Universal Natural Intelligence

This is the document. Rendered from the repository at the commit above, with nothing rewritten for the web. A gate re-renders it on every deploy and fails the build if a single byte differs.

RENDERED FROM STRUCTURED DATA — this page is not a markdown file. Source: gpt/knowledge/K20-constants-ratios-and-nature-ledger.json in uni-cookbook at 575fc93d9d31. The sha256 recorded for this page is the digest of that JSON, not of this text. The renderer is generators/derive_docs.cjs; it copies fields and adds no facts. Every value below is a field of the source, verbatim, including the ones that read NOT RUN and PENDING.

The 88 claims the ledger's own rule ruled out, each with why it is inadmissible and the receipt. This page is the ledger doing its job.

88 rows.

claim	why inadmissible	receipt	origin
A4 = 432 "natural tuning"	class as written: INADMISSIBLE — names no measurand distinguishing it from any other convention	Name a measurable physical or biological quantity that differs between 440 and 432 Hz, and pre-register it.	ledger row (class INADMISSIBLE)
chakra frequency tables	class as written: INADMISSIBLE as physics — no measurand defined	May be recorded as an HONEST/cultural signal. Never as a TRUE/measured one.	ledger row (class INADMISSIBLE)
Δt_OPERA = 60.7 ± 6.9 (stat) ± 7.4 (sys) → 6.5 ± 15	class as written: INADMISSIBLE — OPERA Collab. (2011), arXiv:1109.4897; 2012 re-measurement	FAILED: loose GPS fibre connector (~73.2 ns) + master-clock oscillator off 0.124 ppm. Re-measured 6.5 ± 15 ns = consistent with zero. Receipt carried, claim withdrawn.	ledger row (class INADMISSIBLE)
b₁ = 0.5400 ± 0.0295 (0.5371 ± 0.0305 with T)	class as written: INADMISSIBLE as 'the exponent' — Kolokotrones et al. 2010; objection: MacKay 2011, J Theor Biol 280(1):194–6; reply: Deeds, Savage & Fontana 2011, 280:197–8 — concedes the artifact, calls it irrelevant to curvature; the class stands either way	Derive: under M'=kM, b₁' = b₁ − 2b₂log k.	ledger row (class INADMISSIBLE)
[CO₂] NASA row = 350	class as written: INADMISSIBLE as current — NASA NSSDC Earth Fact Sheet (mirror)	Already refuted by the NOAA row above.	ledger row (class INADMISSIBLE)
Bond albedo + T_e (NASA sheet) = 0.385 → T_e 247.3 (printed); 246.8 recomputed from its own S and α	class as written: INADMISSIBLE as current — NASA NSSDC Earth Fact Sheet (Colorado mirror) — same sheet, same defect as its 350 ppm CO₂ row	Superseded by modern CERES-era planetary albedo ≈0.29–0.31 and TSI ≈1361 W/m²; refute those and this row returns and the greenhouse gap becomes ~41 K.	ledger row (class INADMISSIBLE)
Bathtub Coriolis (as told) = INADMISSIBLE	class as written: INADMISSIBLE — refuted by Ro ~3 × 10² above	Measure sink vorticity sign vs. hemisphere without controlling residual circulation.	ledger row (class INADMISSIBLE)
Jeans swindle = derivation linearises about a background that is not a solution	class as written: INADMISSIBLE as rigorous — Binney & Tremaine, Galactic Dynamics	A derivation retaining the background field that recovers λ_J.	ledger row (class INADMISSIBLE)
G_Polychaos = 670 pg (~655 Gbp)	class as written: INADMISSIBLE — BNID 104470, flagged 'dubious report / outdated value'	Re-measure single nuclei with modern methods.	ledger row (class INADMISSIBLE)
R_B/H,folklore = 10:1 — superseded	class as written: INADMISSIBLE as stated (no method, no denominator) — Sender et al. 2016 (the tracing)	Produce the original measurement it was said to rest on.	ledger row (class INADMISSIBLE)
R_glia,folklore = 10:1 — superseded	class as written: INADMISSIBLE as stated — Herculano-Houzel 2012, PNAS 109(S1):10661–8 (the tracing)	Produce the primary source it rested on.	ledger row (class INADMISSIBLE)
N_neurons,folklore = 100 × 10⁹ — superseded	class as written: INADMISSIBLE as stated — Herculano-Houzel 2012 (the tracing)	As above.	ledger row (class INADMISSIBLE)
N_beats,human@40yr = ~1.5 × 10⁹ (~1.3 SD, 'comfortably inside the band') — WITHDRAWN this pass	class as written: INADMISSIBLE as stated: the 40 yr was unsourced, and matches neither sourced quantity — above hunter-gatherer e₀ (21–37), well below modal adult lifespan (~7 decades) — previously 'Computed here'; the input was never sourced	Produce a sourced pre-industrial lifespan of ~40 yr that is a lifespan and not an e₀.	ledger row (class INADMISSIBLE)
'40–60 bits/s' conscious bandwidth = do not quote as fact	class as written: INADMISSIBLE as a measurement (admissible as an estimate, attributed) — Zimmermann via Nørretranders 1991, The User Illusion	Locate the primary experiment and its stated CI.	ledger row (class INADMISSIBLE)
n_gen,Rice&Salt as one merged figure	class as written: INADMISSIBLE — do not merge — —	Any single generation count offered for 'the Rice & Salt experiment' without naming which of 1988 / 1990 it belongs to.	ledger row (class INADMISSIBLE)
'more evolved'	class as written: INADMISSIBLE — Gould 1996, Full House	Specify an environment-independent measurable. None is known.	ledger row (class INADMISSIBLE)
The golden ratio is a universal design law of nature.	UNFALSIFIABLE AS STATED and cherry-picked: it names no measurand, no scope, and no observation that could refute it. 'Approximately φ' absorbs any value, and the examples are selected after the fact from an unbounded pool. It fails on unfalsifiability, before any question of truth arises — the claim survives every counterexample by redescription, which is the signature of an unfalsifiable claim.	Carried beside its EARNED neighbour, and the pair is the wing's template for every mystical-adjacent item: Douady & Couder (1992) produced golden-angle divergence in a PHYSICAL ferrofluid experiment from repulsion dynamics alone, with no golden-ratio input. And the universality version fails on its own poster child — 18% of scored sunflower counts (136/768) are NON-Fibonacci (Swinton & Ochu 2016, R. Soc. Open Sci. 3:160091). Honour the mechanism; reject the universal law. Recorded, NOT mocked: the underlying intuition points at something real, and the honest answer is the mechanism, not a dismissal.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
432 Hz is nature's / the universe's true tuning.	Names no measurand that could distinguish it from any other convention. Structurally, 432 Hz HAS UNITS — a dimensioned quantity cannot be a scale-free law, because its numerical value changes with the unit: 432 Hz is also 25,920 min⁻¹, and nothing about nature changed. A universal claim must be dimensionless or it is not a universal claim.	Concert pitch is a HUMAN ADMINISTRATIVE CONVENTION. A440 was chosen at a London conference in 1939 (Swinburne's stated argument: 440 factors conveniently and is easy to synthesise electronically), adopted by ISO as Recommendation R 16 (1955), formalised as ISO 16 (1975) at 440 ± 0.5 Hz. Before that France standardised 435 Hz in the 1860s; Austria recommended 435 in 1885. There is no privileged reference in the physics: a pitch standard is a CHOICE OF ORIGIN, and nature has no opinion about where you put your origin. The common 'derives from the Schumann resonance' justification does not survive its own arithmetic: 7.83 × 55 = 430.65, not 432 — and the near-miss is then rounded TOWARD the desired conclusion, the diagnostic signature of post-hoc numerology. What the studies actually show: Calamassi & Pomponi (2019), Explore 15(4):283–290, a double-blind cross-over pilot, n = 33, reported slightly lower heart and respiratory rate under 432 Hz; a CORRIGENDUM was issued (Explore 16(1):8). A single underpowered pilot with a correction is a reason to run a bigger trial, not a physics claim. MAY be recorded as an HONEST/cultural signal — a tuning preference is a real thing that real people really have — NEVER as a TRUE/measured one.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Chakra frequency tables (Hz values assigned to chakras).	INADMISSIBLE AS PHYSICS: no measurand is defined. The tabulated Hz values are not measurements of anything; no instrument produces them and no observation could contradict them.	No measured mechanism assigns those frequencies to those anatomical sites. They MAY be recorded in the HONEST store as an attributed cultural teaching — verbatim, with attribution, never asserted, never mocked — exactly as this program's doctrine handles every attributed teaching. They may NEVER be admitted as a TRUE/measured signal, and may NEVER be wired to a computation or used as a proxy for a measured quantity. Crossing an honest signal into the true store is the cardinal sin; a chakra Hz table entering a calculation IS that crossing. Keeping the stores separate is what lets the honest signal be respected exactly as lived rather than quietly deleted.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Schumann 7.83 Hz entrains the human brain / has human health effects (because alpha is ~8–13 Hz).	INADMISSIBLE AS STATED. Both frequencies are real and separately replicated. The BRIDGE names no mechanism, no dose–response, and no falsifier. The 'alpha ≈ 8 Hz ≈ SR 7.83 Hz' argument additionally leans on a band edge that different labs place anywhere between 7.5 and 8.5 Hz — A COINCIDENCE BETWEEN A MEASUREMENT AND A CONVENTION IS NOT A FINDING.	The geophysics is EARNED and stays earned: ~7.83 Hz is a measured Earth–ionosphere cavity resonance. The health claims are a SEPARATE claim on separate evidence. The physical starting point for anyone who wants to make the case: the SR magnetic amplitude is ~1 pT against a static geomagnetic field of ~30–50 µT — about 10⁷× LARGER — plus every field the subject already lives in. Any mechanism must explain how a signal 7 orders down is detected THROUGH the larger one. The effect size at ambient pT strengths is NOT-MEASURED (no pre-registered, independently replicated dose–response located). NOT REFUTED — not established, and not carried. If re-stated with a measurand and a pre-registered dose–response it becomes admissible as a HYPOTHESIZED row. It is not there yet.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Everything is vibration / everything has a frequency.	UNFALSIFIABLE AS STATED: it names no system, no observable, no scope, and no outcome that could fail. There is no measurement whose result would be 'no'. A statement that cannot lose cannot inform — it has zero surprise and therefore zero evidential content.	A FALSIFIABLE VERSION EXISTS and looks like this: 'System X, driven at f₀ ± δ, exhibits response amplitude ≥ A relative to a tuned off-resonance control at f₀ + Δ, with n ≥ N, pre-registered, sham-controlled.' That version can lose. Note what it costs: a named system, a named f₀, a named δ, a control, and a bar set before the run. St = 0.2–0.4 CAN LOSE. 'Everything has a frequency' CANNOT. That is why one is in the table and the other is here.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Resonance transfers to soft tissue ('find the resonant frequency and it will do X').	A real, load-bearing engineering fact for STRUCTURES that almost never transfers to soft tissue — and the reason is mechanical, not rhetorical.	