It's my understanding that research done at government labs comes with a similar restriction, and that it's been standard practice for years for these researchers to reserve copyright on their publications. If that's true, then this will give Princeton researchers the leverage they need to retain copyright to their publications. Even though there's theoretically a waiver process, I'm going to bet that most authors won't want to go through that process, since the work is more valuable to them if they retain the ability to distribute it themselves.