Typo in Stein's online modular form textbook?

57 Views Asked by At

In Stein's online textbook here, in the section "Explicit Basis of the Eisenstein Subspace", he defines the Eisenstein series

$$E_{k,\chi,\psi}(q)=c_0+\sum_{n=1}^{\infty}\left(\sum_{d|n}\psi(d)\chi(n/d)d^{k-1}\right)q^n$$

It is then stated that if $t$ is a positive integer and $k$ is a natural such that $\psi(-1)\chi(-1)=(-1)^k$ then other than in some exceptional cases $E_{k,\chi,\psi}\left(q^t\right)$ defines an element of $M_k(\Gamma(RLt))$ where $R$ and $L$ are the conductors of $\psi$ and $\chi$ respectively.

I am confused for a couple of reasons

  • If a function is a modular form in $\Gamma(N)$, then it is periodic mod $N$. This means that it should have a $q$ expansion where $q=e^{2\pi i z/N}$. For this reason, raising $q$ to the power of $t$ should decrease the period, not increase it, i.e $E_{k,\chi,\psi}(q^t)$ should be an element of $M_k(\Gamma(RL/t))$ and $E_{k,\chi,\psi}(q^{1/t})$ should be an element of $M_k(\Gamma(RLt))$. Not what is said.

  • Am I correct in assuming that when the modular form is being defined, $q=e^{2\pi i z/RL}$? This is what it should need to be for the periods to match up if I am not mistaken.

1

There are 1 best solutions below

0
On BEST ANSWER

Stein defines the Eisenstein subspace only for the group $\Gamma_1(N)$. There is indeed a typo, it should be $\Gamma_1(RLt)$ instead of $\Gamma(RLt)$. Moreover $q$ denotes $e^{2\pi iz}$ as usual. Also, in general, if $f(q)$ is a modular form on $\Gamma_1(N)$ then $f(q^t)$ is a modular form on $\Gamma_1(Nt)$. These things are also explained in the book A first course on modular forms by Diamond and Shurman.

Here is an example (pointed out by the OP). Take $k \geq 2$ even, $\chi$ is the unique character modulo $2$, and $\psi$ is the trivial character. Then as explained in Stein's textbook, $E_{k,\chi,\psi}$ belongs to $M_k(\Gamma_1(2))$.

What about $E_{k,\chi,\psi}(q^{1/2})$? One can show as an exercise that if $f(z)$ is a modular form on $\Gamma_1(N)$, then $f(q^{1/d})$ is a modular form on $\Gamma(N)$ for every $d$ dividing $N$ (here $q^{1/d} = e^{2\pi iz/d}$). Thus, in your case, $E_{k,\chi,\psi}(q^{1/2})$ is a modular form on $\Gamma(2)$.