\documentclass[11pt,a4paper]{article} \usepackage[T1]{fontenc} \usepackage{lmodern,amsmath,amssymb,amsthm,booktabs,microtype} \usepackage[margin=27mm]{geometry} \usepackage[colorlinks=true,linkcolor=blue!50!black,citecolor=blue!50!black,urlcolor=blue!50!black]{hyperref} \usepackage{xcolor} \setlength{\emergencystretch}{2em} \newtheorem{theorem}{Theorem}[section] \newtheorem{proposition}[theorem]{Proposition} \newtheorem{lemma}[theorem]{Lemma} \newtheorem{corollary}[theorem]{Corollary} \theoremstyle{remark}\newtheorem{remark}[theorem]{Remark} \newcommand{\Q}{\mathbb Q}\newcommand{\Z}{\mathbb Z} \title{Arithmetic dependencies of the construction with 29 variables} \author{}\date{8 November 2026} \begin{document}\maketitle \begin{abstract} We identify the arithmetic results used by the finite-algebra construction with 29 total variables. Its four stated arithmetic inputs follow from the specified elliptic-group, height and pole theorems in the source manuscript and Sun's simultaneous prime-pattern theorem. In particular, the required avoidance at seven, compatibility with an arbitrary fixed congruence progression, and realization with zero preparation parameters require no stronger global existence assertion. The deductions are given explicitly. The deep arithmetic theorems themselves remain cited inputs. \end{abstract} \section{The statements being used} Fix \[ K=\Q(\sqrt2),\quad D=11-6\sqrt2,\quad E:y^2=x^3-Dx, \qquad P=\left(\frac{57-28\sqrt2}{8},\frac{-660+417\sqrt2}{32}\right). \] Heights are absolute logarithmic Weil heights, and rational valuations are normalized by $v_p(p)=1$. We count the distinguished rational input as one of the variables of a polynomial. The source manuscript used here is Version~9 of \cite{GlazerV9}, at commit \begin{center}\small\texttt{14e8f601fe0d9173d4a5f084aedc3e07066a7688}.\end{center} The relevant file is \texttt{manuscripts/quartic\_v9.tex}. Sun's theorem is Theorem~4.1 of the explicitly versioned preprint \cite{Sun2026}, arXiv:2610.20691v2. File hashes and theorem identifiers are recorded in the accompanying source certificate. These pins fix the statements whose implications are proved below. For the underlying restricted-converse argument we incorporate the Version~11 correction \cite{GlazerV11}: the Kato classes are defined on the full index set needed by the 1999 theorem, and their norms are identified with the retained cyclotomic class. This correction changes the proof supporting an inherited input, not the statement of the input used here. \begin{enumerate} \item The elliptic source gives $2E(K)=\Z P$ and $h(x([n]P))\ge n^2-6$ for $n\ne0$. The relevant source identifiers are \nolinkurl{subsec:v5-point-source} and \nolinkurl{eq:o8-indexheight}. \item Put \[ B_* = \prod_{p\le76176}p, \qquad H_{19}=2^{110}210^{1800}B_*^{19}+128. \] For $t=u/v$ in lowest terms, $v>0$, and $t\notin\{0,1,2401,-343/32\}$, let $M_5(t)$ be the product of primes $p>76176$ for which at least one of \[ v_p(v),\quad v_p(u),\quad v_p(u-v),\quad v_p(u-2401v),\quad v_p(32u+343v) \] is positive and odd. The source theorem \nolinkurl{thm:quartic-height19} gives $h(t)\le H_{19}M_5(t)^{19}$. \item The pole results are the source obstruction, simultaneous realization and zero-preparation lemmas, respectively: \begin{quote}\small \nolinkurl{hyp:o8-pole},\\ \nolinkurl{lem:original-seven-simultaneous-realization},\\ \nolinkurl{v5:preparation-zero-bridge}. \end{quote} The last of these has exactly the zero-preparation conclusion used in Section~\ref{sec:poles}. \item The prime-pattern result is Sun's Theorem~4.1, whose implications for the prescribed congruences and the elliptic reduction orders are proved in Section~\ref{sec:avoidance}. The same quantified statement appears as \nolinkurl{thm:v5-simultaneous-primes} in the source manuscript. \end{enumerate} The first two entries involve the identical normalized point and the identical height constant. The exact integer calculation \[ 2^{2093447}0$. Sun's Theorem~4.1 supplies positive integers $u,w$ such that \[ q\mid w,\qquad u/w^2>H_0,\qquad p_j=u+c_jw^2>Y \] are primes, with Gaussian generators whose order norms satisfy \[ \begin{gathered} v_2(d_-(\pi_j))=6,\qquad v_2(d_+(\pi_j))=2,\\ \gcd(d_-(\pi_j),d_+(\pi_h))=4\qquad(1\le j,h\le k), \end{gathered} \] and are all prime to the odd part of $L$. The primes may simultaneously be required to satisfy $p_j\equiv1\pmod{8\cdot3\cdot17}$. \end{lemma} \begin{proof} Take the avoidance modulus to be the product of the odd primes dividing $3\cdot17\cdot L$, with seven omitted. It is squarefree and prime to seven, as required by Sun's theorem. At each of its prime factors, the two adjacent norms of $i$ are both two. They are therefore units. At seven the fixed residue $1+3i$ instead gives \[ d_-(1+3i)=9\equiv2\pmod7,\qquad d_+(1+3i)=13\equiv6\pmod7. \] Thus seven is avoided even when it divides $L$. Write $\pi=9+16a+16bi$. Direct calculation gives \[ d_+(\pi)\equiv4\pmod8, \qquad