I suspect many of us have grappled with this issue.

So you’ve got a research program, with at least two main parts, X and Y. Working out all the issues associated with both X and Y simply cannot be done in an article. It will take at least two, and probably three or four, and eventually will lead to a monograph. But for the moment, you’re trying to publish your results on X and your results on Y, since this will contribute to an ongoing debate, and let you stake your claim to some promising ideas.

At some points, your work on X inevitably depends on what you have to say about Y. But, as I said, you just can’t do it all in one article. And you’re honest about this. At the relevant point, you say, “Here I depend on what I’ve argued elsewhere,” or something to that effect. But referees seem to very much dislike this. They say things like, “The author points out that the argument here depends on other work. But this really needs to be worked out here. Otherwise, the paper isn’t publishable.”

So what’s the solution? Here are some possibilities:

  • (A) Forget the articles. Just publish a book.
  • (B) Appeal to the editor.
  • (C) Keep submitting the articles to good places, hoping for a more sympathetic referee.
  • (D) Don’t be honest. Don’t point out the dependency. Hope the referee doesn’t notice.
  • (E) Give in and write a really long article (c. 20000 words). Hope that the referees don’t say “This is too long. It should be broken up.”

Option A forgoes some potentially nice publications in the short to medium term, which is difficult to do for junior people. Perhaps it’s an easier choice for more senior people, though even they run the risk of getting scooped, especially if the research project deals with a “hot” or emerging topic. The expected utility of option B seems quite low, especially at the very good and elite journals, where they get a surfeit of excellent submissions, in which case they really have little incentive to intervene. I can’t see myself choosing option D. I’d be leery of option E. Option C isn’t so great, but maybe it’s the best of the lot.

What would you do? Is there a better option?