Soft biological tissue is viscoelastic and heavily damped — low Q, high viscous loss. Energy is dissipated WITHIN the cycle instead of accumulating ACROSS cycles, so there is no sharp peak to find and no build-up to exploit. Wakeling, Nigg & Rozitis (2002), J. Appl. Physiol., report that soft-tissue vibration is damped and that MUSCLE ACTIVITY INCREASES THE DAMPING — the body actively suppresses the very resonance the claim depends on. The numeric Q of soft tissue is NOT-MEASURED (NA-06 obtained no sourced value and will not supply one). The tissue that DOES resonate sharply — cochlear hair cells, the basilar membrane's place code — does so because it has evolved dedicated high-Q machinery with active amplification: specialised structure, not general principle.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Crystals hold/emit healing or scalar energy; crystal grids; quartz as a consciousness amplifier.	INADMISSIBLE AS STATED: no mechanism is specified and no observation is named that could refute the claim.	Recorded, not mocked. The EARNED neighbours are close by and are worth more than the claim: quartz PIEZOELECTRICITY is real, measured, and runs the clock in the device you are reading this on; and the P–T phase surface really does let a mineral report the conditions of its own growth (kyanite/andalusite/sillimanite). The distance between those and a 'crystal grid' is the entire method. May be recorded as HONEST/cultural signals — real aesthetic and personal meaning, honestly held — never as TRUE/measured ones. Separate stores, never merged.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Water memory (Benveniste).	INADMISSIBLE AS STATED.	THE RECEIPT IS THE LESSON, NOT THE VERDICT. Davenas et al. (1988), Nature 333:816–818, reported human basophil degranulation triggered by anti-IgE antiserum diluted past the point where any antibody molecule could remain, concluding the CONFIGURATION OF WATER was biologically active. Nature sent Maddox, Randi & Stewart, who re-ran the assay IN THE SAME LAB WITH THE SAME PEOPLE and published 'High-dilution experiments a delusion', Nature 334:287–290. The single load-bearing change was BLINDING: sample codes were sealed and taped to the ceiling, revealed only after degranulation was scored. Under sealed codes the effect vanished. Hirst et al. (1993), Nature 366(6455):525–527, independently found their results contained a source of variation they could not account for, but NO aspect of the data was consistent with the previously published claims. THE LESSON IS NOT 'THEY WERE FRAUDS' — it is that an effect which exists unblinded and dies under sealed codes was located in the OBSERVER, and that the people who found this out did so by RUNNING THE EXPERIMENT rather than by arguing. Physical anchor: liquid water's H-bond network loses its own configuration in ~1.2 ps (Fecko et al. 2003, Science 301:1698–1702).	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Emoto's crystal / intention claims (that intention changes water's structure).	INADMISSIBLE AS A CLAIM ABOUT WATER — and the honest handling is subtler than 'it failed', because ONE CONTROLLED TEST EXISTS AND REPORTED A POSITIVE RESULT.	Radin, Hayssen, Emoto & Kizu (2006), Explore 2(5):408–411, had ~2,000 people in Tokyo direct intention at water in a shielded room in California; crystal images were rated by 100 independent judges, and treated-water crystals scored higher for AESTHETIC APPEAL (P = .001, one-tailed). THE RECEIPT IS WHAT THAT MEASURED: the dependent variable was a SUBJECTIVE AESTHETIC RATING OF SELECTED PHOTOGRAPHS — not any physical property of water or ice. No density, no spectrum, no diffraction, no nucleation temperature. Emoto's protocol has photographers SELECT the most pleasing crystals; ice crystal habit is exquisitely sensitive to supercooling, nucleation and trace contaminants, and Tiller's noted objection is that supercooling was NEITHER CONTROLLED NOR MEASURED. Add: the claimant is a co-author, and no replication outside the proponent group was located. So the claim 'intention changes water's structure' WAS NEVER TESTED. A claim about aesthetic ratings of chosen images was. That is a specific, checkable defect — not a slur. NOT claimed: that Radin et al. 2006 is fraudulent, incompetent, or unpublished.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Structured / EZ water' as a health claim.	SPLIT IT — and the split is the whole point. Collapsing the three parts into one verdict, IN EITHER DIRECTION, is the defect.	(a) EARNED: the EXCLUSION ZONE — a layer near hydrophilic surfaces (notably Nafion) from which plastic microspheres are excluded — is REAL and independently replicated by ~10 laboratories (Elton, Spencer, Riches & Williams 2020, Int J Mol Sci 21(14):5041). A genuine phenomenon in search of an explanation. (b) CONTESTED: Pollack's 'fourth phase / structured water' EXPLANATION is disputed — the same review reports flaws in the supporting birefringence measurements and argues Schurr's diffusiophoresis account explains things Pollack's cannot (EZ growth time-course, pH gradients from the Nafion surface, the optical-tweezer force field). (c) INADMISSIBLE: the HEALTH claim — that drinking 'structured water' does anything. The EZ is a bulk INTERFACIAL effect at a SYNTHETIC hydrophilic polymer; no measurement of any bulk property of ingested structured water, nor of any clinical endpoint, was located (NOT-MEASURED). The phenomenon is real; the explanation is contested; the health claim is neither.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Water spirals down the drain one way in the north and the other in the south.'	INADMISSIBLE AS STATED — and the receipt is a dimensionless number, not a sneer.	Ro = U/(fL) ≈ 3 × 10² for a sink: rotation is ~300× too weak to compete with the residual circulation already in the basin from filling it, its geometry, or the plug being pulled. BUT THE UNDERLYING PHYSICS IS REAL AND WAS MEASURED. Shapiro (1962, Nature 196:1080–1081) built a covered 6-ft tank, filled it with a deliberate CLOCKWISE swirl, let it settle 24 HOURS to kill U, and drained it: the float sat motionless for 12–15 minutes, then turned COUNTERclockwise at ~1 rev per 3–4 s. A Southern-Hemisphere counterpart followed (Nature 207:1084). READ THE METHOD: the 24-hour settle is not fussiness, IT IS THE EXPERIMENT — driving U → 0 drives Ro → 0, the only regime where the effect can win. This is the exact shape of an earned/unearned split: the folk claim is false, the physics is true, and the difference is a controlled variable. NEITHER HALF MAY BE QUOTED WITHOUT THE OTHER.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The scale ladder is a ladder of increasing consciousness / awareness / mind.'	INADMISSIBLE: no measurand, no units, no instrument, no falsifier.	The 'Blanket?' column is not monotone in size — blankets appear at rungs 6 (organelle), 7 (cell) and 10 (organism) and VANISH above and below. A galaxy is enormous and has no blanket; a mitochondrion is a micron across and has one. Size is not what makes a boundary. So even the one structurally checkable property does not ascend with scale. And specific power runs AGAINST the ladder by ~5 × 10⁶ from bacterium to Sun. Recorded, NOT mocked: the intuition that the ladder points somewhere is honest and common. It is simply not a measurement, and dressing it as one is the defect. Writing the ladder as an ascent toward mind is a RHETORICAL MOVE DRESSED AS A MEASUREMENT.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The Earth/biosphere is an organism' (Gaia, as a literal claim). Also: 'Gaia is science' / 'Gaia is nonsense' — both inadmissible as stated.	INADMISSIBLE AS STATED at rung 13. It fails criterion (a) of a rung: no boundary with demonstrated conditional independence is exhibited, and the claim names no falsifier.	THE THREE PARTS TRAVEL SEPARATELY OR THE FENCE HAS FAILED. (1) OBSERVED-REPLICATED: Earth's atmosphere is wildly out of thermodynamic equilibrium (O₂ and CH₄ coexisting is the standard exhibit) — this observation belongs to no theory. (2) HYPOTHESIZED: that the biosphere REGULATES the planet homeostatically 'by and for' itself. The serious objection is structural and unanswered — Doolittle (1981) and Dawkins (1982): natural selection acts on differentially reproducing entities; there is ONE Earth, no population of competing siblings, no reproduction, no inheritance with variation, so planetary homeostasis could not have been SELECTED FOR. Gaia as usually stated requires a selective process for which no unit exists. That is not hostility to the idea; it is naming the missing mechanism. (3) MODELED: Daisyworld (Watson & Lovelock 1983) is an EXISTENCE PROOF THAT A MECHANISM IS POSSIBLE without teleology — which genuinely defuses the 'Gaia needs foresight' objection — but it is NOT evidence that Earth runs that mechanism, and cannot in principle supply it. Gaia as a HEURISTIC about coupled biotic–abiotic feedback names real, measurable couplings and is admissible in that form. Contempt is a defect exactly as much as credulity is; the fence goes on the claim, not the claimant.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Emergence means the whole is more than the sum of its parts.'	INADMISSIBLE AS STATED: unfalsifiable — no observation is specified that could refute it.	The ADMISSIBLE neighbour has three checkable conditions, each with a test: (a) a boundary with real conditional independence forms (the Markov-blanket condition — a factorization claim about the dynamics, checkable); (b) a new characteristic timescale SEPARATES from the level below (Simon 1962, 'The Architecture of Complexity', Proc. Am. Philos. Soc. 106(6):467–482 — near-decomposability; timescale separation is not a description of hierarchy, it is what MAKES one); (c) the level has its own dynamics that do not require tracking the level below (Anderson 1972, 'More Is Different', Science 177(4047):393–396 — rejecting the constructionist hypothesis). Note the honest ordering: (b) CAUSES (c). Emergence here is a statement about SPECTRAL GAPS IN A DYNAMICAL SYSTEM, not about magic. Anderson (1972) is the earned version; the slogan is the unearned one. The distance between them is the method. Both the mystical reading and the dismissive reading are defects.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'A Markov blanket implies a self / awareness / sentience.'	INADMISSIBLE: unfalsifiable as stated — it names no observation that would refute it.	The definition quantifies over conditional independence in a partition and MENTIONS NO EXPERIENCER. The same formalism admits partitions for a Newton's cradle, a burning fuse and the Lorenz attractor — Beck & Ramstead (2025, arXiv:2502.21217) ran blanket detection on exactly those. IF A BLANKET IMPLIED A SELF, IT WOULD IMPLY ONE FOR THE BURNING FUSE. The claim is the Pearl→Friston slide (Bruineberg, Dołęga, Dewhurst & Baltieri 2022, BBS 45:e183 — a PEARL blanket is a formal statistical construct in a graph YOU DREW; a FRISTON blanket is a metaphysical claim that some real system genuinely HAS an inside, a surface and an outside; the field's main error is sliding from the first to the second without paying for it) with a further leap on top. Recorded, never asserted.