2-Adic Obstructions to Presburger-Definable Characterizations of Collatz Cycles
Madhav Dhiman, Rohan Pandey
TL;DR
The paper tackles the Collatz problem from the perspective of definability, showing that integrality constraints distinguishing genuine integer cycles from $2$-adic ghost cycles are not capturable by Presburger arithmetic. It constructs ghost cycles as unique $2$-adic solutions to the cycle equation $n_0(2^x - 3^y) = C(y, \vec{\sigma})$ for admissible parity patterns and demonstrates these cycles form genuine periodic orbits under the $2$-adic Collatz map $T_2$, while the integrality condition required to collapse ghost cycles to integers is non-semi-linear due to unbounded fiber periods. This establishes a fundamental obstruction to Presburger-definable approaches and finite-automaton encodings of integrality, implying that any proof strategy must involve analytic, not purely algebraic, arguments. The results illuminate the limitations of linear arithmetic methods for Collatz and point to the necessity of analytic obstructions in bridging $\mathbb{Z}_2$-solutions and integers, with open questions about the density and structure of ghost cycles and possible generalizations to maps of the form $qn+d$.
Abstract
I investigate structural limitations of Presburger-arithmetic-based approaches to the Collatz problem. I show that the Collatz cycle equation admits a unique solution in the $2$-adic integers, which I term a \emph{ghost cycle}. These ghost cycles are shown to be genuine periodic orbits of the $2$-adic Collatz map, satisfying all local parity constraints. I prove unconditionally that the divisibility predicate $\mathcal{D}_y = \{(x, C) \in \mathbb{N}^2: (2^x - 3^y) \mid C\}$, which acts as the algebraic necessary condition for integrality, is not semilinear for any fixed number of odd steps $y \ge 1$. This result is established by demonstrating that the fibers of $\mathcal{D}_y$ exhibit unbounded periods, an obstruction to Presburger definability. Consequently, strategies relying solely on Presburger arithmetic or finite automata to define the integrality constraint cannot capture the distinction between ghost cycles and genuine integer cycles. I conclude with a heuristic argument suggesting that because ghost cycles satisfy the algebraic cycle equation, the non-existence of integer cycles cannot be proven solely through algebraic manipulation of the cycle equation itself.
