I understand that the Weierstrass function is an example of a function that is continuous everywhere but differentiable nowhere. I was thinking if it is possible to "make the Weierstrass function differentiable" somehow. One approach on this was already discussed here.

My approach is the following: The Weierstrass function (let us call it $w$) can be represented as the set $$W := \{ (x,y) \in \mathbb{R}^2 \mid w(x)=y \}$$

I am wondering if there is a differentiable parametrization $p:\mathbb{R} \to \mathbb{R}^2$ of $W$, in the sense that

  • $W = p(\mathbb{R}) := \{ p(t) \mid t \in \mathbb{R} \}$
  • $p$ is injective
  • $p$ is differentiable (almost everywhere would be fine)

Is there such a parametrization?

My guess is no, because based on a differentiable parametrization, we could probably compute the derivative of $w$ itself. However, I have not found out how to do this yet...


Of course we'd need to have $p'(t)=0$ almost everywhere.

In fact the answer is yes, at least for almost-everywhere differentiability. The "Cantor function" $x:[0,1]\to[0,1]$ is a continuous non-decreasing surjective function with the following property: There exists an open set $E\subset[0,1]$ with $m(E)=1$ such that $x$ is constant on each connected component of $E$. So if we let $p(t)=(x(t), w(x(t)))$ then $p$ is constant on each connected component of $E$, hence $p'=0$ almost everywhere.

  • $\begingroup$ Thank you for your answer! Unfortunately, I forgot to specify that my parametrization should also be injective (otherwise, it would be a strange parametrization). I think in that case, the Cantor function cannot be applied directly? $\endgroup$
    – Peter
    Oct 27 '17 at 16:25

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.