to prepare derivative works based upon the copyrighted work;
reserves the right to prepare "derivative works" to the authors of the original work. The original authors, of course, can authorize such work use of their work to other parties.
17 USC 106, however, is "subject to sections 107 through 122." Section 107 (fair use) allows for an exception to copyright restrictions for the purposes of "research."
In a series of comments to a different question, we have stumbled on a question of whether derivative work, which is performed privately and not shown to anyone, constitutes a violation of copyright. It all seems moot, but I have thought of a scenario where it may become apropos.
purely hypothetical scenario here
Suppose someone dabbles in a study of a foreign language X. Suppose they come across an esoteric industry publication in that language. They invest some personal time to produce a translation of that publication and fix it on paper. This would fall within all the parameters necessary to make this translation a derivative work. However, they do not publish the translation. Nor do they even show it to anyone. They do, however, study the material in it after the translation is completed. Having educated themselves on the subject matter covered in the translation, they come up with an invention which becomes commercially viable. And they develop the invention with a moderate degree of commercial success. A number of years after, someone else uncovers the unpublished manuscript of the translation and reveals the existence of the translation to the original authors.
Do the authors have a copyright violation claim against the translator? It seems like it should fall under the research exception, but the question is whether the research exception has been tested in court for unpublished derivative works.
Rather than speculating what a court might say on the subject, I am curious whether the courts have already said anything on this subject.