I have been reading about uniform spaces and topological groups. There does not look to be a lot of literature on the topic, much less accesible literature, and the books that I have been reading do not mention any examples of uniform spaces, other than metric spaces and topological groups. There is another which I came up with, which is $\mathbb R$ with uniform structure containing $U_r = \lbrace (x,y) \in \mathbb R ^2 \ | \ x = y$ or exactly one of $x,y$ irrational and $|x-y| < r \rbrace$ but this example is a bit trivial since the interior of each $U_r[x]$ is just the point $x$. Can anyone give me some more examples?

  • $\begingroup$ Completely regular spaces are exactly the uniform spaces. $\endgroup$ – Camilo Arosemena-Serrato Oct 8 '13 at 23:16
  • $\begingroup$ Correction: $\:$ If Dependent Choice then completely regular spaces are exactly the uniformizable spaces. $\hspace{.36 in}$ $\endgroup$ – user57159 Oct 8 '13 at 23:22
  • 1
    $\begingroup$ @Ricky: That’s just hyper-pedantry: in this, as in most contexts, full $\mathsf{AC}$ is the default assumption. The real correction is that the completely regular spaces are precisely the uniformizable spaces; they don’t become uniform spaces until they’re given a uniform structure. It seems reasonably clear that Daron is looking for examples of uniform structures. I’ll see if I can come up with a nice example or two. $\endgroup$ – Brian M. Scott Oct 9 '13 at 10:50

For now, at least, I’m going to give only one example, because the first order of business is to explain why your sets

$$U_r=\Delta\cup\{\langle x,y\rangle\in\Bbb R^2:|x-y|<r\text{ and exactly one of }x\text{ and }y\text{ is irrational}\}$$

for $r>0$ do not give you a uniformity, where $\Delta$ is the diagonal in $\Bbb R^2$.

$\mathscr{U}=\{U_r:r>0\}$ is not a uniformity: it doesn’t satisfy the condition that if $U\subseteq V\subseteq\Bbb R^2$, and $U\in\mathscr{U}$, then $V\in\mathscr{U}$. It also fails to satisfy the condition that if $U\in\mathscr{U}$, there is a $V\in\mathscr{U}$ such that $V\circ V\subseteq U$. If $\langle x,y\rangle,\langle y,z\rangle\in V\in\mathscr{U}$, then $x$ and $z$ are either both rational or both irrational; if $x\ne z$, then $\langle x,z\rangle\in(V\circ V)\setminus U$. In other words, $\mathscr{U}$ fails to satisfy conditions $(2)$ and $(4)$ of this definition.

Failure to satisfy $(2)$ isn’t catastrophic: we could just replace $\mathscr{U}$ by $$\mathscr{V}=\{V\subseteq\Bbb R^2:V\supseteq U_r\text{ for some }r>0\}\;,$$ taking $\mathscr{U}$ as a base for the uniformity rather than the uniformity itself. However, $\mathscr{V}$ still isn’t a uniformity: it still fails to satisfy $(4)$, since there is no $V\in\mathscr{V}$ such that $V\circ V\subseteq U_1$, for instance.

I can come up with a uniformity on $\Bbb R$ that is somewhat similar to your attempt. For $r>0$ let

$$U_r=\{\langle x,y\rangle\in\Bbb R^2:x=y\in\Bbb R\setminus\Bbb Q\text{ or }|x-y|<r\}\;,$$

and let $$\mathscr{U}=\{U\subseteq\Bbb R^2:U\supseteq U_r\text{ for some }r>0\}\;.$$ This $\mathscr{U}$ is a uniformity. Most of the conditions are very easy to check. For $(4)$, let $U\in\mathscr{U}$. There is an $r>0$ such that $U_r\subseteq U$, and it’s easy to check that $U_{r/2}\circ U_{r/2}\subseteq U_r\subseteq U$.

Now let’s see what topology $\mathscr{U}$ generates on $\Bbb R$. Call this topology $\tau$, and for each $U\in\mathscr{U}$ and $x\in\Bbb R$ let $$U[x]=\{y\in\Bbb R:\langle x,y\rangle\in U\}\;.$$ Suppose first that $x$ is irrational. Then $U_r[x]=\{x\}$ for each $r>0$, so $\{x\}$ is a nbhd of $x$. Thus, $x$ is in the interior of $\{x\}$, so $\operatorname{int}_\tau\{x\}=\{x\}$, and $\{x\}$ is $\tau$-open: each irrational is an isolated point of $\langle\Bbb R,\tau\rangle$.

Note that I use the term neighborhood (nbhd) in the broad sense, not requiring it to be open: a set $N$ is a nbhd of $x$ if $x$ is in the interior of $N$.

Now suppose that $x$ is rational. Then $$U_r[x]=\{y\in\Bbb R:|x-y|<r\}=(x-r,x+r)$$ for each $r>0$. Since $\{U_r:r>0\}$ is a base for the uniformity $\mathscr{U}$, $\{U_r[x]:r>0\}$ is a local base for the topology $\tau$ at the point $x$, so $\{(x-r,x+r):r>0\}$ is a local base at each rational.

Without using uniformities we can describe the space $\langle\Bbb R,\tau\rangle$ by saying that it’s obtained from the usual topology on $\Bbb R$ by isolating each irrational. The resulting space is a fairly important example in topology; it’s called the Michael line, and you can read about it in Dan Ma’s Topology Blog. Note that this is an example of a uniform space that is not metrizable and cannot be turned into a topological group.


It is sometimes useful to know that for every compact Hausdorff space $X$ there is precisely one uniform structure whose uniform topology is the topologyof $X$. The entourages are just the neighborhoods of the diagonal in $X \times X$.

This is consonant with the fact that on compact Hausdorff spaces, every continuous function to a uniform space is uniformly continuous (I'm going on memory for this last bit).


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.