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'A low F means the model is correct.'	INADMISSIBLE — non sequitur. F bounds surprisal UNDER THE ASSUMED GENERATIVE MODEL; the bound gap is unobservable without the posterior that was intractable to begin with.	F = D_KL[q‖p(s\|o)] − ln p(o): a confidently wrong q under a wrong p(o,s) can sit at low F. Minimizing F never tests whether the generative model was the right one — the whole construction is conditional on p(o,s), which is a CHOICE, not a measurement. Buckley et al. (2017), J. Math. Psych. 81:55–79.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The FEP is a falsifiable empirical theory of the brain, confirmed by the evidence.'	INADMISSIBLE AS STATED — conflates a model STRUCTURE with a model. Specific process-theory models are falsifiable; the structure is not, and cannot inherit their results.	Andrews (2021), Biol. Philos. 36:30, DOI 10.1007/s10539-021-09807-0 — demands of falsifiability on the FEP itself 'rest on a category error'; the FEP should designate a MODEL STRUCTURE onto which construals are added. Take this in BOTH directions: it defends the FEP from a bad objection AND concedes the thing that matters — A MODEL STRUCTURE IS NOT AN EMPIRICAL FINDING. Colombo & Wright (2021), Synthese 198:3463–3488, on its limits as a first principle; Colombo & Palacios (2021), Biology & Philosophy 36(5), on the statistical-physics foundations. Related, carried as CONTESTED not asserted: 'Markov blankets identify the physical boundary of an agent' (Bruineberg et al. 2022, BBS 45:e183) and 'the FEP's assumptions are general' (Aguilera et al. 2022, Phys. Life Rev. 40:24–50 — the blanket + solenoidal conditions hold only in a 'very narrow space of parameters' for linear stochastic systems; the paper drew commentaries and replies, and neither side is adopted).	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'UNI schedules its work by computing G(pi).'	NEGATIVE / drift — no G is computed as a scheduler anywhere in UNI; G(pi) is FRAMING VOCABULARY.	UNI's ledger is the authority on UNI build status. Scheduling is by PENDING-burndown, inadmissible-event catch, migration gating, and read-agency. Prose implying UNI RUNS EFE is drift and is a defect. Stated in NA-02 only to stop that chapter being misread as a capability.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'It looks like a Turing pattern, therefore it is one.'	INADMISSIBLE. Fitting an image is not identifying a mechanism.	The field says so in print: 'even if a spatial patterning is successfully reproduced by a reaction-diffusion model, it may not be clear whether or not a diffusion factor is responsible' — Kondo (2022), Development 149(24):dev200974. RECEIPT OF FAILURE: for ~40 years the theory ran ahead of any identified molecular system. The admissible form requires a DIAGNOSTIC PERTURBATION — an experiment whose outcome the Turing mechanism predicts and the alternatives do not: e.g. laser-ablate a square of zebrafish melanophores and the pattern REGENERATES, mapping short-range activation + long-range inhibition between cell types (Nakamasu et al. 2009, PNAS 106:8429–8434); excise a mouse palatal ruga and new Shh stripes appear NOT at the cut edge but as BIFURCATING stripes branching off the neighbouring stripe — the signature of reaction-diffusion and not of a pre-patterned map (Economou et al. 2012, Nat Genet 44:348–351).	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Turing requires D_inhibitor/D_activator ≥ 10, universally.'	INADMISSIBLE AS STATED. d_c is set by the KINETICS; there is no kinetics-independent threshold, so quoting 'you need 10×' as a law is a category error.	The defensible statements are: d > 1 STRICTLY for a two-species system (set D_u = D_v and the instability condition collapses to tr J > 2√(det J), which contradicts the stability of the well-mixed state — so equal diffusivities can NEVER give a two-species Turing pattern); and 'the required ratios are often implausibly large for similarly-sized morphogens' (the fine-tuning problem). AND EVEN THE STRICT REQUIREMENT IS RELAXED: networks including CELL-AUTONOMOUS (non-diffusing) nodes can pattern with EQUALLY diffusing signals, for any combination of diffusion coefficients (Marcon, Diego, Sharpe & Jaeger 2016, eLife 5:e14022).	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Waddington's epigenetic landscape as a measured object.	INADMISSIBLE AS PHYSICS. It is a metaphor. Nobody measured a landscape.	The honest modern statement: the bistability/bifurcation formalism is MODELED (valleys → stable steady states, ridges → unstable ones, valley-splitting → pitchfork bifurcations), and Ferrell (2012), Current Biology (PMID 22677291), worked it through for cell-fate induction and found the computed landscape DOES NOT qualitatively resemble Waddington's picture. The metaphor survives as intuition and fails as geometry. Recorded as a NEGATIVE for the PICTURE, not for the intuition.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'DNA is a program / source code / blueprint.'	INADMISSIBLE AS STATED. No falsifier accompanies it; it is a metaphor doing argumentative work.	The receipt is in the chapter's own numbers: 1–2% of the human genome is coding; a fork fern (Tmesipteris oblanceolata) carries 160.45 Gbp — 52× the human genome; an onion is ≥95% repetitive at 16.4 Gbp. THE C-VALUE PARADOX finishes the job: genome size does not track organismal complexity — not weakly, not noisily, NOT AT ALL. Keller, The Century of the Gene (Harvard UP, 2000), argues the CELL may as well be read as the program and DNA as (part of) the DATA; Peluffo (2015), Genetics 200(3):685–696, traces the metaphor's genesis and its failures — it omits temporality, mechanical forces in development, symbiosis, and environment. A GPT THAT THINKS DNA IS SOURCE CODE WILL DESIGN WRONG: it will look for a specification where there is a PRIOR, expect compile-time determinism where there is CONTEXT-DEPENDENT CHEMISTRY, and treat 98% of the substrate as dead weight to be optimised away. NOT claimed either way: the ENCODE >80%-functional claim is contested (Graur et al. 2013, Genome Biol Evol 5(3):578–590, argue it conflicts with the <10% under purifying selection); CN-05 takes no position and asserts neither 80% nor 10%.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Sperm race to the egg and the fastest wins.'	INADMISSIBLE AS STATED: it names no falsifier. 'Fastest' is unmeasured in vivo, and the tract is not a racecourse.	Smith et al. 2009 observed a 34 µm/s spread in progressive velocity among high-viscosity migrants that their kinematic parameters DID NOT EXPLAIN. The competing accounts of the ~10⁶-fold attrition (10⁸ launched → ~10² arriving; median 251 recovered from both tubes, range 79–1,386, Williams et al. 1993) remain UNRESOLVED and must be carried, not chosen: (1) SELECTION FILTER — the tract screens (Sakkas et al. 2015, Hum Reprod Update 21(6):711–726); (2) RAFFLE — under sperm competition, fertilisation probability scales with your share of sperm at the ovum; (3) TRANSPORT LOSS — most sperm are simply lost mechanically, and redundancy selects nothing. The human case does not obviously favour the raffle: relative testis mass is the classic discriminator (Harcourt et al. 1981, Nature 293:55–57) and humans sit intermediate-to-low. The race narrative is a story fitted to the outcome.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Kamikaze sperm': that a large fraction of the ejaculate is a specialised non-fertilising caste that blocks or kills rival sperm.	NEGATIVE, with a receipt — and this is the GOOD outcome of the method.	Moore, Martin & Birkhead (1999), Proc R Soc B 266(1436):2343–2350, mixed ejaculates from different males and looked for exactly the predicted effects (differential mortality, agglutination, morphological damage). They found VERY FEW significant changes in sperm aggregation or performance between different-male and same-male mixtures, and NONE CONSISTENT WITH the previously reported findings; they concluded that incapacitation of rival sperm 'seems an unlikely mechanism of sperm competition in humans'. MATCH THE WORDING TO THE SOURCE: 'very few significant changes … none consistent with' is not 'none' — some signal was seen, and the authors' claim is that none of it reproduced the predicted pattern, which is why their own summary hedges to UNLIKELY rather than REFUTED. The hypothesis made a testable prediction, the test was run, the prediction did not reproduce. Recorded because the idea still circulates in popular accounts as though it had survived.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The ant colony is conscious' / 'the colony is a mind.'	INADMISSIBLE: no observation is specified that would refute it. NOT EVALUATED — neither asserted nor denied.	The MEASURED facts are strong and need no help from the annotation: stigmergic coordination (the coordination state was moved OUT of the agents and INTO the medium — Grassé 1959, Insectes Sociaux 6:41–80); a multi-decade collective phenotype under selection (colony ~25 yr vs exterior workers observed ≤ 33 days — two orders of magnitude); threshold-based task allocation with no allocator. Recorded, not mocked: the question is not stupid, it is UNFALSIFIABLE AS POSED, and the fence is on the claim's form, not on the asker. Note the thesis is the REVERSE of a mind claim: no ant knows it, and nothing is asserted to know it in the ant's place.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Swarm intelligence shows emergence beats design.'	INADMISSIBLE AS STATED: names no metric, no baseline, no refuting observation.	Its admissible neighbour is the SOP result — a pre-registered benchmark, a competitor, and a published margin. And M7 applied to the wing's own showpiece gives the opposite answer where it was first pointed: Dorigo & Stützle (2018), Handbook of Metaheuristics, IN THE ORIGINATORS' OWN VOICE — Ant System 'did not prove to be competitive with state-of-the-art algorithms specifically designed for the TSP', and the fix was to make successors 'less and less biologically inspired and more and more motivated by the need of making ACO algorithms better'. Where ACO holds world-class results (sequential ordering problem, Gambardella & Dorigo 2000) the win rides on a HYBRID WITH A HAND-BUILT LOCAL SEARCH. Recorded NEGATIVE for TSP, POSITIVE for SOP-with-hybrid, and never as 'nature-inspired, therefore better'.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Any dinosaur figure quoted as precise with no method and no error bar ('T. rex weighed 7 tonnes', 'T. rex ran at 20 mph').	INADMISSIBLE — not wrong, UNEVALUABLE. And be precise about where the defect is: the defect in '7 tonnes' is NOT the number (~7 t sits inside Hutchinson et al. 2011's own published conclusion of 6,000–8,000 kg). The defect is that NO METHOD AND NO ENVELOPE IS NAMED, so no falsifier exists, so it is not a measurement.	