Continental Genesis: why dry land exists

§1.1The origin question

Mainstream geology explains how the two-tier crust is recycled, but not cleanly why it exists in the first place — why there is light, emergent dry land at all, sitting above heavy ocean floor. This volume asks that origin question under a firewall, and answers it with composition + buoyancy, not age.

Begin from a water-covered world with only a thin mantle skin and no continents. A rift opens; its centre founders into a deep basin floored by bare mantle skin; the displaced skin is pushed sideways, thickened and folded; the lateral push and burial heat the thickened, water-bearing pile; the wet pile partially melts, distilling a light felsic (granitic) fraction that rises, cools, and — being too light to sink back — accumulates as buoyant crust. Dry land is that distilled, accumulated felsic; the ocean is the basic mantle skin it stands above. Land need not rise; the basin subsides, and emergence follows by isostasy.

This reframes "why land" away from the clock and onto the mechanism that makes light crust — the hard part, which is composition.

§1.2The bidirectional chronology firewall (the core rule)

Every past-occurrence statement is [O]; every absolute date is RECORD, forbidden as load-bearing in both directions. The firewall is symmetric: it removes the mainstream's deep-time dates and withholds the convenient young numbers. Neither "billions of years" nor "recent" is load-bearing.

The theory's distinctive content, on every axis, reduces to rate / magnitude / timing / order, which the firewall caps at [O] both ways. Only present-tense, non-rate observables can be load-bearing. This is the constitution; where any later clause conflicts with it, the firewall wins.

§1.3The grading system

Processing degree D0–D5: D0 = raw observation (top); D0–D3 = load-bearing (ACTIVE); D4 (model output) and D5 (absolute chronology) = RECORD, never load-bearing.

Status grades, with precedence [F] > [V] > [L] > [O]:

• [F] — physically forced / permitted by measured constants (a screen result).
• [V] — observed / verified present-tense (measured, or engine-reproduced).
• [L] — leaning: a well-anchored inference, short of forced.
• [O] — open / undecided. Occurrence is capped at [O] forever.

Nothing downstream may launder an [O] into a higher grade. No external interpretation may settle the meaning of an observation — a favoured and a disfavoured interpretation of the same observation get the same grade. Every datum is recomputed from the primary source, not recalled; only the recomputed quantity is [V]. The rule binds the inconvenient direction as well as the welcome one.

Reproducibility is enforced: no fitted parameter in any Tier-A object; SEED = 19 convention; determinism proven by double-SHA-256 self-gates; changing any LOCK constant defines a new version; corrections mark, never erase.

§1.4The specific failure mode this volume must not repeat

A tempting but forbidden move would be to positively refute the 4 Ga zircon ages (thermal reset → "young") to make recent continent-formation load-bearing. That would be the canonical error this firewall forbids — attacking a firewalled date to privilege the convenient direction. This volume does not do that. It holds the U-Pb ages as RECORD, routes when/how-fast to [O], and load-bears only on present-tense mechanism screens. The zircon understanding here is a caveat (isochron non-uniqueness, §8), never a refutation.

§2.1The inherited results, carried in full

The argument uses — and ships — five upstream results so the package is self-contained without Zenodo:

• I1 — c² = B/ρ marginal-substrate fluid. At the unjamming margin the relaxed shear modulus → 0 while the bulk modulus B stays finite; the medium is a fluid by arrangement (the same substrate as the hot mantle, §12).
• I2 — R19 unjamming switch. The void-suction / unjamming-rupture bistable switch: bulk jammed, local unjamming where the threshold is crossed.
• I3 — rotation → three-body C₃ core. Odd-cycle (C₃) frustration forces co-rotation.
• I4 — co-rotation → merger → Ekman through-flow. Co-rotating vortices merge; a coherent rotation drives a convergent Ekman inflow; inflow and updraft are one loop (§18).
• I5 — the firewall. The bidirectional chronology firewall (§1), inherited as the constitution.

The fluid-dynamics foundational PDF and four runnable inherited scripts (marginal_fluidity.py, unjam_inflow.py, axioms.py, corotation.py) are shipped under inherited/; each runs offline and reproduces its inherited mechanism. The physics (DOI 10.5281/zenodo.17932566) and Atlantic-geodynamics (DOI 10.5281/zenodo.17978934) source packages remain external, their load-bearing results stated in inherited/INHERITED_RESULTS_CARD.md.

§9.1The observation (present-tense, [V])

On NW-Pacific bathymetry the deep ocean floor is smooth, while crumpling (folded/thrust relief) is concentrated on and around the felsic-loaded crust — even the marginal seas. The author's reading: the deep floor is not "less active"; it is the bare, hard bottom — there is nothing on it to crumple. This module makes that precise.

§9.2The mechanism — a buoyancy gate ([F])

When a lithospheric column is compressed, its response is gated by its net buoyancy relative to the asthenosphere:

B = Σ_i (ρ_asth − ρ_i) · h_i (>0 = floats / resists subduction).

(B in 10⁶ kg/m²; F_b = g·B in MPa, upward positive.)

• Old ocean skin is net-NEGATIVE (its cold dense lithospheric root beats the thin light basalt crust). Under compression it shifts the load downward — it subducts — rather than building thick relief. Its interior stays smooth. This is the smooth deep-Pacific floor.
• Felsic crust is net-strongly-POSITIVE (thick low-density crust dominates; the cratonic mantle root is ~isopycnic). It cannot subduct, so compression must crumple/thicken it. Crumpling is gated to exactly where felsic is loaded — the land.

Two present-tense reasons the skin stays smooth, pointing the same way: it (a) sheds compression by subduction (this module), and (b) is too thin to buckle into thick relief anyway (Module 04 — buckling threshold rises steeply as the plate thins).

§9.3What this adds — internal economy (the honest value)

The same composition/buoyancy engine that distils continents (M05 wet re-melt → felsic; M06 buoyant accumulation) now also explains where crumpling happens (M04) and why the bare floor is smooth (here). One engine, three facts — distillation, emergence, and the crumple/smooth asymmetry all fall out of "light felsic floats; thin cold skin sinks."

§9.4Honest grade & scope

• The buoyancy gate is [F] physics; the smooth/crumpled contrast is [V] observation.
• DEGENERATE as a discriminator: "continental buoyancy resists subduction" is also mainstream — so this does not uniquely favor the genesis thesis over standard tectonics. Its worth is internal unification, not a new discriminator. (Recorded honestly per the PUZZLE_MAP discipline: do not file this as evidence for the thesis.)
• Rate / when of any compression stays [O] both ways — unchanged.
• Earlier-turn correction logged (mark-never-erase): the smooth/rough contrast was first read as "proximity to active process + sediment"; the author's buoyancy-gate reading is deeper and is adopted here as the load-bearing mechanism, with the proximity/sediment factors retained as secondary.

Ledger: CG-22 (see BLUEPRINT_and_GRADED_LEDGER.md).

§10.1What was tested

CG-22 (M08): negative-buoyancy ocean skin sheds compression by subduction and stays smooth; only positive-buoyancy felsic crust crumples. The promised test: predict where ocean lithosphere flips net-negative, and check whether that boundary is the smooth/crumpled map boundary — a present-tense [V] test that could falsify the gate.

§10.2[1] The sign-flip — computed ([F])

Half-space cooling gives the net buoyancy B(age) = (ρ_m−ρ_cr)·h_cr − ρ_m·α·ΔT·2√(κt/π):

**Sign-flip t\* ≈ 11 Ma (sensitivity 7.5–15.4 Ma over reasonable crust/ΔT), depth ~3.7 km** — a small fraction of the ~180 Ma seafloor range. So most of the deep Pacific floor is older/deeper than t\* and is net-negative, consistent with M08.

§10.3[2] The honest finding — the proposed test MIS-MAPS

The sign-flip is the young-vs-old ocean boundary. But crumpling is gated by felsic LOAD (ocean vs continent), and both young and old ocean lack that load → neither crumples into thick relief. So "sign-flip boundary vs smooth/crumpled boundary" compares different axes. The smooth/crumpled boundary is the ocean/continent (load) boundary (M08), not the ocean buoyancy-sign boundary. Stating this honestly is the result — the test as proposed cannot confirm the gate, because it targets the wrong boundary.

§10.4[3] The real breaker — central Indian Ocean ([V] counter-case)

A documented present-tense falsifier of CG-22's absolute form: in the central Indian Ocean, old (~50–80 Ma, net-negative) oceanic lithosphere is **folding under intraplate compression without subducting — long-wavelength lithospheric folds (~100–300 km), gravity/geoid undulations, reverse faulting, large intraplate earthquakes. So negative-buoyancy bare skin can deform; it does not always shed compression by subduction. The strong claim is falsified**.

§10.5[4] The bounded, surviving claim (the refinement)

• THICK crumpling (mountain-scale relief) remains gated to felsic-loaded (positive-buoyancy) crust — this survives: there are no bare-ocean mountain belts.
• Bare ocean skin under compression either subducts (if a zone is available) or, where it cannot, folds only into long-wavelength, low-amplitude undulations (central Indian Ocean) — never thick relief (it is thin; M04 buckling threshold rises as the plate thins).

→ CG-22 is DOWNGRADED: [F] law → [L] bounded tendency, with a logged exception.

§10.6Honest ledger impact

• CG-22 refined to [L] (bounded tendency), not the [F] law M08 first stated — marked, not erased (M08's strong wording stands with this refinement appended).
• CG-23 added: the central Indian Ocean is the standing falsifier that bounds the gate, and a live test target (does the long-λ/low-amplitude limit hold wherever bare skin is compressed without a subduction outlet?).
• The proposed sharpening strengthened the program's credibility by breaking its own overreach — the firewall working as intended. The thesis's core (thick relief needs felsic load; the deep floor is bare negative-buoyancy skin) is unscathed.

§11.1The law ([F])

The maximum isostatically-compensated relief a crust can support is

h_max ≈ ΔH · (ρ_m − ρ_c) / (ρ_m − ρ_top) (ρ_top = ρ_water submarine, ≈0 subaerial),

where ΔH is the crustal thickening available. Tall compensated relief needs a deep, light root — i.e. thick felsic crust. Computed:

Normal 7 km basaltic skin, even doubled, supports only ~1 km of compensated relief. Felsic collision supports ~5 km (deep light root). Only special igneous over-thickening (LIP) lifts basaltic crust a few km, and even that stays below continental heights (basalt's smaller density contrast).

§11.2The present-tense [V] test set (verified by literature)

Verified anchors:

• Central Indian Ocean: undulation amplitude 1–2 km, wavelength 100–300 km (avg ~200 km), increasing northward with age (40→70 Ma) and flexural rigidity; mean lithospheric yield strength ~400 MPa to make ~1 km amplitude with ~60 km shortening; highest intraplate oceanic seismicity (M up to 8) (Weissel et al. 1980; Zuber 1987; Krishna et al.).
• Gorringe Bank: a 5000 m high massif of uplifted oceanic gabbro + serpentinized peridotite, geoid +9 m, Bouguer ~120 mGal, large positive free-air anomaly; topographic highs sit on shallow flat Moho and are not isostatically balanced — held by an active NW-vergent thrust (≥20 km slip) at the Azores–Gibraltar transpressional (incipient-convergent) boundary (Souriau 1984; Jiménez-Munt et al. 2010; Galindo-Zaldívar et al. 2003).

§11.3Why this matters (and the honest scope)

• Precision, not rough: the bounded claim is now a quantitative law with a numeric cap (~1 km for normal ocean skin) and a single present-tense discriminator — free-air gravity + Moho depth.
• Falsifier (live): a compensated multi-km bare-ocean high with a deep light root (small free-air anomaly, deep Moho, no felsic) would break the law. None is known → it stands, but the test is explicit and checkable everywhere.
• DEGENERATE as a discriminator: compensated-vs-uncompensated topography is textbook isostasy — so this does not newly favor the genesis thesis over mainstream. Its worth is (a) making CG-22/23 precise & falsifiable (the user's goal), and (b) internal economy: the same felsic-buoyancy that distils continents (M05–06) and gates crumpling (M08) also sets why only felsic crust carries compensated mountains. One engine, now four facts.
• Rate / when of any deformation stays [O] both ways.

§11.4Ledger impact

• CG-24 added: the isostatic-compensation law ([F]/[V]), with the free-air-gravity/Moho test and the verified census.
• CG-23 sharpened: the bounded tendency is now stated as this law; its live test is the gravity field; the central Indian Ocean (capped) and Gorringe Bank (uncompensated) are the two anchors that bound it.
• Earlier rough wording retained (mark-never-erase); this module is the precise replacement.

