The Riemann $Ξ$-function from primitive Markovian cycles II: Strip rigidity and divisor identification

Technical Summary: The Riemann $\Xi$-Function from Primitive Markovian Cycles II: Strip Rigidity and Divisor Identification

Problem Statement
This paper addresses a fundamental comparison problem in analytic number theory: establishing the equivalence between two distinct analytic constructions derived from a single primitive dynamical framework. Specifically, it seeks to compare the classical Riemann $\Xi$-function, arising from the Archimedean completion of a scaling-limit trace kernel, against a canonical family of real-entire spectral determinants ($P_N$) derived from the cycle Laplacian of finite reversible Markov chains. The central challenge is to demonstrate that these two objects share the same zero divisor (i.e., the same zeros) on horizontal strips in the complex plane, up to a rigid, zero-free normalization. This comparison is conditional on specific holomorphy and boundary non-vanishing hypotheses, which are posited to be verified in a forthcoming third paper in the trilogy.

Methodology
The paper employs a "strip rigidity" approach, utilizing complex analysis on horizontal strips to link the Laplace-side spectral factorization with the Mellin-side zeta function. The methodology proceeds through three primary stages:

1. Primitive Dynamical Origin: The analysis begins with the scaling limit of finite, reversible, nearest-neighbor Markov dynamics on discrete cycles ($\mathbb{Z}/N\mathbb{Z}$). This yields a renormalized trace kernel with an explicit theta-series representation. Crucially, this construction is purely probabilistic and analytic at this stage, independent of number-theoretic inputs like functional equations or Mellin transforms.
2. Dual Analytic Outputs: From this single kernel, two distinct analytic objects are generated:
   • Laplace Side: A bilateral Laplace transform $B_{\Phi^*}(s)$ of a centered kernel, which admits a Schoenberg–Edrei–Karlin factorization into a zero-free entire function $E(s)$ and a real-entire function $\Psi(s)$ (the reference family $P_N$) with only real zeros.
   • Mellin Side: An Archimedean-completed kernel $\tilde{K}_{\text{arch}}$, whose Mellin transform is identified with the classical Riemann $\Xi$-function ($\Xi(2z)$).
3. Strip Comparison and Rigidity: The core mechanism involves comparing these two functions on overlap strips.
   • Left-Strip Continuation: The paper extends the domain of the Laplace transform $B_{\Phi^*}$ to a left strip ($-1/2 < \Re(s) < 1/4$) using modular splitting and the exact twisted symmetry of the centered kernel.
   • Boundary Identity: A precise identity is established on the boundary line $\Re(s) = -1/2$, linking the continued Laplace transform directly to the Mellin transform of the Archimedean kernel (and thus to $\Xi$).
   • Seam Ratio Analysis: A normalized "seam ratio" $R(w)$ is defined, linking $\Xi(2w)$ to the reference polynomial $P_N(w)$. Under Hypothesis 1.1 (spectral coupling/boundary separation on rectangles) and Hypothesis 1.2 (quantitative sector control), a rigidity lemma is applied. This lemma, rooted in Rouché's theorem and Nevanlinna/Hardy space theory, forces the seam ratio to be a "strip-unit" (a holomorphic, zero-free function of bounded type).

Key Contributions and Results

• Theorem 1.6 (Main Theorem Overview): The paper establishes that, under standing assumptions, the scaling-limit trace kernel yields both a Schoenberg–Edrei–Karlin factorization with a real-zero divisor and an Archimedean completion identified with $\Xi(2z)$. Crucially, it provides the analytic machinery (continuation and boundary identities) to place these two objects in a common strip setting.
• Theorem 1.7 (Conditional Strip-Unit Theorem): This is the paper's primary result. It states that if Hypotheses 1.1 and 1.2 hold, the normalized seam ratio $\tilde{R}_N(w)$ is holomorphic and zero-free on the strip $S_\eta$. Consequently, $\Xi(2w)$ and the reference family $P_N(w)$ share the exact same zero divisor on that strip.
• Analytic Rigidity: The paper introduces a "strip-unit forcing" mechanism. By demonstrating that the difference between the two functions is small relative to the reference function on the boundary of expanding rectangles (Hypothesis 1.1), and that the argument of the ratio is controlled (Hypothesis 1.2), the paper proves that the functions must have identical zero counts in the interior.
• Logical Separation: The work rigorously separates the probabilistic construction of the kernel from the number-theoretic identification. The derivation of the real-entire factorization does not rely on the properties of $\Xi$, and the identification of the Mellin transform does not rely on the spectral factorization, ensuring the argument is non-circular.

Significance and Claims
The paper claims to provide the "missing divisor-identification step" from the preceding work (Paper I). Its significance lies in:

1. Unifying Framework: It demonstrates that the Riemann $\Xi$-function and a spectral determinant from a simple Markov process are analytically linked via a rigid normalization.
2. Divisor Identification: It proves that, conditionally on the verification of boundary estimates, the non-trivial zeros of the Riemann zeta function (encoded in $\Xi$) must coincide with the real zeros of the cycle spectral determinant $P_N$. Since $P_N$ has only real zeros by self-adjointness, this suggests a pathway to proving the Riemann Hypothesis (that all non-trivial zeros lie on the critical line) by forcing the zeros of $\Xi$ to be real in the transformed variable.
3. Methodological Rigor: The paper emphasizes a "stress-tested" approach, explicitly verifying decay rates, symmetry conditions, and the absence of spurious poles to ensure the validity of the analytic interchanges.

Limitations and Dependencies
The paper explicitly states that its main conclusion (divisor identification) is conditional. It relies on:

• Hypothesis 1.1: The existence of a zero-free holomorphic unit $U_{\eta,N}$ that satisfies a strict boundary separation inequality on rectangles $R_T$.
• Hypothesis 1.2: Quantitative sector control on the argument of the normalized seam ratio on the boundary of these rectangles.
  The paper notes that the verification of these hypotheses, including the choice of the admissible regime for the cycle size $N(T)$, is deferred to Paper III. Without the verification of these boundary conditions, the divisor identification remains a conditional result.