6

I'm writing an article which contains some quite long proofs, some of which include claims with proofs of their own.

I dealt with that by adding a claim environment using \newtheorem, but there's a slight inconvenience that arose from that: when I write a proof of a claim, then after \end{proof} a QED symbol is shown much like at the end of the proof of the main theorem, which can be confusing when the claim is in the middle of the proof as it seems like the proof is done while it's actually not, and at the end of the proof, when suddenly there's two QED signs.

I dealt with that by putting a

\renewcommand{\qedsymbol}{$\square$ (claim)}

line before the end of the proof. However, I imagine it could be rather tedious when writing long-ish documents with many claims. So, is there any way to do that for all claims at once?

I would also welcome suggestions as to how to do it “the right way” if I'm doing something terribly wrong.

  • 3
    \newenvironment{claimproof}[1][\proofname]{\proof[#1]\renewcommand{\qedsymbol}{$\square$ (claim)}}{\endproof} and then use {claimproof} environment instead of {proof}. – Manuel Mar 11 '14 at 21:22
  • Most math papers that I've read don't format a claim like a theorem. Normally it is simple text (or possibly a list) within the proof of a theorem or lemma or proposition. Its statement might not even stand on its own. E.g., "I claim that $f$ and $g$ are orthogonal", where these are previously discussed functions. The proof of a claim is usually just ended with a statement that the claim has been proven. A claim usually has a one- or at most two-paragraph proof. Any longer than that and it should probably be a lemma. The same if it needs to be refered to. – Dan Mar 11 '14 at 21:54
  • @Dan: whether you call it a claim or a lemma, it doesn't change much, unless you strive to separate the claims/lemmas from the proof of the theorem, but sometimes it's rather unnatural, when they only refer to some local definitions, limited in scope to the proof itself. I call them claims precisely to differ them from lemmas which are independent statements and can be proven entirely separately. – tomasz Mar 11 '14 at 22:14
  • @Manuel: that seems like a good idea. :) It has the added benefit of distinguishing the claim proofs not only in the rendered document, but also the source. – tomasz Mar 11 '14 at 22:20
  • @tomasz if they are small you can even try something like page 50 to differentiate those “small claims” that lead to the resolution of the theorem. – Manuel Mar 11 '14 at 22:31
5

A basic option would seem to create a new proof-like environment:

\newenvironment{claimproof}[1][\proofname]
  {%
    \proof[#1]%
      \renewcommand*\qedsymbol{‌​$\square$ (claim)}%
  }
  {%
    \endproof%
  }

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.