Skip to content

00. Prologue: Euclid's engine, twins, Riemann and the classical fronts — one rank-parity programme

↑ Repository · 01. The engine (EPMI) →

This is the entry file of the whole programme — the prologue and chief navigator. Beyond it come the numbered chapters prose/NN_*.md (paired prose, 00→56) and the modules EuclidsPath/Engine/*.lean (machine verification). Here we introduce the object, declare the strategy, draw the map of parts I–VIII and — above all — honestly record what exactly is machine-proven, what is conditional on the single axiom, and what remains open. Status legend: 🟢 — machine-proven under the standard Lean/mathlib axioms; 🟡 — AXIOM-TAINTED: conditional on the repository's single axiom step00FirstCause (the first cause 0 → 1 with ONE boundary — the twin node: postulating it = postulating exactly the twins; Riemann, Navier–Stokes and P/NP were detached from the decree (Option A, chapters 38/41/39, machine-checked in Engine/Step00FrontClosureAudit), and the fourth, Collatz boundary was taken and WITHDRAWN after the machine refutation of its law — see chapters 55–56; 16 such declarations, all asserting the twins, with no leaks — the verifier tracks every one at each build); 🔴 — an open node / goal. The goal itself, twin_prime_conjecture, remains sorry. Nothing stronger than machine facts is claimed here. All key notions of the programme are gathered in the glossary.

★ The programme's main theorem

At the summit of the whole construction stands a single statement — higherEnergyIncompatibility_main (Engine/FiniteKnowledgeBarrier), the higher energy incompatibility. Its core is entirely green, with no axiom at all.

Theorem 0.1 (higherEnergyIncompatibility_main). The five faces meet in a single conjunction:

\[ \bigl(\mathsf{KnowsCause}\Rightarrow\mathsf{Engine}\bigr)\ \wedge\ \neg\,\mathsf{KnowsCause}\ \wedge\ W_{\mathrm{fin}}\ \wedge\ W_{\infty}\ \wedge\ \bigl(\neg\,\mathsf{Engine}\wedge\mathsf{Step00}\Rightarrow \#\{\text{twin lowers}\}=\infty\bigr).\tag{0.1} \]

Here \(\mathsf{KnowsCause}\Rightarrow\mathsf{Engine}\) — internal knowledge of the first cause builds a concrete Euclidean engine; \(\neg\,\mathsf{KnowsCause}\) — the cause is unknowable from inside (cause_unknowable); \(W_{\mathrm{fin}}\) — for any sieving system \(S\), scale \(A\), certificate and class \(B\): if \(B\) is equivalent to an element \(\mathit{bad}\) that is not a twin, the system does not certify \(B\) as a twin; \(W_{\infty}\) — if arbitrarily far out every class contains a non-twin, the system does not certify the infinitude of twins; and the fifth face: the absence of an engine together with the causal boundary step00FirstCause entails TwinLowers.Infinite. The first four faces are 🟢; the conclusion of the fifth is 🟡 (conditional on the boundary).

It rests on a simple thought: learning the cause from inside costs energy that a closed system does not have.

The thought unfolds as follows. To know the first cause from inside would mean deriving it from inside — and that would build a perpetual engine, which does not exist. Hence the cause is unknowable in principle.

On the other side stands a second, independent wall. A finite sieving observer knows about a twin only if its entire finite class consists of twins; a twin inside a mixed class is invisible to it. Hence the infinitude of twins, too, cannot be certified from inside.

The two walls — the causal and the epistemic — turn out to be of one nature (two_walls_one_nature): from inside a finite system one can see neither its first cause nor the infinitude of twins. And then the load-bearing, fifth face closes the circle: this very incompatibility (no engines = the cause cannot be learnt), together with the accepted causal boundary, entails the infinitude of twins.

Read as a whole, it is priced energetically: "knowing from inside" costs a perpetual engine, of which there is none; while the infinitude of twins is external knowledge, paid for by the first cause.

Most briefly the theorem reads: one cannot know that the twins are infinite; but if the unknowability of the first cause is accepted as truth — they are infinite, and rigorously so. This is the corollary below.

