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 came across the following problem in my self-study:

Let $f:[0,1] \rightarrow [0,1]$ be the Cantor (ternary) function, and let $g(x) = f(x) + x$. Then:

(a) $g$ is bijection from $[0,1]$ to $[0,2]$, and $h = g^{-1}$ is continuous.

(b) If $C$ is the Cantor (ternary) set, $m(g(C))=1$.

(c) It follows that $g(C)$ contains a Lebesgue nonmeasurable set $A$. Let $B = g^{-1}(A)$, and show that $B$ is Lebesgue measurable but not Borel.

(d) There exist a Lebesgue measurable function $F$ and a continuous function $G$ on $\mathbb{R}$ such that $F \circ G$ is not Lebesgue measurable.

I think I can prove (a), but I am not having any nice ideas on how to proceed for the remaining (3) parts, and I am interested to see if anyone visiting knows how to tackle this interesting exercise.

share|cite|improve this question
For (b) consider $[0,2]-g([0,1]-C)$ which is a countable disjoint union. – Tim kinsella Feb 1 '12 at 6:58
For (c) note that $B$ is a subset of a null set. Use the continuity of $h$ to show that $B$ can not be Borel. – Tim kinsella Feb 1 '12 at 7:07
sorry, i meant for (b) note that $g(C)= [0,2]-g([0,1]-C)$ and that $g([0,1]-C)$ is a countable disjoint union of intervals. – Tim kinsella Feb 1 '12 at 7:23
up vote 2 down vote accepted

For (b), the key observation is that, as $g$ is continuous, the image of an interval is an interval, and so $m(g((a,b)))=g(b)-g(a)$.

Then, as Tim said, we can use that the complement of $C$ is a countable disjoint union of open intervals $C_k$ whose measures add to 1, and that on these intervals we have $f=0$. So, if $C_k=(a_k,b_k)$, then $$ m(g(C_k))=g(b_k)-g(a_k)=f(b_k)-f(a_k)+b_k-a_k=b_k-a_k=m(C_k). $$ So \begin{eqnarray} 2-m(g(C))&=m([0,2]\setminus g(C))=m(g(\bigcup_k C_k))=m(\bigcup_k g(C_k))=\sum_k m(g(C_k)) =\sum_k m(C_k)=1, \end{eqnarray} i.e. $m(g(C))=1$.

For (c), it is an established result that any set of positive measure contains a non-measurable subset. So let $A\subset g(C)$ be non-measurable, and put $B=g^{-1}(A)$. Since $B\subset C$ and $C$ is a null-set, $B$ is also a null-set, and it particular it is measurable. But it is not Borel: since $g^{-1}$ is continuous, the pre-image of a Borel set is Borel; as $A=g(B)$, this would make $A$ Borel and thus measurable, which it's not.

Finally, for (d), the key is that $$ 1_A=1_{g(B)}=1_B\circ g^{-1}. $$ The characteristic $1_A$ is not measurable since $A$ isn't. But $B$ is measurable so $1_B$ is, and $g^{-1}$ is continuous. So the composition of measurable functions may fail to be measurable, even if one of them is continuous.

share|cite|improve this answer

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.