Is there any point in translating ancient texts

All of us know that the internet has revolutionised our access to ancient texts. 

First sites like CCEL came into being, back in the mid-90’s.  This made the Ante-Nicene, Nicene and Post-Nicene Fathers accessible to us all.  Indeed I remember, long ago, seeing a bound set in 38 volumes of that collection, in Mowbrays Bookshop in Kings Parade in Cambridge.  I was interested in the Fathers even then, but such a thing was far beyond my slender financial means.  But with the internet came CCEL, and suddenly we took it for granted. 

Google books came along a few years ago.  I don’t even remember when it arrived, so much do I take it for granted, but after 2005, certainly.  That gave us access to vast amounts of older literature, of scholarly series such as the PL, PG, the Bonn series of Byzantine texts and much, much more.  All this was given freely, and with the generous aid of American universities like Harvard.  European publishers poisoned the gift, and by their bleating for money made it largely inaccessible; but Google meant us all to have it.

Manuscripts are coming online as well, despite much resistance.

Now we have Google translate.  This improves constantly.  For French it is now very good indeed, and doubtless other languages will improve over time.  Latin has been added already.  All this would have been unimaginable as recently as 2005.

Now let us look into the future; a future that may be no further away than a handful of years.  As translate improves, will there be any purpose in providing hand-made translations?

When I first came on the web, CCEL was all there was.  I sought to help, by scanning more translations and placing them online.  Then Google books came along, and made much of this work redundant.  If you go to, or Google books, an OCR of these older translations is generated automatically.  The books are searchable.  Yes, it’s not perfect; but we can always get the text, and often it is very, very good.  So there is now very little purpose in my duplicating this effort, I sometimes feel.

Instead I have been translating stuff, commissioning new translations, and so forth.

But will this go the same way?  Will it too, one day soon, be pointless.

I’m not sure, I admit.  For one thing, digitising texts is still worthwhile.  When I want a text, I rejoice if I find it at Lacus Curtius, accurately typed in and easy to search.  I look there in preference.  Probably other sites like mine are also used in this way.

Will it be the same for man-made as opposed to machine translations? 

Note: I have several interesting emails in my inbox awaiting answers.  Unfortunately I have gone down with the headache bug, so it will be a day or two before I can reply.

6 thoughts on “Is there any point in translating ancient texts

  1. It is still preferable to have the texts available at a website rather than Google Books. I love your sites of course but as a Canadian with some knowledge of French I quite like Remacle with its side by side Greek and French translations. I find it amusing to see how loose and at times superficial the French translations tend to be when compared with the more literal English and German renderings! That’s not all of course. It is all a matter of convenience and you and people like you have made things very convenient for lazy (pseudo)scholars everywhere.

  2. A couple of comments:

    BabelFish (like Google Translate) has been around since before 2005 (I think it may even have originally been made back in the late 90’s), and I’m sure that there have been other machine translation systems before it.

    I suspect Google Translate will never be that good at translating dead languages due to technical limitations. Google Translate works by storing a pile of translations that it already has on record for a word, or phrase, or sentence. It gets these translations by looking at the various works in Google’s databases that already have translations between the two languages. This means that language pairs like English-Spanish or French-English will work fairly well, but pairs like Turkish-Welsh won’t work very well simply because of the numbers of web pages and books available in both languages. This is a huge limitation for dead languages because unless there is already a lot of works in the same language (and even probably from the same approximate time-period) with English translations it simply won’t be able to cope. This can be overcome of course with completely different technology (and may even work for particularly popular dead languages like Latin and Greek, but I’m still doubtful).

    I wouldn’t be expecting any cheap, accurate machine translation for anything but the most popular of dead languages (and perhaps not even them) for decades to come. Of course some sort of revolution in AI may change this, but I wouldn’t be betting that way.

Leave a Reply