Let $X$ be a set, let $\mathscr{F} \subset \mathscr{P}(X)$, and let $\Omega$ be the collection of all $\sigma$-algebras on $X$ which contain $\mathscr{F}$. Then $$\sigma(\mathscr{F}) = \bigcap_{\mathfrak{M} \in \Omega} \mathfrak{M}$$ is the smallest $\sigma$-algebra on $X$ which contains $\mathscr{F}$.
Question: When proving the statement above, Rudin points out that $\Omega$ is nonempty. Why is this relevant?