Simultaneous sign change and equidistribution of signs of Fourier coefficients of two cusp forms (Q1660096): Difference between revisions

Let \(f_1:=\sum_{n=1}^{\infty} \lambda_1 (n) n^{\frac{k_1-1}{2}} q^n\) and \(f_1:=\sum_{n=1}^{\infty} \lambda_2 (n) n^{\frac{k_2-1}{2}} q^n\) in \(S_{k_1}^{\text{new}}(N_1)\) and \(S_{k_2}^{\text{new}}(N_2)\), respectively. For each \(i=1,2\) and prime \(p\) not dividing \(N_1 N_2\) put \(\lambda_i(p)=2 \cos \theta_{i}(p)\) with \(\theta_i \in [0, \pi] \) and assume that \(1\), \(\frac{\theta_1(p)}{2\pi}\), \(\frac{\theta_2(p)}{2\pi}\) are linearly independent over \(\mathbb{Q}\). Then, in the paper under review, the author shows \[ \lim_{x \to \infty} \# \dfrac{ \{ \nu \leq x : \lambda_1 (p^ \nu) \lambda_2 (p^ \nu) \lessgtr 0 \} }{x} = \dfrac{1}{2}. \] This calculates the proportion of positive integers \(\nu\) for which \(\lambda_1(p^\nu)\) and \(\lambda_2(p^\nu)\) have the same sign. The proof is based on the \(2\)-dimensional variant of Weyl's equidistribution theorem. Furthermore, the author studies the sign changes in the sequence \(\{\lambda_1(p^\nu)\lambda_2(p^\nu) \}\) as \(p\) varies over the prime numbers not dividing \(N_1N_2\) and he translates the result from integral weight Hecke eigenforms to half-integral weight Hecke eigenforms via the Shimura correspondence. The last proofs are based on pair-Sato-Tate conjecture.