I know of at least two inference rules for disjunction elimination:
$a \vee b, \neg a \vdash b$
$a \vee b, a \to c, b \to c \vdash c$ (also known as proof by cases)
I know the second follows from the first, because I have proved it. But does the first follow from the second? I have not been able to prove this.
Indeed, assuming $\neg a$ is equivalent to $a \to \bot$ and assuming you're allowed access to $\neg \neg b \to b$. (Just let $c = \bot$.)
Note that you do need some form of the law of excluded middle for this implication, because the second is constructively valid but the first is not.