Irreducibility

How much of a genesis is beyond recovery from its endpoint is an exact quantity here — computational irreducibility with closed-form route posteriors — and it has a critical point the thermodynamics cannot see. The per-move route-entropy rate is analytic in β, a smooth crossover: on a breathless column, one whose period rises at every deepening (ramification 1), it peaks at log 2 exactly at the critical clock β_col (the root of the depth column's thermal normalizer Z_col(β) = 1; ≈ 1.496 for the 3-column); where the column breathes, its period rising only every few deepenings, the peak sits strictly below log 2 (the mode staircase below). What reorganizes at the clock is the genesis mode: the unbounded-depth tail's occupation jumps 0 → 1, and the depth to which the origin stays uncertain — a correlation length — climbs as Θ(ln 1/(β−β_col)), walking the column's transparent prime family — those whose arrival leaves λ where it stood — one entry at a time (the clock gap is governed by the next unincluded family prime: measured slope −1.499 against −β_col = −1.496, R² = 0.99997), so it diverges exactly iff that family is infinite. The structural law: an interior critical point exists iff the fate's normalizer keeps a nonzero unbounded-depth limit — depth's does; breadth's clocks pile up only at the ζ pole. Every finite place carries its own clock: β_col at the locked column's prime q is rooted in that column's transparent set, which is q's residue-and-power data (for odd q: 2 transparent at every column, 5 iff q ≡ 1 mod 4, and a prime uniquely blind to its own column), so the finite places of Q map into the interval (1, β*), β* = 1.7286 the root of ζ(β) = 2, richer transparent sets pulling the clock lower. The top is the 2-adic seat: the q = 2 clock, 1.6045, cancels its own powers and sees only the Fermat primes — the poorest transparent family, the highest clock — so the divergence question at the seat, its transparent family being the Fermat primes, is the Fermat-prime question. The reading survives a change of field: over F₂[x] the clock equation clears to a polynomial in 2^−β (the degree-1 clock solves 2^−β = 1 − 1/√2 exactly; same-degree places share a clock), over Z[√−5] the clock splits by the class-group characters (the 3-column's element clock 1.3527 below its ideal 1.5492, the pair's ratio reading h near the pole), and in every tested field the highest clock sits at the place whose residue field has two elements — Ostrowski's census of the places, read thermodynamically.

Scope. The rate's analyticity and the structural law are exact statements about the closed-form normalizers; the spectrum values, orderings, and the seat law are computed across the tested families; the correlation length's divergence is pinned, column by column, to the finiteness of the column's transparent prime family (the walk is measured at the 3-column; the per-column pinning is read from the ordering of each column's mode steps) — settled where the family is provably finite, open at the 2-adic seat and the 3-column.

verifiers: explore_irreducibility_order.py, explore_irreducibility_places.py, explore_irreducibility_crossfield.py

A column at a place of any global field has a ramification index e, the number of times the place's own uniformizer divides the rational prime p beneath it, and past a finite transient its period λ rises once every e deepenings — a tick. Between ticks the outside entrants of the transparent set stand still while the column's own contribution to it shrinks, so the clock is not one number: it cycles through exactly e of them, and the deep clock is the root of that same normalizer read at the states sitting immediately before a tick. Four layers, and each reads a different datum of the field. THE VALUE reads the limit transparent set alone — which primes eventually arrive transparently, its entrants, and at what exponent, fixed by the field's splitting law and the period tables of its ramified places — and reads nothing of the finite-time detail beneath it, because an entrant's exponent cap is divisibility-monotone in λ and λ never falls, so the limit set is path-free. Two witnesses, one on each side: Q(√3) and Q(ζ₃) carry columns with the same local data (p, e, f) = (3, 2, 1) — prime, ramification, residue degree — and opposite fine structure, yet their clocks 1.4475 and 1.3336 are separated by the transparent set alone: the first admits the single entrant 2 and nothing else ever, the second a long family of split primes, and transplanting either field's unit chain onto the other moves neither root. Q(√2) and Q(√−2) run the other way, sharing one chain and still landing at 1.4861 against 1.6318, their transparent sets parting on whether 3 splits. Same chain, different clock; same set, same clock. What that pair witnesses is the invariance and not a direction: two fields also carry two zeta functions, so the ordering ACROSS fields is not the within-Q rule above that richer sets sit lower. THE ACCUMULATION SET reads e: past the transient, the thermal mass of a state's transparent moves (the term 1 for moving nowhere counted) takes exactly e values — its pre-tick value times the partial sums 1, 1 + p^−β, 1 + p^−β + p^−2β, …, through e terms, one per depth of headroom left before the next tick — and the e clocks are the roots those e masses give, the deep clock's the bare 1. Four values over Q(ζ₈) and one over Q itself, whose 2-column has e = 1 and does not breathe at all. THE PHASE reads torsion — whether the field already contains a p-th root of unity flips which deepenings the pre-tick states fall on. And the finite-time TRANSIENT is the only layer that reads the fine structure of the p-power map on the field's units — the arrival-graded object the Observatory measures, visible in the normalizer at finite depth wherever a field carries one, and nowhere at the limit.

