3
$\begingroup$

The volume of an $n$-ball of radius $1$ is $$V_{n}={\frac {\pi ^{n/2}}{\Gamma {\bigl (}{\tfrac {n}{2}}+1{\bigr )}}}.$$

The functional equation of Riemann zeta function is $${\displaystyle \pi ^{-{s \over 2}}\Gamma \left({s \over 2}\right)\zeta (s)=\pi ^{-{1 \over 2}+{s \over 2}}\Gamma \left({1 \over 2}-{s \over 2}\right)\zeta (1-s)}.$$

Combining the two equalities and replacing the Zeta function with Bernoulli numbers (generalized to negative orders), using the formula $B^+_n=-n\zeta(1-n)$ we get:

$$V_n B^+_n=V_{1-n}B^+_{1-n}.$$

In terms of umbral calculus, denoting the index-lowering operator as $\operatorname{eval}$, we get:

$$\operatorname{eval} V_n (B+1)^n=\operatorname{eval} V_{1-n}(B+1)^{1-n}$$

I can note here that index-lowering operator coresponds to finding scalar part of a quantity, and as such to the Euler's characteristic of a set.

But what does this intuitively mean? We have expressions of volumes of n-balls with radius of Bernoulli umbra (plus 1). What could it even mean intuitively?

P.S. Asked this question in Math.StackExchange several months ago, I think, I can now post it here as well.

$\endgroup$
2
  • $\begingroup$ for positive real $n \le \frac{1}{2}$ this might suggest some kind of duality between the Hausdorff content of the $n$ and $1-n$ dimensional spheres… this is very much a blind idea but perhaps there should exist natural pairs of fractals of $n$ and $1-n$ dimension so that the Hausdorff content of their balls obeys said identity. Unfortunately without a notion of negative dimensional spaces with suitable spheres (I know your lattice construction but it’s unclear how to fit it here) we can’t find such a duality for higher dimensions. $\endgroup$ Commented Oct 7, 2023 at 2:21
  • $\begingroup$ In fact fractal spaces whose “spheres” up to some suitable definition satisfy this could serve as strong examples for this: mathoverflow.net/questions/432880/… $\endgroup$ Commented Oct 7, 2023 at 2:23

0

You must log in to answer this question.