Warning: this is more of a picture of the idea, rather than mathematically rigorous explanation.
The two limits are basically defining a candidate "limit set" by giving a bound from outside or inside.
The lim sup is considering big union of sets, namely $\bigcup_{k \ge n} A_k$, and by taking intersections over $n$, we are throwing away stuff that appears only in finitely many $A_k$. If something appears in infinitely many $A_k$, it is a plausible candidate to be an element of a "limit set".
The lim inf works the other way around: we first select elements that, for $k\geq n$, appear in every $A_k$, that is $\bigcap_{k \ge n} A_k$. Those are definitely good candidates for a "limit set". Then we consider the union of all these intersections.
These two limits are good bounds for a candidate "limit set". On one hand, if $x$ is not in the limsup, then it means that $x$ is contained in only finitely many $A_n$, so it cannot belong to a reasonable "limit set"; the limsup is an "outer bound". On the other hand, anything that is contained in every set (except possibly finitely many) should probably be also in the "limit set"; the liminf is therefore an "inner bound".
To give an example, consider this sequence of sets:
$$
A_n = \begin{cases}
[0,1/2)\quad \mbox{if n is odd}\\
[1/2,1]\quad \mbox{if n is even}\\
\end{cases}
$$
Since $A_n\cap A_{n+1}=\emptyset$, the liminf is the empty set. Makes sense: there's no element that is contained in every set (for $n$ large enough). The limsup on the other hand is $[0,1]$, since every element in it keeps appearing in the sequence.
http://math.stackexchange.com/questions/107931/lim-sup-and-lim-inf-of-sequence-of-sets/107938#107938
– Michael Hardy Jul 02 '15 at 21:31