# The Unapologetic Mathematician

## Uniform Spaces

Now let’s add a little more structure to our topological spaces. We can use a topology on a set to talk about which points are “close” to a subset. Now we want to make a finer comparison by being able to say “the point $a$ is closer to the subset $A$ than $y$ is to $B$.” We’ll do this with a technique similar to neighborhoods. But there we just defined a collection of neighborhoods for each point. Here we will define the neighborhoods of all of our points “uniformly” over the whole space.

To this end, we will equip our set $X$ with a family $\Upsilon$ of subsets of $X\times X$ called the “uniform structure” on our space, and the elements $E\in\Upsilon$ will be “entourages”. We will write $E[x]$ for the set of $y$ so that $(x,y)\in E$, and we want these sets to form a neighborhood filter for $x$ as $E$ varies over $\Upsilon$. Here we go:

• Every entourage $E$ contains the diagonal $\{(x,x)|x\in X\}$.
• If $E$ is an entourage and $E\subseteq F\subseteq X\times X$, then $F$ is an entourage.
• If $E$ and $F$ are entourages, then $E\cap F$ is an entourage.
• If $E$ is an entourage then there is another entourage $F$ so that $(x,y)\in F$ and $(y,z)\in F$ imply $(x,z)\in E$.
• If $E$ is an entourage then its reflection $\bar{E}=\{(y,x)|(x,y)\in E\}$ is also an entourage.

The first of these axioms says that $x\in E[x]$, as we’d hope for a neighborhood. The next two ensure that the collection of all the $E[x]$ forms a neighborhood filter for $x$, but it does so “uniformly” for all the $x\in X$ at once. This means that we can compare neighborhoods of two different points because each of them comes from an entourage, and we can compare the entourages. The fourth axiom is like the one I omitted from my discussion of neighborhoods; every collection of entourages gives rise to a topology, but topologies can only give back uniform structures satisfying this requirement. Finally, the last axiom gives the very reasonable condition that if $y\in E[x]$, then $x\in \bar{E}[y]$. That is, if one point is in a neighborhood of another, then the other point should be in a neighborhood of the first. Sometimes this requirement is omitted to get a “quasi-uniform space”.

Now that we can compare closeness at different points, we can significantly enrich our concept of nets. Before now we talked about a net $x_\alpha$ converging to a point $x$ in the sense that the points $x_\alpha$ eventually got close to $x$. But now we can talk about whether the points of the net are getting closer to each other. That is, for every entourage $E$ there is a $\gamma\in D$ so that for all $\alpha\geq\gamma$ and $\beta\geq\gamma$ the pair $(x_\alpha,x_\beta)$ is in $E$. In this case we say that the net is “Cauchy”.

Now, if the full generality of nets still unnerves you, you can restrict to sequences. Then the condition is that there is some number $N$ so that for any two numbers $m$ and $n$ bigger than $N$ we have $x_m\in E[x_n]$. This gives us the notion of a Cauchy sequence, which some of you may already have heard of.

We can also enrich our notion of continuity. Before we said that a function $f:X\rightarrow Y$ from a topological space defined by a neighborhood system $(X,\mathcal{N}_X)$ to another one $(Y,\mathcal{N}_Y)$ is continuous at a point $x\in X$ if for each neighborhood $V\in\mathcal{N}_Y(f(x))$ contained the image $f(U)$ of some neighborhood $U\in\mathcal{N}_X(x)$, and we said that $f$ was continuous if it was continuous at every point of $X$.

Now our uniform structures allow us to talk about neighborhoods of all points of a space together, so we can adapt our definition to work uniformly. We say that a function $f:X\rightarrow Y$ from a uniform space $(X,\Upsilon_X)$ to another one $(Y,\Upsilon_Y)$ is uniformly continuous if for each entourage $F\in\Upsilon_Y$ there is some entourage $E\in\Upsilon_X$ that gets sent into $F$. More precisely, for every pair $(x_1,x_2)\in E$ the pair $(f(x_1),f(x_2))$ is in $F$.

In particular, any neighborhood of a point $f(x)\in Y$ is of the form $F[f(x)]$ for some entourage $F\in\Upsilon_Y$. Then uniform continuity gives us an entourage $E\in\Upsilon_X$, and thus a neighborhood $E[x]$ which is sent into $F[f(x)]$. Thus uniform continuity implies continuity, but not necessarily the other way around. It is possible that a function is continuous, but that the only ways of picking neighborhoods to satisfy the definition do not come from entourages.

These two extended definitions play well with each other too. Let’s consider a uniformly continuous function $f:X\rightarrow Y$ and a Cauchy net $x_\alpha$ in $X$. Then I assert that the image $f(x_\alpha)$ of this net is again Cauchy. Indeed, for every entourage $F\in\Upsilon_Y$ we want a $\gamma$ so that $\alpha\geq\gamma$ and $\beta\geq\gamma$ imply that the pair $(f(x_\alpha),f(x_\beta))$ is in $F$. But uniform continuity gives us an entourage $E\in\Upsilon_X$ that gets sent into $F$, and the Cauchy property of our net gives us a $\gamma$ so that $(x_\alpha,x_\beta)\in E$ for all $\alpha$ and $\beta$ above $\gamma$. Then $(f(x_\alpha),f(x_\beta))\in F$ and we’re done.