The same specimen (FMNH PR2081, 'Sue') carries a published minimal-to-maximal envelope of 9,502–18,489 kg FROM A SINGLE STUDY — a 1.95× spread, on the same bones, by the same team, in the same paper — so any single unqualified figure is selecting from that range without saying so. M(FMNH PR2081) = 9,500 kg is a CLAIM; 'M = 9,502 kg (minimal volumetric model; authors' preferred region; plausible envelope to 18,489 kg maximal, which the authors judge less plausible; Hutchinson et al. 2011)' is a RESULT.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Dinosaurs were warm-blooded' / 'dinosaurs were cold-blooded.'	BOTH INADMISSIBLE AS STATED: both are underdetermined by current evidence and neither names what would refute it.	The admissible form carries the taxon, the proxy, the number, and the contest. State of play, honestly: multiple independent lines indicate metabolic rates and body temperatures well above modern ectotherms in at least some dinosaur lineages (Eagle et al. 2011, Science 333:443–445: 36–38 °C from clumped isotopes — but 4–7 °C LOWER than a mass-scaling model predicted, which is a falsification event); the ancestral reconstruction is disputed on measurement-uncertainty grounds (Wiemann et al. 2022, Nature 606:522–526 vs Motani, Gold, Carlson & Vermeij 2023, Nature 621(7977):E1–E3, a Matters Arising comment arguing a single relative-intensity value may correspond to metabolic rates differing FIVEFOLD; reply Wiemann et al. 2023, Nature 621:E4–E6); and the dispute is live. GIGANTOTHERMY is the null hypothesis you must beat before invoking endothermy (Paladino, O'Connor & Spotila 1990, Nature 344:858–860). Also retired: LAGs (lines of arrested growth) as evidence of ectothermy — Köhler et al. 2012, Nature 487:358–361, found cyclical growth is UNIVERSAL in homoeothermic endotherms. The bone rings are still there; THE INFERENCE FROM THEM IS GONE. Anyone who gives you a clean answer here is selling something.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Quetzalcoatlus weighed X kg.'	INADMISSIBLE AS A POINT ESTIMATE.	Published values span 70 → 544 kg (~8×) FROM THE SAME FOSSILS at a 10–11 m wingspan: ~70 kg (Chatterjee & Templin 2004), ~200–250 kg (the consensus cluster — Paul 2002; Witton 2008; Witton & Habib 2010; Martin & Palmer 2014), ~544 kg (Henderson 2010, whose analysis concludes it could not sustain powered flight at all). Quoting one reports a PREFERENCE, not a measurement. Quote the spread or say nothing — and quote it WHOLE: rounding 544 down to '~500' quietly reports a smaller disagreement than the literature contains, which is the same defect one rounding-step milder.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The blue whale is nature's most efficient / perfect design.'	UNFALSIFIABLE AS STATED: no observation is specified that could refute it.	The falsifiable neighbours are in the ledger and are what should be cited instead: energetic efficiency RISES with body size in rorquals and FALLS with body size in odontocetes (Goldbogen et al. 2019, Science 366:1367–1372) — a divergence, and the divergence is the finding. Recorded, not mocked.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Whale calls can be heard across an ocean' / 'whales talk to each other across the globe.'	NOT-MEASURED AS STATED, and routinely presented as if measured.	Payne & Webb (1971), Ann. N.Y. Acad. Sci. 188:110–141, CALCULATED a basin-scale range by propagating fin-whale-like 20 Hz signals through a deep-sound-channel model to the ambient-noise floor — they did NOT observe reception. Secondary sources render its headline as ~700 km, 3,500 mi, 4,000 mi and 13,000 mi — a spread of more than an order of magnitude, because the answer depends entirely on the assumed noise floor. WHAT IS MEASURED: whales EMIT at 189 ± 3 dB re 1 µPa @ 1 m (blue, 25–29 Hz) and 189 ± 4 (fin, 15–28 Hz), and hydrophones DETECT them at ~200 km (blue) and ~56 km (fin) — Širović et al. 2007, JASA 122(2):1208–1215. WHAT IS NOT MEASURED: that a whale receives, recognises, and acts on a conspecific's call at basin scale. EMISSION IS NOT COMMUNICATION, AND DETECTION BY AN INSTRUMENT IS NOT RECEPTION BY AN ANIMAL. The best evidence to date is correlational: Podolskiy et al. (2024), Phys. Rev. Research 6:033174, found dive synchronisation among 12 tagged bowheads at separations up to ~100 km persisting up to a week — but their tags recorded DIVES, NOT SOUNDS, so the acoustic mechanism is inferred and a shared environmental driver is not excluded. Recorded as an open question with a stated falsifier, NOT as folklore and NOT with contempt: Payne & Webb's model was good science that has simply never been closed.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Comparing underwater dB to airborne dB directly ('the sperm whale is louder than a jet engine').	NEGATIVE / units trap — wrong by 61.58 dB (~1.4 × 10⁶ in intensity). Not a finding; a missing impedance correction.	Underwater sound is referenced to 1 µPa; air to 20 µPa. Two corrections separate them: the reference-pressure ratio (20² = 400) and the acoustic impedance ratio (ρc water / ρc air ≈ 3600). Together: 10·log₁₀(400 × 3600) = 61.58 dB. Møhl et al. (2003), JASA 114(2):1143–1154, do the conversion THEMSELVES: their 235 dB re 1 µPa rms corresponds to 173 dB SPL re 20 µPa in air. TWO FURTHER UNITS TRAPS FROM THE SAME PAPER: (1) rms vs peak-to-peak — Møhl et al. report TRUE RMS and warn it is 'significantly different (yielding lower values) from the peak-to-peak measures, used in most of the literature on odontocete clicks', so a sperm whale click and a dolphin click quoted from two papers are frequently NOT IN THE SAME UNIT; (2) half-angle vs full beamwidth — they print a half-power HALF-ANGLE of ~4°, and beamwidths are conventionally quoted FULL WIDTH, so the −3 dB beam is ~8°: quoting '4°' is wrong by 2×. A convention is not a detail, and the trap catches the people who know about the trap.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Bats prove that consciousness / intention is required for intelligent behaviour.'	INADMISSIBLE as a claim about the measurements: it names no observation that could refute it.	Nagel's question (what it is like to be a bat) is a REAL philosophical question and is NOT answered, addressed, or bounded by any measurement in CN-10. Recorded, not mocked: it is a good question in the wrong ledger.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Bat sonar demonstrates a natural implementation of the free-energy principle.'	UNFALSIFIABLE AS STATED: no observation is specified that would show a bat NOT implementing it.	The admissible version is narrow and is what CN-10 claims: specific measured behaviours (Doppler-shift compensation, call-duration collapse across the attack, the terminal buzz, the ~6 dB/halving gain schedule) are CONSISTENT WITH ambiguity-minimising action, AND THE FIT HAS NOT BEEN QUANTITATIVELY TESTED. The difference between those two sentences is the whole method. Explicitly NOT claimed: that a bat COMPUTES expected free energy — nothing shows a bat evaluating G(pi) over a policy space. The bat emits, listens, and adjusts; 'minimising the ambiguity term' is OUR DESCRIPTION of its behaviour, in our vocabulary, for our purposes. THE BAT IS NOT DOING OUR ARITHMETIC. Also not claimed: that the active-inference reading is REQUIRED to explain any measurement — classical sensorimotor control, signal-detection theory and plain optimal-foraging accounts predict much of it, and no discriminating experiment is offered.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'10 bacteria for every human cell.'	INADMISSIBLE AS STATED: no denominator, no method.	Sender et al. (2016), PLoS Biol 14(8):e1002533, traced EVERY downstream citation to ONE 1972 back-of-the-envelope (10¹¹ bacteria per gram of gut content × 1 litre of alimentary capacity). That estimate was never wrong so much as it was NEVER A MEASUREMENT. Recorded, not mocked — and the popular retelling omits the honest twist: measured, it is ~1.3 : 1 against ALL human cells, but ~84% of those (2.5 × 10¹³) are ANUCLEATE red blood cells; against NUCLEATED cells the ratio is ~8:1. THE OLD 10:1 FOLKLORE WAS APPROXIMATELY RIGHT — FOR A DENOMINATOR IT NEVER STATED. The revision from 10:1 to 1.3:1 is not a wrong number replaced by a right one; it is an UNSTATED DENOMINATOR. Both ratios are defensible; neither means anything until you say which cells you are counting. (And: each defecation event reduces colonic content by a quarter to a third — a ratio that shifts ~30% after a bowel movement is not a constant of your being, it is a standing inventory.)	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'100 billion neurons and 10× as many glia.'	INADMISSIBLE AS STATED: Herculano-Houzel (2012), PNAS 109(S1):10661–8, could locate NO supporting reference — in her words, 'none other than ballpark estimates'.	Measured, by isotropic fractionator: 86.1 ± 8.1 × 10⁹ neurons and 84.6 ± 9.8 × 10⁹ non-neuronal cells — a glia:neuron ratio of ~1:1, not 10:1 (Azevedo et al. 2009, J Comp Neurol 513:532–541). THIS IS ONE OF THREE INDEPENDENT CASES OF THE SAME FAILURE MODE in the best-studied organism there is: '10¹²–10¹⁶ cells in the body' (no proper reference; measured 3.72 × 10¹³ / 3.0 × 10¹³), '10 bacteria per human cell' (one 1972 envelope; measured ~1.3:1), and this one. A number with no denominator, no method, and no source, repeated until repetition felt like evidence — each fell the first time anyone counted with stated assumptions. IF THE HEADLINE NUMBERS ABOUT US WERE FOLKLORE, TREAT THE INTERPRETATIONS BUILT ON THEM AS SUSPECT UNTIL EACH IS SEPARATELY SOURCED.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The human brain is anomalously large / uniquely encephalised for our size.'	NEGATIVE against the primate line — recorded as a published negative in the M15 sense: this is the credibility, not an embarrassment.	We carry the neuron count EXPECTED of a generic primate brain of our size — the human brain is an ISOMETRICALLY SCALED-UP PRIMATE BRAIN (Azevedo et al. 2009's title claim); it is not an outlier on the primate scaling line, it sits ON it. We are not first in relative brain size, and at ~1.5 kg the human brain is two- to threefold SMALLER than an elephant's (Herculano-Houzel 2012). The apparent specialness is real but RELOCATED: PRIMATES scale differently from rodents, packing neurons more economically as brain size grows; humans inherited that rule and are the largest-brained primate. THE DISTINCTIVE STEP HAPPENED AT THE CLADE, NOT AT US. Related NEGATIVE / inference defect: 'the brain uses 20% of your energy, therefore human cognition is special' — the premise is real, the conclusion does not follow, because per-neuron cost is CONSERVED across primates, so the 20% share follows ARITHMETICALLY from neuron count.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'~10⁹ heartbeats per lifetime is an invariant.'	OBSERVED-CONTESTED, AND LOOSE — inadmissible when quoted as a constant.	