d_-(\pi)\equiv64\pmod{128}. \] The exact powers of two in the statement follow. The exclusion of odd common factors supplied by Sun's theorem then gives every stated cross-sign gcd, including a position paired with itself. Finally the Gaussian norm is one modulo eight, and its residues at three and seventeen are the norm of $i$, namely one. These are precisely the required congruences on $p_j$. \end{proof} \begin{corollary}\label{cor:compatible-progressions} In the situation of Lemma~\ref{lem:arbitrary-avoidance}, put \[ U=\operatorname{lcm}_j\bigl(d_-(\pi_j)/4\bigr),\qquad V=\operatorname{lcm}_j\bigl(d_+(\pi_j)/4\bigr). \] Then \[ \gcd(U,V)=1,\qquad\gcd(U,L)=16, \qquad\gcd(V,L)=1. \] For every integer $h_*$ divisible by $64$, the congruences \[ h\equiv0\pmod U,\qquad h\equiv1\pmod V,\qquad h\equiv h_*\pmod L \] have an arithmetic progression of solutions. \end{corollary} \begin{proof} Division by four makes every minus order have two-adic valuation four and every plus order odd. The cross-sign gcds make their least common multiples coprime. Lemma~\ref{lem:arbitrary-avoidance} excludes all odd factors of $L$ from both least common multiples. The displayed gcds follow. The only nontrivial pairwise compatibility condition is $h_*\equiv0\pmod{16}$, which holds. The generalized Chinese remainder theorem now supplies the progression. \end{proof} The reduction orders at the two places of $K$ above $p_j$ are the unordered pair $d_-(\pi_j),d_+(\pi_j)$, by the source lemma \nolinkurl{lem:v5-two-reductions}. Both reductions have full rational two-torsion. For a finite group of elliptic-curve points of order $d$, full two-torsion implies that the subgroup of doubles has order $d/4$. Every double is therefore killed by $d/4$. This justifies the divisions by four in Corollary~\ref{cor:compatible-progressions}; the source point $P$ is itself a double. \begin{lemma}\label{lem:rational-contacts} The preceding statements also apply to any fixed finite list of rational contact coefficients, with a common square denominator for the slope and with arbitrarily deep fixed denominator divisibility. \end{lemma} \begin{proof} For coefficients $\gamma_j\in\Q$, choose a positive integer $D_0$ so that $D_0^2\gamma_j$ are integers, and remove repeated coefficients. Apply Lemma~\ref{lem:arbitrary-avoidance} with $c_j=D_0^2\gamma_j$ and put $v=D_0^2w^2$. Then $u+c_jw^2=u+\gamma_jv$ and $u/v>D_0^{-2}H_0$. Enlarge the prescribed lower bound before applying the lemma to obtain any desired slope bound. By choosing $q$ divisible by suitable powers of a fixed finite set of primes, one obtains any prescribed lower bounds on their valuations in $v$. Enlarge $Y$ to exclude any fixed finite list of primes. None of these choices changes the quantified form of Sun's theorem. \end{proof} \begin{lemma}\label{lem:real-progression} Let $G$ be the closure of a cyclic subgroup generated by a point $Q$ in a compact real Lie group, and let $G^0$ be its identity component. Suppose $L$ is divisible by the order of $G/G^0$. Any arithmetic progression $h\equiv h_*\pmod L$, with an additional compatible finite collection of congruences, is dense after multiplication by $Q$ in the component $h_*Q+G^0$. Arbitrarily large positive indices can be used. \end{lemma} \begin{proof} The compatible congruences give one coset modulo an integer $L'$ divisible by $L$. A compact connected abelian Lie group is a torus, and multiplication by any nonzero integer is onto it. It follows that $[L']G=G^0$. Thus the two-sided progression has the stated closure. In a compact group, the closure of a positive cyclic orbit is also a group: positive return times approach the identity and consequently approximate the negative multiples. Removing a finite initial segment does not change the closure. This proves the positive-index assertion. \end{proof} For the construction, apply this lemma to the closure of $(P,P^\sigma)$ in the product of the two real elliptic groups. The scalar point, the four integer root labels, the real compact neighborhoods, all contact coefficients, and the modulus $L$ are fixed before the prime-pattern theorem is used. Corollary~\ref{cor:compatible-progressions} then supplies the final shared index without disturbing those neighborhoods. No estimate uniform in the coefficients, $q$, $L$, or the number of contact positions is required. \section{Pole realization with zero preparation} \label{sec:poles} For $b\in\Q^\times$, put $B=17b$. The zero-preparation predicate is \[ \begin{aligned} p^2&=B(1-e)(1+3e),\\ h_0&=e(B-b^2),\\ Bq^2&=(s^2-Bh_0)(s^2+3Bh_0), \end{aligned} \] with rational $e,p,q,s$; $h_0$ is the displayed expression. No nonzero condition is imposed on these four witnesses. The source lemma \nolinkurl{v5:preparation-zero-bridge} states exactly the following two implications. A solution excludes negative odd valuations of $b$ outside $\{2,3,17\}$. Conversely a solution exists if the reduced denominator of $b$ is a square and \[ \begin{gathered} 0