§12.1The refinement and its honest bound

The author's move: the "heavy fluid" that lets the bare deep floor flatten (CG-24) is, in jamming terms, hot — near magma, with only a thin solid skin. The kernel is correct; the degree is bounded by data.

KERNEL — confirmed [F] (jamming language fits). The mantle flows like a heavy fluid on long timescales because it is hot and sits close to its unjamming (melting) threshold — a jammed solid near the transition. Homologous temperature T/T_solidus ≈ 0.7–0.95 (adiabat ~1350 °C vs dry solidus ~1500 °C, upper mantle): hot, close to but below melting. That is why it convects, creeps, and lets the floor flatten. This is the R19 bistable switch exactly: bulk jammed, local unjamming where the threshold is crossed.

BOUND — present-tense [V] (seismology). It is not bulk magma:

1. S-waves. The mantle's shear modulus μ ≈ 10¹¹ Pa > 0, so it transmits S-waves through its whole depth to the core–mantle boundary. A liquid has μ = 0 and passes no S-waves — which is exactly how we know the outer core is liquid (the S-wave shadow). So the mantle is solid; the outer core is the liquid. "Bulk magma mantle" is falsified.
2. Viscosity. Mantle η ≈ 10²¹–10²² Pa·s vs magma η ≈ 10²–10⁶ Pa·s → ratio 10¹⁵–10¹⁹×. "Magma level" overshoots by ~15–20 orders.
3. Maxwell time τ = η/μ ≈ 300 yr (asthenosphere) → solid on short timescales (seismic seconds → S-waves) but fluid on long timescales (convection, Myr). The "heavy fluid" is the long-timescale face of a solid, not a liquid.

The "thin solid skin over fluid" is right — one level up. It maps onto lithosphere (the cold rigid ~100 km skin) over asthenosphere (hot, near-solidus, ~1 % partial melt, flowing), not solid-mantle over whole-mantle-magma. Magma is the local unjammed state (ridges by decompression, plumes by heat, subduction by water-lowered solidus), not the bulk.

So: instinct right (hot, near-unjamming, heavy-fluid, thin rigid skin); single correction = degree (near the threshold, solid that creeps — not past it, magma). The S-wave datum pins it.

§12.2Honest audit — does this "resolve many earlier problems"? Mixed.

The reframe must not be oversold (PUZZLE_MAP discipline). Checked against the earlier strains:

Net: the reframe genuinely clarifies the heavy-fluid behaviour and strengthens the engine's coherence, and partially helps the felsic-seed/scale problem. It does not resolve the ordered-pattern strains (S1–S3), and the strong "magma mantle" form is falsified by S-waves. So: real and load-bearing for one deep point, honestly limited elsewhere.

§12.3Ledger impact

• CG-25 (from the rigidity/density split): rigidity → wavelength (M04 buckling), density+heat → amplitude = no compensated mountains (M10/CG-24); active stress + rigidity → transient uncompensated relief (Gorringe). The amplitude/no-mountain answer is load-bearing on density (heavy fluid), not rigidity.
• CG-26 added: the mantle is a jammed solid near unjamming (T/T_solidus ~0.7–0.95), not magma — bounded by S-waves (μ>0, solid) and viscosity (10¹⁵–10¹⁹× magma); the skin-over-fluid is lithosphere/asthenosphere; magma is the local unjammed state. [F] kernel / [V] bound. Strain audit recorded honestly.
• Module 08 correction (mark-never-erase): its reason "(b) too thin to buckle" is a rigidity argument that sets wavelength, not the amplitude/no-mountain result; the load-bearing reason the bare floor builds no mountains is the heavy-fluid (density) + hot-near-unjamming state here. M08's wording stands with this correction appended.

§13.1The picture, clause by clause

The author's present-state synthesis: below circulates (and, near unjamming, is more mobile than a far-from-melting solid); the skin is cooled/solid; the felsic "oil-scum" gathered on one side; crumpling has largely happened; energy is supplied slowly yet the system is maintained. Checked against present-tense physics:

§13.2The honest verdict (and the one place to push)

• Clauses [1]–[3] are [F]/[V] coherent but DEGENERATE. They are, in plain terms, mantle convection + a rigid lid + an internal heat engine — textbook. So the synthesis is internally consistent, but it is not new evidence for the genesis thesis, and must not be filed as such (PUZZLE_MAP discipline). Its worth is coherence: the chain now has a consistent present state that follows from the inherited R19/near-unjamming engine (M11).
• Clause [4] is [V] in distribution, [O] in rate. The deformation is real and now largely frozen; how fast it formed stays open both ways — the firewall does not let "it's all crumpled now" become "it crumpled fast/once."
• Clause [5] — the hemispheric asymmetry — is the ONE non-degenerate EDGE. Mainstream has no settled "why" for the degree-1 (one-sided) concentration of continents. This is the PUZZLE_MAP #1 edge, and it is exactly what a single-rupture / one-sided-condensation picture (the thesis's natural shape) could speak to. This is where evidence is worth pursuing — not the convection clauses.

"Maintained" is right short-term; honestly it is a slowly-cooling quasi-steady state — so the steady-state framing holds on Myr but not on Gyr. Rate/when stays [O] throughout.

§13.3Ledger impact

• CG-27 added: the present-state synthesis (convecting interior [F], frozen lid [V], slow-but-supplied heat engine [F]/[V], quasi-steady/slowly-cooling). Coherent but DEGENERATE — internal coherence only, not evidence.
• CG-28 added: the hemispheric (degree-1) continental asymmetry — land hemisphere ~81 % of land, no settled mainstream "why" — as the thesis's #1 non-degenerate EDGE and the priority evidence target. [V] observable / [O] explanation.
• This module is the present-state capstone of the mechanism chain (M01–M11); it converts the picture into honest grades and points the next work at clause [5].

§14.1[1] The asymmetry, quantified ([V])

Continents are hemispherically concentrated: the land hemisphere is ~47 % land, the antipodal water hemisphere ~89 % ocean — a degree-1 concentration of ~4.3 : 1. Real, present-tense.

§14.2[2] But as a DISCRIMINATOR it is degenerate — and the strong reading is disfavored

The asymmetry was elevated to "#1 edge" (M12). Tested against data, it does not discriminate the single-rupture thesis from mainstream, and its strong "primordial permanent scar" reading leans against the evidence:

• Pacific = the persistent Panthalassa superocean — the Pacific is the direct continuation of the superocean that surrounded Pangaea. Mainstream explains the one-sidedness cleanly.
• Continents demonstrably reconfigure — Vaalbara → Ur → Kenorland → Columbia → Rodinia → Pannotia → Pangaea (the supercontinent cycle, ~300–500 Myr). The asymmetry's orientation has changed repeatedly → not a fixed scar.
• The asymmetry is FADING — Pangaea was ~⅓ of the surface in ~one hemisphere (more concentrated); today's spread is more even. The one-sidedness is a decreasing post-breakup snapshot, not a permanent feature.

→ The asymmetry fits both accounts via different histories (degenerate), and the permanent-scar reading is disfavored. CG-28 is downgraded: a genuine [V] observable, but not where the thesis diverges. (This corrects its prior elevation — the firewall catching an over-claim.)

§14.3[3] Where it actually diverges — the origin (the author is right)

The supercontinent cycle and Panthalassa assume continental crust exists; they describe how it moves and cycles, not why felsic crust exists or how the first crust formed. That is the genuine divergence:

• Mainstream: felsic crust grows at subduction arcs over ~4 Gyr (water-fluxed arc melting; continents accrete from arcs).
• Thesis: felsic distils from a rupture's wet-remelt face (M05).

Both use wet melting → the divergence is the setting and efficiency, and it is precisely CG-12.

§14.4[4] The present-tense test of the origin — and an honest headwind

The bulk continental crust is andesitic with a Nb–Ta depletion — the "arc signature." This is mainstream's principal present-tense evidence for arc growth, so the origin question, tested this way, currently leans mainstream. For the thesis to genuinely diverge it must either:

• (a) show a rupture/extension path that reproduces the arc signature at scale, or
• (b) find continent-scale felsic without that signature.

This ties directly to CG-12: if pure extension cannot distil continent-scale felsic (Iceland yields only minor rhyolite), the thesis needs a subduction-like deep-water path → it converges with mainstream; if it can, it diverges. CG-12 is the real pivot — not the asymmetry, not the map.

§14.5Honest verdict and ledger impact

• CG-28 downgraded: [V] observable, degenerate discriminator, strong reading disfavored. Removed from the evidence column; kept as an observable (mark-never-erase: M12's elevation stands with this correction).
• CG-29 added: the real divergence = the origin/composition question, precisely CG-12 (rupture vs arc felsic) plus the arc-geochemical-signature test (Nb–Ta depletion in bulk crust), which currently leans mainstream. The thesis's genuine, falsifiable edge lives here. [O] open / present-tense test defined.
• Net: the firewall demoted a hoped-for win (asymmetry) and relocated the real contest to the origin — exactly as the author framed it. The next work is CG-12, with the arc signature as the present-tense bar to clear.

§15.1The recurring breach

The firewall's core rule is: absolute dates and deep-time sequences are RECORD — forbidden as load-bearing in BOTH directions. My violation pattern was singular and consistent: I imported the CHRONOLOGICAL INTERPRETATION of a dataset (age progression, deep-time sequence, secular rate) and scored it as a present-tense fact — almost always to weigh AGAINST the thesis. Mainstream "evidence" is nearly always chronologically framed, and I kept reaching for that framing instead of stripping each dataset to its present-tense core first.

Crucial: the phenomena are never denied (the author's point: we don't deny the phenomena, we want the more fundamental). The present-tense pattern stays load-bearing; only the imported deep-time reading is demoted to RECORD/[O] — for and against.

§15.2The breaches and corrections

§15.3The corrections (mark-never-erase; the originals stand with these appended)

• S1–S3 corrected. The thesis must reproduce the present-tense geometry — the symmetric stripe pattern, the sediment gradient, the bent chain. These remain legitimate [V] constraints. The age progression is [O] both ways and is not a load-bearing strain. (So the "strains" are real but narrower than I stated — geometry, not chronology.)
• Arc signature corrected. The Nb–Ta depletion is a present-tense hydrous-melting signature — consistent with the thesis's wet-remelt path, not exclusive to arcs. Verdict: degenerate, not "leans mainstream." My "leans mainstream" overstated an interpretation as evidence. (CG-29 amended: the origin contest is degenerate on present-tense composition; CG-12 — felsic scale — remains the live pivot.)
• Asymmetry corrected. "Disfavored by the deep-time reconfiguration record" withdrawn — it rested on imported chronology. Only "degenerate" stands (the asymmetry cannot discriminate). So CG-28 is neutral [O], not evidence against the thesis. (CG-28 amended.)
• Secular cooling corrected. Present heat flow is [V]; the cooling rate and "early hotter" are [O] and demoted from any load-bearing role.
• S-waves corrected. Relabeled a model-mediated inference [F] (robust), not raw [V] data. The load-bearing content is the jamming state (T/T_solidus, viscosity), not the word "solid."

§15.4Net effect — honest

Several strains and the "leans mainstream" / "asymmetry disfavored" verdicts rested on imported chronology and are demoted to [O]. The effect on the thesis is that it is less penalized — not validated — on those axes: the firewall, applied to my own analysis, restores the [O] it always required. The thesis's genuine, present-tense contest is unchanged and narrow: CG-12 (can a rupture distil continent-scale felsic?) and reproducing the present-tense stripe/gradient/chain geometry — neither settled, both firewall-clean.

Meta-lesson (recorded for all future work): the firewall is hardest to keep exactly when importing "mainstream evidence," because that evidence arrives chronologically framed. Strip every dataset to its present-tense core before it touches the load-bearing column; the ages and sequences stay RECORD — for and against.

§15.5Ledger impact

• AUDIT-1 added: this self-audit; S1–S3, CG-28, CG-29, and the secular-cooling/S-wave claims amended per above (mark-never-erase).

§16.1The hard honesty first — what "simulation" can mean here

A time-stepped reconstruction is forbidden: it needs absolute ages and a deformation history — both barred (the firewall bars dates; the author's rule bars trusting geology to supply the history). So this is not a movie. What pure physics can assemble is a causal dependency chain: a sequence of entailments of the form "given present-tense facts X and conservation law L, Y is necessary." It is logical/physical necessity, not a timeline. Every step names its forcing law and grade; every place geology would be needed is flagged as a gap, not filled.

