Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm trying to prove that for any $K\subset[0,1]$ such that K is compact and $m(K)=0$ with no isolated points, there exists a continuous, monotonic increasing function $f$ which maps $[0,1]$ onto $[0,1]$, and such that $f'(x)=0$ for every $x\in[0,1]-K$ (i.e. $f'=0$ a.e. $x$).

Now, it is clear that the Cantor-Lebesgue function $F$ will do for $K=\mathscr{C}$, and a modification of $F$ will accommodate any Cantor-like set (sets of constant dissection).

I'm wondering if there are any other sets $K$ with the properties listed. If so, can they be narrowed down into cases and a constructive proof applied to each case?

If not, then how does one go about proving the statement for general $K$?


Here's my current idea.

Let $K$ be as above. Define $K_{0}=[\inf K, \sup K]$. Dissect the middle third of $K_{0}$ into pieces $J_{1}^{1}$ and $J_{1}^{2}$ by removing the open middle segment. Define $K_{1}:=[\inf K, \sup K\cap J_{1}^{1}]\cup[\inf K\cap J_{1}^{2}, \sup K]$. Then dissect this set by removing the middle thirds of the two component intervals, yielding $J_{2}^{1}$, $J_{2}^{2}$, $J_{2}^{3}$, $J_{2}^{4}$. Then put $K_{2}=[\inf K, \sup K\cap J_{2}^{1}]\cup[\inf K\cap J_{2}^{2}, \sup K\cap J_{2}^{2}]\cup[\inf K\cap J_{2}^{3}, \sup K\cap J_{2}^{3}]\cup[\inf K\cap J_{2}^{4}, \sup K]$. Then continue dissecting $K$ as such. The $\sup$ and $\inf$ appearing in each definition exist since every $J$ is compact, and the intersection of two compact sets is again compact.

Now, at the jth step of the construction, we have $K_{j}$ being the union of $2^{j}$ component subintervals. On these subintervals, simply require $f$ to be linearly increasing, and on the corresponding open subinterval which is removed, define $f=\frac{n}{2^{j}}$ where $n$ denotes the nth subinterval going left to right. Then $f$ evaluated at the least element of $K$ is $1$ (if we begin by $n=0$) and $f$ evaluated at the largest element of $K$ is $1$, and the function is monotonic increasing.

I guess now I'm running into the same problem...establishing continuity. I want to say something like \begin{equation*} |f_{j+1}(x)-f_{j}(x)|<\frac{2}{j} \end{equation*} so that I can establish uniform convergence.

These Cantor problems never get any easier ><

share|cite|improve this question
up vote 3 down vote accepted

EDIT: In response to the objection raised in the comments, I think this works:

Since $K$ is compact, its complement in $[0,1]$ is open, thus a countable union of open intervals. Enumerate them any way you like. For $x$ in the closure of interval $I_k$, define $f(x)$ to be $\sum2^{-j}$ where $j$ runs over all those numbers such that $I_j$ is to the left of $I_k$.

share|cite|improve this answer
How will you enumerate the rationals so that you obtain continuity and monotonicity? Just say it so? I like your idea though, let me think about it for a little bit. – Taylor Martin Jun 17 '12 at 1:33
Monotonicity is not a problem. The open intervals don't overlap, so if interval $I$ is left of interval $J$, then $f(x)\lt f(y)$ for all $x$ in $I$ and all $y$ in $J$, and that carries over to the points in $K$. But continuity, that's a problem. I'll leave my answer up so maybe someone can use it as a starting place, but I think continuity is an issue. – Gerry Myerson Jun 17 '12 at 1:56
I might have a better idea. See post. – Taylor Martin Jun 17 '12 at 3:00
Thank you Gerry ~ This is much simpler than my idea! – Taylor Martin Jun 17 '12 at 4:01

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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