“The RWA didn’t explicitly say that it would amend copyright law, but it could could have done so implicitly, or by superseding any parts of current law inconsistent with the new law. Under the NIH policy, authors give permission for OA when they are still the copyright holders. Even when they later transfer some rights to publishers, they retain the right to authorize OA. Hence, OA through NIH is authorized by the relevant rightsholder, in this case by the author. But RWA Section 2.1 would have required publisher consent for that OA. It would have required publisher consent even when the holder of the relevant rights under current law had already consented. A consent which suffices under current copyright law would not suffice under RWA. Either that would violate US copyright law or amend it pro tanto (that is, amend it to the extent necessary to avoid irreconcilable conflict between the old and new statutes).
This may seem like a technical point of law. But it’s the most radical aspect of RWA. Under current law, in the US and around the world, authors are the copyright holders in their work until or unless they decide to transfer rights to someone else, such as a publisher. Copyright consists of a bundle of rights, and authors may lawfully transfer all, some, or none of those rights, as they see fit. If they retain the right to authorize OA, then no other permission is needed. Under RWA, however, publishers would have held a new right, beyond copyright, to overrule the rights exercised by authors under copyright law.
It was an unprecedented power grab by publishers. Unlike past, lopsided legal reforms to benefit publishers, this one was not limited to enhancing the rights of copyright holders against users and consumers. This one would have harmed all copyright holders except publishers, and benefited publishers even when they were not copyright holders….”