CV ≈ 77% on Levine's own 15-species mean (7.3 ± 5.6 × 10⁸), defended only 'within an order of magnitude'. And per Levine's Figure 1 caption, 'Most coordinates represent average values (4–6)', where refs 4–6 are a 1956 biological-data handbook, a 1971 Encyclopaedia Britannica, and a 1938 elephant ECG paper — the species coordinates are AVERAGED AND UNTRACED, and 'life expectancy' is left undefined per species. This is a FOURTH instance of the chapter's own spine (unsourced ballparks repeated until repetition felt like evidence), applied to the chapter's own comparator. Composing the MEASURED exponents also breaks the invariant: beats/lifetime ∝ M^(−0.25) · M^(+0.153) = M^(−0.097), a ~4.8× decline across 7 decades of mammalian mass — NOT M⁰. Humans sit ~4 SD outside at a modern lifespan, and ~3.3 SD outside at a SOURCED evolutionary one (modal adult age at death ≈ 7 decades, Gurven & Kaplan 2007). An outlier is DATA, not a trophy.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The inverted retina is fine / optimal, because Müller cells fix it.'	NEGATIVE / inference defect.	Franze et al. (2007), PNAS 104(20):8287–8292, is real and replicated — Müller glia act as living optical fibres, funnel-shaped, higher refractive index than surrounding tissue, aligned with the light path, carrying the image through the obstructing layers with low distortion and low loss. It REBUTS 'the inversion is a disaster'. It does NOT rebut 'the inversion is a FROZEN ACCIDENT': a patch that recovers most of the loss from a defect is EVIDENCE THE DEFECT WAS REAL AND WORTH PATCHING. Were the inverted retina an optimum, you would not need a dedicated glial waveguide array to undo it. And the patch is incomplete — the blind spot is still there, filled in by the visual system rather than fixed by the optics, and retinal detachment remains a failure mode this architecture invites. The cephalopod built the other arrangement independently: photoreceptors facing the light, axons exiting behind, NO BLIND SPOT. Same problem, same regime, two lineages — and one is wired backwards. This is the wing's standing counterexample to the Panglossian reading, in the organism most prone to being described as nature's summit.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'More evolved' / 'higher on the evolutionary ladder' / 'beyond human' as a scalar.	INADMISSIBLE: no environment-independent measurable is specified. Evolution is a fitness landscape RELATIVE TO AN ENVIRONMENT, not a ladder. If you cannot state the environment, you cannot state the fitness, and 'beyond' has no referent — every claim about a successor that does not name the selective environment it is a successor IN is not a weak claim, it is A CLAIM WITH NO TRUTH CONDITIONS.	Gould (1996), Full House: complexity has a LEFT WALL — life cannot be arbitrarily simpler than the simplest viable self-replicator. Start a random walk hard against a wall and the walk's MAXIMUM drifts away from the wall over time, not because the walk is directional but because the wall is the only barrier. The right tail extends; THE MODE DOES NOT MOVE. The mode of life on Earth is, and has always been, bacteria. Reading 'progress' off the growth of the right tail is READING A BOUNDARY CONDITION AS A DRIVING FORCE. The intuition that there is a ladder is the great chain of being — a pre-Darwinian idea that keeps returning because it is intuitive, not because it survived contact with data. AND THE CANDIDATE SCALARS HAVE ALREADY BEEN MEASURED TO FAIL, TWICE IN ONE YEAR: the African elephant has 257 × 10⁹ neurons (3× the human total) but 97.5% of them (251 × 10⁹) sit in the CEREBELLUM, and its cerebral cortex holds only 5.6 × 10⁹ — about a third of the human count (Herculano-Houzel et al. 2014, Front Neuroanat 8:46); the long-finned pilot whale has ~37.2 × 10⁹ NEOCORTICAL neurons, roughly TWICE the human figure (Mortensen et al. 2014, Front Neuroanat 8:132). Two 2014 papers, same journal, directly OPPOSED conclusions from compatible counts. EQ fails too: Deaner et al. (2007), Brain Behav Evol 70(2):115–124, found ABSOLUTE brain size the best predictor of primate cognitive performance while EQ and other body-size-corrected residuals were not strongly correlated — and EQ is not even well-defined until you declare the fitted exponent and coefficient, which are CHOICES.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The upload would be conscious' (substrate independence AS EXPERIENCE).	INADMISSIBLE-AS-STATED: no observation is specified that distinguishes a system that HAS experience from one behaving identically WITHOUT it.	This is a fence on the claim's FORM, not a verdict on its truth, and not a slight to those who hold it — SERIOUS PEOPLE HOLD THIS POSITION AND THE FENCE IS NOT A REBUTTAL OF THEM; IT IS A REQUEST. Give it a falsifier and it moves out of this section. Contrast the neighbouring claims that ARE falsifiable and are carried as HYPOTHESIZED: substrate independence AS FUNCTIONAL ORGANISATION (supported by a brain function realised on a materially different substrate at matched input/output; refuted by exhibiting a brain function provably unrealisable on any other substrate at any speed); and whole-brain emulation (supported by a simulation built from a fixed structural scan alone that reproduces an animal's behavioural repertoire without further tuning — the substrate now exists to try, via FlyWire's complete adult Drosophila connectome, 139,255 proofread neurons and >50 million synapses, Dorkenwald et al. 2024, Nature; refuted by exhibiting one functionally load-bearing variable not recoverable from any fixed structural scan).	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Conscious bandwidth is 40–60 bits/s.'	INADMISSIBLE AS A MEASUREMENT (admissible as an ATTRIBUTED ESTIMATE; never as a measured constant). The value DRIFTS across retellings — 16, 20, 40, 60 bits/s appear in different retellings of the SAME underlying estimate. A number whose value depends on who is retelling it is not a measurement.	Its lineage runs through Zimmermann's sensory-physiology estimate and Nørretranders' The User Illusion (1991). NOT MOCKED — it is a reasonable estimate that got LAUNDERED INTO A FACT BY REPETITION, which is a failure of citation hygiene, not of the estimator. The current serious treatment is Zheng & Meister (2025), Neuron 113(2) (a Perspective, i.e. a literature synthesis, not a new experiment): sensory systems gather ~10⁹ bits/s; human behavioural throughput ~10 bits/s. AND IT IS ALREADY FORMALLY CONTESTED: Sauerbrei & Pruszynski (2025), Nat Neurosci, 'The brain works at more than 10 bits per second', accept a ceiling on high-level cognition but argue unconscious real-time motor control — occupying most of the CNS — substantially exceeds it. Carry the dispute, do not average it: the contest is not noise, it is the observation that 'throughput' is not a single well-defined quantity for a brain.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Citing Bremermann's limit as a design constraint.	NEGATIVE / rhetorical trap. It is real (c²/h = 1.36 × 10⁵⁰ bits·s⁻¹·kg⁻¹) and it is ~28–37 ORDERS OF MAGNITUDE AWAY. A limit that distant CONSTRAINS NOTHING.	Recorded because invoking it SOUNDS like physics while doing no work. And the distance is itself an accounting choice, one level up: ~28 orders against the Landauer-ceiling accounting, ~34–37 against Laughlin's measured per-bit costs. Any bare single figure for it — a '~39 orders' drops straight out of an undeclared ~10¹¹ bits/s spike-accounting — is the same defect as a bare N above Landauer, and gets declared or it does not get quoted. THE CEILING THAT ACTUALLY BINDS IS HEAT: erasure costs energy, energy becomes heat, and heat leaves through a SURFACE growing as R² while the computing VOLUME grows as R³.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The brain is N orders of magnitude above the Landauer limit.'	NEGATIVE / accounting trap. Any single N quoted WITHOUT ITS ACCOUNTING is a choice wearing a measurement's clothes.	Across the MEASURED per-bit accountings alone (Laughlin, de Ruyter van Steveninck & Anderson 1998, Nat Neurosci 1(1):36–41 — ~10⁴ ATP/bit at a chemical synapse, 10⁶–10⁷ ATP/bit for graded signals or spike coding), N ranges over ~3 ORDERS (5.5–8.5) depending on which signalling mode counts as the bit operation. Against the Landauer-CEILING accounting (20 W ÷ k_B T ln2 = 6.7 × 10²¹ bit-erasures/s), N is 0 BY CONSTRUCTION — a perfect Landauer-limited machine sits 0 orders above the bound by definition, so that accounting is a ceiling, not a measurement of any brain. The full spread is ~5.5–8.5 orders. THE LOAD-BEARING DECISION — what counts as one bit operation in a brain — IS NOT-MEASURED. (Note also the per-bit costs are from a FLY retina and the ΔG anchor is a BACTERIUM: nothing establishes that a fly photoreceptor and a human cortical synapse are interchangeable per bit.)	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Hoyle predicted the carbon-12 resonance because carbon-based life exists' (the anthropic origin story).	INADMISSIBLE AS HISTORY.	Kragh (2010), Archive for History of Exact Sciences 64:721–751, doi:10.1007/s00407-010-0068-8, documents that Hoyle and his contemporaries DID NOT associate the level with life; the anthropic gloss was retrofitted in the 1980s. AND THE TRUE VERSION IS BETTER: 'we exist' is not a quantitative constraint and predicts NO NUMBER. THE COSMIC CARBON ABUNDANCE IS A MEASUREMENT, and it did the work — Hoyle reasoned that carbon exists in the observed abundance, the known triple-alpha rate cannot make it, therefore there must be an unknown resonance in ¹²C just above the ⁸Be + α threshold, near 7.68 MeV with spin-parity 0⁺. Dunbar, Pixley, Wenzel & Whaling (1953), Phys Rev 92:649–650, looked, and found it. Modern value: 7,654.07 ± 0.19 keV. An observation about nature — HOW MUCH CARBON THERE IS — predicted the excitation energy of a nuclear state. Stripping the mysticism off does not diminish the story: it is the difference between an anecdote and a method. Recorded because this is the most-repeated version of the best story in CN-04, and it is wrong.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'GW170817 measured where gold comes from.'	NEGATIVE.	Gillanders et al. (2021), MNRAS 506(3):3560–3577, searched AT2017gfo for EXACTLY THIS and found NO prominent Pt or Au signatures, reporting only upper limits (Au ≲ 10⁻² M☉, Pt ≲ a few × 10⁻³ M☉). What WAS identified is STRONTIUM — a Sr II P Cygni feature at ~8000 Å at 1.4, 2.4 and 3.4 days post-merger (Watson et al. 2019, Nature 574:497–500). THE HONEST CHAIN: strontium (Z = 38) was identified → an r-process runs in neutron-star mergers → gold (Z = 79) is an r-process product BY ABUNDANCE PATTERN AND NUCLEAR THEORY → mergers are INFERRED to be A site of gold production. Every link is defensible, but 'MEASURED' ATTACHES TO STRONTIUM AND NOTHING HEAVIER. 