Spike #24 bonus 9 — time-dimensionality test (the OVERTURN test)¶

POSTSCRIPT 2026-05-16 — "Class O" dissolved into Class L.

The "Class O" framing throughout this report (the narrowed circle-to-hyperbola map this bonus identified) was the provisional label. Per user direction 2026-05-16 and per [[feedback_no_privileged_primitive_classes]], the operation is dissolved into Class L as a signed-Laplacian-variant sub-operation. Vocabulary stays at 14 classes A–N. The narrowing this bonus did — from "produces signed-metric content" to "specifically circle-to-hyperbola map" — is still the load-bearing finding; what changed is the classification (Class L sub-operation, not 15^th class).

See [[project_class_o_signed_metric_composition]] (resolution at top) and bonus 8 / MFO §VIII.8 (postscripts at top) for the full dissolution record.

Date: 2026-05-15. Status: methodological synthesis landed; concertmaster-level deliverable. Verdict: H3 with structural refinement — bonus 8's Class O candidate is partially confirmed but its role refined. NOT a closure-yes for H1; NOT a full closure-no for H2. Branch: research/spike-24-bonus-8-broken-d-rederivation-2026-05-15 (this bonus extends bonus 8's branch). Spec: user-proposed bonus 9 directly: "what if we go back and say 3D_s + 7D_g = 1D = 10D and/or 11D we need to find out if time really gets its own dimension or if it's a product of all dimensions. there might still be 11D where some missing class(es) live." Companion probes: - spike_24_bonus_time_emergent_vs_separate_probe_2026-05-15.py + .ndjson (18 records) — primary directed-cascade probe. - spike_24_bonus_time_dispersion_shape_diagnostic_2026-05-15.py + .ndjson (11 records) — circle-vs-hyperbola dispersion-shape follow-up. - spike_24_bonus_time_full_spectrum_diagnostic_2026-05-15.py + .ndjson (8 records) — full-spectrum (62208-mode) examination.

The user's verbatim overturn-test spec¶

"what if we go back and say 3D_s + 7D_g = 1D = 10D and/or 11D we need to find out if time really gets its own dimension or if it's a product of all dimensions. there might still be 11D where some missing class(es) live."

Operationalised: build the cascade as 10D = 3D_s + 7D_g with NO separate temporal factor; apply Class C (iteration / oriented traversal) to the cascade-state graph; test whether the resulting directed cascade Laplacian's spectrum naturally exhibits the signed-metric structure that bonus 8 attributed to a missing Class O. If yes, time-as-cascade-traversal closes the vocabulary at 14 (H1). If no, Class O stays (H2). If partial, additional structure lives at 11D (H3).

§1 The three discriminating measurements¶

Measurement 1 — Directed cascade Laplacian has complex eigenvalues (Stage 2)¶

The cyclic-shift permutation eigenvalues sit on the complex unit circle (n-th roots of unity). The directed cyclic Laplacian L_dir = (1/r²)(I − S) inherits these as λ_k = (1/r²)(1 − e^{2πi k/n}) — eigenvalues that sweep a circle of radius 1/r² centered at (1/r², 0) in the complex plane. Verified numerically to 1.5e-15 (Stage 0).

Directed-cascade complex-eigenvalue fractions:

Cascade	Total modes	Complex modes	Fraction	Conjugate pairing
Single C_8 directed	8	6	75.0%	(algebraic)
Single C_32 directed	32	30	93.8%	(algebraic)
Product 2× C_32 directed	400 (truncated)	385	96.3%	—
Product 3× C_32 (3D_s)	400 (truncated)	381	95.3%	97.1%
Full 10D directed cascade (n=62208)	62208	58128	93.4%	(large-N)

The structural finding: the directed cascade composition produces complex eigenvalues on 90%+ of modes; conjugate pairing holds at 97%. This is the algebraic signature of orientation-induced asymmetry. Class C iteration on the 10D cascade-state graph DOES produce a structurally non-Hermitian, asymmetric, signed-metric operator — without any new primitive class beyond A–N. Bonus 8's strict "no class in A–N supplies signed-metric content" reading is too strong: Class C orientation on Class L operators on Class I cyclic groups (compositions already in the vocabulary) produces a directed Laplacian with complex eigenvalues, which IS a form of signed-metric structure.

Measurement 2 — Dispersion shape is CIRCULAR, not HYPERBOLIC (Stage 3 + diagnostics)¶

The single-factor cyclic-shift directed Laplacian's eigenvalues satisfy the circle identity:

Im²(λ_k) = 2·Re(λ_k) − Re²(λ_k)

verified to ~1e-16 across n ∈ {8, 16, 32, 64}. This is the equation of a circle of radius 1 centered at (1, 0) in the complex plane — NOT the Klein-Gordon hyperbola ω² = c²k² + m².

