Imagine you're trying to measure the total water flow through a city's pipe network, but every time you use your standard flow meter (the Rényi approach), one particular junction causes the meter to blow a fuse at exactly the setting you need (n=1). For years, everyone has been extrapolating from nearby settings — measuring at n=0.9 and n=1.1 and hoping the limit behaves. This paper says: forget the meter. Trace every pipe individually, decompose the network into independent segments and loops, and add up the flows by hand. That's the mechanism here — a combinatorial decomposition that sidesteps an analytic singularity. The committed claim: the complete computational-basis Born distribution of the critical transverse-field Ising chain can be decomposed, at exactly n=1, into independent long-range Bernoulli edges and conditional cycle-space terms, yielding an all-orders analytic hierarchy for the Shannon entropy. The four-vertex linked coefficient is evaluated in closed form, and for an interval the edge contribution is pinned exactly at γE = log2/4 − 1/16. The logarithmic coefficient γ1 = 0.060020(3) is recovered independently from exact finite-size probabilities. The architecture is pure analytical statistical mechanics — no numerics-first approach, no Monte Carlo, no tensor networks as the primary tool. The method represents the Born probabilities of the full computational basis using Pfaffians and decomposes them via the Shannon chain rule into edge-by-edge and cycle-by-cycle terms. This is a graph-combinatorial framework applied to a free-fermion lattice model. The key structural insight is that the periodic chain's probability distribution factors into independent Bernoulli random variables (the edges) plus conditional corrections from the cycle space of the underlying graph, and these corrections organize themselves into a linked-cluster expansion indexed by vertex support. The ladder here is subtle. The prior art is primarily numerical: the periodic constant and the logarithmic coefficient γ1 were established by high-precision numerics (Stéphan, Misguich, Pasquier and others). This paper does not beat those numbers — it recovers them analytically. The value proposition is not a better number but a better understanding: an explicit finite-dimensional integral representation for every linked coefficient, with the first few sectors already nearly saturating the known constant. For the interval, the exact result γE = log2/4 − 1/16 and the recovery of γ1 = 0.060020(3) serve as self-consistency checks against established numerics rather than competing predictions. Integrity is both a strength and a caveat. The self-consistency checks — recovering known numerical constants from the analytic hierarchy — are strong. The paper is explicit about what is unconditional and what is conditional: the edge anomaly and the exact interval edge contribution are proven, while the full constant for the periodic chain and the remaining logarithmic coefficient rest on stated 'finite-part matching and forest assumptions.' The paper flags these assumptions clearly, which is good scientific practice, but it means the complete result is not yet a theorem in the strict sense. No code is mentioned, no pre-registration applies (this is mathematical physics), and independent verification would require another group working through 86 pages of combinatorics. The milestone question for universal subleading entropies in critical systems is whether full analytic control can extend beyond free-fermion models to interacting systems. This paper establishes the machinery for the simplest nontrivial case — the free-fermion Ising chain. The next concrete target would be closing the conditional assumptions to make the periodic constant fully rigorous, and then attempting the same decomposition for a model outside the free-fermion class, such as the XXZ chain at its critical point. The obvious experiment not run: extending to the XXZ or Potts models where the free-fermion structure is absent. The honest read is (a) — the combinatorial decomposition leans heavily on the Pfaffian structure of free fermions, and generalizing to interacting models would require fundamentally new machinery. This isn't something saved for a sequel; it's a genuinely harder problem that this framework, as constructed, cannot reach without significant extension.