Corollary 0.2 (higherEnergyIncompatibility_twins). Instantiating the fifth face of (0.1) by the decree step00FirstCause yields the form-unconditional

\[ \#\{\text{twin lowers}\}=\infty \qquad(\texttt{TwinLowers.Infinite}).\tag{0.2} \]

🟡: the fifth face, instantiated by the decree step00FirstCause (a decree is the intentional acceptance of a law by axiom rather than by proof; see the glossary). The core above it stays green, but the corollary itself is conditional — it is not a proof of the twin conjecture.

And the whole construction carries a rigorous cosmological reading — the theorems are rigorous, the cosmology is merely their translation (the full step-by-step account is in chapter 33).

The number line and the impossibility of the engine together encode space and time. The strict order of traversal is space with a singularity floor at 0, while the irreversible arrow of time is proven rigorously (engine_never_returns: height is strictly antitone, there is no way back).

The first cause is the instant when the engine emerges from the singularity 0 and sets off forward without turning. It cannot be supplied from inside, for self-ignition would be a perpetual engine (no_internalisedOriginEvent) — hence it is accepted from outside, by the single axiom.

That is why, inside the system, the infinitude of twins can be neither proven nor refuted: either act would mean building a perpetual engine. And there is no tower of universes beneath the first cause — the regress of causes is well-founded (no_rankedMetaFractalBranch); the universe is one.

1. The object: Euclid's engine as well-founded multiplicative descent

The whole edifice stands on one elementary object: the "state" of the Euclidean decomposition process reduces to its height — the index \(m\) of the centre of the pair \((6m-1,\,6m+1)\); a pure descent step decreases the height multiplicatively: \(\mathrm{DescentStep}(A,h,h') :\Leftrightarrow A\cdot h' < h\) (DescentStep, Engine/EPMI).

Euclid's engine is a hypothetical infinite sequence of heights in which every step is of this kind. It does not exist: \(H(t)+t\) does not increase, and a strictly decreasing chain of natural heights breaks off — this is Fermat's infinite descent, rewritten multiplicatively.

Theorem 0.3 (no_infinite_descent). Impossibility of the perpetual engine: a strictly decreasing chain of natural heights breaks off (\(H(t)+t\) does not increase). The structural form no_perpetual_engine, the step dichotomy boundary_dichotomy — all 🟢, without sorry, on the bare Lean 4 kernel (without even mathlib). This is the hardest stone of the foundation: any construction presented as a perpetual engine is automatically false.

2. The strategy: contraposition through the engine

For every goal \(P\) (infinitude of twins; Riemann; the fronts) one builds the bridge \(\neg P \Rightarrow \mathsf{Engine}\) and closes it with the proven EPMI: \((\neg P \Rightarrow \mathsf{Engine}) \land \neg\,\mathsf{Engine} \Rightarrow P\).

For the twins: finiteness of twins (NoNewTwinAbove M0) drives the infinite stream of pure starts into a finite set of old absorbers — a rigid cycle, an engine, a contradiction. For Riemann: a zero off the critical line yields a free directed transfer of mass — again an engine.

We do not write twins ⟹ RH: what the branches share is not an implication but a mechanism — one EPMI, one contraposition and, as §4 will show, one rank-parity invariant.

3. Map of parts I–VIII

Navigator map: one prohibition → the main theorem (pressing on the axiom) → the axiom and its boundaries; every problem with its honest status — taken 🟡 / deferred ⏸️ / fallen ✗.

Navigator map of the programme

I. The engine and its laws (ch. 01–09, 🟢). The impossibility core (Engine/EPMI, ch. 01); the carrier of two no_large_shared_divisor (Engine/Carrier, ch. 02); the 1st law \(XY-ZW=2\)det_law_rank33 (Engine/TwoGap, ch. 03); descent and the boundary law (Engine/Descent, ch. 04); irreversibility, turn ⇒ halt (Engine/Irreversibility, ch. 05); the vanishing of the diagonal (Engine/NoBackward, ch. 06); squeeze, the bounded cycle, factor-repeat rigidity (Engine/Squeeze, Engine/BK, Engine/Cycle, ch. 07–09).

