2
$\begingroup$

Let $f:\mathbb R/\mathbb Z\to \mathbb R$ a function (1-periodic). Is such a function bounded ? (it's the fact that f is defined on the circle that disturb me). Indeed, for such a function (usually at least $L^1$), I often see in my course $\|f\|_{L^\infty }$, but I don't see why it would be well defined if $f$ is not bounded.

For example, does $f(x)=\tan(\pi/2 x)$ is defined on $S^1$ ? I really have problem with this $\mathbb R/\mathbb Z$.

$\endgroup$
  • 1
    $\begingroup$ To be clear, you aren't assuming your functions are continuous? $\endgroup$ – Eric Wofsey Jul 7 '16 at 7:07
  • $\begingroup$ @EricWofsey: no I don't. $\endgroup$ – MathBeginner Jul 7 '16 at 7:09
  • $\begingroup$ $tan(x/(2\pi))$ is unbounded but still defined on (0,1). You could define a function to equal tan(x/2pi) when $x\neq 0$ and something else when $x=0$. $\endgroup$ – mathematician Jul 7 '16 at 7:13
  • $\begingroup$ Eric Wofsey's very nice answer shows this need not be true. However, if $f$ is continuous, then since $S^{1}$ is compact, $f$ must have compact image. Since every compact subset of $\mathbb{R}$ is bounded, $f$ is bounded. Of course, you probably already knew this. :) $\endgroup$ – Alex Wertheim Jul 7 '16 at 7:20
3
$\begingroup$

As Eric Wofsey has said, such functions need not be bounded. But I think it is important to note as well that if the function is continuous, then it will be bounded. This follows from the fact that $\mathbb{R/Z}$ is compact, and so its image under a continuous map is also compact. Since compact subsets of $\mathbb{R}^N$ are closed and bounded...

Anyhow, his function is not continuous, and nor is the stereographic projection as described (at least, it is not continuous over all of $\mathbb{R/Z}$), so this is not a concern.

$\endgroup$
  • $\begingroup$ so $f(x)=\tan(\frac{\pi}{2}x)$ is not continuous on $\mathbb R/\mathbb Z$ ? $\endgroup$ – MathBeginner Jul 7 '16 at 7:33
  • $\begingroup$ Well, it's not even technically well-defined on $\mathbb{R/Z}$ (since we would need to have $0 = f(0) = f(1) "=" \infty$, but those do not agree. But once you fix that concern (e.g. define instead $f(x) = \tan(\pi x)$), then what is $f(1/2)$? $\endgroup$ – Simon Rose Jul 7 '16 at 7:35
  • $\begingroup$ Thank you for this nice answer ! I understand now. Just to be, $f(x)=\tan(\pi x)$ is well defined on $\mathbb R/\mathbb Z$, isn't it ? (even if it not continuous). $\endgroup$ – MathBeginner Jul 7 '16 at 7:37
  • $\begingroup$ It's almost well-defined. What you would have to do is define $$ f(x) = \begin{cases} \tan(\pi x) & x \neq 1/2 \\ c & x = 1/2 \end{cases} $$ for some choice of a real number $c$. Then it is defined on all of $\mathbb{R/Z}$, but not continuous. $\endgroup$ – Simon Rose Jul 7 '16 at 14:39
3
$\begingroup$

There are certainly plenty of unbounded functions $f:\mathbb{R}/\mathbb{Z}\to\mathbb{R}$. For instance, let $x_0,x_1,x_2,\ldots\in\mathbb{R}/\mathbb{Z}$ be any infinite sequence of distinct points. Then you could define $f(x_n)=n$ and $f(x)=0$ if $x\neq x_n$ for all $n$. This $f$ is unbounded, since its image is all of $\mathbb{N}$.

For a slightly less artificial example, you could take a stereographic projection: identify $\mathbb{R}/\mathbb{Z}$ with the unit circle in $\mathbb{R}^2$, and define $f(p)$ for $p\neq (0,1)$ as follows (you can then define $f(0,1)$ however you want). Draw the line from $(0,1)$ to $p$. This line will intersect the $x$-axis at a point $(q,0)$; define $f(p)=q$. This map is in fact surjective, since for any $q\in\mathbb{R}$, the line from $(q,0)$ to $(0,1)$ will intersect the circle at another point $p$, and then $f(p)=q$.

$\endgroup$
  • $\begingroup$ Great examples! Stereographic projection is a natural example which, to my chagrin, didn't immediately come to mind before more bizarre functions. +1 $\endgroup$ – Alex Wertheim Jul 7 '16 at 7:18

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.