Does it suffice to state the law of excluded middle here?
Assume $\Sigma; \alpha$ is not finitely satisfiable. Then there must be some [finite] $\Sigma_0 \subseteq \Sigma$ such that $\Sigma_0 ; \alpha$ is not satisfiable. Let $f$ be a truth assignment which satisfies every finite subset of $\Sigma$ – since $\Sigma$ is finitely satisfiable such an $f$ must exist. If $f$ does not satisfy $\Sigma_0 ; \alpha$, then it must satisfy $\Sigma_0 ; \neg \alpha$, and similarly for the remaining case.
Is that correct? If not, why?
Best Answer
You want to show that $\Sigma,\alpha$ or $\Sigma,\neg\alpha$ is finitely satisfiable. Hint: Assume what would happen if neither of them was finitely satisfiable (you then use some ideas from your post). The full proof is in the end of this post.
I have some problems understanding your proof. It seems to me that you use that $\Sigma$ is satisfiable, instead of finitely satisfiable (secret use of compactness?). Of course, if you assume that $\Sigma$ satisfiable then adding one of $\alpha,\neg\alpha$ is trivial.