Growth over other fields

The same demands run over a number field or a function field, and what each ring charges: the local invariant deciding whether a demand locks or sprawls, the price a class group adds when growth is paid for in elements, and how much of that structure temperature dissolves.

A growth law is a structural demand plus a greedy move: extend the state by the least admissible m ≥ 2, where a tie-break settles any choice among equally cheap admissible moves. Over Z the demands trichotomize into three fates — breadth, depth and mortality (Growth) — and every one of them is stated over a ring. Throughout, λ(N) is the exponent of the unit group, the least L with aL = 1 for every unit a; a prime p's window is its slot in the state, OPENED by seating p and DEEPENED by raising p's exponent; and the door at p is the least move there that raises λ against the whole state at once. The depth demand is dynamics-greed, which asks that λ grow; over Z it locks, collapsing within a bounded number of moves onto one prime's column forever with no rival door falling afterwards. A law's basin map sends each seed to the place its trajectory ends on. Everything below replaces Z.

The price of depth: the module law

Run the depth fate across global fields and the outcomes split: cold dynamics-greed locks in mixed characteristic (Q, Z[√−5], Q(√−23)) and sprawls in equal characteristic (Fq[x]: openings recur forever, under any tie-break). The deciding invariant is local.

The local module law theorem

At a place with residue field Fq, q = pf, ramification e, the 1-unit exponent E(a) = exp(U1/Ua) obeys: in mixed characteristic U1 has Zp-rank ef and the p-power map is an isomorphism UbUb+e past the wild threshold, so E(a+e) = p·E(a) eventually — a linear pump, one λ-tick per e of depth at the eventually constant price qe = pef = prank. In equal characteristic (1+u)pm = 1 + upm gives E(a) = p⌈logp a exactly, with rank infinite — a log clock whose tick prices are unbounded on any recurring column. The lock is the finite-rank case; the sprawl is its rank-∞ limit.

Scope. The two cases are proved (Cohen structure / direct computation); instances brute-verified at Z, Q(√−23), and F2/F3/F9[[t]].

verifier: explore_module_law.py

The law's full chain, wild transient included, is one orbit: every unit's p-power orbit o(n) = v(upn − 1) steps by ψ(i) = min(p·i, i+e) — a Frobenius gear below, a pump gear above — with at most one splice at the Kummer seat i* = e/(p−1) (off the seat a two-line valuation — for a unit 1+x at level i = v(x), the terms p·x and xp of (1+x)p − 1 sit at i+e and p·i, every other term above the smaller; the census over 14 mixed-characteristic fields and 2 equal-characteristic rings is the record: explore_local_clock.py). What the transient reads classically — the ramification of the p-power cyclotomic tower over the base — is the Observatory section's story.

The class-group split pattern

At class number h > 1 the depth fate runs once per currency. Growing by ideals, all three censused fields reproduce Z's geography: every seed locks a window at the module-law price, each quadratic tie a Galois orbit broken by fiat (only Q's greedy needs no tie-break clause), and the cubic field of discriminant −283 locks five distinct places across its census where the headless cubic — one with no head, no place whose column opens with a stretch of wider gaps — has a one-point basin map (the headed walk). Growing by elements — principal states and moves — the world pays its class group in price, and the basin map is a norm race among the principal vehicles the seed's dowry — the windows and the λ it starts with — leaves alive. A nonprincipal window's cheapest vehicle is the free-rider door, the window bundled flat with a partner at the product norm — in Z[√−5] a norm-6 move deepens the 3-window while carrying the 2-window along — and every ride lifts the passenger one depth until (2) wakes and absorbs the trajectory: Z[√−5]'s element census is a (2)-monobasin at 4 per move. The monobasin is field geography, not an h > 1 law. Q(√−23) (h = 3) has two basins: 38 censused seeds ride to (2) at 4 per move, three lock the inert 5 at 25 per move — a pure principal column at the untaxed ideal price, whose very opening carries period 24 and swallows the novelty of every cheaper move; the passenger lift is real but conditional on riding. The cubic (h = 2) has three: 17 seeds to the square of its cheapest place, 3 to a principal norm-17 place tax-free, and 1 to the full rational (2) at 8 = 2³ — the first tail to pay its ring's degree, because that seed's dowry supplies the power of 2 that covers the cheap square's every step, so the square never moves λ and the degree-priced vehicle is the cheapest LEFT: a dowry can edit the vehicle menu. THE ELEMENT PRICE: a tail pays its cheapest 1-move principal vehicle still ticking — still moving λ against the state — the power Pm (m the window's class order, norm pfm, f the residue degree) against the rational (p) (norm pn, n the ring's degree) — so pmin(fm, n) per move wherever both tick. The class-group tax exists only at fm > n, and it is pure myopia — in Q(√−23) the tax-free cube P2³ (P2 a prime above 2) sits at norm 8 forever, three depths per purchase at the untaxed 2 per tick, and a per-move greedy never picks it. And the tax need not buy depth at all: the cubic's taxed square pays 4 = 2² per move while the exponent E above climbs two doubling steps per move — 2 per step, the ideal world's own price at that place — so where the vehicle squares a place whose E-column carries a flat step, the tax buys granularity, two steps a move, and not step price.