§16.2The chain (each step: forcing law → consequence → grade)

STEP 0 — substrate (present-tense, M02/M11). A thin basaltic skin (ρ=2900) over a hot mantle near its unjamming threshold (ρ=3300, flows on long timescales); buoyant felsic (ρ=2800) exists as a distinct light phase. [V] present-tense densities + rheology. (First felsic = gap, CG-12.)

STEP 1 — opening entails closing · forcing: mass/area conservation on a fixed-area sphere. The surface area is fixed (rigid radius): A = 4πR² = 5.101×10¹⁴ m². Any area opened must be exactly balanced by area closed elsewhere: dA_open + dA_close = 0 — an identity, not a model. An opening reaching ~10 % of the surface forces ~5.1×10¹³ m² of convergence on the complementary surface. Rate-free: at 2/5/10 cm/yr opening, closing matches at 2/5/10 cm/yr — independent of the value. GRADE: [F] forced by conservation.

STEP 2 — closing is absorbed by the skin, not the raft · forcing: buoyancy gate, CG-22/24. The convergence meets two materials; net buoyancy decides: the negatively-buoyant basaltic skin sinks (subducts), absorbing convergence; the positively-buoyant felsic raft cannot sink → it is swept/shortened, not consumed. The skin is the conveyor that is destroyed; the felsic is the flotsam carried toward, and piled at, the closing side. GRADE: [F] buoyancy (present-tense). (How much piles depends on supply = gap.)

STEP 3 — buoyant flotsam collects at the sink · forcing: kinematics of a non-subductable float on a subducting sheet. A float that cannot follow the sheet down accumulates where the sheet descends — as non-wetting scum collects at a drain, not by attraction but because the carrier leaves and the float cannot. → felsic concentrates on the closing side; one-sidedness is forced once an opening/closing pair exists. GRADE: [F] kinematic necessity. (The load-bearing link to the asymmetry.)

STEP 4 — collected rafts weld into one mass · forcing: incomplete — GAP. Physics forces collection (STEP 3) but not clean welding into a single rigid block: suturing, accretion, and join strength are material/contingent. Flagged gap: "they touch" is forced; "they become one continent" is not forced by conservation + buoyancy alone. GRADE: [O] — collection [F], welding not entailed.

STEP 5 — the assembly is provisional · forcing: the same conservation law, run either way. The area identity (STEP 1) has no preferred direction: an opening can later become a closing. A one-sided pile is a state the law permits, not a terminus — which is exactly why the present asymmetry is neutral/[O], not a fixed scar (consistent with the M14 audit: permanent-vs-cyclic is [O] both ways). GRADE: [F] reversibility; [O] on which phase "now" is.

§16.3What pure physics forces vs what stays open

Forced (the assembled chain):

opening ⟹ antipodal closing (conservation [F]) ⟹ skin subducts, raft cannot (buoyancy [F]) ⟹ raft collects at the sink (kinematics [F]) ⟹ a one-sided felsic mass is entailed by an opening/closing pair. [F]

Not forced — flagged gaps (geology NOT imported):

• the first felsic (origin) — [O] (CG-12/CG-20)
• the trigger / why this opening — [O]
• clean welding into one rigid continent — [O] (STEP 4)
• any rate, duration, order, or date — [O] (RECORD both ways)
• which phase (opening vs closing) is "now" — [O] (STEP 5)

§16.4Verdict

An opening/closing pair on a fixed-area sphere, acting on a buoyancy-gated skin+raft system, forces a one-sided felsic concentration — this much is pure physics [F], no geology, no clock. Everything historical (first seed, trigger, welding, rate, order) stays [O]. The result is a dependency graph, not a reconstruction — which is the most the data and pure physics permit, and exactly what the firewall allows.

Ledger: CG-30 (the assembly entailment chain).

§17.1Part A — "why one side?" inherited, and honestly bounded

The author's claim: any fluid has imbalance, so a leaning / one-sided outflow logic always occurs — inherit this.

What the fluid-dynamics volume actually locks (read, not assumed): a marginal jammed substrate is a fluid by arrangement — at the isostatic margin the relaxed shear modulus → 0 while the bulk modulus B stays finite (c² = B/ρ, [LOCK] there). "Marginal" means the medium is poised at the threshold with no restoring shear stiffness. This is the same substrate as M11 (mantle: hot, jammed, near unjamming) — so we inherit a substrate property, not a new mechanism. The volume does not anywhere establish that a fluid spontaneously leans to one global side; over-reading it that way would repeat the M14 import-error.

The honest split (firewall + M14 audit):

• Inherited [F] — the capacity. A medium with zero relaxed shear stiffness has no restoring force against a sideways/transport perturbation. So the symmetric state is neutral/unstable: a one-sided channel, once seeded, is not opposed by the substrate. Symmetry-breaking is permitted — even favoured — not forbidden. This is the legitimate inherited answer to why one-sidedness is possible at all: the marginal medium allows it.

• Not inherited [O] — the inevitability. "Always occurs" overstates. Marginality removes the restoring force; it does not by itself force a break (neutral ≠ obligatory), nor pick one global side vs many convection cells (degree-1 vs higher), nor fix the direction or timing. Which / where / when stay [O]. The overstatement conflates "not opposed" (inherited [F]) with "forced" (not inherited, [O]).

How it plugs into the assembly chain (M15/CG-30). M15 showed: given an opening, conservation + buoyancy + kinematics force a one-sided felsic pile [F]. This module supplies the missing upstream permission — the marginal substrate does not resist the initial symmetry break that seeds the opening. So the full chain reads:

marginal substrate (no restoring shear) — permits → a symmetry break (permitted [F], not forced [O]) — if it occurs → opening — forces → antipodal closing — forces → one-sided felsic pile [F]

The break is permitted [F] but not forced [O]; everything after the opening is forced [F]. This is a real strengthening of CG-30: its one unstated premise (why a break is allowed) is now grounded in the inherited c²=B/ρ substrate — without claiming inevitability.

§17.2Part B — "why a long horizontal band?" (the author's own weak hypothesis)

The author flags this as weak and guesses a magnetic-field link. Two honest points, in order:

1. First check the geometry is even a fact. "A long horizontal (E–W) band" is not an established present-tense description of the continental mass. The land is concentrated in one hemisphere (degree-1, M12/M13), and that concentration is fading (M14 audit). It is not a clean E–W zonal band: the Americas run N–S; Africa–Eurasia is broadly E–W; Antarctica sits at the pole. So before explaining a band, the band itself is [O]/not-established — possibly an artifact of one map projection or one supercontinent phase. Do not explain a shape that isn't confirmed.

1. The magnetic guess — graded honestly as weak [O], with a falsifiable form. A magnetic-field control on continental layout faces a hard quantitative wall: the magnetic body force on crustal rock is negligible next to gravity/buoyancy (rock is weakly magnetic; the geomagnetic field ~25–65 µT exerts forces tens of orders below the isostatic forces that govern where light crust sits). So a direct magnetic steering of continents is, on present-tense physics, disfavoured. The defensible falsifiable version is indirect and still [O]: do continental/true-polar-wander geometry and the core's degree-1/degree-2 structure (e.g. the antipodal lower-mantle provinces) share an orientation, beyond chance? That is a present-tense correlation test — but it is not evidence yet, and the body-force route is essentially closed. **Grade: weak hypothesis, [O]; magnetic body-force steering disfavoured by present-tense physics; an indirect orientation-correlation is the only testable residue.**

§17.3Ledger impact

• CG-31 added: inherited symmetry-breaking capacity — a marginal (zero-relaxed-shear) substrate does not restore a transport perturbation, so one-sidedness is permitted/favoured [F]; inevitability/side/count/timing remain [O]. Grounds CG-30's upstream premise. Inherits DOI 10.5281/zenodo.17972568.
• CG-32 added (author's weak hypothesis, recorded honestly): the "long band" geometry is not established ([O]); a magnetic control on layout is disfavoured as a body force, with only an indirect orientation-correlation left as a falsifiable, currently-unsupported [O] test.

§18.1The hypothesis

A stationary "storm" in the mid-Pacific where three co-rotating cores meet forces suction + ejection, venting hot/unjammed mantle upward; the vented material spreads laterally, and the present continental structure is downstream of that primary vent (a "leading"/seeding location). If the centre is near the equator, centrifugal force may give a reason. (Author flags the equator part as tentative.)

This explicitly invokes the inherited mechanism, so it is graded clause-by-clause against what that volume actually forces.

§18.2What the fluid-dynamics volume actually licenses

Read, not assumed (App-K / §11): **odd-cycle (C₃ = three-body) frustration forces co-rotation; co-rotating vortices merge into one larger coherent rotation; and a coherent rotation drives an Ekman through-flow — radial inflow transport ∝ √(ν/Ω). So the volume gives a genuine forced vortex + forced pumped flow. It says nothing about three cells meeting to eject upward**.

§18.3Clause-by-clause grade

§18.4The decisive point — direction of the forced flow (CORRECTED, mark-never-erase)

Correction (v1 → v2). An earlier version of this module split "inflow/suction" and "upward venting" as rival directions and concluded co-rotation forces inflow, not ejection. That was wrong, and the author's pushback is correct. They are one mass-conserving circulation, and the energy source fixes the sign:

• Heat is supplied from below (hot core/mantle, near-unjamming — M11). A flow driven from below cannot vent net-downward — that would transport energy back toward its source, against the gradient. So "would it sink into the core?" → no. The forced sense of a bottom-heated cell is rise at the centre + spread at the top.
• It is one loop: the buoyant column rises at the centre → by mass conservation it pulls fluid in at the base (that inflow is the "suction") → rotation organizes that convergent feed into a coherent vortex (Ekman pumping has an up-leg through the interior) → outflow/spread aloft. Base-inflow and centre-uprise are the same circulation; the v1 error was reporting only the bottom (inflow) leg.

What rotation does vs what buoyancy does (kept honest): buoyancy supplies the lift (hot/near-unjammed mantle rising, M05/M11); rotation supplies the organization (C₃ frustration forces co-rotation; vortices merge; coherent rotation forces the convergent Ekman feed). Together: a bottom-heated rotating cell that pumps up at the axis and spreads at the top. So "three rotations meet → a vent" is correct once the driver is heat-from-below — the rotation does not lift the material, it organizes the rising cell and its feed. The typhoon analogy is apt, and the upward vent is the forced sense; the v1 "inflow-only" reading is retracted.

§18.5On the equator / centrifugal sub-claim

Centrifugal acceleration at the equator is real but ~0.34 % of gravity — it sustains the 21 km equatorial bulge and can weakly bias long-wavelength structure toward the rotation axis. It is far too small to pin a vent location by itself ([O] as a locator), but it is a genuine, present-tense, non-zero symmetry-breaker aligned with the spin axis ([L] as a weak bias). Honest: a tendency, not a cause.

§18.6Assembled verdict

• Forced [F]: energy-from-below ⇒ a bottom-heated cell vents up at the axis and spreads at the top (sinking into the core is excluded by the reductio); rotation forces the coherent vortex + convergent Ekman feed; buoyancy supplies the lift. The upward vent is the forced sense — the v1 inflow-only reading is retracted.
• Still open: the lift is a buoyancy [L] term (needs heat-from-below, which M11 supplies); the vent being fixed/primary for the present continents is [O] (deep-time; continents reconfigure per M14); equator/centrifugal is a [L] weak bias (0.34 % g), [O] as a locator.

The typhoon analogue is apt: a real inherited rotating vortex [F] organizing a buoyancy-driven up-and-out vent. This connects cleanly to CG-30/CG-31: the marginal substrate permits the break, buoyancy (energy from below) drives the rise, and co-rotation forces the organizing vortex with its convergent feed — each grade kept honest.

§18.7Ledger impact

• CG-33 (CORRECTED v2): a bottom-heated rotating cell has a forced UP-and-OUT vent — energy from below cannot sink to the core; buoyancy lifts [L] (M05/M11) and rotation forces the organizing vortex + convergent Ekman feed [F] (DOI .17972568). The earlier 'inherited sign is inflow, not ejection' split is retracted (base-inflow and centre-uprise are one loop). Vent fixity/primacy [O]; equatorial/centrifugal [L] weak bias (0.34 % g) / [O] locator.