'We know where gold comes from because we saw it' is FALSE; 'an r-process runs in neutron-star mergers because we saw strontium, and gold is an r-process element' is TRUE — three links, one of them theory. Also contested: whether mergers are the DOMINANT r-process site (mergers take time to inspiral, yet r-process europium sits in very old metal-poor stars). Site identified; budget open. Recorded against CN-04's own commissioning brief, which contained the error.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'L ∝ M^3.5 is the mass–luminosity relation.'	NEGATIVE. The canonical 3.5 is a rough average over a range that should never have been averaged.	Eker et al. (2018), MNRAS 479(4):5491–5511, Table 4, fitted 509 main-sequence stars from detached eclipsing binaries over 0.179 ≤ M/M☉ ≤ 31 and found a SIX-PIECE relation with break points where the ENERGY-TRANSPORT MECHANISM SHIFTS: 2.028 / 4.572 / 5.743 / 4.329 / 3.967 / 2.865. NO PIECE IS 3.5, and the sequence is NON-MONOTONIC — the exponent runs from 2.0 to 5.7. The break points are the physics. Usable as an order-of-magnitude device; INADMISSIBLE as a measurement. Related INADMISSIBLE: 'stellar L ∝ M^α and Kleiber's B ∝ M^(3/4) reveal a shared scaling principle' — unfalsifiable in that form, and the two are OPPOSITE IN SIGN of α − 1 (Kleiber is SUBLINEAR: a bigger animal burns LESS power per kilogram; the stellar relation is wildly SUPERLINEAR). A power law is the generic output of a scale-free constraint; many unrelated mechanisms produce one, so a shared functional FORM is the weakest available evidence of shared mechanism. Recorded as a RHYME; the temptation is real and the refusal is the method.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'⁵⁶Fe is the most tightly bound nucleus.'	NEGATIVE as stated: true of one quantity, false of the other, and stating it without NAMING THE QUANTITY is the defect.	⁶²Ni has the HIGHEST BINDING ENERGY PER NUCLEON at 8.7945 MeV/nucleon, then ⁵⁸Fe at 8.7922, then ⁵⁶Fe at 8.7903 — first to third spans 4.2 keV/nucleon, 0.048%. ⁵⁶Fe has the LOWEST MASS PER NUCLEON — a DIFFERENT quantity — because ⁶²Ni carries a larger neutron fraction (34/62 vs 30/56) and neutrons are heavier than protons. Both are true of different quantities. (And stars do not make ⁵⁶Fe directly: silicon burning under nuclear statistical equilibrium favours the alpha-conjugate ⁵⁶Ni (Z = N = 28); iron arrives by decay, ⁵⁶Ni → ⁵⁶Co (t½ ≈ 6 d) → ⁵⁶Fe (t½ ≈ 77 d) — and that chain powers the Type Ia light curve.)	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
The Jeans swindle.	INADMISSIBLE AS A RIGOROUS DERIVATION, while remaining in use as a working criterion. Named and used, not hidden.	The Jeans criterion derivation linearises about a UNIFORM, STATIC, INFINITE background — but a uniform static self-gravitating medium is NOT A SOLUTION of the equations; it would itself be collapsing. The background field is silently set to zero to make the algebra work. The defect is known and named (Binney & Tremaine, Galactic Dynamics). The criterion survives because it approximates well where other methods can check it, NOT because the derivation is sound.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Two-thirds of Earth's minerals require life / an oxygenated biosphere.'	NEGATIVE / WITHDRAWN — the originating authors superseded it.	Hazen & Morrison (2022), Am Mineral 107:1262–1287, surveyed 57 paragenetic modes across 5,659 species and reported, EXPLICITLY CONTRARY TO PREVIOUS ESTIMATES, that only ~34% of mineral species form EXCLUSIVELY as a consequence of biological processes — naming the contrast with the two-thirds estimate of Hazen et al. (2008). THE SAME AUTHOR SUPERSEDED HIS OWN HEADLINE NUMBER, and it stood wrong for fourteen years until he re-audited his own claim against a bigger dataset — caught by the author, not by a critic. THAT IS WHAT A WORKING HONESTY FENCE LOOKS LIKE FROM THE INSIDE. What stands, restated correctly: at least 2,707 of 5,659 species (47.8%) form under biological INFLUENCE, >1,900 exclusively — a pervasive planetary biomarker. AND BIOLOGY IS NOT THE MAIN DRIVER, IT IS THE SECOND: water is first, at AT LEAST 4,583 species (81.0%) via water–rock interaction. Both are the authors' explicit LOWER BOUNDS, so this ranks two floors — the margin between them is NOT a measured quantity. And it rests on ONE unreplicated single-team survey using the authors' own PROPOSED 57 paragenetic modes ('we welcome additions and corrections'), so it is fenced MODELED, not OBSERVED-REPLICATED. THE 2022 AUDIT HAS NOT BEEN INDEPENDENTLY CHECKED EITHER — saying so does not weaken the self-correction lesson; it IS the lesson, applied one turn further.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Arithmetic on Mohs hardness numbers (averaging, differencing, taking ratios).	NEGATIVE / scale-abuse. The Mohs scale is ORDINAL and nothing more: the scale is not linear, the integers do not license interpolation, and the scratching procedure was never well defined. Mohs 4 is not twice Mohs 2; there is no meaningful 'mean hardness' of a rock.	RECEIPT (single-crystal, uncontaminated): kyanite (001), at Mohs 5–5.5 — the LOWEST Mohs entry in Whitney, Broz & Cook's (2007, Am Mineral 92:281–288) Table 1 — is HARDER by DSI indentation (14.8 ± 1.4 GPa) than orthoclase at Mohs 6 (9.1 ± 0.6) and periclase at Mohs 6–6.5 (9.4 ± 1.4). A rank inversion. And kyanite carries Mohs 5–5.5 and Mohs 7 ON DIFFERENT FACES OF THE SAME CRYSTAL, with the modulus swinging 186 → 253 GPa: hardness is not a scalar property of a substance, it is DIRECTIONAL. RECEIPT EXPLICITLY NOT USED, recorded so nobody re-imports it: the periclase-vs-orthoclase MICROhardness inversion (5.3 vs 6.9 GPa) looks like the cleanest inversion on the table and is NOT — periclase is a POLYCRYSTALLINE SYNTHETIC reference material (the authors' own footnote: 'comparison with single-crystal values for MgO are more appropriate') while Mohs is a single-crystal scratch test, AND the DSI column of the same table PRESERVES the Mohs order for that pair (9.4 ± 1.4 vs 9.1 ± 0.6, overlapping). It points the right way and carries no weight. MECHANISM: scratch resistance is a COMPOSITE of hardness, fracture toughness AND elastic modulus — three properties collapsed into one integer, so the map is not invertible and no care recovers a physical quantity from a Mohs number (Broz, Cook & Whitney 2006, Am Mineral 91:135–142: none of the measured properties increases consistently or linearly with Mohs number across the scale). KEEP THE SCALE — a superb 200-year-old field test costing one fingernail. Just never do arithmetic on it.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The golden ratio governs crystal form.'	INADMISSIBLE AS PHYSICS: unfalsifiable as stated.	The genuinely interesting five-fold story is THE OPPOSITE OF MYSTICAL and it is in the ledger: five-fold symmetry is FORBIDDEN in all 230 periodic space groups; quasicrystals achieve it BY NOT BEING PERIODIC; and one occurs naturally — icosahedrite, Al₆₃Cu₂₄Fe₁₃, with six five-fold axes (Bindi et al. 2011, Am Mineral 96:928–931), in the Khatyrka meteorite (Bindi et al. 2012, PNAS 109:1396–1401). Real, strange, sourced, and it needs no help. RECORDED FAILURE OF THE FIELD, kept as the reference case for scope-forgetting: a REPLICATED observation (Shechtman, Blech, Gratias & Cahn 1984, Phys Rev Lett 53:1951–1953) was dismissed for years because it violated a completeness theorem whose SCOPE CONDITION ('periodic') had stopped being restated. Pauling: 'There is no such thing as quasicrystals. Just quasiscientists.' Shechtman took the 2011 Nobel Prize in Chemistry. Preserved not to sneer at Pauling but because THE FAILURE MODE IS THE LESSON: a proof is complete only within its axioms, and an axiom nobody says out loud becomes a prejudice — in the best-enumerated science we have.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Water has the highest specific heat capacity of any liquid.'	NEGATIVE / FALSE CLAIM — recorded because the claim is near-universal in teaching material.	LIQUID AMMONIA BEATS IT: 4.700 vs water's 4.181 J/(g·K), both liquids at 25 °C — with ammonia under its ~10 bar saturation pressure (normal bp −33.34 °C), a condition the compilations omit. (Whether ammonia leads water across its ENTIRE liquid range is NOT-VERIFIED; closure: fetch NIST Webbook cₚ along the NH₃ saturation line.) TWO FURTHER CONFLATION TRAPS: (1) PER MOLE, ETHANOL BEATS WATER — 112.4 vs 75.3 J/(mol·K); water's per-gram advantage is PARTLY JUST LOW MOLAR MASS (18.015 g/mol), and quoting the per-gram figure as though it revealed something about the H-bond network, without saying that molar mass is doing half the work, is overclaiming. The same applies to ΔH_vap: ethanol is 38.56 kJ/mol against water's 40.65 — NEARLY THE SAME PER MOLE — but 837 J/g vs 2256 J/g, so water wins 2.7× per gram ALMOST ENTIRELY ON MOLAR MASS. (2) WHAT IS ACTUALLY REMARKABLE IS THE CONJUNCTION: high per gram AND high per cm³ (4.17 J/(cm³·K), ~3,400× dry air) AND liquid across a 100 K span at 1 atm AND the most abundant liquid on the planet. No single one of those is unique.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Ice insulates the lake because ice is a good insulator.'	NEGATIVE / conflation trap.	ICE CONDUCTS ~4× BETTER THAN LIQUID WATER: ~2.2 vs ~0.561 W/(m·K), BOTH AT 0 °C — the temperature that actually matters for a freezing lake. (Quote the two at different temperatures and the ratio drifts: liquid water rises to ~0.607 at 25 °C, which would make it ~3.6×.) The insulation comes from SNOW (as low as ~0.074 W/(m·K)) and, above all, from KILLING CONVECTION. A conductive lid over still water is a far worse heat exporter than a convecting column. THE FLOATING IS WHAT BUYS THE STILL COLUMN. And the teleology is refused: water was not tuned for lakes — the density anomaly (max at 3.983 °C; a swing of just 1.32 × 10⁻⁴, 132 ppm) is a consequence of tetrahedral hydrogen bonding, which is a consequence of two lone pairs and two protons at 104.48°. Life occupied the niche the anomaly created. Per Gould & Lewontin, 'it works out well for us' is the exact shape of the error this book exists to catch.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Capillary action lifts water up a tree.'	NEGATIVE / conflation trap — off by ~75×.	Jurin's law in a 10 µm xylem lumen gives 1.47 m against a measured 112.7 m tree (Sequoia sempervirens). The nanometre-scale menisci in cell walls are OVER-specified by ~25× (a 5 nm pore gives ~2.9 km). Per Koch et al. (2004), Nature 428:851–854, the real ceiling (predicted 122–130 m) is set by LEAF WATER POTENTIAL and its downstream cost to photosynthesis, NOT by the meniscus failing. This is the NA-08 discipline verbatim: FIND THE LENGTH SCALE THAT ACTUALLY CARRIES THE FLUX BEFORE CONCLUDING THE PHYSICS BENT.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Oil and water repel' / the hydrophobic effect as repulsion.	NEGATIVE / the botched classic. There is NO repulsion.	