Consider the first verse of the gospel of John, and considerits usual English translation:
| Εν αρχη ην ο λογος και ο λογος ην προς τον θεον και θεος ην ο λογος | In the beginning was the word, and the word was with God, and the word was God. |
Now consider retroverting the latter back into Greek.Chances are that a translator, lacking any knowledge of theGreek, would produce something like
Εν η αρχη ην ο λογος
και ο λογος ην συν τω θεω
και ο λογος ην ο θεος.
Note that, while the English translation is more or less anadequate rendering of the Greek (except, perhaps, for theinteresting flavour of the Greek preposition προςinstead ofσυνorμετα),it is simply impossible to move from the English to the Greek. Itdoesn't preserve the same attributes.
This is a constant difficulty, and one rarely addressed ineither the manuals of textual criticism or those ontranslation; both leave it for the other.
Fundamentally, to the translator, variants can be classifiedinto four groups based on two criteria:
1. Meaningful variants, and
2. Translatable variants.
The former list is almost the same from language to language;the latter differs from tongue to tongue.
Using English as our target language, let's give examples ofeach class:
It will presumably be evident that variants of the first classare the most important, and variants of the last class can beignored. We will return to this subject later.
More complex are the cases where the distinction is blurred.Take the disciple whose name was either Lebbaeus or Thaddaeus.This is clearly a translatable distinction. But is itmeaningful? Not necessarily, since neither name occurselsewhere in the New Testament. If this disciple had been called"James Francis Edward Stuart the Old Pretender,"it might set us wondering about anachronisms, but it wouldn'taffect the plot, if I may so call it, of the gospels. It wouldaffect synoptic studies, but those should be conducted basedon the Greek text anyway.
Or, similarly, take the parable of the two sons (Matthew 21:28-31).We know that the son who went to the vinyard is the one who did thewill of the father. But is that the first or the second son? This isa difficult question textually. The meaning, however, is thesame either way. Is this a translationally significant variant?
There is also the question of textual support. Matthew 1:16 hasa major variant concerning the paternity of Jesus -- but the realvariant is found only in the Sinai Syriac. Is that enough reasonto note a variant? Or 1 John 5:7-8 -- the work of a known heretic,with no significant textual support at all. Is that worth noting?
So which variants should go in the margin of a translation (if any)?
The answer to this depends very much on the intended audience ofthe translation. Obviously a translation intended for children shouldnot have any marginalia at all if it can be avoided. But a translationfor educated adults certainly should note places where the text isdoubtful.
The number of variants still depends on the intended audience. Aswell as on the style of the translation. A severely literal translation,we should observe, ought to have more textual variants noted in the margin,because readers are trusting it to say what it says. By the same argument,a translation with a high number of marginal notes on the translationshould have a high number of textual notes, because the text affectsthe translation.
The obvious temptation is to take the United Bible Societies'edition -- which, after all, has variants selected for translators --and simply follow the variants there, or perhaps those marked as beingof only the "{C}" level or higher, indicating significantuncertainty. This is presumably why they provided those variants, andfor a translator with no text-critical background, it's certainly betterthan nothing. But there are several problems with this. First is thefact that it is generally conceded that the UBS editors are overconfident --the fourth edition, in particular, marks many variants as more securethan they should be. Second, their selection of variants is somewhatquestionable. And third, there is the problem of how this will beused. My experience is that the notes in a translation are most oftenused by groups such as small Bible Study classes. These groups willusually have several translations in use -- including, perhaps, someonewith a King James Bible. The UBS apparatus omits many variants where itdiffers significantly from the Byzantine tradition behind the King JamesBible. A good translation needs better notes than UBS provides.
A point I don't often see addressed is the different typesof marginal notes. Typically, a translation, if it has notes at all,will feature both notes on the text and notes on the translation. This,of course, is perfectly reasonable -- but it's not obvious that theyshould be grouped together. (Note that most other critical editions with marginal glosses -- e.g. editions of Shakespeare or Chaucer -- havetextual and linguistic glosses firmly separated.) In the case of translational differences,you put the best rendering you can in the text (either you thinkανωθενmeans"again" or it means "from above," but it meanswhat it means). If something is in the margin, it's a less likelyrendering. Textual variants are fundamentally different: Only onecan be correct. There is no doubt of meaning; there is doubt ofreading. It makes a different demand on the reader. A note on thetranslation often makes our understanding of the text richer. Buta note on the text says that there are two different traditions aboutwhat is read here.
Then, too, most editions don't really indicate the nature ofa variant. Is it highly uncertain? Is it included only becauseit's found in the King James Bible (e.g. 1 John 5:7-8)? Admittedlya translation probably shouldn't be a textual commentary. But a strongcase can be made that it should be more than it is: That it shouldinclude nearly every translatable and meaningful variant where thereis significant doubt about the text, and that it should also includetranslatable and meaningful variants where the reading is not reallyin doubt but where some well-known edition has included the readinganyway -- and that these two classes of readings should be clearlydistinguished.
There is no absolute and final rule for how to deal with textualvariants in translations. There is no doubt in my mind, however, thatmore needs to be done.