§19.1A. The spine that holds (present-tense [F]/[V], firewall-clean)

These survive the firewall and the M14 self-audit. Read as one chain:

1. Two-tier crust ⇒ dry land exists — felsic (~2.8) over basalt-skin (~2.9) over mantle (~3.3); the density dichotomy is why land and basins coexist. [V] (CG-1)
2. A marginal, hot, near-unjamming mantle behaves as a heavy fluid — bounded by S-waves (solid, not magma) and viscosity (10¹⁵–10¹⁹× magma); flows on Myr, rigid on seconds. [F]/[V] (CG-26)
3. Energy from below + rotation ⇒ a forced up-and-out vent cell — buoyancy lifts [L], co-rotation (C₃→merge→Ekman feed) organizes the vortex [F]; it cannot sink to the core. [F] skeleton (CG-33)
4. A marginal substrate does not restore a symmetry break — one-sidedness is permitted/favoured, not forbidden. [F] (CG-31)
5. Given an opening, a one-sided felsic mass is forced — fixed-area sphere ⇒ opening entails antipodal closing (conservation) ⇒ skin subducts/raft cannot (buoyancy) ⇒ raft collects at the sink (kinematics). [F] (CG-30)
6. Compression on the closing face is feasible and gated by buoyancy — buckling needs only hundreds of MPa at long λ [F] (CG-8); thick compensated relief needs a felsic root, so bare skin stays smooth (the deep Pacific floor) — the isostatic-compensation law [F] (CG-24).
7. A wet, buried, heated pile distils felsic → granite — water drops the solidus ~300 °C; the submarine start supplies the water. [F] with water (CG-10)
8. Emergence needs no special lift — basin subsidence + buoyancy give freeboard by isostasy. [F] (CG-14)

This is a genuinely connected mechanism from "hot fluid mantle" to "one-sided dry land," with each link forced or measured. That is the achievement, and it is not small: a chronology-free, pure-physics skeleton that does not import deep time anywhere (AUDIT-1 cleaned the places it had crept in).

§19.2B. The one structural weight-bearing gap

Everything in A is conditional on felsic existing. The entire spine routes through one unresolved node:

**CG-12 / CG-20 — can the rupture process actually make continent-scale felsic, and how did the first seed nucleate?**

• CG-12 (scale): pure extension demonstrably makes minor rhyolite (Iceland); whether it reaches continent scale without a subduction-like deep-water path is [L]/open. If no, the model must inherit a deep-water face → it converges with mainstream arc-genesis on mechanism (differing only in setting). If yes, it diverges — and that divergence is the whole point (CG-29).
• CG-20 (first seed): with no land to erode, how did the first light crust nucleate? [O] — and honestly open in mainstream too.

This is the single highest-leverage problem. Resolve CG-12 and the thesis either earns its divergence or honestly merges with the standard account; nothing else changes that verdict.

§19.3C. Remaining tasks, ranked by leverage

Tier 1 — structural (moves the whole edifice):

• T1. CG-12 felsic-scale fork. Build a present-tense screen: the felsic fraction producible by pure decompression melting of hydrated basalt vs by a deep-water (flux) path, anchored on measured rhyolite fractions (Iceland, oceanic plateaus) vs arc andesite. Output: a quantitative threshold for "continent-scale," and which path clears it. This is the next build.
• T2. CG-20 first-seed. Frame the nucleation as a present-tense feasibility question (can a small felsic domain self-concentrate from a basalt skin under repeated wet melting?) without invoking history. Likely stays [O], but sharpen why.

Tier 2 — consolidating (strengthens the spine, lower risk):

• T3. Couple the three [F] engines into one screen. CG-31 (permission) + CG-33 (vent) + CG-30 (assembly) are currently three screens; a single coupled screen showing permission → buoyant vent → opening → one-sided pile as one forced sequence would make the skeleton one object, not three.
• T4. CG-24 live falsifier sweep. The compensation law predicts no compensated multi-km bare-ocean high with a deep light root. A short global check (free-air + Moho) of candidate oceanic highs would either break it or harden it.
• T5. CG-33 vent ↔ buoyancy quantification. Tie the rotation-organized vortex to a buoyancy (Rayleigh/plume) criterion so "vent" has a present-tense threshold, not just a direction.

Tier 3 — honest housekeeping (keeps the firewall clean):

• T6. The S1–S3 present-tense geometry. Per AUDIT-1, the symmetric stripe pattern / sediment gradient / chain geometry remain legitimate present-tense constraints the thesis should reproduce in geometry (not chronology). A screen that states what geometry the rupture model predicts vs observes.
• T7. Inheritance hardening. When the engine bundle (DOIs .17932566 / .17978934 / .17972568) is shipped alongside, re-verify the inherited gates locally (as v28 does), so the package is self-contained.

§19.4D. The honest posture (for both of us)

The spine is real and connected — that is the headline, and prior turns under-stated it by foregrounding caveats. The model is not "mostly holes": it is a forced chain with one decisive load-bearing gap (CG-12) and a set of consolidating tasks. The discipline that makes it trustworthy — pure physics, no imported chronology, mark-never-erase, every result gated — is exactly what lets the spine stand without overclaiming. Next build: T1 (CG-12 felsic-scale fork) — the one test that decides whether the thesis diverges from or merges with the standard account.

§19.5Ledger impact

• CG-34 added: standing synthesis + leverage map. Spine = CG-1/26/33/31/30/24/10/14 (forced/measured); single structural gap = CG-12 (felsic scale) with CG-20 (first seed); next build = T1.

§20.1The move

Spread out the granite map and the limestone map; read the causation. This is exactly the load-bearing gap (CG-12) made geographic — because where a rock concentrates tells you what process makes it.

§20.2What the two maps show (present-tense, from the literature)

Granite (felsic batholiths) concentrates at convergent margins and orogenic "mobile" belts. It is made by partial melting at depth (>70 % silica; lower crust melts; magma rises), where a subduction-like wet zone delivers water and heat from below. Continents grow by marginal accretion of these granite belts, younger away from the ancient cores. Korea is granite-rich precisely because the Pacific plate subducts under Eurasia → wet partial melting above the slab → batholiths.

Limestone (carbonate platforms) concentrates on shallow (<200 m), warm, clear, sunlit shelves of extensional / passive margins and flooded epeiric seas. It is made by surface precipitation (corals, algae, shells + chemical), away from clastic input. It is laid down on top, not melted from below.

§20.3The causal discriminator

granite ← DEEP + hot-from-below + WATER + BURIAL + COMPRESSION (a melt distilled up from a buried, squeezed, wet pile) limestone ← SHALLOW + sunlit + SURFACE + low-clastic + EXTENSION (a precipitate laid down on a quiet, flooded, stretched shelf)

The two rocks are almost orthogonal processes — one is bottom-up from depth, the other top-down at the surface. And the deciding variable is the tectonic face: **granite marks where the crust closed/compressed; limestone marks where it stayed open/extended and shallow.**

§20.4The insight the map gives

The granite map and the limestone map are a readout of the two faces of one rupture engine. A rupture has an opening face (extension → subsiding basin → flooded shelf → limestone) and an antipodal closing face (compression → burial → wet melt → granite). So the maps are not two unrelated facts — they are the squeeze side and the stretch side of the same motion, written in rock.

Locality readout (present-tense): granite on compression faces 4/4 (Korea, Sierra Nevada/Idaho, SE-Australia/W-Europe belts, Andes); limestone on extension faces 4/4 (Bahamas, Arabian shelf, S-China epeiric, Santos pre-salt). The alignment is essentially clean.

§20.5What the map decides for CG-12 (the load-bearing gap)

CG-12 asked: does pure extension make continent-scale felsic, or does felsic need the convergent / deep-water wet path? The granite map answers in a specific direction:

• Continent-scale granite tracks the compression/convergent face, where the deep-water wet-melt path operates — not pure extensional rifts (those build the basin + the carbonate shelf + only minor silicic volcanics).

This is not a falsification of the thesis — it is a sharpening. The VP rupture already has two faces (CG-7), and the thesis already says granite is distilled on the closing/compression face (CG-9/CG-10). So the map resolves CG-12 by splitting it:

So the honest result: the model stops trying to make continent-scale granite from extension alone (the map says that doesn't happen) and locates felsic genesis on the compression face — which is where the thesis put it anyway. What remains genuinely VP-distinctive is the origin of the convergence itself (the closing antipode), now the real open locus.

§20.6Korea as the worked example

Korea is granite-rich and sits on the convergent western-Pacific face (wet melting above the subducting Pacific plate). Per the thesis, Korea is on a compression/closing face → should be granite-dominated → it is. A clean present-tense consistency check. The mirror prediction is equally testable: a thick carbonate-platform region should sit on an extension/passive face — the opposite face of the same engine.

§20.7Ledger impact

• CG-35 added: the granite map vs limestone map are the two faces of the rupture engine — granite = deep wet-melt at compression (bottom-up), limestone = surface precipitate at extension (top-down); discriminator = depth+water+burial+face. This resolves CG-12: continent-scale felsic tracks the convergent wet path (VP converges on melt-mechanism; strong extension-only form disfavoured), while VP's origin claim (convergence = antipodal closing face of a primordial rupture) survives as the real open locus of divergence. Korea fits the compression-face prediction (present-tense). This retires the strong form of CG-12 and re-points the next work at the origin of the convergence, not at extension-only felsic genesis.

§21.1The question

After M19 the named "big remaining problem" was the origin of the convergence — why is there a closing / compression face at all? This module asks, with the firewall on: is that really a remaining problem, or does it dissolve once the logic is closed? The honest answer is that it dissolves — it was an illusion of an un-closed loop.

§21.2The loop, closed (each link graded)

1. A hot fluid sphere cooling at its surface must convect. Rayleigh number ≈ 5.85×10⁶ ≫ critical (~10³) → convection is forced, not optional [F]. And the marginal substrate (relaxed shear → 0, CG-31) removes any stiffness that could lock a stagnant lid → it favours mobile-lid convection.
2. Convection on a closed sphere = upwelling limbs + downwelling limbs. Mass conservation forces both: what rises in one place sinks in another [F]. **→ the convergence is the downwelling limb — the antipodal partner of the opening already in the spine (opening forces antipodal closing, CG-30). The named gap dissolves here:** the convergence is not a separate, unexplained assumption.
3. At a downwelling, the denser hydrated skin sinks and carries water down [F] (cool dense skin sinks; the water-world start made the skin hydrated).
4. Water + heat at depth → flux melting → felsic distilled [F with water] — exactly the convergent wet path the granite map confirmed (CG-35).
5. Felsic is buoyant, can't re-sink → accumulates at downwelling zones [V] (CG-13).
6. Cells are discrete, not uniform → felsic piles in patches (continents) over downwellings; bare skin (ocean) over upwelling/spreading zones [F].

The loop closes: forced convection → down-limb (= convergence) → wet flux-melt → buoyant felsic → permanent, patchy dry land. No step needs an unexplained convergence.

§21.3The dissolutions (honest)

• "Origin of the convergence" → the downwelling limb of forced convection. Not a gap.
• "The first felsic seed" (CG-20) → the first downwelling of hydrated basalt skin flux-melts to make the first felsic; no pre-existing felsic is needed, and thermal convection (not a crustal density contrast) starts it. Largely dissolves too.
• "Does the loop actually run?" → answered by present-tense evidence: continent-scale granite exists at convergent belts, and active arcs make felsic now (Andes, etc.). It runs.

This is the logical work the question demanded. The "big problem" was hard only because the loop hadn't been closed; once closed, it isn't a problem.

§21.4What genuinely remains (and it is not a mechanism gap)

(R1) The quantitative budget — the real, falsifiable test. Does this loop, at realistic rates, distil the observed continental fraction (~40 % felsic), the freeboard, and the Moho contrast (~7 vs ~35 km) — or too little / too much? This is a present-tense number check, not a mystery. It is the genuine next test, and a wrong number could break the theory.

(R2) Identity — what is VP actually adding? The loop (convection → arcs → continents) is essentially mainstream mantle-convection plate tectonics. VP's value-add is not the convection but the upstream physics: c²=B/ρ explains why the mantle flows at all (the jamming/unjamming switch), and the water-world start supplies the flux water. **VP is the foundation beneath plate-convection, not a competitor to it** — the same posture the fluid-dynamics volume took ("the foundation beneath fluid dynamics, not a competitor"). The granite map (CG-35) already pointed here: VP converges with mainstream on mechanism and keeps its distinctiveness upstream.

§21.5The honest bottom line

The "big remaining problem" (origin of the convergence) was an illusion of an un-closed loop. Closed logically: the convergence is the forced downwelling limb, the first seed is the first downwelling's flux-melt, and the granite map shows the loop runs. The only genuine residual is the quantitative budget (R1) — a number to compute — plus the identity clarification (R2): VP is the c²=B/ρ foundation under mantle convection, not a rival to it.

