Yesterday I started googling about Isidore of Pelusium, and I quickly came across a number of papers showing that Dr. Madaline Toca is actively working on Isidore of Pelusium, the manuscripts of his letters, the reception of his work in Latin, and so on. This is good news! Most of these papers are accessible on Academia here, which is even better news. Also among her efforts is an online bibliography for Isidore, here.
This bibliography informed me that a previously unpublished Greek “Life” had been printed back in 1958 in an obscure Greek volume. Thankfully she provided a PDF here.
The author of this publication was none other than a certain Morton Smith. Today Morton Smith is notorious for his “Secret Mark” forgery. But in 1957 he was just another a young scholar, travelling through Greece and the Levant, and searching for manuscripts of the letters of Isidore of Pelusium.
The “Life” printed is basically a transcription of four manuscripts, the oldest 12th century, the rest 16th century; three from Mount Athos, and the other in the National Library in Athens. From the incipit and explicit, it appears to be the text listed in the BHG, vol. 3 (supplements) on p.38 as BHG 2209. The text is in 12 paragraphs, which are those in the manuscripts. No paragraph numbering or subtitles appear.
I was curious what the text said. Unfortunately Smith gave no translation. It does not seem to be in the TLG. But I knew that if I could create an electronic text, then I could work on it with ChatGPT.
What I did was to run it through Abbyy Finereader 15 Pro. Unfortunately this only supports modern Greek, which means that the accentuation gets mangled. But never mind that. Who cares? This is just idle interest. Or so I thought.
Anyway I pulled it, did a certain amount of light correction. Then I pasted it into ChatGPT using the free online web interface, and asked for an English translation. The request that I gave it – the “prompt” in the AI jargon – was this:
Please correct the following text in medieval Greek: ”’ …. ”’
with the Greek between the triple quotes. Triple quotes seem to be the standard way to indicate something in these AI interfaces. I have no idea why.
Well, ChatGPT produced a very nice, perfectly readable translation. Here is the opening section:
The Life and Conduct of Our Holy Father Isidore of Pelusium
This divine Isidore, who lived a wondrous life both in deed and word, shone like the sun in both respects, and even brighter. His homeland was Pelusium, a distinguished city in Egypt. He sprang from a pious lineage, yet he himself flourished as an even more devout offshoot. From his earliest years, he ardently embraced virtue and steadfastly held to it as an inseparable possession. He despised the pleasures of youth, such as childish games, frivolous laughter, indulgence that weakens the soul, and enslavement to the desires of the flesh—everything, in short, that distracts the mind from perceiving true beauty. Instead, he pursued complete detachment and the heights of divine contemplation, embracing temperance in all things and the highest form of education—one that elevates the soul and leads it toward the heavens.
My experiences with ChatGPT as a helper with the translation of the “Life” of John Damascene had demonstrated that ChatGPT does give reasonable accuracy. But also that you do have to check! But how to do so?
My next thought was to try out another AI product, the Chinese DeepSeek, also available for free online. I did the same with this, with the same prompt, and it also produced a translation. Here’s the opening section:
The Life and Conduct of Our Holy Father Isidore of Pelusium
This divine Isidore, who was wondrous in both life and word, shone brighter than the sun in both respects. His homeland was Pelusium, a renowned city in Egypt. From a devout root, a more devout branch sprouted, for from his earliest youth, he wholeheartedly embraced virtue and held it as an inalienable possession, demonstrating his unwavering love for it. He shunned the pleasures that typically delight youth—childish games, youthful frivolities, and indulgences that weaken the soul and enslave it to the desires of the flesh. In short, he avoided everything that hinders the mind from attaining true understanding of what is good. Instead, he devoted himself entirely to what leads to perfect detachment and the highest contemplation of divine things: universal self-control and the best education, which lightens the soul and lifts it to heavenly things.
Various awkwardnesses in the DeepSeek output suggested to me that it was probably closer to the Greek. But there are enough differences there that I felt that I needed to get closer to the Greek.
