When submitting an extended version, is it mandatory to include all authors of the original paper?

Question

In my field, it's a common practice to publish a work at a conference first, then extend it (typically with 25%-40% more content) and submit it for publication in a journal.

Often the extension is driven by a subset of the original authors, for example, because the remaining authors lost interest or don't have time to develop the extension. In this situation, the question arises whether it's mandatory to include the non-contributing authors as authors of the extension.

Arguably, the authors have already received the credit for their contribution with authorship of the original paper. On the other hand, their contribution to the original work is still a contribution of the extended work.

Is there any official guideline of how such cases should be handled?

Answer 1

An extended version of a conference paper supersedes the conference paper. A priori, this means that every author of the conference paper is also an author of the journal paper. In same rare cases it could be appropriate to drop an author (if every part they contributed to gets removed), but as a matter of protocol, I would only consider this on request of the relevant author.

This is different from a subsequent paper building on a prior paper from the same project, where a citation to the previous paper suffices. In such a case, one would not copy over important text anyway.

Answer 2

It would be outrageous if I took one of your papers, added 25–40% more stuff to it and published it as my own, claiming that you'd already received credit for your parts and, anyway, you didn't write any of the new stuff. What you're proposing is basically the same thing.

Answer 3

The correct answer to this varies by field and possibly location. Ethically, so long as you cite the original paper, the people who work on the extended paper are, properly, its authors.

It also depends, of course, on your definition of "extended version". If the new paper includes the old then it is quite different than if it merely cites the old. In the first case the original authors should all be included.

But politically and for keeping open future collaboration possibilities it will cost little to include everyone as author of the extended version. My own preference would be to include people, with their permission. In some places it might even be necessary to mollify a powerful colleague.

But if you haven't already built the new version, you could ask everyone whether they want to participate and become authors. That would already imply, of course, that your preference would be to exclude non participants.