II. Reduction to twins (ch. 10–11, 🟢). infinite_of_unbounded_centers (Engine/NonCover, ch. 10); twin_prime_conjecture_of_blocks (Engine/TwoTransport, ch. 11).

III. Attack lines and the parity wall (ch. 12–17, 🟢 conditionally/model-level). Four-corner (Engine/FourCorner, ch. 12); the model layer (Engine/ModelFourCorner, ch. 13); decomposition of the remainder (Engine/RealFourCorner, ch. 14); the chain to twins (Engine/ToTwins, ch. 15); finite ∧ H ⇒ False (Engine/FiniteContradiction, ch. 16); the ledger (Engine/PaymentLedger, ch. 17).

IV. The final reduction of the twin line (ch. 18–25, 🟢 + a 🔴 node). twin_primes_of_SNOL (Engine/SNOL, ch. 18); old-peel (Engine/OldPeel, ch. 19); NOPSL (ch. 20); regeneration_dichotomy (Engine/Regeneration, ch. 21); residuals/clean-graph (ch. 22–23); the boundary decomposition and the final twin node TheLastStep00Obligation (Engine/BoundaryDecomp, Engine/ConcreteStep00Graph and their retinue, ch. 24); the rigid closure reaches_twin (Engine/RigidClose, ch. 25).

V. Product-core rank descent (ch. 26–29, 🟢 + a 🔴 node). no_productHall (Engine/SeparatingScale, ch. 26); product_core_engine_of_carrier (Engine/ProductCore, ch. 27); factor_rank_le_four (Engine/MkNode, ch. 28); cleanCenters_infinite, engine_of_factorization (Engine/CarrierBridge, ch. 29).

VI. The Riemann branch and the rank-parity bridge (ch. 30–32). riemannHypothesis_of_engine_bridge, not_RH_gives_engine (Engine/RiemannBranch, Engine/RiemannEngine, Engine/RiemannImpossibleEngine(-Off), Engine/RankJumpBridge, ch. 30); liouville_eq_neg_one_pow_rank, riemann_of_liouville_bound (Engine/RiemannLiouville, ch. 31); the single rank-parity node — the epilogue conjecture (ch. 32).

VII. The first cause and the finite-knowledge barrier (ch. 33, written in parallel). The quarantined axiom step00FirstCause and its theorem retinue: step00CausalClosure (now a theorem from the first cause), the honesty step00FirstCause_iff_causalClosure (the markers carry True, all the strength sits in the boundary), the remainder equivalence nonAxiomaticRemainingObligation_iff_lastStep00Obligation (Engine/CausalClosureAxiom); the epistemics — cause_unknowable, two_walls_one_nature and the main theorem higherEnergyIncompatibility_main (Engine/FiniteKnowledgeBarrier) — see the ★ block above.

