A paper that I need to refer to reasonably often has in its arXiv version the authors listed alphabetically by surname (as is the custom in mathematics). The published version lists the authors in a different order, having moved the last author to second place. The last author in the published version is recently told me he doesn't understand why people keep insisting on using the published author order (or to put it bluntly: why they keep listing him last).

What would be the correct way to refer to the joint work, given that the arXiv version is no less prominent than the published version? (Other than in the bibliography, where I would stick to the published order) I'm not sure whether it is more accurate to talk about the work/results of ABCD or of ADBC.

  • 5
    I would use the order corresponding to which version you cite. Thus, use the arXiv order if the citation is to the arXiv appearance and use the published order if the citation is to the published version. Commented Nov 21, 2017 at 10:10
  • 3
    I would actually cite both versions, with cross-references between the two. "[1] Alpher, Bethe, Gamow, and Deltoid. Twinkle twinkle little star. Preprint, October 2009. Arxiv:0980.1234 [astro-ph.CO]. Preliminary version of [2]." followed by "[2] Alpher, Gamow, Bethe, and Deltoid. Twinkle twinkle little star. J. Shiny Objects 14(3):123–194, 2010. Archival version of [1]." Ink is cheap.
    – JeffE
    Commented Nov 21, 2017 at 20:16
  • 3
    @JeffE since the OP has stated the two versions are identical, I'd suggest merging the two citations: [2] Alpher, Gamow, Bethe, and Deltoid. Twinkle twinkle little star. J. Shiny Objects 14(3):123–194, 2010. Also available as arxiv:0980.1234. (Ink might well be cheap, but the reader's time is not. The single citation hopefully suggests the versions are identical, whereas multiple citations might indicate distinct content.)
    – user2768
    Commented Nov 22, 2017 at 9:49
  • 1
    @user2768 But the two versions are not identical; they have different author orders! And it’s impossible to tell which of the author orders is “correct”, so I won’t (and shouldn’t) try. Thank Alpher et al. for wasting the readers time (or the reader for actually reading the bibliography(?!)), not me.
    – JeffE
    Commented Nov 22, 2017 at 14:03
  • 3
    I'm not suspicious of the soundness of the paper, @Dilworth, since I understand the results myself, have generalised them in my own work, and know at least three of the authors personally. I don't know why this happened. Commented Dec 21, 2017 at 0:35

4 Answers 4


As Aaron Montgomery's answer states, if you use the full author list in the text, it should match the author list in the bibliography. However, with four authors, you're usually in a position where you can just say

A et al.

in the text rather than

A, B, C, and D


A, D, B, C

  • 4
    I find this trick particularly useful when writing A et al. [1,2], where [1] A, B, C, and D ..., and [2] A, D, B, and C ...
    – user2768
    Commented Nov 21, 2017 at 8:53
  • 4
    This is really unfair to B, C, and D. Commented Dec 20, 2017 at 14:06
  • 7
    Writing "A et al" in a field that (usually!) uses alphabetical order of authors is not a good practice. Commented Dec 20, 2017 at 14:48
  • Are there a lot of papers in such fields that have lots of authors?
    – aeismail
    Commented Dec 20, 2017 at 19:40
  • 2
    In this paper alpha.math.uga.edu/~pete/Main_Theorem.pdf I refer to a result of "Alon-Kleitman-Lipton-Meshulam-Rabin-Spencer." It is rare but not unheard of for math papers to have more than six authors. But I have never seen a paper in pure mathematics with so many authors that it would add any significant length to one's paper to list all the names in full. Commented Dec 21, 2017 at 6:13

I would defer to the order in the version that shows up in your bibliography, presuming that one of those made it in. If not, or if you're somehow citing both, I personally would use the peer-reviewed ordering, but I doubt anyone would quibble either way.

  • I doubt anyone would quibble either way one of the authors is, as I noted in my question.... And I would stick the arXiv reference with the published version in my bibliography as a matter of principle, since it is a freely available copy: there's no reason to cite both separately. Commented Nov 21, 2017 at 1:56
  • I mean, if you're worried about personally offending one of the authors, I suppose you have your answer. Commented Nov 21, 2017 at 1:59

This is probably an error on the part of the journal. This happened to us once on a four author paper, because the journal got confused because they typeset 4-author papers in a 2-by-2 grid and then messed up the bibliographic information when they read it off the grid. We’ve tried to get it fixed, but it’s essentially impossible. Go with the version that the authors have control over.

  • 1
    Didn't think of that. I could ask A, the senior author (coincidence in this case that they are alphabetically first) if that happened. Commented Dec 20, 2017 at 23:49

I have spoken with D, who was placed in second position in the author listing for the published version without their knowledge. They've said they don't mind me referring to the paper placing them last, as in the arXiv version. I'm not going to hack BibTeX to fiddle with author order in the citation keys, but if I use the authors in a sentence I will use "A, B, C and D".

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .