Citing into the future on ArXiv – good or bad idea?

Question

I have coauthored two papers, which are strongly related. The younger paper is thus citing the first and the older paper is annoucing the second along the lines of “a study of aspect X will be published elsewhere“. As ArXiv enables you to update your papers, it would be possible to include a citation to the new paper in the old paper after the aforementioned sentence. This might save a reader of the old paper some time with finding the new paper.

However, this breaks some paradigms that were inherently fulfilled by any pre-internet citation, i.e., that you could not cite future work¹ and that there are no loops in citation graphs (i.e., there can be no papers A₁, …, such that A₁ cites A₂, which cites A₃, which cites …, which cites A₁). Thus I find it conceivable that such a citation into the future may cause some problems, for example some weird software behavior (ignoring for the example’s sake that this would arguably be the software’s fault).

Is there any such issue, which would make the aforementioned citation into the future a problem?

---

^{¹ Of course, will be published elswhere existed before, but it could not be accompanied by a regular citation.}

Answer 1

Occasionally two related articles are published simultaneously and cite each other, so loops in the citation graph are OK. See for example this article in The Scientist which describes two papers which do so. I was able to verify that they both cite each other through my university's library.

Answer 2

Well, arXiv paper is on the same level as any other preprint. When a paper is on arXiv, it's somehow not quite different from you putting it on your personal website. Whence I don't see any problem here.

Answer 3

I understand that when you are "updating" your paper, you are creating a new version in the same way a new edition to a book or a republication of the article in another journal or event. If memory serves arxiv takes track of the version numbers. Does it not?

The problem is: a 2nd edition should be cited by its new publication date and location. The same way we do with books. When a republication happens on journals there is usually a note in the header informing the original publication date and location of the article. I do not think this is an internet related phenomena. Version tracking is common in books. Republications are not common in printed journals, but they do happen.

In short: I do not see any problem. You are actually citing the past on a new edition. I think a note in the header in the document explaining that this is always an good idea.