VIII. The classical fronts (ch. 34–37, written in parallel).Mersenne (ch. 34): the identity \(M_p = 6c+1\)mersenne_eq_sixCenter_add_one, the conditional bridge twinLowersInfinite_of_mersenneTwins, the honest gate noTwinsToMersenneImplicationClaimed (Engine/MersenneBranch); the payment conflict — soundness_forbids_mersennePrimePaymentConflict, twinLowersInfinite_of_infiniteMersenneSupply (Engine/MersennePaymentConflict); peel pressure — mersenneCenter_base4PeelStep, absence_forces_peelCoverage_or_paymentLaw_defect (Engine/MersennePeelPressure); the forward front of 34 bricks (Engine/MersenneForwardFront) — ⚠️ the late noEngine packages are uninhabited (vacuity no. 3, see §5). — P vs NP, locally (ch. 35): verificationEasy_always, localPSuccess_iff_semanticFlowLedgerCollisionResolves, localP_success_detects_twin, the unconditional small-scale incompressibility concrete_localSearchIncompressible_smallScale (Engine/LocalPNPNode); the classical bridge — genealogyLanguage_in_NP, classicalSeparation_of_localIncompressible, the scope gate bridgeScopeGuard_ok (Engine/ClassicalPNPBridge); the canonical self-reduction — extracts_local_success_of_selfReduction, the anti-vacuity gate trivialFrame_not_faithful (Engine/CanonicalSelfReduction); the front's routes — extracts_local_success_of_bitwiseEncoding (Engine/ClassicalFrontierRoutes); the rank discipline — no_rankBoundaryEngine, taxonomyExpansion_not_strictProgress (Engine/RankClosureFront). — Navier–Stokes (ch. 36): ns_no_infinite_dissipative_cascade, ns_no_infinite_dissipative_cascade_of_balance, kineticEnergy_nonneg (Engine/NavierStokes, the cascade skeleton — Engine/DissipativeCascade) — the structural "no infinite dissipative cascade", not the Clay regularity problem. — Riemann fronts (ch. 37): the entrance is closed — trivialBelowZeroClassification 🟢 (Engine/RiemannTrivialZeros, mathlib's functional equation): RH is conditional only on EngineBridge (riemannHypothesis_of_engineBridge_only) or TwoTransportBridge; the rank-projection route and its exposure (Engine/RiemannRankProjection, Engine/RiemannRankProjectionAudit — vacuity no. 2, see §5); the two-transport form and its honesty coherentTwoTransportBridge_iff_RH (Engine/RiemannTwoTransportFront); the arithmetic atom \(+2\)riemannHypothesis_of_arithmeticTwoTransport (Engine/RiemannArithmeticTwoTransport); the spectral audits — front_pair_iff_RH, no_single_atom_anchors_two_distinct_invariants (Engine/RiemannSpectralAnchorAudit), the origin-blind firewall no_identity_with_residues_555_111_plus_two (Engine/RiemannLayerBoxFront), the terminal rank front — no_free_origin_for_distinct_zeros, the checklist of 66 lines / 11 balances (Engine/RiemannTerminalRankFront).

4. The single rank-parity node

Both original conjectures reduce to one invariant — the parity of the rank \(\mathrm{rank}(n)=\Omega(n)\).

The twin side: the exclusivity of two (no_large_shared_divisor) forbids the cross term \(xy\) in the rank generating function and forces the four-corner inequality \(N_{00}N_{33}\le N_{03}N_{30}\) (N33_lt_N00_of_four_corner, Engine/FourCorner) — the balance of twins is a rank-parity balance.

The Riemann side: \(\lambda(n)=(-1)^{\mathrm{rank}(n)}\) (liouville_eq_neg_one_pow_rank), deleteFactor flips the sign (liouville_flip_of_mul_prime), and RH is the smallness of \(L(x)=\sum_{n\le x}(-1)^{\mathrm{rank}(n)}\) (LiouvilleBound).

Conclusion. One invariant, one operator, one parity wall — which is why we speak of one rank-parity node. The fronts of part VIII are attempts to get around this wall from different sides; their audits (ch. 37) show machine-wise where the bypass is genuine and where it is a repackaging of RH.

5. Honest status by branch

Twins — 🔴 a single node, narrowed to the utmost. The whole branch is machine-reduced to TheLastStep00Obligation (Engine/ConcreteStep00Graph): twinLowersInfinite_of_lastStep00Obligation 🟢. The node is a twin detector: twin_above_of_resolves (an input at scale M0 itself presents a twin above M0 — scale for scale it is no weaker than the goal); the whole family of ~15 equivalent forms (energy / nested / seam / gauge / compression …) is machine-⟺ to it.

⚠️ Narrowing (an adversarial probe): the branch A ≤ 4 is refuted — the 5-adic chain \(c(k{+}1)=5c(k)+1\) yields infinitely many admissible genealogies with no twin hypotheses (smallScale_branch_of_lastStep00Obligation_refuted); ∃A survives only at A ≥ 5 (lastStep00Obligation_forces_scale_ge_five).