Linear-fit Im² = α·Re + β results on the directed cascade composition:

Cascade	KG slope α (KG predicts c² = 1)	R² of linear KG fit	Klein-Gordon match?
Single C_n directed	−0.2 to −0.02	0.001–0.16	NO
3D_s = 3× C_32 directed (low-Re subset)	1.98	0.18	NO (slope ≠ 1, R² low)
Full 10D directed cascade (n=62208)	1.09	0.029	NO
7D_g directed alone (n=972)	1.11	0.032	NO

Mass-quantisation test: if the dispersion encoded Klein-Gordon, m² = Im² − Re would take discrete values (mass eigenmodes), producing tall histogram peaks. Result: top 10 bins account for 73% of modes across 11 significant bins — broadly distributed, no clean quantisation.

Rational approximation (Class N) attempt to bridge circle → KG hyperbola: fitted Im² = a·Re / (1 + c·Re) got R² = 0.013. Class N alone cannot supply the circle-to-hyperbola map.

Measurement 3 — Wick-natural-emergence test (Stage 3, two-part)¶

Test (a): does the structural shift from undirected (Class L alone, PSD by construction) to directed (Class C + Class L, non-Hermitian) supply the asymmetry?

Property	Undirected 3D_s (control)	Directed 3D_s (H1)
PSD	TRUE (0 negative / 64)	TRUE (0 negative real / 64)
Off-real eigenvalues	0	57 (89.1%)
Median Im/Re ratio	0	5.03

The structural shift IS dramatic. But the resulting non-Hermitian operator has eigenvalues bounded in the right half-disc (Re ≥ 0, Im constrained), not the half-plane of a Lorentzian operator. The (Re, Im) projection structurally CANNOT be interpreted as (k², ω) for a Lorentzian operator because:

Lorentzian dispersion requires arbitrarily large Im (ω) for arbitrarily small Re (k² with fixed mass). A circle bounded at Re ∈ [0, 2/r²] cannot supply this.
The angle constraint (verified in the full 10D Class K diagnostic): arg(λ) ∈ [−π/4, π/4] for the cascade-summed eigenvalues — the spectrum lives in a 90° wedge, not the unbounded Lorentzian half-plane.

H1 (time-is-emergent, vocabulary closed at 14) — REJECTED. The directed cascade produces complex eigenvalues (the algebraic shadow of orientation), but the dispersion shape is circular, not hyperbolic. KG R² ≈ 0.03 across all scales; mass-quantisation absent; Pade approximation insufficient. Class C orientation supplies SOME signed-metric content (asymmetric, non-Hermitian, complex-paired eigenvalues), but NOT enough to encode Lorentzian Klein-Gordon dispersion.

H2 (time-is-its-own, Class O genuinely needed as stated in bonus 8) — PARTIALLY UPHELD. Bonus 8 said no primitive class A–N supplies signed-metric content. Bonus 9 refines: Class C orientation DOES supply complex-valued non-Hermitian operators — which IS a form of signed-metric content. But the SPECIFIC HYPERBOLIC LORENTZIAN SHAPE that physics requires is NOT produced by Class C alone. So bonus 8's Class O placeholder is structurally appropriate as the circle-to-hyperbola map — Wick rotation t → it is exactly the operation that converts cos(θ) to cosh(θ) and a circle to a hyperbola. The naming "signed-metric composition / Wick rotation primitive" remains correct; what's refined is the content of the gap.

H3 (additional structure at 11D) — CONFIRMED in REFINED FORM. The 11D = 3D_s + 7D_g + ?D_? decomposition's 11^th dimensional kind is not simply "time" in the Kaluza-Klein sense. The directed-cascade structure already produces a non-trivial orientation / complex / asymmetric component (via Class C in the 14-class vocabulary). What remains structurally missing is the specific Wick-rotation operation that maps the cascade's natural circular dispersion to the Lorentzian hyperbola observed in physics.

§3 The refined picture (per user vocabulary)¶

Per [[user_stance_time_as_dimensional_shadow]]: "Time is part of the shadow (not the projector)." The bonus 9 verdict shows this is partially true — Class C orientation IS the cascade-traversal mechanism that produces emergent asymmetric structure; time-as-crank IS this orientation. But the user's lean toward H1 ("time emerges entirely from cascade-state traversal") doesn't survive the dispersion-shape test: the cascade's natural traversal-dispersion is circular, not hyperbolic. Something additional — the Wick-rotation operation, or its discrete equivalent — is needed to convert circular shadow to Lorentzian shadow.

Per [[user_stance_fractal_shadow]]: integer-cyclic upstream / continuous downstream. The directed cyclic-shift eigenvalues ARE the upstream integer-cyclic primitive content; the circular dispersion downstream IS their projection-shadow. What's separately needed is the operation that further projects circular-shadow → hyperbolic-shadow, which is the Wick-rotation step.