So I put in another prompt to DeepSeek:
For each sentence, give the Greek, then the English translation; then analyse the syntax, with meanings; then give a list (in table format) of all the words with the grammar, lemma, part of speech, case, number, gender, and meaning for each word.
Doing so produces quite a bit of output. The “translation” gets quite a bit closer to the Greek as you do this, and of course you can read its explanation of the syntax, you can lookup individual words in Logeion or somewhere using the lemma, and generally wrestle with the sentence until you believe what you have.
But back to the AI output.
If you run this prompt in ChatGPT, it just does one sentence only, and then it asks if you want to continue. If you say “yes”, then it does another. However ChatGPT has no idea what a sentence is! So it does a dozen or so words. DeepSeek is the same, but I quickly found that the length of a “sentence” was much shorter. Rather nervously I asked if it could “do the same but for two sentences” and I got longer outputs. So that worked. When I asked if it could do four “sentences”, it went a bit funny. So I went back to two.
At one sentence a time, all this becomes very tedious. Copying and pasting the output to a word document takes a lot of time. Indeed I have spent the whole day doing this. But DeepSeek did a fine job. It was no worse than ChatGPT.
It’s generally best to do this, one paragraph at a time. It doesn’t feel so oppressive, and you can go off for a break at the end of each paragraph. You don’t want to sicken yourself, and it takes too long to do the whole thing in one go, even for just 4 pages of Greek.
For the curious, I attach a sample file with what I got for one of the paragraphs. How reliable the output is, well, I will find out in due course! It’s here:
- 8 deepseek (Word .docx)
But I did get tired. So I wondered if it was possible to do this process from the computer command line, thereby saving myself a lot of time. You can indeed connect to the website using the “API,” which would allow you to write a program. But… alas… they want money for that! The restrictions on the free web interface are deliberate.
You can also download for free a DeepSeek “model” (jargon word) and run it on your PC. But unless you have an awful lot of memory fitted, you will find yourself working with “distilled” versions which are not nearly as good. The process is fairly technical, and although I got it to work, I’d need to spend a lot more time on this. Whether my fairly powerful PC would handle a full-size model is something that I don’t yet know. So I went back to the free web interface.
One place where DeepSeek is definitely superior to ChatGPT is that it recognises when it reaches the end of the passage. ChatGPT does not. It will quite happily continue beyond the end, “translating” random Greek garbage. So every so often you have to take the last word translated, and check that it is still in the text!
Doing this led to an interesting discovery. I always ask for the Greek, the English, and then the syntax analysis. I found that DeepSeek was silently fixing up the garbage Greek text that I had got from the OCR. It was adding the missing accentuation.
So I tried asking it explicitly to do so.
Please correct the following text in medieval Greek: ”’…”’
And it did, and then translated it. A quick look at the original PDF suggests that it is doing a good job. Well, well.
Update 27 March 2025. I did find a couple of places in the “corrected” Greek text where it had mysteriously introduced a full-stop. It also capitalised proper names without my asking it to! But still interesting.
One thing that is really important – divide your text, however short, into chunks of no more than half-a-dozen sentences, and work on each chunk in turn in separate documents. If you think, as I did, that the document is too short to bother, you will quickly get into a morass. It’s psychologically necessary to have some positive reward every few sentences, or you get depressed and give up. In this case I ended up simply numbering the paragraphs and taking each as a “chapter.”
Here is a section you posted previously which I ran through grok.com, with the command “Translate from Greek and justify your translation”:
Thought for 53s
Expand for details
Here is the translation of the provided Greek text into English, followed by a justification for the translation:
Translation
(371) TOA’. To Pansophius.
If the advocate of Marcion’s blasphemy presents the so-called Gospel among them, take it and read, and you will find immediately in the introductions the absurdity. For they cut off the very genealogy tracing Christ from David and Abraham; and proceeding a little later, you will see another evil intention. For having altered the words of the Lord, who said, “I did not come to abolish the law or the prophets,” they made it: “Do you think that I came to fulfill the law or the prophets? I came to abolish, but not to fulfill.” From these things, you will know how they fabricate hostility to the two Testaments, having contrived that Christ is foreign to the law.
