How can Springer and Wiley put a 12 month embargo on posting post-review revisions to the arXiv and similar?

The copyright agreement does not require that you leave your mistakes uncorrected, it simply requires you to make corrections in a particular way. The simplest method, if you discover mistakes in the published version, is to distribute some note of correction that specifically addresses your original error, which can be done without violating the copyright agreement. The copyright agreement may allow you to post post-review versions of the paper, since the agreement may contain a clause that "Prior versions of the Contribution published on non-commercial pre-print servers like ArXiv/CoRR and HAL can remain on these servers and/or can be updated with Author’s accepted version". There is always a clause where the author warranties that the work has not been previously published (which includes online distribution via ArXiv), and this is why they need to expressly say that you are also permitted to post the submitted version.

There are two notions of "enforceability" relevant to your agreement with the publisher. The stronger one involves litigation, where the publisher sues you for damages (when you breach a contract). You should hire a Swiss lawyer to get advice on whether Springer is likely to prevail in court. The other notion is "having negative consequences". A simple negative consequence would be that Springer refuses to publish any more of your work, if you flagrantly violate the terms of the agreement. Whether or not your breach involves copyright violation depends on whether or not you give them a license to publish, versus transfer copyright. (In interpreting the Lecture Notes in CS agreement, you would want to consult the Swiss attorney for a precise interpretation of granting and assigning the exclusive, sole right to copy).

Your comment about the review process is tangential, and suggests that you are unaware of review protocol, You ask "Why one should review for free for a journal that won't let the author share the results of your review with the community in as timely a fashion as possible". Reviews are not only anonymous, but also privileged communication between the reviewer and the editor (often -- though not always -- shared verbatim with the author). So you should not distribute reviewer comments, unless that is expressly permitted by the journal.

Springer (copyright transfer form) actually says (my emphasis):

"Authors may self-archive the author’s accepted manuscript of their articles on their own websites. Authors may also deposit this version of the article in any repository, provided it is only made publicly available 12 months after official publication or later. He/ she may not use the publisher's version (the final article), which is posted on SpringerLink and other Springer websites, for the purpose of self-archiving or deposit. Furthermore, the author may only post his/her version provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be provided by inserting the DOI number of the article in the following sentence: “The final publication is available at Springer via http://dx.doi.org/[insert DOI]”."

The only restrictions are on the accepted version and the published version. Pre-prints and revisions prior to acceptance are not precluded. However, links to the Springer website/doi should be provided on publication. See: http://www.springer.com/gp/open-access/authors-rights/self-archiving-policy/2124

Note that it is untrue to say that: "you are obliged to leave uncorrected any mistakes that are uncovered during the review process for a protracted period of time". The corrected version appears in the journal. Just update the arxiv paper with a link to the journal paper, stating that that is the corrected version. And you can update the preprint server with corrections anyway: you just can't breach your publisher's copyright agreement. If necessary, find a second way to phrase the correction, or if that's too onerous, add a note saying that corrections are forthcoming in {this journal}. But all this seems by-the-by, as no one's obliging you to use this publisher: I'm sure any number of open-access publishers will be happy to take your cash in exchange for very few restrictions.

You ask whether this agreement is legally enforceable. It doesn't matter. When someone breaches the agreement they have made with a publisher, they've shown that they are untrustworthy. The route you've identified, of someone deliberately breaching an embargo that they know they are going to sign up to, would be an act of bad faith that would demonstrate that they are untrustworthy.

Academia is built on trust relationships. So if anyone was considering reneging on an agreement, they should first give careful consideration to all their other current agreements and all their future possible agreements, with media, academia, voluntary sector, industry, and government.

Would a publisher pull a paper if the embargo had been breached? Well, they're completely within their moral and legal rights to do so. And consider the business case: if the one thing that they were going to get out of the deal was exclusive distribution of the paper for a year, and someone took that one thing away, there's nothing left for them in the deal.

Note that there are lots of publishing models out there. If you don't like one, and your funder doesn't oblige you to use it, then don't use it. There's a market there: it's pointless to to try to coerce everyone to follow one publishing model, regardless of how fashionable or ideologically pure that publishing model is.