1
$\begingroup$

The theta function $\theta_\chi(t)$ of a Dirichlet character $\chi$ is defined to be $\theta_\chi(t) = \frac{1}{2} \sum\limits_{n=-\infty}^\infty \chi(n) e^{2\pi i n^2 t}$ if $\chi(-1) = 1$ (i.e., $\chi$ is even), or $\frac{1}{2} \sum\limits_{n=-\infty}^\infty n\chi(n) e^{2\pi i n^2 t}$ if $\chi(-1) = -1$ (i.e., $\chi$ is odd).

I understand the definition when $\chi$ is odd can't be the same as the definition when $\chi$ is even because the formula would have vanished identically. Multiplying an additional factor of $n$ indeed could solve the problem since $n\chi(n)$ is once again even.

Is there a more intrinsic reason why the theta function for an odd Dirichlet character is twisted in this specific way? Theoretically speaking, one can twist by any odd function of $n$. Is there a deeper reason why this form of twist is preferred over others?

$\endgroup$

1 Answer 1

3
$\begingroup$

The point is that $\theta_\chi(t)$ is a modular form whose coefficients are essentially $\chi(n)$. If we twist $\chi(n)$ by some random odd smooth function of $n$, we don't get a modular form. (Perhaps there are some examples other than twisting by $n$, but they will be artificial. That is, in a sense, $\theta_\chi(t)$ is the only natural example obtainable this way.)

$\endgroup$
3
  • 1
    $\begingroup$ Is there a precise mathematical statement you have in mind with your parenthetical remark? $\endgroup$ Commented May 29, 2023 at 11:30
  • 2
    $\begingroup$ @AnuragSahay: For example, twisting with an odd polynomial of $n$ results in a linear combination of derivatives of $\theta_\chi(t)$. Twisting with a rapidly convergent odd power series of $n$ results in an infinite linear combination of derivatives of $\theta_\chi(t)$. $\endgroup$
    – GH from MO
    Commented May 29, 2023 at 13:51
  • 1
    $\begingroup$ ah, fair enough. Thanks for the clarification! $\endgroup$ Commented May 30, 2023 at 14:09

You must log in to answer this question.

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