2
$\begingroup$

Let $f:\mathbb R\to\mathbb R$ be a (bounded, if necessary) Lipschitz continuous function. Are we able to show that $\partial f^{-1}\left(\left\{0\right\}\right)$ has Lebesgue measure $0$. If not, are there mild conditions under which the claim holds true?

I don't have much to contribute, since I struggle to find a good starting point.

EDIT: The question seems to be related to the notion of Hausdorff measures and maybe Sard's theorem. Since I've never heard about Hausdorff measures before reading the Wikipedia article, I hope there is a solution to this problem which doesn't need this concept.

Improve this question
$\endgroup$
  • $\begingroup$ Hint: Every closed subset can be realized as zero level set of 1-Lipschitz function. $\endgroup$ – Moishe Kohan May 5 '19 at 16:50
  • 1
    $\begingroup$ @MoisheKohan That's clear to me. If $(E,d)$ is a metric space and $A\subseteq E$, then $E\ni x\mapsto d(x,A)$ is $1$-Lipschitz continuous. And if $A$ is closed, its zero level set is precisely $A$. But I don't see how this helps here. $\endgroup$ – 0xbadf00d May 5 '19 at 17:57
  • $\begingroup$ Take a Cantor set of positive linear measure... $\endgroup$ – Moishe Kohan May 5 '19 at 21:04
  • $\begingroup$ @MoisheKohan Could you elaborate on your approach? $\endgroup$ – 0xbadf00d May 6 '19 at 4:29
3
$\begingroup$

Let $C\subset {\mathbb R}$ be a fat Cantor set, i.e. a subset of ${\mathbb R}$ which has positive Lebesgue measure and is homeomorphic to the standard Cantor set (i.e. is nonempty, compact, perfect and has empty interior). Let $f(x)=d(x,C)$ be the distance function to $C$. As you know, $f$ is 1-Lipschitz. At the same time, $C=f^{-1}(0)= \partial C$ (since $C$ has empty interior). With a bit more work one can replace $f$ with a function $g$ which is infinitely differentiable on ${\mathbb R}$ and still have $C=g^{-1}(0)$.

As for Sard's theorem, it is about images not preimages, so is unrelated to your question.

Related: The Boundary of a Lipschitz domain has Lebesgue measure zero?

Improve this answer
$\endgroup$
  • $\begingroup$ Any set which is closed and nowhere dense and of positive measure will do . So if we want $f$ to be bounded , replace $C$ with $C\cup \Bbb Z.$................+1 $\endgroup$ – DanielWainfleet May 6 '19 at 14:04
  • $\begingroup$ @DanielWainfleet sure $\endgroup$ – Moishe Kohan May 6 '19 at 15:14
  • $\begingroup$ Does the situation change if $f$ is assumed to be continuously differentiable? In that case, $f^{-1}(\{0\})$ is a $(d-1)$-dimensional submanifold of $\mathbb R^d$. $\endgroup$ – 0xbadf00d May 8 '19 at 8:39
  • $\begingroup$ @0xbadf00d No, it is not a submanifold. You can even get infinitely differentiable function. $\endgroup$ – Moishe Kohan May 8 '19 at 11:41
  • $\begingroup$ Sorry, I've intended to assume that $0$ is a regular value of $f$. Asked for that here: math.stackexchange.com/q/3218322/47771. $\endgroup$ – 0xbadf00d May 8 '19 at 11:45

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.