Let me give another example type which never gets in the way of the users,
What happened to Textures and Bluesky Research?
Now this is clearly a very important question due to its answer not as a question itself. If you look at the typically-complained questions, they are either grandma category or big list of LaTeX editors question type for which everybody has an opinion.
This is simply the same problem with removing Thanks in disguise. Whenever there is room for speculation, it rubs the disciplined users the wrong way. But they are not alone in this environment. Moreover, as we know from the every-now-and-then happening edit frenzies some users only care about the structure but not the content itself.
This is by far the most importantly a social medium. Clear set of rules etc. don't apply but only guide. It is especially important on TeX.SE which is the main fabric of our
<insert all the nice things here, friendliness, kindness>.
We can deal with uncertainty and hance we deal with them case-by-case basis. There is no need to impose structure on an already nicely-working system led by its entire user base. Example; it takes a few comments to make a (proper!) case and make a question reopened. Anyone is free to try the same on other SE networks which have clear set of rules for reopening that render reopening impossible.
I would like to remind everyone who participates in the discussion and cites FAQ and help pages that originates from SO-templates, that we have fiddled with most of the rules and expectations. So every SE site has its own habitat. Arguing via SO-based dictations might be misleading.
If it's fun we don't spoil the fun. Sometimes we can't just close it because it's fun. Sometimes we don't let it become fun. I don't know how that mechanism works. But I don't need to anyway.
Long story short there is no fixed rule set hence there is no way for a rigorous explanation of a complicated social interaction. It just doesn't work that way