Let $A=\{A_1,\ldots,A_s\}$ be a family of subsets of $\{1,\ldots,n\}$. Then, $A$ is said to be an intersecting family if any two elements in $A$ have a nontrivial intersection. If we also place the restriction that all subsets in $A$ have the same cardinality, say $A$ is an intersecting family of $r$-subsets of $\{1,\ldots,n\}$ , then how large can $A$ be?
One way to generate an intersecting family of $r$-subsets of $\{1,\ldots,n\}$ is to pick an element, say $1$, to include in each of the $r$-subsets. The remaining $r-1$ elements can be chosen in ${n-1 \choose r-1}$ ways. Thus, there exists an intersecting family of $r$-subsets of $\{1,\ldots,n\}$ of cardinality ${n-1 \choose r-1}$, namely the set of all $r$-subsets which contain the element $1$. The Erdos-Ko-Rado theorem asserts that this value is the maximal size of an intersecting family of $r$-subsets and that every intersecting family of $r$-subsets of maximal size is exactly of the form just mentioned: the set of all $r$-subsets which contain one particular element.
The symbol $\partial$ is probably the notation for something like the shadow (not the partial derivative). You can look at your reference (or a text on set-systems) for the exact definition.