Linear independence of Eisenstein Space Basis

160 Views Asked by At

I'm currently working throught Diamond and Shurman's book "A First Course on Modular Forms". On page 129 Theorem 4.5.2 states

Let $ N $ be a positive integer and let $ k \geq 3 $. The set $$ \left\{E_{k}^{\psi,\varphi,t}\ |\ (\psi,\varphi,t) \in A_{N,k} \right\} $$ is a basis of $ \mathcal{E}_{k}(\Gamma_{1}(N)) $,

where $ A_{N,k} $ is defined as the set of triples $ (\psi,\varphi,t) $ such that $ \psi $ and $ \varphi $ are primitive Dirichlet characters modulo $ u $ and $ v $ respectively such that $\varphi(-1) \psi(-1) = (-1)^k$, and $ t $ a positive integer such that $ tuv \mid N $. We are given that $ \textit{dim}(\mathcal{E}_{k}(\Gamma_{1}(N)) = |A_{N,k}|$, so all that remains is to prove linear independence. We have that $$ E_{k}^{\psi,\varphi,t} = \delta(\psi)L(1-k,\varphi) + 2\sum_{n=1}^{\infty}\left(\sum_{\substack{m \mid n \\ m > 0}} \psi(n/m)\varphi(m)m^{k-1}\right)q^{nt} $$ where $$ \delta(\psi) = \left\{\begin{array}{c} 1 \textit{ if } \psi \textit{ is trivial }\\ 0 \textit{ otherwise } \end{array} \right\} $$ and $ q = e^{2\pi iz} $. So I have started by observing that the first term is nice i.e $$ \delta(\psi)L(1-k,\varphi) + 2q^{t} .$$ I want to say that no further values of $ n $ cancel out the $ n = 1 $ value, and so just prove the linear independence of the $ n = 1 $ term. However I am unable to establish whether this is the case. The book skips the proof and I have been unable to find anything on the internet. If anybody is able to prove the linear independence or knows of a proof then I would be most grateful.

1

There are 1 best solutions below

1
On BEST ANSWER

Following up on the hint in @cfairwea's comment, it's not too hard to check that if $(\varphi', \psi') \ne (\varphi, \psi)$, then for any $t$ and $t'$ the series $E_k^{\varphi, \psi, t}$ and $E_k^{\varphi', \psi', t'}$ have different eigenvalues for the Hecke operators away from $N$, and hence there can't be any nontrivial linear dependence relation between them. So it suffices to show that the $E_k^{\varphi, \psi, t}$ for a fixed pair $(\varphi, \psi)$ and different values of $t$ are linearly independent; and this is not so difficult (you've already made a good start on it).