⚠️ Vacuity no. 1 (exposed, repaired): the degenerate peel p = p·1 made the node refutable; the patch properDiv + targetPos — now the twin hypothesis is load-bearing, and satisfiability at A ≥ 5 is genuinely open. The conditional closure through the first cause (higherEnergyIncompatibility_twins and kin) — 🟡. Step00.twin_prime_conjecture remains sorry.

Riemann — 🔴 the fronts' inputs, 🟡 through the extended decree. 🟢: trivialBelowZeroClassification (every zero with Re ≤ 0 is trivial) — RH is conditional only on EngineBridge/TwoTransportBridge/LiouvilleBound.

⚠️ Vacuity no. 2 (exposed, not embellished): the goal of the rank-jump route is not anchored — the full package LiouvilleToTwinLocalization is inhabited with zero input (fullLocalization_noInput); the honest wall is exactly ¬LiouvilleViolation (wall_global) — of RH strength; the window bookkeeping, meanwhile, is closed honestly (relevantViolation_gives_window). The machine honesty of the bridges: offCriticalBridge_iff_RH, coherentTwoTransportBridge_iff_RH, no_coherent_twoTransportLaw — the decompositions are maps of obligations, not a non-circular path.

✳️ Through the first cause (chapter 38): RH is run through the same rank machine as the twins — a zero off the line = an unpaid deviation; the manifestation law (the Riemann boundary, subsequently withdrawn from the decree, Option A — a detached front, not a live decree boundary; manifestation is the principle "a deviation must show itself", see the glossary) + the green impossibility of supply at permitted scales ⟹ riemannHypothesis_from_firstCause (conditional, WITHDRAWN dead code). Not a proof of RH; the price is disclosed: riemannManifestation_asserts_RH — under the boundary the law is ⟺ RH.

P vs NP — a local architecture; the separation in the rank model is a 🟢 theorem (chapter 39). 🟢: verification is always easy (verificationEasy_always), local success ⟺ the semantic twin node (localPSuccess_iff_semanticFlowLedgerCollisionResolves), at the small scale A ≤ 4 incompressibility is unconditional (concrete_localSearchIncompressible_smallScale).

✳️ The reading "NP = full payment of rank certificates, P = rank-fast passage" is a theorem: pnp_rank_separation_smallScale 🟢 and concrete_localPSuccess_iff_fullPayment 🟢 (chapter 39); the split across scales — the decree at A ≥ 5 itself pays all the certificates (decreedScale_fullPayment 🟡); there is machine-checkedly no P/NP boundary of the decree (the trilemma — the mandatory three-branch test of a boundary candidate, see the glossary: two candidates are refutable, the third is vacuous). The decree's only live boundary is the twins.

⚠️ Vacuity no. 4: the decider-gated extraction fronts are classically empty (PDecider is free) — their separation conclusions are vacuous. The classical separation (classicalSeparation_of_localIncompressible) is conditional on the InP-gated bridge; the frames are plastic (allPFrame/constantsFrame) — this is not a proof of P ≠ NP, and it is not claimed to be.

Navier–Stokes — a structural result, not Clay. 🟢: ns_no_infinite_dissipative_cascade — under energy balance an infinite dissipative cascade is impossible (the same well-founded EPMI mechanism in \(\mathbb{R}_{\ge0}\)). Nothing is claimed about global regularity of smooth solutions.

✳️ Smoothness through the cascade + the integral taken (chapter 41): a singularity = a singular cascade = a perpetual engine; noSingularCascade_of_energyBalance 🟢 — there are no blow-up points of cascade type (the surrogate ≠ C^∞ — disclosed); the integral is taken — energy_identity_of_energyBalance 🟢 (the identity as an equality) and isNSSolution_integral_form 🟢 (the mild form, an E3-valued FTC). There is no Navier–Stokes boundary of the decree — the trilemma of chapter 41 (cookedFlow / the zero solution / a forged profile cascade).

