[Paper Review] $n$-correlation with restricted support

J. Brian Conrey, N. C. Snaith|arXiv (Cornell University)|Dec 21, 2012

Analytic Number Theory Research3 citations

TL;DR

This paper provides a new proof that the $n$-correlation of zeros of $L$-functions matches random matrix theory predictions by leveraging ratios of characteristic polynomials in unitary matrices. By analyzing the structure of these ratios under support restrictions, the authors identify surviving terms that exactly match Rudnick and Sarnak’s result, demonstrating consistency between number theory and random matrix theory via a method that generalizes easily to $L$-functions.

ABSTRACT

We give a new proof that Rudnick and Sarnak's calculation of the $n$-correlation of zeros of $L$-functions matches random matrix theory. We do this using the derivation of $n$-correlation for eigenvalues of random unitary matrices from our previous paper. There we detailed a method using ratios of characteristic polynomials that is far less elegant than standard random matrix techniques, but which has the advantage of translating easily into the number theory case by applying the Ratios Conjectures for ratios of $L$-functions. The formulae in our previous paper are unwieldy in comparison with the standard determinantal random matrix version, but we show here that their form allows immediate identification of which terms remain when restrictions are placed on the support of the test function. After this it is straightforward to show that the surviving terms match Rudnick and Sarnak's result.

Motivation & Objective

• To provide an alternative proof that the $n$-correlation of $L$-function zeros aligns with random matrix theory predictions.
• To demonstrate how the ratios of characteristic polynomials method, though less elegant, enables straightforward analysis under restricted support conditions.
• To identify which terms persist in the $n$-correlation formula when the test function has compactly supported Fourier transform.
• To bridge number theory and random matrix theory by showing that the surviving terms after restriction match Rudnick and Sarnak’s result exactly.

Proposed method

• Utilize the ratios of characteristic polynomials approach from prior work to compute $n$-correlation for eigenvalues of random unitary matrices.
• Apply the same formalism to $L$-functions by invoking the Ratios Conjectures for ratios of $L$-functions.
• Analyze the algebraic structure of the resulting formulae to isolate terms that survive when the test function has restricted support.
• Identify the specific terms in the $n$-correlation expression that remain under support constraints by examining the functional form of the ratios.
• Show that the surviving terms in the number theory setting exactly match the known result from Rudnick and Sarnak’s calculation.
• Leverage the unwieldy but systematic form of the ratios-based formula to simplify the analysis under restrictions, unlike standard random matrix techniques.

Experimental results

Research questions

• RQ1 How can the $n$-correlation of $L$-function zeros be re-derived using a method that generalizes to number theory?
• RQ2 Which terms in the $n$-correlation formula persist when the test function has restricted support?
• RQ3 Can the ratios of characteristic polynomials method yield the same result as standard random matrix theory under support constraints?
• RQ4 Does the structure of the ratios-based formula allow for clear identification of surviving contributions under support restrictions?
• RQ5 Is there a direct match between the surviving terms in the $L$-function case and Rudnick and Sarnak’s result when support is restricted?

Key findings

• The $n$-correlation of $L$-function zeros matches the prediction from random matrix theory when the test function has restricted support.
• The ratios of characteristic polynomials method, while algebraically complex, cleanly isolates the terms that survive under support restrictions.
• The surviving terms in the $n$-correlation expression after restriction are identical to those derived by Rudnick and Sarnak using standard techniques.
• The method enables a direct translation from random matrix theory to number theory via the Ratios Conjectures, preserving structural clarity under constraints.
• The form of the ratios-based formula allows immediate identification of relevant contributions, unlike more elegant but less transparent standard random matrix approaches.
• The analysis confirms the robustness of the $n$-correlation prediction across different mathematical frameworks and support conditions.