Table of Contents
Fetching ...

Fake Mobius-type Functions on the Unit Circle

Ali Saraeb

TL;DR

The paper extends the fake Möbius framework by allowing complex unit-circle prime-power data and derives a universal zeta-factorization F_f(s)=ζ(s)^z ζ(2s)^w G_f(s). It develops a rigorous analytic apparatus—complex-power bounds for ζ(s), contour shifts with Hankel detours, and Perron-type smoothing—to obtain explicit Laplace-representations and Watson-type expansions for the smoothed summatory function A_f^{exp}(x). A central contribution is a precise bias analysis at the critical x^{1/2} (log x)^{w-1} scale, distinguishing persistent, apparent, and no-bias regimes under the RH/SZC framework and mild zero-spacing assumptions. The work clarifies how the leading singular data (z,w) governs the bias, while higher-order ε_k affect only lower-order terms, and sets the stage for future exploration of complementary parameter regions.

Abstract

We study \emph{fake Mobius-type functions on the unit circle}: multiplicative functions $\mathfrak f$ determined by prime-power data $\mathfrak f(p^k)=\varepsilon_k\in\mathbb S^1\cup\{0\}$, independent of $p$. Their Dirichlet series admits the Euler product \[ F_{\mathfrak f}(s)=\sum_{n\ge1}\frac{\mathfrak f(n)}{n^s} =\prod_p g(p^{-s}),\qquad g(u)=\sum_{k\ge0}\varepsilon_k u^k, \] and a canonical zeta-factorization \[ F_{\mathfrak f}(s)=ζ(s)^{\,z}\,ζ(2s)^{\,w}\,G_{\mathfrak f}(s), \qquad z=\varepsilon_1,\ \ w=\varepsilon_2-\frac{\varepsilon_1(\varepsilon_1+1)}{2}, \] where $G_{\mathfrak f}(s)$ is a holomorphic Euler product on $\Re s>1/3$. Assuming the Riemann hypothesis and simplicity of zeros, we derive a general explicit and asymptotic formula for the summatory functions $A_{\mathfrak f}^{\exp}(x)$ of the form \[ A_{\mathfrak f}^{\exp}(x) -Δ_1(x;z,w) = Δ_{1/2}(x;z,w)\;+\;\sum_ρΔ_ρ(x;z,w,\mathfrak f)\;+\;\mathcal E(x), \] where $Δ_1(x;z,w)$ is the main term coming from the $s=1$ singularity of $ζ(s)^z$, $Δ_{1/2}(x;z,w)$ is the secondary term coming from the $s=1/2$ singularity of $ζ(2s)^w$, the terms $Δ_ρ(x;z,w,\mathfrak f)$ arise from the singularities of $ζ(s)^z$ at nontrivial zeros $ρ$ of $ζ(s)$, and $\mathcal E(x)$ is an error term. Furthermore, we introduce a notion of \emph{bias} at the natural scale $x^{1/2}(\Log x)^{w-1}$ and obtain an explicit criteria distinguishing persistent, apparent, and unbiased behavior in this regime. The present paper treats the major parameter region \[ -1\le\Re(z)\le1,\qquad -2\le\Re(w)<1, \] and we choose to defer the complementary ranges of $(z,w)$ to a future paper as a different analytic approach is required there.

Fake Mobius-type Functions on the Unit Circle

TL;DR

The paper extends the fake Möbius framework by allowing complex unit-circle prime-power data and derives a universal zeta-factorization F_f(s)=ζ(s)^z ζ(2s)^w G_f(s). It develops a rigorous analytic apparatus—complex-power bounds for ζ(s), contour shifts with Hankel detours, and Perron-type smoothing—to obtain explicit Laplace-representations and Watson-type expansions for the smoothed summatory function A_f^{exp}(x). A central contribution is a precise bias analysis at the critical x^{1/2} (log x)^{w-1} scale, distinguishing persistent, apparent, and no-bias regimes under the RH/SZC framework and mild zero-spacing assumptions. The work clarifies how the leading singular data (z,w) governs the bias, while higher-order ε_k affect only lower-order terms, and sets the stage for future exploration of complementary parameter regions.

Abstract

We study \emph{fake Mobius-type functions on the unit circle}: multiplicative functions determined by prime-power data , independent of . Their Dirichlet series admits the Euler product and a canonical zeta-factorization where is a holomorphic Euler product on . Assuming the Riemann hypothesis and simplicity of zeros, we derive a general explicit and asymptotic formula for the summatory functions of the form where is the main term coming from the singularity of , is the secondary term coming from the singularity of , the terms arise from the singularities of at nontrivial zeros of , and is an error term. Furthermore, we introduce a notion of \emph{bias} at the natural scale and obtain an explicit criteria distinguishing persistent, apparent, and unbiased behavior in this regime. The present paper treats the major parameter region and we choose to defer the complementary ranges of to a future paper as a different analytic approach is required there.

Paper Structure

This paper contains 26 sections, 23 theorems, 273 equations, 2 figures.

Key Result

Proposition 2.8

For $s\in\Omega$, Proof. The identity holds on $\{\Re s>1\}$ and the right hand side is holomorphic on $\Omega$, so extend by the identity theorem. ∎

Figures (2)

  • Figure 1: Figure $1.3.4:$ The region of the $(\Re(z),\Re(w))$ parameter space covered in this paper.
  • Figure 2: Figure 3.2.1: Contour of integration $\Gamma_T$

Theorems & Definitions (88)

  • Definition 2.1: Generalized binomial coefficient
  • Definition 2.2
  • Definition 2.3
  • Definition 2.4: Definition of $\log \zeta(s)$ on $\Omega$
  • Definition 2.5: Logarithm of $(s-1)\zeta(s)$ and the function $Z(s;z)$
  • Remark 2.6
  • Definition 2.7
  • Proposition 2.8: Analytic continuation
  • Remark 2.9
  • Remark 2.10
  • ...and 78 more