Yang–Mills — a structural result, not Clay (chapter 40). 🟢: masslessness = a perpetual engine (the halving ladder — the very ℝ-counterexample of the cascade warning); quantization of the spectrum ⟹ a mass gap (massGap_of_quantizationLaw: ladder + rank = ℕ-descent, killed by EPMI).

There is no Yang–Mills boundary of the decree — the trilemma is machine-checked (the universal is refutable, the existential is vacuous, the Riemann mirror is incompatible with the accepted boundary — the ladder is presentable, unlike the zero; the collapse quantizationLaw_iff_massGap — the law is ⟺ the gap, green without the boundary). The decree's world is gapped in the language of supplies (decreedScale_no_deviationSupply 🟡). The 🔴 input — a data-anchored constructed YM spectrum.

Hodge — a structural result, not Clay (chapter 42). 🟢: a Hodge class = a quantized (p,p)-charge (an ℕ-height), a cycle = a payment; the engine (a chain of unpaid descent) is dead unconditionally (isEmpty_unpaidDescentChain — quantization is built into the model, the asymmetry with YM is disclosed); hodgeProperty_of_descentLaw — the descent law ⟹ the model's conjecture (strong induction on height).

The collapse descentLaw_iff_hodgeProperty (the converse side is vacuous — disclosed) — there is no Hodge boundary of the decree; the trilemma of chapter 42: the universal is refutable (cookedUnpaid), the existential is proven with no axioms at all, the chain form degenerates into the green one (V2′), manifestation over a presentable class is incompatible with the boundary. mathlib has no Hodge theory; the 🔴 input — DescentLaw for genuine (p,p)-classes.

Mersenne — conditional bridges + vacuity no. 3. 🟢: the arithmetic of centres, the conditional export twinLowersInfinite_of_mersenneTwins and the payment-peel defect dichotomies.

⚠️ Vacuity no. 3 (exposed, recorded in the module header): in Engine/MersenneForwardFront the late noEngine packages (the NoForbiddenPrimePaymentEngine family and kin) are uninhabited — the tokens carry a free field witness : Prop, the "engine" is built trivially, the headline conclusions of these bricks are vacuous; the routes in this form cannot be instantiated. The branch has no unconditional strong conclusions.

First cause — 🟡 quarantine, ONE boundary. The single axiom step00FirstCause carries only the twin node (causalBoundary): postulating it = postulating exactly the twins (the two-sided, non-vacuous equivalence of the boundary and the conjecture, machine-checked in Engine/Step00FrontClosureAudit). Riemann (chapter 38), Navier–Stokes (chapter 41) and P/NP (chapter 39) once projected onto the decree, but were detached (Option A): they lack a two-sided boundary equivalent to their conjecture, so they were withdrawn from the axiom and live as honest green/red conditional fronts — not results. The fourth, Collatz boundary was taken and WITHDRAWN after the machine refutation of its law (ropeLaw_universal_refuted, witness n = 27; the tripwire fired, the decree overpaid into falsehood — chapter 56). 16 AXIOM-TAINTED declarations, all asserting the twins (13 in the generated flow formulation + the corollary higherEnergyIncompatibility_twins + 2 of the chapter 49 geometry through the same twin boundary); the verifier recounts every one at each build. That the P/NP, Yang–Mills and Hodge boundaries are NOT taken is machine-checked (the trilemmas of chapters 39–42). Internalisation of the first cause is impossible — an engine (no_internalSelfDerivation_step00CausalClosure, 🟢).

The honest bottom line. The 🟢 corpus is a genuinely verified machine: the engine, the reductions, the audits, the main theorem (its core). The 🟡 layer is exactly what is paid for by the first cause, and it is fenced off by the quarantine. 🔴 — TheLastStep00Obligation at A ≥ 5, the RH inputs, the classical fronts. A reduction is not a proof; a sorry cannot be faked.

