MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top


let $F$ and $G$ be two primitive functions of the Selberg class, and let $\mathbb{A}$ be the set of values taken by the function which maps a prime number $p$ to $a_{p}(F)\overline{a_{p}(G)}$. $\mathbb{A}$ is finite or countable, so that there exists $I\subset\mathbb{N}^*$ such that $\displaystyle{\mathbb{A}=\bigcup_{i\in I}\{A_{i}\}}$, with $A_i\neq A_j$ whenever $i\neq j$. Now, let $\delta$ be the function that maps $i\in I$ to $\displaystyle{\lim_{x\to\infty}\frac{#\{p\leq x, a_{p}(F)\overline{a_{p}(G)}=A_i\}}{\pi(x)}}$.

Does Selberg's orthonormality conjecture imply $\displaystyle{\sum_{i}\delta(i)A_i=\delta_{F,G}}$, where $\delta_{F,G}=1$ if $F=G$ and $\delta_{F,G}=0$ otherwise? If so, is such an implication in fact an equivalence? Thank you in advance.

share|cite|improve this question
up vote 3 down vote accepted

No. (This is very unlikely to hold for anything except $L$-functions with only finitely many distinct coefficients; the simplest counter-example is to take an elliptic curve $E$ and its associated $L$-function $L(E,s)$, normalized to have critical line at $1/2$, and $F=G=L(E,s)$; for any non-zero real number $x>0$, due to the renormalization -- which by the way makes the question rather strange --, any prime $p$ for which $|a(p)|^2=x$, the Hasse-Weil coefficient $a_p$ of the elliptic curve must satisfy $a_p^2=px$; in particular $\sqrt{px}$ is an integer, and there are rather obviously only finitely many $p$ which can satisfy this for a given $x$. In other words, your function $\delta$ is identically zero, despite the fact that $\delta_{F,F}=1$.)

One can also of course just take the $L$-function of the Ramanujan delta function, replacing just squareroots with powers $11/2$ in the right places.

share|cite|improve this answer
Indeed I thought about this question considering the case of Dirichlet L-functions and wondering whether it would hold for the whole Selberg class or not. Thank you much for your answer. – Sylvain JULIEN Mar 27 '11 at 18:53

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.