Per [[user_stance_pi_as_projection]]: pi enters via the discrete-to-continuous projection. The cyclic-shift eigenvalue circle IS the integer-tooth-count projection on the unit circle — that's where pi appears. The Wick rotation is then a SEPARATE projection from the unit circle to the Lorentzian hyperbola via cos → cosh.

§4 What this means for bonus 8's verdict¶

Bonus 8 verdict status: PARTIALLY OVERTURNED, PARTIALLY CONFIRMED.

Partially overturned: bonus 8's strict claim that NO composition of A–N supplies any signed-metric content was too strong. Class C orientation on Class I cyclic groups produces a non-Hermitian operator with complex eigenvalues — which IS signed-metric content (asymmetric, off-real, paired). The bonus 8 probe missed this because it used UNDIRECTED Laplacians throughout (symmetric tridiagonal cyclic with periodic wrap), which are necessarily PSD. The choice of undirected operator was an implicit framing choice, not a forced one.
Partially confirmed: bonus 8's named gap (signed-metric composition / Wick rotation primitive) remains REAL but its role is refined. It is NOT the entire signed-metric production (Class C already supplies the asymmetric orientation); it IS the specific circle-to-hyperbola map that converts the cascade's natural CIRCULAR dispersion to the LORENTZIAN HYPERBOLIC dispersion physics observes.

The refined Class O candidate:

Class O — Wick-rotation primitive (refined). Operation that maps a circular dispersion (Im² = 2·Re − Re², natural product of Class C-oriented cyclic cascades) to a hyperbolic dispersion (Im² = Re + m², Klein-Gordon Lorentzian). Algebraically: cos(θ) → cosh(θ) projection, equivalent to multiplication of the dispersion variable by i. Not the entire production of signed-metric content (that's Class C); the specific signature-conversion projection that produces Lorentzian rather than Euclidean dispersion.

This is the same algebraic operation bonus 8 named, but with the structural understanding that the directed cascade ALREADY produces asymmetric / complex / oriented content via Class C; Class O is the FURTHER projection that converts circular to hyperbolic dispersion.

§5 What this means for `[[project_space_gauge_time_framework]]`¶

The 11D = 3D_s + 7D_g + 1D_t decomposition needs refinement, not rejection:

The 10D 3D_s + 7D_g substrate is well-defined and the directed cascade on it is structurally rich.
The "11^th dimension" carries traversal-parameter content (the τ of Class C iteration), which is part of the cascade-state graph's directed structure, not a separate cascade factor.
AND carries signature-conversion content (the Wick-rotation Class O candidate), which is a primitive operation, not a cascade factor.

Proposed reframing: 11D = 3D_s + 7D_g (10D cascade-state substrate) + τ (traversal parameter, Class C content) + Class O (signature-conversion primitive). The "+1D" of bonus 5's original framing is not simply "+1D_t"; it splits into a traversal parameter (already in the 14-class vocabulary as Class C) and a signature-conversion operation (Class O candidate).

This is HONEST to both bonus 8 (Class O is real) and bonus 9 (the cascade-traversal mechanism is also real and separate from Class O). The user's intuition that "time really gets its own dimension" was partially right (Class O is a separable primitive operation that introduces the Lorentzian signature) and partially wrong (the temporal cascade factor C_64 is not load-bearing; Class C orientation on the 10D cascade-state graph already supplies the traversal).

§6 What this means for MFO §VIII.8 (the bonus 8 landing)¶

The MFO §VIII.8 entry (the bonus 8 closure-test finding) needs an annotation rather than a rewrite:

The closure-test verdict (FAILURE; Class O located) stands as the strict result for cascade-composition with UNDIRECTED Laplacians.
A new MFO §VIII.9 should land the bonus 9 refinement: the directed-cascade test shows complex-eigenvalue structure is naturally present via Class C orientation, but the specific Lorentzian dispersion shape requires Class O (or equivalent circle-to-hyperbola map).
The reframed central computation §XIII.1 now becomes: "Find the directed cascade composition of cyclic groups in the 10D = 3D_s + 7D_g substrate whose Klein-Gordon-projected spectrum (via Class O Wick rotation) matches SM mass² ratios." The Class O is necessary; the directed cascade is the substrate it acts on.