Nonpolar solutes and water ATTRACT each other by dispersion forces like everything else. The segregation is driven by WHAT HAPPENS TO THE WATER, and Chandler (2005), Nature 437(7059):640–647, is explicit that the effect is MULTIFACETED — it depends on scale, and the crossover is at ~1 nm, which is (not coincidentally) the characteristic length of protein structure. SMALL SOLUTES (< ~1 nm) — ENTROPY-dominated: the hydrogen-bond network is NOT broken; in Huang & Chandler's words, 'hydrogen bonds simply go around the solute'. The cost is entropic — the solute restricts the configuration space available to the surrounding water's hydrogen bonding. Water pays in disorder, not in bond energy. LARGE SURFACES (> 1 nm) — ENTHALPY-dominated: 'it is impossible to maintain a hydrogen bond network adjacent to an extended surface'; water density is DEPLETED near the surface — it partially dries — and the free-energy cost becomes largely energetic, scaling with interfacial area. NOTE THE PATTERN: NaCl dissolves for ENTROPIC reasons too — dissolution is ENDOTHERMIC (ΔH_soln ≈ +3.9 kJ/mol; salt dissolving makes the water slightly COLDER, and you can feel it). Water does not 'pull the lattice apart' energetically; it barely breaks even — lattice enthalpy (+787 kJ/mol) and summed ion hydration enthalpies (Na⁺ ≈ −406, Cl⁻ ≈ −363) are two ~800 kJ/mol terms that NEARLY CANCEL. In both cases the naive energetic story is wrong, and in both cases THE SOLVENT — NOT THE SOLUTE — IS WHERE THE THERMODYNAMICS LIVES.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'CO₂ is 350 ppm' (or any undated CO₂ figure).	INADMISSIBLE. The defect is not the number; it is THE MISSING DATE ON A ROW THAT MOVES ~2 ppm/yr.	A live NASA-derived fact sheet carries 350 ppm while the measured Mauna Loa monthly mean for June 2026 was 431.44 ppm (NOAA GML, updated 05 Jul 2026) — roughly 80 ppm stale. AN UNDATED CO₂ FIGURE IS NOT A NUMBER. The discipline in miniature: a NASA composition table is a GOOD source for the rows that do not move and a WRONG source for the row that does. AND THE SAME SHEET IS STALE IN TWO ROWS, NOT ONE — it also prints Bond albedo 0.385 and S = 1367.6 W/m², giving a black-body temperature of 247.3 K and a greenhouse gap of ~41 K rather than 33 K. CN-03 made a virtue of catching the stale row that SUPPORTED its argument, so the stale albedo row, which does NOT, gets the identical receipt or the method was never a method. CATCHING ONLY THE STALE ROW THAT FLATTERS YOU IS NOT DISCIPLINE; IT IS PREFERENCE WEARING DISCIPLINE'S CLOTHES.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The sky is blue because of 1/λ⁴, therefore it should be violet.'	NEGATIVE / conflation trap — the cleanest example of A CORRECT LAW PRODUCING A WRONG ANSWER THROUGH A DROPPED TERM.	Rayleigh's 1/λ⁴ is right: violet (400 nm) against red (650 nm) → 7.0×; blue (450 nm) against red → 4.4×. So why is the sky not violet? BECAUSE 1/λ⁴ IS ONE FACTOR IN A PRODUCT. The solar spectrum has less power in the violet, and the human cone response is weakest there. SKY COLOUR IS THE SCATTERING CROSS-SECTION × THE SOURCE SPECTRUM × THE OBSERVER — and only the first is Rayleigh's. Note the law's own fence: it holds for scatterers ≪ λ. Cloud droplets are not — which is why clouds are white, and it is the same physics DECLINING TO APPLY.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Absorption in air goes as f².'	NEGATIVE as a general statement: true of the classical Stokes–Kirchhoff term, FALSE of real air below ~10 kHz, where N₂/O₂ VIBRATIONAL RELAXATION (catalysed by water molecules) dominates.	1 → 10 kHz at 20 °C / 50% RH gives 34×, NOT the 100× f² demands. And humidity dependence at 10 kHz is NON-MONOTONIC — 190 → 240 → 95 dB/km at 10 → 30 → 90% RH, PEAKING AT 280 near 20% RH — which no f² law can produce. One caveat usually dropped: absorption sits ON TOP OF geometric spreading (−6 dB per distance doubling), which is FREQUENCY-INDEPENDENT — at low f and short range spreading dominates and absorption is a rounding error; at high f and long range absorption is the whole story. Quoting either alone is the error.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Giant Carboniferous insects prove high O₂ allowed gigantism' — AND its mirror image, 'the 2026 paper shows oxygen was irrelevant.'	BOTH INADMISSIBLE. NOT admissible as settled, IN EITHER DIRECTION. Class: OBSERVED-CONTESTED — print both citations or neither.	FOR: Harrison, Kaiser & VandenBrooks (2010), Proc. R. Soc. B 277(1690):1937–1946, review the case that tracheal O₂ limitation constrains maximum insect size and late-Palaeozoic hyperoxia (GEOCARBSULF's ~30% Permian peak — itself a MODEL output, a mass balance over isotopic and weathering parameters, not a measurement) permitted gigantism; beetles' tracheal investment rises disproportionately with size, exactly the compensation the hypothesis predicts. Meganeuropsis permiana reached a ~71 cm wingspan. AGAINST: Snelling et al. (2026), Nature, doi:10.1038/s41586-026-10291-3, argue at the DELIVERY ENDPOINT — tracheoles occupy ~1% or less of flight-muscle volume in most species (capillaries take ~10× that in bird and mammal cardiac muscle), and relative tracheolar space scales up only ~1.8-fold across a 10,000-FOLD body-mass range, holding when extended to M. permiana. If O₂ delivery were binding, the giants should compensate at the tracheoles. They do not. BUT the 2026 result attacks ONE MECHANISM with a morphometric scaling measurement; it does not show O₂ was irrelevant to Palaeozoic insect size, nor refute the GEOCARBSULF curve. HONEST POSITION: the correlation is real, the classical mechanism is under serious and recent attack, the replacement is not settled.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Daisyworld shows Earth self-regulates.'	NEGATIVE / category error — and the most common one here.	Watson & Lovelock (1983), Tellus B 35(4):284–289, built a world of black and white daisies, each growing only for its OWN LOCAL FITNESS with no regulatory intent, and showed planetary temperature nonetheless stabilises across a range of solar luminosity. It IS an EXISTENCE PROOF THAT A MECHANISM IS POSSIBLE — global homeostasis emerging from purely local selfish fitness, without teleology — which genuinely defuses the 'Gaia requires foresight' objection. It is NOT evidence that Earth uses that mechanism, and it does not answer Doolittle and Dawkins in general: it builds a world where the coupling HAPPENS TO BE tight and sign-correct, which is an assumption about Daisyworld, not a finding about Earth. A MODEL SHOWING X CAN OCCUR AND AN OBSERVATION THAT X DOES OCCUR ARE DIFFERENT CLAIMS. Daisyworld supplies only the first, and CANNOT IN PRINCIPLE supply the second.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'10⁻¹⁰ is DNA polymerase's error rate.'	NEGATIVE / conflation trap — the single most common error in this material. It is the rate AFTER THREE LAYERS.	The polymerase alone is ~10⁻⁴–10⁻⁵ in vitro (Kunkel & Bebenek 2000; Kunkel 2004). THE PRINCIPLE THAT SURVIVES, AND IT GENERALISES FAR BEYOND BIOLOGY: nature did not build one perfect component. IT STACKED THREE CHEAP IMPERFECT ONES, each individually unimpressive, and multiplied their INDEPENDENT failure probabilities — a ~10⁻⁵ part, a ~10⁻² filter and a ~10⁻³ filter compose to 10⁻¹⁰. No single element in the chain is remarkable; THE CHAIN IS. If you set out to build a 10⁻¹⁰ component you will fail; if you set out to build three cheap layers whose failures are uncorrelated, you will succeed. THE INDEPENDENCE IS THE WHOLE TRICK — correlated failures collapse the product back toward the worst layer. BUT THE RUNGS ARE THE SOFTEST NUMBERS IN THE CHAPTER: the round decades are the REVIEW's convenience (Fijalkowska, Schaaper & Jonczyk 2012, FEMS Microbiol Rev 36(6):1105–1121 — explicitly hedged with its own 'roughly'), NOT Schaaper's measurements. Schaaper (1993), JBC 268(32):23762–23765, reports FOLD-DISCRIMINATIONS: base selection 200,000–2,000,000×, proofreading 40–200×, MMR 20–400×. Composing his MEASURED factors onto a 10⁻⁵ rung spans ~10⁻⁷·⁹ to ~10⁻⁹·⁹, reaching 10⁻¹⁰ ONLY AT ITS MOST GENEROUS EDGE. THE DECADES ARE ROUND BECAUSE THEY WERE CHOSEN TO LAND ON 10⁻¹⁰ — not because three measurements happened to multiply out that way. And St Charles et al. (2015), DNA Repair 31:41–51, measured in vivo base selectivity MORE THAN 100× HIGHER than in vitro. THREE POSITIONS SIT ON THE FIRST RUNG ALONE, unreconciled. Related trap: 1.2 × 10⁻⁸ PER GENERATION (Kong et al. 2012) set beside 10⁻¹⁰ PER REPLICATION — DIFFERENT DENOMINATORS; a generation contains many germline divisions, and their ratio means nothing.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
The 30 nm chromatin fibre drawn as an established in vivo structural level.	NEGATIVE. It is real IN VITRO. Its status as an IN VIVO structural level is contested and must not be drawn in a diagram without the fence.	TWO INDEPENDENT NEAR-NATIVE METHODS FAIL TO FIND IT. Maeshima, Hihara & Eltsov (2010), Curr Opin Cell Biol (PMID 20346642), asked directly — 'Chromatin structure: does the 30-nm fibre exist in vivo?' — and report that cryo-EM of VITRIFIED human mitotic cells, imaged close to native state, found NO 30-nm fibres. Ou et al. (2017), Science 357(6349):eaag0025, using ChromEMT electron tomography, found chromatin to be 'a disordered 5- to 24-nanometer-diameter curvilinear chain' packed at varying 3D concentration — again no 30-nm fibre. ChromEMT requires fixation, dehydration, heavy-metal staining and plastic embedding; cryo-ET does not. THE TWO METHODS HAVE DIFFERENT ARTIFACTS AND AGREE ANYWAY, which is why this is strong. Textbook diagrams showing it inside a living nucleus are AHEAD OF THE EVIDENCE.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
