11
$\begingroup$

I slacked off and went to sleep last night with this Suggested Edit review pending on my screen.

When I returned this morning, I voted to reject the proposed replacement of tag "proof-writing" by "fake-proofs". There was already one vote to reject when I cast mine, as "conflicts with the author's intent", so I was surprised to see the red box pop-up informing me that the edit had already been approved! That's the price paid for falling asleep!

Anyway, not to rollback without discussion, the current text of the tag info for "fake-proofs" leans into the notion of "Seemingly flawless arguments are often presented to prove obvious fallacies (such as 0=1)."

In this Question the issue is rather one of an incorrect proof of a true result (not of an "obvious fallacy"), namely that squares of real numbers are nonnegative.

As with many discussions Martin Sleziak was there first with this meta Question, What is the best tag for "spot the mistake" questions?

The other tag Martin mentioned in the older discussion was "solution-verification" or "proof-verification", noting the drawback that in these situations is aware that a mistake exists and is not asking for confirmation that a proof/solution is correct. Jyrki made a Comment there that perhaps no special tag is needed for such Questions beyond the subject matter itself. The original tags for the reviewed post were "proof-writing" and "logic", which seem appropriate even if one might accept tagging with "fake-proofs".

My question is whether the "fake-proofs" tag should be strictly limited to cases of obvious(?) fallacies. If not, then presumably the tag info should be updated to acknowledge the expanded scope of wrong proofs of true statements.

UPDATE: I replaced the "fake-proofs" tag on the linked Question with the "proof-verification" tag, and I made slight edits to both tags to reflect the slight sharpening of scope per the voting on this post.

$\endgroup$
  • 1
    $\begingroup$ I think I was a little bit sloppy when I submitted the edit suggestion: I haven't read carefully the tag excerpt that (proof-verification) can be used when asking to identify errors. So (proof-verification) should be more suitable here. Thanks for pointing out. $\endgroup$ – YuiTo Cheng Feb 9 at 17:06
  • 2
    $\begingroup$ For what it's worth, regardless of whether that question should be tagged (fake-proofs) or (proof-verification), the tag (proof-writing) is definitely inappropriate and should not be restored. (proof-writing) is for questions about the writing of proofs (writing style, how to clearly convey your logic, etc), not about the correctness of a particular line of reasoning. $\endgroup$ – Eric Wofsey Feb 9 at 19:41
  • $\begingroup$ @EricWofsey: The main thrust of the exercise was surely to find the mistake in the proposed proof, but also the title asks about "proof by contradiction". The tag "proof-writing" strikes me as weak tea for most searches, but if someone wanted to ask about writing a direct proof versus a proof by contradiction, I'd say it could be so tagged on the basis of being about the formulation of a proof. $\endgroup$ – hardmath Feb 9 at 20:05
  • 3
    $\begingroup$ I would say "fake-proofs" should be limited to proofs of contradictions, and proofs of conjectures that are "too easy", and the OP wants the answer-er to find the error. (By "too easy", I mean something like a 4 line proof the of the Collatz Conjecture. By using the tag, the OP is signalling they do not believe they solved the conjecture, but are asking where the proof fails.) $\endgroup$ – PyRulez Feb 9 at 21:07

You must log in to answer this question.

Browse other questions tagged .