9
$\begingroup$

In The Basel problem revisited? a question about the function, similar to sinc, $f(x)$ was asked:

$$f(x) = \prod_{n=1}^\infty \left ( 1+ \frac{x^3}{n^3} \right ) = \prod_{n=1}^\infty \left ( 1+ \frac{x}{n} \right ) \left ( 1+ \frac{x}{n}\omega \right )\left ( 1+ \frac{x}{n} \omega^2 \right )$$

which was shown to be equal to:

$$f(x) = \sum_{k=0}^\infty x^{3k}\zeta_k(3)$$

where we have defined:

$$\zeta_k(3) = \sum_{1\le n_1 <n_2 \cdots <n_k} \left(\frac{1}{n_1 n_2 \cdots n_k} \right) ^3$$

and which is also equal to:

$$f(x) = \frac{1}{\Gamma(1+x)\Gamma(1+\omega x) \Gamma(1+\omega^2 x)}$$

Let me define similar to $\sin,\cos$ functions $c_0,c_1,c_2$ where $c_0$ corresponds to $\sin$:

Let $c_0(x) = x f(\frac{x}{\pi})$. Then $c_0( \omega x ) = \omega x f(\frac{\omega x}{\pi}) = \omega x f(\frac{x}{\pi}) = \omega c_0(x)$ and similariliy $c_0(\omega^2x) = \omega^2 c_0(x)$.

From this it follows that $$c_0(x) + c_0(\omega x) + c_0(\omega^2 x ) = (1+\omega + \omega^2) c_0(x) = 0$$.

We have: \begin{align} c_0'(x) &= f(\frac{x}{\pi}) + \frac{x}{\pi}f'(\frac{x}{\pi})\\ &= \sum_{k=0}^\infty \frac{x^{3k}}{\pi^{3k}} \zeta_k(3) + \frac{x}{\pi}\sum_{k=0}^\infty \frac{3k x^{3k-1}}{\pi^{3k-1}} \zeta_k(3)\\ &= \sum_{k=0}^\infty \frac{(3k+1)x^{3k}}{\pi^{3k}} \zeta_k(3) \\ &= \sum_{k=0}^\infty \frac{(3k+1)\zeta_k(3)}{\pi^{3k}} x^{3k} \end{align}

Let

$$c_1(x) := c_0(\frac{\pi}{3}+\omega x), c_2(x) := c_1(\frac{\pi}{3}+\omega x) = c_0 ( \frac{\pi}{3}+\omega \frac{\pi}{3} + \omega^2 x)$$

then we also have:

$$c_2(\frac{\pi}{3}+\omega x) = c_0(x)$$

Question:

  1. Do the $c_0,c_1,c_2$ satisfy an "addition formula" similar to $\sin,\cos$?
  2. Is there a relationship, similar to $\sin,\cos$, between the derivatives of these functions?
  3. Is there a geometric interpretation of these functions?

Edit: I have plotted $c_0(x)$ ($c_1(x)$ and $c_2(x)$ look similar) and it shows a three-fold symmetry, like there are three waves which meet at a threefold star and do a physical refraction:

threefold_symmetry_waves

Second edit: As a number theoretic application one can deduce some equalities concering Aperys constant $\zeta(3)$:

$$\zeta(3) = \frac{81}{\pi} c_2(\frac{2 \pi}{3}) \omega -27 - \sum_{n=1}^\infty \frac{\zeta_{n+1}(3)}{27^n}$$

Also:

$$\zeta_k(3) = \frac{c_2^{(3k+1)}(\frac{\pi}{3})\omega \pi^{3k}}{(3k+1)!}, k=0,1,2,\cdots$$

especially we get:

$$\zeta(3) = \frac{c_2^{(4)}(\frac{\pi}{3})\omega \pi^{3}}{24}$$

$\endgroup$
4
  • $\begingroup$ Also asked here: physics.stackexchange.com/questions/683999/… $\endgroup$ Commented Dec 17, 2021 at 18:44
  • 1
    $\begingroup$ A meta-reason why the answer to these questions would be negative is that a positive answer to any of these questions would lead to a new identity or interpretation of the Gamma function, one of the most intensively studied special functions in mathematics, and this would be very unlikely to be discovered "by accident". Another reason to doubt any nice new identities is that (as mentioned in my answer to the linked question) the set of zeroes/poles of these functions obey far fewer symmetries (or near-symmetries) than those of the trigonometric or pure Gamma functions. $\endgroup$
    – Terry Tao
    Commented Dec 18, 2021 at 16:53
  • $\begingroup$ @TerryTao: I had tried an attack of this sort on the Apery constant a few years earlier without success, but still the mathematics was kind of fun: mathoverflow.net/questions/227161/… . I guess that not all properties of $\sin,\cos$ can be generalised, but a few of them will, which depends on the group $C_3$ acting on the roots in contrast to the group $C_2$ which acts on the $sinc$ roots. To reflect what I mean by group action on the roots, I have framed a question, which might be the reason why the "near-symmetries" are there $\endgroup$ Commented Dec 18, 2021 at 17:27
  • $\begingroup$ @TerryTao: In my previous comment, I meant to refer to this question but the comment was too long: math.stackexchange.com/questions/2779590/… $\endgroup$ Commented Dec 18, 2021 at 17:28

1 Answer 1

11
$\begingroup$

There is no addition formula: functions satisfying an algebraic addition formula have been completely characterized,

Painlevé, P. Sur les fonctions qui admettent un théorème d’addition, Acta Math. 27, 1-54 (1903).

P. J. Myrberg, Über Systeme analytischer Funktionen, welche ein Additionstheorem besitzen. Preisschriften gekrönt und herausgegebenen von der Fürstlich Jablonowskischen Gesellschaft zu Leipzig, Teubner, Leipzig, 1922.

Abe, Yukitaka, A statement of Weierstrass on meromorphic functions which admit an algebraic addition theorem, J. Math. Soc. Japan 57 (2005), no. 3, 709–723.

$\endgroup$
5
  • $\begingroup$ (+1) Thanks for the pointer. I did not know this result. $\endgroup$ Commented Dec 18, 2021 at 6:31
  • 3
    $\begingroup$ There are plenty of results called addition formulas that are not covered by this classification, for instance Koornwinder's addition formula for Jacobi polynomials. $\endgroup$ Commented Dec 18, 2021 at 8:16
  • $\begingroup$ @Hjagmar Rosengren 2: sure. But I meant an addition theorem similar to that of sin, cos, namely $f(x+y)=F(f(x),f(y))$, where f is a vector function. $\endgroup$ Commented Dec 18, 2021 at 11:16
  • $\begingroup$ @AlexandreEremenko: What is $F$? I have some functions similar to $sinh,cosh$ which I can show satisfiy an addition theorem: mathoverflow.net/questions/227161/… $\endgroup$ Commented Dec 18, 2021 at 13:04
  • 1
    $\begingroup$ @stackExchangeUser: F is a rational, or more generally, an algebraic function. $\endgroup$ Commented Dec 19, 2021 at 0:43

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .