Does the following notion (in some sense reverse to continuity) have a name? What are interesting properties of this concept?

$f$ is "anticontinuous" in $x$ if $$\forall \epsilon>0 \exists \delta>0:f[(x-\epsilon;x+\epsilon)]\supseteq (f(x)-\delta;f(x)+\delta).$$

(Here $f[X]$ is the image of a set $X$ under a map $f$.)


As Berci commented, when we demand that this hold for all $x$ in the domain of the function we're just saying that $f$ maps open sets to open sets - this is the definition of an open map. Similarly, a closed map is a map sending closed sets to closed sets. Closedness, openness, and continuity are fundamentally independent:

  • If $X$ has more than one element, the identity map on $X$ is closed and open but not continuous when we equip the domain with the indiscrete topology and the codomain with the discrete topology.

  • The "left inclusion" map $$\mathbb{R}\rightarrow\mathbb{R}^2:a\mapsto (a,0)$$ (with the usual topologies on $\mathbb{R}$ and $\mathbb{R}^2$) is closed and continuous but not open.

  • The "left projection" map $$\mathbb{R}^2\rightarrow\mathbb{R}: (a,b)\mapsto a$$ is continuous and open but not closed.

It's a good exercise to check each of these examples.

(Meanwhile, the obvious fourth notion - "preimages of closed sets are closed" - is equivalent to continuity, since the complement of the preimage is the preimage of the complement).

I've seen in conversation the term "open at $x$" used to refer to the situation you describe, but I haven't seen that in a formal text so I don't know if it's universally accepted.


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.