In my last post I'd gotten through Mark's fifth post. In the sixth one, he discusses the difficulty of translating into a changing language. Linguists tell us that English has changed less since the standardization of spelling and grammar in formal media , which lessens Mark's point a little bit, but he's still right. As English changes, younger generations will have less familiarity with the forms of language in an older translation, and the NIV is old enough that it has forms that sounded ok to its translators, many of whom were old in the 1970s when the NIV was completed. Mark points out that those who grew up with the NIV wouldn't notice this, because they learned English with those expressions as part of it, but the biblically illiterate have much less of this. This point can be taken to show more than it does show, but Mark simply makes it without concluding much yet, so I'll leave it at that for now.
Mark gives a brief explanation in post 7 of why it's inaccurate to describe translations as more literal or less literal but better to describe them as formal equivalence or dynamic equivalence. Formal equivalence pays more attention to the meanings of atomistic words, and dynamic equivalence pays more attention to the meanings of larger phrases.
Consider Luke 22:31. Jesus says to Peter, "I will sift you as wheat." This is the closest way tro render the form in English, because it doesn't involve supplying extra words. However, one element is lost. The word for 'you' is plural. Peter is not plural. It's not just Peter who will be sifted as wheat. The TNIV changes the NIV reading to "all of you". There's no "all" in the original, so this is a more dynamic translation, but what it sacrifices in that respect is less than what it gains by translating the number of the direct object, an aspect of the form of the original that can only come out in a dynamic translsation due to English's insensitivity to a distinction in Greek.
I don't disagree with what Mark says here, but there's a little bit more to say about why it's inaccurate to call formal equivalence translations more literal. The primary reason is that you can lose as much of the literal meaning in translating formally as you can in translating dynamically. You don't carry through the entire meaning in either case in some passages, and decisions between the two aren't about literalness. If a word's contribution to the meaning of a sentence is as part of an idiom, then translating formally loses the plain meaning of the text in its original context. Calling a dynamic translation (in this sort of case) less literal just misses the point. The literal meaning of the word doesn't play a role in the meaning of that phrase, so a literal translation of the sentence does not involve literally translating the words of the sentence.
The issue is over whether you translate the words literally or whether you translate the larger units literally. Whether a word is being used metaphorically often depends on the larger context, and idioms are not always metaphors (as 'driving me up the wall' and 'go against the grain' are) but sometimes have a literal meaning (as with 'skinny dip', which is not some metaphor about thin people being dipped or 'I gotta go' in the context of a full bladder). In non-metaphorical cases, the meaning of the idiom does not depend on the literal meaning of the individual words. 'I gotta go' has a literal meaning. It's an ellipsis for 'I gotta go to the bathroom', which is itself an idiom for 'I need to empty my bladder in the bathroom'. None of that is non-literal. It's still an idiom that you wouldn't translate into another language by translating each word literally. You don't arrive at the literal meaning of the expression by thinking about it that way. Another way to think of this is that 'he is driving me up the wall' has a literal meaning, and it's not the one people mean when they say that, whereas 'I gotta go' has a literal meaning, but the meaning of the phrase is more than the literal meaning of the combination of words according to grammatical combinational rules.
So that's why I'd say a little more than Mark does against the idea that formal equivalence translations are more literal. They're more literal if you care only about literal meanings of individual words, but they're less literal if you care more about literal meanings of larger units of text. Mark goes on in post 8 to illustrate one way dynamic and form equivalence translations will tend to differ. He lists a whole bunch of translations of a verse about Jesus reclining, and even the most formally equivalent translations suppy "at table" or some such expression. They also tend to translate "was reclining with" as things carrying more meaning than what we think of as reclining, which is simply lying down. Since for Jews at the time, reclining involved a position on one's side eating, what's often called the literal translation just fails to capture what reclining in that context means. So the so-called literal translations might translate what the ESV translates as "was reclining with" as "was dining with" or "came as guests to eat with".
