1
$\begingroup$

As a takeaway of this post we have the following result.

P. Let $f:[0,1]\to\mathbb{R}$ be infinitely differentiable such that for all $x\in[0,1]$ the sequence $\{f^{(n)}(x)\}$ is eventually $0$. Then $f$ is a polynomial.

Now take $f:[0,1]\to\mathbb{R}$ be infinitely differentiable and be $S$ the set of $x$ such that the sequence $\{f^{(n)}(x)\}$ is not eventually $0$.

I got interested in answering this question: which are valid $S$? How small can $S$ be?

Property. $S$ has no isolated points.

This is a consequence of P.

Property. Let $N=\{x\in[0,1]:f\text{ is not a polynomial locally at }x\}$. Then $N$ is a perfect set (possibly empty).

Observe that $S\subset N$.

I realised that $S$ could have zero measure. In this comment I explained the construction of an infinitely differentiable function whose derivatives eventually vanish except in the Cantor set. (I.e. for which $N=\text{ Cantor set }$).

  • Does anyone know of any research material pointed in this direction?
  • Has the function described above appeared in the literature before?

Thank you in advance.

$\endgroup$
4
  • $\begingroup$ As to the last question, it has appeared at least here on MO :) mathoverflow.net/questions/94038/… $\endgroup$ Commented Apr 9 at 19:39
  • $\begingroup$ Cool. I find it just amazing that SE's algorithm put me in contact with a previous ideator of one of my ideas. $\endgroup$
    – aleph2
    Commented Apr 9 at 20:49
  • $\begingroup$ Your idea is very nice, and as to priority, time is but an illusion ;) $\endgroup$ Commented Apr 10 at 7:30
  • $\begingroup$ So is yours ;) And I agree! $\endgroup$
    – aleph2
    Commented Apr 10 at 19:58

0

You must log in to answer this question.

Browse other questions tagged .