§21.6Ledger impact

• CG-36 added: the named big gap (origin of the convergence) dissolves — the convergence is the downwelling limb of forced convection (Ra ≫ critical [F]; mass conservation forces both limbs [F]; = the antipodal closing of CG-30). The first seed (CG-20) largely dissolves (first downwelling flux-melts hydrated skin; no pre-existing felsic needed). The loop runs by present-tense evidence (granite at convergent belts; active arcs now). Genuine residual = (R1) the quantitative budget (observed ~40 % felsic / freeboard / Moho — falsifiable number check) and (R2) identity (VP = the c²=B/ρ foundation beneath mantle convection, not a competitor). The next build is R1, a budget screen — not a mechanism hunt.

§22.1The decisive test, reformulated to stay firewall-clean

R1 was the test named at M20/CG-36: does the buoyancy loop, at realistic rates, distil the observed continental crust? The naive form — production rate × time = standing volume — needs a duration, which the firewall caps at [O]. This is exactly where mainstream "crustal growth curves" smuggle deep time in. So R1 is reformulated as a present-tense fixed-point check: do the measured densities, thicknesses, and water volume reproduce the measured freeboard, hypsometry, and a sustainable steady-state volume? No time enters. A wrong number could break the thesis; it passes.

§22.2[0] The rock relief reproduces the prior screen

Airy isostasy gives continent-top minus ocean-floor (no water) = 4455 m, reproducing screen 4 / CG-14.

§22.3[B1] Freeboard — the lever is the felsic density deficit

Isostasy + the measured ocean water volume (filling the measured ocean area) gives the freeboard:

The textbook upper-crust value overshoots to +1.9 km; the measured bulk continental density (~2835–2870, because the lower crust is mafic/dense) or thickness lands at +840…+1170 m, hitting the observed mean freeboard (~+0.84 km). This is a real, falsifiable test — the answer could have been the wrong sign or 10× off — and it passes. Crucially the freeboard lever is the density deficit (ρ_m − ρ_c) and the crustal thickness: exactly the distilled-felsic buoyancy the thesis asserts. Mean ocean depth from the same budget = 3686 m (observed ~3.7 km). Grade [F]/[V] (sign + order + mechanism); the exact value below 0.5 km is sensitive (±~1 km) to bulk density and shelf treatment.

§22.4[B2] Bimodal hypsometry — the two peaks

The same isostatic system yields the two elevation peaks (relative to sea level, ρ_c = 2835): land +1544 m, ocean −3686 m, separation 5231 m ≈ 5.2 km (observed ~5 km). Sea level sits ~70 % of the way up the relief, so continents are near-marginal — emergent by only the top ~30 %. The deep basin lets the rest stand proud; land does not rise, the basin subsides. Grade [F]/[V].

§22.5[A] Volume steady state — a ratio of present-tense fluxes (no integration)

Comparing only the ratio of present-tense fluxes (no time): felsic production (juvenile arc addition, now) ~ 1–3 km³/yr against felsic destruction (sediment subduction + subduction erosion + lower-crust delamination, now) ~ 1.5–3 km³/yr gives P/D ~ O(1) → continental volume is near a fixed point now, and the ~40 % is sustainable as a balance. Falsifiable: P ≫ D ⇒ rapid growth (freeboard rising fast — not seen); P ≪ D ⇒ collapse (not seen). Grade [L] (the rates carry real uncertainty and are stated as a present-tense ratio, never as a date).

§22.6What this closes, and what stays open (honest)

• Closes (present-tense, falsifiable, passes): freeboard sign + order + mechanism [F]/[V]; bimodal hypsometry + ~5 km separation [F]/[V]; near-marginal emergence [F]; volume steady-state [L].
• Stays open: the exact freeboard below 0.5 km (measured bulk density + thermal subsidence + shelf hypsometry); the area fraction value is treated in module 22; occurrence/timing stays [O] forever.
• Firewall held: no deep-time growth curve is used anywhere; every rate is present-tense; "how it grew over time" is [O]; only present-tense observables are load-bearing.

§22.7Ledger impact

• CG-37 added: the present-tense budget (B1 freeboard, B2 bimodal hypsometry, A steady-state volume) reproduces the observed crust with no fitted parameter and no deep-time curve; freeboard and hypsometry [F]/[V], volume balance [L]. R1 is addressed on present-tense grounds; the lever is the distilled-felsic buoyancy. The exact freeboard (<0.5 km) and the area-fraction value remain open.

§23.1The question R1 left

Module 21 showed the freeboard, hypsometry, and steady-state volume close present-tense. The one remaining number is the continental area fraction (~40 %). Is it derivable, or only a rate balance?

§23.2The insight: continents are not the instantaneous downwelling footprint

Trenches are thin lines; continents are broad. Felsic is buoyant and accumulates — it is not confined to where a downwelling sits now. The geometric ceiling is set by the ocean network: the basaltic skin must stay connected (ridge → trench) to keep subducting. Continents grow until the ocean is at the edge of percolation; a fragmenting ocean throttles its own subduction (weak return-flow) → felsic production stalls → the fraction is pinned at the percolation threshold, independent of the rate constants. This is the same marginal/critical signature as the c²=B/ρ substrate.

§23.3The evidence, in four layers (all SEED=19, no fit)

1. Planar percolation (self-contained, in the screen). A stdlib union-find percolation: the ocean stops spanning at a continental fraction **f\* ≈ 0.41 (site p_c ≈ 0.5927 → 1 − p_c ≈ 0.407). The observed continental fraction ~0.41 sits on this geometric threshold**.
2. A self-organizing loop (simulations_session/percolation_attractor.py). Felsic accretes at active (connected) ocean margins, throttled by ocean-connectivity health, recycled at rate r. Scanning the production/destruction ratio over 4× (8…30) gives steady f = 0.39–0.41 — a band of ±0.01 centred on f\ and on observed, far* narrower than the rate variation. The percolation ceiling is a rate-insensitive attractor, not a rate artefact; the connectivity throttle is a physical dependence (a fragmenting ocean subducts weakly), not a tuned knob.
3. A 3D-convection stress test (simulations_session/rbc3d.py + tracer). An infinite-Pr (mantle) 3D Boussinesq solver, validated against the analytic onset Ra_c = 27π⁴/4 ≈ 657.5 (single-mode growth matches linear theory to ~1e-7; onset crosses zero at 657.6), produces a realistic cellular surface (downwelling area fraction 0.46, ~20 cells). Driving the tracer with this real convective surface flow: the percolation throttle still bounds f and compresses the rate-dependence (the attractor survives in 3D), but the value is coupling-sensitive — static 2D = 0.39–0.41; frozen-flow stirring = 0.24–0.32 (convective stirring disperses continents, lowering the ceiling).
4. The two-way-coupled 3D run — EXECUTED (simulations_session/coupled_rbc3d.py + coupled_driver.py). Continent insulation feeds back into the buoyancy field (zero-mean = self-limiting) and the convection co-evolves with the continents (which ride the large-scale flow = raft rigidity). The coupled solver is re-validated: with C=0 it reproduces Ra_c = 657.5 exactly (the coupling does not corrupt it). Within the stable window the coupled fraction settles in a statistical steady state around f ≈ 0.30 (~0.27–0.36), robust to the coupling strength (q=0 baseline and self-limiting insulation alike). This **confirms the frozen-flow lower bound and the percolation mechanism — it does not reach the observed 0.41.** Numerical caveat: at the strongly-supercritical Ra=10⁴, 64×64×12 is stable only for a finite window (~7000 steps) — even the q=0 baseline eventually blows up — so the long-time limit is the under-resolved convection, not the coupling; a fully-converged long run needs higher resolution / implicit (or hyperviscous / adaptive-dt) stabilization.

§23.4Honest reading (falsification = discovery)

The two-way-coupled run did not confirm the convenient 0.41 — it broke that hope and confirmed ~0.30. Vigorous convective stirring disperses a passive continental tracer, holding f near ~0.30; and real Earth's mantle is at far higher Ra (~10⁶–10⁷), so it stirs even more and would push a passive tracer lower, not higher. So the observed 0.41 must be sustained by physics this passive model omits: continental coherence/strength resisting dispersal (a real rheology, not a passive scalar), felsic production concentrated at convergent margins, sphere geometry, and internal heating (narrower downwellings). The coupled run did the honest thing — it confirmed the lower bound and sharpened the residual into a specific, testable list, rather than rubber-stamping the target.

Grade [L]. The percolation attractor is now confirmed as a robust bounding mechanism (~0.25–0.41) across three independent settings: static 2D (0.39–0.41), 3D-frozen (0.24–0.32), and 3D-two-way-coupled (~0.30). Not [F]: the connectivity threshold is model-dependent (planar → 0.41; the real plate-network's effective connectivity giving f\* ≈ observed is a falsifiable prediction), and the coupled value (~0.30) sits below observed 0.41. The decisive run is done; the residual is no longer "run the coupled case" but "supply the omitted continental rheology + convergent-margin production (and a numerically-stable high-Ra long run) to test whether they lift ~0.30 to the observed 0.41."

§23.5R2 — the identity, resolved: three faces of one marginal-criticality kernel

Three present-tense observables are the same marginal/critical self-organized attractor:

So continental genesis — where the land is (over downwellings), how high it stands (marginal freeboard), and how much there is (marginal percolation) — is governed by marginal-criticality self-organization, the c²=B/ρ kernel's signature. This is R2: the convection loop itself is essentially mainstream; **VP is the criticality foundation beneath mantle convection, not a competitor to it.** Average-theory is kept out precisely here — VP does not re-explain convection, it supplies the marginal physics under it.

§23.6Ledger impact

• CG-38 added: the ~40 % area fraction is a self-organized marginal-connectivity (percolation) attractor of the complementary ocean network — robust and rate-insensitive (band 0.39–0.41 in 2D across a 4× rate range; planar f\* ≈ 0.407). A validated infinite-Pr 3D Boussinesq solver (Ra_c = 657.5) shows the attractor survives in real convection but the value is coupling-sensitive. The two-way-coupled 3D run is now EXECUTED (insulation feedback + raft rigidity co-evolving with convection; re-validated C=0 → Ra_c = 657.5): the coupled fraction settles around ~0.30 (~0.27–0.36), robust to coupling strength — confirming the frozen-flow lower bound and the percolation mechanism, not reaching observed 0.41. The attractor is thus confirmed as a robust bounding mechanism (~0.25–0.41) across three settings: static 2D (0.39–0.41), 3D-frozen (0.24–0.32), 3D-coupled (~0.30). Grade [L]; the residual is reframed — not "run the coupled case" (done) but supply the omitted continental rheology + convergent-margin production (real Earth's higher Ra would stir a passive tracer even lower). The three marginal attractors (substrate / freeboard / area) unify continental genesis under the c²=B/ρ kernel — R2: VP is the marginal-criticality foundation beneath mantle convection, not a rival.

§24.1The question CG-38 left

The two-way-coupled passive tracer settles ~0.30 < observed 0.41 (CG-38). The residual was named, not hand-waved: the passive scalar disperses under vigorous convective stirring, and real Earth's higher Ra would push a passive tracer lower, so the 0.30→0.41 gap must be closed by physics the passive model omits — continental coherence/strength resisting dispersal (a real rheology, not a scalar) + felsic production concentrated at convergent margins. CG-39 supplies exactly those two ingredients and asks whether they lift ~0.30 to 0.41.

§24.2The two ingredients (each physical, neither tuned)

1. Continental coherence (rheology). Interior continent cells resist advective dispersal — raft rigidity at the cell scale: mobility = 1/(1 + κ·(#continent neighbours)). A cell deep inside a continent (8 continent neighbours) is nearly frozen; an isolated cell advects freely. κ = 0 recovers the passive tracer. This is the "real rheology, not a passive scalar" lever, encoded with one scanned strength κ.
2. Convergent-margin-concentrated production. Felsic born preferentially where the surface flow is downwelling (convergent), versus uniform at all active-ocean margins.

§24.3The result (scanned, not tuned)

Sim 1 — the lift from below (continental_rheology.py; real 3D convective stirring; f vs κ, 3 seeds):

The passive point (κ=0, uniform) = 0.29 reproduces the established coupled ~0.30 deficit. Coherence lifts f monotonically toward the ceiling, recovering ~75 % of the 0.30→0.41 gap, but saturates ~0.38 (uniform) / ~0.34 (convergent) at cell-scale rigidity — it does not reach 0.41. (Convergent-only production with no coherence gives ~0, because continents form at downwellings and are immediately swept away; coherence is what lets convergent production accumulate at all.)

