2
$\begingroup$

I want to have an example of a non-constant function which has uncountably many zeros?

Is the following function continuous?

$f(x)=\prod\limits_{\alpha\in \mathbb{R\setminus Q}}(x-\alpha)$ If it so, it will be the required example.

$\endgroup$
  • $\begingroup$ You are trying to find a continuous function on the line that is $0$ at all irrationals? That function by continuity would be everywhere $0.$ Take a look at $f(x)=0, x\in [0,1], f(x) = x, x<0, f(x) = x-1,x>0.$ $\endgroup$ – zhw. Sep 21 '15 at 16:04
  • $\begingroup$ I assume we're only talking about functions $\Bbb R \to \Bbb R$? $\endgroup$ – pjs36 Sep 21 '15 at 16:32
8
$\begingroup$

Let $C$ be any uncountable closed set, for example the Cantor set, and for every $x \in \mathbb{R}$, define $f(x) = d(x,C) = \inf\{d(x,c) : c \in C\}$. Since $C$ is closed, $f(x) = 0$ if and only if $x \in C$. Since $C$ is uncountable, we only need to verify that $f$ is continuous. To see this, fix $x,y \in \mathbb{R}$, and $c \in C$. Then by the triangle inequality, we have $$d(x,c) \leq d(x,y) + d(y,c)$$ Taking the infimum over $c \in C$ gives us $$f(x) \leq d(x,y) + f(y)$$ Similarly (interchanging the roles of $x$ and $y$ above), we have $$f(y) \leq d(x,y) + f(x)$$ Combining these two inequalities, we conclude that $$|f(x) - f(y)| \leq d(x,y)$$ which shows that $f$ is continuous (in fact, uniformly continuous).

P.S. I used the notation $d(x,y)$ above for the distance between $x$ and $y$, because the same proof works in any metric space that has an uncountable closed set $C$. In the case of $\mathbb{R}$, of course, we set $d(x,y) = |x-y|$.

$\endgroup$
2
$\begingroup$

Your product does not converge for $x \in \mathbb{Q}$, so it does not define a continuous function. In fact the irrationals are dense, so a continuous function which is constant on the irrationals is constant everywhere.

If you are OK with a function which is locally constant in some regions, you can just take a continuous function with compact support, such as $f(x)=\max \{ 1-|x|,0 \}$. In some sense this feels like cheating to me, because it only achieves the goal by gluing a constant function to a non-constant function.

If you want a function which is nowhere locally constant and yet has uncountably many zeros, I can give a somewhat nonconstructive example: the trajectories of a Brownian motion have this property with probability $1$.

$\endgroup$

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.