It wouldn’t surprise me if one could turn this around like we did for neighborhoods. Given a map $f:X\rightarrow Y$ which is not uniformly continuous use the uniform structure $\Upsilon_X$ as a directed set and construct a net on it which is Cauchy in $X$, but whose image is not Cauchy in $Y$. Then one could define uniform continuity as preservation of Cauchy nets and derive the other definition from it. However I’ve been looking at this conjecture for about an hour now and don’t quite see how to prove it. So for now I’ll just leave it, but if anyone else knows the right construction offhand I’d be glad to hear it.

November 23, 2007 - Posted by | Point-Set Topology, Topology

1. Motivating examples of uniform spaces being given by metric spaces and by topological groups. (In particular, compare the fourth axiom on entourages with the triangle inequality.) Probably you’ll be talking about that?

Uniform continuity is something stronger than preservation of Cauchy nets; consider the fact that the squaring map R –> R takes Cauchy nets to Cauchy nets but is not uniformly continuous. I think somehow nets in the usual sense are too “local” (e.g., they converge to a single point in Hausdorff spaces), or uniform continuity too global, for uniform continuity to be easily captured by nets.

Unless: you change the concept of net so that there is a notion of convergence to the diagonal of X; e.g.: instead of nets D –> X which converge to a point in X, consider “uniform nets” f: D –> Rel(X, X) valued in the set of binary relations on X which converge to the diagonal of X, or even more general relations. Here’s an offhand attempt at definition: say that the uniform net f converges to a binary relation R if for every entourage E, there exists x in D such that x <= y implies that the binary relation f(y) is contained in E o R (the composite of the relations E and R). This definition may have to be tweaked a bit to make everything come out just right.

By the way: spurred in part by your posts, I’m thinking a bit about another approach to topology similar to nets, but based on ultrafilter convergence. (In particular, I wanted to understand better this relational beta-module business, which turns out to be a very attractive piece of lax algebra [in the categorical sense].) Two papers have caught my eye: one by Walter Tholen, which gives a uniform treatment of ordered sets, metric spaces, and general topological spaces (and it seems to me uniform spaces can also be fit within his framework). Another is by Claudio Pisani which characterizes exponentiable topological spaces in lax algebraic terms, and which usefully gives complete proofs of things including Barr’s relational beta-module characterization of topological spaces.

Comment by Todd Trimble | November 23, 2007 | Reply

2. That’s a great counterexample, Todd. Thanks. And of course I’ll be talking about topological groups, and particularly about ordered groups, which will finally open the road to my first official mention of the real numbers. Have you noticed I’ve gotten all this way without talking about them yet? :D

Comment by John Armstrong | November 24, 2007 | Reply

3. [...] Uniform Spaces Okay, in a uniform space we have these things called “Cauchy nets”, which are ones where the points of the net [...]

Pingback by Complete Uniform Spaces « The Unapologetic Mathematician | November 29, 2007 | Reply

4. [...] correspond to our usual notion of magnitudes like distances, let’s refine our concept of a uniform space to take into account this idea of [...]

Pingback by Metric Spaces and continuity of real-valued functions « The Unapologetic Mathematician | December 10, 2007 | Reply

5. [...] many other ones which give rise to different distance functions, but the same topology and the same uniform structure. And often it’s the topology that we’ll be most interested [...]

Pingback by Products of Metric Spaces « The Unapologetic Mathematician | August 19, 2008 | Reply

6. [...] an inverse , and it only makes sense that these be homeomorphisms. And to capture this, we put a uniform structure on our space. That is, we specify what the neighborhoods are of , and just translate them around to [...]

Pingback by Topological Vector Spaces, Normed Vector Spaces, and Banach Spaces « The Unapologetic Mathematician | May 12, 2010 | Reply

7. [...] we say that the sequence is Cauchy a.e. if there exists a set of measure zero so that is a Cauchy sequence of real numbers for all . That is, given and there is some natural number depending on and so [...]

Pingback by Convergence Almost Everywhere « The Unapologetic Mathematician | May 14, 2010 | Reply

8. [...] then , , and itself are all uniformly continuous. [...]

Pingback by The Metric Space of a Measure Ring « The Unapologetic Mathematician | August 6, 2010 | Reply

9. You left out an axiom: that there exists an entourage at all (or equivalently, in light of the other axioms, that X × X is an entourage).

Comment by Toby Bartels | June 1, 2011 | Reply

10. That’s a good point, though does the existence of an “empty” uniform structure lead to any huge problems?

Comment by John Armstrong | June 1, 2011 | Reply