An interesting, rarely surfaced point, in a [reflection post](https://unlockingresearch-blog.lib.cam.ac.uk/) from the University of Cambridge's working group on open research in the humanities, on the drawbacks of the "green," accepted-manuscript repository route to OA. For arts and humanities (A&H), self-archiving might be
> … problematic because it neglects the import of the editing process in A&H research. Without undergoing this process, ‘accepted manuscripts’ are very vulnerable to errors, especially in the case of the very many scholars who regularly work in languages that are not their first, or in the case of early career scholars who are less familiar with critical processes and how to evidence them, or in the case of colleagues with various kinds of disabilities such as dyslexia.
And:
> In cases where scholars receive an acceptance that is subject to improvement, the final ‘date of acceptance’ is ambiguous for legal purposes. And in cases where the work in question uses copyrighted material, further legal issues emerge about when and how it may be possible to circulate this. In all these senses, then, many A&H colleagues simply dislike the thought of their ‘accepted manuscript’ circulating. In the case of institutional repositories, there seems to be a direct and obvious tension between the goals of open research and quality control.
There's a lot of [exciting momentum](https://www.coar-repositories.org/news-updates/coar-welcomes-significant-funding-for-the-notify-project/) around swapping out the journal system for an interwoven repository infrastructure—with various publish-then-review schemes intended to sidestep the prestige lock-in propping up the current publishing oligopoly. In broad terms, the idea is that repositories could serve as an article's iterative resting place, rather than as a "green" way station. The repository solution, if that's what it is, seems to address most of the [Cambridge group's questions](https://unlockingresearch-blog.lib.cam.ac.uk/)—and even the copyrighted-material concern could be answered with a culture of aggressive fair dealing/fair use invocation.