The programme in seven stages

Each stage is anchored by a registered result: the Rational-Fibre Theorem (T1), the common-future theorem (T6), the Hasse structure of the reachability graph (S2.10), the incompatibility theorem (T15), exact scaling (T51), single-prime dominance (T60) and the effective bounds (T69–T72).

1When a number has more than one history

The number 72 appears ordinary. Yet within the UNNS generative framework it has three distinct incoming routes and four complete histories. The number is the same; the structures that produce it are not.

The rule that generates these structures is deliberately small. Take a number m and a nest size N that divides it. The nest rule sends m to m(N+1)²/N. Every integer therefore has as many successors as it has divisors, and the whole system is a single directed graph on the positive integers.

Ask which numbers lead to 72, and the answer is a surprise. The number 18 reaches it with N = 1, the number 16 with N = 2, and the number 10 with N = 5. Three different generating steps produce one and the same value. Follow each of them back to where it began and there are four complete histories, because 16 can itself be reached in two ways.

Figure 1. One value, many generative histories. In the UNNS Substrate, numerical equality does not erase generative distinction. The integer 72 has three distinct incoming routes but four complete histories in the SELECT-II arithmetic graph. This demonstrates that a numerical endpoint, its immediate generating routes, and its complete ancestry represent different levels of mathematical structure.

The diagram below is not an illustration drawn by hand: it is computed from the rule. Every edge is a genuine step of the nest rule, and the four histories are read off the graph.

That is the opening discovery, and the rest of the programme follows from taking it seriously: UNNS distinguishes numerical equality from generative identity. The question is no longer only what value a process produces, but by which structures — and what those structures force about the future.

2From a generative framework to two mathematical volumes

The idea has now been developed into a complete two-volume treatise, The Mathematical Foundations of the UNNS Substrate. Each volume has seven chapters, and every statement in both is a registered record: it carries an identifier, a scope, a standing, and the list of earlier records it depends on. Altogether the registry holds 128 records, 60 in Volume I and 68 in Volume II.

Two features make the treatise unusually checkable. First, a proof audit: for each Volume II record, the manuscript states whether the result is a full proof, an exact computation, a definition, a scope statement, a registered correction or an open problem — and where in the text it can be found. Second, reproducible verification: the numerical claims of Volume II are backed by 49 scripted checks, and all 49 pass.

The registry also records scope. Some results concern pure rule arithmetic and make no choice about interaction at all; others hold only inside a named interaction grammar. Where an early statement was too broad, the registry records a correction as a precise mathematical scope statement, so that every claim in the volumes says exactly what it covers.

3Generative multiplicity has an exact arithmetic structure

Volume I starts by making “more than one route to the same value” precise. For a rational value x = a/b in lowest terms, the occurrence fibre is the set of generating steps that produce it. The Rational-Fibre Theorem says this set is exactly D(x) = { d ≥ 2 : d² divides a, and d ≡ 1 (mod b) }, with nest size N = d − 1.

For integers the count has a closed form. If 72 = 2³·3², the number of routes is (⌊3/2⌋+1)(⌊2/2⌋+1) − 1 = 3: the square divisors 2, 3 and 6. The sources of the graph, the numbers with no incoming route, are exactly the squarefree integers. Writing a = s²t with t squarefree reduces every fibre to its square-root part s, so a whole value is governed by a much smaller object.

The arithmetic is also honest about where it is not tidy. For an integer value, the set of routes is closed under least common multiples. For a proper fraction it need not be. The value x = 400/3 has routes d = 4 and d = 10, but their lcm 20 is not a route, because gcd(4, 10) = 2 is not congruent to 1 mod 3. The theorem that survives is exact: an lcm of routes is again a route precisely when their gcd is ≡ 1 mod b. Volume I states the integer case and the fraction case separately.

4The hidden geometry of generative routes

Distinct occurrences of one value do not drift apart. They share common futures, and those futures form a lattice. For two routes d and d′ with nest sizes N and N′, put L = lcm(d, d′) and g = gcd(N, N′). The values that both routes can reach, at or after x, are exactly x + jΔ for j ≥ 0, where Δ = L²/g.

Take 72 again. Its routes d = 2 and d = 3 have L = 6 and g = gcd(1, 2) = 1, so Δ = 36 and the two routes keep meeting at 72, 108, 144, and onward. The meeting index x/Δ is always a positive integer, and every positive integer occurs as such an index.

Taking all routes of a fibre together gives a collective rendezvous step Δ_coll = lcm(d_i²)/gcd(N_i), and a coherence integer γ(x) = x/Δ_coll, which equals the gcd of the pairwise meeting indices. This integer γ will return in Volume II as the coordinate in which the dynamics becomes simple.

