This "legal loophole" is trivial for journals to close.
In fact, I wouldn't be surprised if most journals hadn't, in effect, already pre-emptively closed it.
Basically, any publishing agreement vetted by a halfway competent lawyer will already have a clause requiring the author to assert that they in fact do have the authority to grant the journal the rights that the agreement says they do. For exclusive copyright transfers, this is likely to include an explicit or at least implicit assertion that the author has not granted any prior conflicting licenses to the article.
For example, here's a quote from a sample publishing agreement I found on the first page of Google results:
Author Representations/Ethics and Disclosure
I affirm [that...]
- Except as expressly set out in this Journal Publishing Agreement, the Article is not subject to any prior rights or licenses and, if my or any of my co-authors’ institution has a policy that might restrict my ability to grant exclusive rights under this Journal Publishing Agreement, a written waiver of that policy has been obtained.
If it turns out that the author has, in fact, previously granted a non-exclusive license to the article to a third party, in a way that is not covered by the explicit permissions described in the publishing agreement, then the author will have violated the publishing agreement, and may be considered to have negotiated in bad faith (since they presumably knew about the previous license when they signed the publishing agreement). What happens next may vary, but at least the journal has perfectly valid grounds to retract the paper if the publishing agreement is breached.
Of course, all this is negotiable.
If you do, in fact, want to both publish your paper in a traditional journal and also distribute it through alternative channels (like, say, arXiv, an institutional repository, your personal website, etc.), you can just tell the journal that you want to include an exception for that in the publishing agreement. You may even, if you like, grant such an alternative distribution license already before submitting your paper to the journal, in which case it's basically a "take it or leave it" deal for the journal — they can't force you to cancel the earlier license, but they can just decide that they don't want your paper.
In fact, many journals do already expressly permit, or at least implicitly tolerate, various forms of third-party article distribution. For example, the same publishing agreement I quoted above also explicitly permits "Scholarly Sharing", which is defined as:
Preprint: Sharing of Preprints by an author on any website or repository at any time. When the Article is accepted, the author is encouraged to include a link to the formal publication through the relevant DOI. The author can also update the preprint on arXiv or RePEc with the Accepted Manuscript.
Accepted Manuscript:
(i) immediately on acceptance: sharing of the Accepted Manuscript by an author:
- via the author’s non-commercial personal homepage or blog
- via the author’s research institute or institutional repository for Internal Institutional Use or as part of an invitation-only research collaboration work-group
- directly by providing copies to the author’s students or to research collaborators for their personal use
- for private scholarly sharing as part of an invitation-only work group on commercial sites with which the publisher has a hosting agreement
(ii) after the embargo period: an author may share the Accepted Manuscript via non-commercial hosting platforms (such as the author’s institutional repository) and via commercial sites with which the publisher has a hosting agreement.
While this is not a carte blanche to let you license your paper to just anybody under any terms, it does allow you to e.g. submit your paper to an institutional preprint repository, like the UC open access policy mandates. It even goes beyond that, and lets you submit the final, published version of your paper to such a repository, too, provided that you wait out the embargo period until doing so.
Of course, journal publishers don't grant such generous permissions just because they feel kind and altruistic — they do so because authors have actively insisted on such permissions, and because they judge that any potential loss of income from such non-commercial paper distribution will cost them less than the loss of manuscripts from boycotting authors, and/or the PR cost of sending lawyers after respected academics who are just doing what they consider to be reasonable and customary sharing of scholarly information. The publishers know that they can only go so far in demanding exclusive distribution rights without risking a major backlash, and they're generally smart enough to adapt and make the best of the situation by explicitly releasing those rights that they couldn't practically enforce anyway, and making it look like benevolence on their part.
So, in short, there is no legal trick to "disable" distribution restrictions by publishers — at least not one that would work more than once per publisher. What does have a chance of working is politics, media and activism: publicizing the issue, exerting pressure on publishers, and voting with your feet. A CC license grant in your back pocket, however, isn't going to be a silver bullet.
Addendum: So what is the point of an institutional Open Access policy, like the University of California has recently enacted, then, if it's not a legal trick to get around publishers' agreements?
The way I see it, this policy accomplishes several things:
Making free preprint distribution the default rather than an exception. For a long time, many publishers have already permitted such distribution, but since it involves extra work for authors (both figuring out whether such distribution is permitted, and then actually uploading the papers), a lot of them may not have bothered. By making preprint distribution the institutional default, and requiring authors to get an explicit waiver to opt out, submitting papers to the repository becomes the "path of least resistance" for authors.
Centralization: Even when authors do make their preprints available, it's often on their personal web pages, where they may be hard to find and which may or may not persist. The UC policy requires all preprints to be submitted to a single centralized repository, where they can be reliably stored and indexed.
Building up a critical mass: If an individual author e-mails a journal that doesn't normally allow preprint sharing, and asks for an exception to their standard publishing agreement, the publisher may be tempted to just brush them off as not worth the trouble. With an institutional policy, publishers with restrictive publishing agreements now face a choice between a) relaxing their terms to make them compatible with the policy, or b) requiring all authors from UC to obtain waivers. As more and bigger institutions adopt such policies, option (a) starts looking more and more attractive to publishers.
Institutional support: At institutions and departments with no prior tradition of doing so, an author who wanted to upload their papers to a preprint server may have been essentially on their own — they'd need to figure out all the legal and technical details themselves, and if they needed to negotiate the matter with a publisher, they'd basically be one person wrangling with a big corporation. If they asked for help from colleagues or admin staff, they might be met with shrugs and blank stares.
With the new policy, authors now have the whole university behind them, both in negotiations and simply in dealing with the technicalities. They can also have confidence that, if they should happen to accidentally violate an ambiguous publishing agreement by uploading their paper to the institutional repository, the university will support them in negotiating a settlement with the publisher.
Establishing a presumption of preprint sharing. Arguably, by setting up the open access policy and announcing it widely, the university is creating a presumption that publishers should know about the policy, and thus know that they need to ask authors from UC to submit a waiver if they don't want to allow institutional preprint distribution. If the journal accepted a manuscript from a UC author, without clearly communicating the need for a waiver, they might be argued to have implicitly consented to the preprint distribution, especially if the situation was otherwise ambiguous. This is a legal trick of sorts, but it's not really a loophole so much as simply stacking the playing field, and it's only possible because the university is big enough that publishers might be reasonably expected to be aware of their general policies.