Syracuse Dynamics on the Trefoil Knot: The $(2,3)$-Torus Knot as the Natural Geometry of the Collatz Conjecture

The $(2,3)$-torus knot---the trefoil---lives on the very torus that carries thefinite residue dynamics of the Syracuse map: its two meridional windings encodethe powers of $2$, its three longitudinal windings the powers of $3$, and itsrational slope $3/2$ is a continued-fraction convergent of the irrational slope$\logb 3$ that governs the dynamics. We use this geometry as an organizing lensfor a set of recently \emph{proved} results about the Syracuse valuationcocycle \cite{JanikFSS}: the parity--residue coupling$\chi_2(S(x))=(-1)^{\val(3x+1)}$, which makes the quadratic character of$(\ZZ/3^r)^\times$ an exact coboundary and biases the mod-$3$ residues in ratio$1\!:\!2$; the classification of coboundaries and the uniform spectral gap ofthe synchronized transfer operator (the finite ``H23a'' theorem); the keystonearrow (every divergent or cycling orbit casts a finite non-coboundary charactershadow); mean-criticality (any counterexample's valuation word must track theslope $\logb 3$ exactly, on pain of descent); the unconditional exclusion ofall Sturmian and subexponential-complexity valuation words by a repetitionrigidity argument whose elementary core is machine-verified in Lean~4; and theeffective exclusion of cycles via linear forms in logarithms. On the torusthese results say: a counterexample must live inside the empirically observedbranch-locus band along the irrational foliation, may not follow anylow-complexity path within it, and cannot close up. We also revisit theempirical $k=729$ branch locus from $10^{11}$ trajectories and reinterpret its``ghost island'' of $251$ cells: it is the geometric shadow of the scale-inducednear-relation $2^{729}\approx 3^{460}$ (deficit $-0.083$), which explains why itappears at $3^6$ and at no smaller power of $3$.