§7 Discipline guards honoured¶

Spectral-graph falsifier (per [[feedback_antiquity_not_greek]]): all three measurements (complex-eigenvalue count, dispersion shape via linear / quadratic / Pade fit, mass-quantisation histogram) are Class L spectral-graph operations on directed cascade Laplacians.
Per [[user_stance_fractal_shadow]]: the circular-dispersion-as-projection-shadow framing is preserved. Cyclic groups upstream; their projection produces unit-circle eigenvalues; the cascade product produces a Minkowski-sum-of-circles dispersion shape.
Per [[feedback_trauma_informed_defensive_scope]]: structural / methodological only.
Per [[user_stance_string_theory_instrument_first]]: 10D / 11D appear via MFO §III.5 Witten 1981 7D minimum-isometry and the user's compression-expression duality, not via imported string-theoretic constructions.
Per [[feedback_no_lineage_claims_in_notebook]]: technical citations only (Klein-Gordon dispersion per MFO Part II.1, Witten 1981 7D minimum-isometry, directed-graph Laplacian theory). No lineage claims about external researchers.
Per [[feedback_ndjson_over_bloated_json]]: 18 + 11 + 8 = 37 NDJSON records across three probes, one per line.
Per [[feedback_pdf_extraction_citation_discipline]]: primary references only. MFO Part II.1 KG equations inherited from bonus 5. Bonus 8 reference inherited from this branch's prior synthesis.
stdlib + numpy + scipy only. CPU substrate per user direction.
NO new primitive class invented unilaterally. Class O candidate is refined per bonus 9; whether to update its description in [[project_class_o_signed_metric_composition]] is a conductor decision.
Antiquity-not-Greek language. "Primitive classes" not "primitives" for canonical A–N references.

§8 Three fermatas for the conductor¶

Update Class O memory? The [[project_class_o_signed_metric_composition]] memory's description currently reads as if Class O is the entire missing signed-metric content. Bonus 9 refines: Class C supplies the asymmetric/oriented content; Class O is the specific circle-to-hyperbola projection. Should the memory be updated, or kept as bonus 8's original framing with the bonus 9 refinement landed only in this synthesis?
Soften [[project_space_gauge_time_framework]]? The 11D = 3D_s + 7D_g + 1D_t framing is partially correct but understates the role of Class C orientation. Proposed refinement: 10D cascade substrate + Class C traversal + Class O signature-conversion. Not a full rewrite; a structural annotation.
Counterpoint dispatch? The directed-cascade probe found H3 because the dispersion shape is circular, not hyperbolic. A counterpoint subagent could test alternative compositions (e.g., Class K modulus-angle projection of the directed eigenvalues followed by Class N rational approximation of the angle distribution) that might supply the circle-to-hyperbola map without invoking Class O. The fermata is whether to dispatch this verification or take the H3 verdict as standing.

§9 The one surprise¶

The single-factor cyclic-shift directed Laplacian's eigenvalues satisfy the circle identity Im² = 2·Re − Re² to machine precision across all tested n. This is structurally beautiful: the algebraic shadow of Class C orientation on a Class I cyclic group is the unit circle in the complex plane, and the directed Laplacian sweeps that circle into a translated circle centered at (1/r², 0). The cascade composition (Class E direct-product sum) deforms this individual-factor circle into a Minkowski sum of circles — which is broadly circular but with mass distributed over a 2D wedge in the (Re, Im) plane, NOT a hyperbola.

The H1 hypothesis would have required this dispersion to be hyperbolic. Bonus 9's data shows it is not: KG R² ≈ 0.03 across all scales; no mass-quantisation; Pade approximation insufficient. The pattern held: Klein-Gordon dispersion is NOT a natural consequence of cascade composition + Class C orientation alone. Class O (or an equivalent circle-to-hyperbola map) is needed.

The deeper surprise is that bonus 8's finding was both right and wrong: right that A–N cannot directly produce Lorentzian dispersion; wrong that A–N cannot produce ANY signed-metric content (Class C orientation supplies asymmetric complex-paired eigenvalues, which IS signed-metric content of a kind, just not the specific Lorentzian kind). The 14-class vocabulary is RICHER than bonus 8's undirected-only audit suggested, and the gap is more specific (a circle-to-hyperbola map) than signed-metric in general.

§10 Closing one-sentence verdict¶

The Spike #24 14-class vocabulary A–N is empirically partial-rich for cascade-composition signed-metric structure (Class C orientation on Class L on Class I produces non-Hermitian, complex-paired, asymmetric operators — a circular dispersion), AND empirically incomplete for cascade-composition Lorentzian-KG-dispersion (the circle-to-hyperbola map is genuinely outside A–N), refining bonus 8's verdict by locating the gap more precisely (Class O is the specific Wick-rotation / cos→cosh / circle-to-hyperbola map, not the entire production of signed-metric content), with H3 winning over both H1 (time-is-fully-emergent vocabulary closed) and H2 (time-is-its-own dimension with Class O as the entire gap).

The math doesn't lie. Class C orientation on the 10D cascade-state graph produces real, asymmetric, complex-paired eigenvalues. The dispersion shape this produces is circular. Klein-Gordon dispersion is hyperbolic. A circle-to-hyperbola primitive is required, and it is structurally the Class O candidate from bonus 8 with refined content.