A simple (albeit indirect) way to obtain a sigma algebra, is to start with some sets, and then to add all the sets needed to make the starting set closed under the various operations.
The generated sigma algebra for a set of subsets is the smallest sigma algbera containing the set of subsets. We must prove the existence and uniqueness of this sigma algebra.
On the other hand, the union of a set of sigma algebras can fail to be a sigma algebra.
We know of one sigma algebra containing $\mathcal{A} $: the power set of $A$. Thus, the set of sigma algebras containing $\mathcal{A} $ is not empty. Proposition~\ref{intersectsigmaalgebras} implies the intersection of all such sigma algebras (containing $\mathcal{A} $) is a sigma algebra. The intersection contains $\mathcal{A} $, and is contained in all other sigma algebras with this property, so is a smallest sigma algebra containing $\mathcal{A} $. If $\mathcal{B} , \mathcal{C} $ were two smallest sigma algebras, then $\mathcal{B} \subset \mathcal{C} $ and $\mathcal{C} \subset \mathcal{B} $, but then $\mathcal{B} = \mathcal{C} $; thus the smallest sigma algebra is unique.
Let $A$ be a set and $\mathcal{A} \subset \powerset{A}$. We denote the sigma algebra generated by $\mathcal{A} $ by $\sigma (\mathcal{A} )$.