Tomorrow is St Nicholas’ Day. I’ve been working hard to complete my translation of the earliest Latin “life” of St Nicholas, by John the Deacon. I pulled it all into one Word document at the weekend. My intention was to read through it today a couple of times, and then get it out of the door. At the moment various urgent but unimportant chores are getting in the way, unfortunately.
But this morning, when I woke up, I found myself thinking about the opening sentence. I’m always wary of translationese, and I started turning it over in my head, thinking about what the writer was actually saying, and how to best express that in English. Gradually I came up with a form of words, not very different from the translationese, and I rushed to the computer to jot it down. This I did, and compared it with the Latin.
But I wrote that material some months ago. Since then I have got far more familiar with John’s word order trickery. And it struck me, at once, that I had made a couple of significant mistakes. In the very first sentence. Disaster! Because if that’s wrong, how many more are wrong?
I must reluctantly conclude that I had better read through all the sentences again, and check. So the release of the translation will be delayed.
* * * *
As I’ve worked on translating the text, I’ve also been learning how to translate it. The process is not linear, but circular. You have to attempt the translation, from wherever you are, and with whatever tools you have. The process forces you to learn how to translate it. You then have to redo your initial attempts.
This used to happen to me sometimes in my programming days. I would attempt to make some kind of change to a mass of someone else’s code. I’d wade into it. And, instead of it getting simpler, more and more problems would arise. I’d learn that I’d gone about it the wrong way.
Unwary programmers would press on, and sometimes I would be called to assist some sad-looking junior, sunk deep in a morass from which they could not extricate themselves.
But I knew better than to do that. This was the benefit of experience, of the mental toolset that I had acquired of ways to do things. I would simply abandon what I had done, and went back to the last version of the code that I knew worked. Then I would start again, but this time with a greater awareness of the problems, and the places where it would all go wrong.
It’s been the same with John. I’ve had to simply charge in and do stuff, somehow, and then find out, repeatedly, that I was wrong.
When I first started, I had the choice of two Latin texts. There was the Mombritius edition of 1477, which was not punctuated in any normal way. And there was the much more friendly looking Falconius edition of 1751, with chapters, punctuation. So I chose that. There was also an 1822 edition by Angelo Mai, but this, I knew, was abbreviated. At this time I knew nothing of the modern Corsi edition, which I was not to obtain access to for many months, nor of the 16th century Lippomano edition.
It was only once I had translated the whole Falconius text that I started looking at the Mombritius text. I did so only out of laziness; because in some places the text is weird, and I wondered if the other was a better variant.
It all started when I got into second sentence of chapter 2, the tail end of which was basically incomprehensible. Falconius had one reading; Mombritius another. I wondered what the manuscripts said, and I found that none of them followed Falconius, and that they varied a lot in that sentence. Falconius wasn’t comprehensible either. But it became obvious from this that Falconius had silently altered the text at this point, without manuscript authority, and probably wrongly.
This led me to scan the whole Mombritius text into a word file, so that I could do a machine comparison with the Falconius text. Of course I had to find out how to do this! I tried various bits of software, which all had too steep a learning curve. My skills are in linux anyway, so I ended up using something odd called dwdiff, and then writing a bit of software to display the output in colourised form. I had to strip out the punctuation, normalise the text to lower case, turn all the v into u, and j into i. Yes, I had to create my own method and teach myself how to compare two Latin texts.
I also scanned the Mai edition for good measure, to see just how different that was; and also the Corsi edition once I had it.
But once I had this output, the extent of the differences between the two editions became very obvious. There they were, line upon line, in black and white – or rather, in red and green!
It was very interesting to see how few of these changes actually affected the meaning in any way.
In the process, it became increasingly clear that the reading in Falconius could most easily be explained as derivative; that the real text of John’s work was a text of Mombritus type, which was modified. The diff also showed that the abbreviated text printed by Mai had plainly been produced by a scribe who had a Mombritius-type text in front of him, chopping out the irrelevant moralising reflections that John added to his story; and, that after a few chapters, the scribe had simply paraphrased what John wrote. Corsi had also started with the Falconius edition, and used a Berlin manuscript, but his edition leaned strongly toward the latter.
We have all sorts of manuscripts online these days. So naturally it’s not too difficult just to download one. And … suddenly I find myself wanting to collate these, to engage in stemmatics, recensio.
But I shall have to find a way to teach myself how to do these next!