The indicator function $\chi$ of a set $A$ is that function for which

$$\chi_A(x) = \begin{cases} 1 & \text{if $x\in A$} \\ 0 & \text{otherwise} \end{cases}$$

(Some people prefer $\mathbf{1}_A$ as a symbol instead.)

When dealing with a universal set $\Bbb U$ (usually $\Bbb R$ I presume) the measure of the set $A\subset\Bbb U$ is, as I understand it,

$$m(A) = \int_\Bbb{U} \chi_A(x) \, dx$$

(where $x$ is simply a dummy variable).

Of course, sets of two dimensions have measures. For example, $m([4,6]\times[0,10])=20$, which I could tell you having taken simple geometry. However, in general terms, how is the measure of a two-dimensional set defined in terms of an indicator function and integration?

Naturally, the element of each set will be an ordered pair, and I anticipate that the measure of the two-dimensional set is not necessarily the product of the measures of each dimension.

  • $\begingroup$ Based on what you've written, I'm not sure what you think is missing. Why not just take your general statement in the case $\mathbb{U}=\mathbb{R}^2$? $\endgroup$ – Eric Wofsey Oct 30 '17 at 2:25
  • $\begingroup$ @EricWofsey I'm not sure how that would look. Wouldn't I need a double integral then? $\endgroup$ – gen-z ready to perish Oct 30 '17 at 2:26
  • 1
    $\begingroup$ Well that depends on your definitions of integrals. Usually, people define measures first and then define integrals in terms of them, not the other way around. $\endgroup$ – Eric Wofsey Oct 30 '17 at 2:27
  • $\begingroup$ @EricWofsey I'm trying to sort through the eclectic assortment of lecture notes over measure theory I can find while I look for a decent book, but I'm pretty lost ¯\_(ツ)_/¯ $\endgroup$ – gen-z ready to perish Oct 30 '17 at 2:29

The measure of a two-dimensional set can be defined exactly as you have written it -- if the characteristic, or indicator, function $\chi_A(x)$ is defined to be $1$ over the set $A$ and $0$ elsewhere, the measure of a set $A$ is indeed equal to $$m(A) = \int_\mathbb{U} \chi_A dm$$

This is the notation I learned in Rudin -- instead of integrating with respect to a bound variable x ($dx$) you integrate with respect to the measure $m$. Does this seem circular?

The thing is -- if I am remembering correctly -- we can resolve this with exactly the sort of simple set you bring up. The measure of a rectangle in $\mathbb{R}^2$, or a box in $\mathbb{R}^n$ (regardless of we include or exclude the vertices/edges/faces/etc.) is defined to be the product in $\mathbb{R}$ of the lengths of the sides. This is useful because every set that can be written as the union of a countable collection of disjoint boxes can be measured by breaking it into boxes, measuring those with the defined measure, and summing.

By the way, if you're looking for a book to introduce measure theory as it is used in abstract integration, Real and Complex Analysis by Walter Rudin is pretty standard.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.