Justification of Translation
To ensure accuracy and clarity, I approached the translation systematically, breaking the text into manageable parts and analyzing the grammar, vocabulary, and context. Below, I justify the choices made for each section of the text:
Header: (371) ΤΟΑʹ. ΠΑΝΣΟΦΙῼ.
ΤΟΑʹ: This appears to be a numerical or sectional marker, likely “371” in a sequence, possibly indicating a chapter, section, or letter number. The symbol “ʹ” is a Greek numeral marker, and “ΤΟΑ” corresponds to 371 (Τ = 300, Ο = 70, Α = 1). I retained “(371) TOA'” to reflect this.
ΠΑΝΣΟΦΙῼ: This is a dative form of “Pansophius,” a proper name, indicating the recipient of the text. Translated as “To Pansophius,” which is standard for dedicatory phrases in Greek epistolary or argumentative texts.
First Sentence: Εἰ προΐσχεται ὁ τῆς Μαρκίωνος συνήγορος βλασφημίας, τὸ παρ’ ἐκείνοις ὀνομαζόμενον Εὐαγγέλιον, λαβὼν ἀνάγνωθι, καὶ εὑρήσεις εὐθὺς ἐν προοιμίοις τὴν ἀτοπίαν.
Εἰ προΐσχεται: “Εἰ” means “if,” and “προΐσχεται” is a third-person singular present middle/passive of “προΐστημι,” meaning “to put forward” or “to present.” In this context, “presents” fits as it suggests offering something for consideration.
ὁ τῆς Μαρκίωνος συνήγορος βλασφημίας: “ὁ συνήγορος” means “the advocate” or “defender,” and “τῆς Μαρκίωνος” is a genitive indicating possession, “of Marcion.” “Βλασφημίας” (genitive of “βλασφημία,” blasphemy) modifies “συνήγορος,” suggesting “the advocate of Marcion’s blasphemy,” i.e., someone defending Marcion’s blasphemous views. Marcion was a known heretic who rejected the Old Testament, providing context for “blasphemy.”
τὸ παρ’ ἐκείνοις ὀνομαζόμενον Εὐαγγέλιον: “τὸ Εὐαγγέλιον” is “the Gospel,” and “παρ’ ἐκείνοις ὀνομαζόμενον” means “called/named among them,” with “ἐκείνοις” referring to Marcion’s followers. Translated as “the so-called Gospel among them” to indicate it’s their version, distinct from the canonical Gospel.
λαβὼν ἀνάγνωθι: “Λαβών” (aorist participle of “λαμβάνω”) means “having taken” or “take,” and “ἀνάγνωθι” (imperative of “ἀναγινώσκω”) means “read.” Together, “take and read” is a natural imperative sequence in English.
καὶ εὑρήσεις εὐθὺς ἐν προοιμίοις τὴν ἀτοπίαν: “εὑρήσεις” (future of “εὑρίσκω”) is “you will find,” “εὐθὺς” means “immediately,” “ἐν προοιμίοις” is “in the introductions” (prefaces or opening sections), and “τὴν ἀτοπίαν” is “the absurdity” (from “ἀτοπία,” meaning something out of place or absurd). Combined as “and you will find immediately in the introductions the absurdity,” reflecting a discovery of something nonsensical.
Second Sentence: Αὐτὴν γὰρ τὴν κατάγουσαν ἐπὶ Χριστὸν ἁπὸ Δαβὶδ καὶ Ἀβραὰμ γενεαλογίαν ἀπέτεμον•
Αὐτὴν γὰρ: “For this very,” with “γὰρ” introducing an explanation and “αὐτὴν” emphasizing the object.
τὴν κατάγουσαν ἐπὶ Χριστὸν ἁπὸ Δαβὶδ καὶ Ἀβραὰμ γενεαλογίαν: “κατάγουσαν” (present participle of “κατάγω,” to lead down) in genealogical contexts means “tracing” or “leading to.” “ἐπὶ Χριστὸν” is “to Christ,” and “ἁπὸ Δαβὶδ καὶ Ἀβραὰμ” is “from David and Abraham.” “Γενεαλογίαν” is “genealogy.” Translated as “the genealogy tracing Christ from David and Abraham,” reflecting the lineage in canonical Gospels (e.g., Matthew 1).