Scope. Censused wall to wall at Z[√−5] (h = 2) and Q(√−23) (h = 3) and over the norm ≤ 40 belt at the cubic field of discriminant −283 (h = 2), both currencies; the basin counts are each field's census. The element price is the pattern the block is badged at: three fields, seven censused tail species, no proof; the still-ticking clause is load-bearing, the cubic's dowry edit being its counterexample otherwise. The pricing it rides is the module law above.

verifiers: explore_number_field_lock.py, explore_module_law.py, explore_headed_cubic_walk.py

Every lock above is a census result: a walk over a field's seeds, wall to wall, finding a lock at the end of each. Whether a lock must EXIST, along some trajectory nobody has enumerated, is a question no census reaches, and it has a corpus of its own (The cascade boundary).

The function-field melt

Run the thermal law — pick admissible m with probability proportional to |m|β rather than taking the argmin (Thermal and density) — over F2[x], where |m| = 2deg m is the norm at the infinite place of F2(x) and the weight is |m|β, and the apparatus transfers with closed forms. One target of the transfer is the crystal, the squarefree limit a breadth trajectory reaches when every irreducible of the world enters once, at depth 1, and none deeper.

The function-field melt rule

The partition function is rational: ζF2[x](β) − 1 = r/(1−r) with r = 21−β, and the crystal probability is exactly 1 − 21−β — 1/2 at β = 2, against Z's 1/ζ(2) ≈ 0.608: no transcendental content survives. The hot-limit theorem transfers, so thermal dynamics reaches the full profinite completion almost surely at every β > 1; the sprawl's geography is a β = ∞ artifact exactly as Z's lock is. Over Z temperature's gift to the dynamics demand was depth (the radical); over F2[x], where cold growth already runs one column's radical while starving breadth, the gift is breadth — the shadowed siblings open.

Scope. The closed forms are exact — partition function, crystal probability, degree law. The hot-limit transfer is the theorem's own argument run here — any thermal growth whose admissible sets are never empty and closed upward in divisibility reaches the full completion almost surely: λ is divisibility-monotone over F2[x] too, so admissibility is closed upward, and the chance that a pick deepens a given irreducible g by at least k is at least |g|βk whatever the state.

verifier: explore_function_field_melt.py

The mortality split rule

What temperature never touches is the module law's diverging costs. Bound the alphabet — moves of degree ≤ D only — and every F2[x] thermal trajectory halts, the death authored by the clock c — the exponent of the dyadic frontier the trajectory climbs — concentrating from a light seed (one starting well below the mode) at c*(D) = ⌊log2 D⌋ + 2; halt times obey T ≤ (D−1) + (cfinalc0). Z, by contrast, has an immortal bounded-alphabet witness: along the pure 2-column, m = 2 rides the linear pump forever. The dichotomy is the module law thermally dressed: selection-authored structure (geography, route, tie-breaks, the lock) melts at β = ∞ exactly; module-authored structure (the cost divergence) is temperature-invariant.

Scope. Bounded-alphabet model over F2[x], light seed. The mode c*(D) is β-free; the tail distribution is not (priced by the dyadic phase of D).

verifiers: explore_halt_clock.py, explore_function_field_melt.py

The cascade boundary

These results are now a page of their own. What they settle: the escape's demand is a conjunction over characteristics rather than a disjunction, so one characteristic closed closes the ring; a budget inequality kills the characteristics 2, 3 and 5 outright and a direct walk kills every one below 1000, still dying when the multiplier is widened by the factor an element move pays; the principal places that close a field unwidened reach at least 23 times further than that factor does and are floored at |D|/4, so finitely many fields either way — that arithmetic now stands on Principal places. What stays open: the residual is inhabited at every bound by an elementary construction, cheap to enter for any ring written down, and no argument built on small-prime divisibility can tell a close covering every ring from the weaker one that would suffice.