5

I would like to cite a theorem in my paper. It's not famous, but it is used by many people.
So, I don't know who first proves the theorem.
Is it okay to reference the theorem in the source that might not be the first paper to prove it?

alryosha
  • 153
  • 4

1 Answers1

11

Is it okay to reference the theorem in the source that might not be the first paper to prove it?

It’s not only “okay”, but you must cite at least one source you are aware of, and preferably the chronologically earliest one. If you can’t trace the exact provenance of the theorem, then do the next best thing, which is to describe the earliest possible source, and explain that you were unable to determine if that represents the original source or a later source.

Of course, saying this carries with it the assumption that you made reasonable efforts to determine the source, for example by asking your adviser, colleagues, or emailing the author of the paper where you found the theorem if they are still around and easily reachable. If you didn’t make such efforts, then that’s what you should do first of all. But if you did make an effort, then sure, proceeding as I described above is “okay” and is the standard way to proceed.

Dan Romik
  • 189,176
  • 42
  • 427
  • 636
  • 3
    In general i agree, OTOH if the theorem was used in a new field, if may be appropriate not to cite the original proof but the first paper where it was applied to a new problem, however the referencing should make that clear. – Sascha Sep 21 '20 at 11:58
  • Not to deliberately pass the buck, but reviewers may be able to provide a more appropriate citation for the theorem. – chepner Sep 21 '20 at 16:05
  • @Sascha I think in that case I would cite both the original and also the first use in the relevant field – GageMartin Sep 21 '20 at 20:36
  • @GageMartin I actually like things like "[n] and references x-y cited therein. It can make the genesis of something more clear, since it indicates that the author got the specific idea to use it from a specific paper. – Sascha Sep 27 '20 at 12:50
  • @Sascha Oh yes something like that sounds nice. I just meant that I wouldn't Only cite [n] without also citing what [n] drew on – GageMartin Sep 27 '20 at 15:18