ἀπέτεμον: Third-person plural aorist of “ἀποτέμνω,” meaning “they cut off” or “excised.” Marcion’s rejection of Christ’s Jewish roots supports this as “they cut off.”
Third Sentence: καὶ μικρὸν ὕστερον προϊὼν ἄλλην ὄψει κακόνοιαν.
καὶ μικρὸν ὕστερον προϊὼν: “καὶ” is “and,” “μικρὸν ὕστερον” is “a little later,” and “προϊὼν” (participle of “προϊέναι”) is “proceeding.” Together, “and proceeding a little later” suggests moving further in the text.
ἄλλην ὄψει κακόνοιαν: “ὄψει” (future of “ὁράω”) is “you will see,” “ἄλλην” is “another,” and “κακόνοιαν” (from “κακόνοια”) means “evil intention” or “malice.” Translated as “you will see another evil intention,” indicating a further issue.
Fourth Sentence: Ἀμείψαντες γὰρ τὴν τοῦ Κυρίου φωνήν, «Οὐκ ἦλθον, λέγοντος, καταλῦσαι τὸν νόμον ἢ τοὺς προφήτας,» ἐποίησαν• Δοκεῖτε, ὅτι ἦλθον πληρῶσαι τὸν νόμον ἢ τοὺς προφήτας; Ἦλθον καταλῦσαι, ἀλλ’ οὐ πληρῶσαι.
Ἀμείψαντες γὰρ τὴν τοῦ Κυρίου φωνήν: “ἀμείψαντες” (aorist participle of “ἀμείβω,” to change) is “having altered,” “γὰρ” is “for,” and “τὴν τοῦ Κυρίου φωνήν” is “the voice of the Lord,” meaning His words. Translated as “for having altered the words of the Lord.”
«Οὐκ ἦλθον, λέγοντος, καταλῦσαι τὸν νόμον ἢ τοὺς προφήτας,»: “λέγοντος” (participle) is “saying,” “οὐκ ἦλθον” is “I did not come,” “καταλῦσαι” (from “καταλύω”) is “to abolish,” and “τὸν νόμον ἢ τοὺς προφήτας” is “the law or the prophets.” This echoes Matthew 5:17, translated as “who said, ‘I did not come to abolish the law or the prophets.'”
ἐποίησαν: “They made,” from “ποιέω,” indicating they changed it to something else.
Δοκεῖτε, ὅτι ἦλθον πληρῶσαι τὸν νόμον ἢ τοὺς προφήτας; Ἦλθον καταλῦσαι, ἀλλ’ οὐ πληρῶσαι: “δοκεῖτε” is “do you think,” “ὅτι ἦλθον πληρῶσαι” is “that I came to fulfill,” “ἦλθον καταλῦσαι” is “I came to abolish,” and “ἀλλ’ οὐ πληρῶσαι” is “but not to fulfill.” This reverses Matthew 5:17’s meaning, translated as “Do you think that I came to fulfill the law or the prophets? I came to abolish, but not to fulfill,” with quotation marks to show the altered text.
Fifth Sentence: Ἐκ τούτων δὲ εἴσῃ, ὅπως ἔχθραν ταῖς δυσὶ Διαθήκαις κατασκευάζουσι, ξένον εἶναι τοῦ νόμου τὸν Χριστὸν σχεδιάσαντες.
Ἐκ τούτων δὲ εἴσῃ: “ἐκ τούτων” is “from these things,” “δὲ” is “and,” and “εἴσῃ” (future of “οἶδα”) is “you will know.” Translated as “from these things, you will know.”
ὅπως ἔχθραν ταῖς δυσὶ Διαθήκαις κατασκευάζουσι: “ὅπως” is “how,” “ἔχθραν” is “hostility,” “ταῖς δυσὶ Διαθήκαις” is “to the two Testaments” (Old and New, from “διαθήκη,” covenant), and “κατασκευάζουσι” (from “κατασκευάζω”) is “they fabricate.” Translated as “how they fabricate hostility to the two Testaments,” aligning with Marcion’s theology.
