-63.

This particular manuscript was originally catalogued as being from the 4th century, but later scholarship points to it being a 9th C copy, carefully imitating a 4th C hand, which was known to be practiced at the White Monastery at that time.

The manuscript itself is at the Vatican website here.

Marvellous!

New Mithras book from Csaba Szabo on Mithras in Dacia

Posted on by Roger Pearse

Csaba Szabo, one of the most active current Mithras scholars, has a new book out from Cambridge Scholars Publishing on Mithras in Dacia.  His blog, with details, is here.

Details:

Csaba Szabó, Mithras in Roman Dacia: Local Appropriations of a Universal God. Cambridge Scholars (2026). ISBN13: 978-1-0364-7770-7.

Blurb:

The Roman cult of Mithras ranks among the ancient religious traditions that have left some of the most striking archaeological remains in Central and Eastern Europe. This volume examines the epigraphic and figurative evidence of the Roman cult of Mithras from the territory of the Roman province of Dacia (modern Romania) through a new and innovative methodology grounded in an interdisciplinary approach. Particular attention is devoted to the centuries-long historiographical tradition of Mithraic scholarship, contemporary developments in the archaeology of religion, and the phenomenon of religious glocalism. Within this framework, Mithras is analysed not merely as a universal and global deity within the religious landscape of the Roman Empire, but also as a glocal divinity actively shaped by local individuals and communities of Dacia.

I’m not sure that “glocal” is a word, but never mind.

Cover:

Likely to be good.  Available for pre-order, or so I gather.

Tertullian Matters: Two new English translations, and a new copy of the Junius 1597 edition at Tresoar

Posted on by Roger Pearse

I learn today that Brepols have published some new English translations of two works by Tertullian: the Ad Martyras and the De Fuga in Persecutione.  Both are translated by Thomas J. Heffernan for the Brepols Library of Christian Sources.  The publisher page is here.  I’ve added this to the Tertullian Project pages also.  That makes the first update since 2023.

Another interesting email reached me from Anne Popkema, at a library of which I had never heard – Tresoar in Frisia, in the Netherlands.  It turns out that they hold a rare item – a copy of the 1597 Junius edition of the works of Tertullian, complete with a great number of annotations, and also a letter from Junius’ grandson:

We hold one of the few copies (on the Continent) here at Tresoar, shelfmark F 1532. It was formerly part of the Dublin King’s Inns Library. Parts of it are quite heavily annotated in pencil, in what at first glance seems a late 18C hand to me. Also, we have a letter (5 August 1657) by Francis Junius F.N., a grandchild of Francis Junius the Elder, to Theodorus Saeckma (d. 1666), thanking him for sending him a copy of his grandfather’s commentaries on Tertullian.

Again I have added a note to the Tertullian Project pages.

It is very good to know about this stuff!

Searching for “Giorgi Fragmentum Copticum”

Posted on by Roger Pearse

One of the earliest enthusiasts for Coptic literature was Cardinal Stefano Borgia (1731-1804), a Papal bureaucrat who created a museum in the Borgia Palace in his home town of Velletri.  He spent considerable sums acquiring material for it, and especially material from Egypt.

There is a 1781 publication which describes how he obtained some Coptic fragments which we now know to come from the White Monastery in Sohag in middle Egypt.  The publication is referenced by Stephen Emmel in his ground-breaking Shenoute’s Literary Corpus, in vol. 1 (2004), p.20 with the bibliography in vol. 2, p.964.  The same material appears in other papers, including one by Paola Buzi (online here).1  Buzi gives the reference thus:

GIORGI A., Fragmentum Copticum ex Actis S. Coluthi Martyris Erutum ex Membranis Vetustissimis Saeculi V ac. Latine Redditum, Romae 1781, p. 3-4.

This is not an easy item to find.  Indeed at one point I wondered if it even existed.  But it does exist, and indeed it is online here.

However it took me two days (!) to locate it, trying various extracts of the title.  In the end the search phrase “Fragmentum Copticum ex Actis” produced it, although I am sure that I had tried this before.

But before I located the 1781 publication, I stumbled upon a 1783 reprint in a four-volume series of what we would today call academic papers, publishing previously unpublished primary materials from various libraries.  The name of Stefano Borgia appears in several places within it.

The series is titled Anecdota litteraria ex mss. codicibus eruta, published in Rome and edited by Giovanni Cristoforo Amaduzzi and G. L. Bianconi.  But reference volumes such as the British Museum Catalogue of Printed Books also reference it under the name of Stefano Borgia.

The series volumes may be found here:

Volume 4 contains the Fragmentum Copticum.  The table of contents (“Elenchus”) on page XII (page 17 of the PDF) gives the “Fragmentum Copticum…” as the second item, on pages 47-229.  Page 49 (page 74 of the PDF) features an address to the reader by Stefano Borgia, and the details that we want are on page 50.  After this, page 75 has a “monitum” (notice) by “Augustinus Antonius Georgius” – i.e. our friend “Giorgi” about the text, but this does not indicate that he is the author of the remarks.

Attributing this work to “A. Giorgi” is unhelpful, it seems, at least as far as Google is concerned.  It is probably better to attribute it to Stefano Borgia.

Title page
  1. Paola Buzi, “The Borgia Coptic manuscripts collection preserved in Naples. A new catalogue,” in Huitième Congres International d’Études Coptes, Paris, 27 giugno – 3 luglio 2004, (Paris 2007), pp. 41-50 (online here).[↩]

The 1412 inventory of the manuscripts of Amelungsborn Abbey

Posted on by Roger Pearse

I reported yesterday on the discovery of two new sermons by St Augustine in a Latin manuscript in a monastery in Poland.  One statement in the press release is also of great interest.  The discoverer, Prof. Christian Tornau, of the University of Würzburg, stated:

“An old catalogue from the monastery mentions a text that bears the same headings and has the same sequence of content as our manuscript. It could have served as a model,” the researcher explains. Tornau cannot confirm this assumption with absolute certainty, however, as the entire library holdings of Amelungsborn Monastery were destroyed by fire during the Thirty Years’ War (1618-1648).

The “old catalogue” in question is the inventory of 1412 of the Cistercian Abbey of Amelungsborn. This takes a little locating.

The inventory is listed in a catalogue of surviving medieval inventories, published by Theodor Gottlieb, Ueber mittelalterliche Bibliotheken (1890 – online here).  Amelungsborn is on p.18-19:

The footnote reads:

Im Anniversarienbuch sind 7 Schenkungen von Büchern aufgeführt; dies wird pg. 18 gesagt. – Ferner ist auf einer Seite des Catal. ein Verzeichniss von Büchern mit der Ueberschr.: Hii de Corbeja, (von woher sie nach Dürres Vermuthung  zur Abschr. entlehnt waren).

The anniversary book lists seven donations of books; this is mentioned on page 18. Furthermore, on one page of the catalogue, there is a list of books with the heading: “Hii de Corbeja” (from where, according to Dürre’s presumption, they were borrowed for copying).

The inventory of 1412 is preserved, then, in a manuscript, once in the Ducal Library in Wolfenbüttel, today in the Niedersächsisches Landesarchiv (NLA) where it has the shelfmark “Niedersächsisches Landesarchiv Wolfenbüttel VII. B. 111.”  A monochrome microfilm is online.  Each image is two pages: the catalogue is right at the end, on pages 101-105, i.e. images 54-56.  Here is the start of it:

Niedersächsisches Landesarchiv Wolfenbüttel VII. B. 111, folio 101r, top (image 54)

The inventory has been printed, but those seeking it may find that it can take a bit of time to locate thanks to a pleasant custom in 19th century German scholarship, whereby the author publishes the article twice under the same title in the same year but with different contents.  So it has been here.

The actual article, that contains a transcription of the inventory of 1412, is this:

H. Dürre, “Beiträge zur Geschichte der Cistercienserabtei Amelungsborn”, Programm des herzoglichen Gymnasiums zu Holzminden. Ostern 1876, p. 3-24; esp. 19-24.

This may be downloaded from here, and the inventory starts on page 19 thus:

But the unwary or unlucky researcher will search Google for “Dürre Beiträge zur Geschichte der Cistercienserabtei Amelungsborn” or something of the sort and will be presented with this:

H. Dürre, “Beiträge zur Geschichte der Cistercienserabtei Amelungsborn”, Zeitschrift des Historischen Vereins für Niedersachsen. 1876

In this volume, Dr Dürre’s article is the fourth, begins on p.179, and ends on 212, omitting the transcription of the inventory altogether.  It was only when you start to search for Programm des herzoglichen Gymnasiums zu Holzminden that you will make progress.

Looking at the inventory, the entry for the new sermons is on p.22, line 220 of the printed inventory.

The manuscript:

NLA WO VIII B 111, fol. 103r, column 2, part way down.

So there is the reference explained.

I think we will all be very interested to see the text of Dr Tornau’s discovery in full!

Another two sermons of St Augustine discovered: on the Witch of Endor

Posted on by Roger Pearse

Excellent news from the University of Würzburg, where a researcher has discovered two unknown sermons by St Augustine in a Latin manuscript in Poland!   The reporting (by Martin Brandstätter) is unusually good:

One day in 2024, the phone rang for Professor Christian Tornau, a Latinist at the University of Würzburg: An employee of the Bad Doberan Monastery Association in Mecklenburg-Western Pomerania asked him to decipher a 12th-century manuscript that originally belonged to Bad Doberan Abbey but is now kept in its daughter monastery, Pelplin, in Poland. The manuscript contained six sermons by the Christian Church Father Augustine of Hippo (354-430).

What initially seemed like a routine philological assignment turned into a discovery: “Two of the six sermons are previously undiscovered writings by Augustine,” Professor Tornau rejoiced at the unexpected find. He is currently working with Professor Dorothea Weber and Dr. Clemens Weidmann from the CSEL series (Corpus Scriptorum Ecclesiasticorum Latinorum) on an edition of the two Latin sermons.  …

The newly discovered sermons deal with the Old Testament story of the Witch of Endor from the First Book of Samuel. “Saul believes himself to be in a hopeless situation shortly before a battle against the Philistines. God does not hear his prayers. He turns to a witch,” explains Tornau. At Saul’s request, she summons the supposed spirit of the deceased prophet Samuel, who foretells his death in battle.

The story raises a theological question: “Why can a necromancer summon the spirit of a prophet? This, in turn, opens up the problem of theodicy: How can an omnipotent God allow this, or is he not truly omnipotent?” the Latinist explained. Theologically, there are two interpretations: Either it must be a case of deception by the witch, or God permitted the summoning to warn Saul of certain death.

The sermons play with these interpretations. “The first sermon was given during the Sunday service and ends with the question of theodicy and its interpretations. Only the second sermon, on the following Wednesday, weighed the options,” Tornau explained. The congregation was thus given a certain degree of freedom to form their own thoughts on the biblical passage.  …

There have already been cases in which supposed writings by Augustine turned out to be forgeries. Therefore, the professor of Latin studies proceeded cautiously: He analyzed the text together with the expert Clemens Weidmann, conducted meticulous research, and organized a summer school in Vienna in the autumn of 2025. Twenty other Latinists attended to discuss and verify the authenticity of the text. In the end, everyone agreed: The sermons are genuine.

Reconstructing the transmission history of the sermons proved challenging. “First of all, the creation of such a manuscript in the 12th century is unusual. A copy from the beginning of the 8th or 9th century would be more typical,” says Latinist Tornau. He therefore considers it very likely that the manuscript is based on an earlier version from the Lower Saxon monastery of Amelungsborn.

