Baker-Montgomery Conjecture on L-Function Zeros
- Baker-Montgomery Conjecture is a hypothesis on the asymptotic behavior of real zeros of L'(s,χ₋d), predicting their increase at a rate proportional to log log |d|.
- The conjecture connects analytic number theory with random matrix theory by employing Selberg's short Dirichlet polynomial approximations and probabilistic models to estimate zero distributions.
- Recent work by Lamzouri and Nath has nearly achieved an unconditional resolution up to a polylogarithmic factor, highlighting both upper and lower bounds for real zeros in the critical interval.
The Baker-Montgomery conjecture concerns the asymptotic behavior of the number of real zeros of the derivative of quadratic Dirichlet -functions, specifically , within the critical interval . Given its focus on fundamental discriminants and the distribution of zeros near the critical line, the conjecture is a central problem in the analytic study of -functions and their derivatives, linking zero distribution phenomena to random matrix theory and sieving methods. Recent advances, notably by Lamzouri and Nath, have nearly resolved the conjecture unconditionally up to a polylogarithmic factor, with a full resolution conditional on well-motivated analytic hypotheses.
1. Formal Statement of the Baker–Montgomery Conjecture
Let be a fundamental discriminant and the associated primitive quadratic Dirichlet character. Define as the number of real zeros of in the interval . The conjecture posits:
For almost all fundamental discriminants ,
for some positive constant ; in particular, .
This predicts that the number of real zeros of in for almost all grows like the double logarithm of .
2. Definitions and Foundational Objects
- Quadratic Dirichlet Character : The primitive character modulo defined via the Kronecker symbol .
- Fundamental Discriminant : The discriminant of the quadratic field , with and squarefree, or , and squarefree.
- Dirichlet -function:
Analytically continued to ; its derivative is .
3. Unconditional Results: Upper and Lower Bounds
Upper Bound (Lamzouri–Nath):
Let as , and fundamental discriminants in in a fixed progression, . Theorem 1.1 asserts:
for almost all . Thus, outside a negligible exceptional set, has at most real zeros in .
Proof strategy summary:
- Approximate by short Dirichlet polynomials (Selberg approximation).
- Develop a probabilistic random model by replacing with i.i.d. random variables.
- Compare moments via discrepancy bounds.
- Use a discretized net ; control uniformly.
- Apply Jensen’s formula over concentric circles to bound the count of real zeros.
Lower Bound (Klurman–Lamzouri–Munsch):
For almost all ,
Moreover, all such zeros lie in .
Combined Order of Magnitude (Corollary 1.2):
For almost all ,
for some , exhibiting the optimal shape up to a polylogarithmic factor.
4. Conditional Results and Hypotheses
Assuming a mild zero-repulsion hypothesis on the low-lying zeros of quadratic Dirichlet -functions—precisely, that for each there exists so that has no zeros in a disk centered at of suitable small radius (Hypothesis )—a full resolution of the conjecture is obtained up to the factor:
Theorem 1.3 (Conditional): Assuming Hypothesis for almost all ,
When GRH and the one-level density conjecture of Katz and Sarnak hold (established for support up to by Özlük and Snyder), Hypothesis is satisfied for almost all .
Key technical tools employed include:
- Orthogonality relations for (Lemma 2.1)
- Zero-density theorems in the critical strip (Lemma 2.7)
- Selberg's short Dirichlet polynomial bounds (Lemma 3.1)
- Discrepancy bounds versus random models (Proposition 4.2)
- Moment estimates for near and away from $1/2$ (Lemmas 4.3, 5.1, Proposition 5.2)
5. Error Terms, Polylogarithmic Factors, and Localization
Error Source and Estimates
- Exceptional set for Theorem 1.1: size , hence for slowly diverging , nearly all are included.
- For Theorem 1.3: the exceptional proportion is .
- The factor originates from:
- Multiple applications of Jensen’s formula on nets of points (with ), each incurring an additive term;
- Discrepancy error terms near the boundary of the critical interval, scaling like , with .
Support for "All Zeros to the Right":
Under Hypothesis , almost all real zeros of in satisfy
hence asymptotically all such zeros are confined to . Methods employ moment bounds on near , Jensen's formula on nested circles, hole-size control by the zero-free disk, and Cauchy-integral mean-value estimates.
6. Remaining Obstacles and Steps Toward a Full Unconditional Proof
Unconditional results leave open the “near-critical” interval , offering no control over zeros in this sliver, and have a slack from the predicted asymptotic. Removal of the last polylogarithmic factor and full control as would require:
- Stronger uniform discrepancy estimates for for ,
- Sharper understanding of large-value behavior of very close to the $1/2$-line,
- Improved zero-density estimates within the near-critical strip.
As of Lamzouri and Nath’s work, methods reach within a polylogarithmic factor of confirming the conjecture, with a full conditional resolution when natural hypotheses on the distribution of low-lying zeros are assumed.
7. Broader Context and Related Developments
The Baker–Montgomery conjecture is part of a broader program investigating the vertical and horizontal distribution of zeros of -functions and their derivatives. The moment-discrepancy methodology, as refined by Lamzouri and Nath, uses a random model to approximate -function behavior and integrates zero-density, orthogonality, and polynomial approximation techniques. The work interacts with the one-level density conjecture and Katz–Sarnak random matrix models, providing strong evidence for deep connections between analytic multiplicities, random matrix statistics, and arithmetic geometry.
A plausible implication is that refinements of random-model comparisons and explicit moment computations for Dirichlet -functions may generalize to higher-degree -functions or to derivatives of automorphic -functions, strengthening the bridge between analytic number theory and statistical mechanics models.
Selected works:
- R. C. Baker & H. L. Montgomery, “Oscillations of quadratic L-functions,” 1989.
- O. Klurman, Y. Lamzouri & M. Munsch, “Sign changes of short character sums and real zeros of Fekete polynomials,” (Klurman et al., 4 Mar 2024).
- J. B. Conrey & K. Soundararajan, “Real zeros of quadratic Dirichlet L-functions,” Invent. Math. 150 (2002).
- A. E. Özlük & C. Snyder, “On the distribution of the nontrivial zeros of quadratic L-functions close to the real axis,” Acta Arith. 91 (1999).