Let $f=\sum_{n\ge1} a_nq^n$ be an Hecke eigenform cusp form of integral weight $k$ on $\Gamma_0(N)$ with character $\varepsilon\pmod N,$ and let $M_f$ be the subfield of $\mathbb{C}$ generated by the image of $\varepsilon$ and all $a_p$ ($p$ runs on primes) $$M_f:=\mathbb{Q}(\{a_p\}_p,\varepsilon)$$ On what condition we have $\dfrac{a_p}{p^{(k-2)/2}}\in M_f$ ?

  • $\begingroup$ If $k$ is even, $p^{(k-2)/2} \in \mathbb Q$, so the answer is "always" in this case. So basically you're interested in the case of odd weight with necessarily non-trivial $\epsilon$. Is it how you intended the question? Or perhaps did you mean it with $(k-2)$ replaced by $(k-1)$? $\endgroup$ – Joël Feb 19 '17 at 14:12
  • $\begingroup$ @Joël, I am interested in the case when $k$ is odd, and $\varepsilon$ non-trivial. Thanks for pointing this out. $\endgroup$ – User 101794987 Feb 19 '17 at 14:17

When $k$ is even the answer is always since $p^{(k-2)/2} \in \mathbb Q$. When $k$ is odd the answer is never since if this was true, then for all $p$ (prime not divdidig $N$), we would have $p^{1/2} \in M_f$, so $M_f$ would contain the field of infinite degree over $\mathbb Q$ generated by the square roots of those prime, contradicting the fact that $M_f$ is a finite extension of $\mathbb Q$.

| cite | improve this answer | |
  • 1
    $\begingroup$ Tiny nitpick: this doesn't quite work as stated for all $p \nmid N$, but just for those where $a_p \ne 0$; this set can be somewhat thinner than all $p\nmid N$ (CM forms) but of course it is still infinite, which is all you need for the conclusion. $\endgroup$ – David Loeffler Feb 20 '17 at 9:51
  • $\begingroup$ @DavidLoeffler, Thank you very much for pointing this out. $\endgroup$ – User 101794987 Feb 21 '17 at 0:42

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.