Sim 2 — the ceiling is not exceeded (coherence_ceiling.py; bit-deterministic cat-map stirring; f vs κ): f stays at/below ~0.41 for all κ (0.43 → 0.41 as coherence rises) and never runs away above it. The percolation ceiling is robust to coherence.

§24.4Honest reading (falsification = discovery)

The two sims bound the value from both sides, and the answer is the middle of the CG-39 fork — both branches are partly right:

• The percolation ceiling (the c²=B/ρ kernel's connectivity margin) sets the binding upper bound ~0.41. Observed 0.41 is not coincidental — it sits on the geometric threshold (planar f\*≈0.407).
• Continental coherence (rheology) is the confirmed lever that lifts the vigorously-stirred system up from the dispersal-suppressed ~0.30 toward that ceiling.
• Coherence cannot push f past the ceiling.

So where the value comes from is now resolved in structure: the kernel fixes the ceiling, rheology determines how close the stirred Earth gets to it. The experiment did not rubber-stamp 0.41 — it confirmed the direction and the ceiling-as-bound and sharpened the residual: the last ~0.38→0.41 needs a true yield rheology (continental lithosphere has finite strength, stronger than cell-scale raft rigidity), plus sphere geometry and internal heating (narrower downwellings).

Grade [L]. The mechanism — coherence lifts the stirred fraction toward a binding percolation ceiling — is confirmed across the 3D run and the deterministic ceiling loop; the exact 0.41 is approached, not nailed, at the modelled rigidity. No fitted parameter (κ scanned). Occurrence/timing stays [O] forever.

§24.5Ledger impact

• CG-39↑ added (amendment to the v1.4 CG-39 residual; mark-never-erase): the omitted physics is supplied and scanned. Coherence lifts the passive 0.29 toward the ceiling (recovers ~75 % of the gap, saturates ~0.38), and cannot exceed the percolation ceiling ~0.41. The ceiling (kernel) sets the bound; coherence (rheology) is the confirmed lever toward it. Grade [L]; residual sharpened to a true yield rheology + sphere geometry + internal heating.
• This strengthens R2: the area fraction's bound is the kernel's percolation margin, and its approach is a rheology lever — both downstream of c²=B/ρ, not a rival to mantle convection.

§24.6v1.6 refinement — the yield-strength candidate is TESTED and FALSIFIED (CG-39↑↑, screen 21)

The v1.5 record above named a true yield rheology (> the cell-scale mobility proxy) as the leading candidate to close the ~0.38→0.41 residual. v1.6 implements it and tests it: simulations_session/continental_yield.py gives continents a finite yield strength — a block deforms only where the convective strain rate exceeds yield Y, and below yield it translates rigidly (block-mean velocity, edges included — the rigid-body behaviour the neighbour-count mobility proxy structurally could not produce, since mobility always leaves rim cells mobile). Y is scanned to the parameter-free rigid limit (Y→∞) — no tuning.

Result (f vs Y/Srms, 3 seeds; Srms = RMS strain rate of the surface flow):

Rising Y lifts f but it saturates ~0.34 (uniform) / ~0.33 (convergent) and stays there all the way to the rigid limit. Making continents perfectly rigid rafts does not reach the ceiling — rigid blocks are still advected into the downwelling/convergence pattern, which throttles accretion below ~0.41. (The convergent column stays at ~0 for low Y because convergent-only continents deform and disperse before they can cohere; only above yield do they survive, then saturate ~0.33.)

This falsifies the v1.5 conjecture. Two independent rheology encodings — the cell-scale mobility proxy (~0.38) and the stress-based yield model (~0.34, even fully rigid) — both saturate short of 0.41. Continental rheology is a confirmed partial lever (it lifts the stirred 0.29–0.30 to 0.34–0.38) but is not sufficient to close the gap from below. Combined with screen 20 (coherence cannot exceed the ceiling from above), the picture tightens: the percolation ceiling ~0.41 is the binding bound, approached from below by rheology and not exceeded from above by coherence; observed 0.41 sits on it.

The residual is therefore NOT continental rheology. It points to the two untested candidates — sphere geometry (a closed surface, free of periodic-box stirring artifacts) and internal heating (narrower downwellings, a different cell pattern) — or, more simply, the planar percolation threshold 0.407 is itself the bound, which a flat stirred box only approaches from below while the real spherical, internally-heated Earth sits at it.

Grade [O] for the closure (reopened: the named closer was tested and rejected) / [L] for the bound and for rheology-as-partial-lever. No fitted parameter (Y scanned; the rigid limit is parameter-free). falsification = discovery — a prior conjecture of this package was put to a machine test and rejected.

§25.1The question this closes

v1.6 (module 23 / screen 21) left the residual as an either/or: does the stirred fraction (~0.34–0.38) close to 0.41 with untested physics (sphere geometry, internal heating), or is the planar percolation ceiling 0.407 the real bound with observed 0.41 sitting on it? A full spherical convection solver is out of scope in this session, so the geometry candidate is tested where it is actually decidable: a percolation threshold depends on the network's coordination number z, and changing the surface geometry (square grid → triangulated sphere → hexagonal tiling) changes z. So the sharp, answerable form of "does geometry matter" is: **is the ceiling f\* universal, or is it the value for z = 4?**

§25.2The result (verified, not recalled)

Continent ceiling f\* = 1 − p_c (the fraction where the ocean network stops spanning) vs coordination z, by one union-find:

f\ rises strongly with connectivity (higher z → the ocean spans more easily → it tolerates more* continent before disconnecting). The observed continental fraction ~0.41 matches only the z = 4 case.

§25.3Honest reading (falsification = discovery)

The percolation ceiling is NOT a geometry-universal constant. It is the z-dependent threshold, ranging ~0.41 (z=4) → ~0.61 (z=8) over reasonable connectivities. Three consequences, stated plainly:

1. 0.41 is the z = 4 value — the 4-connected ocean the simulation grid uses (scipy.label default = 4-neighbour). The earlier statement "observed 0.41 sits on the percolation ceiling" holds only for 4-connectivity.

1. Reaching exactly 0.41 is therefore NOT a missing-convection-physics gap. Sphere geometry or internal heating would not "close" 0.34→0.41 — the 0.41 target is itself the z=4 threshold. The stirred-box undershoot (0.34–0.38; screens 20–21) sits below the z=4 ceiling because vigorous stirring + a finite box hold it there; it is not evidence of a missing mechanism that lifts the ceiling.

1. The match to observation becomes a falsifiable empirical claim: the effective coordination of the real subductable-ocean network is ≈ 4. This is not obviously right — a naive plate-adjacency graph (each major plate borders ~5–6 others) is z ≈ 5–6 and would predict f\ ≈ 0.45–0.50, not* 0.41. Whether the right model is a continuum/grid ocean (z ≈ 4) or a plate-adjacency network (z ≈ 5–6) is open, and it decides the predicted value. This is exactly the falsifiable prediction the ledger flagged earlier ("the real plate-network connectivity giving f\*≈observed"), now made quantitative.

§25.4What this does to the grade

• The area fraction is a percolation bound — confirmed and now characterised across geometries (the mechanism is robust). [L].
• The specific value 0.41 is connectivity-contingent, requiring z ≈ 4. This downgrades the certainty of the value relative to the v1.5 "attractor at 0.41" framing — the honest direction. [O] for the value, pending the real-network connectivity.

This is a deliberate downgrade: a machine test of geometry-robustness showed the value is less forced than v1.5 implied. The mechanism survives; the number is contingent.

§25.5Remaining untested levers (named, not faked)

Two candidates from v1.6 remain untested here and would refine — but not overturn — this conclusion:

• Internal heating in the Cartesian box (narrower downwellings): a convection-pattern test of whether the stirred value approaches the z-appropriate ceiling more closely. It changes the approach, not the ceiling's z-dependence.
• Full spherical convection (a closed-surface solver): out of reliable scope in this session; flagged honestly rather than approximated. It would fix the effective z of a real convecting spherical ocean — the very quantity this module shows is decisive.

§25.6Ledger impact

• CG-39↑↑↑ added (mark-never-erase): the ceiling is connectivity-dependent (~0.41 at z=4 → ~0.61 at z=8); 0.41 is the z=4 value; the residual either/or resolves to "the ceiling is the bound, its value is contingent on connectivity"; the observed match is a falsifiable claim that the real ocean network is effectively ~4-connected (naive plate-adjacency z≈5–6 would predict ~0.45–0.50). Grade [L] bound / [O] value. The area-fraction value certainty is downgraded from v1.5 — falsification = discovery.

§26.1The question this closes

Screen 22 (module 24) showed the percolation ceiling is connectivity-dependent — f\* ≈ 0.41 at z = 4, ≈ 0.50 at z = 6 — and left one decidable question: is the real ocean/plate network z = 4 (→ 0.41) or z ≈ 6 (→ 0.50)? That single number sets the predicted value. This module answers it not by guessing but by exact topology.

§26.2The answer (exact, not estimated)

A sphere tiled into F plates whose boundaries meet at triple junctions (the generic, mechanically stable case — quadruple junctions are unstable) satisfies Euler's formula V − E + F = 2 together with the triple-junction count 2E = 3V. Eliminating V gives E = 3F − 6, hence the mean number of neighbours per plate:

$$ z \;=\; \frac{2E}{F} \;=\; 6 - \frac{12}{F}. $$

This is exact. For realistic plate counts:

The ~15 commonly-cited plates give z = 5.2; even the full 52-plate catalogue gives 5.8. The natural plate network is ~6-connected, not 4 — closer to a triangular tiling (z = 6, f\ ≈ 0.50) than to the square grid (z = 4, f\ ≈ 0.41) the simulation happened to use.

§26.3Honest reading (falsification = discovery)

A z ≈ 5–6 network sits between the bracketing integer connectivities, so its implied percolation ceiling is **f\* ≈ 0.46–0.50 — which overshoots** observed 0.41. Three plain consequences:

1. The answer to "z = 4 or z ≈ 6" is z ≈ 6. So the clean "0.41 = z = 4 threshold" coincidence (screen 22) is not the natural network's prediction. The natural network predicts a higher ceiling than observed.

1. Observed 0.41 sits below the natural-network ceiling. It lands in the band where the convective stirring holds the fraction down (the 0.30–0.38 undershoot, screens 20–21). So 0.41 is best read as (connectivity ceiling ≈ 0.46–0.50) − (a stirring undershoot) — an interplay, not a single forced number.

1. What survives is the claim that was actually load-bearing. The continental fraction is a percolation-bounded quantity in the correct regime — a sub-majority ~0.4–0.5, not ~0.1 or ~0.9. The kernel chain (c²=B/ρ → felsic accumulates and cannot subduct → the ocean skin must stay connected to subduct) fixes that regime. It does not uniquely pin 0.41; the exact value lies inside a ~0.30–0.50 model bracket spanned by (connectivity choice) × (stirring vigour).

§26.4Final grade and closeout

• [L] — the continental fraction is a percolation-bounded sub-majority (~0.4–0.5), forced by the kernel. This is robust across every test in screens 18–23.
• [O] — the exact value 0.41 is set by the connectivity-ceiling-minus-stirring interplay and sits within model uncertainty; it is not uniquely forced by any single clean argument.

CG-39 is CLOSED. The residual that began as "lift the coupled 0.30 to 0.41" was chased to its floor across five tests — coherence (recovers ~75 %, saturates 0.38), yield rheology (saturates 0.34 even rigid, falsified as the closer), the connectivity-dependence of the ceiling, and finally the exact plate topology. The honest end state is regime-yes, exact-value-no: the mechanism is real and predicts the right order; the precise number is contingent. Occurrence/timing stays [O] forever.

§26.5Ledger impact

• CG-39 ✓ (CLOSED) added (mark-never-erase): plate topology gives z = 6 − 12/F ≈ 5–6, so the natural-network ceiling f\* ≈ 0.46–0.50 overshoots observed 0.41; 0.41 is regime-correct (~0.4–0.5, kernel-fixed, [L]) but not uniquely forced (exact value within a ~0.30–0.50 bracket, [O]). CG-39 closes — falsification = discovery.

§27.10. What this document asserts

Continental Genesis derives its full chain — hot fluid mantle → convection → one-sided felsic mass → granite → emergent dry land — from a single inherited kernel:

• the marginal-substrate relation c² = B/ρ (a medium at the unjamming margin: relaxed shear → 0, bulk modulus B finite), and
• the R19 jammed ⇄ unjammed bistable switch (void-suction / unjamming rupture),

inherited, not re-derived, from the VP physics (DOI 10.5281/zenodo.17932566), Atlantic-geodynamics (DOI 10.5281/zenodo.17978934), and Configured-Continuum (DOI 10.5281/zenodo.17972568) volumes.