Scope. The blindness is proved from the two limit facts stated, given that a state's transparent moves are the divisors of one cofactor, and inherits the tier of the local period tables it is computed over; the layer readings are a rule in range over eight columns — Q, Q(√2), Q(√−2), Q(i), Q(√−5), Q(ζ₃), Q(ζ₈), Q(√3) — each deep clock computed three ways, through the field's true unit chain, through a counterfactual chain carrying no fine structure, and at the limit set, agreeing to 10⁻⁹; convergence is exact rather than asymptotic, the pre-tick normalizer equalling its limit past the last entrant threshold. Residue degree above 1 and the non-abelian cubics are outside the census.

verifier: explore_clock_factorization.py

On a breathing column — ramification e ≥ 2 — the genesis mode does not jump once: it quantizes into e geometric steps, one at each of the column's e clocks. The mode's occupation climbs 0, 1/e, …, 1, beginning at the bracket clock — the one rooted at the just-ticked states, the breath's other end from the deep clock — and completing at the deep clock: the pre-tick states join the mode last. In a mixed-characteristic window the depths that have joined between consecutive clocks form a residue stripe mod e, and the stripe's phase — which residues are in — reads the field's unit chain: Q(√3) and Q(ζ₃), the twin columns whose clock values the chain provably cannot move, carry complementary stripes (even against odd), and transplanting either chain onto the other flips the stripe while every step's clock holds to 10⁻⁹. The mode's geography reads the torsion that the clock values, all e of them, are blind to. And breathing is taxed: the per-move route-entropy rate stays analytic through every step — all e transitions are geometric — and its peak sits strictly below log 2 whenever e ≥ 2 (by 3.3·10⁻³ at √3 up to 9.8·10⁻³ at ζ₈): a peak at log 2 needs a fair per-move coin, and the clocks being distinct, at most one stripe's can sit at 1/2 at any one temperature. Perfect per-move irreducibility is exclusive to breathless columns. Each step's correlation length walks that step's own transparent family — constant at √3, whose family is provably finite (a first-order-like jump, bounded unconditionally); the Fermat entry thresholds at the 2-adic columns (over Q the onset climbs 4, 6, 10, 18, the depths at which 5, 17, 257 and 65537 enter — bounded past 65537 iff the Fermat set is finite); ζ₃'s family 2·3ⁱ + 1 exactly, the onset depth climbing linearly in the family index — so whether a column's transition is a sharp jump or long-ranged is pinned to its family's finiteness: settled at √3, an open prime-family problem at the others, with the standard heuristics putting the 2-adic seat (its Fermat sum convergent: a sharp jump) and the ζ₃ column (its family sum divergent: long-ranged) on opposite sides.

Scope. The step ordering is proved from the accumulation set; the quantization, the stripes, the tax and the correlation-length walks are a rule in range over the tested columns (√3, ζ₃, ζ₈, the 2-adic columns); the chain-transplant invariance is measured to 10⁻⁹; the transition order is settled where the transparent family is provably finite and open where its finiteness is a prime-family question.

verifiers: explore_mode_staircase.py, explore_astar_slope.py