Erik McDonald has expressed some doubts about my take on the quickly aging Gessen review of AK, so here goes–I’m quoting from his blog XIX vek, of which he sent me a snippet.
“I personally love trying to figure out what’s causing a whole group of translators to read something differently than I read it […], and Gessen’s way of tying the two scenes together [Anna’s arm described early in the book and then very differently later–RSV] seems convincing to me. But Russell Scott Valentino thinks language should be beneath the critic’s notice:
[this is McDonald quoting me] Some readers of Gessen’s review will hear the authority of someone who knows the source and, as a result, they will essentially cede their own authority to make judgments because, well, she knows the source so she must know the right answer! They are hearing the voice of the translation police, which lurks behind every example and in fact informs the entire approach. This approach mostly involves calling up a variety of largely unconnected individual lexical items, selected by the reviewer and held forth with relative approval or disapproval almost as if she were teaching a foreign-language class and telling us which words mean what the translator has said they mean and which do not — but on the basis of a text that exists only in the reviewer’s mind. We certainly don’t have access to it. In fact, no one does. Remember, if you read the original Russian, you’re just a reader of the original Russian; the translation is created in the act of writing by the translator.
[back to McDonald here] I’m left wondering how any act of communication at all is possible in Valentino’s model. Sure, we don’t have access to a reviewer’s or translator’s or author’s mind, and written words aren’t identical to what the writer was thinking at any given moment. But we can make inferences about what other people think, know, or feel based on what they write, or why would we read or write anything?”
I (this is me now) am not quite sure where to start here. Language should never be beneath a critic’s notice. I did not claim that it should. But let me be more explicit and say that making comparisons to the foreign source in a review whose audience is likely not to know the source language is an implicit argument from authority, which says listen to me because I know something you don’t. The fact that others did in fact read her review this way is made clear in the subsequent Slate radio interview. Basically the two interviewers bow down before the authority of the person who knows the source. They don’t know it. She must therefore be right when she makes comparisons to it and pronounces some solutions simply “wrong.”
This last points to another problem, which McDonald’s comment about communication, and Gessen’s about “wrong” translations, both point to, though they get to it by different routes. Gessen gets there by a simple argument from authority. (I know what’s right and I’m going to give you a glimpse of it by telling you the correct translation of these three words.) McDonald does so by means of what 20th-century critics referred to as the intentional fallacy, which basically reduces the meaning of a work to what the author wanted to say. He defends this way of approaching a text by wondering about how communication is possible in my “model.” This conflates several kinds of actions that are not the same. Translation combines interpretation and writing in really interesting ways, but it makes very little use of the common tools of interpersonal communication, especially when the author of one’s text has been dead for over a hundred years.
Thinking about the meaning of a work as being “what the author intended” has a tendency to limit meaning and blind the interpreter to the other, often widely divergent, meanings that other readers in other times and places have found in the very same words. The intentional fallacy has little patience with polyvalence. It likes to have God, or Homer, or Shakespeare mean pretty much one thing. It likes to pronounce other interpretations wrong because they don’t understand what the author intended. This is why I emphasize the text over the author’s intention. The text is what we have. The author’s intention is what we imagine. The text, because it is made up of words that no one owns, can mean many different things to people in different parts of the world at different times. The author’s intention is limited historically, linguistically, geographically. Shakespeare did not mean to say anything about America, yet Americans find plenty of meaning in Shakespeare’s words. This is not because Shakespeare intended us to.
In no way do I want to say that comparisons of multiple translations with each other or with the source text are not worth making, or that they are “beneath the critic’s notice.” Such comparisons and explications are the sorts of things I have engaged in for much of my professional life. And when teaching translation, I frequently use comparative methods, looking at the source when my students have the linguistic means to discuss translation choices on its basis. But textual explications and comparative translation teaching methods are not the same as writing a review.
I am fascinated by the nuances of foreign words and take endless pleasure in comparing them with English ones. But I also know quite well that the moment I begin to talk or write about the details of the source in the company of people who mostly do not share that pleasure, we leave common ground behind, almost as if one of us has read the book but the others have only seen the movie.
There are much better ways to review translations than this.
Russell Scott Valentino 
making comparisons to the foreign source in a review whose audience is likely not to know the source language is an implicit argument from authority, which says listen to me because I know something you don’t.
But surely what you say is true of anything, not just translation. A physicist who reviews a book on physics or a Byzantine scholar who reviews a book on medieval Constantinople and discusses details that the lay reader has no way of checking is also making “an implicit argument from authority, which says listen to me because I know something you don’t.” Isn’t that part of the point of reading reviews, to learn something from a reviewer who knows things you don’t? It would seem that the implication of your theory is either that books should never be reviewed by specialists or that such specialists should never refer to their specialist knowledge, a depressing idea to me. I loved Gessen’s review, and I can’t for the life of me see how readers are poorly served by explaining ways in which translations misrepresent the original book.