“An old catalog from the monastery mentions a text that bears the same headings and has the same sequence of content as our manuscript. It could have served as a model,” the researcher explains. Tornau cannot confirm this assumption with absolute certainty, however, as the entire library holdings of Amelungsborn Monastery were destroyed by fire during the Thirty Years’ War (1618-1648).

The first edition will address the transmission of the texts and contextualize their content and authenticity. “This isn’t a sensational find like the 30 Augustinian writings discovered in Mainz in 1990. But we are supplementing Augustine’s extensive body of work with two further fascinating texts in a critical edition,” Tornau summarizes. The edition is expected to be published by CSEL at the end of 2026.

The article even includes a legible photograph of the first page of the manuscript, and clearly identifies it – a first, in my experience.  The red title reads, I think: “Sermones Augustini de phitonissa et de resuscitatione samuelis”, (i.e. de pythonissa…?)

Excerpt from the manuscript (in red): “Augustine’s sermons on the fortune teller and the appearance of Samuel.” Source: Pelplin, Diocesan Library, Codex 114 (195), fol. 14r. (Image: Biblioteka Diecezjalna im. Biskupa Jana Bernarda Szlagi w Pelplinie)

The town of Pelplin originated from the monastery, itself populated by monks from Bad Doberan in Mecklenburg, according to Wikipedia.  The Biblioteka Diecezjalna library has a website here, although it does not seem to have any manuscript material online.  There is a website for the “Friends of Bad Doberan monastery,” mentioned in the article, here.  I was unable to locate a publication of the “old catalogue” of Amelungsborn unfortunately.

It’s very good news, and it will be most interesting to see what Augustine had to say on this curious passage of the bible.

Update (16 June 2026): The “old catalogue” of Amelungsborn, mentioned by Dr. Tornau, is discussed in my next post.

Let’s all agree that Amelineau needs to be flogged!

Posted on by Roger Pearse

Today I received an email asking me about a letter of St Shenouda of Atripe, the 4th century Coptic abbot.  The letter in question was apparently written to a nun, saying “I knew you long ago.”  The email asked if I knew the source.

Well, I have almost no familiarity with the works of Shenouda (or Shenoute), so I thought that I would poke around a bit.  The email author was French, so I wondered what existed in French.

Interestingly Emile Amelineau translated some material by Shenouda into French as long ago as 1907, in his “Oeuvres de Schenoudi”.  Vol. 1 is here; vol. 2 here.  But… neither has a table of contents!  Cunningly, the editor has also ensured that none of the names of the works appears in the running titles.  This would not be great in printed volumes; in a PDF it is impossible to gain an overview of the volume and its contents without significant work.  Flinders Petrie had hard words for Amelineau’s destructive ineptitude as an archaeologist at Abydos; a century later, I feel ever so slightly vexed at such carelessness.

On reading his interminable introduction, in search of information, I find that he is not really editing anything.  What he is doing is printing chunks of fragments from papyri.  So, for him, each section is just whatever was found in such a manuscript.  That may be better than nothing, but it does not help us here.

It seems that the reconstruction of Shenouda’s works had to wait until modern times, and the labours of Stephen Emmel.1.  But that will have to wait for another day.

  1. See https://bmcr.brynmawr.edu/2023/2023.08.22/[↩]

Garbage in… Greek out? Experiments with Deepseek using OCR’d Italian containing embedded Greek.

Posted on by Roger Pearse

The letters of the 6th century sophist Aeneas of Gaza have been sitting in a folder on my desktop for a month or two now, and I want to make some progress with making a translation into English.

It’s not a big text.  Each letter is only a short paragraph, and there are only twenty-five letters.  So the whole text would fill less than a dozen pages perhaps.  I have the 1962 edition with Italian translation by Lidia Massa Positano, which is more than a hundred pages.  There seems to be a rule that editions of tiny texts can be obese!  I have never forgotten the Gerlo edition of Tertullian’s “De Pallio” – a very short text of a page or two – which filled two lengthy volumes.  But Gerlo published in 1940 in the Netherlands, and it may have been expedient for him to be engaged in such a project at that time.  Anyway the Positano edition is not that daft, and consists of an introduction, the Greek text, the Italian translation of each letter with commentary and footnotes.

