After reading this answer, and the subsequent comments, I found myself wondering if it is a rule for publishers to get author approval after edits before publishing a paper? I would assume so because of copyright issues but I would like to know if there are legal or ethical guidelines in the publishing industry about this.
In my experience, editors will make edits but they have always asked me for my approval before actually publishing the edited piece.
In the answer you are referring to there is nothing to say that approval is not required. The issue is rather that an editor can request certain things can be removed from the paper so the author cannot demand everything written is finally printed. Editors have the right and obligation to weed out materials that are, for example, offensive just as we can flag posts that contain such phrasings on Stackexchange. Jokes, which was the point of the post you referred to, are not edited out by any means as a rule but can be misunderstood or considered offensive and might be more likely to be edited out than other material during the review and editorial process. In such a case, the author will certainly be part of the process and asked to revise or accept changes. the point here is that the editor has the final say through approval of the paper as well as a responsibility to make sure the publication is free from material that violates, for example, ethical of the publisher. As an example, in Sweden, where I reside, I as an editor am responsible by law for what is printed in "my" journal. This responsibility may lie with the publisher in other countries.
So, authors will definitely be part of the revision process and sign off on final proofs, it is just that some changes may not be subject to additional discussion and will have to be accepted. Does this happen often? Not at all. In my case, I have had to deal with this in heated "Letters to the Editor" where authors have been in a written, published, exchange over issues in a published paper.
Edit (to expand based on earthling's comment): Basically all changes except typos and changes to adhere to journal or publisher style are subject to an "OK" from the author (although it would not hurt to also get an OK on these). The publishing of a paper is a joint venture between an author and the journal, so with a functioning working relationship both parties should agree on changes. But, again, some changes are non-negotiable, which does not mean they can be imposed without the knowledge of the author who always has the option to go elsewhere to publish.
So changes can come in three forms, (1) those that can be changed without consent (e.g. typos), (2) those that must be made without discussion (subject to ethical considerations) and (3) those that can go either way after, say, clarification (most of the normal changes). In cases 2 and 3 the author need to approve the changes it is just that for case 2 the option might be accept or withdraw the paper.
When I took over being the Graphics Editor / Typesetter of one journal, I realized that there used to be no proofreading stage for the articles! I of course insisted that the articels are proofread.
However, we give the authors only a limited amount of time to claim any corrections, but not less than 2 weeks. This is because the journal is quite small and we need to "schedule" the articles into the issues to make the issues reasonably equal in size.
Even after the approval, small changes appear; the Chief Editor goes quickly through the text just before publishing to catch last small typos and mistakes: missing full stops after figures' captions, corrupted text-flow on the page (widows/orphans standing out badly etc.), wrong capilatization, and stuff like this. However, we never modify the text itself in this phase, not even by adding or removing an article (since this can change the meaning of the text).
In general, by law, nobody is allowed to publish anything signed by your name without your approval. The reality is that this rule is violated quite often and it's difficult to do something about it. As well, remember that by submitting an article, you agree with the policy of the journal that can state for instance: "By submitting an article, you give your permission for it to be published, and you confirm that all authors have agreed to this. The Editor's Office is allowed to make modification to the text without the authors' agreement."
In my experience, even corrected typos, grammar mistakes and changes necessary for correct typesetting are usually approved (or rejected) by the authors. Thats what the print proof is for. In my opinion that is absolutely necessary, because even changing a few letters could change the meaning of a sentence, and as an author I have the full responsibility for the whole paper.
I've once made the weird experience that I found a paper listed in pubmed before we had received acceptance notification or print proofs. While nothing bad had happened to the paper's text, we decided never to go with that publisher again.