The architectural claim splits into two independent properties, kept separate because a theory can hold one without the other:

• NT-1 (no tuning). No Tier-A object contains a fitted parameter. Every LOCK constant is a measured present-tense quantity or a mathematical constant. SEED = 19 + double-SHA-256 make this machine-checkable.
• NT-2 (no addition). Every stage is an application of the one kernel — the opening face and the compression face of the same switch, the same buoyancy-sign law throughout — not a new mechanism introduced to clear that stage.

NT-1 and NT-2 are graded separately below. Where the kernel forces a result the grade is [F]; where it only permits one, [F]-permitted / [O]-forced; where the inference is well-anchored but short of forced, [L]; occurrence and rate stay [O] by construction of the firewall. The grade carries the modality. Nothing here is stated above its grade, and nothing established is stated below it.

§27.21. NT-1 — no tuning, and why the firewall makes it structural

A growth-curve account of continental volume has one adjustable degree of freedom that does the work: duration (rate × time = standing volume). The bidirectional chronology firewall caps exactly that quantity at [O] — for and against. The thesis therefore cannot tune the deep-time integral, because it forbids itself the integral. This is the structural source of NT-1: the one knob that an "explains-anything" account turns is removed at the constitution level, not by promise.

Consequently R1 (the quantitative budget) is posed as a present-tense fixed point, not an integration (module 21 / screen 17 / CG-37). Inputs and outputs are all measured-now; no time enters.

Worked example — the freeboard test is where no-fit is demonstrated, not asserted. Airy isostasy plus the measured ocean water volume (V_W = 1.335×10¹⁸ m³) and measured densities (ρ_m = 3300, ρ_w = 1027, ρ_o = 2900 kg/m³, H_o = 7 km) gives:

The crustal density was not moved to hit the target: the textbook value overshoots, and the measured bulk value lands on the observation. The freeboard lever is the distilled-felsic density deficit (ρ_m − ρ_c) — the kernel's own product. The same isostatic system reproduces the bimodal hypsometry (land peak +1544 m, ocean peak −3686 m, separation 5231 m; observed ~4.5–5 km) and places sea level 70 % up the relief, so continents are near-marginal — emergent by only the top ~30 %. Grade [F]/[V] on sign, order, and mechanism.

Machine-checkable. All 23 screens print REPRO GATE: PASS; MANIFEST.sha256 verifies 89 / 89 files. A reader regenerates the no-fit claim in five minutes offline; there is no place a parameter could be quietly fitted that the hashes would not expose.

NT-1 is therefore a verified fact, not a stylistic preference.

§27.32. NT-2 — one kernel across the chain

The spine is a single buoyancy / unjamming cascade. The relevant table is not the list of stages (that is in WHITEPAPER.md) but, per stage, whether it is the same engine or a new mechanism, and at what grade.

Two facts in this table are the substance of NT-2:

(i) The same switch is the opening face and the compression face. Subduction-versus-crumpling is not two mechanisms; it is one buoyancy-sign law applied to net-negative skin versus net-positive felsic (CG-22). The rift that floors a basin and the lateral push that thickens its skin are the two ledgers of one motion (CG-7).

(ii) M20 removes an assumption rather than adding one. The "origin of the convergence" looked like a separate posit. It is not: a hot fluid sphere must convect (Ra ≈ 5.85×10⁶ ≫ critical, [F]); closed-sphere convection has upwelling and downwelling limbs by conservation ([F]); the convergence is the downwelling limb (CG-36). The first felsic seed (CG-20), formerly the deepest [O], largely dissolves the same way: the first downwelling flux-melts hydrated skin. NT-2's strongest evidence is therefore a subtraction — an apparent extra mechanism shown to be already entailed by the one engine.

R2 — the unifying statement. Three present-tense observables are the same marginal-criticality signature:

This is what "one kernel" means quantitatively: where land sits (over downwellings), how high it stands (marginal freeboard), and how much there is (marginal percolation) are three faces of the same self-organized criticality. VP is the criticality foundation beneath mantle convection, not a competitor to it — the convection loop is mainstream; the marginal physics under it is the kernel.

§27.43. The exact extent — three grade boundaries

The grade system reports three boundaries. Stating them at grade is the objectivity: the [F]/[V] parts above are established; the parts below are capped exactly where the evidence caps them. Neither is inflated, neither is apologized for.

(a) Permitted versus forced. The kernel forces some links (CG-30 entailment, CG-36 downwelling limb, CG-14 emergence) and only permits others. CG-31 is explicit: the marginal substrate permits one-sidedness (it offers no restoring force) but does not force which side, how many, or when — those stay [O] ("not opposed" ≠ "forced"). The flux water (CG-11) is [L] and its origin is a problem the thesis assumes, not sources. So the architecture is feasible-to-forced depending on the link, and the grades say which. This is a statement of resolution, not a deficiency.

(b) Degeneracy: coherence is not evidence. Most single-kernel results in the §2 table are degenerate — mainstream predicts them too (CG-22 gating, CG-24 compensation law, CG-9 subduct-vs-resist, CG-27 present state). The no-tuning chain passing a degenerate stage demonstrates coherence (the one engine accommodates it without a new knob) — which is necessary — but not evidence (a competing account accommodates it equally). The package's design keeps degenerate items out of the evidence column (PUZZLE_MAP.md). Evidential weight therefore lives only in the non-degenerate edges: hemispheric continental asymmetry (#14; ~81 % of land on one hemisphere, no settled mainstream "why") and freeboard constancy (#3). The correct reading is precise: the no-tuning property is architectural; its evidential payoff is localized, by design, to the edges. "Passes every stage" and "wins every stage" are different statements, and only the first is claimed across the board.

(c) The one quantitative undershoot — stated flat. The area-fraction value is the single place where the passive kernel reaches a number below observation. The mechanism — a marginal-connectivity (percolation) attractor of the complementary ocean network — is [L] and robust: the ocean must stay connected (ridge → trench) to subduct, so continents grow until the ocean is at percolation, a rate-insensitive ceiling (planar f\ ≈ 0.407 from site p_c ≈ 0.5927; 2D self-organized band 0.39–0.41 over a 4× rate range). The value* is bracketed across three settings of an independently validated infinite-Pr 3D Boussinesq solver (onset Ra_c = 27π⁴/4 ≈ 657.5, matched to ~1e-7):

The two-way-coupled run (insulation feedback + raft rigidity co-evolving with convection; re-validated C=0 → Ra_c = 657.5 exact) settles ~0.30, robust to coupling strength, and does not reach the observed 0.41. Vigorous stirring disperses a passive tracer; real Earth's far higher Ra (~10⁶–10⁷) would push a passive tracer lower, not higher. So the observed 0.41 requires physics the passive scalar omits — continental rheology (coherence resisting dispersal, a material law not derived from c²=B/ρ) and felsic production concentrated at convergent margins. This is the single point at which reaching the observed number currently needs an added material law. It is reported at [L], the residual is named, and the run that produced 0.30 was kept although 0.41 was the convenient target — falsification = discovery.

§27.54. Why the architecture is the substantive claim

A theory that clears N stages with N tunable knobs constrains nothing; with enough freedom any chain can be made to pass. A theory that clears them with zero fitted parameters and one engine is constrained — it can be broken by a single wrong number. The freeboard / hypsometry / Moho fixed point is exactly such a breakable number, and with measured inputs it comes out right in sign and order.

The firewall makes this account less free than a mainstream growth-curve narrative, which retains the deep-time integral as an adjustable degree of freedom. Removing that freedom is the epistemic content: fewer degrees of freedom carrying the same present-tense fit is a stronger, not weaker, position. This is stated at full strength because it is what the grade system supports.

It is capped just as exactly. NT-1 and NT-2 establish that the chain is feasible, unfitted, and singly-sourced; they do not establish that the sequence occurred ([O] forever), and they do not make VP a rival to mantle convection (R2: it is the foundation beneath it). "Constrained and breakable" is not "confirmed." Both halves are stated at grade.

§27.65. Falsification surface

The no-tuning thesis is broken by any one of:

• a Tier-A object that requires a fitted parameter (NT-1 fails);
• a stage that demonstrably needs a second, independent engine not reducible to the switch (NT-2 fails);
• a present-tense fixed point — freeboard or bimodal hypsometry — that comes out wrong sign or order under measured inputs (the breakable number fails);
• a compensated multi-km bare-ocean high with a deep light root (small free-air anomaly, deep Moho, no felsic), which breaks the compensation law CG-24;
• the omitted continental rheology + convergent-margin production failing to lift the coupled area fraction from ~0.30 toward observed 0.41 once supplied (the §3c residual fails).

None of these has fired. The fourth has been probed (central Indian Ocean caps at 1–2 km; Gorringe Bank is uncompensated with a positive free-air anomaly) and not broken.

The no-tuning property — one inherited kernel, no fitted parameter, no per-stage mechanism — is the load-bearing architectural claim of this package; §3 states its exact extent and §5 its falsifiers. Occurrence and timing stay [O] forever. falsification = discovery.

§28.1Verdict legend

• EDGE — non-degenerate illumination or a distinct testable prediction → strengthens the thesis; look for evidence here.
• DEGEN — both the hypothesis and mainstream explain it → [O]; claiming it as evidence is overreach.
• SHARED-GAP — genuinely open; the hypothesis does NOT solve it (or assumes it).
• STRAIN — present-tense data the hypothesis must accommodate, or it is contradicted → the falsification = discovery frontier (most valuable).
• N/A — the hypothesis says nothing.

§28.2A. Crust & continent origin

§28.3B. Tectonics & deformation

§28.4C. Water & oceans

§28.5D. Deep Earth & geophysics

§28.6E. Surface, sedimentary, biological

§28.7F. Geochronology

§28.8G. STRAIN — the present-tense data the thesis MUST confront (falsification = discovery)

These are not "puzzles the thesis solves"; they are ordered present-tense observations that test it. Surviving them is what strengthens the thesis — listing wins does not.

Firewall note: S1-S3 are about order/pattern (present-tense), not absolute dates — so they are load-bearing as observations even while rates stay [O]. The thesis does not get to wave them away with the chronology firewall; it must show the ordered geometry is reproducible.

§28.9H. Honest synthesis

The map does NOT show "many puzzles solved." It shows:

• 3-4 genuine EDGES (non-degenerate, evidence-worthy): #14 land/water hemisphere asymmetry (top), #3 continental freeboard constancy, #23 tectonic-mode/water-enabler, and the #1 two-tier-origin framing. These are where pursuing evidence actually strengthens the thesis.
• ~12 DEGENERATE items (mainstream explains them equally) → [O]; claiming them as evidence is the overreach to avoid.
• ~7 SHARED-GAP items the thesis does NOT solve (incl. the water origin it assumes and the first-seed it leaves open).
• 6 STRAIN items — the real frontier. The thesis is strengthened by surviving S1-S3 (the ordered seafloor/hotspot patterns) and by closing S5-S6 (first seed, felsic scale), not by enumerating wins.

Recommended use: drop the DEGEN/N-A claims (they invite the "explains everything" trap). Put the work on the EDGES (esp. #14 hemisphere asymmetry — a real, under-explained, chronology-free present-tense observable) and on surviving the STRAIN (S1-S3). That is where "the thesis gets stronger and more evidence appears" — honestly, and falsifiably.

§29.1The 23 gated screens

Every quantitative screen is self-contained, deterministic (SEED = 19), and prints REPRO GATE: PASS on its last line; its displayed numbers are recomputed and checked against a double-SHA-256 gate, so a wrong number cannot pass silently. The screens, with the chapter each supports and its gate prefix:

§29.2Verify in five minutes (offline, no network)

cd repro
for s in *.py; do python3 "$s" | tail -1; done   # expect 23x REPRO GATE: PASS

Integrity of the whole tree, from the package root:

sha256sum -c MANIFEST.sha256   # expect: all OK

§29.3The genuine computational-reproducibility caveat

One quantity has a real reproducibility obstacle and is graded accordingly. The area-fraction value from the fully-coupled high-Rayleigh-number run is computationally bounded: at the strongly-supercritical Ra = 10⁴, a 64×64×12 grid is numerically stable only for a finite window (~7000 steps) — even the q = 0 baseline eventually blows up — so the long-time limit is the under-resolved convection, not the coupling. A fully-converged long run needs higher resolution or implicit / hyperviscous / adaptive-time-step stabilisation. The coupled solver, driver, and recorded run are shipped in repro/simulations_session/ so the bound is reproducible even though the converged value is not. The area-fraction value is therefore reported as a bracket, not a point, and graded [O] for the exact value (§26).