ξένον εἶναι τοῦ νόμου τὸν Χριστὸν σχεδιάσαντες: “ξένον” is “foreign,” “εἶναι” is “to be,” “τοῦ νόμου” is “of the law,” “τὸν Χριστὸν” is “Christ,” and “σχεδιάσαντες” (from “σχεδιάζω,” to contrive) is “having contrived.” Translated as “having contrived that Christ is foreign to the law,” reflecting Marcion’s separation of Christ from Jewish law.
Overall Rationale
The translation reflects a critique of Marcion’s Gospel, which altered canonical texts to fit his theology—rejecting the Old Testament and Christ’s connection to it. I used precise grammatical analysis, consulted the historical context (Marcion’s heresy), and compared the altered quote to Matthew 5:17 to highlight the reversal. Each phrase was rendered naturally in English while preserving the Greek’s intent, ensuring the critique’s sharpness is clear.
Roger: I have had success using gImageReader to OCR ancient Greek texts. Its a free download. Also, I just happened to by Morton Smith’s last TA before retired at Columbia back in the 80s. Keep up in the interesting work!
Glen Thompson (www.fourthcentury.com).
What a fascinating and honestly inspiring read. The potential for AI to unlock ancient texts and broaden our access to understanding this period of history is immense! I am following your experiments with great interest.
Thank you! It’s odd that I don’t see more about people doing this. But it’s incredible that it’s even possible!
That’s a very interesting prompt and output – thank you!
(Sorry – comment vanished into spam folder, which I have just looked at. Very sorry!)
Hi Roger, since you’re a programmer, you could automate the process. Just send one or two sentences at the same time, and save the response, repeat. Talking to an LLM is quite easy, although through the API it won’t behave exactly like ChatGPT, as you need to work a bit to get parameters and prompt similar, but it also allows much more fine-tuning.
But the API costs money, I believe?
Obviously I can get the prompt the same. But I don’t understand the “parameters” bit. Do you have a link you can point me to?
Here’s an example script of running aider (recommended tool):
===================
#!/usr/bin/env bash
set -o errexit -o pipeline -o nounset
prompt=”$1″
if [ -z “$prompt” ]
then
echo 1>&2 Pass prompt as first parameter
exit 1
fi
aider \
–model o3-mini \
–architect \
–reasoning-effort high \
–editor-model sonnet \
–no-detect-urls \
–yes-always \
–message “$prompt”
===================
Parameters are things like initial prompt, reasoning effort, token size, etc.
Thank you! Looks good. I had not heard of aider. Will try!
Aider might be a bit too much of a programming tool in your case. But use it to build the tool for you. I think AI might be the perfect tool for you in this case as your tools would be highly specialised. So you work with aider to create a tool that:
Accepts text.
Splits in paragraphs.
Translates it.
Puts it together in tables.
You can ask it to build a UI around this.
BTW, if Aider has too much of a programming slant, you can simply call ChatGPT directly:
=============
const prompt = I want to translate “${text}” from Latin to Greek. Only return the Greek text. Only return the Greek text
const OPENAI_API_KEY = process.env.OPENAI_API_KEY
const response = await fetch(“https://api.openai.com/v1/chat/completions”, {
method: “POST”,
headers: {
“Content-Type”: “application/json”,
Authorization:
Bearer ${OPENAI_API_KEY},},
body: JSON.stringify({
model: “gpt-4o”, // Changed from o3-mini to gpt-4o
messages: [{ role: “user”, content: prompt }],
max_completion_tokens: 1024,
}),
})
const data = await response.json()
=============
I appreciate the ideas and samples – thank you. I did get aider working, and found that it was very programming-focused. But as you say, that might actually be a good thing, if I get it to generate the script(s) that I want it to run. I did communicate with DeepSeek via OpenRouter OK also. But then… I realised that I needed to design precisely what I wanted it to do! And yes… let’s get AI to generate code to do some of this.
Apologies for the formatting of your comment. I did try to format it correctly, but the wretched comments editor won’t play.