Suppose we have a measure $\mu$ on $\left( \mathbb{R},\mathcal{B}(\mathbb{R}) \right)$ with $\mu((0,1])=1$, and $\mu$ invariant under translations i.e. $\mu(A) = \mu(A+c)$ for every $c \in \mathbb{R}$ and Borel set $A \subseteq \mathbb{R}$.

Now letting $\lambda$ be the Lebesgue measure, in the first part of the question I was able to show that for any interval $A=(a,b]$ with $b-a \in \mathbb{Q}$ that $\mu(A) = \lambda(A)$.

The second part of the question asks to extend this to all Borel sets, i.e. $\forall A \in \mathcal{B}(\mathbb{R})$ we get that$\mu(A) = \lambda(A)$, but I am really struggling to think of a practical way of achieving this.

Clearly using the first part of the question is the correct approach, so I was attempting to come up with some ways of using this.

The first thing I thought of was maybe to approximate all of the intervals in $\mathcal{B}(\mathbb{R})$ by these rational intervals in the first part of the question, but I could not think of a rigorous way of doing this and I do not think this is the correct approach.

My next thought was to potentially show that the intervals from the first part form a $\pi$-system of some sort and maybe generate the Borel sets and so agreeing only on these intervals is sufficient to show they agree on the whole space, but I am not too confident in arguing this so any help would be appreciated thanks.


Having $\mu (a,b]=b-a$ for rational $a,b$ gives $\mu (a,b)=b-a$ for all real $a<b.$ Proof: Write $(a,b)$ as the increasing union of $(a_n,b_n]$ for appropritate rational $a_n,b_n.$ Standard measure theory with your result for rationals then gives the result.

Since every open set in $\mathbb R $ is the disjoint union of open intervals, we see $\mu(U) = \lambda (U)$ for all open $U\subset \mathbb R.$

I'll stop here for now. Ask questions if you like.

  • $\begingroup$ Hi thanks for the comment, I was thinking about this approach at first but had some concerns about the fact that it's not neccesary that $a_n$ and $b_n$ are rationals, since only their difference need be rational - is there still a way to make this method work? $\endgroup$ – UsernameInvalid Nov 21 '18 at 22:09
  • 1
    $\begingroup$ But if both $a_n,b_n$ are rational, you have it covered. $\endgroup$ – zhw. Nov 21 '18 at 22:11
  • $\begingroup$ Ah yes of course, I suppose I was looking into it too much, thanks for the help :) $\endgroup$ – UsernameInvalid Nov 21 '18 at 22:14

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.