Polychaos dubium at 670 pg (~655 Gbp) as 'the largest genome'.	INADMISSIBLE. Still the top hit in popular sources; still wrong to cite.	BNID 104470 carries it flagged 'dubious report' and 'outdated value'; the measurement used 1960s methods assaying WHOLE CELLS RATHER THAN ISOLATED NUCLEI, and has never been repeated with modern methods. Excluding it, the records are Tmesipteris oblanceolata (160.45 Gbp) and Lepidosiren paradoxa (91 Gbp, sequenced). Recorded, NOT mocked — it was an honest 1960s measurement, and THE DEFECT IS IN RE-CITING IT IN 2026, not in having made it.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Ants solve the travelling salesman problem' / 'deposition + evaporation → shortest path.'	NEGATIVE / conflation traps. They do not, and it is not.	Goss et al. (1989), Naturwissenschaften 76:579–581, tested a TWO-BRANCH BINARY CHOICE at r ∈ {1, 1.4, 2} in Iridomyrmex humilis (= Linepithema humile), 11 colonies. That is ONE DECISION WITH TWO OPTIONS. The TSP is a different object, and the ACO that addresses it needed a graph, a memory, quality-proportional deposition, and evaporation — NONE of which the ants have or the 1989 model contains. AND THE MECHANISM IS NOT DEPOSITION-PLUS-EVAPORATION: the model that MADE the discovery EXPLICITLY IGNORES EVAPORATION (the experiment's timescale is of the order of the pheromone's ~30 min mean lifetime). The mechanism is a DELAY — ants returning along the short branch re-mark it after 20 s, ants on the long branch only after 20r s; between those two moments the short branch is marked AT BOTH ENDS while the long branch is marked at one. That asymmetry (Dorigo's 'differential path length effect') is amplified by the n = 2 autocatalysis in the choice function P_s = (20+S)²/[(20+S)² + (20+L)²]. EVAPORATION WAS ADDED BY DORIGO FOR A STATED ENGINEERING REASON: 'From a practical point of view, pheromone evaporation is needed to avoid a too rapid convergence of the algorithm towards a sub-optimal region. It implements a useful form of forgetting.' The engineers added forgetting BECAUSE THEY HAD SEEN THE TRAP — and the trap is measured: at r = 2 with the short branch added AFTER the long trail was established, the colony chose the short branch in only 2 OF 18 TRIALS. IT CANNOT SWITCH. The same n = 2 autocatalysis that finds the shortcut LOCKS IN THE FIRST ANSWER. Any distributed design lifted from this experiment inherits the lock-in.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Colony metabolism scales sublinearly (0.81)' as evidence for the superorganism reading.	NEGATIVE / the CI test applied to the wing's own showpiece. This is the exact defect M2 exists to prevent.	It is NOT what Hou et al. (2010), PNAS 107(8):3634–3638, established. THEIR OWN SENTENCE: the slope 'is statistically indistinguishable from the predicted value of 0.75, but it is also statistically indistinguishable from unity' — 95% CI 0.55–1.08, r² = 0.82, and n = 12 COLONIES in that figure. THE CI INCLUDES 1.0. The metabolic result does not establish sublinearity. THE ROW THAT CARRIES THE ARGUMENT IS PRODUCTION: colony biomass production scales as M^0.83, 95% CI 0.68–0.98 (n = 16) — that interval EXCLUDES 1.0. And Pequeno & Glazier (2025), J Anim Ecol 94(6):1285–1293, across 51 ant species, found the exponent DIVERGES with trophic level and caste polymorphism (herbivorous 0.69 [0.58–0.79] vs predaceous 0.81 [0.74–0.89]; monomorphic 0.75 [0.68–0.82] vs polymorphic 0.89 [0.79–1.00]) and REFUTE suggestions of a single colony-level metabolic scaling exponent. VERDICT: production scaling supports the superorganism reading; metabolic scaling does not settle it and there is no single exponent to appeal to.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Kleiber's law is a law.'	NEGATIVE — recorded as a naming defect with a receipt. It is a robust REGULARITY OVER A RANGE, which is a real and useful thing, and IS NOT A LAW.	51% of compiled exponents fall OUTSIDE 0.7–0.8, with the distribution running from <0.5 to >1.0 (Glazier 2005, Biol Rev 80:611–662, re-examining Peters's 146 relations — a compilation that is itself 72% vertebrate and 44% birds and mammals, in a world where most animals are invertebrates). Intraspecific scaling is worse: 0.3 to 1.8, mode 0.667 (Withers 1992, 220 species). THE SAME PAPER's binned and unbinned fits disagree on whether 3/4 survives: Savage et al. (2004) is titled 'The predominance of quarter-power scaling in biology', yet its own fit to all 626 UNBINNED species yields a CI of [0.699, 0.724] that REJECTS 3/4 — the 3/4 conclusion arrives only after binning by 0.1 log-unit mass intervals. THIS IS M2 IN THE WILD: point estimates of 0.712 and 0.737 look like agreement; their confidence intervals reach OPPOSITE VERDICTS. AND THE LOG-LOG RELATION CARRIES SIGNIFICANT CONVEX CURVATURE — it is not a power law at all (Kolokotrones et al. 2010, Nature 464:753–756: b₂ = 0.0322 ± 0.0053, P = 9.0 × 10⁻¹⁰). Fit a line to a mouse-heavy dataset and you get 2/3; fit one to a dog-and-up dataset and you get 3/4. Both fits are locally correct and globally meaningless. NEITHER TEAM WAS INCOMPETENT. BOTH WERE FITTING A LINE TO A CURVE. And the resolution is ITSELF contested (MacKay 2011, J Theor Biol 280(1):194–196), with a published reply by three of the four original authors (Deeds, Savage & Fontana 2011, 280:197–198) — so the claim, the resolution, AND the objection to the resolution are all contested. Most decisively, Kleiber is not universal ACROSS the ladder at all: DeLong et al. (2010), PNAS 107(29):12941–12945, fitted across ~16 orders of magnitude and found the exponent changes AT THE MAJOR EVOLUTIONARY TRANSITIONS — prokaryotes 1.7 (SUPERlinear), protists 1.0 (linear), metazoans 0.76 (sublinear). Quoting Kleiber's 3/4 as a law of life is the single most over-extrapolated number in this wing.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'b₁ = 0.54 is the metabolic exponent.'	INADMISSIBLE. It is the LOCAL SLOPE AT M₀ = 1 g and changes value if you measure mass in kilograms.	Under a change of mass scale M' = kM, b₁' = b₁ − 2b₂·log k. Only b₂ and the local slope AT A NAMED MASS are unit-invariant. MacKay (2011), J Theor Biol 280(1):194–196, raises this, and the reply by three of the four original authors CONCEDES THE ARTIFACT AND ARGUES IT IRRELEVANT — 'the only parameter necessary for the assessment of curvature in the data is b₂, which MacKay himself agrees is scale-invariant' (Deeds, Savage & Fontana 2011, 280:197–198). Read carefully, that ABANDONS b₁ RATHER THAN RESCUING IT. This is why NA-05 states local slopes at named masses (0.57 at ~3.6 g; 2/3 at ~160 g; 3/4 at ~4.2 kg; 0.87 at ~460 kg) instead of b₁. Note the last row of that ladder — slope = 1 at ~74 t — is an EXTRAPOLATION ~2.2 decades beyond the fitted data, which end at ~460 kg, and Kolokotrones et al. hedge it TWICE themselves: the unbounded slope rise 'may be due to the paucity of data for large animals', and the size-limit reading holds only 'If this is correct'.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'The model predicts 3/4 and we observe 3/4, therefore the model is right' (elastic similarity vs WBE).	NEGATIVE / the exponent does not discriminate — recorded because this is the most common malformed argument in this literature.	McMahon's elastic similarity (1973, Science 179:1201–1204) and WBE's fractal network (1997, Science 276:122–126) are INCOMPATIBLE MECHANISMS THAT BOTH PREDICT 3/4. Per M7, if a discriminator does not collapse the alternatives, the agreement is not a result. AN EXPONENT THAT MULTIPLE UNRELATED MECHANISMS PRODUCE IS WEAK EVIDENCE FOR ANY OF THEM. And elastic similarity MOSTLY FAILED ITS OWN EMPIRICAL TEST: Alexander et al. (1979), J Zool 189:305–314, measured femora, tibiae, humeri and radii across 32 mammal species from 0.020 to 3500 kg and found length ∝ M^0.31 and diameter ∝ M^0.35 — against elastic similarity's predicted M^0.25 and M^0.375, and isometry's M^0.333 for both. MAMMALIAN LONG BONES SCALE CLOSE TO GEOMETRIC SIMILARITY. The famous derivation of Kleiber from elasticity rests on a premise the bones do not obey. Separately, WBE's specific 1997 geometry is refuted on a SIGN ERROR, not a poor fit: for finite animals the model yields M = c₀B + c₁B^(4/3), and with both coefficients positive this predicts CONCAVE curvature. The data are CONVEX. What is refuted is the specific 1997 geometry, NOT network explanations in general — modified variants (moving the pulsatile/smooth flow transition a constant FRACTION of levels from the heart) do produce convex curvature and fit nearly as well as the quadratic. That distinction is the whole discipline.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'~10⁹ heartbeats' / any 'quarter-power family' member quoted as measured (heart rate ∝ M^(−1/4), lifespan ∝ M^(1/4)).	NEGATIVE / the family is quoted as if every member were measured. LIFESPAN ∝ M^(1/4) DOES NOT SURVIVE.	de Magalhães, Costa & Church (2007), J Gerontol A 62(2), fit 856 mammals (excluding cetaceans): t_max = 4.88·M^0.153 years, R² = 0.66. FOR 518 BIRDS: t_max = 5.22·M^0.218, R² = 0.70. NEITHER EXPONENT IS 1/4. The mammalian value, 0.153, is not close. Compose the two MEASURED exponents and the most famous invariant in comparative physiology comes apart: beats/lifetime ∝ M^(−0.25) · M^(+0.153) = M^(−0.097) — NOT invariant; it declines with mass by ~4.8× across 7 decades of mammalian mass. The 'constant ~10⁹ heartbeats' is a CONSEQUENCE OF ASSUMING the lifespan exponent is exactly +1/4, which cancels the −1/4 of heart rate. Insert the measured 0.153 and the cancellation fails. ALSO: there is no 'the' metabolic exponent even for one animal — VO₂max scales as M^0.872 across 34 eutherian species (7 g – 500 kg) against basal ~0.70; THE EXPONENT DEPENDS ON WHICH METABOLIC STATE YOU MEASURE. And the heart-rate PREFACTOR usually quoted alongside (f = 241·M^(−0.25) bpm) is NOT-MEASURED: the number 241 appears NOWHERE in the cited source (Lindstedt & Hoppeler 2023), and the Calder/Stahl primary was not read — SOURCING A VALUE TO NEITHER THE CITATION NOR THE PRIMARY IS THE THIRD STATE THIS WING FORBIDS.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
OPERA superluminal neutrinos (2011).	INADMISSIBLE — FAILED observation. Recorded with the receipt, NOT mocked.	OPERA (2011, arXiv:1109.4897) measured muon neutrinos arriving 60.7 ± 6.9 (stat) ± 7.4 (sys) ns EARLIER than light over the 730 km CERN→Gran Sasso baseline, i.e. (v−c)/c ≈ 2.48 × 10⁻⁵. Traced to a LOOSE GPS FIBRE-OPTIC CONNECTOR between the GPS antenna and the master clock (~73.2 ns) plus a master-clock oscillator off by 0.124 ppm in the opposite direction. The 2012 re-measurement gave 6.5 ± 15 ns: CONSISTENT WITH ZERO. The claim died. THE METHOD WORKED EXACTLY AS DESIGNED — the collaboration published the anomaly WITH its full error budget, explicitly DECLINED TO INTERPRET IT, and then published its own refutation. Both were correct scientific acts. And it worked BECAUSE the provenance set (what was measured, on what instrument, by whom, when, THROUGH WHICH CABLE) had been kept well enough to find the cable. THE ANOMALY'S DEATH WAS NOT A FAILURE OF THE METHOD; IT WAS THE METHOD.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'
'Einstein confirmed' (the 1919 eclipse).	Not the result. The verdict is the CI bound that excludes the threshold, NEVER the point estimate.	Two model predictions for starlight deflection at the solar limb: general relativity 1.75″, the Newtonian half-deflection 0.87″. Measured: Sobral 4-inch 1.98 ± 0.12″; Príncipe 1.61 ± 0.31″. Computing from those published values: Sobral sits ~9.3σ above the Newtonian value and ~1.9σ ABOVE THE GR VALUE; Príncipe ~2.4σ above Newtonian and ~0.5σ from GR. THE HONEST READING: 1919 SEVERELY EXCLUDED THE NEWTONIAN VALUE and was CONSISTENT WITH, BUT DID NOT PIN, THE GR VALUE. That is a real result, and it is not the result 'Einstein confirmed'. The excluding bound is the verdict; the point estimate is decoration. And note what travels with it: the Sobral 16-inch plate set was EXCLUDED as 'diffused and apparently out of focus'. That exclusion may well be correct on instrument grounds — and it is exactly the researcher degree of freedom that Simmons requirement #5 exists to force into daylight. AN EXCLUSION IS NOT A SIN. AN UNDISCLOSED EXCLUSION IS.	chapter section: 'Recorded INADMISSIBLE / NEGATIVE (first-class, inline)'

sha256 e443dca15319ea0b — of the original file, so what was ingested stays checkable.

INADMISSIBLE — claims the ledger rejected

How to read this page

Precise — the source document