2
$\begingroup$

Question 1: In $\mathbb{R}^2$, let $l_1$,$l_2$ be two parallel lines and $l_3$ another line which is not parallel to $l_1$. Given two measurable sets $E_1$ and $E_2$ in $l_1$ and $l_2$ respectively, both of which have positive (1-dimension) Lebesgue measure. Denote by $l_{u,v}$ the line passing through two points $u,v$. Now, we define the third set $E_3$ in $l_3$ as follows. $E_3=\lbrace{p\in l_3: p=l_{u,v}\cap l_3, u\in E_1, v\in E_2\rbrace}$.

Can we say that $E_3$ contains an interval (or say "a segment") in $l_3$?

EDIT. Joseph Van Name answered Question 1 affirmatively. The following Question 2 is a generalization of Question 1, which is not answered.

Question 2: In $\mathbb{R}^n$, let $L_1$,$L_2$ be two parallel (n-1)-dimensional hyperplanes and $L_3$ another (n-1)-hyperplane which is not parallel to $L_1$. Given two measurable sets $E_1$ and $E_2$ in $L_1$ and $L_2$ respectively, both of which have positive ((n-1)-dimension) Lebesgue measure. Denote by $l_{u,v}$ the line passing through two points $u,v$. Now, we define the third set $E_3$ in $L_3$ as follows. $E_3=\lbrace{p\in L_3: p=l_{u,v}\cap L_3, u\in E_1, v\in E_2\rbrace}$.

Can we say that $E_3$ contains an $(n-1)$-ball in $L_3$?

It seems that the result in "An elementary proof and an extension of a theorem of Steinhaus,Kuczma, Marcin E.; Kuczma, Marek, Glasnik Mat. Ser. III 6(26) (1971), 11–18." noted by Joseph Van Name does not apply to the general case $n\geq 3$ of Question 2.

$\endgroup$
3
  • $\begingroup$ What is $w$ and what is $p$? $\endgroup$
    – user26107
    Jan 12, 2013 at 18:20
  • $\begingroup$ I think theorem 2.1 in the paper A GENERALIZATION OF STEINHAUS THEOREM AND SOME NONMEASURABLE SET could be used to prove this result. $\endgroup$ Jan 12, 2013 at 19:08
  • $\begingroup$ Better yet, theorem 2 in the paper "An elementary proof and an extension of a theorem of Steinhaus" probably works. $\endgroup$ Jan 12, 2013 at 20:01

1 Answer 1

2
$\begingroup$

We will need the following result from the paper that I mentioned in the comments "An elementary proof and an extension of a theorem of Steinhaus" by M. Kuczma and M. Kuczma that generalizes the Steinhaus theorem.

Thm:(Kuczma and Kuczma) Let $A,B$ be sets with positive inner Lebesgue measure and let $f(x,y)$ be a real-valued function of class $C^{1}$ in a region $D\supseteq A\times B$ with $f_{x}\neq 0,f_{y}\neq 0$ in $D$. Then the set $f(A\times B)$ contains an interval.

Now, to prove our result we shall assume without loss of generality that $\ell_{1}=\{(x,0)|x\in\mathbb{R}\},\ell_{2}=\{(x,1)|x\in\mathbb{R}\}$ and $\ell_{3}=\{(0,y)|x\in\mathbb{R}\}$. Let $U=\{(x,y)\in\mathbb{R}^{2}|x\neq y\}$. Then $U$ is the union of the two regions $\{(x,y)|x<y\}$ and $\{(x,y)|x>y\}.$ Let $f:U\rightarrow\mathbb{R}$ be the function where if $(x,y)\in U$, then $f(x,y)$ is the unique real number such that the points $(x,0),(y,1),(0,f(x,y))$ all fall on the same line. Then any reasonable high school algebra student should be able to tell you that $f(x,y)=\frac{x}{x-y}$. Therefore $f$ is real-analytic on $U$. Also, any calculus 3 student should also be able to tell you that $f_{x}(x,y)=\frac{-y}{(x-y)^{2}}$ and $f_{y}(x,y)=\frac{x}{(x-y)^{2}}$.` Therefore the first partial derivatives of $f$ with respect to $x$ and $y$ do not vanish on the set $U$.

If $E_{1},E_{2}\subseteq\mathbb{R}$ are sets with positive Lebesgue measure, then there are subsets $A_{1}\subseteq E_{1},A_{2}\subseteq E_{2}$ of positive Lebesgue measure such that either $a_{1}<a_{2}$ for $a_{1}\in A_{1},a_{2}\in A_{2}$ or $a_{1}>a_{2}$ for $a_{1}\in A_{1},a_{2}\in A_{2}$. Therefore the set $A_{1}\times A_{2}$ is completely contained in some region in $U$, so $f(A_{1}\times A_{2})$ contains some interval. We therefore conclude that the set $E_{3}$ defined in the problem must contain some interval as well.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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