I managed to finish reading Faulkner's Light in August by the end of August, with just a day to spare (only 8 years late according to Oprah's schedule for her summer of Faulkner, not that I had signed up for that or anything). So now I turn my attention to Proust. I have been carrying around these tomes for many years, and I have read the introduction and first few pages a few times, though I always knew I wasn't quite ready to start. Now I finally am...
The introduction mostly discusses the need for a new translation (this the the early 80s updating of the Scott Moncrieff translation from the 1930s). Essentially there was a major overhaul of the text of À la recherche du temps perdu by the French publishers in the 50s, and this was an attempt by Kilmartin to update the translation to this new text, starting from Moncrieff's version. Now the French publishers have continued to update/correct the text, and perhaps it was inevitable that they would issue a "definitive" version in 1987. In short order, Modern Library issued a translation by D.J. Enright that used the Moncrieff-Kilmartin as its starting point. (Penguin is going a different tack, with a different translator for each volume. This strikes me as a terrible idea, and I have no interest in these translations.)
My gut feeling is that the improvements to the text between the original and the 1954 version are crucial, and must be incorporated into the final translation, but that the improvements from 1954 to 1987 are much more subtle and only a marginal improvement (with some critics clinging (stubbornly?) to the 1954 version). What I do think is important is that Moncrieff serve as the base. Essentially all English mono-linguists have come to Proust through Moncrieff or a revision of Moncrieff, and I don't see any good reason to break with this tradition. More to the point, I know I'll never have the time to read a second translation of Proust, so it's either stick with the Moncrieff-Kilmartin that I already have or switch over to Moncrieff-Kilmartin-Enright. But Proust for me is the silver-colored three volume set of Moncrieff-Kilmartin, so the issue is settled fairly easily.
I feel even more strongly about when studying the Bible as literature, one needs to use the King James Version, since that is the version that influenced 400 years of English literature. I think I have mentioned this before where a professor decided to use the Revised Standard Version in a class literally titled The Bible as Literature. Now I suppose the RSV came out in 1901, but I still don't see it having the cultural impact of the King James Version -- and I still stick to my guns that the professor was simply misguided in his approach. But I guess you just can't be too precious about this. Shakespeare wasn't influenced by the King James Version, since he had written all of his plays before it was issued in 1611 (aside from possibly The Tempest, Henry VIII and The Two Noble Kinsmen (written with John Fletcher)). More to the point, Shakespeare would have been influenced by the version that he was available when he was a child and young man (most likely this would have been the Geneva Bible). I was particularly curious about Faulkner, since his books are just infused with Old Testament values of vengeance and over-the-top spirituality. It turns out his copy of the Bible was a Holman's Edition Bible (though the version Faulkner would have had seems to basically be a minor variant of KJV). Melville is perhaps the other American author that immediately springs to mind when I think of being deeply influenced by Biblical themes and language, and he clearly was brought up on various editions of the KJV Bible.
Now you can take this too far (only reading texts that other writers would have had access to/been influenced by). Most 20th Century readers of Dostoevsky (and Tolstoy) came to him through the Constance Garnett translations, but there are some problems with her approach. She did flatten the language a bit. In those cases I have really investigated, the Pevear and Volokhonsky translations have always been the best,* though for old times' sake, I'll hang onto my Garnett translation of Crime and Punishment. Indeed, whenever I fit Crime and Punishment back into my reading schedule (2020 perhaps?), this might be another time where I actually read two translations back-to-back (I did this for Bulgakov's Master and Margarita as well), which can be rewarding, if exhausting. But normally you don't have time for this level of effort -- and it makes you think too much about the fact that you can't access the original, which can be a bit depressing. There is no chance that I will ever read Russian, but if I have all the time in the world and do revisit Proust, I would go ahead and read (in French) one of the corrected versions from Gallimard that came out in the 90s. But not the one that essentially doubled its length by including all textual variants -- life is definitely too short for that!
There are lots of nooks and crannies of literature where I just don't have a strong opinion** -- the best translation of Homer? Of the remaining Greek plays? Or Virgil? I try to just pick one and then not get too fussed about it. I definitely am not going to keep seeking out multiple translations of ancient texts, though Homer is "important" enough that I will try to read both the Lattimore and Fitzgerald translations and see if I can settle on one. (Notice that again, I will stop with translations that were current (or at least around if just a bit dusty) from my days at university and generally not take too much interest in more current ones, with the exception of Pevear/Volokhonsky and Grossman's newish translation of Don Quixote.)
Edit (12/17/2016): I'm considering writing a separate post on Kafka, where there are quite a few dueling translations. I'm also trying to decide just what to do for my own reading program. For myself (and most Americans), the translations by Edwin and Willa Muir (with supplemental material added to The Trial and The Castle) in the 1950s are the ones that basically informed my understanding of Kafka.
Essentially everyone agrees that the Michael Hofmann translation is considerably better than the previous version of Amerika (aside from the great Edward Gorey cover -- incidentally this site has some of the other great Gorey covers).
It is more debatable whether the other Muir translations have been truly surpassed by Mark Harman’s translation of The Castle and Breon Mitchell’s translation of The Trial (both published by Schocken), or indeed the somewhat unheralded translation of both by John Williams in Wordsworth Classics' The Essential Kafka. I think for old times' sake, I will reread the Muir version of The Castle (just as I already reread their Trial last year), and then in a few years I will compare the Schocken and the Worthsworth editions and decide. The Trial is short enough I could probably read both translations, but I think I would pick only one Castle.
* These debates can get quite heated with some of the translators with a lot at stake getting quite peeved at each other, though that may be nothing compared to their fans! Here is an interesting article on Russian translations, though the piece suffers a bit from seeming to side too strongly with Pevear and Volokhonsky.
** Of course I reserve the right to have a strong (if not necessarily a well-informed) opinion on translations outside my area of expertise. So for instance, Dante should be translated as poetry (of course) and the only one that ever seemed to do the job was by John Ciardi, and I don't see any point in seeking out the newer one by Pinsky (and indeed it seems he didn't get around to translating the entire Divine Comedy!). Not being able to read Italian doesn't prevent me from deciding definitively on this.