§29.4Why the [O] grades are not reproducibility failures

This volume's [O] grades are an epistemic cap, not a computational one: occurrence, rate, magnitude, order, and absolute date are held [O] by construction of the firewall (§1), for and against, regardless of how reproducible any number is. They are not entries in a calculation-blocked ledger; they are quantities the constitution forbids from carrying the argument. The one exception that is a computational limit — the coupled area-fraction value — is stated above. Everything load-bearing is present-tense and reproducible.

§30.1Graded ledger

§30.2Roadmap (next falsifiable tests, chronology-free)

FINALIZATION UPDATE (supersedes the priority notes below). The mechanism chain now closes logically (CG-36 / M20): the convergence is the forced downwelling limb of mantle convection, and the granite map (CG-35 / M19) shows the loop runs. Consequently CG-12 is resolved-in-direction (continent-scale felsic tracks the convergent wet path, not pure extension — strong form disfavoured) and CG-20 is largely dissolved (first downwelling flux-melts the hydrated skin). The single decisive next test is now R1 — the quantitative budget (does the loop distil the observed ~40 % felsic fraction, freeboard, and ~7 vs ~35 km Moho? a falsifiable number). Tasks 2-4 and 6 below remain useful consolidating checks. Task 1 (CG-12) and Task 5 (CG-20) are retained for the record but are no longer the decisive fork.

v1.2 UPDATE (supersedes the note above; the prior note is retained per mark-never-erase). R1 is now addressed on present-tense grounds (CG-37 / M21 / screen 17): isostasy + the measured water volume reproduces the observed freeboard (within measured crustal density/thickness), the bimodal hypsometry (~5 km separation), and a sustainable steady-state volume — no fitted parameter, no deep-time curve. The area-fraction value (CG-38 / M22 / screen 18) is now understood as a self-organized marginal-connectivity (percolation) attractor (planar f\*≈0.407; 2D self-organized band 0.39–0.41; a validated 3D convection solver, Ra_c≈657.5, shows the attractor survives but the value is coupling-sensitive, 0.24–0.41). R2 (identity) is resolved by the three marginal attractors (substrate/freeboard/area). The single decisive open residual is now the two-way-coupled 3D run (continents ↔ convection) to pin the area-fraction value; the validated solver is shipped, so it adds only a continent buoyancy term. Occurrence/timing stays [O] forever.

v1.3 UPDATE (supersedes the v1.2 note above; the prior notes are retained per mark-never-erase). The two-way-coupled 3D run is now EXECUTED (CG-38 updated; M22 layer 4; simulations_session/coupled_rbc3d.py + coupled_driver.py). The coupled solver is re-validated (C=0 → Ra_c=657.5 exact). Result: the coupled continental fraction settles ~0.30 (~0.27–0.36), robust to coupling strength — it confirms the frozen-flow lower bound and the percolation mechanism, but does NOT reach the observed 0.41 (falsification = discovery). The percolation attractor is thus a robust bound (~0.25–0.41) across three settings (2D 0.39–0.41 / 3D-frozen 0.24–0.32 / 3D-coupled ~0.30). The residual is reframed: not "run the coupled case" (done) but supply the omitted continental rheology + convergent-margin-concentrated production (a passive tracer disperses under vigorous stirring; real Earth's far higher Ra would lower it further), plus a numerically-stable high-Ra long run (at Ra=10⁴, 64×64×12 is stable only for a finite window — even the q=0 baseline eventually blows up). Grade stays [L]; occurrence/timing stays [O] forever.

v1.4 UPDATE (supersedes the v1.3 note above; prior notes retained per mark-never-erase). The NT-1/NT-2 falsifier surface is now machine-gated (repro/no_tuning_falsifier_screen.py, screen 19, SEED=19 double-SHA-256). NT-1 (no fitted parameter) and NT-2 (no second engine: the convergence is the forced downwelling limb, Ra/Ra_c ≈ 8900) survive their falsifiers on measured present-tense input, and each test is non-vacuous — a counterfactual fires every one of F1–F4 (e.g. ρ_c=3400>ρ_m flips the freeboard sign; Ra=100<Ra_c would force an added convergence mechanism; a compensated 5 km bare-ocean high would break CG-24). The single open item — the area-fraction value (coupled passive tracer ~0.30 < observed 0.41) — is registered as CG-39: the named residual requiring the omitted continental rheology + convergent-margin-concentrated production, reported open (F5), not laundered into a pass. The architectural claim is documented in NO_TUNING_THESIS.md (parallel to WHITEPAPER/GOVERNANCE). (Registry note: screen 19 is registered in the README screen table, 00_MASTER_SEED.md §4, and MANIFEST.sha256 as of the v1.4 re-seal.) Grade stays [O] for the residual / [L] for the bound; occurrence/timing stays [O] forever. falsification = discovery.

v1.5 UPDATE (supersedes the v1.4 note above; prior notes retained per mark-never-erase). The CG-38 residual is now tested, not just named (CG-39↑; M23; screen 20; simulations_session/continental_rheology.py + coherence_ceiling.py). Continental coherence (a real rheology resisting dispersal, κ scanned not tuned) lifts the passive 0.29 toward the percolation ceiling — recovering ~75% of the 0.30→0.41 gap but saturating ~0.38 (uniform)/~0.34 (convergent), and a deterministic cat-map loop shows coherence cannot exceed the ceiling ~0.41. Both CG-39 forks are partly right: the kernel's percolation margin sets the binding bound 0.41 (observed 0.41 is on the threshold), and rheology is the confirmed lever toward it; the sharpened residual ~0.38→0.41 needs a true yield rheology + sphere geometry + internal heating. Grade [L]; the experiment confirmed the direction and did not rubber-stamp 0.41 (falsification = discovery). (Registry: screen 20 + module 23 registered in README, 00_MASTER_SEED.md §4, MANIFEST.sha256 at the v1.5 re-seal.) Occurrence/timing stays [O] forever.

v1.6 UPDATE (supersedes the v1.5 note above; prior notes retained per mark-never-erase). The v1.5 conjecture — that a true yield rheology closes the ~0.38→0.41 residual — is tested and FALSIFIED (CG-39↑↑; module 23; screen 21; simulations_session/continental_yield.py). A finite yield strength scanned to the parameter-free rigid limit lifts f to only ~0.34 (uniform)/~0.33 (convergent) — perfectly rigid continents are still advected into the downwelling pattern and do not reach the ceiling. Two independent rheology encodings (mobility ~0.38, yield-rigid ~0.34) both saturate SHORT of 0.41, so continental rheology is a confirmed partial lever, not the closer; the percolation ceiling ~0.41 is the binding bound (not exceeded from above per screen 20, not attained from below here), with observed 0.41 on it. The residual is not rheology — it reopens toward the untested sphere geometry + internal heating, or the planar ceiling 0.407 is the bound a flat stirred box only approaches. Grade [O] closure / [L] bound; falsification = discovery — a prior conjecture of this package was machine-tested and rejected. (Registry: screen 21 + module-23 refinement registered in README, 00_MASTER_SEED.md §4, MANIFEST.sha256 at the v1.6 re-seal.) Occurrence/timing stays [O] forever.

v1.7 UPDATE (supersedes the v1.6 note above; prior notes retained per mark-never-erase). The CG-39 residual either/or is resolved (CG-39↑↑↑; module 24; screen 22). A full spherical convection solver is out of scope, so the 'sphere geometry' candidate is tested where it is decidable — the percolation ceiling depends on coordination z: f*=1−p_c rises ~0.41 (z=4) → ~0.50 (z=6) → ~0.61 (z=8). So 0.41 is not a geometry-universal bound; it is the z=4 (4-connected ocean) value, which the simulation grid uses and which matches observed 0.41 only for that connectivity. Reaching exactly 0.41 is therefore not a missing-convection-physics gap — sphere/heating would not 'close' it; the z=4 ceiling already IS 0.41, and the stirred undershoot (0.34–0.38) sits below it because stirring + a finite box hold it there. The match to observation is now a falsifiable empirical claim: the effective coordination of the real subductable-ocean network is ~4 (a naive plate-adjacency graph, z≈5–6, would predict ~0.45–0.50). The certainty of the value 0.41 is DOWNGRADED vs v1.5 — the mechanism (percolation bound) is [L], the specific value is [O], pending the real-network connectivity. falsification = discovery. (Remaining untested levers, named not faked: internal heating in the box, full spherical convection — both refine the approach to the z-appropriate ceiling, neither changes its z-dependence.) (Registry: screen 22 + module 24 registered in README, 00_MASTER_SEED.md §4, MANIFEST.sha256 at the v1.7 re-seal.) Occurrence/timing stays [O] forever.

v1.8 UPDATE — CG-39 CLOSEOUT (supersedes the v1.7 note above; prior notes retained per mark-never-erase). The connectivity question v1.7 left open is decided by exact topology (CG-39 ✓; module 25; screen 23). A sphere tiled into F plates with triple junctions satisfies Euler V−E+F=2 with 2E=3V, giving mean coordination z = 6 − 12/F — ~5–6 for any realistic plate count (5.2 at ~15 plates, 5.8 at 52). The natural plate network is ~6-connected, NOT 4. Its percolation ceiling is therefore **f*≈0.46–0.50, which overshoots observed 0.41. So the clean 0.41 = z=4 match (screen 22) is not the natural network's prediction — it predicts a higher ceiling, and observed 0.41 sits below it, in the band where stirring holds the fraction down (the 0.30–0.38 undershoot, screens 20–21). The honest CG-39 verdict: the continental fraction is a percolation-bounded quantity in the correct REGIME (a sub-majority ~0.4–0.5), which the kernel fixes; the EXACT value 0.41 = (connectivity ceiling ~0.46–0.50) − (a stirring undershoot), within a ~0.30–0.50 model bracket, and is NOT uniquely forced. Grade [L] regime / [O] exact value. CG-39 is CLOSED**: the residual was chased to its floor and the answer is regime-yes, exact-value-no — falsification = discovery. (Registry: screen 23 + module 25 registered in README, 00_MASTER_SEED.md §4, MANIFEST.sha256 at the v1.8 re-seal.) Occurrence/timing stays [O] forever.

1. CG-12 — continent-scale felsic from pure extension. The decisive open mechanism. Screen the felsic fraction producible by submarine extension alone (Iceland rhyolite % is the present-tense datum) vs what a deep-water/subduction-like path yields. If pure extension cannot reach continent-scale, the model requires a deep-water face — a real, testable fork. [L] → [F]/[O].
2. CG-8/CG-18 — push budget. Does the inherited unjamming engine deliver the stress the buckling bar requires at the observed fold wavelengths? Couple screen 1 to the engine's stress output. First place "huge push" becomes pass/fail.
3. CG-9/CG-10 — heat budget. Does shear heating + blanketing actually raise a wet lower crust over the ~650 °C solidus? A thermal-budget screen (no age). Pass → CG-10 firms to [F] standalone; fail → needs an external heat source (plume), one more card.
4. CG-15 — Moho test sharpened. Predict the freeboard-vs-crustal-thickness relation across real cratons/margins and check against measured Moho + elevation. Present-tense [V] test that can break the thin-skin/distilled picture.
5. CG-20 — first seed. The honest deepest open question; hold [O], do not paper over. falsification = discovery.

1. CG-24 falsifier (live) — search globally for a compensated multi-km bare-ocean high with a deep light root (small free-air anomaly, deep Moho, no felsic): it would break the compensation law. [DONE so far: central Indian Ocean fits (capped ~1-2 km); Gorringe fits (uncompensated +FAA). None breaks it yet.]
2. CG-22 sign-boundary (superseded note) — the buoyancy-sign boundary (where ocean skin flips net-negative) and test it against the smooth/crumpled map boundary: a present-tense [V] test that can break the gate reading.

PRIORITY (finalized): the mechanism chain closes (CG-36); the decisive remaining test is R1 — the quantitative budget (observed ~40 % felsic / freeboard / Moho). VP's distinctiveness is upstream (c²=B/ρ explains why the mantle flows; the water-world start supplies the flux water) — VP is the foundation beneath mantle convection, not a competitor (R2).

Recommended first cut: R1 — the budget screen (the natural screen 17). It is the only remaining test that can falsify the thesis on present-tense grounds; everything upstream of it now closes logically.