I am a postdoc in theoretical aspects of CS/ML working on a quite theoretical problem. I discussed it extensively with two colleagues for a while. We all dedicated working hours to it, however, none of us got any results. The problem was formulated by me; I was the one majorly driving it.

After about a month almost everyone gave up. But I kept working on it, and I have finally solved it. The solution does not involve any ideas from my colleagues, but they did invest their time initially on it. Should I inform them and ask if they would like to join the paper? Or should I just publish on my own?

  • 27
    Remember that exploring dead ends is also a valuable contribution, even when no part of that exploration gets used. Because without it, you might have had to explore all those dead ends yourself. Jan 27 at 9:56

3 Answers 3


My best guess (pure math, cs background) is that you should be first author and the others have likely gained an authorship position as it is likely that they contributed to the insight you gained, even if by cutting off unprofitable avenues. At a minimum they have earned a strong acknowledgement.

Yes, you should contact them and discuss this. For a lot of reasons, ethics among them. But if working together has produced something now, working with them in the future is likely advantageous.

The path to a theoretical result is seldom straight, with many unprofitable side paths that must be explored. You did that with your colleagues. Acknowledge it.

And even if this seems a bit generous, it is profitable to all in the long run. And long term thinking is good for a researcher and academic.

  • 6
    I wonder if the notion of "first author" exists in the OP's subfield. Jan 27 at 20:48
  • 1
    @MarcGlisse Even if it is unconventional in OP's field, it can be explicitly stated in an "Authors contributions" section at the end of their paper.
    – Perry
    Jan 29 at 12:39

@Buffy's answer sounds reasonable, and most likely no harm will come to you from adopting its recommendations. Potentially there will be benefits, since you will make your collaborators happy and that can end up helping you in various ways.

However, let me add this: if you honestly think that your coauthors genuinely contributed nothing intellectually, then personally I don't see why they should be coauthors. Offering them coauthorship would be a form of gift authorship, and would be ethically dubious. The fact that each of them spent a few hours trying to make a contribution is irrelevant.

Well, I'm not sure if you actually think they contributed literally nothing. By your description, this sounds like just enough of a borderline case that if you make the decision of what to do on purely utilitarian grounds, then whatever that decision is I think no one is likely to view it as some kind of moral transgression. However, from a more utopian or idealistic point of view, my point is that the most intellectually honest thing to do would be to not offer coauthorship to a person who didn't contribute either to the formulation of the problem or to its solution. If you did everything yourself, I'd prefer to live in a world where you get all the credit.

  • For all we know, the intellectual contributions and discussions may have helped the OP arrive at the solution more quickly, or at least, not waste their time on fruitless ideas.
    – Mari-Lou A
    Jan 27 at 7:32
  • 5
    @Mari-LouA sure, we don't know the details of what happened, so we can't give a definite answer. But I suspect that OP does know whether the collaborators' suggestions helped or not. If they didn't, my answer would apply.
    – Dan Romik
    Jan 27 at 7:34

I think @Buffy's answer is generally excellent, as it seems so does the SE community.

However, I'ld like to suggest a supplemental caution/refinement.

I'm wondering; even "quite a theoretical problem" in "theoretical aspects of CS/ML" may turn out to have practical applications which lead to other events related to awards (with monetary and career-building impacts) and intellectual property filings later on, and later transactions related to said IP, which could be subject to becoming unexpectedly contested.

Those may not be specifically limited to this one result, but might also apply to some derivative work based on it.

While a bit of "generosity" can certainly be "profitable to all in the long run" in the academy, a lack of accuracy and/or clarity in a published attribution of effort may have unforeseen effects later on.

@Buffy says, correctly:

At a minimum they have earned a strong acknowledgement.

Acknowledgments offer the opportunity to define and constrain the nature and extent of the acknowledged contributions.

Some journals provide a space to outline the extent and nature of the contributions of each author beyond what we can infer from the ordering of the names. You can consider where you'd like to publish, and if it offers such a space and authorship is offered to others, it's best to agree upon the wording there at the same time, rather than surprising folks later and suddenly have a potential second-order disagreement on your hands.

You must log in to answer this question.

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