# Why define the Cantor set with an intersection?

Define $E_n$ as

$E_1 = \left[0,\frac{1}{3}\right] \cup \left[\frac{2}{3},1\right]$

$E_2 = \left[0,\frac{1}{9}\right] \cup \left[\frac{2}{9},\frac{3}{9}\right] \cup \left[\frac{6}{9},\frac{7}{9}\right] \cup \left[\frac{8}{9},1\right]$

and so on.

I usually see the Cantor set defined as $C = \bigcap_{n=1}^\infty E_n$. Why use this limit with an intersection, instead of the seemingly more natural $C = \lim_{n\to\infty} E_n$ ?

As far as I can tell, when the limit isn't involved, the intersection is unnecessary:

$E_N = \bigcap_{n=1}^N E_n$

-
And how would you define the limit of these sets? – Asaf Karagila Oct 26 '12 at 10:37
It's just the formal way of defining what you'd intuitively imagine the "limit" of the Cantor sets to be. – user123123 Oct 26 '12 at 10:39
If you want to discuss limits of subsets of the reals, I would suggest looking up "Hausdorff distance". In fact, $C$ is a limit of the sequence $(E_{n})$ in this topology. Here's a link to Wikipedia: en.wikipedia.org/wiki/Hausdorff_distance – T. Eskin Oct 26 '12 at 11:21

Limits are associated with topology; whereas the first time I saw the Cantor set constructed, I was in my first semester. Topology had yet to come out behind the curtain.

For a freshman, limits are for sequences of numbers, or functions of real and complex numbers. Limits are preserved for objects which are not sets in their essence. On the other hands, when you have a family of sets you obviously can discuss their unions and intersections.

So from a pedagogical point of view, this is indeed a reasonable approach to avoid the discussion about continuous operations on sets in the topological space $\mathcal P(\Bbb R)$ (the power set of the real numbers).

Much much later I have learned that a function from ordinals into sets is called continuous if at limit stages we have $\bigcup_{\alpha<\delta} f(\alpha)=f(\bigcup_{\alpha<\delta}\alpha)=f(\lim_{\alpha\to\delta}\alpha)=f(\delta)$.

In this aspect, $E_n$ makes somewhat of a continuous sequence of length $\omega$, whose limit point is the Cantor set. The above definition fails here because we wish to discuss intersection and the continuity defined above discussed unions, alas this is not an important matter as we can always talk about $D_n=\mathbb R\setminus E_n$ instead.

To sum up my ramble above, yes it is possible to discuss limits instead of intersections, but the limit is the intersection (at least in the Cantor set case), but from a teaching point of view it is often the case where the Cantor set is introduced the students have a proper grasp of topology, sufficient to discuss limits of sets. In this case the use of intersections which is much clearer to understand when discussing sets is better.

-