Abstract
We establish a result of Bombieri–Vinogradov type for the Dirichlet coefficients at prime ideals of the standard $L$-function associated to a self-dual cuspidal automorphic representation $\pi $ of $\operatorname {GL}_n$ over a number field $F$ when $\pi$ is not a quadratic twist of itself. Our result does not rely on any unproven progress towards the generalized Ramanujan conjecture or the nonexistence of Landau–Siegel zeros. In particular, when $\pi $ is fixed and not equal to a quadratic twist of itself, we prove the first unconditional Siegel-type lower bound for the twisted $L$-values $|L(1,\pi \otimes \chi )|$ in the $\chi $-aspect, where $\chi $ is a primitive quadratic Hecke character over $F$. Our result improves the levels of distribution in other works that relied on these unproven hypotheses. As applications, when $n=2,3,4$, we prove a $\textrm {GL}_n$ analogue of the Titchmarsh divisor problem and a nontrivial bound for a certain $\textrm {GL}_n\times \textrm {GL}_2$ shifted convolution sum.
1. Introduction
In analytic number theory, the Bombieri–Vinogradov (B-V) theorem is often described as a "poor man’s Riemann Hypothesis." While the Generalized Riemann Hypothesis (GRH) provides precise predictions for the distribution of primes in arithmetic progressions, it remains one of the most significant unproven conjectures. The B-V theorem offers a result nearly as potent as GRH by showing that the expected distribution holds "on average" across different moduli.
Scaling this theorem from the integers ($\mathbb{Q}$) to higher-rank groups (such as $\operatorname{GL}_n$) and general number fields is a formidable challenge. Historically, such extensions have been hindered by two major obstacles: the Generalized Ramanujan Conjecture (GRC) and the potential existence of Landau–Siegel Zeros. This paper presents a breakthrough by establishing a version of the B-V theorem for $\operatorname{GL}_n$ that is entirely unconditional.
2. Methodology and Proof Structure
The authors utilize a sophisticated combination of analytic techniques to bypass traditional dependencies on unproven hypotheses. The proof is structured to handle various types of sums and error terms systematically, moving from preliminary reductions to specific sieve estimates.
Key methodological components include:
- Generalized Vaughan Identity: A combinatorial sieve method adapted for $L$-function coefficients associated with $\operatorname{GL}_n$, used to decompose sums over primes into more manageable "Type I" and "Type II" sums.
- Zero-Free Regions: To address the Siegel zero problem, the authors establish a new unconditional lower bound for the twisted $L$-function at $s=1$. This allows for a "Siegel–Walfisz" type result without assuming the nonexistence of exceptional zeros.
- Large Sieve Estimates: Utilizing inequalities for number fields (originally developed by Huxley) to bound error terms arising from the Vaughan identity.
- Dual Pieri Rule: A tool from representation theory and combinatorial lemmas used to bound certain divisor functions without assuming GRC.
3. Key Insights
3.1 Breaking the Dependency on GRC
Most previous work relied on GRC to bound the magnitude of the mathematical objects being summed. By leveraging Rankin–Selberg theory and combinatorial identities, the authors demonstrate that sufficient cancellation can be achieved in the sums to prove the theorem without knowing if GRC is true.
3.2 The Siegel Zero Workaround
A significant technical achievement of this work is Theorem 1.2, which provides an unconditional lower bound for $L$-functions in the $\chi$-aspect. This addresses a long-standing technical gap and is a critical component in making the B-V theorem unconditional for higher-rank groups.
3.3 Level of Distribution
The paper achieves a "level of distribution"—a measure of the maximum size of the moduli—of $x^{1/\eta}$, where $\eta = \max\{2, n/2\}$. Notably, for $\operatorname{GL}_2$, this matches the classical "square root" barrier of $1/2$.
4. Results and Applications
The primary result, Theorem 1.1, establishes the Bombieri–Vinogradov theorem for self-dual cuspidal automorphic representations of $\operatorname{GL}_n$ over a number field $F$. Beyond the main theorem, the authors provide several significant applications:
- Titchmarsh Divisor Problem: They prove a $\operatorname{GL}_n$ version for $n=2, 3, 4$, bounding the sum $\sum_{p \le x} \lambda_\pi(p) d(p-1)$ non-trivially, where $d(n)$ is the divisor function.
- Shifted Convolution Sums: The work provides non-trivial bounds for $\operatorname{GL}_n \times \operatorname{GL}_2$ sums, which are critical in the study of quantum unique ergodicity and other areas of mathematical physics.
5. Limitations and Future Outlook
While groundbreaking, the results have specific constraints. The main theorem requires the representation $\pi$ to be self-dual and not a quadratic twist of itself. Furthermore, the constants in the lower bounds are "ineffective," meaning the proof guarantees their existence but does not provide a method for calculation. Future research may focus on extending these results to non-self-dual cases and improving the level of distribution for higher ranks ($n > 4$) toward the Elliott-Halberstam range.
Summary updated