At some earlier point I seem to have run the Positano book through Abbyy Finereader 15 software to create a Word document, which is 77k in size.  So today I extracted the portion to do with the letters.  The OCR language was Italian, so the portions of the commentary that contained quotes from the Greek were gibberish at those points.

Anche Procopio scrive lettere riguardanti prestiti di libri. Per esempio, dall’ep. LXIII, diretta a Pizio, risulta che Procopio si rammarica di non possedere un libro chiestogli in prestito da co­stui. Nell’ep. CHI, diretta a Stefano, Procopio vivamente lo rim­provera di un grave ritardo nella restituzione di un libro (p. 572,46 : èpuO-piàv cpiQaec^ w; napa^à; tVjv òrtoo/eaiv xaì xà? auv^xaj uirepi- Swv xal tò pi^Xfov i/wv Tpkov v) ré-capTov toutI ó p^SèrpiTOV xaftéÉEiv è7raYYeiXàp,Evo<;) e cita come esempio di solerzia e di zelo proprio un Giovanni (p. 573,12 àXX’ où/ 3 y£ aocpuiTaxo; ’IwàvvTj;

The printed text is as follows:

I thought, as a first step, that I would load the Word document into ChatGPT and get an AI translation of the lot, commentary, footnotes etc.  The idea was only to allow me to skim the material, and decide what to focus on.

But ChatGPT started complaining that it was a long file, a very very long file, etc etc:

The document you uploaded is very large (roughly 130,000 characters / 60+ pages). Translating the entire file accurately would be too long for a single response.

Then it offered various ways to make my life difficult if I carried on.

Well, I didn’t.  I popped it into Deepseek instead.  My prompt was “Translate this Italian text into English” – nothing exotic.   Deepseek made no objections and speedily output an English version of the whole thing.  I did have to copy and paste the output to Word, but that was not too burdensome.

But as I copied and pasted, I noticed something strange.  My eye was drawn to the gibberish sections of the Greek.  Here is the same passage, converted from exactly that gibberish above:

Procopius also writes letters concerning loans of books. For example, from ep. LXIII, addressed to Pyzius, it appears that Procopius regrets not possessing a book requested from him by the latter. In ep. CIII, addressed to Stephen, Procopius severely reproaches him for a serious delay in returning a book (p. 572,46: ἐρυθριᾷν φῂς ὡς παρελθὼν τὴν ὑπόσχεσιν καὶ τὰς συνθήκας ὑπερβὰς καὶ τὸ ἐμὸν χρέος τεταρτεῦσαι τοῦθ᾽ ὁ μηδὲ τρίτον καθέξειν ἐπαγγειλάμενος) and cites as an example of diligence and zeal a certain John (p. 573,12 ἀλλ᾽ οὐχ ὁ γε σοφώτατος Ἰωάννης…

It has recognised that the text is Greek – I did not tell it so – and it produced that accented Greek output.  The Greek is not actually completely correct.  But it’s very close!

In computing, there is no magic.  If it looks like magic, it means only that you don’t understand what is going on.  The input that I gave it was not enough to produce that Greek output.  It was just the attempts of an OCR engine to make Italian out of Greek.  So the Greek was retrieved from elsewhere, and the garbage string used to search for it.

I then tried the following prompt with a page of the garbled Greek:

This comes from a book in Italian, displaying Greek. Correct the Greek. ”’….”’

This it proceeded to do, with an interesting commentary underneath:

So the garbled text is being used for a look-up of some sort.

We know that the databases in these “AI” engines – the Large Language Models (LLMs) – are essentially a search database made by pirating vast amounts of books and everything else. In this case it is perhaps using the garbled strings to look up stuff in the Thesaurus Linguae Graecae database.

But, as ever with AI, you just cannot trust what it gives you.  You have to check, and checking can take longer than doing it yourself.

Interesting, and frustrating.  As ever!