Figure 2. The internal geometry of generative multiplicity. Distinct occurrences of the same rational value possess precisely calculable relationships through common-future lattices and weighted fibre graphs. Three-route configurations can be classified by their divisibility structure and weighted-graph symmetries, revealing both rigid and nonrigid arrangements. Generative multiplicity therefore carries mathematical organization beyond the mere number of representations.

Three routes: a complete classification

With three routes the geometry is a weighted triangle, and Volume I classifies every case. A shape theorem shows that a three-route fibre is exactly one of a short list: a chain, a product configuration, a single comparable pair, or an antichain. Integer fibres with three routes are always non-rigid — the routes are {p, q, pq} or {p, p², p³}, and the triangle has a non-trivial symmetry. Rational fibres can be rigid: for x = 20736/5 the routes are {6, 16, 36} with pair indices {1, 9, 16}, and the triangle has no symmetry at all.

A signature criterion turns this into a test: a three-route fibre is rigid exactly when its three pair signatures are pairwise distinct. And rigidity is not an isolated accident. Volume I constructs infinite rigid families, including scaled copies of any rigid primitive and a family indexed by primes in an arithmetic progression.

5Reachability is not the same as numerical equality

Volume I examines the way routes might be combined from two directions. In the first, an anchor-route model, a joint operation on routes is proved to be the same thing as a reachability order: choosing how routes combine is choosing an order. Requiring the combination to be minimal and to absorb the generative moves singles out one operation, but the volume is careful to record that this is an additional rule, not a consequence of nesting alone.

In the second, the graph is declared first and its structure is read off. This is the arithmetic graph of Section 1, and its properties are exact. Each integer m has exactly τ(m) successors, and every step at least quadruples the value, so the graph is acyclic. The in-degree of t equals its nest multiplicity. Every integer has a squarefree ancestor, and ancestor chains are no longer than log₄ t.

No universal join

One might hope any two numbers have a least common future. They do not. The numbers 1 and 2 have exactly two common futures up to 100, namely 81 and 100, and 81 does not reach 100. So there is no least common future, and the reachability order is not a join-semilattice. Volume I records the counts as data: among the pairs 1 ≤ a < b ≤ 40, most have at least two minimal common futures, and none of those has a join.

A Hasse diagram that is not graded

Every edge of the graph is a cover relation: no edge can be shortcut by a longer path. The graph is therefore the Hasse diagram of its own reachability order, and the predecessors of any value form an antichain. Yet it is not graded. From 6 to 784 there are paths of different lengths, so path length cannot serve as a rank.

Finally, histories. The history count H(t) is at least the route count R(t), and for 72 it is strictly more: H(72) = 4 against R(72) = 3. A Markov no-go explains why this information stays hidden: the continuation of a history depends only on its endpoint, so anything that looks only forward cannot see how a value was reached. The route and history structure is retrospective — and a history-sensitive operation has to be defined by reading backwards.

6Why interaction cannot simply be assumed

Suppose two routes into the same value are allowed to interact, and the result must be a new value. Which one? Volume I answers with a theorem of nonselection. If an interaction of two routes is required to produce a value in their common future, and to satisfy the basic admissibility conditions, then it must have the form x + j(r, s)Δ, where j is any positive integer the operation chooses. The arithmetic permits the whole family; it does not pick a member. Asking for minimality selects j = 1, but that is a choice added from outside.

Natural requirements then collide. Three are especially natural: that interactions land in common futures (common-future admissibility), that the law is unchanged under shifting both routes together (stationarity), and partner independence, in its weak form: the outcome must not be tied to the labels of the particular routes that are paired. The incompatibility theorem proves that no value-valued interaction defined on the full occurrence-pair domain can satisfy all three at once.

Figure 4. Interaction nonselection and mathematical incompatibility. UNNS arithmetic determines admissible common futures but does not uniquely prescribe how distinct generative routes must interact. The Volume I incompatibility theorem proves that common-future admissibility, stationarity under paired route shifts, and weak partner independence cannot simultaneously hold for a value-valued interaction on the full occurrence-pair domain. The result identifies precise constraints on the construction of additional interaction grammars.

Two consequences are worth stating plainly. Operations that ignore route labels are necessarily collective: they depend only on the value and its fibre. And the two natural families split along the incompatibility. A route-aware interaction can be stationary but uses labels; a collective interaction is label-blind but non-stationary. Neither is forced. A grammar has to be chosen, and the choice has to be stated.

7The transition from generative structure to collective dynamics

Volume II adopts the collective map. Given a value x with fibre routes d₁, …, d_k, let L be their least common multiple and G the gcd of the nest sizes d_i − 1. Then

What changes with an explicit grammar is that questions about the future become questions with answers. The coherence integer γ = Gx/L² is a positive integer that tracks the orbit, and a plateau law says that while the fibre stays unchanged, γ simply rises by one at each step. Fibres can only grow along an orbit, and G can only shrink to a divisor of itself.

