Citing papers that are in a language that you do not read. I have a basic question that others have definitely considered.
Often there are papers that originally appeared in a language that one might not understand (and I mean a natural language here). I would like to cite the original paper, because that is where the credit belongs. But on the other hand, doing so violates the golden-rule of read that paper that you cite! What should I do to overcome this dilemma?
So far, I have always cited the original, and if possible some other related work that has appeared in English---but sometimes, reviewers write back that I should not be citing papers written in a language different from English, which is what motivated me to ask this question.
Thanks for any useful advice.
 A: I think a common-sense approach is to cite the original paper (whatever the language) in order to give credit and attribution but only rely on arguments from papers you can understand in your proofs (so you don't violate the golden rule). 
Regarding reviewers, the worst that can happen (I think) is that you use a crucial argument from a paper you can understand but that the reviewer cannot understand. In that case, I think the problem is the same whether the reviewer cannot understand it because he is unfamiliar with the math or because he is unfamiliar with the natural language. In both cases, you, as  the author, should try to present relatively clear references, and that includes translations when appropriate I guess, but ultimately this is a failure of the reviewer. If I were reviewing a paper and found myself in this situation, I would politely ask the author if there is a translation available. If not, I would tell the editors I am not competent, but wouldn't blame the author.
It is a bad idea to upset reviewers, but banning reference in languages other than English (or any other language) even for attribution purpose is an outrageous suggestion that should not be complied with.
A: Partly with the point of summarizing other answers/remarks: there are several points mixed together here. There is the issue of recognizing "prior art" (whether or not one's present work depends upon it). There is the issue of recognizing "necessary background" (often logically prior to the work in question). There is the issue of "competing, perhaps incomplete work" (to be civil, as well as being a mensch and informing readers about it). There is "compliance-necessitated reference", as in "icons" which, if not cited, will make the skeptical quasi-expert reader doubt the competence of the author.
Of course, there have been times in which otherwise-intelligent people did not realize they should write in English. (Sorry, part of my reason for making an "answer" was to publicly form that sentence. I did grow up in midwestern, mid-20th-century U.S., and was led to believe that people who didn't speak English were just being obstructionist jerks, since obviously English was the universal natural language... whew!)
Another perversity: more than one historically-famous mathematician has let slip to me in conversation that they take the viewpoint that not looking at someone else's paper relieves them from any obligation to cite it [even if it is prior work]. Not to my taste.
More perverse than pretending to ignore others' work is the actualy benefit their work may give you/one, even, or perhaps especially, if one hasn't assimilated it yet. As a historical example, it confused me for a decade or two that Siegel and Harish-Chandra (both at IAS) had apparently not communicated... ever?... so that the "holomorphic discrete series", visible in Siegel's and H. Braun's work in the late 1930's, were not visible... and, then, contrariwise, Shimura and other "modern" automorphic forms people seemed functionally oblivious to H-C's work 10-15 years earlier.
Must/should one personally certify anything cited? Well, sure, obviously, this would be desirable. Also, not feasible. Of course, if you're counting on the correctness of an obscure paper, you are on thin ice. If you're counting on correctness of an already-often-cited paper, then you are in better shape. No mystery here. Refereeing does not assure correctness, just makes it a tad more likely. If you claim to prove something scandalous, people will revisit all those innocent-seeming papers you cite. "Just the obvious." (But, contrary to the professional pose that ... oh, ... published papers are first-order predicate-logic ... correct? A needless conceit, of course, and we should not slip into it.) 
The editorial pressures are corruptive. Yes. The professional competitive pressures are potentially corruptive, yes. But I think if we are honest with ourselves we can see what acknowledgements we should make. Prior work, even if we don't use it. Even if we are competitors. Give the reader (supposing they care!) a guide about how to arrive ... here. 
Depending on how one understands the word, I think that, happily (to me) "honesty" is a good guide.
Oop, "was it in English?" What? Um... My own discussion, and all others, indicate that this cannot possibly be a legitimate issue. Write to the editor. But it is a "stimulating" question. :)
A: I think it's completely unreasonable to expect authors to have read and completely understood the details of every paper that they cite in their references. Most authors certainly do not do that. We'd never have time to write any of our own papers if we had to thoroughly digest every single other paper that is relevant to our work. As I see it, the rules should be: (1) don't cite anything you don't refer to in the paper. (2) only cite those papers which are relevant to your work, although it's okay to have some papers in your references which are only cited in the sense of "papers [x] and [y] and [z] also consider similar problems..." and (3) you should completely understand the statement of any result you use. If you haven't actually checked the details of the proof, most mathematicians with experience in their fields can skim over a proof and decide whether or not they believe it. If the reference has been published and later turns out to be incorrect (which sadly happens all too often) then you cannot be held responsible for that.
Of course, the language of the paper you cite is completely irrelevant. I've never before heard of a referee caring about this.
