We should allow as much work as any well-meaning non-OP editor is willing to give away for that post. Within very few and wide limits.
A significant well-meant edit is a chore the editor offers, why should we prohibit or inhibit this generosity?
Questions can be edited, and for a reason. Sometimes OPs change the content and aim and scope quite significantly. Hopefully for improving the quality also significantly.
Questions can be edited, by other-than-OP users, and for a reason. And if there is something significantly 'wrong' or a good chance to improve the post, then it should be edited.
The only problem I see here is that the guidelines simply say that edits should not deviate too much from the original intent of the OP. 'How much deviation' is a question for which the answer is not easily inferred from guidelines: That's vague in the first place and on the other hand: how should anyone of us really know the 'true' intent? Sometimes an OP is just a bit hapless and lost for words, ending up quite grateful for a good edit, even if it seems to deviate a lot from the original.
I wish we had some simple heuristics to go by here.
But I disagree, to a degree, on this:
The problem I have here is there's no indication the original poster had any say in this.
Not "any say" seems incompatible with the OP having the 'power to accept' the edits and the 'power to roll back' if they don't? That is more obviously 'a problem' if these users then disappear.
These two powers are fostered if in case of newbies explanatory comments highlight the fact that they are encouraged to either accept the edit or have it their way, since such changes can seem unfriendly or intimidating to newbies.
That looks like an easy fix. If the edits are indeed well intentioned. If OPs are still around that is. Such significant and deviating edits appear inherently more problematic on old questions…
An edit is perhaps well meant but not well done if it completely changes and the question and OP disagrees on a fresh question, or if community members disagree on an old question, or the changes invalidate the answers on an old question.
Thus, if it is obvious that the OP has had the chance to improve (active on site: has read the criticism) and especially if an OP was obstinate to fail to improve (trolling, arguing in comments, rollbacks…) and a 'good core' was present, then we should encourage others to re-use the seen-as-valuable' good core or topic and then post 'a good question' based on it.
The above assumes a default of well-meaning question and poster.
The situation changes if the post is really bad in the sense of not a good-faith effort. In such a case a 'good edit' enables the wrong. Such a 'good edit' will give reputation points to a contributor for which the system was initially designed to prevent those kind of users from getting any rep and privileges.
This kind of 'rep for bad effect should be prevented' is of course quite difficult to put into a decision heuristic on what to do with a post, but it should be kept in mind when evaluating.