There are occasional Science Ethics questions that must be confronted. Sometimes they are big: someone falsifies their data in which case a big storm ensues after it is reported. Sometimes they are small: someone misspells your name in a presentation and you want it fixed.
And then sometimes they are interesting from the point of view of `what is right’. So here is one of those.
- A publishes a paper with a partial analysis on day 1.
- B, C, D, E publish much more complete analysis between day 10 and day 60.
- A revises their paper with a complete analysis on day 61 (essentially writing a new paper, but not quite) and does not cite B,C,D,E nor any subsequent paper to day 1. The rationale being that the papers came after the posting date of the original paper and the work was mostly done anyhow.
- E notices the replacement and believes that A should have cited B,C,D,E.
Because A just filed a replacement, this will probably go under the radar unless E speaks up. Unless the referee of the journal is very aware of all the details of the history submission of various versions of the paper etc, he will probably not notice either and then B,C,D,E will look like partial analysis after the big analysis, etc etc.
The moral issue is twofold:
What should E do? Assume that E contacted A and A sticks by their rationale.
Is A right? (is their rationale a `good’ rationale or is it flawed?)
My two cents: primary role of citations is as service to the reader. Obviously BCDE papers would be interesting to the readers of A, assuming they are correct and interesting, so they should be cited. This would be more of a dilemma in my eyes if we’d be considering issues of credit.
But it is an issue of credit. There is some revisionism of history. Especially if B,C,D,E try to publish their papers and they get a response like
‘This was already done in paper by A’.
I don’t know if this is the case.
The issue at stake is also whether the revision should really have been a new paper, rather than a revision. Sometimes the results can be quite different from a partial analysis to a full analysis.
The main problem I see here is a lack of tracking of changes to the original paper. Because A was allowed to replace their original paper such that it looked like the revision was the original, all subsequent work based on the original version unfairly looks derivative, redundant, etc.
It is entirely possible that A’s rationale was accurate, that the analysis A did was independent of the work of B, C, D, and E. I don’t know if scientific paper publishing culture generally requires the mention of closely related, but independent, work done by others, but if not, then A’s rationale could be perfectly valid. If A’s rationale is not valid — if, in fact, A’s analysis wasn’t independent of the work of B, C, D, and E — then there is a bigger ethical concern at work here.
In my opinion, E has a valid grievance to bring to the journal referees and editors. The journal’s policy of allowing A (or, presumably, other authors) to file a major revision as a “replacement” without some sort of indication of a change noted for the record can be detrimental to the professional reputation of any scientist who publishes based on the original version, but not the revised version. The policy could, and in this case has, distorts the historical record.
E should ask, and should receive, at the minimum an editors note attached to the paper that mentions the original publication and revision dates. That way, anyone following up on the citations in B, C, D, and E’s papers could see the dating discrepancy (that their papers were published after the initial version but before the revision) and understand that B, C, D, and E’s work wasn’t redundant.
That’s just my opinion.
Agreed, but since it is not purely an issue of credit, I can weasel out of this one…
I should have pointed out that ‘publishing’ can also mean publishing a preprint on the preprint arxiv, where there is a track record of past versions.
hm, interesting. The `right’ answer must depend on the technical details: typographical errors do not require new papers, and major errors in the main result do. In between there is a continous spectrum, and at some point people of good faith start disagreeing as to how significant a result is. One person’s trivial details is another person’s rigorous result, and this may be due to an unavoidable difference in scientific style.
A related interesting question is if the main result is wrong, and crucially wrong, but due to an essentially trivial error. An example would be the sign of a beta function. Where would credit belong then, to those who did the hard calculation or to those who found a relatively trivial mistake?
What E is always free to do if they feel strongly enough is to draw attention to the issue in future papers (or a revision of their own paper), by conspicuously citing the two arxiv versions of A separately and as a service to the reader explaining the difference. Whether they want a public fight by doing this is another matter.
piscator
ps this post doesn’t relate to m2 branes does it?
Hi Piscator:
No relation to M2 branes.
Also, the option of a public fight would bring even more attention to the paper and hence more citations… making the credit issue even more skewed. Not to mention that the prestige currency is very important in scientific communities. This goes also by the reputation of how much you like to fight, etc.
Also, there is no `right’ answer in an obvious sense. This comes from a real life issue that I am aware of, but I’m a third party. Not any of A, B, C, D, E.
The whole point of the exercise is to get people to draw some lines and justify where to draw them and the whole situation involves various shades of gray, so any argument needs to be nuanced.
A should write something in the intro out of courtesy to both the other authors and to the reader, saying (after this paper was submitted, several related papers B-E appeared).
Thats pretty standard.
I can’t simply write a paper dealing with progress on solving the Poincare conjecture, wait for Perelman to solve it, and then rewrite and resubmit the paper without citing him. That would be absurd.
I’d err on that side of the argument, even if B-E’s progress is relatively trivial.
Citation demands are a known irratant in publishing, and well thats part of the price you have to pay for a functioning system.
The way I see it, there’s no way A can honourably avoid citing B – E. Either A’s first version is a good partial analysis unto itself, and should then be published as such (A can then write a full analysis in which he cites all partial analyses A – E), or the first version doesn’t stand on its own, in which case the replacement should contain the standard acknowledgement Haelfix suggested. That way the citations fulfil their primary purpose as a service to the reader, and due credit is given to everyone without anyone’s contribution looking less than it is. So no, A’s rationale doesn’t hold water.
The question then is what E should do in the present situation. That answer is not so clear-cut, and in my opinion depends mostly on how important this particular citation is to him. If he feels that this is a very serious omission, he may contact the journal and discuss the issue with the editor in a friendly manner. There’s no guarantee that this would work, and there’s certainly a risk of making enemies and of making an impression of being excessively whiny. Only E can weigh the pros and cons of that course of action.
I have myself seen a few publications that could — maybe even should — very well have cited some of my work, but didn’t. I have not complained about it, and will not either. It has never seemed anywhere near worth the trouble.
Here is a similar dilemma which is purely about credit: A and B send out a paper with a beautiful new idea, but they are not successful in implementing it with all the details coming out correctly, and the paper is riddled with mistakes. C and D send out, shortly afterward, an implementation which not only nail all the numerical factors correctly, but makes important contributions necessary for any other implementation.
Now, the Noble price is only given for 3 authors, which ones?
I think that most people who don’t have something to gain from the situation would reach the same consensus that has been reached here, (credit where credit is due), but as has been pointed out, the award system is somewhat broken.
Physicists don’t seem to mind tacking on an ever-increasing list of somewhat temporally disconnected names to an important idea, (WKBJ, DBI…) but this sort of acknowledgment alone doesn’t put “food on your family”. Why don’t we see more granularity in the Nobel prizes, for example?
Lionel:
The issue with Nobel prizes is that if you give out more, then the prize becomes less important. That is fine if you just want to recognize good work, but it is lousy if you want a Nobel prize winner to be able to have influence with politicians etc, and be able to better advance the cause of science.
Having ambassadors to science that are widely recognized as being important and that should be paid attention to is much more important than trying to be fair (the benefits to the cause of science are much greater than the perceived damage).