We recently submitted a paper to an IEEE conference, where it was presented and appears in the proceedings. We now wish to submit an extended version of the same work onto ArXiv.

In addition to acknowledging that a short version has appeared in the IEEE conference, are there any restrictions on the kind of license that we can tie to the extended version?

  • I think that the only restriction would be that once you have issued a permissive license on a work, you can't issue a more restrictive license on the same work. The old license is still valid.
    – Buffy
    Jul 23, 2019 at 14:26
  • @Buffy I see. Just to clarify, you are saying that the IEEE license from the short version of paper directly applies to the extended version, so anything that complies with that should be fine?
    – Television
    Jul 25, 2019 at 16:52

1 Answer 1


Let me expand my comment, since I think it was misinterpreted. Suppose you produce some work and license it with a permissive license. In the future, you can relicense that work with an even more permissive one, but not a more restrictive one. People can rely on the earlier permissive license for that work, but only for that work.

However, if you produce a new work, you can license that however you wish, or not at all. You can even put "Copyright - all rights reserved" on the new work. But that doesn't change the license to the old one. Anyone relying on only the older work can rely on its original license. But if someone wants to use the new work somehow, they have to abide by its license (or lack thereof). Of course the two works would need to be sufficiently different for this to have any real meaning.

For example you could publish a textbook under a Creative Commons license (CC-BY-SA, say) and then a second edition with all rights reserved. The first edition still has its permissive license. Only the material in the second has the restrictive license.

Note that I am not a lawyer. But the basic idea is that you can't "lock up" IP once you make it available, but you can make it more available if you like.

And, in general, a work that isn't specifically licensed is considered in many (most?) places to be copyrighted. But laws do differ.

So, your best option is to be explicit about the new work's license.

An example, sparked by a now deleted comment:

UC Berekely created a version of UNIX, the BSD (Berkeley Software Distribution) with its own permissive license. Apple built OS X (macOS) on top of BSD and used many of its elements. But macOS does not have a permissive license, but a restrictive one (or none - all rights reserved - proprietary software). But Apple didn't "lock up" BSD and anyone can still depend on that original license for their own work. But the copyright holders of BSD cannot, now, put more restrictions on their own license that would, now, invalidate OS X. They could, however, make it even more permissive.

  • I understand. I am no expert, but by having a paper accepted to an IEEE conference, I think I am "handing over" the rights on the work to IEEE. In that sense it would probably be wrong to redistribute the work with no license. In the context of the BSD / Apple example of yours, my question is: if BSD was very restrictively licensed, is Apple allowed to redistribute macOS (built on BSD) with a more permissive license?
    – Television
    Aug 2, 2019 at 10:30

You must log in to answer this question.

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