Mark summarizes, "Because though formal equivalence preserves the word meanings and forms of the original, it cannot render these in the most natural and intelligible English. Thus formal equivalence translators consistently abandon a strict word-for-word approach in order to make their translations intelligible and readable." It makes little sense to accuse these translations of adding to or subtracting from God's word. They're translating what the expression means. They're thus getting a more accurate translation for most English speakers than the ESV's more word-literal translation. For someone like me who knows what reclining to them was, the ESV is the best translation. Most Bible translations aren't geared toward me, though. They're geared toward the average English speaker.
It's worth noting, however, as Mark does, that dynamic translations can go too far in supplying information that would not just be assumed by the ordinary reader of the text in its original context. The NLT does this by saying that Matthew invited Jesus. Maybe he did, but the text doesn't say that. The NLT goes too far in speculative reconstruction of the situation. This kind of thing is even inappropriate in a commentary, unless it's specified that it's no more than an intelligent guess. Putting it in a translation is even worse. This is a danger of dynamic creation. At the same time, the opponents of dynamic translation like to points this out as if it makes dynamic translation in principle less accurate. Nothing could be further from the truth. As the example above shows, dynamic translations is often required to get an accurate rendering of the meaning of the text.
Mark makes a side comment about the NIV using scare quotes for 'sinners', while the TNIV removes them. The reason the NIV uses them is because in the overall context of the gospel it's clear that sometimes people in Jesus' day would use that term inappropriately, but since it was the label used the gospel writers used it. They didn't mean the people were really sinners. It was sort of an ironic use. We sometimes signal such irony with scare quotes, but we often don't. The TNIV translators acknowledged that since the original text doesn't signal the irony in such a way, and you have to pick it up from context, you therefore should have to pick it up in context in English if a translation is to be accurate in that respect.
Mark considers this odd. I don't. The TNIV makes quite a number of changes from the NIV that move it more back in the direction of formal equivalence. This is one of them. Every one I know of was deliberate and in response to criticisms of the NIV. There are many ways the TNIV is much more accurate than the NIV, some because of moves like this toward formal translation, some because of moves toward more dynamic translation. The favorable reviews of the TNIV on the TNIV website detail these changes, and very few of them have anything to do with gender. I'll give a couple examples and refer you to Craig Blomberg's review of the TNIV for more.
'Miracles' becomes 'signs' or 'works'. 'Peter' is now 'Cephas' when that's how the Greek refers to him. Rather than translitering the Greek 'Christos' as 'Christ', the TNIV translates it as 'Messiah'. The feminine form of a pronoun in Matt 1:16 makes it clear that Jesus was born of Mary and not Mary and Joseph. The form of the NIV's translation obscured this. The TNIV fixes it. Luke 1:15 in the NIV translated a word having to do with from the womb as "from birth". The TNIV is more accurate. The NIV left out a whole clause in II Cor 1:23 that the TNIV now htranslates. Phil 3:6 is "legalistic righteousness" in the NIV and "righteousness based on the law" in the TNIV. The latter is much closer to what's normally referred to as a literal translation.
This isn't to say that all the changes in the TNIV are in the direction of being more formally equivalent. Some are more dynamic than the NIV, and sometimes I think it means more accuracy. One of my biggest annoyances with the NIV is its translation of Philemon 6. Paul is talking about the fellowship he has with Philemon in the faith, and the NIV translates this as "the sharing of your faith". I've seen many an evangelist taking that out of context as referring to evangelism, and that's in fact what the average contemporary reader of English would take the NIV's rendering to mean. It's less formally equivalent to translate the expression as "your partnership in the faith", as the TNIV does, but it's much more accurate. The Greek 'koinonia' is most accurately rendered by the English 'sharing' if you don't factor in contextual hints as to meaning. The problem is that the NIV rendering of it does have contextual hints that shift the meaning from fellowship to evangelism, and it's simply inaccurate.