The orbit of 72 shows the machinery working. Its fibre is {2, 3, 6}, with L = 6, G = 1 and Δ = 36, so C(72) = 108. The table lists the first states.

This is the simplest regime, where G = 1, and it can be solved completely. The values run through the cycle γ = 2, 3, 1 forever, the number of routes climbs by two every three steps, and from the start x_n+3 = 4x_n — an exact scaling law, visible in the table (72 → 288, 108 → 432, 144 → 576). Everything that follows asks what survives of this picture when G is larger and the dynamics are richer.

8Infinite growth, yet restricted possibilities

This is genuine unbounded growth, and one might expect it to bring chaos. It does not. Volume II proves a pair of restrictions that sit alongside the growth.

Reachability is sparse. In the regime G = 1 the number of values an orbit family can reach below a bound X is given exactly, and the set has density zero. In the persistent regimes treated later the counting function is only of order log X while the route count grows linearly. The orbit explores a vanishingly thin part of the integers.

Collapse is not universal. A natural conjecture would be that G always falls to 1. For small values it does: G = 2 collapses within 7 steps, G = 3 within 2 and G = 6 within 29, so a permanently stabilised G cannot be 2, 3 or 6. But the conjecture is false in general. For every g ≥ 5 coprime to 6 there are persistent regimes, collective orbits whose G equals g at every step and never reaches 1. A later theorem extends persistence beyond the coprime case, while showing that a permanently stabilised G greater than 1 can never divide 24.

The persistent orbits are also confined: after a finite time their γ-coordinate stays inside a fixed window of size at most twice a modulus determined by g. Unbounded numbers, bounded coordinate.

9Exact scaling: order within unbounded evolution

The γ-coordinate gives a one-step law for persistent orbits. From γ the next state is computed from a single integer: the next fibre, whether the orbit persists, and the next γ are all determined by γ + 1 and a few arithmetic invariants. Applied to a concrete case, g = 11, it gives the eight-step theorem: from a specific starting value the orbit satisfies x_n+8 = 9x_n, with γ running through the cycle 1, …, 8 and one 3-carry per eight steps.

The general statement is a closure theorem. The state that matters — the class of the orbit — can take only finitely many values, so every persistent orbit is eventually periodic in its class. Periodic classes force an exact numerical law.

It is worth pausing on how strong this is. The orbit never stops growing. Yet its growth is not merely “geometric-like”: after the transient it multiplies by a fixed integer ratio ρ² exactly, every P steps, forever. Unbounded numerical evolution is compatible with precisely organised structural recurrence.

10Structural rigidity: why one prime survives

The scaling factor ρ is not arbitrary. To see why, Volume II looks inside the dynamics at its carries: the events at which the fibre acquires a new route and γ resets. Which carries occur is itself an arithmetic question, and the answers are rigid.

• Carries are eventually complete, and each one has a definite length, one less than the square of its size.
• Eventual carries are prime powers. Every carry in the eventual regime has the form p^j.
• The decoupled odometer law. Each active prime behaves as an independent counter; at γ = 1 the next carry is the smallest pending prime power.
• A carry scheduler. At every reset state the next carry is the prime with the smallest pending carry, ties are impossible, and the orbit stays persistent exactly while that carry stays inside the window set by g.

Put the pieces together and the competition between primes has a winner, and only one.

Take g = 31. Its single active prime is 2. One period of the orbit consists of three carries, 2², 2² and 2¹, which occupy 15, 15 and 3 steps, so P = 33. The exponent of 2 rises by 2 + 2 + 1 = 5 per period, so ρ = 2⁵ = 32 and the orbit satisfies x_n+33 = 1024 x_n for every n in the eventual regime.

Figure 5. Exact scaling and the emergence of single-prime rigidity. Under the explicitly defined UNNS collective dynamics, expanding arithmetic trajectories eventually acquire an exact recurring structure. The single-prime dominance theorem establishes that only one prime carry persists indefinitely, forcing the eventual scaling factor to be a prime power. Thus, unbounded numerical evolution becomes compatible with precisely organized structural recurrence.

Further results make the picture robust. For an odd prime g, two saturated clocks always have coprime periods, so the primes cannot lock into a common rhythm. A general q-carry theorem handles other primes, for example a 3-carry at g = 11 with an eight-step cycle and a 7-carry at g = 367 with a forty-eight-step cycle.

11From eventual behaviour to certified prediction

“Eventually” is a weak word for a predictor. Volume II closes the gap in three steps: it identifies the survivor, certifies the identification from finite data, and bounds the time it takes.

