MathOverflow is a question and answer site for professional mathematicians. 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

Let $X$, $Y$, and $Z$ be topological spaces. Let $f:X \rightarrow Y$. Further assume that for every continuous function $g:Y \rightarrow Z$, $g \circ f$ is continuous.

Question: Under what conditions on the topology of $Y$ and/or $Z$ can we conclude that $f$ must then be continuous?

This is easily achieved if $Y$ is (essentially) homeomorphic with $Z$, but this seems like drastic overkill.

To me, this seems like some kind of 'universal property' for the continuity of $f$; or maybe some kind of generalized "uniformly continuous" condition? Note: original question due to Nick James, but he's not a big web user, so I am asking on his behalf.

share|cite|improve this question
It is enough that $Y$ be homeomorphic to a subset of some product of $Z$, so a simple sufficient condition is that $Y$ be completely regular and $Z=[0,1]$. More generally, it is sufficient that there be enough continuous functions from $Y$ to $Z$ to separate points in $Y$ from disjoint closed subsets of $Y$. – Bill Johnson Jan 3 '12 at 19:32
@Bill: thanks. But somehow that still seems like asking 'too much' for what you get. For example, why 'regular' but not Hausdorff or something weaker still? – Jacques Carette Jan 3 '12 at 20:15
The second condition is sufficient with no assumptions on the topologies, Jacques. The first condition is just the observation that the second condition applies when $Y$ is completely regular and $Z=[0,1]$, which I (perhaps wrongly) assumed would take care of any situation that arose in Computer Science. $$ $$ I haven't thought about it, but would guess that the second condition is necessary in order for what you want to be true for all $X$ and all $f$. – Bill Johnson Jan 4 '12 at 18:44
Thanks for the expansion. The situation where this arises concerns 'generalized Stream' spaces, $Y = \mathbb{T}\rightarrow Q$ where $\mathbb{T}$ is time-like (might be $\mathbb{R}$, or $\mathbb{R}^+$ but also much weirder spaces), and $Q$ is some kind of (topological) algebra. – Jacques Carette Jan 4 '12 at 20:12
up vote 9 down vote accepted

It's possible to find a sufficient condition without putting any restraint on the spaces $X$ or $Y$.

Let $Z$ be the Sierpinski space, that is $Z$ has $\{0,1\}$ as a base set, and $\tau = \{\emptyset, \{0\}, \{0,1\}\}$ as a topology.

If $f: X\to Y$ is not continuous, then there is $V\subseteq Y$ open such that $f^{-1}(V)$ is not open in $X$. Defining $g_V: Y\to Z$ as mapping $V$ to $0$ and $Y\setminus V$ to $1$ we immediately see that $g_V\circ f: X \to Z$ is not continuous.

An interesting (and probably solved) question would be how to characterise the spaces $Z$ that have this universal property for all spaces $X, Y$.

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.