# Intrersection of open set is closed

In the first chapter of Probability wih Martingales (Willams) I came across the following relation

$$(-\infty,x]=\bigcap_{n\in N}(-\infty,\, x+n^{-1}).$$ Even though this is intuitive to me I would like to know if this is a rigorous statement. Is the intersection of a infinite number of sets defined in the same way as for a countable number of sets ? or by some limiting operation like liminf or limsup?

Thank you.

-
To answer the title, the intersection of an infinite number of open sets is not necessarily open. – Jean Hominal Aug 20 '13 at 8:15

The intersection of any number of sets is defined in the following way:

$$\bigcap_{i\in I}A_i=\{a\mid\forall i\in I.a\in A_i\}.$$

In order to see that this equation is true, note that for every $y>x$ there is some $n$ such that $y-x>\frac1n$.

-
thank you. I understand it. – triomphe Aug 20 '13 at 9:35

The intersection is defined in the usual way: a real number $\alpha$ is in the intersection if and only if it belongs to each of the rays $(\leftarrow,x+n^{-1})$. That is the case precisely when $\alpha\le x$, so the intersection is indeed $(\leftarrow,x]$.

-
thank you. nice explanation. – triomphe Aug 20 '13 at 9:36
@MLT: You’re welcome. – Brian M. Scott Aug 20 '13 at 16:57
Did you invent this arrow notation? I quite like it. – Clive Newstead Aug 20 '13 at 21:53
@Clive: No, I picked it up somewhere; it was years ago, though, and I no longer have any idea where. – Brian M. Scott Aug 21 '13 at 1:45