Survivor selection and static certification

A minimax selection theorem shows that for a fixed set of independent odometers the survivor is the prime p that minimises p^d, where d is the largest cyclic gap in that prime’s pattern of availability — independent of the phases. A quotient-residue formula computes that pattern. An exact carry-level predictor then gives, from any reset state, the next carry and the next state.

A static survivor certificate (T66, proved for prime g) turns this into a finite test. If three explicit conditions S3–S5 hold at a reset state, then only the certified prime carries forever and the scaling law holds for all later steps. The certificate depends only on finite data at one state.

Effective transient bounds

How long can the transient last? Volume II gives four answers, T69–T72, each stronger in a different direction.

• T69, reset descent. In a constant-G segment starting at γ₀, the first reset occurs after at most κ(γ₀) carries, where κ(a) is the least k with 4^k > (3a − 4)/2. This bound is sharp: it is attained for every κ from 1 to 11. For g = 103 and γ₀ = 1,345,610, the orbit needs exactly 11 carries.
• T70, class counting. After the first reset the class sequence repeats within H(g) steps, where H(g) = ∏(2o_p − 1) over primes with p² < g.
• T71, polynomial bound. The entry index is at most K_poly(g), a bound that is cubic in the multiplicative orders instead of exponential.
• T72, dynamic G. From any state, the chain of segments with changing G has at most Ω(g₀) segments with G > 1, and the total gives an explicit bound R(g₀, γ₀).

For concrete numbers take g = 211. The class-counting bound is H(211) = 24,165,141,422,031 and the polynomial bound is K_poly(211) = 113,903,360. Recorded entry indices are much smaller: 208 in the recorded run at g = 211, and a sample maximum of 194 at g = 197 among 150 sampled orbits.

Figure 6. From structural theorems to effective arithmetic prediction. Volume II establishes computable upper bounds on the transient preceding eventual arithmetic organization. For g = 211, the class-count bound H(211) = 24,165,141,422,031 is reduced to K_poly(211) = 113,903,360. Recorded entry indices are much smaller: 208 in a source run at g = 211, and a sample maximum of 194 at g = 197. These comparisons illustrate both the power of finite mathematical certification and the opportunity to derive substantially sharper practical bounds. The illustration is conceptual and not drawn to scale; the computed chart below plots the exact values.

12Seven revelations about generative mathematics

Taken together, the two volumes support seven short discoveries. Each is stated once, with the reason it matters.

13How UNNS relates to established mathematics

None of this stands outside mathematics. The volumes lean throughout on established tools — divisibility and congruences, orbits and recurrence, partial orders, finite-state periodicity, scheduling arguments and exact checking — and the programme is best read as a development that connects them around a specific generative system. The table lists the shared concepts and what is specific to the UNNS treatment.

The aim is to show how UNNS builds on these fields while naming the specific results and structures that the treatise develops. It does not replace any of them, and the use of classical techniques inside the proofs is part of what makes the results checkable. Two standard references frame the comparison: the integer sequence of square-divisor counts, OEIS A046951, which is the classical multiplicity formula underlying the fibre counts, and Silverman's The Arithmetic of Dynamical Systems for the established arithmetic-dynamical context.

14Toward practical predictive computation

The most direct computational use of this mathematics is a UNNS Certified Dynamics Analyzer. It would accept an initial rational state, calculate its structural invariants — the fibre, L, G and γ — apply the proved carry and scaling theorems, and return a mathematical certificate of eventual behaviour where the hypotheses permit: the surviving prime, the period, the scale factor ρ², and an explicit bound on the transient.

That is a pathway from pure mathematical theory to usable software, and the dashboard embedded above is a first step in that direction: it already exposes the registry, the machine checks and the bound computations in one inspectable place.

The wider objective

Behind the analyzer sits a broader question, the programme of Structural Complexity Reduction: whether structural constraints of this kind can reduce search spaces in independently defined computational problems. The status of this question should be stated precisely. The mathematics establishes certification within the selected UNNS system. Measurable computational advantage, and transfer to systems defined independently of UNNS, remain research questions — now with a rigorous base to be tested against.

15What has been established

The mathematical foundations of the UNNS Substrate now form a coherent development from generative arithmetic to structural reachability, interaction constraints, collective dynamics, exact recurrence, prime rigidity, and finite prediction.

The significance of this progression is that structure is no longer merely described. Within explicitly defined generative systems, its organisation can be derived, its restrictions proved, its eventual behaviour characterised, and aspects of its future certified.

The next challenge is to determine how far these mathematical principles extend beyond the systems in which they have been established.

The programme remains open to further grammar development, independent evaluation and practical investigation. The volumes, the registry and the checks are public precisely so that all three can proceed.