Where we are and where next

We have introduced the object (no_infinite_descent), declared the strategy (contraposition through the engine), charted parts I–VIII and set higherEnergyIncompatibility_main at the top of the map: knowledge from inside costs a perpetual engine, which does not exist. The next chapter, 01. EPMI, builds the foundation literally — on the bare Lean kernel; chapters 33–42 unfold the first cause (one boundary — the twins; Riemann and the NS gate were detached, Option A; the fate of the Collatz one — chapters 55–56) and the classical fronts.

Philosophical digression: one physical law beneath seven problems

It is worth spelling out what binds the whole programme together, because that bond is not a literary device but its load-bearing structure. At the foundation lies the hardest of physical laws — the impossibility of a perpetual engine: one cannot draw work out of nothing, one cannot descend forever without paying. In Euclid–Fermat form this is no_infinite_descent: a strictly decreasing chain of natural heights breaks off.

And the whole programme is the discovery that the seven great questions are one and the same prohibition, worn by different objects:

  • Twins: finiteness of twins would drive the infinite stream of pure starts into a finite cage — a rigid cycle, an engine.
  • Riemann: a zero off the line is an unstable mode of a hidden quantum spectrum (Hilbert–Pólya), a self-amplifying oscillation, an engine.
  • Yang–Mills: a massless spectrum is a tower of arbitrarily cheap excitations of the vacuum, a gratuitous transfer, an engine; and the mass gap is the fee the vacuum charges for the first excitation.
  • Navier–Stokes: a singularity is a Kolmogorov cascade that has reached infinitely small scales in finite time, infinite energy out of finite, an engine; and viscosity is the rope that always arrives in time at the bottom.
  • Hodge: an unpaid quantized charge is an eternal regress of ever smaller denominators, an engine on the integer lattice.
  • P/NP: full payment of all certificates costs more than rank-fast passage — the thermodynamic price of information.
  • Collatz: a damped drift whose mean is reconciled and whose floor is absorbing; a counterexample = a perpetual engine (an orbit tail from the minimum), and it cannot be decided from inside — only by checking; the decree on the rope law was taken and withdrawn after the machine refutation (chapters 55–56).

One root, seven branches. And in each of them — the same honest wall: the structural half is proven green (where the engine prohibition holds, there is no deviation), while the last step — attachment to the genuine object (a QFT spectrum, a Hamiltonian of the primes, a Leray solution, (p,p)-classes, a Turing machine) — is either accepted by the decree of the first cause under an honestly disclosed price, or remains a 🔴 input.

Section takeaway. The programme does not solve the millennium problems; it shows that all of them are shadows of one physical "cannot", and presents exactly the part that does not depend on the concrete object.

Mathematics here turns out to be thermodynamics rewritten in the language of ranks: where the physicist says "there is no free energy", the formalist proves no_infinite_descent — and this alone suffices to come right up to the boundary of the knowable seven times over.

Beyond the seven branches comes the arithmetic zoo (chapters 4448, all 🟢): the same manifestation apparatus, run through Polignac's cousins and sexy primes, Sophie Germain, Goldbach, Legendre, perfect numbers and Fermat numbers. No decree fields are taken there — intentionally (§17): the trilemmas are passed, but an honest boundary also demands a stake one need not be ashamed of.

One theorem fell out green and unconditional — the Euler–Lagrange pearl: Sophie Germain primes with p ≡ 3 (mod 4) divide Mersenne numbers. And the geometry of the path (chapter 49) reads the descent graph itself as curved spacetime: an arrow of time, computed curvature, and a violation of Euclid's second postulate.

The eighth mask — Birch–Swinnerton-Dyer (chapter 53) — enters differently: here the engine prohibition does not guard a deviation but serves as a method — Fermat's infinite descent proves finiteness of the rank (Mordell–Weil, on a real mathlib curve), while the parity of the rank turns out to be the same rank-parity node that stands behind Riemann. The analytic bridge (rank = ord L), meanwhile, is honestly 🔴; no boundary is added (the trilemma) — BSD does not raise the tariff.

The full summation of this thought — where the single prohibition is read as a possible structure of spacetime, up to a conjecture of a theory of everything — is carried out to the concluding coda (chapter 50).


↑ Repository · 01. The engine (EPMI) →