MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

Let $(X,\mathcal U)$ be a uniform space and let $U\in \mathcal U$. Is this statement true? $$\forall V\in \mathcal U, \exists W\in \mathcal U, U\circ W\subseteq V\circ U$$

I think if the above statement is true then we can easily prove: $$\forall V\in \mathcal U, \exists W\in \mathcal U, W\circ U\subseteq U\circ V$$

share|cite|improve this question
1) I am a bit confused between the text and the title of your "question". – András Bátkai Apr 2 '13 at 14:21
What do you mean exactly by $VU$ and $UW$? The product of these entourages? Are you working in a topological group? – Filippo Alberto Edoardo Apr 2 '13 at 14:30
$UV$ is $UoV$. I'll edit my question. – user31968 Apr 2 '13 at 14:50
I edited the header too. thanks. – user31968 Apr 2 '13 at 14:52
Need some motivation. Why do you care about this result? Why is this "research-level?" – David White Apr 2 '13 at 14:59
up vote 1 down vote accepted

The second statement follows from the first one by passing to inverses (i.e. reflecting along the diagonal).

Now consider the uniform structure on $\mathbb{R}$ consisting of all subsets of $\mathbb{R}^2$ that contain an open neighborhood of the diagonal.

Just to avoid confusion with left and right, I want to use the notation $U\circ V = \{(x,z)| \exists y: (x,y)\in U \wedge (y,z)\in V\}.$

Let $U:=\{(x,y)| |y-x|<1 \}\cup \mathbb{R}\times \{0\}$. $V:=\{(x,y)| |y-x|<1 \}$. Then $V\circ U=\{(x,y)| |y-x|<2 \}\cup \mathbb{R}\times \{0\}$.

Now let $W$ be any entourage and let $Z$ be some open neighborhood of $0$ such that $\{0\}\times Z\subset W$. Then $U\circ W$ will contain $\mathbb{R}\times Z$. Thus $U\circ W$ will not be contained in $V\circ U$. Since $W$ was arbitrary this should be a counterexample.

share|cite|improve this answer
My tries teached me: $U o W = (W^{-1}oU^{-1})^{-1}=(WoU)^{-1}$ and not $WoU$. too common mistake I made too! – user31968 Apr 3 '13 at 14:02
And for second part $(100,0.5)\in VoU$ because $(100,0)\in U$ and $(0,0.5)\in V$ but $(100,0.5)\notin \{(x,y)| |y-x|<2 \}\cup (\mathbb{R}\times \{0\})$. So $$V\circ U \ne \{(x,y)| |y-x|<2 \}\cup (\mathbb{R}\times \{0\})$$ – user31968 Apr 3 '13 at 14:14
@CC Thank you comments. I agree with your first comment. However your second comment suggests that your definition of $U\circ V$ is my definition of $V \circ U$. (To avoid this I included my definition of $U\circ V$). I still think this is a counterexample. – HenrikRüping Apr 3 '13 at 16:46

I recently found an answer to a similar question. Suppose: $$(\forall V\in \mathcal U)(\exists W\in \mathcal U)(U\circ W\subseteq V\circ U)$$ By axiom of choice,for each $V \in \mathcal U$, there's some symmetric $D_V\in \mathcal U$ such that $$D_V\circ D_V\subseteq V$$

and there's some symmetric $W_V\in \mathcal U$ such that:

$$U\circ W_V \subseteq D_V\circ U$$ and $$W_V\subseteq D_V$$

so $$W_V\circ U\circ W_V \subseteq W_V\circ D_V\circ U\subseteq D_V\circ D_V\circ U\subseteq V\circ U$$


$$\overline U=\bigcap_{W\in \mathcal U}W\circ U \circ W\subseteq \bigcap_{V\in \mathcal U}W_V\circ U \circ W_V\subseteq \bigcap_{V\in \mathcal U}V\circ U\subseteq U\circ U$$

Now see this thread.

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.